diff options
Diffstat (limited to 'drivers/gpu/drm/amd')
473 files changed, 45711 insertions, 8151 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/Kconfig b/drivers/gpu/drm/amd/amdgpu/Kconfig index 12adca8c7819..b91e79c721e2 100644 --- a/drivers/gpu/drm/amd/amdgpu/Kconfig +++ b/drivers/gpu/drm/amd/amdgpu/Kconfig @@ -69,6 +69,16 @@ config DRM_AMDGPU_USERPTR This option selects CONFIG_HMM and CONFIG_HMM_MIRROR if it isn't already selected to enabled full userptr support. +config DRM_AMDGPU_WERROR + bool "Force the compiler to throw an error instead of a warning when compiling" + depends on DRM_AMDGPU + depends on EXPERT + depends on !COMPILE_TEST + default n + help + Add -Werror to the build flags for amdgpu.ko. + Only enable this if you are warning code for amdgpu.ko. + source "drivers/gpu/drm/amd/acp/Kconfig" source "drivers/gpu/drm/amd/display/Kconfig" source "drivers/gpu/drm/amd/amdkfd/Kconfig" diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile index 415a7fa395c4..8d16f280b695 100644 --- a/drivers/gpu/drm/amd/amdgpu/Makefile +++ b/drivers/gpu/drm/amd/amdgpu/Makefile @@ -39,6 +39,26 @@ ccflags-y := -I$(FULL_AMD_PATH)/include/asic_reg \ -I$(FULL_AMD_DISPLAY_PATH)/amdgpu_dm \ -I$(FULL_AMD_PATH)/amdkfd +subdir-ccflags-y := -Wextra +subdir-ccflags-y += -Wunused +subdir-ccflags-y += -Wmissing-prototypes +subdir-ccflags-y += -Wmissing-declarations +subdir-ccflags-y += -Wmissing-include-dirs +subdir-ccflags-y += -Wold-style-definition +subdir-ccflags-y += -Wmissing-format-attribute +# Need this to avoid recursive variable evaluation issues +cond-flags := $(call cc-option, -Wunused-but-set-variable) \ + $(call cc-option, -Wunused-const-variable) \ + $(call cc-option, -Wstringop-truncation) \ + $(call cc-option, -Wpacked-not-aligned) +subdir-ccflags-y += $(cond-flags) +subdir-ccflags-y += -Wno-unused-parameter +subdir-ccflags-y += -Wno-type-limits +subdir-ccflags-y += -Wno-sign-compare +subdir-ccflags-y += -Wno-missing-field-initializers +subdir-ccflags-y += -Wno-override-init +subdir-ccflags-$(CONFIG_DRM_AMDGPU_WERROR) += -Werror + amdgpu-y := amdgpu_drv.o # add KMS driver @@ -60,7 +80,7 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.o \ amdgpu_umc.o smu_v11_0_i2c.o amdgpu_fru_eeprom.o amdgpu_rap.o \ amdgpu_fw_attestation.o amdgpu_securedisplay.o \ amdgpu_eeprom.o amdgpu_mca.o amdgpu_psp_ta.o amdgpu_lsdma.o \ - amdgpu_ring_mux.o + amdgpu_ring_mux.o amdgpu_xcp.o amdgpu-$(CONFIG_PROC_FS) += amdgpu_fdinfo.o @@ -78,7 +98,7 @@ amdgpu-y += \ vega20_reg_init.o nbio_v7_4.o nbio_v2_3.o nv.o arct_reg_init.o mxgpu_nv.o \ nbio_v7_2.o hdp_v4_0.o hdp_v5_0.o aldebaran_reg_init.o aldebaran.o soc21.o \ sienna_cichlid.o smu_v13_0_10.o nbio_v4_3.o hdp_v6_0.o nbio_v7_7.o hdp_v5_2.o lsdma_v6_0.o \ - nbio_v7_9.o + nbio_v7_9.o aqua_vanjaram_reg_init.o # add DF block amdgpu-y += \ @@ -183,12 +203,14 @@ amdgpu-y += \ vcn_v2_5.o \ vcn_v3_0.o \ vcn_v4_0.o \ + vcn_v4_0_3.o \ amdgpu_jpeg.o \ jpeg_v1_0.o \ jpeg_v2_0.o \ jpeg_v2_5.o \ jpeg_v3_0.o \ - jpeg_v4_0.o + jpeg_v4_0.o \ + jpeg_v4_0_3.o # add ATHUB block amdgpu-y += \ @@ -203,6 +225,7 @@ amdgpu-y += \ smuio_v11_0.o \ smuio_v11_0_6.o \ smuio_v13_0.o \ + smuio_v13_0_3.o \ smuio_v13_0_6.o # add reset block @@ -228,6 +251,7 @@ amdgpu-y += \ amdgpu_amdkfd_gfx_v9.o \ amdgpu_amdkfd_arcturus.o \ amdgpu_amdkfd_aldebaran.o \ + amdgpu_amdkfd_gc_9_4_3.o \ amdgpu_amdkfd_gfx_v10.o \ amdgpu_amdkfd_gfx_v10_3.o \ amdgpu_amdkfd_gfx_v11.o diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 02b827785e39..a84bd4a0c421 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -107,8 +107,9 @@ #include "amdgpu_fdinfo.h" #include "amdgpu_mca.h" #include "amdgpu_ras.h" +#include "amdgpu_xcp.h" -#define MAX_GPU_INSTANCE 16 +#define MAX_GPU_INSTANCE 64 struct amdgpu_gpu_instance { @@ -212,6 +213,8 @@ extern int amdgpu_noretry; extern int amdgpu_force_asic_type; extern int amdgpu_smartshift_bias; extern int amdgpu_use_xgmi_p2p; +extern int amdgpu_mtype_local; +extern bool enforce_isolation; #ifdef CONFIG_HSA_AMD extern int sched_policy; extern bool debug_evictions; @@ -242,9 +245,10 @@ extern int amdgpu_num_kcq; extern int amdgpu_vcnfw_log; extern int amdgpu_sg_display; +extern int amdgpu_user_partt_mode; + #define AMDGPU_VM_MAX_NUM_CTX 4096 #define AMDGPU_SG_THRESHOLD (256*1024*1024) -#define AMDGPU_DEFAULT_GTT_SIZE_MB 3072ULL /* 3GB by default */ #define AMDGPU_WAIT_IDLE_TIMEOUT_IN_MS 3000 #define AMDGPU_MAX_USEC_TIMEOUT 100000 /* 100 ms */ #define AMDGPU_FENCE_JIFFIES_TIMEOUT (HZ / 2) @@ -282,6 +286,7 @@ extern int amdgpu_sg_display; #define AMDGPU_SMARTSHIFT_MAX_BIAS (100) #define AMDGPU_SMARTSHIFT_MIN_BIAS (-100) +struct amdgpu_xcp_mgr; struct amdgpu_device; struct amdgpu_irq_src; struct amdgpu_fpriv; @@ -463,6 +468,8 @@ struct amdgpu_fpriv { struct mutex bo_list_lock; struct idr bo_list_handles; struct amdgpu_ctx_mgr ctx_mgr; + /** GPU partition selection */ + uint32_t xcp_id; }; int amdgpu_file_to_fpriv(struct file *filp, struct amdgpu_fpriv **fpriv); @@ -573,6 +580,8 @@ struct amdgpu_asic_funcs { /* query video codecs */ int (*query_video_codecs)(struct amdgpu_device *adev, bool encode, const struct amdgpu_video_codecs **codecs); + /* encode "> 32bits" smn addressing */ + u64 (*encode_ext_smn_addressing)(int ext_id); }; /* @@ -607,6 +616,9 @@ void amdgpu_cgs_destroy_device(struct cgs_device *cgs_device); typedef uint32_t (*amdgpu_rreg_t)(struct amdgpu_device*, uint32_t); typedef void (*amdgpu_wreg_t)(struct amdgpu_device*, uint32_t, uint32_t); +typedef uint32_t (*amdgpu_rreg_ext_t)(struct amdgpu_device*, uint64_t); +typedef void (*amdgpu_wreg_ext_t)(struct amdgpu_device*, uint64_t, uint32_t); + typedef uint64_t (*amdgpu_rreg64_t)(struct amdgpu_device*, uint32_t); typedef void (*amdgpu_wreg64_t)(struct amdgpu_device*, uint32_t, uint64_t); @@ -657,7 +669,7 @@ enum amd_hw_ip_block_type { MAX_HWIP }; -#define HWIP_MAX_INSTANCE 28 +#define HWIP_MAX_INSTANCE 44 #define HW_ID_MAX 300 #define IP_VERSION(mj, mn, rv) (((mj) << 16) | ((mn) << 8) | (rv)) @@ -665,6 +677,17 @@ enum amd_hw_ip_block_type { #define IP_VERSION_MIN(ver) (((ver) >> 8) & 0xFF) #define IP_VERSION_REV(ver) ((ver) & 0xFF) +struct amdgpu_ip_map_info { + /* Map of logical to actual dev instances/mask */ + uint32_t dev_inst[MAX_HWIP][HWIP_MAX_INSTANCE]; + int8_t (*logical_to_dev_inst)(struct amdgpu_device *adev, + enum amd_hw_ip_block_type block, + int8_t inst); + uint32_t (*logical_to_dev_mask)(struct amdgpu_device *adev, + enum amd_hw_ip_block_type block, + uint32_t mask); +}; + struct amd_powerplay { void *pp_handle; const struct amd_pm_funcs *pp_funcs; @@ -750,6 +773,7 @@ struct amdgpu_device { struct amdgpu_acp acp; #endif struct amdgpu_hive_info *hive; + struct amdgpu_xcp_mgr *xcp_mgr; /* ASIC */ enum amd_asic_type asic_type; uint32_t family; @@ -797,6 +821,8 @@ struct amdgpu_device { amdgpu_wreg_t pcie_wreg; amdgpu_rreg_t pciep_rreg; amdgpu_wreg_t pciep_wreg; + amdgpu_rreg_ext_t pcie_rreg_ext; + amdgpu_wreg_ext_t pcie_wreg_ext; amdgpu_rreg64_t pcie_rreg64; amdgpu_wreg64_t pcie_wreg64; /* protects concurrent UVD register access */ @@ -830,7 +856,7 @@ struct amdgpu_device { dma_addr_t dummy_page_addr; struct amdgpu_vm_manager vm_manager; struct amdgpu_vmhub vmhub[AMDGPU_MAX_VMHUBS]; - unsigned num_vmhubs; + DECLARE_BITMAP(vmhubs_mask, AMDGPU_MAX_VMHUBS); /* memory management */ struct amdgpu_mman mman; @@ -962,6 +988,7 @@ struct amdgpu_device { /* soc15 register offset based on ip, instance and segment */ uint32_t *reg_offset[MAX_HWIP][HWIP_MAX_INSTANCE]; + struct amdgpu_ip_map_info ip_map; /* delayed work_func for deferring clockgating during resume */ struct delayed_work delayed_init_work; @@ -1020,6 +1047,9 @@ struct amdgpu_device { struct pci_saved_state *pci_state; pci_channel_state_t pci_channel_state; + /* Track auto wait count on s_barrier settings */ + bool barrier_has_auto_waitcnt; + struct amdgpu_reset_control *reset_cntl; uint32_t ip_versions[MAX_HWIP][HWIP_MAX_INSTANCE]; @@ -1050,6 +1080,8 @@ struct amdgpu_device { bool job_hang; bool dc_enabled; + /* Mask of active clusters */ + uint32_t aid_mask; }; static inline struct amdgpu_device *drm_to_adev(struct drm_device *ddev) @@ -1081,11 +1113,18 @@ size_t amdgpu_device_aper_access(struct amdgpu_device *adev, loff_t pos, void amdgpu_device_vram_access(struct amdgpu_device *adev, loff_t pos, void *buf, size_t size, bool write); +uint32_t amdgpu_device_wait_on_rreg(struct amdgpu_device *adev, + uint32_t inst, uint32_t reg_addr, char reg_name[], + uint32_t expected_value, uint32_t mask); uint32_t amdgpu_device_rreg(struct amdgpu_device *adev, uint32_t reg, uint32_t acc_flags); +u32 amdgpu_device_indirect_rreg_ext(struct amdgpu_device *adev, + u64 reg_addr); void amdgpu_device_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v, uint32_t acc_flags); +void amdgpu_device_indirect_wreg_ext(struct amdgpu_device *adev, + u64 reg_addr, u32 reg_data); void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev, uint32_t reg, uint32_t v); void amdgpu_mm_wreg8(struct amdgpu_device *adev, uint32_t offset, uint8_t value); @@ -1137,6 +1176,8 @@ int emu_soc_asic_init(struct amdgpu_device *adev); #define WREG32_PCIE(reg, v) adev->pcie_wreg(adev, (reg), (v)) #define RREG32_PCIE_PORT(reg) adev->pciep_rreg(adev, (reg)) #define WREG32_PCIE_PORT(reg, v) adev->pciep_wreg(adev, (reg), (v)) +#define RREG32_PCIE_EXT(reg) adev->pcie_rreg_ext(adev, (reg)) +#define WREG32_PCIE_EXT(reg, v) adev->pcie_wreg_ext(adev, (reg), (v)) #define RREG64_PCIE(reg) adev->pcie_rreg64(adev, (reg)) #define WREG64_PCIE(reg, v) adev->pcie_wreg64(adev, (reg), (v)) #define RREG32_SMC(reg) adev->smc_rreg(adev, (reg)) @@ -1204,7 +1245,8 @@ int emu_soc_asic_init(struct amdgpu_device *adev); /* * ASICs macro. */ -#define amdgpu_asic_set_vga_state(adev, state) (adev)->asic_funcs->set_vga_state((adev), (state)) +#define amdgpu_asic_set_vga_state(adev, state) \ + ((adev)->asic_funcs->set_vga_state ? (adev)->asic_funcs->set_vga_state((adev), (state)) : 0) #define amdgpu_asic_reset(adev) (adev)->asic_funcs->reset((adev)) #define amdgpu_asic_reset_method(adev) (adev)->asic_funcs->reset_method((adev)) #define amdgpu_asic_get_xclk(adev) (adev)->asic_funcs->get_xclk((adev)) @@ -1235,6 +1277,10 @@ int emu_soc_asic_init(struct amdgpu_device *adev); #define amdgpu_inc_vram_lost(adev) atomic_inc(&((adev)->vram_lost_counter)); +#define for_each_inst(i, inst_mask) \ + for (i = ffs(inst_mask) - 1; inst_mask; \ + inst_mask &= ~(1U << i), i = ffs(inst_mask) - 1) + #define MIN(X, Y) ((X) < (Y) ? (X) : (Y)) /* Common functions */ @@ -1348,6 +1394,12 @@ struct amdgpu_afmt_acr amdgpu_afmt_acr(uint32_t clock); /* amdgpu_acpi.c */ +struct amdgpu_numa_info { + uint64_t size; + int pxm; + int nid; +}; + /* ATCS Device/Driver State */ #define AMDGPU_ATCS_PSC_DEV_STATE_D0 0 #define AMDGPU_ATCS_PSC_DEV_STATE_D3_HOT 3 @@ -1365,15 +1417,32 @@ int amdgpu_acpi_power_shift_control(struct amdgpu_device *adev, u8 dev_state, bool drv_state); int amdgpu_acpi_smart_shift_update(struct drm_device *dev, enum amdgpu_ss ss_state); int amdgpu_acpi_pcie_notify_device_ready(struct amdgpu_device *adev); +int amdgpu_acpi_get_tmr_info(struct amdgpu_device *adev, u64 *tmr_offset, + u64 *tmr_size); +int amdgpu_acpi_get_mem_info(struct amdgpu_device *adev, int xcc_id, + struct amdgpu_numa_info *numa_info); void amdgpu_acpi_get_backlight_caps(struct amdgpu_dm_backlight_caps *caps); bool amdgpu_acpi_should_gpu_reset(struct amdgpu_device *adev); void amdgpu_acpi_detect(void); +void amdgpu_acpi_release(void); #else static inline int amdgpu_acpi_init(struct amdgpu_device *adev) { return 0; } +static inline int amdgpu_acpi_get_tmr_info(struct amdgpu_device *adev, + u64 *tmr_offset, u64 *tmr_size) +{ + return -EINVAL; +} +static inline int amdgpu_acpi_get_mem_info(struct amdgpu_device *adev, + int xcc_id, + struct amdgpu_numa_info *numa_info) +{ + return -EINVAL; +} static inline void amdgpu_acpi_fini(struct amdgpu_device *adev) { } static inline bool amdgpu_acpi_should_gpu_reset(struct amdgpu_device *adev) { return false; } static inline void amdgpu_acpi_detect(void) { } +static inline void amdgpu_acpi_release(void) { } static inline bool amdgpu_acpi_is_power_shift_control_supported(void) { return false; } static inline int amdgpu_acpi_power_shift_control(struct amdgpu_device *adev, u8 dev_state, bool drv_state) { return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c index aeeec211861c..385c6acb5728 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c @@ -26,6 +26,7 @@ #include <linux/acpi.h> #include <linux/backlight.h> #include <linux/slab.h> +#include <linux/xarray.h> #include <linux/power_supply.h> #include <linux/pm_runtime.h> #include <linux/suspend.h> @@ -38,6 +39,45 @@ #include "amd_acpi.h" #include "atom.h" +/* Declare GUID for AMD _DSM method for XCCs */ +static const guid_t amd_xcc_dsm_guid = GUID_INIT(0x8267f5d5, 0xa556, 0x44f2, + 0xb8, 0xb4, 0x45, 0x56, 0x2e, + 0x8c, 0x5b, 0xec); + +#define AMD_XCC_HID_START 3000 +#define AMD_XCC_DSM_GET_NUM_FUNCS 0 +#define AMD_XCC_DSM_GET_SUPP_MODE 1 +#define AMD_XCC_DSM_GET_XCP_MODE 2 +#define AMD_XCC_DSM_GET_VF_XCC_MAPPING 4 +#define AMD_XCC_DSM_GET_TMR_INFO 5 +#define AMD_XCC_DSM_NUM_FUNCS 5 + +#define AMD_XCC_MAX_HID 24 + +struct xarray numa_info_xa; + +/* Encapsulates the XCD acpi object information */ +struct amdgpu_acpi_xcc_info { + struct list_head list; + struct amdgpu_numa_info *numa_info; + uint8_t xcp_node; + uint8_t phy_id; + acpi_handle handle; +}; + +struct amdgpu_acpi_dev_info { + struct list_head list; + struct list_head xcc_list; + uint16_t bdf; + uint16_t supp_xcp_mode; + uint16_t xcp_mode; + uint16_t mem_mode; + uint64_t tmr_base; + uint64_t tmr_size; +}; + +struct list_head amdgpu_acpi_dev_list; + struct amdgpu_atif_notification_cfg { bool enabled; int command_code; @@ -801,6 +841,343 @@ int amdgpu_acpi_smart_shift_update(struct drm_device *dev, enum amdgpu_ss ss_sta return r; } +#ifdef CONFIG_ACPI_NUMA +static inline uint64_t amdgpu_acpi_get_numa_size(int nid) +{ + /* This is directly using si_meminfo_node implementation as the + * function is not exported. + */ + int zone_type; + uint64_t managed_pages = 0; + + pg_data_t *pgdat = NODE_DATA(nid); + + for (zone_type = 0; zone_type < MAX_NR_ZONES; zone_type++) + managed_pages += + zone_managed_pages(&pgdat->node_zones[zone_type]); + return managed_pages * PAGE_SIZE; +} + +static struct amdgpu_numa_info *amdgpu_acpi_get_numa_info(uint32_t pxm) +{ + struct amdgpu_numa_info *numa_info; + int nid; + + numa_info = xa_load(&numa_info_xa, pxm); + + if (!numa_info) { + struct sysinfo info; + + numa_info = kzalloc(sizeof *numa_info, GFP_KERNEL); + if (!numa_info) + return NULL; + + nid = pxm_to_node(pxm); + numa_info->pxm = pxm; + numa_info->nid = nid; + + if (numa_info->nid == NUMA_NO_NODE) { + si_meminfo(&info); + numa_info->size = info.totalram * info.mem_unit; + } else { + numa_info->size = amdgpu_acpi_get_numa_size(nid); + } + xa_store(&numa_info_xa, numa_info->pxm, numa_info, GFP_KERNEL); + } + + return numa_info; +} +#endif + +/** + * amdgpu_acpi_get_node_id - obtain the NUMA node id for corresponding amdgpu + * acpi device handle + * + * @handle: acpi handle + * @numa_info: amdgpu_numa_info structure holding numa information + * + * Queries the ACPI interface to fetch the corresponding NUMA Node ID for a + * given amdgpu acpi device. + * + * Returns ACPI STATUS OK with Node ID on success or the corresponding failure reason + */ +static acpi_status amdgpu_acpi_get_node_id(acpi_handle handle, + struct amdgpu_numa_info **numa_info) +{ +#ifdef CONFIG_ACPI_NUMA + u64 pxm; + acpi_status status; + + if (!numa_info) + return_ACPI_STATUS(AE_ERROR); + + status = acpi_evaluate_integer(handle, "_PXM", NULL, &pxm); + + if (ACPI_FAILURE(status)) + return status; + + *numa_info = amdgpu_acpi_get_numa_info(pxm); + + if (!*numa_info) + return_ACPI_STATUS(AE_ERROR); + + return_ACPI_STATUS(AE_OK); +#else + return_ACPI_STATUS(AE_NOT_EXIST); +#endif +} + +static struct amdgpu_acpi_dev_info *amdgpu_acpi_get_dev(u16 bdf) +{ + struct amdgpu_acpi_dev_info *acpi_dev; + + if (list_empty(&amdgpu_acpi_dev_list)) + return NULL; + + list_for_each_entry(acpi_dev, &amdgpu_acpi_dev_list, list) + if (acpi_dev->bdf == bdf) + return acpi_dev; + + return NULL; +} + +static int amdgpu_acpi_dev_init(struct amdgpu_acpi_dev_info **dev_info, + struct amdgpu_acpi_xcc_info *xcc_info, u16 bdf) +{ + struct amdgpu_acpi_dev_info *tmp; + union acpi_object *obj; + int ret = -ENOENT; + + *dev_info = NULL; + tmp = kzalloc(sizeof(struct amdgpu_acpi_dev_info), GFP_KERNEL); + if (!tmp) + return -ENOMEM; + + INIT_LIST_HEAD(&tmp->xcc_list); + INIT_LIST_HEAD(&tmp->list); + tmp->bdf = bdf; + + obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0, + AMD_XCC_DSM_GET_SUPP_MODE, NULL, + ACPI_TYPE_INTEGER); + + if (!obj) { + acpi_handle_debug(xcc_info->handle, + "_DSM function %d evaluation failed", + AMD_XCC_DSM_GET_SUPP_MODE); + ret = -ENOENT; + goto out; + } + + tmp->supp_xcp_mode = obj->integer.value & 0xFFFF; + ACPI_FREE(obj); + + obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0, + AMD_XCC_DSM_GET_XCP_MODE, NULL, + ACPI_TYPE_INTEGER); + + if (!obj) { + acpi_handle_debug(xcc_info->handle, + "_DSM function %d evaluation failed", + AMD_XCC_DSM_GET_XCP_MODE); + ret = -ENOENT; + goto out; + } + + tmp->xcp_mode = obj->integer.value & 0xFFFF; + tmp->mem_mode = (obj->integer.value >> 32) & 0xFFFF; + ACPI_FREE(obj); + + /* Evaluate DSMs and fill XCC information */ + obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0, + AMD_XCC_DSM_GET_TMR_INFO, NULL, + ACPI_TYPE_PACKAGE); + + if (!obj || obj->package.count < 2) { + acpi_handle_debug(xcc_info->handle, + "_DSM function %d evaluation failed", + AMD_XCC_DSM_GET_TMR_INFO); + ret = -ENOENT; + goto out; + } + + tmp->tmr_base = obj->package.elements[0].integer.value; + tmp->tmr_size = obj->package.elements[1].integer.value; + ACPI_FREE(obj); + + DRM_DEBUG_DRIVER( + "New dev(%x): Supported xcp mode: %x curr xcp_mode : %x mem mode : %x, tmr base: %llx tmr size: %llx ", + tmp->bdf, tmp->supp_xcp_mode, tmp->xcp_mode, tmp->mem_mode, + tmp->tmr_base, tmp->tmr_size); + list_add_tail(&tmp->list, &amdgpu_acpi_dev_list); + *dev_info = tmp; + + return 0; + +out: + if (obj) + ACPI_FREE(obj); + kfree(tmp); + + return ret; +} + +static int amdgpu_acpi_get_xcc_info(struct amdgpu_acpi_xcc_info *xcc_info, + u16 *bdf) +{ + union acpi_object *obj; + acpi_status status; + int ret = -ENOENT; + + obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0, + AMD_XCC_DSM_GET_NUM_FUNCS, NULL, + ACPI_TYPE_INTEGER); + + if (!obj || obj->integer.value != AMD_XCC_DSM_NUM_FUNCS) + goto out; + ACPI_FREE(obj); + + /* Evaluate DSMs and fill XCC information */ + obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0, + AMD_XCC_DSM_GET_VF_XCC_MAPPING, NULL, + ACPI_TYPE_INTEGER); + + if (!obj) { + acpi_handle_debug(xcc_info->handle, + "_DSM function %d evaluation failed", + AMD_XCC_DSM_GET_VF_XCC_MAPPING); + ret = -EINVAL; + goto out; + } + + /* PF xcc id [39:32] */ + xcc_info->phy_id = (obj->integer.value >> 32) & 0xFF; + /* xcp node of this xcc [47:40] */ + xcc_info->xcp_node = (obj->integer.value >> 40) & 0xFF; + /* PF bus/dev/fn of this xcc [63:48] */ + *bdf = (obj->integer.value >> 48) & 0xFFFF; + ACPI_FREE(obj); + obj = NULL; + + status = + amdgpu_acpi_get_node_id(xcc_info->handle, &xcc_info->numa_info); + + /* TODO: check if this check is required */ + if (ACPI_SUCCESS(status)) + ret = 0; +out: + if (obj) + ACPI_FREE(obj); + + return ret; +} + +static int amdgpu_acpi_enumerate_xcc(void) +{ + struct amdgpu_acpi_dev_info *dev_info = NULL; + struct amdgpu_acpi_xcc_info *xcc_info; + struct acpi_device *acpi_dev; + char hid[ACPI_ID_LEN]; + int ret, id; + u16 bdf; + + INIT_LIST_HEAD(&amdgpu_acpi_dev_list); + xa_init(&numa_info_xa); + + for (id = 0; id < AMD_XCC_MAX_HID; id++) { + sprintf(hid, "%s%d", "AMD", AMD_XCC_HID_START + id); + acpi_dev = acpi_dev_get_first_match_dev(hid, NULL, -1); + /* These ACPI objects are expected to be in sequential order. If + * one is not found, no need to check the rest. + */ + if (!acpi_dev) { + DRM_DEBUG_DRIVER("No matching acpi device found for %s", + hid); + break; + } + + xcc_info = kzalloc(sizeof(struct amdgpu_acpi_xcc_info), + GFP_KERNEL); + if (!xcc_info) { + DRM_ERROR("Failed to allocate memory for xcc info\n"); + return -ENOMEM; + } + + INIT_LIST_HEAD(&xcc_info->list); + xcc_info->handle = acpi_device_handle(acpi_dev); + acpi_dev_put(acpi_dev); + + ret = amdgpu_acpi_get_xcc_info(xcc_info, &bdf); + if (ret) { + kfree(xcc_info); + continue; + } + + dev_info = amdgpu_acpi_get_dev(bdf); + + if (!dev_info) + ret = amdgpu_acpi_dev_init(&dev_info, xcc_info, bdf); + + if (ret == -ENOMEM) + return ret; + + if (!dev_info) { + kfree(xcc_info); + continue; + } + + list_add_tail(&xcc_info->list, &dev_info->xcc_list); + } + + return 0; +} + +int amdgpu_acpi_get_tmr_info(struct amdgpu_device *adev, u64 *tmr_offset, + u64 *tmr_size) +{ + struct amdgpu_acpi_dev_info *dev_info; + u16 bdf; + + if (!tmr_offset || !tmr_size) + return -EINVAL; + + bdf = (adev->pdev->bus->number << 8) | adev->pdev->devfn; + dev_info = amdgpu_acpi_get_dev(bdf); + if (!dev_info) + return -ENOENT; + + *tmr_offset = dev_info->tmr_base; + *tmr_size = dev_info->tmr_size; + + return 0; +} + +int amdgpu_acpi_get_mem_info(struct amdgpu_device *adev, int xcc_id, + struct amdgpu_numa_info *numa_info) +{ + struct amdgpu_acpi_dev_info *dev_info; + struct amdgpu_acpi_xcc_info *xcc_info; + u16 bdf; + + if (!numa_info) + return -EINVAL; + + bdf = (adev->pdev->bus->number << 8) | adev->pdev->devfn; + dev_info = amdgpu_acpi_get_dev(bdf); + if (!dev_info) + return -ENOENT; + + list_for_each_entry(xcc_info, &dev_info->xcc_list, list) { + if (xcc_info->phy_id == xcc_id) { + memcpy(numa_info, xcc_info->numa_info, + sizeof(*numa_info)); + return 0; + } + } + + return -ENOENT; +} + /** * amdgpu_acpi_event - handle notify events * @@ -1054,6 +1431,36 @@ void amdgpu_acpi_detect(void) } else { atif->backlight_caps.caps_valid = false; } + + amdgpu_acpi_enumerate_xcc(); +} + +void amdgpu_acpi_release(void) +{ + struct amdgpu_acpi_dev_info *dev_info, *dev_tmp; + struct amdgpu_acpi_xcc_info *xcc_info, *xcc_tmp; + struct amdgpu_numa_info *numa_info; + unsigned long index; + + xa_for_each(&numa_info_xa, index, numa_info) { + kfree(numa_info); + xa_erase(&numa_info_xa, index); + } + + if (list_empty(&amdgpu_acpi_dev_list)) + return; + + list_for_each_entry_safe(dev_info, dev_tmp, &amdgpu_acpi_dev_list, + list) { + list_for_each_entry_safe(xcc_info, xcc_tmp, &dev_info->xcc_list, + list) { + list_del(&xcc_info->list); + kfree(xcc_info); + } + + list_del(&dev_info->list); + kfree(dev_info); + } } #if IS_ENABLED(CONFIG_SUSPEND) @@ -1092,16 +1499,20 @@ bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev) * S0ix even though the system is suspending to idle, so return false * in that case. */ - if (!(acpi_gbl_FADT.flags & ACPI_FADT_LOW_POWER_S0)) - dev_warn_once(adev->dev, + if (!(acpi_gbl_FADT.flags & ACPI_FADT_LOW_POWER_S0)) { + dev_err_once(adev->dev, "Power consumption will be higher as BIOS has not been configured for suspend-to-idle.\n" "To use suspend-to-idle change the sleep mode in BIOS setup.\n"); + return false; + } #if !IS_ENABLED(CONFIG_AMD_PMC) - dev_warn_once(adev->dev, + dev_err_once(adev->dev, "Power consumption will be higher as the kernel has not been compiled with CONFIG_AMD_PMC.\n"); -#endif /* CONFIG_AMD_PMC */ + return false; +#else return true; +#endif /* CONFIG_AMD_PMC */ } #endif /* CONFIG_SUSPEND */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index 0385f7f69278..b4fcad0e62f7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -53,7 +53,6 @@ int amdgpu_amdkfd_init(void) amdgpu_amdkfd_total_mem_size *= si.mem_unit; ret = kgd2kfd_init(); - amdgpu_amdkfd_gpuvm_init_mem_limits(); kfd_initialized = !ret; return ret; @@ -143,6 +142,8 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) int i; int last_valid_bit; + amdgpu_amdkfd_gpuvm_init_mem_limits(); + if (adev->kfd.dev) { struct kgd2kfd_shared_resources gpu_resources = { .compute_vmid_bitmap = @@ -162,7 +163,7 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) * clear */ bitmap_complement(gpu_resources.cp_queue_bitmap, - adev->gfx.mec.queue_bitmap, + adev->gfx.mec_bitmap[0].queue_bitmap, KGD_MAX_QUEUES); /* According to linux/bitmap.h we shouldn't use bitmap_clear if @@ -427,14 +428,23 @@ uint32_t amdgpu_amdkfd_get_fw_version(struct amdgpu_device *adev, } void amdgpu_amdkfd_get_local_mem_info(struct amdgpu_device *adev, - struct kfd_local_mem_info *mem_info) + struct kfd_local_mem_info *mem_info, + struct amdgpu_xcp *xcp) { memset(mem_info, 0, sizeof(*mem_info)); - mem_info->local_mem_size_public = adev->gmc.visible_vram_size; - mem_info->local_mem_size_private = adev->gmc.real_vram_size - + if (xcp) { + if (adev->gmc.real_vram_size == adev->gmc.visible_vram_size) + mem_info->local_mem_size_public = + KFD_XCP_MEMORY_SIZE(adev, xcp->id); + else + mem_info->local_mem_size_private = + KFD_XCP_MEMORY_SIZE(adev, xcp->id); + } else { + mem_info->local_mem_size_public = adev->gmc.visible_vram_size; + mem_info->local_mem_size_private = adev->gmc.real_vram_size - adev->gmc.visible_vram_size; - + } mem_info->vram_width = adev->gmc.vram_width; pr_debug("Address base: %pap public 0x%llx private 0x%llx\n", @@ -497,7 +507,7 @@ int amdgpu_amdkfd_get_dmabuf_info(struct amdgpu_device *adev, int dma_buf_fd, struct amdgpu_device **dmabuf_adev, uint64_t *bo_size, void *metadata_buffer, size_t buffer_size, uint32_t *metadata_size, - uint32_t *flags) + uint32_t *flags, int8_t *xcp_id) { struct dma_buf *dma_buf; struct drm_gem_object *obj; @@ -541,6 +551,8 @@ int amdgpu_amdkfd_get_dmabuf_info(struct amdgpu_device *adev, int dma_buf_fd, if (bo->flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED) *flags |= KFD_IOC_ALLOC_MEM_FLAGS_PUBLIC; } + if (xcp_id) + *xcp_id = bo->xcp_id; out_put: dma_buf_put(dma_buf); @@ -732,17 +744,19 @@ int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct amdgpu_device *adev, if (adev->family == AMDGPU_FAMILY_AI) { int i; - for (i = 0; i < adev->num_vmhubs; i++) + for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) amdgpu_gmc_flush_gpu_tlb(adev, vmid, i, 0); } else { - amdgpu_gmc_flush_gpu_tlb(adev, vmid, AMDGPU_GFXHUB_0, 0); + amdgpu_gmc_flush_gpu_tlb(adev, vmid, AMDGPU_GFXHUB(0), 0); } return 0; } int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct amdgpu_device *adev, - uint16_t pasid, enum TLB_FLUSH_TYPE flush_type) + uint16_t pasid, + enum TLB_FLUSH_TYPE flush_type, + uint32_t inst) { bool all_hub = false; @@ -750,7 +764,7 @@ int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct amdgpu_device *adev, adev->family == AMDGPU_FAMILY_RV) all_hub = true; - return amdgpu_gmc_flush_gpu_tlb_pasid(adev, pasid, flush_type, all_hub); + return amdgpu_gmc_flush_gpu_tlb_pasid(adev, pasid, flush_type, all_hub, inst); } bool amdgpu_amdkfd_have_atomics_support(struct amdgpu_device *adev) @@ -758,11 +772,32 @@ bool amdgpu_amdkfd_have_atomics_support(struct amdgpu_device *adev) return adev->have_atomics_support; } +void amdgpu_amdkfd_debug_mem_fence(struct amdgpu_device *adev) +{ + amdgpu_device_flush_hdp(adev, NULL); +} + void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev, bool reset) { amdgpu_umc_poison_handler(adev, reset); } +int amdgpu_amdkfd_send_close_event_drain_irq(struct amdgpu_device *adev, + uint32_t *payload) +{ + int ret; + + /* Device or IH ring is not ready so bail. */ + ret = amdgpu_ih_wait_on_checkpoint_process_ts(adev, &adev->irq.ih); + if (ret) + return ret; + + /* Send payload to fence KFD interrupts */ + amdgpu_amdkfd_interrupt(adev, payload); + + return 0; +} + bool amdgpu_amdkfd_ras_query_utcl2_poison_status(struct amdgpu_device *adev) { if (adev->gfx.ras && adev->gfx.ras->query_utcl2_poison_status) @@ -770,3 +805,28 @@ bool amdgpu_amdkfd_ras_query_utcl2_poison_status(struct amdgpu_device *adev) else return false; } + +int amdgpu_amdkfd_check_and_lock_kfd(struct amdgpu_device *adev) +{ + return kgd2kfd_check_and_lock_kfd(); +} + +void amdgpu_amdkfd_unlock_kfd(struct amdgpu_device *adev) +{ + kgd2kfd_unlock_kfd(); +} + + +u64 amdgpu_amdkfd_xcp_memory_size(struct amdgpu_device *adev, int xcp_id) +{ + u64 tmp; + s8 mem_id = KFD_XCP_MEM_ID(adev, xcp_id); + + if (adev->gmc.num_mem_partitions && xcp_id >= 0 && mem_id >= 0) { + tmp = adev->gmc.mem_partitions[mem_id].size; + do_div(tmp, adev->xcp_mgr->num_xcp_per_mem_partition); + return ALIGN_DOWN(tmp, PAGE_SIZE); + } else { + return adev->gmc.real_vram_size; + } +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h index 01ba3589b60a..2d0406bff84e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h @@ -30,10 +30,12 @@ #include <linux/kthread.h> #include <linux/workqueue.h> #include <linux/mmu_notifier.h> +#include <linux/memremap.h> #include <kgd_kfd_interface.h> #include <drm/ttm/ttm_execbuf_util.h> #include "amdgpu_sync.h" #include "amdgpu_vm.h" +#include "amdgpu_xcp.h" extern uint64_t amdgpu_amdkfd_total_mem_size; @@ -97,10 +99,13 @@ struct amdgpu_amdkfd_fence { struct amdgpu_kfd_dev { struct kfd_dev *dev; - int64_t vram_used; - uint64_t vram_used_aligned; + int64_t vram_used[MAX_XCP]; + uint64_t vram_used_aligned[MAX_XCP]; bool init_complete; struct work_struct reset_work; + + /* HMM page migration MEMORY_DEVICE_PRIVATE mapping */ + struct dev_pagemap pgmap; }; enum kgd_engine_type { @@ -151,6 +156,8 @@ void amdgpu_amdkfd_interrupt(struct amdgpu_device *adev, void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev); void amdgpu_amdkfd_device_init(struct amdgpu_device *adev); void amdgpu_amdkfd_device_fini_sw(struct amdgpu_device *adev); +int amdgpu_amdkfd_check_and_lock_kfd(struct amdgpu_device *adev); +void amdgpu_amdkfd_unlock_kfd(struct amdgpu_device *adev); int amdgpu_amdkfd_submit_ib(struct amdgpu_device *adev, enum kgd_engine_type engine, uint32_t vmid, uint64_t gpu_addr, @@ -160,7 +167,8 @@ bool amdgpu_amdkfd_have_atomics_support(struct amdgpu_device *adev); int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct amdgpu_device *adev, uint16_t vmid); int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct amdgpu_device *adev, - uint16_t pasid, enum TLB_FLUSH_TYPE flush_type); + uint16_t pasid, enum TLB_FLUSH_TYPE flush_type, + uint32_t inst); bool amdgpu_amdkfd_is_kfd_vmid(struct amdgpu_device *adev, u32 vmid); @@ -224,7 +232,8 @@ int amdgpu_amdkfd_remove_gws_from_process(void *info, void *mem); uint32_t amdgpu_amdkfd_get_fw_version(struct amdgpu_device *adev, enum kgd_engine_type type); void amdgpu_amdkfd_get_local_mem_info(struct amdgpu_device *adev, - struct kfd_local_mem_info *mem_info); + struct kfd_local_mem_info *mem_info, + struct amdgpu_xcp *xcp); uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct amdgpu_device *adev); uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct amdgpu_device *adev); @@ -234,13 +243,15 @@ int amdgpu_amdkfd_get_dmabuf_info(struct amdgpu_device *adev, int dma_buf_fd, struct amdgpu_device **dmabuf_adev, uint64_t *bo_size, void *metadata_buffer, size_t buffer_size, uint32_t *metadata_size, - uint32_t *flags); + uint32_t *flags, int8_t *xcp_id); uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct amdgpu_device *dst, struct amdgpu_device *src); int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct amdgpu_device *dst, struct amdgpu_device *src, bool is_min); int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct amdgpu_device *adev, bool is_min); +int amdgpu_amdkfd_send_close_event_drain_irq(struct amdgpu_device *adev, + uint32_t *payload); /* Read user wptr from a specified user address space with page fault * disabled. The memory must be pinned and mapped to the hardware when @@ -279,7 +290,8 @@ int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev, void amdgpu_amdkfd_gpuvm_release_process_vm(struct amdgpu_device *adev, void *drm_priv); uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *drm_priv); -size_t amdgpu_amdkfd_get_available_memory(struct amdgpu_device *adev); +size_t amdgpu_amdkfd_get_available_memory(struct amdgpu_device *adev, + uint8_t xcp_id); int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( struct amdgpu_device *adev, uint64_t va, uint64_t size, void *drm_priv, struct kgd_mem **mem, @@ -310,6 +322,7 @@ int amdgpu_amdkfd_gpuvm_import_dmabuf(struct amdgpu_device *adev, uint64_t *mmap_offset); int amdgpu_amdkfd_gpuvm_export_dmabuf(struct kgd_mem *mem, struct dma_buf **dmabuf); +void amdgpu_amdkfd_debug_mem_fence(struct amdgpu_device *adev); int amdgpu_amdkfd_get_tile_config(struct amdgpu_device *adev, struct tile_config *config); void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev, @@ -319,9 +332,18 @@ void amdgpu_amdkfd_block_mmu_notifications(void *p); int amdgpu_amdkfd_criu_resume(void *p); bool amdgpu_amdkfd_ras_query_utcl2_poison_status(struct amdgpu_device *adev); int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, - uint64_t size, u32 alloc_flag); + uint64_t size, u32 alloc_flag, int8_t xcp_id); void amdgpu_amdkfd_unreserve_mem_limit(struct amdgpu_device *adev, - uint64_t size, u32 alloc_flag); + uint64_t size, u32 alloc_flag, int8_t xcp_id); + +u64 amdgpu_amdkfd_xcp_memory_size(struct amdgpu_device *adev, int xcp_id); + +#define KFD_XCP_MEM_ID(adev, xcp_id) \ + ((adev)->xcp_mgr && (xcp_id) >= 0 ?\ + (adev)->xcp_mgr->xcp[(xcp_id)].mem_id : -1) + +#define KFD_XCP_MEMORY_SIZE(adev, xcp_id) amdgpu_amdkfd_xcp_memory_size((adev), (xcp_id)) + #if IS_ENABLED(CONFIG_HSA_AMD) void amdgpu_amdkfd_gpuvm_init_mem_limits(void); @@ -352,6 +374,17 @@ void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo) { } #endif + +#if IS_ENABLED(CONFIG_HSA_AMD_SVM) +int kgd2kfd_init_zone_device(struct amdgpu_device *adev); +#else +static inline +int kgd2kfd_init_zone_device(struct amdgpu_device *adev) +{ + return 0; +} +#endif + /* KGD2KFD callbacks */ int kgd2kfd_quiesce_mm(struct mm_struct *mm, uint32_t trigger); int kgd2kfd_resume_mm(struct mm_struct *mm); @@ -372,6 +405,8 @@ int kgd2kfd_post_reset(struct kfd_dev *kfd); void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry); void kgd2kfd_set_sram_ecc_flag(struct kfd_dev *kfd); void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint64_t throttle_bitmask); +int kgd2kfd_check_and_lock_kfd(void); +void kgd2kfd_unlock_kfd(void); #else static inline int kgd2kfd_init(void) { @@ -437,5 +472,14 @@ static inline void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint64_t throttle_bitmask) { } + +static inline int kgd2kfd_check_and_lock_kfd(void) +{ + return 0; +} + +static inline void kgd2kfd_unlock_kfd(void) +{ +} #endif #endif /* AMDGPU_AMDKFD_H_INCLUDED */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c index 4485bb29bec9..60f9e027fb66 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c @@ -23,6 +23,149 @@ #include "amdgpu_amdkfd.h" #include "amdgpu_amdkfd_arcturus.h" #include "amdgpu_amdkfd_gfx_v9.h" +#include "gc/gc_9_4_2_offset.h" +#include "gc/gc_9_4_2_sh_mask.h" +#include <uapi/linux/kfd_ioctl.h> + +/* + * Returns TRAP_EN, EXCP_EN and EXCP_REPLACE. + * + * restore_dbg_registers is ignored here but is a general interface requirement + * for devices that support GFXOFF and where the RLC save/restore list + * does not support hw registers for debugging i.e. the driver has to manually + * initialize the debug mode registers after it has disabled GFX off during the + * debug session. + */ +static uint32_t kgd_aldebaran_enable_debug_trap(struct amdgpu_device *adev, + bool restore_dbg_registers, + uint32_t vmid) +{ + uint32_t data = 0; + + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1); + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0); + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, 0); + + return data; +} + +/* returns TRAP_EN, EXCP_EN and EXCP_REPLACE. */ +static uint32_t kgd_aldebaran_disable_debug_trap(struct amdgpu_device *adev, + bool keep_trap_enabled, + uint32_t vmid) +{ + uint32_t data = 0; + + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, keep_trap_enabled); + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0); + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, 0); + + return data; +} + +static int kgd_aldebaran_validate_trap_override_request(struct amdgpu_device *adev, + uint32_t trap_override, + uint32_t *trap_mask_supported) +{ + *trap_mask_supported &= KFD_DBG_TRAP_MASK_FP_INVALID | + KFD_DBG_TRAP_MASK_FP_INPUT_DENORMAL | + KFD_DBG_TRAP_MASK_FP_DIVIDE_BY_ZERO | + KFD_DBG_TRAP_MASK_FP_OVERFLOW | + KFD_DBG_TRAP_MASK_FP_UNDERFLOW | + KFD_DBG_TRAP_MASK_FP_INEXACT | + KFD_DBG_TRAP_MASK_INT_DIVIDE_BY_ZERO | + KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH | + KFD_DBG_TRAP_MASK_DBG_MEMORY_VIOLATION; + + if (trap_override != KFD_DBG_TRAP_OVERRIDE_OR && + trap_override != KFD_DBG_TRAP_OVERRIDE_REPLACE) + return -EPERM; + + return 0; +} + +/* returns TRAP_EN, EXCP_EN and EXCP_RPLACE. */ +static uint32_t kgd_aldebaran_set_wave_launch_trap_override(struct amdgpu_device *adev, + uint32_t vmid, + uint32_t trap_override, + uint32_t trap_mask_bits, + uint32_t trap_mask_request, + uint32_t *trap_mask_prev, + uint32_t kfd_dbg_trap_cntl_prev) + +{ + uint32_t data = 0; + + *trap_mask_prev = REG_GET_FIELD(kfd_dbg_trap_cntl_prev, SPI_GDBG_PER_VMID_CNTL, EXCP_EN); + trap_mask_bits = (trap_mask_bits & trap_mask_request) | + (*trap_mask_prev & ~trap_mask_request); + + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1); + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, trap_mask_bits); + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, trap_override); + + return data; +} + +static uint32_t kgd_aldebaran_set_wave_launch_mode(struct amdgpu_device *adev, + uint8_t wave_launch_mode, + uint32_t vmid) +{ + uint32_t data = 0; + + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, LAUNCH_MODE, wave_launch_mode); + + return data; +} + +#define TCP_WATCH_STRIDE (regTCP_WATCH1_ADDR_H - regTCP_WATCH0_ADDR_H) +static uint32_t kgd_gfx_aldebaran_set_address_watch( + struct amdgpu_device *adev, + uint64_t watch_address, + uint32_t watch_address_mask, + uint32_t watch_id, + uint32_t watch_mode, + uint32_t debug_vmid) +{ + uint32_t watch_address_high; + uint32_t watch_address_low; + uint32_t watch_address_cntl; + + watch_address_cntl = 0; + watch_address_low = lower_32_bits(watch_address); + watch_address_high = upper_32_bits(watch_address) & 0xffff; + + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + MODE, + watch_mode); + + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + MASK, + watch_address_mask >> 6); + + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + VALID, + 1); + + WREG32_RLC((SOC15_REG_OFFSET(GC, 0, regTCP_WATCH0_ADDR_H) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_high); + + WREG32_RLC((SOC15_REG_OFFSET(GC, 0, regTCP_WATCH0_ADDR_L) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_low); + + return watch_address_cntl; +} + +static uint32_t kgd_gfx_aldebaran_clear_address_watch(struct amdgpu_device *adev, + uint32_t watch_id) +{ + return 0; +} const struct kfd2kgd_calls aldebaran_kfd2kgd = { .program_sh_mem_settings = kgd_gfx_v9_program_sh_mem_settings, @@ -42,5 +185,14 @@ const struct kfd2kgd_calls aldebaran_kfd2kgd = { kgd_gfx_v9_get_atc_vmid_pasid_mapping_info, .set_vm_context_page_table_base = kgd_gfx_v9_set_vm_context_page_table_base, .get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy, - .program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings + .enable_debug_trap = kgd_aldebaran_enable_debug_trap, + .disable_debug_trap = kgd_aldebaran_disable_debug_trap, + .validate_trap_override_request = kgd_aldebaran_validate_trap_override_request, + .set_wave_launch_trap_override = kgd_aldebaran_set_wave_launch_trap_override, + .set_wave_launch_mode = kgd_aldebaran_set_wave_launch_mode, + .set_address_watch = kgd_gfx_aldebaran_set_address_watch, + .clear_address_watch = kgd_gfx_aldebaran_clear_address_watch, + .get_iq_wait_times = kgd_gfx_v9_get_iq_wait_times, + .build_grace_period_packet_info = kgd_gfx_v9_build_grace_period_packet_info, + .program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings, }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c index 4191af5a3f13..625db444df1c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c @@ -26,6 +26,7 @@ #include "amdgpu.h" #include "amdgpu_amdkfd.h" #include "amdgpu_amdkfd_arcturus.h" +#include "amdgpu_reset.h" #include "sdma0/sdma0_4_2_2_offset.h" #include "sdma0/sdma0_4_2_2_sh_mask.h" #include "sdma1/sdma1_4_2_2_offset.h" @@ -48,6 +49,8 @@ #include "amdgpu_amdkfd_gfx_v9.h" #include "gfxhub_v1_0.h" #include "mmhub_v9_4.h" +#include "gc/gc_9_0_offset.h" +#include "gc/gc_9_0_sh_mask.h" #define HQD_N_REGS 56 #define DUMP_REG(addr) do { \ @@ -276,6 +279,117 @@ int kgd_arcturus_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd, return 0; } +/* + * Helper used to suspend/resume gfx pipe for image post process work to set + * barrier behaviour. + */ +static int suspend_resume_compute_scheduler(struct amdgpu_device *adev, bool suspend) +{ + int i, r = 0; + + for (i = 0; i < adev->gfx.num_compute_rings; i++) { + struct amdgpu_ring *ring = &adev->gfx.compute_ring[i]; + + if (!(ring && ring->sched.thread)) + continue; + + /* stop secheduler and drain ring. */ + if (suspend) { + drm_sched_stop(&ring->sched, NULL); + r = amdgpu_fence_wait_empty(ring); + if (r) + goto out; + } else { + drm_sched_start(&ring->sched, false); + } + } + +out: + /* return on resume or failure to drain rings. */ + if (!suspend || r) + return r; + + return amdgpu_device_ip_wait_for_idle(adev, AMD_IP_BLOCK_TYPE_GFX); +} + +static void set_barrier_auto_waitcnt(struct amdgpu_device *adev, bool enable_waitcnt) +{ + uint32_t data; + + WRITE_ONCE(adev->barrier_has_auto_waitcnt, enable_waitcnt); + + if (!down_read_trylock(&adev->reset_domain->sem)) + return; + + amdgpu_amdkfd_suspend(adev, false); + + if (suspend_resume_compute_scheduler(adev, true)) + goto out; + + data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_CONFIG)); + data = REG_SET_FIELD(data, SQ_CONFIG, DISABLE_BARRIER_WAITCNT, + !enable_waitcnt); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_CONFIG), data); + +out: + suspend_resume_compute_scheduler(adev, false); + + amdgpu_amdkfd_resume(adev, false); + + up_read(&adev->reset_domain->sem); +} + +/* + * restore_dbg_registers is ignored here but is a general interface requirement + * for devices that support GFXOFF and where the RLC save/restore list + * does not support hw registers for debugging i.e. the driver has to manually + * initialize the debug mode registers after it has disabled GFX off during the + * debug session. + */ +static uint32_t kgd_arcturus_enable_debug_trap(struct amdgpu_device *adev, + bool restore_dbg_registers, + uint32_t vmid) +{ + mutex_lock(&adev->grbm_idx_mutex); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true); + + set_barrier_auto_waitcnt(adev, true); + + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, false); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; +} + +/* + * keep_trap_enabled is ignored here but is a general interface requirement + * for devices that support multi-process debugging where the performance + * overhead from trap temporary setup needs to be bypassed when the debug + * session has ended. + */ +static uint32_t kgd_arcturus_disable_debug_trap(struct amdgpu_device *adev, + bool keep_trap_enabled, + uint32_t vmid) +{ + + mutex_lock(&adev->grbm_idx_mutex); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true); + + set_barrier_auto_waitcnt(adev, false); + + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, false); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; +} const struct kfd2kgd_calls arcturus_kfd2kgd = { .program_sh_mem_settings = kgd_gfx_v9_program_sh_mem_settings, .set_pasid_vmid_mapping = kgd_gfx_v9_set_pasid_vmid_mapping, @@ -294,6 +408,15 @@ const struct kfd2kgd_calls arcturus_kfd2kgd = { kgd_gfx_v9_get_atc_vmid_pasid_mapping_info, .set_vm_context_page_table_base = kgd_gfx_v9_set_vm_context_page_table_base, + .enable_debug_trap = kgd_arcturus_enable_debug_trap, + .disable_debug_trap = kgd_arcturus_disable_debug_trap, + .validate_trap_override_request = kgd_gfx_v9_validate_trap_override_request, + .set_wave_launch_trap_override = kgd_gfx_v9_set_wave_launch_trap_override, + .set_wave_launch_mode = kgd_gfx_v9_set_wave_launch_mode, + .set_address_watch = kgd_gfx_v9_set_address_watch, + .clear_address_watch = kgd_gfx_v9_clear_address_watch, + .get_iq_wait_times = kgd_gfx_v9_get_iq_wait_times, + .build_grace_period_packet_info = kgd_gfx_v9_build_grace_period_packet_info, .get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy, .program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gc_9_4_3.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gc_9_4_3.c new file mode 100644 index 000000000000..5b4b7f8b92a5 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gc_9_4_3.c @@ -0,0 +1,384 @@ +/* + * Copyright 2021 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ +#include "amdgpu.h" +#include "amdgpu_amdkfd.h" +#include "amdgpu_amdkfd_gfx_v9.h" +#include "gc/gc_9_4_3_offset.h" +#include "gc/gc_9_4_3_sh_mask.h" +#include "athub/athub_1_8_0_offset.h" +#include "athub/athub_1_8_0_sh_mask.h" +#include "oss/osssys_4_4_2_offset.h" +#include "oss/osssys_4_4_2_sh_mask.h" +#include "v9_structs.h" +#include "soc15.h" +#include "sdma/sdma_4_4_2_offset.h" +#include "sdma/sdma_4_4_2_sh_mask.h" + +static inline struct v9_sdma_mqd *get_sdma_mqd(void *mqd) +{ + return (struct v9_sdma_mqd *)mqd; +} + +static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev, + unsigned int engine_id, + unsigned int queue_id) +{ + uint32_t sdma_engine_reg_base = + SOC15_REG_OFFSET(SDMA0, GET_INST(SDMA0, engine_id), + regSDMA_RLC0_RB_CNTL) - + regSDMA_RLC0_RB_CNTL; + uint32_t retval = sdma_engine_reg_base + + queue_id * (regSDMA_RLC1_RB_CNTL - regSDMA_RLC0_RB_CNTL); + + pr_debug("RLC register offset for SDMA%d RLC%d: 0x%x\n", engine_id, + queue_id, retval); + return retval; +} + +static int kgd_gfx_v9_4_3_hqd_sdma_load(struct amdgpu_device *adev, void *mqd, + uint32_t __user *wptr, struct mm_struct *mm) +{ + struct v9_sdma_mqd *m; + uint32_t sdma_rlc_reg_offset; + unsigned long end_jiffies; + uint32_t data; + uint64_t data64; + uint64_t __user *wptr64 = (uint64_t __user *)wptr; + + m = get_sdma_mqd(mqd); + sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id, + m->sdma_queue_id); + + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL, + m->sdmax_rlcx_rb_cntl & (~SDMA_RLC0_RB_CNTL__RB_ENABLE_MASK)); + + end_jiffies = msecs_to_jiffies(2000) + jiffies; + while (true) { + data = RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_CONTEXT_STATUS); + if (data & SDMA_RLC0_CONTEXT_STATUS__IDLE_MASK) + break; + if (time_after(jiffies, end_jiffies)) { + pr_err("SDMA RLC not idle in %s\n", __func__); + return -ETIME; + } + usleep_range(500, 1000); + } + + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_DOORBELL_OFFSET, + m->sdmax_rlcx_doorbell_offset); + + data = REG_SET_FIELD(m->sdmax_rlcx_doorbell, SDMA_RLC0_DOORBELL, + ENABLE, 1); + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_DOORBELL, data); + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_RPTR, + m->sdmax_rlcx_rb_rptr); + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_RPTR_HI, + m->sdmax_rlcx_rb_rptr_hi); + + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_MINOR_PTR_UPDATE, 1); + if (read_user_wptr(mm, wptr64, data64)) { + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_WPTR, + lower_32_bits(data64)); + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_WPTR_HI, + upper_32_bits(data64)); + } else { + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_WPTR, + m->sdmax_rlcx_rb_rptr); + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_WPTR_HI, + m->sdmax_rlcx_rb_rptr_hi); + } + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_MINOR_PTR_UPDATE, 0); + + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_BASE, m->sdmax_rlcx_rb_base); + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_BASE_HI, + m->sdmax_rlcx_rb_base_hi); + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_RPTR_ADDR_LO, + m->sdmax_rlcx_rb_rptr_addr_lo); + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_RPTR_ADDR_HI, + m->sdmax_rlcx_rb_rptr_addr_hi); + + data = REG_SET_FIELD(m->sdmax_rlcx_rb_cntl, SDMA_RLC0_RB_CNTL, + RB_ENABLE, 1); + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL, data); + + return 0; +} + +static int kgd_gfx_v9_4_3_hqd_sdma_dump(struct amdgpu_device *adev, + uint32_t engine_id, uint32_t queue_id, + uint32_t (**dump)[2], uint32_t *n_regs) +{ + uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, + engine_id, queue_id); + uint32_t i = 0, reg; +#undef HQD_N_REGS +#define HQD_N_REGS (19+6+7+12) +#define DUMP_REG(addr) do { \ + if (WARN_ON_ONCE(i >= HQD_N_REGS)) \ + break; \ + (*dump)[i][0] = (addr) << 2; \ + (*dump)[i++][1] = RREG32(addr); \ + } while (0) + + *dump = kmalloc_array(HQD_N_REGS * 2, sizeof(uint32_t), GFP_KERNEL); + if (*dump == NULL) + return -ENOMEM; + + for (reg = regSDMA_RLC0_RB_CNTL; reg <= regSDMA_RLC0_DOORBELL; reg++) + DUMP_REG(sdma_rlc_reg_offset + reg); + for (reg = regSDMA_RLC0_STATUS; reg <= regSDMA_RLC0_CSA_ADDR_HI; reg++) + DUMP_REG(sdma_rlc_reg_offset + reg); + for (reg = regSDMA_RLC0_IB_SUB_REMAIN; + reg <= regSDMA_RLC0_MINOR_PTR_UPDATE; reg++) + DUMP_REG(sdma_rlc_reg_offset + reg); + for (reg = regSDMA_RLC0_MIDCMD_DATA0; + reg <= regSDMA_RLC0_MIDCMD_CNTL; reg++) + DUMP_REG(sdma_rlc_reg_offset + reg); + + WARN_ON_ONCE(i != HQD_N_REGS); + *n_regs = i; + + return 0; +} + +static bool kgd_gfx_v9_4_3_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd) +{ + struct v9_sdma_mqd *m; + uint32_t sdma_rlc_reg_offset; + uint32_t sdma_rlc_rb_cntl; + + m = get_sdma_mqd(mqd); + sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id, + m->sdma_queue_id); + + sdma_rlc_rb_cntl = RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL); + + if (sdma_rlc_rb_cntl & SDMA_RLC0_RB_CNTL__RB_ENABLE_MASK) + return true; + + return false; +} + +static int kgd_gfx_v9_4_3_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd, + unsigned int utimeout) +{ + struct v9_sdma_mqd *m; + uint32_t sdma_rlc_reg_offset; + uint32_t temp; + unsigned long end_jiffies = (utimeout * HZ / 1000) + jiffies; + + m = get_sdma_mqd(mqd); + sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id, + m->sdma_queue_id); + + temp = RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL); + temp = temp & ~SDMA_RLC0_RB_CNTL__RB_ENABLE_MASK; + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL, temp); + + while (true) { + temp = RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_CONTEXT_STATUS); + if (temp & SDMA_RLC0_CONTEXT_STATUS__IDLE_MASK) + break; + if (time_after(jiffies, end_jiffies)) { + pr_err("SDMA RLC not idle in %s\n", __func__); + return -ETIME; + } + usleep_range(500, 1000); + } + + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_DOORBELL, 0); + WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL, + RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL) | + SDMA_RLC0_RB_CNTL__RB_ENABLE_MASK); + + m->sdmax_rlcx_rb_rptr = + RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_RPTR); + m->sdmax_rlcx_rb_rptr_hi = + RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_RPTR_HI); + + return 0; +} + +static int kgd_gfx_v9_4_3_set_pasid_vmid_mapping(struct amdgpu_device *adev, + u32 pasid, unsigned int vmid, uint32_t xcc_inst) +{ + unsigned long timeout; + unsigned int reg; + unsigned int phy_inst = GET_INST(GC, xcc_inst); + /* Every two XCCs share one AID */ + unsigned int aid = phy_inst / 2; + + /* + * We have to assume that there is no outstanding mapping. + * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because + * a mapping is in progress or because a mapping finished + * and the SW cleared it. + * So the protocol is to always wait & clear. + */ + uint32_t pasid_mapping = (pasid == 0) ? 0 : (uint32_t)pasid | + ATC_VMID0_PASID_MAPPING__VALID_MASK; + + WREG32(SOC15_REG_OFFSET(ATHUB, 0, + regATC_VMID0_PASID_MAPPING) + vmid, pasid_mapping); + + timeout = jiffies + msecs_to_jiffies(10); + while (!(RREG32(SOC15_REG_OFFSET(ATHUB, 0, + regATC_VMID_PASID_MAPPING_UPDATE_STATUS)) & + (1U << vmid))) { + if (time_after(jiffies, timeout)) { + pr_err("Fail to program VMID-PASID mapping\n"); + return -ETIME; + } + cpu_relax(); + } + + WREG32(SOC15_REG_OFFSET(ATHUB, 0, + regATC_VMID_PASID_MAPPING_UPDATE_STATUS), + 1U << vmid); + + reg = RREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_LUT_INDEX)); + /* Every 4 numbers is a cycle. 1st is AID, 2nd and 3rd are XCDs, + * and the 4th is reserved. Therefore "aid * 4 + (xcc_inst % 2) + 1" + * programs _LUT for XCC and "aid * 4" for AID where the XCC connects + * to. + */ + WREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_LUT_INDEX), + aid * 4 + (phy_inst % 2) + 1); + WREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT) + vmid, + pasid_mapping); + WREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_LUT_INDEX), + aid * 4); + WREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT_MM) + vmid, + pasid_mapping); + WREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_LUT_INDEX), reg); + + return 0; +} + +static inline struct v9_mqd *get_mqd(void *mqd) +{ + return (struct v9_mqd *)mqd; +} + +static int kgd_gfx_v9_4_3_hqd_load(struct amdgpu_device *adev, void *mqd, + uint32_t pipe_id, uint32_t queue_id, + uint32_t __user *wptr, uint32_t wptr_shift, + uint32_t wptr_mask, struct mm_struct *mm, uint32_t inst) +{ + struct v9_mqd *m; + uint32_t *mqd_hqd; + uint32_t reg, hqd_base, hqd_end, data; + + m = get_mqd(mqd); + + kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst); + + /* HQD registers extend to CP_HQD_AQL_DISPATCH_ID_HI */ + mqd_hqd = &m->cp_mqd_base_addr_lo; + hqd_base = SOC15_REG_OFFSET(GC, GET_INST(GC, inst), regCP_MQD_BASE_ADDR); + hqd_end = SOC15_REG_OFFSET(GC, GET_INST(GC, inst), regCP_HQD_AQL_DISPATCH_ID_HI); + + for (reg = hqd_base; reg <= hqd_end; reg++) + WREG32_RLC(reg, mqd_hqd[reg - hqd_base]); + + + /* Activate doorbell logic before triggering WPTR poll. */ + data = REG_SET_FIELD(m->cp_hqd_pq_doorbell_control, + CP_HQD_PQ_DOORBELL_CONTROL, DOORBELL_EN, 1); + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), regCP_HQD_PQ_DOORBELL_CONTROL), + data); + + if (wptr) { + /* Don't read wptr with get_user because the user + * context may not be accessible (if this function + * runs in a work queue). Instead trigger a one-shot + * polling read from memory in the CP. This assumes + * that wptr is GPU-accessible in the queue's VMID via + * ATC or SVM. WPTR==RPTR before starting the poll so + * the CP starts fetching new commands from the right + * place. + * + * Guessing a 64-bit WPTR from a 32-bit RPTR is a bit + * tricky. Assume that the queue didn't overflow. The + * number of valid bits in the 32-bit RPTR depends on + * the queue size. The remaining bits are taken from + * the saved 64-bit WPTR. If the WPTR wrapped, add the + * queue size. + */ + uint32_t queue_size = + 2 << REG_GET_FIELD(m->cp_hqd_pq_control, + CP_HQD_PQ_CONTROL, QUEUE_SIZE); + uint64_t guessed_wptr = m->cp_hqd_pq_rptr & (queue_size - 1); + + if ((m->cp_hqd_pq_wptr_lo & (queue_size - 1)) < guessed_wptr) + guessed_wptr += queue_size; + guessed_wptr += m->cp_hqd_pq_wptr_lo & ~(queue_size - 1); + guessed_wptr += (uint64_t)m->cp_hqd_pq_wptr_hi << 32; + + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), regCP_HQD_PQ_WPTR_LO), + lower_32_bits(guessed_wptr)); + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), regCP_HQD_PQ_WPTR_HI), + upper_32_bits(guessed_wptr)); + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), regCP_HQD_PQ_WPTR_POLL_ADDR), + lower_32_bits((uintptr_t)wptr)); + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), + regCP_HQD_PQ_WPTR_POLL_ADDR_HI), + upper_32_bits((uintptr_t)wptr)); + WREG32(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), regCP_PQ_WPTR_POLL_CNTL1), + (uint32_t)kgd_gfx_v9_get_queue_mask(adev, pipe_id, + queue_id)); + } + + /* Start the EOP fetcher */ + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), regCP_HQD_EOP_RPTR), + REG_SET_FIELD(m->cp_hqd_eop_rptr, + CP_HQD_EOP_RPTR, INIT_FETCHER, 1)); + + data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1); + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), regCP_HQD_ACTIVE), data); + + kgd_gfx_v9_release_queue(adev, inst); + + return 0; +} + +const struct kfd2kgd_calls gc_9_4_3_kfd2kgd = { + .program_sh_mem_settings = kgd_gfx_v9_program_sh_mem_settings, + .set_pasid_vmid_mapping = kgd_gfx_v9_4_3_set_pasid_vmid_mapping, + .init_interrupts = kgd_gfx_v9_init_interrupts, + .hqd_load = kgd_gfx_v9_4_3_hqd_load, + .hiq_mqd_load = kgd_gfx_v9_hiq_mqd_load, + .hqd_sdma_load = kgd_gfx_v9_4_3_hqd_sdma_load, + .hqd_dump = kgd_gfx_v9_hqd_dump, + .hqd_sdma_dump = kgd_gfx_v9_4_3_hqd_sdma_dump, + .hqd_is_occupied = kgd_gfx_v9_hqd_is_occupied, + .hqd_sdma_is_occupied = kgd_gfx_v9_4_3_hqd_sdma_is_occupied, + .hqd_destroy = kgd_gfx_v9_hqd_destroy, + .hqd_sdma_destroy = kgd_gfx_v9_4_3_hqd_sdma_destroy, + .wave_control_execute = kgd_gfx_v9_wave_control_execute, + .get_atc_vmid_pasid_mapping_info = + kgd_gfx_v9_get_atc_vmid_pasid_mapping_info, + .set_vm_context_page_table_base = + kgd_gfx_v9_set_vm_context_page_table_base, + .program_trap_handler_settings = + kgd_gfx_v9_program_trap_handler_settings +}; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c index 9378fc79e9ea..8ad7a7779e14 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c @@ -21,6 +21,7 @@ */ #include "amdgpu.h" #include "amdgpu_amdkfd.h" +#include "amdgpu_amdkfd_gfx_v10.h" #include "gc/gc_10_1_0_offset.h" #include "gc/gc_10_1_0_sh_mask.h" #include "athub/athub_2_0_0_offset.h" @@ -31,6 +32,7 @@ #include "v10_structs.h" #include "nv.h" #include "nvd.h" +#include <uapi/linux/kfd_ioctl.h> enum hqd_dequeue_request_type { NO_ACTION = 0, @@ -79,7 +81,7 @@ static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmi uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, - uint32_t sh_mem_bases) + uint32_t sh_mem_bases, uint32_t inst) { lock_srbm(adev, 0, 0, 0, vmid); @@ -91,7 +93,7 @@ static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmi } static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, - unsigned int vmid) + unsigned int vmid, uint32_t inst) { /* * We have to assume that there is no outstanding mapping. @@ -135,7 +137,8 @@ static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, * but still works */ -static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id) +static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id, + uint32_t inst) { uint32_t mec; uint32_t pipe; @@ -205,7 +208,7 @@ static inline struct v10_sdma_mqd *get_sdma_mqd(void *mqd) static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t __user *wptr, uint32_t wptr_shift, - uint32_t wptr_mask, struct mm_struct *mm) + uint32_t wptr_mask, struct mm_struct *mm, uint32_t inst) { struct v10_compute_mqd *m; uint32_t *mqd_hqd; @@ -286,9 +289,9 @@ static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd, static int kgd_hiq_mqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, - uint32_t doorbell_off) + uint32_t doorbell_off, uint32_t inst) { - struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; + struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring; struct v10_compute_mqd *m; uint32_t mec, pipe; int r; @@ -303,7 +306,7 @@ static int kgd_hiq_mqd_load(struct amdgpu_device *adev, void *mqd, pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n", mec, pipe, queue_id); - spin_lock(&adev->gfx.kiq.ring_lock); + spin_lock(&adev->gfx.kiq[0].ring_lock); r = amdgpu_ring_alloc(kiq_ring, 7); if (r) { pr_err("Failed to alloc KIQ (%d).\n", r); @@ -330,7 +333,7 @@ static int kgd_hiq_mqd_load(struct amdgpu_device *adev, void *mqd, amdgpu_ring_commit(kiq_ring); out_unlock: - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&adev->gfx.kiq[0].ring_lock); release_queue(adev); return r; @@ -338,7 +341,7 @@ out_unlock: static int kgd_hqd_dump(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, - uint32_t (**dump)[2], uint32_t *n_regs) + uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst) { uint32_t i = 0, reg; #define HQD_N_REGS 56 @@ -469,7 +472,7 @@ static int kgd_hqd_sdma_dump(struct amdgpu_device *adev, static bool kgd_hqd_is_occupied(struct amdgpu_device *adev, uint64_t queue_address, uint32_t pipe_id, - uint32_t queue_id) + uint32_t queue_id, uint32_t inst) { uint32_t act; bool retval = false; @@ -510,7 +513,7 @@ static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd) static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, - uint32_t queue_id) + uint32_t queue_id, uint32_t inst) { enum hqd_dequeue_request_type type; unsigned long end_jiffies; @@ -673,7 +676,7 @@ static bool get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev, static int kgd_wave_control_execute(struct amdgpu_device *adev, uint32_t gfx_index_val, - uint32_t sq_cmd) + uint32_t sq_cmd, uint32_t inst) { uint32_t data = 0; @@ -708,8 +711,295 @@ static void set_vm_context_page_table_base(struct amdgpu_device *adev, adev->gfxhub.funcs->setup_vm_pt_regs(adev, vmid, page_table_base); } +/* + * GFX10 helper for wave launch stall requirements on debug trap setting. + * + * vmid: + * Target VMID to stall/unstall. + * + * stall: + * 0-unstall wave launch (enable), 1-stall wave launch (disable). + * After wavefront launch has been stalled, allocated waves must drain from + * SPI in order for debug trap settings to take effect on those waves. + * This is roughly a ~3500 clock cycle wait on SPI where a read on + * SPI_GDBG_WAVE_CNTL translates to ~32 clock cycles. + * KGD_GFX_V10_WAVE_LAUNCH_SPI_DRAIN_LATENCY indicates the number of reads required. + * + * NOTE: We can afford to clear the entire STALL_VMID field on unstall + * because current GFX10 chips cannot support multi-process debugging due to + * trap configuration and masking being limited to global scope. Always + * assume single process conditions. + * + */ + +#define KGD_GFX_V10_WAVE_LAUNCH_SPI_DRAIN_LATENCY 110 +static void kgd_gfx_v10_set_wave_launch_stall(struct amdgpu_device *adev, uint32_t vmid, bool stall) +{ + uint32_t data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL)); + int i; + + data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_VMID, + stall ? 1 << vmid : 0); + + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data); + + if (!stall) + return; + + for (i = 0; i < KGD_GFX_V10_WAVE_LAUNCH_SPI_DRAIN_LATENCY; i++) + RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL)); +} + +uint32_t kgd_gfx_v10_enable_debug_trap(struct amdgpu_device *adev, + bool restore_dbg_registers, + uint32_t vmid) +{ + + mutex_lock(&adev->grbm_idx_mutex); + + kgd_gfx_v10_set_wave_launch_stall(adev, vmid, true); + + /* assume gfx off is disabled for the debug session if rlc restore not supported. */ + if (restore_dbg_registers) { + uint32_t data = 0; + + data = REG_SET_FIELD(data, SPI_GDBG_TRAP_CONFIG, + VMID_SEL, 1 << vmid); + data = REG_SET_FIELD(data, SPI_GDBG_TRAP_CONFIG, + TRAP_EN, 1); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_CONFIG), data); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA0), 0); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA1), 0); + + kgd_gfx_v10_set_wave_launch_stall(adev, vmid, false); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; + } + + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0); + + kgd_gfx_v10_set_wave_launch_stall(adev, vmid, false); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; +} + +uint32_t kgd_gfx_v10_disable_debug_trap(struct amdgpu_device *adev, + bool keep_trap_enabled, + uint32_t vmid) +{ + mutex_lock(&adev->grbm_idx_mutex); + + kgd_gfx_v10_set_wave_launch_stall(adev, vmid, true); + + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0); + + kgd_gfx_v10_set_wave_launch_stall(adev, vmid, false); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; +} + +int kgd_gfx_v10_validate_trap_override_request(struct amdgpu_device *adev, + uint32_t trap_override, + uint32_t *trap_mask_supported) +{ + *trap_mask_supported &= KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH; + + /* The SPI_GDBG_TRAP_MASK register is global and affects all + * processes. Only allow OR-ing the address-watch bit, since + * this only affects processes under the debugger. Other bits + * should stay 0 to avoid the debugger interfering with other + * processes. + */ + if (trap_override != KFD_DBG_TRAP_OVERRIDE_OR) + return -EINVAL; + + return 0; +} + +uint32_t kgd_gfx_v10_set_wave_launch_trap_override(struct amdgpu_device *adev, + uint32_t vmid, + uint32_t trap_override, + uint32_t trap_mask_bits, + uint32_t trap_mask_request, + uint32_t *trap_mask_prev, + uint32_t kfd_dbg_trap_cntl_prev) +{ + uint32_t data, wave_cntl_prev; + + mutex_lock(&adev->grbm_idx_mutex); + + wave_cntl_prev = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL)); + + kgd_gfx_v10_set_wave_launch_stall(adev, vmid, true); + + data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK)); + *trap_mask_prev = REG_GET_FIELD(data, SPI_GDBG_TRAP_MASK, EXCP_EN); + + trap_mask_bits = (trap_mask_bits & trap_mask_request) | + (*trap_mask_prev & ~trap_mask_request); + + data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK, EXCP_EN, trap_mask_bits); + data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK, REPLACE, trap_override); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), data); + + /* We need to preserve wave launch mode stall settings. */ + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), wave_cntl_prev); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; +} + +uint32_t kgd_gfx_v10_set_wave_launch_mode(struct amdgpu_device *adev, + uint8_t wave_launch_mode, + uint32_t vmid) +{ + uint32_t data = 0; + bool is_mode_set = !!wave_launch_mode; + + mutex_lock(&adev->grbm_idx_mutex); + + kgd_gfx_v10_set_wave_launch_stall(adev, vmid, true); + + data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2, + VMID_MASK, is_mode_set ? 1 << vmid : 0); + data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2, + MODE, is_mode_set ? wave_launch_mode : 0); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL2), data); + + kgd_gfx_v10_set_wave_launch_stall(adev, vmid, false); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; +} + +#define TCP_WATCH_STRIDE (mmTCP_WATCH1_ADDR_H - mmTCP_WATCH0_ADDR_H) +uint32_t kgd_gfx_v10_set_address_watch(struct amdgpu_device *adev, + uint64_t watch_address, + uint32_t watch_address_mask, + uint32_t watch_id, + uint32_t watch_mode, + uint32_t debug_vmid) +{ + uint32_t watch_address_high; + uint32_t watch_address_low; + uint32_t watch_address_cntl; + + watch_address_cntl = 0; + + watch_address_low = lower_32_bits(watch_address); + watch_address_high = upper_32_bits(watch_address) & 0xffff; + + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + VMID, + debug_vmid); + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + MODE, + watch_mode); + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + MASK, + watch_address_mask >> 7); + + /* Turning off this watch point until we set all the registers */ + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + VALID, + 0); + + WREG32((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_CNTL) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_cntl); + + WREG32((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_ADDR_H) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_high); + + WREG32((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_ADDR_L) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_low); + + /* Enable the watch point */ + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + VALID, + 1); + + WREG32((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_CNTL) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_cntl); + + return 0; +} + +uint32_t kgd_gfx_v10_clear_address_watch(struct amdgpu_device *adev, + uint32_t watch_id) +{ + uint32_t watch_address_cntl; + + watch_address_cntl = 0; + + WREG32((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_CNTL) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_cntl); + + return 0; +} + + +/* kgd_gfx_v10_get_iq_wait_times: Returns the mmCP_IQ_WAIT_TIME1/2 values + * The values read are: + * ib_offload_wait_time -- Wait Count for Indirect Buffer Offloads. + * atomic_offload_wait_time -- Wait Count for L2 and GDS Atomics Offloads. + * wrm_offload_wait_time -- Wait Count for WAIT_REG_MEM Offloads. + * gws_wait_time -- Wait Count for Global Wave Syncs. + * que_sleep_wait_time -- Wait Count for Dequeue Retry. + * sch_wave_wait_time -- Wait Count for Scheduling Wave Message. + * sem_rearm_wait_time -- Wait Count for Semaphore re-arm. + * deq_retry_wait_time -- Wait Count for Global Wave Syncs. + */ +void kgd_gfx_v10_get_iq_wait_times(struct amdgpu_device *adev, + uint32_t *wait_times) + +{ + *wait_times = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_IQ_WAIT_TIME2)); +} + +void kgd_gfx_v10_build_grace_period_packet_info(struct amdgpu_device *adev, + uint32_t wait_times, + uint32_t grace_period, + uint32_t *reg_offset, + uint32_t *reg_data) +{ + *reg_data = wait_times; + + /* + * The CP cannont handle a 0 grace period input and will result in + * an infinite grace period being set so set to 1 to prevent this. + */ + if (grace_period == 0) + grace_period = 1; + + *reg_data = REG_SET_FIELD(*reg_data, + CP_IQ_WAIT_TIME2, + SCH_WAVE, + grace_period); + + *reg_offset = SOC15_REG_OFFSET(GC, 0, mmCP_IQ_WAIT_TIME2); +} + static void program_trap_handler_settings(struct amdgpu_device *adev, - uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr) + uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr, + uint32_t inst) { lock_srbm(adev, 0, 0, 0, vmid); @@ -750,5 +1040,14 @@ const struct kfd2kgd_calls gfx_v10_kfd2kgd = { .get_atc_vmid_pasid_mapping_info = get_atc_vmid_pasid_mapping_info, .set_vm_context_page_table_base = set_vm_context_page_table_base, + .enable_debug_trap = kgd_gfx_v10_enable_debug_trap, + .disable_debug_trap = kgd_gfx_v10_disable_debug_trap, + .validate_trap_override_request = kgd_gfx_v10_validate_trap_override_request, + .set_wave_launch_trap_override = kgd_gfx_v10_set_wave_launch_trap_override, + .set_wave_launch_mode = kgd_gfx_v10_set_wave_launch_mode, + .set_address_watch = kgd_gfx_v10_set_address_watch, + .clear_address_watch = kgd_gfx_v10_clear_address_watch, + .get_iq_wait_times = kgd_gfx_v10_get_iq_wait_times, + .build_grace_period_packet_info = kgd_gfx_v10_build_grace_period_packet_info, .program_trap_handler_settings = program_trap_handler_settings, }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.h new file mode 100644 index 000000000000..e6b70196071a --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.h @@ -0,0 +1,55 @@ +/* + * Copyright 2023 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ + +uint32_t kgd_gfx_v10_enable_debug_trap(struct amdgpu_device *adev, + bool restore_dbg_registers, + uint32_t vmid); +uint32_t kgd_gfx_v10_disable_debug_trap(struct amdgpu_device *adev, + bool keep_trap_enabled, + uint32_t vmid); +int kgd_gfx_v10_validate_trap_override_request(struct amdgpu_device *adev, + uint32_t trap_override, + uint32_t *trap_mask_supported); +uint32_t kgd_gfx_v10_set_wave_launch_trap_override(struct amdgpu_device *adev, + uint32_t vmid, + uint32_t trap_override, + uint32_t trap_mask_bits, + uint32_t trap_mask_request, + uint32_t *trap_mask_prev, + uint32_t kfd_dbg_trap_cntl_prev); +uint32_t kgd_gfx_v10_set_wave_launch_mode(struct amdgpu_device *adev, + uint8_t wave_launch_mode, + uint32_t vmid); +uint32_t kgd_gfx_v10_set_address_watch(struct amdgpu_device *adev, + uint64_t watch_address, + uint32_t watch_address_mask, + uint32_t watch_id, + uint32_t watch_mode, + uint32_t debug_vmid); +uint32_t kgd_gfx_v10_clear_address_watch(struct amdgpu_device *adev, + uint32_t watch_id); +void kgd_gfx_v10_get_iq_wait_times(struct amdgpu_device *adev, uint32_t *wait_times); +void kgd_gfx_v10_build_grace_period_packet_info(struct amdgpu_device *adev, + uint32_t wait_times, + uint32_t grace_period, + uint32_t *reg_offset, + uint32_t *reg_data); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c index ba21ec6b35e0..8c8437a4383f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c @@ -22,6 +22,7 @@ #include <linux/mmu_context.h> #include "amdgpu.h" #include "amdgpu_amdkfd.h" +#include "amdgpu_amdkfd_gfx_v10.h" #include "gc/gc_10_3_0_offset.h" #include "gc/gc_10_3_0_sh_mask.h" #include "oss/osssys_5_0_0_offset.h" @@ -80,7 +81,7 @@ static void program_sh_mem_settings_v10_3(struct amdgpu_device *adev, uint32_t v uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, - uint32_t sh_mem_bases) + uint32_t sh_mem_bases, uint32_t inst) { lock_srbm(adev, 0, 0, 0, vmid); @@ -93,7 +94,7 @@ static void program_sh_mem_settings_v10_3(struct amdgpu_device *adev, uint32_t v /* ATC is defeatured on Sienna_Cichlid */ static int set_pasid_vmid_mapping_v10_3(struct amdgpu_device *adev, unsigned int pasid, - unsigned int vmid) + unsigned int vmid, uint32_t inst) { uint32_t value = pasid << IH_VMID_0_LUT__PASID__SHIFT; @@ -105,7 +106,8 @@ static int set_pasid_vmid_mapping_v10_3(struct amdgpu_device *adev, unsigned int return 0; } -static int init_interrupts_v10_3(struct amdgpu_device *adev, uint32_t pipe_id) +static int init_interrupts_v10_3(struct amdgpu_device *adev, uint32_t pipe_id, + uint32_t inst) { uint32_t mec; uint32_t pipe; @@ -177,7 +179,7 @@ static inline struct v10_sdma_mqd *get_sdma_mqd(void *mqd) static int hqd_load_v10_3(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t __user *wptr, uint32_t wptr_shift, - uint32_t wptr_mask, struct mm_struct *mm) + uint32_t wptr_mask, struct mm_struct *mm, uint32_t inst) { struct v10_compute_mqd *m; uint32_t *mqd_hqd; @@ -273,9 +275,9 @@ static int hqd_load_v10_3(struct amdgpu_device *adev, void *mqd, static int hiq_mqd_load_v10_3(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, - uint32_t doorbell_off) + uint32_t doorbell_off, uint32_t inst) { - struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; + struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring; struct v10_compute_mqd *m; uint32_t mec, pipe; int r; @@ -290,7 +292,7 @@ static int hiq_mqd_load_v10_3(struct amdgpu_device *adev, void *mqd, pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n", mec, pipe, queue_id); - spin_lock(&adev->gfx.kiq.ring_lock); + spin_lock(&adev->gfx.kiq[0].ring_lock); r = amdgpu_ring_alloc(kiq_ring, 7); if (r) { pr_err("Failed to alloc KIQ (%d).\n", r); @@ -317,7 +319,7 @@ static int hiq_mqd_load_v10_3(struct amdgpu_device *adev, void *mqd, amdgpu_ring_commit(kiq_ring); out_unlock: - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&adev->gfx.kiq[0].ring_lock); release_queue(adev); return r; @@ -325,7 +327,7 @@ out_unlock: static int hqd_dump_v10_3(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, - uint32_t (**dump)[2], uint32_t *n_regs) + uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst) { uint32_t i = 0, reg; #define HQD_N_REGS 56 @@ -456,7 +458,7 @@ static int hqd_sdma_dump_v10_3(struct amdgpu_device *adev, static bool hqd_is_occupied_v10_3(struct amdgpu_device *adev, uint64_t queue_address, uint32_t pipe_id, - uint32_t queue_id) + uint32_t queue_id, uint32_t inst) { uint32_t act; bool retval = false; @@ -498,7 +500,7 @@ static bool hqd_sdma_is_occupied_v10_3(struct amdgpu_device *adev, static int hqd_destroy_v10_3(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, - uint32_t queue_id) + uint32_t queue_id, uint32_t inst) { enum hqd_dequeue_request_type type; unsigned long end_jiffies; @@ -586,7 +588,7 @@ static int hqd_sdma_destroy_v10_3(struct amdgpu_device *adev, void *mqd, static int wave_control_execute_v10_3(struct amdgpu_device *adev, uint32_t gfx_index_val, - uint32_t sq_cmd) + uint32_t sq_cmd, uint32_t inst) { uint32_t data = 0; @@ -628,7 +630,8 @@ static void set_vm_context_page_table_base_v10_3(struct amdgpu_device *adev, } static void program_trap_handler_settings_v10_3(struct amdgpu_device *adev, - uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr) + uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr, + uint32_t inst) { lock_srbm(adev, 0, 0, 0, vmid); @@ -652,142 +655,6 @@ static void program_trap_handler_settings_v10_3(struct amdgpu_device *adev, unlock_srbm(adev); } -#if 0 -uint32_t enable_debug_trap_v10_3(struct amdgpu_device *adev, - uint32_t trap_debug_wave_launch_mode, - uint32_t vmid) -{ - uint32_t data = 0; - uint32_t orig_wave_cntl_value; - uint32_t orig_stall_vmid; - - mutex_lock(&adev->grbm_idx_mutex); - - orig_wave_cntl_value = RREG32(SOC15_REG_OFFSET(GC, - 0, - mmSPI_GDBG_WAVE_CNTL)); - orig_stall_vmid = REG_GET_FIELD(orig_wave_cntl_value, - SPI_GDBG_WAVE_CNTL, - STALL_VMID); - - data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_RA, 1); - WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data); - - data = 0; - WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), data); - - WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), orig_stall_vmid); - - mutex_unlock(&adev->grbm_idx_mutex); - - return 0; -} - -uint32_t disable_debug_trap_v10_3(struct amdgpu_device *adev) -{ - mutex_lock(&adev->grbm_idx_mutex); - - WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0); - - mutex_unlock(&adev->grbm_idx_mutex); - - return 0; -} - -uint32_t set_wave_launch_trap_override_v10_3(struct amdgpu_device *adev, - uint32_t trap_override, - uint32_t trap_mask) -{ - uint32_t data = 0; - - mutex_lock(&adev->grbm_idx_mutex); - - data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL)); - data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_RA, 1); - WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data); - - data = 0; - data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK, - EXCP_EN, trap_mask); - data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK, - REPLACE, trap_override); - WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), data); - - data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL)); - data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_RA, 0); - WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data); - - mutex_unlock(&adev->grbm_idx_mutex); - - return 0; -} - -uint32_t set_wave_launch_mode_v10_3(struct amdgpu_device *adev, - uint8_t wave_launch_mode, - uint32_t vmid) -{ - uint32_t data = 0; - bool is_stall_mode; - bool is_mode_set; - - is_stall_mode = (wave_launch_mode == 4); - is_mode_set = (wave_launch_mode != 0 && wave_launch_mode != 4); - - mutex_lock(&adev->grbm_idx_mutex); - - data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2, - VMID_MASK, is_mode_set ? 1 << vmid : 0); - data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2, - MODE, is_mode_set ? wave_launch_mode : 0); - WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL2), data); - - data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL)); - data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, - STALL_VMID, is_stall_mode ? 1 << vmid : 0); - data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, - STALL_RA, is_stall_mode ? 1 : 0); - WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data); - - mutex_unlock(&adev->grbm_idx_mutex); - - return 0; -} - -/* kgd_get_iq_wait_times: Returns the mmCP_IQ_WAIT_TIME1/2 values - * The values read are: - * ib_offload_wait_time -- Wait Count for Indirect Buffer Offloads. - * atomic_offload_wait_time -- Wait Count for L2 and GDS Atomics Offloads. - * wrm_offload_wait_time -- Wait Count for WAIT_REG_MEM Offloads. - * gws_wait_time -- Wait Count for Global Wave Syncs. - * que_sleep_wait_time -- Wait Count for Dequeue Retry. - * sch_wave_wait_time -- Wait Count for Scheduling Wave Message. - * sem_rearm_wait_time -- Wait Count for Semaphore re-arm. - * deq_retry_wait_time -- Wait Count for Global Wave Syncs. - */ -void get_iq_wait_times_v10_3(struct amdgpu_device *adev, - uint32_t *wait_times) - -{ - *wait_times = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_IQ_WAIT_TIME2)); -} - -void build_grace_period_packet_info_v10_3(struct amdgpu_device *adev, - uint32_t wait_times, - uint32_t grace_period, - uint32_t *reg_offset, - uint32_t *reg_data) -{ - *reg_data = wait_times; - - *reg_data = REG_SET_FIELD(*reg_data, - CP_IQ_WAIT_TIME2, - SCH_WAVE, - grace_period); - - *reg_offset = mmCP_IQ_WAIT_TIME2; -} -#endif - const struct kfd2kgd_calls gfx_v10_3_kfd2kgd = { .program_sh_mem_settings = program_sh_mem_settings_v10_3, .set_pasid_vmid_mapping = set_pasid_vmid_mapping_v10_3, @@ -805,12 +672,13 @@ const struct kfd2kgd_calls gfx_v10_3_kfd2kgd = { .get_atc_vmid_pasid_mapping_info = get_atc_vmid_pasid_mapping_info_v10_3, .set_vm_context_page_table_base = set_vm_context_page_table_base_v10_3, .program_trap_handler_settings = program_trap_handler_settings_v10_3, -#if 0 - .enable_debug_trap = enable_debug_trap_v10_3, - .disable_debug_trap = disable_debug_trap_v10_3, - .set_wave_launch_trap_override = set_wave_launch_trap_override_v10_3, - .set_wave_launch_mode = set_wave_launch_mode_v10_3, - .get_iq_wait_times = get_iq_wait_times_v10_3, - .build_grace_period_packet_info = build_grace_period_packet_info_v10_3, -#endif + .get_iq_wait_times = kgd_gfx_v10_get_iq_wait_times, + .build_grace_period_packet_info = kgd_gfx_v10_build_grace_period_packet_info, + .enable_debug_trap = kgd_gfx_v10_enable_debug_trap, + .disable_debug_trap = kgd_gfx_v10_disable_debug_trap, + .validate_trap_override_request = kgd_gfx_v10_validate_trap_override_request, + .set_wave_launch_trap_override = kgd_gfx_v10_set_wave_launch_trap_override, + .set_wave_launch_mode = kgd_gfx_v10_set_wave_launch_mode, + .set_address_watch = kgd_gfx_v10_set_address_watch, + .clear_address_watch = kgd_gfx_v10_clear_address_watch }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v11.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v11.c index 7e80caa05060..91c3574ebed3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v11.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v11.c @@ -30,6 +30,7 @@ #include "soc15d.h" #include "v11_structs.h" #include "soc21.h" +#include <uapi/linux/kfd_ioctl.h> enum hqd_dequeue_request_type { NO_ACTION = 0, @@ -78,7 +79,7 @@ static void program_sh_mem_settings_v11(struct amdgpu_device *adev, uint32_t vmi uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, - uint32_t sh_mem_bases) + uint32_t sh_mem_bases, uint32_t inst) { lock_srbm(adev, 0, 0, 0, vmid); @@ -89,7 +90,7 @@ static void program_sh_mem_settings_v11(struct amdgpu_device *adev, uint32_t vmi } static int set_pasid_vmid_mapping_v11(struct amdgpu_device *adev, unsigned int pasid, - unsigned int vmid) + unsigned int vmid, uint32_t inst) { uint32_t value = pasid << IH_VMID_0_LUT__PASID__SHIFT; @@ -101,7 +102,8 @@ static int set_pasid_vmid_mapping_v11(struct amdgpu_device *adev, unsigned int p return 0; } -static int init_interrupts_v11(struct amdgpu_device *adev, uint32_t pipe_id) +static int init_interrupts_v11(struct amdgpu_device *adev, uint32_t pipe_id, + uint32_t inst) { uint32_t mec; uint32_t pipe; @@ -162,7 +164,7 @@ static inline struct v11_sdma_mqd *get_sdma_mqd(void *mqd) static int hqd_load_v11(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t __user *wptr, uint32_t wptr_shift, uint32_t wptr_mask, - struct mm_struct *mm) + struct mm_struct *mm, uint32_t inst) { struct v11_compute_mqd *m; uint32_t *mqd_hqd; @@ -258,9 +260,9 @@ static int hqd_load_v11(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, static int hiq_mqd_load_v11(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, - uint32_t doorbell_off) + uint32_t doorbell_off, uint32_t inst) { - struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; + struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring; struct v11_compute_mqd *m; uint32_t mec, pipe; int r; @@ -275,7 +277,7 @@ static int hiq_mqd_load_v11(struct amdgpu_device *adev, void *mqd, pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n", mec, pipe, queue_id); - spin_lock(&adev->gfx.kiq.ring_lock); + spin_lock(&adev->gfx.kiq[0].ring_lock); r = amdgpu_ring_alloc(kiq_ring, 7); if (r) { pr_err("Failed to alloc KIQ (%d).\n", r); @@ -302,7 +304,7 @@ static int hiq_mqd_load_v11(struct amdgpu_device *adev, void *mqd, amdgpu_ring_commit(kiq_ring); out_unlock: - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&adev->gfx.kiq[0].ring_lock); release_queue(adev); return r; @@ -310,7 +312,7 @@ out_unlock: static int hqd_dump_v11(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, - uint32_t (**dump)[2], uint32_t *n_regs) + uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst) { uint32_t i = 0, reg; #define HQD_N_REGS 56 @@ -445,7 +447,7 @@ static int hqd_sdma_dump_v11(struct amdgpu_device *adev, } static bool hqd_is_occupied_v11(struct amdgpu_device *adev, uint64_t queue_address, - uint32_t pipe_id, uint32_t queue_id) + uint32_t pipe_id, uint32_t queue_id, uint32_t inst) { uint32_t act; bool retval = false; @@ -486,7 +488,7 @@ static bool hqd_sdma_is_occupied_v11(struct amdgpu_device *adev, void *mqd) static int hqd_destroy_v11(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, - uint32_t queue_id) + uint32_t queue_id, uint32_t inst) { enum hqd_dequeue_request_type type; unsigned long end_jiffies; @@ -571,7 +573,7 @@ static int hqd_sdma_destroy_v11(struct amdgpu_device *adev, void *mqd, static int wave_control_execute_v11(struct amdgpu_device *adev, uint32_t gfx_index_val, - uint32_t sq_cmd) + uint32_t sq_cmd, uint32_t inst) { uint32_t data = 0; @@ -606,6 +608,183 @@ static void set_vm_context_page_table_base_v11(struct amdgpu_device *adev, adev->gfxhub.funcs->setup_vm_pt_regs(adev, vmid, page_table_base); } +/* + * Returns TRAP_EN, EXCP_EN and EXCP_REPLACE. + * + * restore_dbg_registers is ignored here but is a general interface requirement + * for devices that support GFXOFF and where the RLC save/restore list + * does not support hw registers for debugging i.e. the driver has to manually + * initialize the debug mode registers after it has disabled GFX off during the + * debug session. + */ +static uint32_t kgd_gfx_v11_enable_debug_trap(struct amdgpu_device *adev, + bool restore_dbg_registers, + uint32_t vmid) +{ + uint32_t data = 0; + + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1); + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0); + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, 0); + + return data; +} + +/* Returns TRAP_EN, EXCP_EN and EXCP_REPLACE. */ +static uint32_t kgd_gfx_v11_disable_debug_trap(struct amdgpu_device *adev, + bool keep_trap_enabled, + uint32_t vmid) +{ + uint32_t data = 0; + + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, keep_trap_enabled); + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0); + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, 0); + + return data; +} + +static int kgd_gfx_v11_validate_trap_override_request(struct amdgpu_device *adev, + uint32_t trap_override, + uint32_t *trap_mask_supported) +{ + *trap_mask_supported &= KFD_DBG_TRAP_MASK_FP_INVALID | + KFD_DBG_TRAP_MASK_FP_INPUT_DENORMAL | + KFD_DBG_TRAP_MASK_FP_DIVIDE_BY_ZERO | + KFD_DBG_TRAP_MASK_FP_OVERFLOW | + KFD_DBG_TRAP_MASK_FP_UNDERFLOW | + KFD_DBG_TRAP_MASK_FP_INEXACT | + KFD_DBG_TRAP_MASK_INT_DIVIDE_BY_ZERO | + KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH | + KFD_DBG_TRAP_MASK_DBG_MEMORY_VIOLATION; + + if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(11, 0, 4)) + *trap_mask_supported |= KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_START | + KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_END; + + if (trap_override != KFD_DBG_TRAP_OVERRIDE_OR && + trap_override != KFD_DBG_TRAP_OVERRIDE_REPLACE) + return -EPERM; + + return 0; +} + +static uint32_t trap_mask_map_sw_to_hw(uint32_t mask) +{ + uint32_t trap_on_start = (mask & KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_START) ? 1 : 0; + uint32_t trap_on_end = (mask & KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_END) ? 1 : 0; + uint32_t excp_en = mask & (KFD_DBG_TRAP_MASK_FP_INVALID | + KFD_DBG_TRAP_MASK_FP_INPUT_DENORMAL | + KFD_DBG_TRAP_MASK_FP_DIVIDE_BY_ZERO | + KFD_DBG_TRAP_MASK_FP_OVERFLOW | + KFD_DBG_TRAP_MASK_FP_UNDERFLOW | + KFD_DBG_TRAP_MASK_FP_INEXACT | + KFD_DBG_TRAP_MASK_INT_DIVIDE_BY_ZERO | + KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH | + KFD_DBG_TRAP_MASK_DBG_MEMORY_VIOLATION); + uint32_t ret; + + ret = REG_SET_FIELD(0, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, excp_en); + ret = REG_SET_FIELD(ret, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_START, trap_on_start); + ret = REG_SET_FIELD(ret, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_END, trap_on_end); + + return ret; +} + +static uint32_t trap_mask_map_hw_to_sw(uint32_t mask) +{ + uint32_t ret = REG_GET_FIELD(mask, SPI_GDBG_PER_VMID_CNTL, EXCP_EN); + + if (REG_GET_FIELD(mask, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_START)) + ret |= KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_START; + + if (REG_GET_FIELD(mask, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_END)) + ret |= KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_END; + + return ret; +} + +/* Returns TRAP_EN, EXCP_EN and EXCP_REPLACE. */ +static uint32_t kgd_gfx_v11_set_wave_launch_trap_override(struct amdgpu_device *adev, + uint32_t vmid, + uint32_t trap_override, + uint32_t trap_mask_bits, + uint32_t trap_mask_request, + uint32_t *trap_mask_prev, + uint32_t kfd_dbg_trap_cntl_prev) +{ + uint32_t data = 0; + + *trap_mask_prev = trap_mask_map_hw_to_sw(kfd_dbg_trap_cntl_prev); + + data = (trap_mask_bits & trap_mask_request) | (*trap_mask_prev & ~trap_mask_request); + data = trap_mask_map_sw_to_hw(data); + + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1); + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, trap_override); + + return data; +} + +static uint32_t kgd_gfx_v11_set_wave_launch_mode(struct amdgpu_device *adev, + uint8_t wave_launch_mode, + uint32_t vmid) +{ + uint32_t data = 0; + + data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, LAUNCH_MODE, wave_launch_mode); + + return data; +} + +#define TCP_WATCH_STRIDE (regTCP_WATCH1_ADDR_H - regTCP_WATCH0_ADDR_H) +static uint32_t kgd_gfx_v11_set_address_watch(struct amdgpu_device *adev, + uint64_t watch_address, + uint32_t watch_address_mask, + uint32_t watch_id, + uint32_t watch_mode, + uint32_t debug_vmid) +{ + uint32_t watch_address_high; + uint32_t watch_address_low; + uint32_t watch_address_cntl; + + watch_address_cntl = 0; + watch_address_low = lower_32_bits(watch_address); + watch_address_high = upper_32_bits(watch_address) & 0xffff; + + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + MODE, + watch_mode); + + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + MASK, + watch_address_mask >> 7); + + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + VALID, + 1); + + WREG32_RLC((SOC15_REG_OFFSET(GC, 0, regTCP_WATCH0_ADDR_H) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_high); + + WREG32_RLC((SOC15_REG_OFFSET(GC, 0, regTCP_WATCH0_ADDR_L) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_low); + + return watch_address_cntl; +} + +static uint32_t kgd_gfx_v11_clear_address_watch(struct amdgpu_device *adev, + uint32_t watch_id) +{ + return 0; +} + const struct kfd2kgd_calls gfx_v11_kfd2kgd = { .program_sh_mem_settings = program_sh_mem_settings_v11, .set_pasid_vmid_mapping = set_pasid_vmid_mapping_v11, @@ -622,4 +801,11 @@ const struct kfd2kgd_calls gfx_v11_kfd2kgd = { .wave_control_execute = wave_control_execute_v11, .get_atc_vmid_pasid_mapping_info = NULL, .set_vm_context_page_table_base = set_vm_context_page_table_base_v11, + .enable_debug_trap = kgd_gfx_v11_enable_debug_trap, + .disable_debug_trap = kgd_gfx_v11_disable_debug_trap, + .validate_trap_override_request = kgd_gfx_v11_validate_trap_override_request, + .set_wave_launch_trap_override = kgd_gfx_v11_set_wave_launch_trap_override, + .set_wave_launch_mode = kgd_gfx_v11_set_wave_launch_mode, + .set_address_watch = kgd_gfx_v11_set_address_watch, + .clear_address_watch = kgd_gfx_v11_clear_address_watch }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c index e83cb1c09610..6bf448ab3dff 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c @@ -78,7 +78,7 @@ static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmi uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, - uint32_t sh_mem_bases) + uint32_t sh_mem_bases, uint32_t inst) { lock_srbm(adev, 0, 0, 0, vmid); @@ -91,7 +91,7 @@ static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmi } static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, - unsigned int vmid) + unsigned int vmid, uint32_t inst) { /* * We have to assume that there is no outstanding mapping. @@ -114,7 +114,8 @@ static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, return 0; } -static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id) +static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id, + uint32_t inst) { uint32_t mec; uint32_t pipe; @@ -158,7 +159,7 @@ static inline struct cik_sdma_rlc_registers *get_sdma_mqd(void *mqd) static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t __user *wptr, uint32_t wptr_shift, - uint32_t wptr_mask, struct mm_struct *mm) + uint32_t wptr_mask, struct mm_struct *mm, uint32_t inst) { struct cik_mqd *m; uint32_t *mqd_hqd; @@ -202,7 +203,7 @@ static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd, static int kgd_hqd_dump(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, - uint32_t (**dump)[2], uint32_t *n_regs) + uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst) { uint32_t i = 0, reg; #define HQD_N_REGS (35+4) @@ -318,7 +319,7 @@ static int kgd_hqd_sdma_dump(struct amdgpu_device *adev, static bool kgd_hqd_is_occupied(struct amdgpu_device *adev, uint64_t queue_address, uint32_t pipe_id, - uint32_t queue_id) + uint32_t queue_id, uint32_t inst) { uint32_t act; bool retval = false; @@ -358,7 +359,7 @@ static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd) static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, - uint32_t queue_id) + uint32_t queue_id, uint32_t inst) { uint32_t temp; enum hqd_dequeue_request_type type; @@ -494,7 +495,7 @@ static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd, static int kgd_wave_control_execute(struct amdgpu_device *adev, uint32_t gfx_index_val, - uint32_t sq_cmd) + uint32_t sq_cmd, uint32_t inst) { uint32_t data; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c index 870f352837fc..cd06e4a6d1da 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c @@ -72,7 +72,7 @@ static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmi uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, - uint32_t sh_mem_bases) + uint32_t sh_mem_bases, uint32_t inst) { lock_srbm(adev, 0, 0, 0, vmid); @@ -85,7 +85,7 @@ static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmi } static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, - unsigned int vmid) + unsigned int vmid, uint32_t inst) { /* * We have to assume that there is no outstanding mapping. @@ -109,7 +109,8 @@ static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, return 0; } -static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id) +static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id, + uint32_t inst) { uint32_t mec; uint32_t pipe; @@ -153,7 +154,7 @@ static inline struct vi_sdma_mqd *get_sdma_mqd(void *mqd) static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t __user *wptr, uint32_t wptr_shift, - uint32_t wptr_mask, struct mm_struct *mm) + uint32_t wptr_mask, struct mm_struct *mm, uint32_t inst) { struct vi_mqd *m; uint32_t *mqd_hqd; @@ -226,7 +227,7 @@ static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd, static int kgd_hqd_dump(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, - uint32_t (**dump)[2], uint32_t *n_regs) + uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst) { uint32_t i = 0, reg; #define HQD_N_REGS (54+4) @@ -350,7 +351,7 @@ static int kgd_hqd_sdma_dump(struct amdgpu_device *adev, static bool kgd_hqd_is_occupied(struct amdgpu_device *adev, uint64_t queue_address, uint32_t pipe_id, - uint32_t queue_id) + uint32_t queue_id, uint32_t inst) { uint32_t act; bool retval = false; @@ -390,7 +391,7 @@ static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd) static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, - uint32_t queue_id) + uint32_t queue_id, uint32_t inst) { uint32_t temp; enum hqd_dequeue_request_type type; @@ -540,7 +541,7 @@ static bool get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev, static int kgd_wave_control_execute(struct amdgpu_device *adev, uint32_t gfx_index_val, - uint32_t sq_cmd) + uint32_t sq_cmd, uint32_t inst) { uint32_t data = 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c index e92b93557c13..51d93fb13ea3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c @@ -38,6 +38,7 @@ #include "soc15d.h" #include "gfx_v9_0.h" #include "amdgpu_amdkfd_gfx_v9.h" +#include <uapi/linux/kfd_ioctl.h> enum hqd_dequeue_request_type { NO_ACTION = 0, @@ -46,29 +47,29 @@ enum hqd_dequeue_request_type { SAVE_WAVES }; -static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe, - uint32_t queue, uint32_t vmid) +static void kgd_gfx_v9_lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe, + uint32_t queue, uint32_t vmid, uint32_t inst) { mutex_lock(&adev->srbm_mutex); - soc15_grbm_select(adev, mec, pipe, queue, vmid); + soc15_grbm_select(adev, mec, pipe, queue, vmid, GET_INST(GC, inst)); } -static void unlock_srbm(struct amdgpu_device *adev) +static void kgd_gfx_v9_unlock_srbm(struct amdgpu_device *adev, uint32_t inst) { - soc15_grbm_select(adev, 0, 0, 0, 0); + soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, inst)); mutex_unlock(&adev->srbm_mutex); } -static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id, - uint32_t queue_id) +void kgd_gfx_v9_acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id, + uint32_t queue_id, uint32_t inst) { uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); - lock_srbm(adev, mec, pipe, queue_id, 0); + kgd_gfx_v9_lock_srbm(adev, mec, pipe, queue_id, 0, inst); } -static uint64_t get_queue_mask(struct amdgpu_device *adev, +uint64_t kgd_gfx_v9_get_queue_mask(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id) { unsigned int bit = pipe_id * adev->gfx.mec.num_queue_per_pipe + @@ -77,28 +78,28 @@ static uint64_t get_queue_mask(struct amdgpu_device *adev, return 1ull << bit; } -static void release_queue(struct amdgpu_device *adev) +void kgd_gfx_v9_release_queue(struct amdgpu_device *adev, uint32_t inst) { - unlock_srbm(adev); + kgd_gfx_v9_unlock_srbm(adev, inst); } void kgd_gfx_v9_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid, uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, - uint32_t sh_mem_bases) + uint32_t sh_mem_bases, uint32_t inst) { - lock_srbm(adev, 0, 0, 0, vmid); + kgd_gfx_v9_lock_srbm(adev, 0, 0, 0, vmid, inst); - WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_CONFIG), sh_mem_config); - WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_BASES), sh_mem_bases); + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmSH_MEM_CONFIG), sh_mem_config); + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmSH_MEM_BASES), sh_mem_bases); /* APE1 no longer exists on GFX9 */ - unlock_srbm(adev); + kgd_gfx_v9_unlock_srbm(adev, inst); } int kgd_gfx_v9_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, - unsigned int vmid) + unsigned int vmid, uint32_t inst) { /* * We have to assume that there is no outstanding mapping. @@ -156,7 +157,8 @@ int kgd_gfx_v9_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, * but still works */ -int kgd_gfx_v9_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id) +int kgd_gfx_v9_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id, + uint32_t inst) { uint32_t mec; uint32_t pipe; @@ -164,13 +166,13 @@ int kgd_gfx_v9_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id) mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); - lock_srbm(adev, mec, pipe, 0, 0); + kgd_gfx_v9_lock_srbm(adev, mec, pipe, 0, 0, inst); - WREG32_SOC15(GC, 0, mmCPC_INT_CNTL, + WREG32_SOC15(GC, GET_INST(GC, inst), mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK | CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK); - unlock_srbm(adev); + kgd_gfx_v9_unlock_srbm(adev, inst); return 0; } @@ -220,7 +222,8 @@ static inline struct v9_sdma_mqd *get_sdma_mqd(void *mqd) int kgd_gfx_v9_hqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t __user *wptr, uint32_t wptr_shift, - uint32_t wptr_mask, struct mm_struct *mm) + uint32_t wptr_mask, struct mm_struct *mm, + uint32_t inst) { struct v9_mqd *m; uint32_t *mqd_hqd; @@ -228,21 +231,22 @@ int kgd_gfx_v9_hqd_load(struct amdgpu_device *adev, void *mqd, m = get_mqd(mqd); - acquire_queue(adev, pipe_id, queue_id); + kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst); /* HQD registers extend from CP_MQD_BASE_ADDR to CP_HQD_EOP_WPTR_MEM. */ mqd_hqd = &m->cp_mqd_base_addr_lo; - hqd_base = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR); + hqd_base = SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_MQD_BASE_ADDR); for (reg = hqd_base; - reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++) + reg <= SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_HQD_PQ_WPTR_HI); reg++) WREG32_RLC(reg, mqd_hqd[reg - hqd_base]); /* Activate doorbell logic before triggering WPTR poll. */ data = REG_SET_FIELD(m->cp_hqd_pq_doorbell_control, CP_HQD_PQ_DOORBELL_CONTROL, DOORBELL_EN, 1); - WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL), data); + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_HQD_PQ_DOORBELL_CONTROL), + data); if (wptr) { /* Don't read wptr with get_user because the user @@ -271,43 +275,43 @@ int kgd_gfx_v9_hqd_load(struct amdgpu_device *adev, void *mqd, guessed_wptr += m->cp_hqd_pq_wptr_lo & ~(queue_size - 1); guessed_wptr += (uint64_t)m->cp_hqd_pq_wptr_hi << 32; - WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_LO), + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_HQD_PQ_WPTR_LO), lower_32_bits(guessed_wptr)); - WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI), + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_HQD_PQ_WPTR_HI), upper_32_bits(guessed_wptr)); - WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR), + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_HQD_PQ_WPTR_POLL_ADDR), lower_32_bits((uintptr_t)wptr)); - WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI), + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_HQD_PQ_WPTR_POLL_ADDR_HI), upper_32_bits((uintptr_t)wptr)); - WREG32_SOC15(GC, 0, mmCP_PQ_WPTR_POLL_CNTL1, - (uint32_t)get_queue_mask(adev, pipe_id, queue_id)); + WREG32_SOC15(GC, GET_INST(GC, inst), mmCP_PQ_WPTR_POLL_CNTL1, + (uint32_t)kgd_gfx_v9_get_queue_mask(adev, pipe_id, queue_id)); } /* Start the EOP fetcher */ - WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_EOP_RPTR), + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_HQD_EOP_RPTR), REG_SET_FIELD(m->cp_hqd_eop_rptr, CP_HQD_EOP_RPTR, INIT_FETCHER, 1)); data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1); - WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE), data); + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_HQD_ACTIVE), data); - release_queue(adev); + kgd_gfx_v9_release_queue(adev, inst); return 0; } int kgd_gfx_v9_hiq_mqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, - uint32_t doorbell_off) + uint32_t doorbell_off, uint32_t inst) { - struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; + struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[inst].ring; struct v9_mqd *m; uint32_t mec, pipe; int r; m = get_mqd(mqd); - acquire_queue(adev, pipe_id, queue_id); + kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst); mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1; pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec); @@ -315,7 +319,7 @@ int kgd_gfx_v9_hiq_mqd_load(struct amdgpu_device *adev, void *mqd, pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n", mec, pipe, queue_id); - spin_lock(&adev->gfx.kiq.ring_lock); + spin_lock(&adev->gfx.kiq[inst].ring_lock); r = amdgpu_ring_alloc(kiq_ring, 7); if (r) { pr_err("Failed to alloc KIQ (%d).\n", r); @@ -342,15 +346,15 @@ int kgd_gfx_v9_hiq_mqd_load(struct amdgpu_device *adev, void *mqd, amdgpu_ring_commit(kiq_ring); out_unlock: - spin_unlock(&adev->gfx.kiq.ring_lock); - release_queue(adev); + spin_unlock(&adev->gfx.kiq[inst].ring_lock); + kgd_gfx_v9_release_queue(adev, inst); return r; } int kgd_gfx_v9_hqd_dump(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, - uint32_t (**dump)[2], uint32_t *n_regs) + uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst) { uint32_t i = 0, reg; #define HQD_N_REGS 56 @@ -365,13 +369,13 @@ int kgd_gfx_v9_hqd_dump(struct amdgpu_device *adev, if (*dump == NULL) return -ENOMEM; - acquire_queue(adev, pipe_id, queue_id); + kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst); - for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR); - reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++) + for (reg = SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_MQD_BASE_ADDR); + reg <= SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_HQD_PQ_WPTR_HI); reg++) DUMP_REG(reg); - release_queue(adev); + kgd_gfx_v9_release_queue(adev, inst); WARN_ON_ONCE(i != HQD_N_REGS); *n_regs = i; @@ -481,23 +485,23 @@ static int kgd_hqd_sdma_dump(struct amdgpu_device *adev, bool kgd_gfx_v9_hqd_is_occupied(struct amdgpu_device *adev, uint64_t queue_address, uint32_t pipe_id, - uint32_t queue_id) + uint32_t queue_id, uint32_t inst) { uint32_t act; bool retval = false; uint32_t low, high; - acquire_queue(adev, pipe_id, queue_id); - act = RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE); + kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst); + act = RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_ACTIVE); if (act) { low = lower_32_bits(queue_address >> 8); high = upper_32_bits(queue_address >> 8); - if (low == RREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE) && - high == RREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE_HI)) + if (low == RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_PQ_BASE) && + high == RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_PQ_BASE_HI)) retval = true; } - release_queue(adev); + kgd_gfx_v9_release_queue(adev, inst); return retval; } @@ -522,7 +526,7 @@ static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd) int kgd_gfx_v9_hqd_destroy(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, - uint32_t queue_id) + uint32_t queue_id, uint32_t inst) { enum hqd_dequeue_request_type type; unsigned long end_jiffies; @@ -532,10 +536,10 @@ int kgd_gfx_v9_hqd_destroy(struct amdgpu_device *adev, void *mqd, if (amdgpu_in_reset(adev)) return -EIO; - acquire_queue(adev, pipe_id, queue_id); + kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst); if (m->cp_hqd_vmid == 0) - WREG32_FIELD15_RLC(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0); + WREG32_FIELD15_RLC(GC, GET_INST(GC, inst), RLC_CP_SCHEDULERS, scheduler1, 0); switch (reset_type) { case KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN: @@ -552,22 +556,22 @@ int kgd_gfx_v9_hqd_destroy(struct amdgpu_device *adev, void *mqd, break; } - WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_DEQUEUE_REQUEST), type); + WREG32_RLC(SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_HQD_DEQUEUE_REQUEST), type); end_jiffies = (utimeout * HZ / 1000) + jiffies; while (true) { - temp = RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE); + temp = RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_ACTIVE); if (!(temp & CP_HQD_ACTIVE__ACTIVE_MASK)) break; if (time_after(jiffies, end_jiffies)) { pr_err("cp queue preemption time out.\n"); - release_queue(adev); + kgd_gfx_v9_release_queue(adev, inst); return -ETIME; } usleep_range(500, 1000); } - release_queue(adev); + kgd_gfx_v9_release_queue(adev, inst); return 0; } @@ -624,14 +628,14 @@ bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev, int kgd_gfx_v9_wave_control_execute(struct amdgpu_device *adev, uint32_t gfx_index_val, - uint32_t sq_cmd) + uint32_t sq_cmd, uint32_t inst) { uint32_t data = 0; mutex_lock(&adev->grbm_idx_mutex); - WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_INDEX, gfx_index_val); - WREG32_SOC15(GC, 0, mmSQ_CMD, sq_cmd); + WREG32_SOC15_RLC_SHADOW(GC, GET_INST(GC, inst), mmGRBM_GFX_INDEX, gfx_index_val); + WREG32_SOC15(GC, GET_INST(GC, inst), mmSQ_CMD, sq_cmd); data = REG_SET_FIELD(data, GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES, 1); @@ -640,12 +644,271 @@ int kgd_gfx_v9_wave_control_execute(struct amdgpu_device *adev, data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SE_BROADCAST_WRITES, 1); - WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_INDEX, data); + WREG32_SOC15_RLC_SHADOW(GC, GET_INST(GC, inst), mmGRBM_GFX_INDEX, data); mutex_unlock(&adev->grbm_idx_mutex); return 0; } +/* + * GFX9 helper for wave launch stall requirements on debug trap setting. + * + * vmid: + * Target VMID to stall/unstall. + * + * stall: + * 0-unstall wave launch (enable), 1-stall wave launch (disable). + * After wavefront launch has been stalled, allocated waves must drain from + * SPI in order for debug trap settings to take effect on those waves. + * This is roughly a ~96 clock cycle wait on SPI where a read on + * SPI_GDBG_WAVE_CNTL translates to ~32 clock cycles. + * KGD_GFX_V9_WAVE_LAUNCH_SPI_DRAIN_LATENCY indicates the number of reads required. + * + * NOTE: We can afford to clear the entire STALL_VMID field on unstall + * because GFX9.4.1 cannot support multi-process debugging due to trap + * configuration and masking being limited to global scope. Always assume + * single process conditions. + */ +#define KGD_GFX_V9_WAVE_LAUNCH_SPI_DRAIN_LATENCY 3 +void kgd_gfx_v9_set_wave_launch_stall(struct amdgpu_device *adev, + uint32_t vmid, + bool stall) +{ + int i; + uint32_t data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL)); + + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1)) + data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_VMID, + stall ? 1 << vmid : 0); + else + data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_RA, + stall ? 1 : 0); + + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data); + + if (!stall) + return; + + for (i = 0; i < KGD_GFX_V9_WAVE_LAUNCH_SPI_DRAIN_LATENCY; i++) + RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL)); +} + +/* + * restore_dbg_registers is ignored here but is a general interface requirement + * for devices that support GFXOFF and where the RLC save/restore list + * does not support hw registers for debugging i.e. the driver has to manually + * initialize the debug mode registers after it has disabled GFX off during the + * debug session. + */ +uint32_t kgd_gfx_v9_enable_debug_trap(struct amdgpu_device *adev, + bool restore_dbg_registers, + uint32_t vmid) +{ + mutex_lock(&adev->grbm_idx_mutex); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true); + + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, false); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; +} + +/* + * keep_trap_enabled is ignored here but is a general interface requirement + * for devices that support multi-process debugging where the performance + * overhead from trap temporary setup needs to be bypassed when the debug + * session has ended. + */ +uint32_t kgd_gfx_v9_disable_debug_trap(struct amdgpu_device *adev, + bool keep_trap_enabled, + uint32_t vmid) +{ + mutex_lock(&adev->grbm_idx_mutex); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true); + + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, false); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; +} + +int kgd_gfx_v9_validate_trap_override_request(struct amdgpu_device *adev, + uint32_t trap_override, + uint32_t *trap_mask_supported) +{ + *trap_mask_supported &= KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH; + + /* The SPI_GDBG_TRAP_MASK register is global and affects all + * processes. Only allow OR-ing the address-watch bit, since + * this only affects processes under the debugger. Other bits + * should stay 0 to avoid the debugger interfering with other + * processes. + */ + if (trap_override != KFD_DBG_TRAP_OVERRIDE_OR) + return -EINVAL; + + return 0; +} + +uint32_t kgd_gfx_v9_set_wave_launch_trap_override(struct amdgpu_device *adev, + uint32_t vmid, + uint32_t trap_override, + uint32_t trap_mask_bits, + uint32_t trap_mask_request, + uint32_t *trap_mask_prev, + uint32_t kfd_dbg_cntl_prev) +{ + uint32_t data, wave_cntl_prev; + + mutex_lock(&adev->grbm_idx_mutex); + + wave_cntl_prev = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL)); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true); + + data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK)); + *trap_mask_prev = REG_GET_FIELD(data, SPI_GDBG_TRAP_MASK, EXCP_EN); + + trap_mask_bits = (trap_mask_bits & trap_mask_request) | + (*trap_mask_prev & ~trap_mask_request); + + data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK, EXCP_EN, trap_mask_bits); + data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK, REPLACE, trap_override); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), data); + + /* We need to preserve wave launch mode stall settings. */ + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), wave_cntl_prev); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; +} + +uint32_t kgd_gfx_v9_set_wave_launch_mode(struct amdgpu_device *adev, + uint8_t wave_launch_mode, + uint32_t vmid) +{ + uint32_t data = 0; + bool is_mode_set = !!wave_launch_mode; + + mutex_lock(&adev->grbm_idx_mutex); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true); + + data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2, + VMID_MASK, is_mode_set ? 1 << vmid : 0); + data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2, + MODE, is_mode_set ? wave_launch_mode : 0); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL2), data); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, false); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; +} + +#define TCP_WATCH_STRIDE (mmTCP_WATCH1_ADDR_H - mmTCP_WATCH0_ADDR_H) +uint32_t kgd_gfx_v9_set_address_watch(struct amdgpu_device *adev, + uint64_t watch_address, + uint32_t watch_address_mask, + uint32_t watch_id, + uint32_t watch_mode, + uint32_t debug_vmid) +{ + uint32_t watch_address_high; + uint32_t watch_address_low; + uint32_t watch_address_cntl; + + watch_address_cntl = 0; + + watch_address_low = lower_32_bits(watch_address); + watch_address_high = upper_32_bits(watch_address) & 0xffff; + + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + VMID, + debug_vmid); + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + MODE, + watch_mode); + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + MASK, + watch_address_mask >> 6); + + /* Turning off this watch point until we set all the registers */ + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + VALID, + 0); + + WREG32_RLC((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_CNTL) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_cntl); + + WREG32_RLC((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_ADDR_H) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_high); + + WREG32_RLC((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_ADDR_L) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_low); + + /* Enable the watch point */ + watch_address_cntl = REG_SET_FIELD(watch_address_cntl, + TCP_WATCH0_CNTL, + VALID, + 1); + + WREG32_RLC((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_CNTL) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_cntl); + + return 0; +} + +uint32_t kgd_gfx_v9_clear_address_watch(struct amdgpu_device *adev, + uint32_t watch_id) +{ + uint32_t watch_address_cntl; + + watch_address_cntl = 0; + + WREG32_RLC((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_CNTL) + + (watch_id * TCP_WATCH_STRIDE)), + watch_address_cntl); + + return 0; +} + +/* kgd_gfx_v9_get_iq_wait_times: Returns the mmCP_IQ_WAIT_TIME1/2 values + * The values read are: + * ib_offload_wait_time -- Wait Count for Indirect Buffer Offloads. + * atomic_offload_wait_time -- Wait Count for L2 and GDS Atomics Offloads. + * wrm_offload_wait_time -- Wait Count for WAIT_REG_MEM Offloads. + * gws_wait_time -- Wait Count for Global Wave Syncs. + * que_sleep_wait_time -- Wait Count for Dequeue Retry. + * sch_wave_wait_time -- Wait Count for Scheduling Wave Message. + * sem_rearm_wait_time -- Wait Count for Semaphore re-arm. + * deq_retry_wait_time -- Wait Count for Global Wave Syncs. + */ +void kgd_gfx_v9_get_iq_wait_times(struct amdgpu_device *adev, + uint32_t *wait_times) + +{ + *wait_times = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_IQ_WAIT_TIME2)); +} + void kgd_gfx_v9_set_vm_context_page_table_base(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { @@ -682,10 +945,11 @@ static void unlock_spi_csq_mutexes(struct amdgpu_device *adev) * @queue_idx: Index of queue in the queue-map bit-field * @wave_cnt: Output parameter updated with number of waves in flight * @vmid: Output parameter updated with VMID of queue whose wave count - * is being collected + * is being collected + * @inst: xcc's instance number on a multi-XCC setup */ static void get_wave_count(struct amdgpu_device *adev, int queue_idx, - int *wave_cnt, int *vmid) + int *wave_cnt, int *vmid, uint32_t inst) { int pipe_idx; int queue_slot; @@ -700,12 +964,12 @@ static void get_wave_count(struct amdgpu_device *adev, int queue_idx, *wave_cnt = 0; pipe_idx = queue_idx / adev->gfx.mec.num_queue_per_pipe; queue_slot = queue_idx % adev->gfx.mec.num_queue_per_pipe; - soc15_grbm_select(adev, 1, pipe_idx, queue_slot, 0); - reg_val = RREG32_SOC15_IP(GC, SOC15_REG_OFFSET(GC, 0, mmSPI_CSQ_WF_ACTIVE_COUNT_0) + + soc15_grbm_select(adev, 1, pipe_idx, queue_slot, 0, inst); + reg_val = RREG32_SOC15_IP(GC, SOC15_REG_OFFSET(GC, inst, mmSPI_CSQ_WF_ACTIVE_COUNT_0) + queue_slot); *wave_cnt = reg_val & SPI_CSQ_WF_ACTIVE_COUNT_0__COUNT_MASK; if (*wave_cnt != 0) - *vmid = (RREG32_SOC15(GC, 0, mmCP_HQD_VMID) & + *vmid = (RREG32_SOC15(GC, inst, mmCP_HQD_VMID) & CP_HQD_VMID__VMID_MASK) >> CP_HQD_VMID__VMID__SHIFT; } @@ -718,9 +982,10 @@ static void get_wave_count(struct amdgpu_device *adev, int queue_idx, * @adev: Handle of device from which to get number of waves in flight * @pasid: Identifies the process for which this query call is invoked * @pasid_wave_cnt: Output parameter updated with number of waves in flight that - * belong to process with given pasid + * belong to process with given pasid * @max_waves_per_cu: Output parameter updated with maximum number of waves - * possible per Compute Unit + * possible per Compute Unit + * @inst: xcc's instance number on a multi-XCC setup * * Note: It's possible that the device has too many queues (oversubscription) * in which case a VMID could be remapped to a different PASID. This could lead @@ -756,7 +1021,7 @@ static void get_wave_count(struct amdgpu_device *adev, int queue_idx, * Reading registers referenced above involves programming GRBM appropriately */ void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev, int pasid, - int *pasid_wave_cnt, int *max_waves_per_cu) + int *pasid_wave_cnt, int *max_waves_per_cu, uint32_t inst) { int qidx; int vmid; @@ -772,13 +1037,13 @@ void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev, int pasid, DECLARE_BITMAP(cp_queue_bitmap, KGD_MAX_QUEUES); lock_spi_csq_mutexes(adev); - soc15_grbm_select(adev, 1, 0, 0, 0); + soc15_grbm_select(adev, 1, 0, 0, 0, inst); /* * Iterate through the shader engines and arrays of the device * to get number of waves in flight */ - bitmap_complement(cp_queue_bitmap, adev->gfx.mec.queue_bitmap, + bitmap_complement(cp_queue_bitmap, adev->gfx.mec_bitmap[0].queue_bitmap, KGD_MAX_QUEUES); max_queue_cnt = adev->gfx.mec.num_pipe_per_mec * adev->gfx.mec.num_queue_per_pipe; @@ -787,8 +1052,8 @@ void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev, int pasid, for (se_idx = 0; se_idx < se_cnt; se_idx++) { for (sh_idx = 0; sh_idx < sh_cnt; sh_idx++) { - amdgpu_gfx_select_se_sh(adev, se_idx, sh_idx, 0xffffffff); - queue_map = RREG32_SOC15(GC, 0, mmSPI_CSQ_WF_ACTIVE_STATUS); + amdgpu_gfx_select_se_sh(adev, se_idx, sh_idx, 0xffffffff, inst); + queue_map = RREG32_SOC15(GC, inst, mmSPI_CSQ_WF_ACTIVE_STATUS); /* * Assumption: queue map encodes following schema: four @@ -808,10 +1073,11 @@ void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev, int pasid, continue; /* Get number of waves in flight and aggregate them */ - get_wave_count(adev, qidx, &wave_cnt, &vmid); + get_wave_count(adev, qidx, &wave_cnt, &vmid, + inst); if (wave_cnt != 0) { pasid_tmp = - RREG32(SOC15_REG_OFFSET(OSSSYS, 0, + RREG32(SOC15_REG_OFFSET(OSSSYS, inst, mmIH_VMID_0_LUT) + vmid); if (pasid_tmp == pasid) vmid_wave_cnt += wave_cnt; @@ -820,8 +1086,8 @@ void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev, int pasid, } } - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); - soc15_grbm_select(adev, 0, 0, 0, 0); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, inst); + soc15_grbm_select(adev, 0, 0, 0, 0, inst); unlock_spi_csq_mutexes(adev); /* Update the output parameters and return */ @@ -830,28 +1096,51 @@ void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev, int pasid, adev->gfx.cu_info.max_waves_per_simd; } +void kgd_gfx_v9_build_grace_period_packet_info(struct amdgpu_device *adev, + uint32_t wait_times, + uint32_t grace_period, + uint32_t *reg_offset, + uint32_t *reg_data) +{ + *reg_data = wait_times; + + /* + * The CP cannont handle a 0 grace period input and will result in + * an infinite grace period being set so set to 1 to prevent this. + */ + if (grace_period == 0) + grace_period = 1; + + *reg_data = REG_SET_FIELD(*reg_data, + CP_IQ_WAIT_TIME2, + SCH_WAVE, + grace_period); + + *reg_offset = SOC15_REG_OFFSET(GC, 0, mmCP_IQ_WAIT_TIME2); +} + void kgd_gfx_v9_program_trap_handler_settings(struct amdgpu_device *adev, - uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr) + uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr, uint32_t inst) { - lock_srbm(adev, 0, 0, 0, vmid); + kgd_gfx_v9_lock_srbm(adev, 0, 0, 0, vmid, inst); /* * Program TBA registers */ - WREG32_SOC15(GC, 0, mmSQ_SHADER_TBA_LO, + WREG32_SOC15(GC, GET_INST(GC, inst), mmSQ_SHADER_TBA_LO, lower_32_bits(tba_addr >> 8)); - WREG32_SOC15(GC, 0, mmSQ_SHADER_TBA_HI, + WREG32_SOC15(GC, GET_INST(GC, inst), mmSQ_SHADER_TBA_HI, upper_32_bits(tba_addr >> 8)); /* * Program TMA registers */ - WREG32_SOC15(GC, 0, mmSQ_SHADER_TMA_LO, + WREG32_SOC15(GC, GET_INST(GC, inst), mmSQ_SHADER_TMA_LO, lower_32_bits(tma_addr >> 8)); - WREG32_SOC15(GC, 0, mmSQ_SHADER_TMA_HI, + WREG32_SOC15(GC, GET_INST(GC, inst), mmSQ_SHADER_TMA_HI, upper_32_bits(tma_addr >> 8)); - unlock_srbm(adev); + kgd_gfx_v9_unlock_srbm(adev, inst); } const struct kfd2kgd_calls gfx_v9_kfd2kgd = { @@ -871,6 +1160,15 @@ const struct kfd2kgd_calls gfx_v9_kfd2kgd = { .get_atc_vmid_pasid_mapping_info = kgd_gfx_v9_get_atc_vmid_pasid_mapping_info, .set_vm_context_page_table_base = kgd_gfx_v9_set_vm_context_page_table_base, + .enable_debug_trap = kgd_gfx_v9_enable_debug_trap, + .disable_debug_trap = kgd_gfx_v9_disable_debug_trap, + .validate_trap_override_request = kgd_gfx_v9_validate_trap_override_request, + .set_wave_launch_trap_override = kgd_gfx_v9_set_wave_launch_trap_override, + .set_wave_launch_mode = kgd_gfx_v9_set_wave_launch_mode, + .set_address_watch = kgd_gfx_v9_set_address_watch, + .clear_address_watch = kgd_gfx_v9_clear_address_watch, + .get_iq_wait_times = kgd_gfx_v9_get_iq_wait_times, + .build_grace_period_packet_info = kgd_gfx_v9_build_grace_period_packet_info, .get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy, .program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings, }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h index c7ed3bc9053c..5f54bff0db49 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h @@ -20,41 +20,81 @@ * OTHER DEALINGS IN THE SOFTWARE. */ - - void kgd_gfx_v9_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid, uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit, - uint32_t sh_mem_bases); + uint32_t sh_mem_bases, uint32_t inst); int kgd_gfx_v9_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid, - unsigned int vmid); -int kgd_gfx_v9_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id); + unsigned int vmid, uint32_t inst); +int kgd_gfx_v9_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id, + uint32_t inst); int kgd_gfx_v9_hqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t __user *wptr, uint32_t wptr_shift, uint32_t wptr_mask, - struct mm_struct *mm); + struct mm_struct *mm, uint32_t inst); int kgd_gfx_v9_hiq_mqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, - uint32_t doorbell_off); + uint32_t doorbell_off, uint32_t inst); int kgd_gfx_v9_hqd_dump(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, - uint32_t (**dump)[2], uint32_t *n_regs); + uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst); bool kgd_gfx_v9_hqd_is_occupied(struct amdgpu_device *adev, uint64_t queue_address, uint32_t pipe_id, - uint32_t queue_id); + uint32_t queue_id, uint32_t inst); int kgd_gfx_v9_hqd_destroy(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int utimeout, uint32_t pipe_id, - uint32_t queue_id); + uint32_t queue_id, uint32_t inst); int kgd_gfx_v9_wave_control_execute(struct amdgpu_device *adev, uint32_t gfx_index_val, - uint32_t sq_cmd); + uint32_t sq_cmd, uint32_t inst); bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev, uint8_t vmid, uint16_t *p_pasid); - void kgd_gfx_v9_set_vm_context_page_table_base(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base); void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev, int pasid, - int *pasid_wave_cnt, int *max_waves_per_cu); + int *pasid_wave_cnt, int *max_waves_per_cu, uint32_t inst); void kgd_gfx_v9_program_trap_handler_settings(struct amdgpu_device *adev, - uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr); + uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr, + uint32_t inst); +void kgd_gfx_v9_acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id, + uint32_t queue_id, uint32_t inst); +uint64_t kgd_gfx_v9_get_queue_mask(struct amdgpu_device *adev, + uint32_t pipe_id, uint32_t queue_id); +void kgd_gfx_v9_release_queue(struct amdgpu_device *adev, uint32_t inst); +void kgd_gfx_v9_set_wave_launch_stall(struct amdgpu_device *adev, + uint32_t vmid, + bool stall); +uint32_t kgd_gfx_v9_enable_debug_trap(struct amdgpu_device *adev, + bool restore_dbg_registers, + uint32_t vmid); +uint32_t kgd_gfx_v9_disable_debug_trap(struct amdgpu_device *adev, + bool keep_trap_enabled, + uint32_t vmid); +int kgd_gfx_v9_validate_trap_override_request(struct amdgpu_device *adev, + uint32_t trap_override, + uint32_t *trap_mask_supported); +uint32_t kgd_gfx_v9_set_wave_launch_mode(struct amdgpu_device *adev, + uint8_t wave_launch_mode, + uint32_t vmid); +uint32_t kgd_gfx_v9_set_wave_launch_trap_override(struct amdgpu_device *adev, + uint32_t vmid, + uint32_t trap_override, + uint32_t trap_mask_bits, + uint32_t trap_mask_request, + uint32_t *trap_mask_prev, + uint32_t kfd_dbg_trap_cntl_prev); +uint32_t kgd_gfx_v9_set_address_watch(struct amdgpu_device *adev, + uint64_t watch_address, + uint32_t watch_address_mask, + uint32_t watch_id, + uint32_t watch_mode, + uint32_t debug_vmid); +uint32_t kgd_gfx_v9_clear_address_watch(struct amdgpu_device *adev, + uint32_t watch_id); +void kgd_gfx_v9_get_iq_wait_times(struct amdgpu_device *adev, uint32_t *wait_times); +void kgd_gfx_v9_build_grace_period_packet_info(struct amdgpu_device *adev, + uint32_t wait_times, + uint32_t grace_period, + uint32_t *reg_offset, + uint32_t *reg_data); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index 83a83ced2439..f61527b800e6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -35,7 +35,9 @@ #include "amdgpu_dma_buf.h" #include <uapi/linux/kfd_ioctl.h> #include "amdgpu_xgmi.h" +#include "kfd_priv.h" #include "kfd_smi_events.h" +#include <drm/ttm/ttm_tt.h> /* Userptr restore delay, just long enough to allow consecutive VM * changes to accumulate @@ -110,13 +112,16 @@ void amdgpu_amdkfd_gpuvm_init_mem_limits(void) struct sysinfo si; uint64_t mem; + if (kfd_mem_limit.max_system_mem_limit) + return; + si_meminfo(&si); mem = si.freeram - si.freehigh; mem *= si.mem_unit; spin_lock_init(&kfd_mem_limit.mem_limit_lock); kfd_mem_limit.max_system_mem_limit = mem - (mem >> 4); - kfd_mem_limit.max_ttm_mem_limit = (mem >> 1) - (mem >> 3); + kfd_mem_limit.max_ttm_mem_limit = ttm_tt_pages_limit() << PAGE_SHIFT; pr_debug("Kernel memory limit %lluM, TTM limit %lluM\n", (kfd_mem_limit.max_system_mem_limit >> 20), (kfd_mem_limit.max_ttm_mem_limit >> 20)); @@ -148,16 +153,20 @@ void amdgpu_amdkfd_reserve_system_mem(uint64_t size) * @size: Size of buffer, in bytes, encapsulated by B0. This should be * equivalent to amdgpu_bo_size(BO) * @alloc_flag: Flag used in allocating a BO as noted above + * @xcp_id: xcp_id is used to get xcp from xcp manager, one xcp is + * managed as one compute node in driver for app * - * Return: returns -ENOMEM in case of error, ZERO otherwise + * Return: + * returns -ENOMEM in case of error, ZERO otherwise */ int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, - uint64_t size, u32 alloc_flag) + uint64_t size, u32 alloc_flag, int8_t xcp_id) { uint64_t reserved_for_pt = ESTIMATE_PT_SIZE(amdgpu_amdkfd_total_mem_size); size_t system_mem_needed, ttm_mem_needed, vram_needed; int ret = 0; + uint64_t vram_size = 0; system_mem_needed = 0; ttm_mem_needed = 0; @@ -172,6 +181,17 @@ int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, * 2M BO chunk. */ vram_needed = size; + /* + * For GFX 9.4.3, get the VRAM size from XCP structs + */ + if (WARN_ONCE(xcp_id < 0, "invalid XCP ID %d", xcp_id)) + return -EINVAL; + + vram_size = KFD_XCP_MEMORY_SIZE(adev, xcp_id); + if (adev->gmc.is_app_apu) { + system_mem_needed = size; + ttm_mem_needed = size; + } } else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) { system_mem_needed = size; } else if (!(alloc_flag & @@ -191,8 +211,8 @@ int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, kfd_mem_limit.max_system_mem_limit && !no_system_mem_limit) || (kfd_mem_limit.ttm_mem_used + ttm_mem_needed > kfd_mem_limit.max_ttm_mem_limit) || - (adev && adev->kfd.vram_used + vram_needed > - adev->gmc.real_vram_size - reserved_for_pt)) { + (adev && xcp_id >= 0 && adev->kfd.vram_used[xcp_id] + vram_needed > + vram_size - reserved_for_pt)) { ret = -ENOMEM; goto release; } @@ -202,9 +222,11 @@ int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, */ WARN_ONCE(vram_needed && !adev, "adev reference can't be null when vram is used"); - if (adev) { - adev->kfd.vram_used += vram_needed; - adev->kfd.vram_used_aligned += ALIGN(vram_needed, VRAM_AVAILABLITY_ALIGN); + if (adev && xcp_id >= 0) { + adev->kfd.vram_used[xcp_id] += vram_needed; + adev->kfd.vram_used_aligned[xcp_id] += adev->gmc.is_app_apu ? + vram_needed : + ALIGN(vram_needed, VRAM_AVAILABLITY_ALIGN); } kfd_mem_limit.system_mem_used += system_mem_needed; kfd_mem_limit.ttm_mem_used += ttm_mem_needed; @@ -215,7 +237,7 @@ release: } void amdgpu_amdkfd_unreserve_mem_limit(struct amdgpu_device *adev, - uint64_t size, u32 alloc_flag) + uint64_t size, u32 alloc_flag, int8_t xcp_id) { spin_lock(&kfd_mem_limit.mem_limit_lock); @@ -225,9 +247,19 @@ void amdgpu_amdkfd_unreserve_mem_limit(struct amdgpu_device *adev, } else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) { WARN_ONCE(!adev, "adev reference can't be null when alloc mem flags vram is set"); + if (WARN_ONCE(xcp_id < 0, "invalid XCP ID %d", xcp_id)) + goto release; + if (adev) { - adev->kfd.vram_used -= size; - adev->kfd.vram_used_aligned -= ALIGN(size, VRAM_AVAILABLITY_ALIGN); + adev->kfd.vram_used[xcp_id] -= size; + if (adev->gmc.is_app_apu) { + adev->kfd.vram_used_aligned[xcp_id] -= size; + kfd_mem_limit.system_mem_used -= size; + kfd_mem_limit.ttm_mem_used -= size; + } else { + adev->kfd.vram_used_aligned[xcp_id] -= + ALIGN(size, VRAM_AVAILABLITY_ALIGN); + } } } else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) { kfd_mem_limit.system_mem_used -= size; @@ -237,8 +269,8 @@ void amdgpu_amdkfd_unreserve_mem_limit(struct amdgpu_device *adev, pr_err("%s: Invalid BO type %#x\n", __func__, alloc_flag); goto release; } - WARN_ONCE(adev && adev->kfd.vram_used < 0, - "KFD VRAM memory accounting unbalanced"); + WARN_ONCE(adev && xcp_id >= 0 && adev->kfd.vram_used[xcp_id] < 0, + "KFD VRAM memory accounting unbalanced for xcp: %d", xcp_id); WARN_ONCE(kfd_mem_limit.ttm_mem_used < 0, "KFD TTM memory accounting unbalanced"); WARN_ONCE(kfd_mem_limit.system_mem_used < 0, @@ -254,14 +286,16 @@ void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo) u32 alloc_flags = bo->kfd_bo->alloc_flags; u64 size = amdgpu_bo_size(bo); - amdgpu_amdkfd_unreserve_mem_limit(adev, size, alloc_flags); + amdgpu_amdkfd_unreserve_mem_limit(adev, size, alloc_flags, + bo->xcp_id); kfree(bo->kfd_bo); } /** - * @create_dmamap_sg_bo: Creates a amdgpu_bo object to reflect information + * create_dmamap_sg_bo() - Creates a amdgpu_bo object to reflect information * about USERPTR or DOOREBELL or MMIO BO. + * * @adev: Device for which dmamap BO is being created * @mem: BO of peer device that is being DMA mapped. Provides parameters * in building the dmamap BO @@ -285,7 +319,7 @@ create_dmamap_sg_bo(struct amdgpu_device *adev, ret = amdgpu_gem_object_create(adev, mem->bo->tbo.base.size, 1, AMDGPU_GEM_DOMAIN_CPU, AMDGPU_GEM_CREATE_PREEMPTIBLE | flags, - ttm_bo_type_sg, mem->bo->tbo.base.resv, &gem_obj); + ttm_bo_type_sg, mem->bo->tbo.base.resv, &gem_obj, 0); amdgpu_bo_unreserve(mem->bo); @@ -527,6 +561,12 @@ kfd_mem_dmamap_dmabuf(struct kfd_mem_attachment *attachment) { struct ttm_operation_ctx ctx = {.interruptible = true}; struct amdgpu_bo *bo = attachment->bo_va->base.bo; + int ret; + + amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU); + ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); + if (ret) + return ret; amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT); return ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); @@ -659,11 +699,10 @@ kfd_mem_dmaunmap_userptr(struct kgd_mem *mem, static void kfd_mem_dmaunmap_dmabuf(struct kfd_mem_attachment *attachment) { - struct ttm_operation_ctx ctx = {.interruptible = true}; - struct amdgpu_bo *bo = attachment->bo_va->base.bo; - - amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU); - ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); + /* This is a no-op. We don't want to trigger eviction fences when + * unmapping DMABufs. Therefore the invalidation (moving to system + * domain) is done in kfd_mem_dmamap_dmabuf. + */ } /** @@ -804,7 +843,7 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem, * if peer device has large BAR. In contrast, access over xGMI is * allowed for both small and large BAR configurations of peer device */ - if ((adev != bo_adev) && + if ((adev != bo_adev && !adev->gmc.is_app_apu) && ((mem->domain == AMDGPU_GEM_DOMAIN_VRAM) || (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL) || (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP))) { @@ -1599,23 +1638,42 @@ out_unlock: return ret; } -size_t amdgpu_amdkfd_get_available_memory(struct amdgpu_device *adev) +size_t amdgpu_amdkfd_get_available_memory(struct amdgpu_device *adev, + uint8_t xcp_id) { uint64_t reserved_for_pt = ESTIMATE_PT_SIZE(amdgpu_amdkfd_total_mem_size); ssize_t available; + uint64_t vram_available, system_mem_available, ttm_mem_available; spin_lock(&kfd_mem_limit.mem_limit_lock); - available = adev->gmc.real_vram_size - - adev->kfd.vram_used_aligned + vram_available = KFD_XCP_MEMORY_SIZE(adev, xcp_id) + - adev->kfd.vram_used_aligned[xcp_id] - atomic64_read(&adev->vram_pin_size) - reserved_for_pt; + + if (adev->gmc.is_app_apu) { + system_mem_available = no_system_mem_limit ? + kfd_mem_limit.max_system_mem_limit : + kfd_mem_limit.max_system_mem_limit - + kfd_mem_limit.system_mem_used; + + ttm_mem_available = kfd_mem_limit.max_ttm_mem_limit - + kfd_mem_limit.ttm_mem_used; + + available = min3(system_mem_available, ttm_mem_available, + vram_available); + available = ALIGN_DOWN(available, PAGE_SIZE); + } else { + available = ALIGN_DOWN(vram_available, VRAM_AVAILABLITY_ALIGN); + } + spin_unlock(&kfd_mem_limit.mem_limit_lock); if (available < 0) available = 0; - return ALIGN_DOWN(available, VRAM_AVAILABLITY_ALIGN); + return available; } int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( @@ -1624,6 +1682,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( uint64_t *offset, uint32_t flags, bool criu_resume) { struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv); + struct amdgpu_fpriv *fpriv = container_of(avm, struct amdgpu_fpriv, vm); enum ttm_bo_type bo_type = ttm_bo_type_device; struct sg_table *sg = NULL; uint64_t user_addr = 0; @@ -1631,6 +1690,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( struct drm_gem_object *gobj = NULL; u32 domain, alloc_domain; uint64_t aligned_size; + int8_t xcp_id = -1; u64 alloc_flags; int ret; @@ -1639,9 +1699,17 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( */ if (flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) { domain = alloc_domain = AMDGPU_GEM_DOMAIN_VRAM; - alloc_flags = AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE; - alloc_flags |= (flags & KFD_IOC_ALLOC_MEM_FLAGS_PUBLIC) ? + + if (adev->gmc.is_app_apu) { + domain = AMDGPU_GEM_DOMAIN_GTT; + alloc_domain = AMDGPU_GEM_DOMAIN_GTT; + alloc_flags = 0; + } else { + alloc_flags = AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE; + alloc_flags |= (flags & KFD_IOC_ALLOC_MEM_FLAGS_PUBLIC) ? AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED : 0; + } + xcp_id = fpriv->xcp_id == ~0 ? 0 : fpriv->xcp_id; } else if (flags & KFD_IOC_ALLOC_MEM_FLAGS_GTT) { domain = alloc_domain = AMDGPU_GEM_DOMAIN_GTT; alloc_flags = 0; @@ -1693,17 +1761,19 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( amdgpu_sync_create(&(*mem)->sync); - ret = amdgpu_amdkfd_reserve_mem_limit(adev, aligned_size, flags); + ret = amdgpu_amdkfd_reserve_mem_limit(adev, aligned_size, flags, + xcp_id); if (ret) { pr_debug("Insufficient memory\n"); goto err_reserve_limit; } - pr_debug("\tcreate BO VA 0x%llx size 0x%llx domain %s\n", - va, (*mem)->aql_queue ? size << 1 : size, domain_string(alloc_domain)); + pr_debug("\tcreate BO VA 0x%llx size 0x%llx domain %s xcp_id %d\n", + va, (*mem)->aql_queue ? size << 1 : size, + domain_string(alloc_domain), xcp_id); ret = amdgpu_gem_object_create(adev, aligned_size, 1, alloc_domain, alloc_flags, - bo_type, NULL, &gobj); + bo_type, NULL, &gobj, xcp_id + 1); if (ret) { pr_debug("Failed to create BO on domain %s. ret %d\n", domain_string(alloc_domain), ret); @@ -1728,6 +1798,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( (*mem)->domain = domain; (*mem)->mapped_to_gpu_memory = 0; (*mem)->process_info = avm->process_info; + add_kgd_mem_to_kfd_bo_list(*mem, avm->process_info, user_addr); if (user_addr) { @@ -1759,7 +1830,7 @@ err_node_allow: /* Don't unreserve system mem limit twice */ goto err_reserve_limit; err_bo_create: - amdgpu_amdkfd_unreserve_mem_limit(adev, aligned_size, flags); + amdgpu_amdkfd_unreserve_mem_limit(adev, aligned_size, flags, xcp_id); err_reserve_limit: mutex_destroy(&(*mem)->lock); if (gobj) @@ -1855,11 +1926,14 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu( } /* Update the size of the BO being freed if it was allocated from - * VRAM and is not imported. + * VRAM and is not imported. For APP APU VRAM allocations are done + * in GTT domain */ if (size) { - if ((mem->bo->preferred_domains == AMDGPU_GEM_DOMAIN_VRAM) && - (!is_imported)) + if (!is_imported && + (mem->bo->preferred_domains == AMDGPU_GEM_DOMAIN_VRAM || + (adev->gmc.is_app_apu && + mem->bo->preferred_domains == AMDGPU_GEM_DOMAIN_GTT))) *size = bo_size; else *size = 0; @@ -2282,8 +2356,9 @@ int amdgpu_amdkfd_gpuvm_import_dmabuf(struct amdgpu_device *adev, (*mem)->dmabuf = dma_buf; (*mem)->bo = bo; (*mem)->va = va; - (*mem)->domain = (bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM) ? + (*mem)->domain = (bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM) && !adev->gmc.is_app_apu ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT; + (*mem)->mapped_to_gpu_memory = 0; (*mem)->process_info = avm->process_info; add_kgd_mem_to_kfd_bo_list(*mem, avm->process_info, false); @@ -2445,7 +2520,9 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info, ret = -EAGAIN; goto unlock_out; } - mem->invalid = 0; + /* set mem valid if mem has hmm range associated */ + if (mem->range) + mem->invalid = 0; } unlock_out: @@ -2577,8 +2654,15 @@ static int confirm_valid_user_pages_locked(struct amdkfd_process_info *process_i list_for_each_entry_safe(mem, tmp_mem, &process_info->userptr_inval_list, validate_list.head) { - bool valid = amdgpu_ttm_tt_get_user_pages_done( - mem->bo->tbo.ttm, mem->range); + bool valid; + + /* keep mem without hmm range at userptr_inval_list */ + if (!mem->range) + continue; + + /* Only check mem with hmm range associated */ + valid = amdgpu_ttm_tt_get_user_pages_done( + mem->bo->tbo.ttm, mem->range); mem->range = NULL; if (!valid) { @@ -2586,7 +2670,12 @@ static int confirm_valid_user_pages_locked(struct amdkfd_process_info *process_i ret = -EAGAIN; continue; } - WARN(mem->invalid, "Valid BO is marked invalid"); + + if (mem->invalid) { + WARN(1, "Valid BO is marked invalid"); + ret = -EAGAIN; + continue; + } list_move_tail(&mem->validate_list.head, &process_info->userptr_valid_list); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c index ac6fe0ae4609..ef4b9a41f20a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c @@ -272,6 +272,7 @@ static int convert_atom_mem_type_to_vram_type(struct amdgpu_device *adev, break; case ATOM_DGPU_VRAM_TYPE_HBM2: case ATOM_DGPU_VRAM_TYPE_HBM2E: + case ATOM_DGPU_VRAM_TYPE_HBM3: vram_type = AMDGPU_VRAM_TYPE_HBM; break; case ATOM_DGPU_VRAM_TYPE_GDDR6: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c index 30c28a69e847..b582b83c4984 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c @@ -104,9 +104,8 @@ static bool igp_read_bios_from_vram(struct amdgpu_device *adev) adev->bios = NULL; vram_base = pci_resource_start(adev->pdev, 0); bios = ioremap_wc(vram_base, size); - if (!bios) { + if (!bios) return false; - } adev->bios = kmalloc(size, GFP_KERNEL); if (!adev->bios) { @@ -133,9 +132,8 @@ bool amdgpu_read_bios(struct amdgpu_device *adev) adev->bios = NULL; /* XXX: some cards may return 0 for rom size? ddx has a workaround */ bios = pci_map_rom(adev->pdev, &size); - if (!bios) { + if (!bios) return false; - } adev->bios = kzalloc(size, GFP_KERNEL); if (adev->bios == NULL) { @@ -168,9 +166,9 @@ static bool amdgpu_read_bios_from_rom(struct amdgpu_device *adev) header[AMD_VBIOS_SIGNATURE_END] = 0; if ((!AMD_IS_VALID_VBIOS(header)) || - 0 != memcmp((char *)&header[AMD_VBIOS_SIGNATURE_OFFSET], - AMD_VBIOS_SIGNATURE, - strlen(AMD_VBIOS_SIGNATURE))) + memcmp((char *)&header[AMD_VBIOS_SIGNATURE_OFFSET], + AMD_VBIOS_SIGNATURE, + strlen(AMD_VBIOS_SIGNATURE)) != 0) return false; /* valid vbios, go on */ @@ -264,7 +262,7 @@ static int amdgpu_atrm_call(acpi_handle atrm_handle, uint8_t *bios, status = acpi_evaluate_object(atrm_handle, NULL, &atrm_arg, &buffer); if (ACPI_FAILURE(status)) { - printk("failed to evaluate ATRM got %s\n", acpi_format_exception(status)); + DRM_ERROR("failed to evaluate ATRM got %s\n", acpi_format_exception(status)); return -ENODEV; } @@ -363,7 +361,7 @@ static bool amdgpu_acpi_vfct_bios(struct amdgpu_device *adev) struct acpi_table_header *hdr; acpi_size tbl_size; UEFI_ACPI_VFCT *vfct; - unsigned offset; + unsigned int offset; if (!ACPI_SUCCESS(acpi_get_table("VFCT", 1, &hdr))) return false; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c index 6be30dcb029d..d34037b85cf8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c @@ -593,11 +593,20 @@ static int amdgpu_connector_set_property(struct drm_connector *connector, switch (val) { default: - case DRM_MODE_SCALE_NONE: rmx_type = RMX_OFF; break; - case DRM_MODE_SCALE_CENTER: rmx_type = RMX_CENTER; break; - case DRM_MODE_SCALE_ASPECT: rmx_type = RMX_ASPECT; break; - case DRM_MODE_SCALE_FULLSCREEN: rmx_type = RMX_FULL; break; + case DRM_MODE_SCALE_NONE: + rmx_type = RMX_OFF; + break; + case DRM_MODE_SCALE_CENTER: + rmx_type = RMX_CENTER; + break; + case DRM_MODE_SCALE_ASPECT: + rmx_type = RMX_ASPECT; + break; + case DRM_MODE_SCALE_FULLSCREEN: + rmx_type = RMX_FULL; + break; } + if (amdgpu_encoder->rmx_type == rmx_type) return 0; @@ -799,12 +808,21 @@ static int amdgpu_connector_set_lcd_property(struct drm_connector *connector, } switch (value) { - case DRM_MODE_SCALE_NONE: rmx_type = RMX_OFF; break; - case DRM_MODE_SCALE_CENTER: rmx_type = RMX_CENTER; break; - case DRM_MODE_SCALE_ASPECT: rmx_type = RMX_ASPECT; break; + case DRM_MODE_SCALE_NONE: + rmx_type = RMX_OFF; + break; + case DRM_MODE_SCALE_CENTER: + rmx_type = RMX_CENTER; + break; + case DRM_MODE_SCALE_ASPECT: + rmx_type = RMX_ASPECT; + break; default: - case DRM_MODE_SCALE_FULLSCREEN: rmx_type = RMX_FULL; break; + case DRM_MODE_SCALE_FULLSCREEN: + rmx_type = RMX_FULL; + break; } + if (amdgpu_encoder->rmx_type == rmx_type) return 0; @@ -1127,7 +1145,8 @@ amdgpu_connector_dvi_detect(struct drm_connector *connector, bool force) /* assume digital unless load detected otherwise */ amdgpu_connector->use_digital = true; lret = encoder_funcs->detect(encoder, connector); - DRM_DEBUG_KMS("load_detect %x returned: %x\n",encoder->encoder_type,lret); + DRM_DEBUG_KMS("load_detect %x returned: %x\n", + encoder->encoder_type, lret); if (lret == connector_status_connected) amdgpu_connector->use_digital = false; } @@ -1991,7 +2010,7 @@ amdgpu_connector_add(struct amdgpu_device *adev, if (amdgpu_connector->hpd.hpd == AMDGPU_HPD_NONE) { if (i2c_bus->valid) { connector->polled = DRM_CONNECTOR_POLL_CONNECT | - DRM_CONNECTOR_POLL_DISCONNECT; + DRM_CONNECTOR_POLL_DISCONNECT; } } else connector->polled = DRM_CONNECTOR_POLL_HPD; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 2eb2c66843a8..d9503882ea97 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -112,6 +112,9 @@ static int amdgpu_cs_p1_ib(struct amdgpu_cs_parser *p, if (r < 0) return r; + if (num_ibs[r] >= amdgpu_ring_max_ibs(chunk_ib->ip_type)) + return -EINVAL; + ++(num_ibs[r]); p->gang_leader_idx = r; return 0; @@ -192,7 +195,7 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p, uint64_t *chunk_array_user; uint64_t *chunk_array; uint32_t uf_offset = 0; - unsigned int size; + size_t size; int ret; int i; @@ -285,6 +288,7 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p, case AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES: case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_WAIT: case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_SIGNAL: + case AMDGPU_CHUNK_ID_CP_GFX_SHADOW: break; default: @@ -305,7 +309,7 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p, } p->gang_leader = p->jobs[p->gang_leader_idx]; - if (p->ctx->vram_lost_counter != p->gang_leader->vram_lost_counter) { + if (p->ctx->generation != p->gang_leader->generation) { ret = -ECANCELED; goto free_all_kdata; } @@ -393,7 +397,7 @@ static int amdgpu_cs_p2_dependencies(struct amdgpu_cs_parser *p, { struct drm_amdgpu_cs_chunk_dep *deps = chunk->kdata; struct amdgpu_fpriv *fpriv = p->filp->driver_priv; - unsigned num_deps; + unsigned int num_deps; int i, r; num_deps = chunk->length_dw * 4 / @@ -464,7 +468,7 @@ static int amdgpu_cs_p2_syncobj_in(struct amdgpu_cs_parser *p, struct amdgpu_cs_chunk *chunk) { struct drm_amdgpu_cs_chunk_sem *deps = chunk->kdata; - unsigned num_deps; + unsigned int num_deps; int i, r; num_deps = chunk->length_dw * 4 / @@ -482,7 +486,7 @@ static int amdgpu_cs_p2_syncobj_timeline_wait(struct amdgpu_cs_parser *p, struct amdgpu_cs_chunk *chunk) { struct drm_amdgpu_cs_chunk_syncobj *syncobj_deps = chunk->kdata; - unsigned num_deps; + unsigned int num_deps; int i, r; num_deps = chunk->length_dw * 4 / @@ -502,7 +506,7 @@ static int amdgpu_cs_p2_syncobj_out(struct amdgpu_cs_parser *p, struct amdgpu_cs_chunk *chunk) { struct drm_amdgpu_cs_chunk_sem *deps = chunk->kdata; - unsigned num_deps; + unsigned int num_deps; int i; num_deps = chunk->length_dw * 4 / @@ -536,7 +540,7 @@ static int amdgpu_cs_p2_syncobj_timeline_signal(struct amdgpu_cs_parser *p, struct amdgpu_cs_chunk *chunk) { struct drm_amdgpu_cs_chunk_syncobj *syncobj_deps = chunk->kdata; - unsigned num_deps; + unsigned int num_deps; int i; num_deps = chunk->length_dw * 4 / @@ -575,6 +579,26 @@ static int amdgpu_cs_p2_syncobj_timeline_signal(struct amdgpu_cs_parser *p, return 0; } +static int amdgpu_cs_p2_shadow(struct amdgpu_cs_parser *p, + struct amdgpu_cs_chunk *chunk) +{ + struct drm_amdgpu_cs_chunk_cp_gfx_shadow *shadow = chunk->kdata; + int i; + + if (shadow->flags & ~AMDGPU_CS_CHUNK_CP_GFX_SHADOW_FLAGS_INIT_SHADOW) + return -EINVAL; + + for (i = 0; i < p->gang_size; ++i) { + p->jobs[i]->shadow_va = shadow->shadow_va; + p->jobs[i]->csa_va = shadow->csa_va; + p->jobs[i]->gds_va = shadow->gds_va; + p->jobs[i]->init_shadow = + shadow->flags & AMDGPU_CS_CHUNK_CP_GFX_SHADOW_FLAGS_INIT_SHADOW; + } + + return 0; +} + static int amdgpu_cs_pass2(struct amdgpu_cs_parser *p) { unsigned int ce_preempt = 0, de_preempt = 0; @@ -617,6 +641,11 @@ static int amdgpu_cs_pass2(struct amdgpu_cs_parser *p) if (r) return r; break; + case AMDGPU_CHUNK_ID_CP_GFX_SHADOW: + r = amdgpu_cs_p2_shadow(p, chunk); + if (r) + return r; + break; } } @@ -729,6 +758,7 @@ static void amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev, if (used_vis_vram < total_vis_vram) { u64 free_vis_vram = total_vis_vram - used_vis_vram; + adev->mm_stats.accum_us_vis = min(adev->mm_stats.accum_us_vis + increment_us, us_upper_bound); @@ -1047,9 +1077,8 @@ static int amdgpu_cs_patch_ibs(struct amdgpu_cs_parser *p, /* the IB should be reserved at this point */ r = amdgpu_bo_kmap(aobj, (void **)&kptr); - if (r) { + if (r) return r; - } kptr += va_start - (m->start * AMDGPU_GPU_PAGE_SIZE); @@ -1356,7 +1385,7 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p, /* Cleanup the parser structure */ static void amdgpu_cs_parser_fini(struct amdgpu_cs_parser *parser) { - unsigned i; + unsigned int i; amdgpu_sync_free(&parser->sync); for (i = 0; i < parser->num_post_deps; i++) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c index c6d4d41c4393..23d054526e7c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c @@ -106,3 +106,41 @@ int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm, ttm_eu_backoff_reservation(&ticket, &list); return 0; } + +int amdgpu_unmap_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm, + struct amdgpu_bo *bo, struct amdgpu_bo_va *bo_va, + uint64_t csa_addr) +{ + struct ww_acquire_ctx ticket; + struct list_head list; + struct amdgpu_bo_list_entry pd; + struct ttm_validate_buffer csa_tv; + int r; + + INIT_LIST_HEAD(&list); + INIT_LIST_HEAD(&csa_tv.head); + csa_tv.bo = &bo->tbo; + csa_tv.num_shared = 1; + + list_add(&csa_tv.head, &list); + amdgpu_vm_get_pd_bo(vm, &list, &pd); + + r = ttm_eu_reserve_buffers(&ticket, &list, true, NULL); + if (r) { + DRM_ERROR("failed to reserve CSA,PD BOs: err=%d\n", r); + return r; + } + + r = amdgpu_vm_bo_unmap(adev, bo_va, csa_addr); + if (r) { + DRM_ERROR("failed to do bo_unmap on static CSA, err=%d\n", r); + ttm_eu_backoff_reservation(&ticket, &list); + return r; + } + + amdgpu_vm_bo_del(adev, bo_va); + + ttm_eu_backoff_reservation(&ticket, &list); + + return 0; +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.h index 524b4437a021..7dfc1f2012eb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.h @@ -34,6 +34,9 @@ int amdgpu_allocate_static_csa(struct amdgpu_device *adev, struct amdgpu_bo **bo int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm, struct amdgpu_bo *bo, struct amdgpu_bo_va **bo_va, uint64_t csa_addr, uint32_t size); +int amdgpu_unmap_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm, + struct amdgpu_bo *bo, struct amdgpu_bo_va *bo_va, + uint64_t csa_addr); void amdgpu_free_static_csa(struct amdgpu_bo **bo); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c index d2139ac12159..0dc9c655c4fb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c @@ -222,8 +222,19 @@ static int amdgpu_ctx_init_entity(struct amdgpu_ctx *ctx, u32 hw_ip, drm_prio = amdgpu_ctx_to_drm_sched_prio(ctx_prio); hw_ip = array_index_nospec(hw_ip, AMDGPU_HW_IP_NUM); - scheds = adev->gpu_sched[hw_ip][hw_prio].sched; - num_scheds = adev->gpu_sched[hw_ip][hw_prio].num_scheds; + + if (!(adev)->xcp_mgr) { + scheds = adev->gpu_sched[hw_ip][hw_prio].sched; + num_scheds = adev->gpu_sched[hw_ip][hw_prio].num_scheds; + } else { + struct amdgpu_fpriv *fpriv; + + fpriv = container_of(ctx->ctx_mgr, struct amdgpu_fpriv, ctx_mgr); + r = amdgpu_xcp_select_scheds(adev, hw_ip, hw_prio, fpriv, + &num_scheds, &scheds); + if (r) + goto cleanup_entity; + } /* disable load balance if the hw engine retains context among dependent jobs */ if (hw_ip == AMDGPU_HW_IP_VCN_ENC || @@ -255,7 +266,8 @@ error_free_entity: return r; } -static ktime_t amdgpu_ctx_fini_entity(struct amdgpu_ctx_entity *entity) +static ktime_t amdgpu_ctx_fini_entity(struct amdgpu_device *adev, + struct amdgpu_ctx_entity *entity) { ktime_t res = ns_to_ktime(0); int i; @@ -268,6 +280,8 @@ static ktime_t amdgpu_ctx_fini_entity(struct amdgpu_ctx_entity *entity) dma_fence_put(entity->fences[i]); } + amdgpu_xcp_release_sched(adev, entity); + kfree(entity); return res; } @@ -303,6 +317,7 @@ static int amdgpu_ctx_get_stable_pstate(struct amdgpu_ctx *ctx, static int amdgpu_ctx_init(struct amdgpu_ctx_mgr *mgr, int32_t priority, struct drm_file *filp, struct amdgpu_ctx *ctx) { + struct amdgpu_fpriv *fpriv = filp->driver_priv; u32 current_stable_pstate; int r; @@ -318,7 +333,7 @@ static int amdgpu_ctx_init(struct amdgpu_ctx_mgr *mgr, int32_t priority, ctx->reset_counter = atomic_read(&mgr->adev->gpu_reset_counter); ctx->reset_counter_query = ctx->reset_counter; - ctx->vram_lost_counter = atomic_read(&mgr->adev->vram_lost_counter); + ctx->generation = amdgpu_vm_generation(mgr->adev, &fpriv->vm); ctx->init_priority = priority; ctx->override_priority = AMDGPU_CTX_PRIORITY_UNSET; @@ -331,6 +346,7 @@ static int amdgpu_ctx_init(struct amdgpu_ctx_mgr *mgr, int32_t priority, else ctx->stable_pstate = current_stable_pstate; + ctx->ctx_mgr = &(fpriv->ctx_mgr); return 0; } @@ -399,7 +415,7 @@ static void amdgpu_ctx_fini(struct kref *ref) for (j = 0; j < AMDGPU_MAX_ENTITY_NUM; ++j) { ktime_t spend; - spend = amdgpu_ctx_fini_entity(ctx->entities[i][j]); + spend = amdgpu_ctx_fini_entity(adev, ctx->entities[i][j]); atomic64_add(ktime_to_ns(spend), &mgr->time_spend[i]); } } @@ -416,6 +432,7 @@ int amdgpu_ctx_get_entity(struct amdgpu_ctx *ctx, u32 hw_ip, u32 instance, u32 ring, struct drm_sched_entity **entity) { int r; + struct drm_sched_entity *ctx_entity; if (hw_ip >= AMDGPU_HW_IP_NUM) { DRM_ERROR("unknown HW IP type: %d\n", hw_ip); @@ -439,7 +456,14 @@ int amdgpu_ctx_get_entity(struct amdgpu_ctx *ctx, u32 hw_ip, u32 instance, return r; } - *entity = &ctx->entities[hw_ip][ring]->entity; + ctx_entity = &ctx->entities[hw_ip][ring]->entity; + r = drm_sched_entity_error(ctx_entity); + if (r) { + DRM_DEBUG("error entity %p\n", ctx_entity); + return r; + } + + *entity = ctx_entity; return 0; } @@ -570,12 +594,15 @@ static int amdgpu_ctx_query2(struct amdgpu_device *adev, if (ctx->reset_counter != atomic_read(&adev->gpu_reset_counter)) out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_RESET; - if (ctx->vram_lost_counter != atomic_read(&adev->vram_lost_counter)) + if (ctx->generation != amdgpu_vm_generation(adev, &fpriv->vm)) out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_VRAMLOST; if (atomic_read(&ctx->guilty)) out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_GUILTY; + if (amdgpu_in_reset(adev)) + out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_RESET_IN_PROGRESS; + if (adev->ras_enabled && con) { /* Return the cached values in O(1), * and schedule delayed work to cache diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h index 0fa0e56daf67..85376baaa92f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h @@ -47,7 +47,7 @@ struct amdgpu_ctx { struct amdgpu_ctx_mgr *mgr; unsigned reset_counter; unsigned reset_counter_query; - uint32_t vram_lost_counter; + uint64_t generation; spinlock_t ring_lock; struct amdgpu_ctx_entity *entities[AMDGPU_HW_IP_NUM][AMDGPU_MAX_ENTITY_NUM]; bool preamble_presented; @@ -57,6 +57,7 @@ struct amdgpu_ctx { unsigned long ras_counter_ce; unsigned long ras_counter_ue; uint32_t stable_pstate; + struct amdgpu_ctx_mgr *ctx_mgr; }; struct amdgpu_ctx_mgr { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c index f60753f97ac5..56e89e76ff17 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c @@ -56,14 +56,14 @@ * * Bit 62: Indicates a GRBM bank switch is needed * Bit 61: Indicates a SRBM bank switch is needed (implies bit 62 is - * zero) + * zero) * Bits 24..33: The SE or ME selector if needed * Bits 34..43: The SH (or SA) or PIPE selector if needed * Bits 44..53: The INSTANCE (or CU/WGP) or QUEUE selector if needed * * Bit 23: Indicates that the PM power gating lock should be held - * This is necessary to read registers that might be - * unreliable during a power gating transistion. + * This is necessary to read registers that might be + * unreliable during a power gating transistion. * * The lower bits are the BYTE offset of the register to read. This * allows reading multiple registers in a single call and having @@ -76,7 +76,7 @@ static int amdgpu_debugfs_process_reg_op(bool read, struct file *f, ssize_t result = 0; int r; bool pm_pg_lock, use_bank, use_ring; - unsigned instance_bank, sh_bank, se_bank, me, pipe, queue, vmid; + unsigned int instance_bank, sh_bank, se_bank, me, pipe, queue, vmid; pm_pg_lock = use_bank = use_ring = false; instance_bank = sh_bank = se_bank = me = pipe = queue = vmid = 0; @@ -136,10 +136,10 @@ static int amdgpu_debugfs_process_reg_op(bool read, struct file *f, } mutex_lock(&adev->grbm_idx_mutex); amdgpu_gfx_select_se_sh(adev, se_bank, - sh_bank, instance_bank); + sh_bank, instance_bank, 0); } else if (use_ring) { mutex_lock(&adev->srbm_mutex); - amdgpu_gfx_select_me_pipe_q(adev, me, pipe, queue, vmid); + amdgpu_gfx_select_me_pipe_q(adev, me, pipe, queue, vmid, 0); } if (pm_pg_lock) @@ -169,10 +169,10 @@ static int amdgpu_debugfs_process_reg_op(bool read, struct file *f, end: if (use_bank) { - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); } else if (use_ring) { - amdgpu_gfx_select_me_pipe_q(adev, 0, 0, 0, 0); + amdgpu_gfx_select_me_pipe_q(adev, 0, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); } @@ -208,7 +208,7 @@ static int amdgpu_debugfs_regs2_open(struct inode *inode, struct file *file) { struct amdgpu_debugfs_regs2_data *rd; - rd = kzalloc(sizeof *rd, GFP_KERNEL); + rd = kzalloc(sizeof(*rd), GFP_KERNEL); if (!rd) return -ENOMEM; rd->adev = file_inode(file)->i_private; @@ -221,6 +221,7 @@ static int amdgpu_debugfs_regs2_open(struct inode *inode, struct file *file) static int amdgpu_debugfs_regs2_release(struct inode *inode, struct file *file) { struct amdgpu_debugfs_regs2_data *rd = file->private_data; + mutex_destroy(&rd->lock); kfree(file->private_data); return 0; @@ -262,14 +263,14 @@ static ssize_t amdgpu_debugfs_regs2_op(struct file *f, char __user *buf, u32 off } mutex_lock(&adev->grbm_idx_mutex); amdgpu_gfx_select_se_sh(adev, rd->id.grbm.se, - rd->id.grbm.sh, - rd->id.grbm.instance); + rd->id.grbm.sh, + rd->id.grbm.instance, rd->id.xcc_id); } if (rd->id.use_srbm) { mutex_lock(&adev->srbm_mutex); amdgpu_gfx_select_me_pipe_q(adev, rd->id.srbm.me, rd->id.srbm.pipe, - rd->id.srbm.queue, rd->id.srbm.vmid); + rd->id.srbm.queue, rd->id.srbm.vmid, rd->id.xcc_id); } if (rd->id.pg_lock) @@ -295,12 +296,12 @@ static ssize_t amdgpu_debugfs_regs2_op(struct file *f, char __user *buf, u32 off } end: if (rd->id.use_grbm) { - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, rd->id.xcc_id); mutex_unlock(&adev->grbm_idx_mutex); } if (rd->id.use_srbm) { - amdgpu_gfx_select_me_pipe_q(adev, 0, 0, 0, 0); + amdgpu_gfx_select_me_pipe_q(adev, 0, 0, 0, 0, rd->id.xcc_id); mutex_unlock(&adev->srbm_mutex); } @@ -319,18 +320,45 @@ end: static long amdgpu_debugfs_regs2_ioctl(struct file *f, unsigned int cmd, unsigned long data) { struct amdgpu_debugfs_regs2_data *rd = f->private_data; + struct amdgpu_debugfs_regs2_iocdata v1_data; int r; + mutex_lock(&rd->lock); + switch (cmd) { + case AMDGPU_DEBUGFS_REGS2_IOC_SET_STATE_V2: + r = copy_from_user(&rd->id, (struct amdgpu_debugfs_regs2_iocdata_v2 *)data, + sizeof(rd->id)); + if (r) + r = -EINVAL; + goto done; case AMDGPU_DEBUGFS_REGS2_IOC_SET_STATE: - mutex_lock(&rd->lock); - r = copy_from_user(&rd->id, (struct amdgpu_debugfs_regs2_iocdata *)data, sizeof rd->id); - mutex_unlock(&rd->lock); - return r ? -EINVAL : 0; + r = copy_from_user(&v1_data, (struct amdgpu_debugfs_regs2_iocdata *)data, + sizeof(v1_data)); + if (r) { + r = -EINVAL; + goto done; + } + goto v1_copy; default: - return -EINVAL; - } - return 0; + r = -EINVAL; + goto done; + } + +v1_copy: + rd->id.use_srbm = v1_data.use_srbm; + rd->id.use_grbm = v1_data.use_grbm; + rd->id.pg_lock = v1_data.pg_lock; + rd->id.grbm.se = v1_data.grbm.se; + rd->id.grbm.sh = v1_data.grbm.sh; + rd->id.grbm.instance = v1_data.grbm.instance; + rd->id.srbm.me = v1_data.srbm.me; + rd->id.srbm.pipe = v1_data.srbm.pipe; + rd->id.srbm.queue = v1_data.srbm.queue; + rd->id.xcc_id = 0; +done: + mutex_unlock(&rd->lock); + return r; } static ssize_t amdgpu_debugfs_regs2_read(struct file *f, char __user *buf, size_t size, loff_t *pos) @@ -343,6 +371,136 @@ static ssize_t amdgpu_debugfs_regs2_write(struct file *f, const char __user *buf return amdgpu_debugfs_regs2_op(f, (char __user *)buf, *pos, size, 1); } +static int amdgpu_debugfs_gprwave_open(struct inode *inode, struct file *file) +{ + struct amdgpu_debugfs_gprwave_data *rd; + + rd = kzalloc(sizeof *rd, GFP_KERNEL); + if (!rd) + return -ENOMEM; + rd->adev = file_inode(file)->i_private; + file->private_data = rd; + mutex_init(&rd->lock); + + return 0; +} + +static int amdgpu_debugfs_gprwave_release(struct inode *inode, struct file *file) +{ + struct amdgpu_debugfs_gprwave_data *rd = file->private_data; + mutex_destroy(&rd->lock); + kfree(file->private_data); + return 0; +} + +static ssize_t amdgpu_debugfs_gprwave_read(struct file *f, char __user *buf, size_t size, loff_t *pos) +{ + struct amdgpu_debugfs_gprwave_data *rd = f->private_data; + struct amdgpu_device *adev = rd->adev; + ssize_t result = 0; + int r; + uint32_t *data, x; + + if (size & 0x3 || *pos & 0x3) + return -EINVAL; + + r = pm_runtime_get_sync(adev_to_drm(adev)->dev); + if (r < 0) { + pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); + return r; + } + + r = amdgpu_virt_enable_access_debugfs(adev); + if (r < 0) { + pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); + return r; + } + + data = kcalloc(1024, sizeof(*data), GFP_KERNEL); + if (!data) { + pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); + amdgpu_virt_disable_access_debugfs(adev); + return -ENOMEM; + } + + /* switch to the specific se/sh/cu */ + mutex_lock(&adev->grbm_idx_mutex); + amdgpu_gfx_select_se_sh(adev, rd->id.se, rd->id.sh, rd->id.cu, rd->id.xcc_id); + + if (!rd->id.gpr_or_wave) { + x = 0; + if (adev->gfx.funcs->read_wave_data) + adev->gfx.funcs->read_wave_data(adev, rd->id.xcc_id, rd->id.simd, rd->id.wave, data, &x); + } else { + x = size >> 2; + if (rd->id.gpr.vpgr_or_sgpr) { + if (adev->gfx.funcs->read_wave_vgprs) + adev->gfx.funcs->read_wave_vgprs(adev, rd->id.xcc_id, rd->id.simd, rd->id.wave, rd->id.gpr.thread, *pos, size>>2, data); + } else { + if (adev->gfx.funcs->read_wave_sgprs) + adev->gfx.funcs->read_wave_sgprs(adev, rd->id.xcc_id, rd->id.simd, rd->id.wave, *pos, size>>2, data); + } + } + + amdgpu_gfx_select_se_sh(adev, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, rd->id.xcc_id); + mutex_unlock(&adev->grbm_idx_mutex); + + pm_runtime_mark_last_busy(adev_to_drm(adev)->dev); + pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); + + if (!x) { + result = -EINVAL; + goto done; + } + + while (size && (*pos < x * 4)) { + uint32_t value; + + value = data[*pos >> 2]; + r = put_user(value, (uint32_t *)buf); + if (r) { + result = r; + goto done; + } + + result += 4; + buf += 4; + *pos += 4; + size -= 4; + } + +done: + amdgpu_virt_disable_access_debugfs(adev); + kfree(data); + return result; +} + +static long amdgpu_debugfs_gprwave_ioctl(struct file *f, unsigned int cmd, unsigned long data) +{ + struct amdgpu_debugfs_gprwave_data *rd = f->private_data; + int r = 0; + + mutex_lock(&rd->lock); + + switch (cmd) { + case AMDGPU_DEBUGFS_GPRWAVE_IOC_SET_STATE: + if (copy_from_user(&rd->id, + (struct amdgpu_debugfs_gprwave_iocdata *)data, + sizeof(rd->id))) + r = -EFAULT; + goto done; + default: + r = -EINVAL; + goto done; + } + +done: + mutex_unlock(&rd->lock); + return r; +} + + + /** * amdgpu_debugfs_regs_pcie_read - Read from a PCIE register @@ -863,7 +1021,7 @@ static ssize_t amdgpu_debugfs_sensor_read(struct file *f, char __user *buf, * The offset being sought changes which wave that the status data * will be returned for. The bits are used as follows: * - * Bits 0..6: Byte offset into data + * Bits 0..6: Byte offset into data * Bits 7..14: SE selector * Bits 15..22: SH/SA selector * Bits 23..30: CU/{WGP+SIMD} selector @@ -907,13 +1065,13 @@ static ssize_t amdgpu_debugfs_wave_read(struct file *f, char __user *buf, /* switch to the specific se/sh/cu */ mutex_lock(&adev->grbm_idx_mutex); - amdgpu_gfx_select_se_sh(adev, se, sh, cu); + amdgpu_gfx_select_se_sh(adev, se, sh, cu, 0); x = 0; if (adev->gfx.funcs->read_wave_data) - adev->gfx.funcs->read_wave_data(adev, simd, wave, data, &x); + adev->gfx.funcs->read_wave_data(adev, 0, simd, wave, data, &x); - amdgpu_gfx_select_se_sh(adev, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF); + amdgpu_gfx_select_se_sh(adev, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0); mutex_unlock(&adev->grbm_idx_mutex); pm_runtime_mark_last_busy(adev_to_drm(adev)->dev); @@ -1001,17 +1159,17 @@ static ssize_t amdgpu_debugfs_gpr_read(struct file *f, char __user *buf, /* switch to the specific se/sh/cu */ mutex_lock(&adev->grbm_idx_mutex); - amdgpu_gfx_select_se_sh(adev, se, sh, cu); + amdgpu_gfx_select_se_sh(adev, se, sh, cu, 0); if (bank == 0) { if (adev->gfx.funcs->read_wave_vgprs) - adev->gfx.funcs->read_wave_vgprs(adev, simd, wave, thread, offset, size>>2, data); + adev->gfx.funcs->read_wave_vgprs(adev, 0, simd, wave, thread, offset, size>>2, data); } else { if (adev->gfx.funcs->read_wave_sgprs) - adev->gfx.funcs->read_wave_sgprs(adev, simd, wave, offset, size>>2, data); + adev->gfx.funcs->read_wave_sgprs(adev, 0, simd, wave, offset, size>>2, data); } - amdgpu_gfx_select_se_sh(adev, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF); + amdgpu_gfx_select_se_sh(adev, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0); mutex_unlock(&adev->grbm_idx_mutex); pm_runtime_mark_last_busy(adev_to_drm(adev)->dev); @@ -1339,6 +1497,15 @@ static const struct file_operations amdgpu_debugfs_regs2_fops = { .llseek = default_llseek }; +static const struct file_operations amdgpu_debugfs_gprwave_fops = { + .owner = THIS_MODULE, + .unlocked_ioctl = amdgpu_debugfs_gprwave_ioctl, + .read = amdgpu_debugfs_gprwave_read, + .open = amdgpu_debugfs_gprwave_open, + .release = amdgpu_debugfs_gprwave_release, + .llseek = default_llseek +}; + static const struct file_operations amdgpu_debugfs_regs_fops = { .owner = THIS_MODULE, .read = amdgpu_debugfs_regs_read, @@ -1416,6 +1583,7 @@ static const struct file_operations amdgpu_debugfs_gfxoff_residency_fops = { static const struct file_operations *debugfs_regs[] = { &amdgpu_debugfs_regs_fops, &amdgpu_debugfs_regs2_fops, + &amdgpu_debugfs_gprwave_fops, &amdgpu_debugfs_regs_didt_fops, &amdgpu_debugfs_regs_pcie_fops, &amdgpu_debugfs_regs_smc_fops, @@ -1429,9 +1597,10 @@ static const struct file_operations *debugfs_regs[] = { &amdgpu_debugfs_gfxoff_residency_fops, }; -static const char *debugfs_regs_names[] = { +static const char * const debugfs_regs_names[] = { "amdgpu_regs", "amdgpu_regs2", + "amdgpu_gprwave", "amdgpu_regs_didt", "amdgpu_regs_pcie", "amdgpu_regs_smc", @@ -1447,7 +1616,7 @@ static const char *debugfs_regs_names[] = { /** * amdgpu_debugfs_regs_init - Initialize debugfs entries that provide - * register access. + * register access. * * @adev: The device to attach the debugfs entries to */ @@ -1459,7 +1628,7 @@ int amdgpu_debugfs_regs_init(struct amdgpu_device *adev) for (i = 0; i < ARRAY_SIZE(debugfs_regs); i++) { ent = debugfs_create_file(debugfs_regs_names[i], - S_IFREG | S_IRUGO, root, + S_IFREG | 0444, root, adev, debugfs_regs[i]); if (!i && !IS_ERR_OR_NULL(ent)) i_size_write(ent->d_inode, adev->rmmio_size); @@ -1470,7 +1639,7 @@ int amdgpu_debugfs_regs_init(struct amdgpu_device *adev) static int amdgpu_debugfs_test_ib_show(struct seq_file *m, void *unused) { - struct amdgpu_device *adev = (struct amdgpu_device *)m->private; + struct amdgpu_device *adev = m->private; struct drm_device *dev = adev_to_drm(adev); int r = 0, i; @@ -1494,12 +1663,12 @@ static int amdgpu_debugfs_test_ib_show(struct seq_file *m, void *unused) kthread_park(ring->sched.thread); } - seq_printf(m, "run ib test:\n"); + seq_puts(m, "run ib test:\n"); r = amdgpu_ib_ring_tests(adev); if (r) seq_printf(m, "ib ring tests failed (%d).\n", r); else - seq_printf(m, "ib ring tests passed.\n"); + seq_puts(m, "ib ring tests passed.\n"); /* go on the scheduler */ for (i = 0; i < AMDGPU_MAX_RINGS; i++) { @@ -1581,7 +1750,7 @@ static int amdgpu_debugfs_benchmark(void *data, u64 val) static int amdgpu_debugfs_vm_info_show(struct seq_file *m, void *unused) { - struct amdgpu_device *adev = (struct amdgpu_device *)m->private; + struct amdgpu_device *adev = m->private; struct drm_device *dev = adev_to_drm(adev); struct drm_file *file; int r; @@ -1978,7 +2147,7 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev) amdgpu_debugfs_ring_init(adev, ring); } - for ( i = 0; i < adev->vcn.num_vcn_inst; i++) { + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { if (!amdgpu_vcnfw_log) break; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 5c7d40873ee2..e25f085ee886 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -707,6 +707,48 @@ u32 amdgpu_device_indirect_rreg(struct amdgpu_device *adev, return r; } +u32 amdgpu_device_indirect_rreg_ext(struct amdgpu_device *adev, + u64 reg_addr) +{ + unsigned long flags, pcie_index, pcie_index_hi, pcie_data; + u32 r; + void __iomem *pcie_index_offset; + void __iomem *pcie_index_hi_offset; + void __iomem *pcie_data_offset; + + pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev); + pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev); + if (adev->nbio.funcs->get_pcie_index_hi_offset) + pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev); + else + pcie_index_hi = 0; + + spin_lock_irqsave(&adev->pcie_idx_lock, flags); + pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4; + pcie_data_offset = (void __iomem *)adev->rmmio + pcie_data * 4; + if (pcie_index_hi != 0) + pcie_index_hi_offset = (void __iomem *)adev->rmmio + + pcie_index_hi * 4; + + writel(reg_addr, pcie_index_offset); + readl(pcie_index_offset); + if (pcie_index_hi != 0) { + writel((reg_addr >> 32) & 0xff, pcie_index_hi_offset); + readl(pcie_index_hi_offset); + } + r = readl(pcie_data_offset); + + /* clear the high bits */ + if (pcie_index_hi != 0) { + writel(0, pcie_index_hi_offset); + readl(pcie_index_hi_offset); + } + + spin_unlock_irqrestore(&adev->pcie_idx_lock, flags); + + return r; +} + /** * amdgpu_device_indirect_rreg64 - read a 64bits indirect register * @@ -747,8 +789,6 @@ u64 amdgpu_device_indirect_rreg64(struct amdgpu_device *adev, * amdgpu_device_indirect_wreg - write an indirect register address * * @adev: amdgpu_device pointer - * @pcie_index: mmio register offset - * @pcie_data: mmio register offset * @reg_addr: indirect register offset * @reg_data: indirect register data * @@ -774,12 +814,50 @@ void amdgpu_device_indirect_wreg(struct amdgpu_device *adev, spin_unlock_irqrestore(&adev->pcie_idx_lock, flags); } +void amdgpu_device_indirect_wreg_ext(struct amdgpu_device *adev, + u64 reg_addr, u32 reg_data) +{ + unsigned long flags, pcie_index, pcie_index_hi, pcie_data; + void __iomem *pcie_index_offset; + void __iomem *pcie_index_hi_offset; + void __iomem *pcie_data_offset; + + pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev); + pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev); + if (adev->nbio.funcs->get_pcie_index_hi_offset) + pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev); + else + pcie_index_hi = 0; + + spin_lock_irqsave(&adev->pcie_idx_lock, flags); + pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4; + pcie_data_offset = (void __iomem *)adev->rmmio + pcie_data * 4; + if (pcie_index_hi != 0) + pcie_index_hi_offset = (void __iomem *)adev->rmmio + + pcie_index_hi * 4; + + writel(reg_addr, pcie_index_offset); + readl(pcie_index_offset); + if (pcie_index_hi != 0) { + writel((reg_addr >> 32) & 0xff, pcie_index_hi_offset); + readl(pcie_index_hi_offset); + } + writel(reg_data, pcie_data_offset); + readl(pcie_data_offset); + + /* clear the high bits */ + if (pcie_index_hi != 0) { + writel(0, pcie_index_hi_offset); + readl(pcie_index_hi_offset); + } + + spin_unlock_irqrestore(&adev->pcie_idx_lock, flags); +} + /** * amdgpu_device_indirect_wreg64 - write a 64bits indirect register address * * @adev: amdgpu_device pointer - * @pcie_index: mmio register offset - * @pcie_data: mmio register offset * @reg_addr: indirect register offset * @reg_data: indirect register data * @@ -840,6 +918,13 @@ static uint32_t amdgpu_invalid_rreg(struct amdgpu_device *adev, uint32_t reg) return 0; } +static uint32_t amdgpu_invalid_rreg_ext(struct amdgpu_device *adev, uint64_t reg) +{ + DRM_ERROR("Invalid callback to read register 0x%llX\n", reg); + BUG(); + return 0; +} + /** * amdgpu_invalid_wreg - dummy reg write function * @@ -857,6 +942,13 @@ static void amdgpu_invalid_wreg(struct amdgpu_device *adev, uint32_t reg, uint32 BUG(); } +static void amdgpu_invalid_wreg_ext(struct amdgpu_device *adev, uint64_t reg, uint32_t v) +{ + DRM_ERROR("Invalid callback to write register 0x%llX with 0x%08X\n", + reg, v); + BUG(); +} + /** * amdgpu_invalid_rreg64 - dummy 64 bit reg read function * @@ -942,7 +1034,8 @@ static int amdgpu_device_asic_init(struct amdgpu_device *adev) { amdgpu_asic_pre_asic_init(adev); - if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(11, 0, 0)) + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3) || + adev->ip_versions[GC_HWIP][0] >= IP_VERSION(11, 0, 0)) return amdgpu_atomfirmware_asic_init(adev, true); else return amdgpu_atom_asic_init(adev->mode_info.atom_context); @@ -998,7 +1091,7 @@ void amdgpu_device_program_register_sequence(struct amdgpu_device *adev, if (array_size % 3) return; - for (i = 0; i < array_size; i +=3) { + for (i = 0; i < array_size; i += 3) { reg = registers[i + 0]; and_mask = registers[i + 1]; or_mask = registers[i + 2]; @@ -1090,7 +1183,8 @@ static int amdgpu_device_doorbell_init(struct amdgpu_device *adev) * doorbells are in the first page. So with paging queue enabled, * the max num_kernel_doorbells should + 1 page (0x400 in dword) */ - if (adev->asic_type >= CHIP_VEGA10) + if (adev->ip_versions[SDMA0_HWIP][0] >= IP_VERSION(4, 0, 0) && + adev->ip_versions[SDMA0_HWIP][0] < IP_VERSION(4, 2, 0)) adev->doorbell.num_kernel_doorbells += 0x400; } @@ -1291,6 +1385,15 @@ int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev) return 0; } +static bool amdgpu_device_read_bios(struct amdgpu_device *adev) +{ + if (hweight32(adev->aid_mask) && (adev->flags & AMD_IS_APU)) { + return false; + } + + return true; +} + /* * GPU helpers function. */ @@ -1310,6 +1413,9 @@ bool amdgpu_device_need_post(struct amdgpu_device *adev) if (amdgpu_sriov_vf(adev)) return false; + if (!amdgpu_device_read_bios(adev)) + return false; + if (amdgpu_passthrough(adev)) { /* for FIJI: In whole GPU pass-through virtualization case, after VM reboot * some old smc fw still need driver do vPost otherwise gpu hang, while @@ -1547,7 +1653,7 @@ static int amdgpu_device_check_arguments(struct amdgpu_device *adev) dev_warn(adev->dev, "sched jobs (%d) must be at least 4\n", amdgpu_sched_jobs); amdgpu_sched_jobs = 4; - } else if (!is_power_of_2(amdgpu_sched_jobs)){ + } else if (!is_power_of_2(amdgpu_sched_jobs)) { dev_warn(adev->dev, "sched jobs (%d) must be a power of 2\n", amdgpu_sched_jobs); amdgpu_sched_jobs = roundup_pow_of_two(amdgpu_sched_jobs); @@ -2194,7 +2300,7 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) total = true; for (i = 0; i < adev->num_ip_blocks; i++) { if ((amdgpu_ip_block_mask & (1 << i)) == 0) { - DRM_ERROR("disabled ip block: %d <%s>\n", + DRM_WARN("disabled ip block: %d <%s>\n", i, adev->ip_blocks[i].version->funcs->name); adev->ip_blocks[i].status.valid = false; } else { @@ -2220,14 +2326,16 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) return r; /* Read BIOS */ - if (!amdgpu_get_bios(adev)) - return -EINVAL; + if (amdgpu_device_read_bios(adev)) { + if (!amdgpu_get_bios(adev)) + return -EINVAL; - r = amdgpu_atombios_init(adev); - if (r) { - dev_err(adev->dev, "amdgpu_atombios_init failed\n"); - amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_INIT_FAIL, 0, 0); - return r; + r = amdgpu_atombios_init(adev); + if (r) { + dev_err(adev->dev, "amdgpu_atombios_init failed\n"); + amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_INIT_FAIL, 0, 0); + return r; + } } /*get pf2vf msg info at it's earliest time*/ @@ -2376,6 +2484,8 @@ static int amdgpu_device_init_schedulers(struct amdgpu_device *adev) } } + amdgpu_xcp_update_partition_sched_list(adev); + return 0; } @@ -2533,8 +2643,10 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) goto init_failed; /* Don't init kfd if whole hive need to be reset during init */ - if (!adev->gmc.xgmi.pending_reset) + if (!adev->gmc.xgmi.pending_reset) { + kgd2kfd_init_zone_device(adev); amdgpu_amdkfd_device_init(adev); + } amdgpu_fru_get_product_info(adev); @@ -2759,8 +2871,9 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev) DRM_ERROR("enable mgpu fan boost failed (%d).\n", r); /* For passthrough configuration on arcturus and aldebaran, enable special handling SBR */ - if (amdgpu_passthrough(adev) && ((adev->asic_type == CHIP_ARCTURUS && adev->gmc.xgmi.num_physical_nodes > 1)|| - adev->asic_type == CHIP_ALDEBARAN )) + if (amdgpu_passthrough(adev) && + ((adev->asic_type == CHIP_ARCTURUS && adev->gmc.xgmi.num_physical_nodes > 1) || + adev->asic_type == CHIP_ALDEBARAN)) amdgpu_dpm_handle_passthrough_sbr(adev, true); if (adev->gmc.xgmi.num_physical_nodes > 1) { @@ -3089,7 +3202,7 @@ static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev) } adev->ip_blocks[i].status.hw = false; /* handle putting the SMC in the appropriate state */ - if(!amdgpu_sriov_vf(adev)){ + if (!amdgpu_sriov_vf(adev)) { if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC) { r = amdgpu_dpm_set_mp1_state(adev, adev->mp1_state); if (r) { @@ -3608,6 +3721,8 @@ int amdgpu_device_init(struct amdgpu_device *adev, adev->smc_wreg = &amdgpu_invalid_wreg; adev->pcie_rreg = &amdgpu_invalid_rreg; adev->pcie_wreg = &amdgpu_invalid_wreg; + adev->pcie_rreg_ext = &amdgpu_invalid_rreg_ext; + adev->pcie_wreg_ext = &amdgpu_invalid_wreg_ext; adev->pciep_rreg = &amdgpu_invalid_rreg; adev->pciep_wreg = &amdgpu_invalid_wreg; adev->pcie_rreg64 = &amdgpu_invalid_rreg64; @@ -3633,6 +3748,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, mutex_init(&adev->srbm_mutex); mutex_init(&adev->gfx.pipe_reserve_mutex); mutex_init(&adev->gfx.gfx_off_mutex); + mutex_init(&adev->gfx.partition_mutex); mutex_init(&adev->grbm_idx_mutex); mutex_init(&adev->mn_lock); mutex_init(&adev->virt.vf_errors.lock); @@ -3708,8 +3824,6 @@ int amdgpu_device_init(struct amdgpu_device *adev, DRM_INFO("register mmio base: 0x%08X\n", (uint32_t)adev->rmmio_base); DRM_INFO("register mmio size: %u\n", (unsigned)adev->rmmio_size); - amdgpu_device_get_pcie_info(adev); - if (amdgpu_mcbp) DRM_INFO("MCBP is enabled\n"); @@ -3725,6 +3839,8 @@ int amdgpu_device_init(struct amdgpu_device *adev, /* detect hw virtualization here */ amdgpu_detect_virtualization(adev); + amdgpu_device_get_pcie_info(adev); + r = amdgpu_device_get_job_timeout_settings(adev); if (r) { dev_err(adev->dev, "invalid lockup_timeout parameter syntax\n"); @@ -3753,21 +3869,24 @@ int amdgpu_device_init(struct amdgpu_device *adev, } /* enable PCIE atomic ops */ - if (amdgpu_sriov_vf(adev)) - adev->have_atomics_support = ((struct amd_sriov_msg_pf2vf_info *) - adev->virt.fw_reserve.p_pf2vf)->pcie_atomic_ops_support_flags == - (PCI_EXP_DEVCAP2_ATOMIC_COMP32 | PCI_EXP_DEVCAP2_ATOMIC_COMP64); + if (amdgpu_sriov_vf(adev)) { + if (adev->virt.fw_reserve.p_pf2vf) + adev->have_atomics_support = ((struct amd_sriov_msg_pf2vf_info *) + adev->virt.fw_reserve.p_pf2vf)->pcie_atomic_ops_support_flags == + (PCI_EXP_DEVCAP2_ATOMIC_COMP32 | PCI_EXP_DEVCAP2_ATOMIC_COMP64); /* APUs w/ gfx9 onwards doesn't reply on PCIe atomics, rather it is a * internal path natively support atomics, set have_atomics_support to true. */ - else if ((adev->flags & AMD_IS_APU) && - (adev->ip_versions[GC_HWIP][0] > IP_VERSION(9, 0, 0))) + } else if ((adev->flags & AMD_IS_APU) && + (adev->ip_versions[GC_HWIP][0] > IP_VERSION(9, 0, 0))) { adev->have_atomics_support = true; - else + } else { adev->have_atomics_support = !pci_enable_atomic_ops_to_root(adev->pdev, PCI_EXP_DEVCAP2_ATOMIC_COMP32 | PCI_EXP_DEVCAP2_ATOMIC_COMP64); + } + if (!adev->have_atomics_support) dev_info(adev->dev, "PCIE atomic ops is not supported\n"); @@ -3783,7 +3902,8 @@ int amdgpu_device_init(struct amdgpu_device *adev, amdgpu_reset_init(adev); /* detect if we are with an SRIOV vbios */ - amdgpu_device_detect_sriov_bios(adev); + if (adev->bios) + amdgpu_device_detect_sriov_bios(adev); /* check if we need to reset the asic * E.g., driver was not cleanly unloaded previously, etc. @@ -3835,25 +3955,27 @@ int amdgpu_device_init(struct amdgpu_device *adev, } } - if (adev->is_atom_fw) { - /* Initialize clocks */ - r = amdgpu_atomfirmware_get_clock_info(adev); - if (r) { - dev_err(adev->dev, "amdgpu_atomfirmware_get_clock_info failed\n"); - amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_GET_CLOCK_FAIL, 0, 0); - goto failed; - } - } else { - /* Initialize clocks */ - r = amdgpu_atombios_get_clock_info(adev); - if (r) { - dev_err(adev->dev, "amdgpu_atombios_get_clock_info failed\n"); - amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_GET_CLOCK_FAIL, 0, 0); - goto failed; + if (adev->bios) { + if (adev->is_atom_fw) { + /* Initialize clocks */ + r = amdgpu_atomfirmware_get_clock_info(adev); + if (r) { + dev_err(adev->dev, "amdgpu_atomfirmware_get_clock_info failed\n"); + amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_GET_CLOCK_FAIL, 0, 0); + goto failed; + } + } else { + /* Initialize clocks */ + r = amdgpu_atombios_get_clock_info(adev); + if (r) { + dev_err(adev->dev, "amdgpu_atombios_get_clock_info failed\n"); + amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_GET_CLOCK_FAIL, 0, 0); + goto failed; + } + /* init i2c buses */ + if (!amdgpu_device_has_dc_support(adev)) + amdgpu_atombios_i2c_init(adev); } - /* init i2c buses */ - if (!amdgpu_device_has_dc_support(adev)) - amdgpu_atombios_i2c_init(adev); } fence_driver_init: @@ -4019,7 +4141,7 @@ static void amdgpu_device_unmap_mmio(struct amdgpu_device *adev) adev->mman.aper_base_kaddr = NULL; /* Memory manager related */ - if (!adev->gmc.xgmi.connected_to_cpu) { + if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu) { arch_phys_wc_del(adev->gmc.vram_mtrr); arch_io_free_memtype_wc(adev->gmc.aper_base, adev->gmc.aper_size); } @@ -4049,7 +4171,7 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev) /* disable all interrupts */ amdgpu_irq_disable_all(adev); - if (adev->mode_info.mode_config_initialized){ + if (adev->mode_info.mode_config_initialized) { if (!drm_drv_uses_atomic_modeset(adev_to_drm(adev))) drm_helper_force_disable_all(adev_to_drm(adev)); else @@ -4714,42 +4836,42 @@ disabled: int amdgpu_device_mode1_reset(struct amdgpu_device *adev) { - u32 i; - int ret = 0; + u32 i; + int ret = 0; - amdgpu_atombios_scratch_regs_engine_hung(adev, true); + amdgpu_atombios_scratch_regs_engine_hung(adev, true); - dev_info(adev->dev, "GPU mode1 reset\n"); + dev_info(adev->dev, "GPU mode1 reset\n"); - /* disable BM */ - pci_clear_master(adev->pdev); + /* disable BM */ + pci_clear_master(adev->pdev); - amdgpu_device_cache_pci_state(adev->pdev); + amdgpu_device_cache_pci_state(adev->pdev); - if (amdgpu_dpm_is_mode1_reset_supported(adev)) { - dev_info(adev->dev, "GPU smu mode1 reset\n"); - ret = amdgpu_dpm_mode1_reset(adev); - } else { - dev_info(adev->dev, "GPU psp mode1 reset\n"); - ret = psp_gpu_reset(adev); - } + if (amdgpu_dpm_is_mode1_reset_supported(adev)) { + dev_info(adev->dev, "GPU smu mode1 reset\n"); + ret = amdgpu_dpm_mode1_reset(adev); + } else { + dev_info(adev->dev, "GPU psp mode1 reset\n"); + ret = psp_gpu_reset(adev); + } - if (ret) - dev_err(adev->dev, "GPU mode1 reset failed\n"); + if (ret) + dev_err(adev->dev, "GPU mode1 reset failed\n"); - amdgpu_device_load_pci_state(adev->pdev); + amdgpu_device_load_pci_state(adev->pdev); - /* wait for asic to come out of reset */ - for (i = 0; i < adev->usec_timeout; i++) { - u32 memsize = adev->nbio.funcs->get_memsize(adev); + /* wait for asic to come out of reset */ + for (i = 0; i < adev->usec_timeout; i++) { + u32 memsize = adev->nbio.funcs->get_memsize(adev); - if (memsize != 0xffffffff) - break; - udelay(1); - } + if (memsize != 0xffffffff) + break; + udelay(1); + } - amdgpu_atombios_scratch_regs_engine_hung(adev, false); - return ret; + amdgpu_atombios_scratch_regs_engine_hung(adev, false); + return ret; } int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, @@ -5478,7 +5600,7 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev) adev->pm.pcie_mlw_mask = amdgpu_pcie_lane_cap; /* covers APUs as well */ - if (pci_is_root_bus(adev->pdev->bus)) { + if (pci_is_root_bus(adev->pdev->bus) && !amdgpu_passthrough(adev)) { if (adev->pm.pcie_gen_mask == 0) adev->pm.pcie_gen_mask = AMDGPU_DEFAULT_PCIE_GEN_MASK; if (adev->pm.pcie_mlw_mask == 0) @@ -5959,6 +6081,7 @@ void amdgpu_device_halt(struct amdgpu_device *adev) struct pci_dev *pdev = adev->pdev; struct drm_device *ddev = adev_to_drm(adev); + amdgpu_xcp_dev_unplug(adev); drm_dev_unplug(ddev); amdgpu_irq_disable_all(adev); @@ -6079,3 +6202,31 @@ bool amdgpu_device_has_display_hardware(struct amdgpu_device *adev) return true; } } + +uint32_t amdgpu_device_wait_on_rreg(struct amdgpu_device *adev, + uint32_t inst, uint32_t reg_addr, char reg_name[], + uint32_t expected_value, uint32_t mask) +{ + uint32_t ret = 0; + uint32_t old_ = 0; + uint32_t tmp_ = RREG32(reg_addr); + uint32_t loop = adev->usec_timeout; + + while ((tmp_ & (mask)) != (expected_value)) { + if (old_ != tmp_) { + loop = adev->usec_timeout; + old_ = tmp_; + } else + udelay(1); + tmp_ = RREG32(reg_addr); + loop--; + if (!loop) { + DRM_WARN("Register(%d) [%s] failed to reach value 0x%08x != 0x%08xn", + inst, reg_name, (uint32_t)expected_value, + (uint32_t)(tmp_ & (mask))); + ret = -ETIMEDOUT; + break; + } + } + return ret; +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c index 0ecce0b92b82..8e1cfc87122d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c @@ -30,6 +30,7 @@ #include "soc15.h" #include "gfx_v9_0.h" +#include "gfx_v9_4_3.h" #include "gmc_v9_0.h" #include "df_v1_7.h" #include "df_v3_6.h" @@ -76,12 +77,15 @@ #include "jpeg_v3_0.h" #include "vcn_v4_0.h" #include "jpeg_v4_0.h" +#include "vcn_v4_0_3.h" +#include "jpeg_v4_0_3.h" #include "amdgpu_vkms.h" #include "mes_v10_1.h" #include "mes_v11_0.h" #include "smuio_v11_0.h" #include "smuio_v11_0_6.h" #include "smuio_v13_0.h" +#include "smuio_v13_0_3.h" #include "smuio_v13_0_6.h" #define FIRMWARE_IP_DISCOVERY "amdgpu/ip_discovery.bin" @@ -200,14 +204,44 @@ static int hw_id_map[MAX_HWIP] = { [PCIE_HWIP] = PCIE_HWID, }; -static int amdgpu_discovery_read_binary_from_vram(struct amdgpu_device *adev, uint8_t *binary) +static int amdgpu_discovery_read_binary_from_sysmem(struct amdgpu_device *adev, uint8_t *binary) +{ + u64 tmr_offset, tmr_size, pos; + void *discv_regn; + int ret; + + ret = amdgpu_acpi_get_tmr_info(adev, &tmr_offset, &tmr_size); + if (ret) + return ret; + + pos = tmr_offset + tmr_size - DISCOVERY_TMR_OFFSET; + + /* This region is read-only and reserved from system use */ + discv_regn = memremap(pos, adev->mman.discovery_tmr_size, MEMREMAP_WC); + if (discv_regn) { + memcpy(binary, discv_regn, adev->mman.discovery_tmr_size); + memunmap(discv_regn); + return 0; + } + + return -ENOENT; +} + +static int amdgpu_discovery_read_binary_from_mem(struct amdgpu_device *adev, + uint8_t *binary) { uint64_t vram_size = (uint64_t)RREG32(mmRCC_CONFIG_MEMSIZE) << 20; - uint64_t pos = vram_size - DISCOVERY_TMR_OFFSET; + int ret = 0; - amdgpu_device_vram_access(adev, pos, (uint32_t *)binary, - adev->mman.discovery_tmr_size, false); - return 0; + if (vram_size) { + uint64_t pos = vram_size - DISCOVERY_TMR_OFFSET; + amdgpu_device_vram_access(adev, pos, (uint32_t *)binary, + adev->mman.discovery_tmr_size, false); + } else { + ret = amdgpu_discovery_read_binary_from_sysmem(adev, binary); + } + + return ret; } static int amdgpu_discovery_read_binary_from_file(struct amdgpu_device *adev, uint8_t *binary) @@ -280,6 +314,7 @@ static void amdgpu_discovery_harvest_config_quirk(struct amdgpu_device *adev) case 0xCF: case 0xDF: adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1; + adev->vcn.inst_mask &= ~AMDGPU_VCN_HARVEST_VCN1; break; default: break; @@ -301,33 +336,30 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) if (!adev->mman.discovery_bin) return -ENOMEM; - r = amdgpu_discovery_read_binary_from_vram(adev, adev->mman.discovery_bin); - if (r) { - dev_err(adev->dev, "failed to read ip discovery binary from vram\n"); - r = -EINVAL; - goto out; - } - - if (!amdgpu_discovery_verify_binary_signature(adev->mman.discovery_bin) || amdgpu_discovery == 2) { - /* ignore the discovery binary from vram if discovery=2 in kernel module parameter */ - if (amdgpu_discovery == 2) - dev_info(adev->dev,"force read ip discovery binary from file"); - else - dev_warn(adev->dev, "get invalid ip discovery binary signature from vram\n"); - - /* retry read ip discovery binary from file */ + /* Read from file if it is the preferred option */ + if (amdgpu_discovery == 2) { + dev_info(adev->dev, "use ip discovery information from file"); r = amdgpu_discovery_read_binary_from_file(adev, adev->mman.discovery_bin); + if (r) { dev_err(adev->dev, "failed to read ip discovery binary from file\n"); r = -EINVAL; goto out; } - /* check the ip discovery binary signature */ - if(!amdgpu_discovery_verify_binary_signature(adev->mman.discovery_bin)) { - dev_warn(adev->dev, "get invalid ip discovery binary signature from file\n"); - r = -EINVAL; + + } else { + r = amdgpu_discovery_read_binary_from_mem( + adev, adev->mman.discovery_bin); + if (r) goto out; - } + } + + /* check the ip discovery binary signature */ + if (!amdgpu_discovery_verify_binary_signature(adev->mman.discovery_bin)) { + dev_err(adev->dev, + "get invalid ip discovery binary signature\n"); + r = -EINVAL; + goto out; } bhdr = (struct binary_header *)adev->mman.discovery_bin; @@ -471,11 +503,11 @@ void amdgpu_discovery_fini(struct amdgpu_device *adev) adev->mman.discovery_bin = NULL; } -static int amdgpu_discovery_validate_ip(const struct ip *ip) +static int amdgpu_discovery_validate_ip(const struct ip_v4 *ip) { - if (ip->number_instance >= HWIP_MAX_INSTANCE) { - DRM_ERROR("Unexpected number_instance (%d) from ip discovery blob\n", - ip->number_instance); + if (ip->instance_number >= HWIP_MAX_INSTANCE) { + DRM_ERROR("Unexpected instance_number (%d) from ip discovery blob\n", + ip->instance_number); return -EINVAL; } if (le16_to_cpu(ip->hw_id) >= HW_ID_MAX) { @@ -493,7 +525,7 @@ static void amdgpu_discovery_read_harvest_bit_per_ip(struct amdgpu_device *adev, struct binary_header *bhdr; struct ip_discovery_header *ihdr; struct die_header *dhdr; - struct ip *ip; + struct ip_v4 *ip; uint16_t die_offset, ip_offset, num_dies, num_ips; int i, j; @@ -510,29 +542,41 @@ static void amdgpu_discovery_read_harvest_bit_per_ip(struct amdgpu_device *adev, ip_offset = die_offset + sizeof(*dhdr); for (j = 0; j < num_ips; j++) { - ip = (struct ip *)(adev->mman.discovery_bin + ip_offset); + ip = (struct ip_v4 *)(adev->mman.discovery_bin + ip_offset); if (amdgpu_discovery_validate_ip(ip)) goto next_ip; - if (le16_to_cpu(ip->harvest) == 1) { + if (le16_to_cpu(ip->variant) == 1) { switch (le16_to_cpu(ip->hw_id)) { case VCN_HWID: (*vcn_harvest_count)++; - if (ip->number_instance == 0) + if (ip->instance_number == 0) { adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN0; - else + adev->vcn.inst_mask &= + ~AMDGPU_VCN_HARVEST_VCN0; + adev->jpeg.inst_mask &= + ~AMDGPU_VCN_HARVEST_VCN0; + } else { adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1; + adev->vcn.inst_mask &= + ~AMDGPU_VCN_HARVEST_VCN1; + adev->jpeg.inst_mask &= + ~AMDGPU_VCN_HARVEST_VCN1; + } break; case DMU_HWID: adev->harvest_ip_mask |= AMD_HARVEST_IP_DMU_MASK; break; default: break; - } - } + } + } next_ip: - ip_offset += struct_size(ip, base_address, ip->num_base_address); + if (ihdr->base_addr_64_bit) + ip_offset += struct_size(ip, base_address_64, ip->num_base_address); + else + ip_offset += struct_size(ip, base_address, ip->num_base_address); } } } @@ -564,10 +608,15 @@ static void amdgpu_discovery_read_from_harvest_table(struct amdgpu_device *adev, switch (le16_to_cpu(harvest_info->list[i].hw_id)) { case VCN_HWID: (*vcn_harvest_count)++; - if (harvest_info->list[i].number_instance == 0) - adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN0; - else - adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1; + adev->vcn.harvest_config |= + (1 << harvest_info->list[i].number_instance); + adev->jpeg.harvest_config |= + (1 << harvest_info->list[i].number_instance); + + adev->vcn.inst_mask &= + ~(1U << harvest_info->list[i].number_instance); + adev->jpeg.inst_mask &= + ~(1U << harvest_info->list[i].number_instance); break; case DMU_HWID: adev->harvest_ip_mask |= AMD_HARVEST_IP_DMU_MASK; @@ -577,6 +626,14 @@ static void amdgpu_discovery_read_from_harvest_table(struct amdgpu_device *adev, 1 << (le16_to_cpu(harvest_info->list[i].number_instance)); (*umc_harvest_count)++; break; + case GC_HWID: + adev->gfx.xcc_mask &= + ~(1U << harvest_info->list[i].number_instance); + break; + case SDMA0_HWID: + adev->sdma.sdma_mask &= + ~(1U << harvest_info->list[i].number_instance); + break; default: break; } @@ -836,9 +893,40 @@ static void ip_disc_release(struct kobject *kobj) kfree(ip_top); } +static uint8_t amdgpu_discovery_get_harvest_info(struct amdgpu_device *adev, + uint16_t hw_id, uint8_t inst) +{ + uint8_t harvest = 0; + + /* Until a uniform way is figured, get mask based on hwid */ + switch (hw_id) { + case VCN_HWID: + harvest = ((1 << inst) & adev->vcn.inst_mask) == 0; + break; + case DMU_HWID: + if (adev->harvest_ip_mask & AMD_HARVEST_IP_DMU_MASK) + harvest = 0x1; + break; + case UMC_HWID: + /* TODO: It needs another parsing; for now, ignore.*/ + break; + case GC_HWID: + harvest = ((1 << inst) & adev->gfx.xcc_mask) == 0; + break; + case SDMA0_HWID: + harvest = ((1 << inst) & adev->sdma.sdma_mask) == 0; + break; + default: + break; + } + + return harvest; +} + static int amdgpu_discovery_sysfs_ips(struct amdgpu_device *adev, struct ip_die_entry *ip_die_entry, - const size_t _ip_offset, const int num_ips) + const size_t _ip_offset, const int num_ips, + bool reg_base_64) { int ii, jj, kk, res; @@ -852,10 +940,10 @@ static int amdgpu_discovery_sysfs_ips(struct amdgpu_device *adev, size_t ip_offset = _ip_offset; for (jj = 0; jj < num_ips; jj++) { - struct ip *ip; + struct ip_v4 *ip; struct ip_hw_instance *ip_hw_instance; - ip = (struct ip *)(adev->mman.discovery_bin + ip_offset); + ip = (struct ip_v4 *)(adev->mman.discovery_bin + ip_offset); if (amdgpu_discovery_validate_ip(ip) || le16_to_cpu(ip->hw_id) != ii) goto next_ip; @@ -903,22 +991,35 @@ static int amdgpu_discovery_sysfs_ips(struct amdgpu_device *adev, return -ENOMEM; } ip_hw_instance->hw_id = le16_to_cpu(ip->hw_id); /* == ii */ - ip_hw_instance->num_instance = ip->number_instance; + ip_hw_instance->num_instance = ip->instance_number; ip_hw_instance->major = ip->major; ip_hw_instance->minor = ip->minor; ip_hw_instance->revision = ip->revision; - ip_hw_instance->harvest = ip->harvest; + ip_hw_instance->harvest = + amdgpu_discovery_get_harvest_info( + adev, ip_hw_instance->hw_id, + ip_hw_instance->num_instance); ip_hw_instance->num_base_addresses = ip->num_base_address; - for (kk = 0; kk < ip_hw_instance->num_base_addresses; kk++) - ip_hw_instance->base_addr[kk] = ip->base_address[kk]; + for (kk = 0; kk < ip_hw_instance->num_base_addresses; kk++) { + if (reg_base_64) + ip_hw_instance->base_addr[kk] = + lower_32_bits(le64_to_cpu(ip->base_address_64[kk])) & 0x3FFFFFFF; + else + ip_hw_instance->base_addr[kk] = ip->base_address[kk]; + } kobject_init(&ip_hw_instance->kobj, &ip_hw_instance_ktype); ip_hw_instance->kobj.kset = &ip_hw_id->hw_id_kset; res = kobject_add(&ip_hw_instance->kobj, NULL, "%d", ip_hw_instance->num_instance); next_ip: - ip_offset += struct_size(ip, base_address, ip->num_base_address); + if (reg_base_64) + ip_offset += struct_size(ip, base_address_64, + ip->num_base_address); + else + ip_offset += struct_size(ip, base_address, + ip->num_base_address); } } @@ -972,7 +1073,7 @@ static int amdgpu_discovery_sysfs_recurse(struct amdgpu_device *adev) return res; } - amdgpu_discovery_sysfs_ips(adev, ip_die_entry, ip_offset, num_ips); + amdgpu_discovery_sysfs_ips(adev, ip_die_entry, ip_offset, num_ips, !!ihdr->base_addr_64_bit); } return 0; @@ -983,6 +1084,9 @@ static int amdgpu_discovery_sysfs_init(struct amdgpu_device *adev) struct kset *die_kset; int res, ii; + if (!adev->mman.discovery_bin) + return -EINVAL; + adev->ip_top = kzalloc(sizeof(*adev->ip_top), GFP_KERNEL); if (!adev->ip_top) return -ENOMEM; @@ -1082,7 +1186,7 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) struct binary_header *bhdr; struct ip_discovery_header *ihdr; struct die_header *dhdr; - struct ip *ip; + struct ip_v4 *ip; uint16_t die_offset; uint16_t ip_offset; uint16_t num_dies; @@ -1098,6 +1202,10 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) return r; } + adev->gfx.xcc_mask = 0; + adev->sdma.sdma_mask = 0; + adev->vcn.inst_mask = 0; + adev->jpeg.inst_mask = 0; bhdr = (struct binary_header *)adev->mman.discovery_bin; ihdr = (struct ip_discovery_header *)(adev->mman.discovery_bin + le16_to_cpu(bhdr->table_list[IP_DISCOVERY].offset)); @@ -1121,7 +1229,7 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) le16_to_cpu(dhdr->die_id), num_ips); for (j = 0; j < num_ips; j++) { - ip = (struct ip *)(adev->mman.discovery_bin + ip_offset); + ip = (struct ip_v4 *)(adev->mman.discovery_bin + ip_offset); if (amdgpu_discovery_validate_ip(ip)) goto next_ip; @@ -1131,7 +1239,7 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) DRM_DEBUG("%s(%d) #%d v%d.%d.%d:\n", hw_id_names[le16_to_cpu(ip->hw_id)], le16_to_cpu(ip->hw_id), - ip->number_instance, + ip->instance_number, ip->major, ip->minor, ip->revision); @@ -1145,23 +1253,33 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) adev->vcn.vcn_config[adev->vcn.num_vcn_inst] = ip->revision & 0xc0; ip->revision &= ~0xc0; - if (adev->vcn.num_vcn_inst < AMDGPU_MAX_VCN_INSTANCES) + if (adev->vcn.num_vcn_inst < + AMDGPU_MAX_VCN_INSTANCES) { adev->vcn.num_vcn_inst++; - else + adev->vcn.inst_mask |= + (1U << ip->instance_number); + adev->jpeg.inst_mask |= + (1U << ip->instance_number); + } else { dev_err(adev->dev, "Too many VCN instances: %d vs %d\n", adev->vcn.num_vcn_inst + 1, AMDGPU_MAX_VCN_INSTANCES); + } } if (le16_to_cpu(ip->hw_id) == SDMA0_HWID || le16_to_cpu(ip->hw_id) == SDMA1_HWID || le16_to_cpu(ip->hw_id) == SDMA2_HWID || le16_to_cpu(ip->hw_id) == SDMA3_HWID) { - if (adev->sdma.num_instances < AMDGPU_MAX_SDMA_INSTANCES) + if (adev->sdma.num_instances < + AMDGPU_MAX_SDMA_INSTANCES) { adev->sdma.num_instances++; - else + adev->sdma.sdma_mask |= + (1U << ip->instance_number); + } else { dev_err(adev->dev, "Too many SDMA instances: %d vs %d\n", adev->sdma.num_instances + 1, AMDGPU_MAX_SDMA_INSTANCES); + } } if (le16_to_cpu(ip->hw_id) == UMC_HWID) { @@ -1169,20 +1287,38 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) adev->umc.node_inst_num++; } + if (le16_to_cpu(ip->hw_id) == GC_HWID) + adev->gfx.xcc_mask |= + (1U << ip->instance_number); + for (k = 0; k < num_base_address; k++) { /* * convert the endianness of base addresses in place, * so that we don't need to convert them when accessing adev->reg_offset. */ - ip->base_address[k] = le32_to_cpu(ip->base_address[k]); + if (ihdr->base_addr_64_bit) + /* Truncate the 64bit base address from ip discovery + * and only store lower 32bit ip base in reg_offset[]. + * Bits > 32 follows ASIC specific format, thus just + * discard them and handle it within specific ASIC. + * By this way reg_offset[] and related helpers can + * stay unchanged. + * The base address is in dwords, thus clear the + * highest 2 bits to store. + */ + ip->base_address[k] = + lower_32_bits(le64_to_cpu(ip->base_address_64[k])) & 0x3FFFFFFF; + else + ip->base_address[k] = le32_to_cpu(ip->base_address[k]); DRM_DEBUG("\t0x%08x\n", ip->base_address[k]); } for (hw_ip = 0; hw_ip < MAX_HWIP; hw_ip++) { - if (hw_id_map[hw_ip] == le16_to_cpu(ip->hw_id)) { + if (hw_id_map[hw_ip] == le16_to_cpu(ip->hw_id) && + hw_id_map[hw_ip] != 0) { DRM_DEBUG("set register base offset for %s\n", hw_id_names[le16_to_cpu(ip->hw_id)]); - adev->reg_offset[hw_ip][ip->number_instance] = + adev->reg_offset[hw_ip][ip->instance_number] = ip->base_address; /* Instance support is somewhat inconsistent. * SDMA is a good example. Sienna cichlid has 4 total @@ -1193,69 +1329,22 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) * example. On most chips there are multiple instances * with the same HWID. */ - adev->ip_versions[hw_ip][ip->number_instance] = + adev->ip_versions[hw_ip][ip->instance_number] = IP_VERSION(ip->major, ip->minor, ip->revision); } } next_ip: - ip_offset += struct_size(ip, base_address, ip->num_base_address); + if (ihdr->base_addr_64_bit) + ip_offset += struct_size(ip, base_address_64, ip->num_base_address); + else + ip_offset += struct_size(ip, base_address, ip->num_base_address); } } - amdgpu_discovery_sysfs_init(adev); - return 0; } -int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, int number_instance, - int *major, int *minor, int *revision) -{ - struct binary_header *bhdr; - struct ip_discovery_header *ihdr; - struct die_header *dhdr; - struct ip *ip; - uint16_t die_offset; - uint16_t ip_offset; - uint16_t num_dies; - uint16_t num_ips; - int i, j; - - if (!adev->mman.discovery_bin) { - DRM_ERROR("ip discovery uninitialized\n"); - return -EINVAL; - } - - bhdr = (struct binary_header *)adev->mman.discovery_bin; - ihdr = (struct ip_discovery_header *)(adev->mman.discovery_bin + - le16_to_cpu(bhdr->table_list[IP_DISCOVERY].offset)); - num_dies = le16_to_cpu(ihdr->num_dies); - - for (i = 0; i < num_dies; i++) { - die_offset = le16_to_cpu(ihdr->die_info[i].die_offset); - dhdr = (struct die_header *)(adev->mman.discovery_bin + die_offset); - num_ips = le16_to_cpu(dhdr->num_ips); - ip_offset = die_offset + sizeof(*dhdr); - - for (j = 0; j < num_ips; j++) { - ip = (struct ip *)(adev->mman.discovery_bin + ip_offset); - - if ((le16_to_cpu(ip->hw_id) == hw_id) && (ip->number_instance == number_instance)) { - if (major) - *major = ip->major; - if (minor) - *minor = ip->minor; - if (revision) - *revision = ip->revision; - return 0; - } - ip_offset += struct_size(ip, base_address, ip->num_base_address); - } - } - - return -EINVAL; -} - static void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev) { int vcn_harvest_count = 0; @@ -1266,7 +1355,8 @@ static void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev) * so read harvest bit per IP data structure to set * harvest configuration. */ - if (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 2, 0)) { + if (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 2, 0) && + adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 3)) { if ((adev->pdev->device == 0x731E && (adev->pdev->revision == 0xC6 || adev->pdev->revision == 0xC7)) || @@ -1425,6 +1515,7 @@ static int amdgpu_discovery_get_mall_info(struct amdgpu_device *adev) mall_size += mall_size_per_umc; } adev->gmc.mall_size = mall_size; + adev->gmc.m_half_use = half_use; break; default: dev_err(adev->dev, @@ -1706,6 +1797,7 @@ static int amdgpu_discovery_set_smu_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(13, 0, 3): case IP_VERSION(13, 0, 4): case IP_VERSION(13, 0, 5): + case IP_VERSION(13, 0, 6): case IP_VERSION(13, 0, 7): case IP_VERSION(13, 0, 8): case IP_VERSION(13, 0, 10): @@ -1804,6 +1896,11 @@ static int amdgpu_discovery_set_gc_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(9, 4, 2): amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); break; + case IP_VERSION(9, 4, 3): + if (!amdgpu_exp_hw_support) + return -EINVAL; + amdgpu_device_ip_block_add(adev, &gfx_v9_4_3_ip_block); + break; case IP_VERSION(10, 1, 10): case IP_VERSION(10, 1, 2): case IP_VERSION(10, 1, 1): @@ -1939,7 +2036,6 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(3, 1, 1): case IP_VERSION(3, 1, 2): case IP_VERSION(3, 0, 2): - case IP_VERSION(3, 0, 192): amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block); if (!amdgpu_sriov_vf(adev)) amdgpu_device_ip_block_add(adev, &jpeg_v3_0_ip_block); @@ -1952,7 +2048,11 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(4, 0, 4): amdgpu_device_ip_block_add(adev, &vcn_v4_0_ip_block); amdgpu_device_ip_block_add(adev, &jpeg_v4_0_ip_block); - return 0; + break; + case IP_VERSION(4, 0, 3): + amdgpu_device_ip_block_add(adev, &vcn_v4_0_3_ip_block); + amdgpu_device_ip_block_add(adev, &jpeg_v4_0_3_ip_block); + break; default: dev_err(adev->dev, "Failed to add vcn/jpeg ip block(UVD_HWIP:0x%x)\n", @@ -2000,6 +2100,17 @@ static int amdgpu_discovery_set_mes_ip_blocks(struct amdgpu_device *adev) return 0; } +static void amdgpu_discovery_init_soc_config(struct amdgpu_device *adev) +{ + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 4, 3): + aqua_vanjaram_init_soc_config(adev); + break; + default: + break; + } +} + int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) { int r; @@ -2177,6 +2288,9 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) break; } + amdgpu_discovery_init_soc_config(adev); + amdgpu_discovery_sysfs_init(adev); + switch (adev->ip_versions[GC_HWIP][0]) { case IP_VERSION(9, 0, 1): case IP_VERSION(9, 2, 1): @@ -2387,6 +2501,12 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(13, 0, 2): adev->smuio.funcs = &smuio_v13_0_funcs; break; + case IP_VERSION(13, 0, 3): + adev->smuio.funcs = &smuio_v13_0_3_funcs; + if (adev->smuio.funcs->get_pkg_type(adev) == AMDGPU_PKG_TYPE_APU) { + adev->flags |= AMD_IS_APU; + } + break; case IP_VERSION(13, 0, 6): case IP_VERSION(13, 0, 8): adev->smuio.funcs = &smuio_v13_0_6_funcs; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h index 8563dd4a7dc2..3a2f347bd50d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h @@ -24,12 +24,10 @@ #ifndef __AMDGPU_DISCOVERY__ #define __AMDGPU_DISCOVERY__ -#define DISCOVERY_TMR_SIZE (4 << 10) +#define DISCOVERY_TMR_SIZE (8 << 10) #define DISCOVERY_TMR_OFFSET (64 << 10) void amdgpu_discovery_fini(struct amdgpu_device *adev); -int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, int number_instance, - int *major, int *minor, int *revision); int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev); #endif /* __AMDGPU_DISCOVERY__ */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c index d60fe7eb5579..b702f499f5fb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c @@ -98,7 +98,7 @@ static void amdgpu_display_flip_callback(struct dma_fence *f, static bool amdgpu_display_flip_handle_fence(struct amdgpu_flip_work *work, struct dma_fence **f) { - struct dma_fence *fence= *f; + struct dma_fence *fence = *f; if (fence == NULL) return false; @@ -1252,21 +1252,21 @@ const struct drm_mode_config_funcs amdgpu_mode_funcs = { .fb_create = amdgpu_display_user_framebuffer_create, }; -static const struct drm_prop_enum_list amdgpu_underscan_enum_list[] = -{ { UNDERSCAN_OFF, "off" }, +static const struct drm_prop_enum_list amdgpu_underscan_enum_list[] = { + { UNDERSCAN_OFF, "off" }, { UNDERSCAN_ON, "on" }, { UNDERSCAN_AUTO, "auto" }, }; -static const struct drm_prop_enum_list amdgpu_audio_enum_list[] = -{ { AMDGPU_AUDIO_DISABLE, "off" }, +static const struct drm_prop_enum_list amdgpu_audio_enum_list[] = { + { AMDGPU_AUDIO_DISABLE, "off" }, { AMDGPU_AUDIO_ENABLE, "on" }, { AMDGPU_AUDIO_AUTO, "auto" }, }; /* XXX support different dither options? spatial, temporal, both, etc. */ -static const struct drm_prop_enum_list amdgpu_dither_enum_list[] = -{ { AMDGPU_FMT_DITHER_DISABLE, "off" }, +static const struct drm_prop_enum_list amdgpu_dither_enum_list[] = { + { AMDGPU_FMT_DITHER_DISABLE, "off" }, { AMDGPU_FMT_DITHER_ENABLE, "on" }, }; @@ -1496,8 +1496,7 @@ int amdgpu_display_get_crtc_scanoutpos(struct drm_device *dev, ret |= DRM_SCANOUTPOS_ACCURATE; vbl_start = vbl & 0x1fff; vbl_end = (vbl >> 16) & 0x1fff; - } - else { + } else { /* No: Fake something reasonable which gives at least ok results. */ vbl_start = mode->crtc_vdisplay; vbl_end = 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c index 0c001bb8fc2b..12210598e5b8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c @@ -149,7 +149,7 @@ static struct sg_table *amdgpu_dma_buf_map(struct dma_buf_attachment *attach, if (!bo->tbo.pin_count) { /* move buffer into GTT or VRAM */ struct ttm_operation_ctx ctx = { false, false }; - unsigned domains = AMDGPU_GEM_DOMAIN_GTT; + unsigned int domains = AMDGPU_GEM_DOMAIN_GTT; if (bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM && attach->peer2peer) { @@ -336,7 +336,7 @@ amdgpu_dma_buf_create_obj(struct drm_device *dev, struct dma_buf *dma_buf) ret = amdgpu_gem_object_create(adev, dma_buf->size, PAGE_SIZE, AMDGPU_GEM_DOMAIN_CPU, flags, - ttm_bo_type_sg, resv, &gobj); + ttm_bo_type_sg, resv, &gobj, 0); if (ret) goto error; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h index 8fd11497faba..f637574644c0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h @@ -59,7 +59,7 @@ struct amdgpu_doorbell_index { uint32_t gfx_ring1; uint32_t gfx_userqueue_start; uint32_t gfx_userqueue_end; - uint32_t sdma_engine[8]; + uint32_t sdma_engine[16]; uint32_t mes_ring0; uint32_t mes_ring1; uint32_t ih; @@ -86,6 +86,8 @@ struct amdgpu_doorbell_index { uint32_t max_assignment; /* Per engine SDMA doorbell size in dword */ uint32_t sdma_doorbell_range; + /* Per xcc doorbell size for KIQ/KCQ */ + uint32_t xcc_doorbell_range; }; typedef enum _AMDGPU_DOORBELL_ASSIGNMENT @@ -164,7 +166,15 @@ typedef enum _AMDGPU_VEGA20_DOORBELL_ASSIGNMENT AMDGPU_VEGA20_DOORBELL64_FIRST_NON_CP = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE0, AMDGPU_VEGA20_DOORBELL64_LAST_NON_CP = AMDGPU_VEGA20_DOORBELL64_VCE_RING6_7, - AMDGPU_VEGA20_DOORBELL_MAX_ASSIGNMENT = 0x18F, + /* kiq/kcq from second XCD. Max 8 XCDs */ + AMDGPU_VEGA20_DOORBELL_XCC1_KIQ_START = 0x190, + /* 8 compute rings per GC. Max to 0x1CE */ + AMDGPU_VEGA20_DOORBELL_XCC1_MEC_RING0_START = 0x197, + + /* AID1 SDMA: 0x1D0 ~ 0x1F7 */ + AMDGPU_VEGA20_DOORBELL_AID1_sDMA_START = 0x1D0, + + AMDGPU_VEGA20_DOORBELL_MAX_ASSIGNMENT = 0x1F7, AMDGPU_VEGA20_DOORBELL_INVALID = 0xFFFF } AMDGPU_VEGA20_DOORBELL_ASSIGNMENT; @@ -301,6 +311,36 @@ typedef enum _AMDGPU_DOORBELL64_ASSIGNMENT AMDGPU_DOORBELL64_INVALID = 0xFFFF } AMDGPU_DOORBELL64_ASSIGNMENT; +typedef enum _AMDGPU_DOORBELL_ASSIGNMENT_LAYOUT1 { + /* XCC0: 0x00 ~20, XCC1: 20 ~ 2F ... */ + + /* KIQ/HIQ/DIQ */ + AMDGPU_DOORBELL_LAYOUT1_KIQ_START = 0x000, + AMDGPU_DOORBELL_LAYOUT1_HIQ = 0x001, + AMDGPU_DOORBELL_LAYOUT1_DIQ = 0x002, + /* Compute: 0x08 ~ 0x20 */ + AMDGPU_DOORBELL_LAYOUT1_MEC_RING_START = 0x008, + AMDGPU_DOORBELL_LAYOUT1_MEC_RING_END = 0x00F, + AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_START = 0x010, + AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_END = 0x01F, + AMDGPU_DOORBELL_LAYOUT1_XCC_RANGE = 0x020, + + /* SDMA: 0x100 ~ 0x19F */ + AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START = 0x100, + AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_END = 0x19F, + /* IH: 0x1A0 ~ 0x1AF */ + AMDGPU_DOORBELL_LAYOUT1_IH = 0x1A0, + /* VCN: 0x1B0 ~ 0x1D4 */ + AMDGPU_DOORBELL_LAYOUT1_VCN_START = 0x1B0, + AMDGPU_DOORBELL_LAYOUT1_VCN_END = 0x1D4, + + AMDGPU_DOORBELL_LAYOUT1_FIRST_NON_CP = AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START, + AMDGPU_DOORBELL_LAYOUT1_LAST_NON_CP = AMDGPU_DOORBELL_LAYOUT1_VCN_END, + + AMDGPU_DOORBELL_LAYOUT1_MAX_ASSIGNMENT = 0x1D4, + AMDGPU_DOORBELL_LAYOUT1_INVALID = 0xFFFF +} AMDGPU_DOORBELL_ASSIGNMENT_LAYOUT1; + u32 amdgpu_mm_rdoorbell(struct amdgpu_device *adev, u32 index); void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v); u64 amdgpu_mm_rdoorbell64(struct amdgpu_device *adev, u32 index); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 7681f79f462e..43613569801b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -50,6 +50,7 @@ #include "amdgpu_ras.h" #include "amdgpu_xgmi.h" #include "amdgpu_reset.h" +#include "../amdxcp/amdgpu_xcp_drv.h" /* * KMS wrapper. @@ -110,9 +111,11 @@ * 3.52.0 - Add AMDGPU_IDS_FLAGS_CONFORMANT_TRUNC_COORD, add device_info fields: * tcp_cache_size, num_sqc_per_wgp, sqc_data_cache_size, sqc_inst_cache_size, * gl1c_cache_size, gl2c_cache_size, mall_size, enabled_rb_pipes_mask_hi + * 3.53.0 - Support for GFX11 CP GFX shadowing + * 3.54.0 - Add AMDGPU_CTX_QUERY2_FLAGS_RESET_IN_PROGRESS support */ #define KMS_DRIVER_MAJOR 3 -#define KMS_DRIVER_MINOR 52 +#define KMS_DRIVER_MINOR 54 #define KMS_DRIVER_PATCHLEVEL 0 unsigned int amdgpu_vram_limit = UINT_MAX; @@ -150,7 +153,7 @@ uint amdgpu_pg_mask = 0xffffffff; uint amdgpu_sdma_phase_quantum = 32; char *amdgpu_disable_cu; char *amdgpu_virtual_display; - +bool enforce_isolation; /* * OverDrive(bit 14) disabled by default * GFX DCS(bit 19) disabled by default @@ -191,6 +194,7 @@ int amdgpu_smartshift_bias; int amdgpu_use_xgmi_p2p = 1; int amdgpu_vcnfw_log; int amdgpu_sg_display = -1; /* auto */ +int amdgpu_user_partt_mode = AMDGPU_AUTO_COMPUTE_PARTITION_MODE; static void amdgpu_drv_delayed_reset_work_handler(struct work_struct *work); @@ -820,6 +824,13 @@ module_param_named(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm #endif /** + * DOC: mtype_local (int) + */ +int amdgpu_mtype_local; +MODULE_PARM_DESC(mtype_local, "MTYPE for local memory (0 = MTYPE_RW (default), 1 = MTYPE_NC, 2 = MTYPE_CC)"); +module_param_named(mtype_local, amdgpu_mtype_local, int, 0444); + +/** * DOC: pcie_p2p (bool) * Enable PCIe P2P (requires large-BAR). Default value: true (on) */ @@ -948,6 +959,28 @@ MODULE_PARM_DESC(smu_pptable_id, "specify pptable id to be used (-1 = auto(default) value, 0 = use pptable from vbios, > 0 = soft pptable id)"); module_param_named(smu_pptable_id, amdgpu_smu_pptable_id, int, 0444); +/** + * DOC: partition_mode (int) + * Used to override the default SPX mode. + */ +MODULE_PARM_DESC( + user_partt_mode, + "specify partition mode to be used (-2 = AMDGPU_AUTO_COMPUTE_PARTITION_MODE(default value) \ + 0 = AMDGPU_SPX_PARTITION_MODE, \ + 1 = AMDGPU_DPX_PARTITION_MODE, \ + 2 = AMDGPU_TPX_PARTITION_MODE, \ + 3 = AMDGPU_QPX_PARTITION_MODE, \ + 4 = AMDGPU_CPX_PARTITION_MODE)"); +module_param_named(user_partt_mode, amdgpu_user_partt_mode, uint, 0444); + + +/** + * DOC: enforce_isolation (bool) + * enforce process isolation between graphics and compute via using the same reserved vmid. + */ +module_param(enforce_isolation, bool, 0444); +MODULE_PARM_DESC(enforce_isolation, "enforce process isolation between graphics and compute . enforce_isolation = on"); + /* These devices are not supported by amdgpu. * They are supported by the mach64, r128, radeon drivers */ @@ -1615,6 +1648,7 @@ static const u16 amdgpu_unsupported_pciidlist[] = { 0x5874, 0x5940, 0x5941, + 0x5b70, 0x5b72, 0x5b73, 0x5b74, @@ -1660,7 +1694,7 @@ static const u16 amdgpu_unsupported_pciidlist[] = { }; static const struct pci_device_id pciidlist[] = { -#ifdef CONFIG_DRM_AMDGPU_SI +#ifdef CONFIG_DRM_AMDGPU_SI {0x1002, 0x6780, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI}, {0x1002, 0x6784, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI}, {0x1002, 0x6788, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI}, @@ -2017,6 +2051,11 @@ static const struct pci_device_id pciidlist[] = { .class_mask = 0xffffff, .driver_data = CHIP_IP_DISCOVERY }, + { PCI_DEVICE(0x1002, PCI_ANY_ID), + .class = PCI_CLASS_ACCELERATOR_PROCESSING << 8, + .class_mask = 0xffffff, + .driver_data = CHIP_IP_DISCOVERY }, + {0, 0, 0} }; @@ -2161,6 +2200,10 @@ retry_init: goto err_pci; } + ret = amdgpu_xcp_dev_register(adev, ent); + if (ret) + goto err_pci; + /* * 1. don't init fbdev on hw without DCE * 2. don't init fbdev if there are no connectors @@ -2233,6 +2276,7 @@ amdgpu_pci_remove(struct pci_dev *pdev) struct drm_device *dev = pci_get_drvdata(pdev); struct amdgpu_device *adev = drm_to_adev(dev); + amdgpu_xcp_dev_unplug(adev); drm_dev_unplug(dev); if (adev->pm.rpm_mode != AMDGPU_RUNPM_NONE) { @@ -2818,6 +2862,33 @@ static const struct drm_driver amdgpu_kms_driver = { .patchlevel = KMS_DRIVER_PATCHLEVEL, }; +const struct drm_driver amdgpu_partition_driver = { + .driver_features = + DRIVER_GEM | DRIVER_RENDER | DRIVER_SYNCOBJ | + DRIVER_SYNCOBJ_TIMELINE, + .open = amdgpu_driver_open_kms, + .postclose = amdgpu_driver_postclose_kms, + .lastclose = amdgpu_driver_lastclose_kms, + .ioctls = amdgpu_ioctls_kms, + .num_ioctls = ARRAY_SIZE(amdgpu_ioctls_kms), + .dumb_create = amdgpu_mode_dumb_create, + .dumb_map_offset = amdgpu_mode_dumb_mmap, + .fops = &amdgpu_driver_kms_fops, + .release = &amdgpu_driver_release_kms, + + .prime_handle_to_fd = drm_gem_prime_handle_to_fd, + .prime_fd_to_handle = drm_gem_prime_fd_to_handle, + .gem_prime_import = amdgpu_gem_prime_import, + .gem_prime_mmap = drm_gem_prime_mmap, + + .name = DRIVER_NAME, + .desc = DRIVER_DESC, + .date = DRIVER_DATE, + .major = KMS_DRIVER_MAJOR, + .minor = KMS_DRIVER_MINOR, + .patchlevel = KMS_DRIVER_PATCHLEVEL, +}; + static struct pci_error_handlers amdgpu_pci_err_handler = { .error_detected = amdgpu_pci_error_detected, .mmio_enabled = amdgpu_pci_mmio_enabled, @@ -2885,9 +2956,11 @@ static void __exit amdgpu_exit(void) amdgpu_amdkfd_fini(); pci_unregister_driver(&amdgpu_kms_pci_driver); amdgpu_unregister_atpx_handler(); + amdgpu_acpi_release(); amdgpu_sync_fini(); amdgpu_fence_slab_fini(); mmu_notifier_synchronize(); + amdgpu_xcp_drv_release(); } module_init(amdgpu_init); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h index 8178323e4bef..5bc2cb661af7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h @@ -42,6 +42,8 @@ #define DRIVER_DESC "AMD GPU" #define DRIVER_DATE "20150101" +extern const struct drm_driver amdgpu_partition_driver; + long amdgpu_drm_ioctl(struct file *filp, unsigned int cmd, unsigned long arg); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_encoders.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_encoders.c index 27a782a9dc72..3aaeed2d3562 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_encoders.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_encoders.c @@ -70,6 +70,7 @@ void amdgpu_encoder_set_active_device(struct drm_encoder *encoder) drm_for_each_connector_iter(connector, &iter) { if (connector->encoder == encoder) { struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector); + amdgpu_encoder->active_device = amdgpu_encoder->devices & amdgpu_connector->devices; DRM_DEBUG_KMS("setting active device to %08x from %08x %08x for encoder %d\n", amdgpu_encoder->active_device, amdgpu_encoder->devices, @@ -165,12 +166,12 @@ void amdgpu_panel_mode_fixup(struct drm_encoder *encoder, { struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder); struct drm_display_mode *native_mode = &amdgpu_encoder->native_mode; - unsigned hblank = native_mode->htotal - native_mode->hdisplay; - unsigned vblank = native_mode->vtotal - native_mode->vdisplay; - unsigned hover = native_mode->hsync_start - native_mode->hdisplay; - unsigned vover = native_mode->vsync_start - native_mode->vdisplay; - unsigned hsync_width = native_mode->hsync_end - native_mode->hsync_start; - unsigned vsync_width = native_mode->vsync_end - native_mode->vsync_start; + unsigned int hblank = native_mode->htotal - native_mode->hdisplay; + unsigned int vblank = native_mode->vtotal - native_mode->vdisplay; + unsigned int hover = native_mode->hsync_start - native_mode->hdisplay; + unsigned int vover = native_mode->vsync_start - native_mode->vdisplay; + unsigned int hsync_width = native_mode->hsync_end - native_mode->hsync_start; + unsigned int vsync_width = native_mode->vsync_end - native_mode->vsync_start; adjusted_mode->clock = native_mode->clock; adjusted_mode->flags = native_mode->flags; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c index f52d0ba91a77..c694b41f6461 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c @@ -42,7 +42,6 @@ #include "amdgpu_reset.h" /* - * Fences * Fences mark an event in the GPUs pipeline and are used * for GPU/CPU synchronization. When the fence is written, * it is expected that all buffers associated with that fence @@ -140,7 +139,7 @@ static u32 amdgpu_fence_read(struct amdgpu_ring *ring) * Returns 0 on success, -ENOMEM on failure. */ int amdgpu_fence_emit(struct amdgpu_ring *ring, struct dma_fence **f, struct amdgpu_job *job, - unsigned flags) + unsigned int flags) { struct amdgpu_device *adev = ring->adev; struct dma_fence *fence; @@ -174,11 +173,11 @@ int amdgpu_fence_emit(struct amdgpu_ring *ring, struct dma_fence **f, struct amd adev->fence_context + ring->idx, seq); /* Against remove in amdgpu_job_{free, free_cb} */ dma_fence_get(fence); - } - else + } else { dma_fence_init(fence, &amdgpu_fence_ops, &ring->fence_drv.lock, adev->fence_context + ring->idx, seq); + } } amdgpu_ring_emit_fence(ring, ring->fence_drv.gpu_addr, @@ -377,14 +376,11 @@ signed long amdgpu_fence_wait_polling(struct amdgpu_ring *ring, uint32_t wait_seq, signed long timeout) { - uint32_t seq; - - do { - seq = amdgpu_fence_read(ring); - udelay(5); - timeout -= 5; - } while ((int32_t)(wait_seq - seq) > 0 && timeout > 0); + while ((int32_t)(wait_seq - amdgpu_fence_read(ring)) > 0 && timeout > 0) { + udelay(2); + timeout -= 2; + } return timeout > 0 ? timeout : 0; } /** @@ -396,7 +392,7 @@ signed long amdgpu_fence_wait_polling(struct amdgpu_ring *ring, * Returns the number of emitted fences on the ring. Used by the * dynpm code to ring track activity. */ -unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring) +unsigned int amdgpu_fence_count_emitted(struct amdgpu_ring *ring) { uint64_t emitted; @@ -475,7 +471,7 @@ void amdgpu_fence_update_start_timestamp(struct amdgpu_ring *ring, uint32_t seq, */ int amdgpu_fence_driver_start_ring(struct amdgpu_ring *ring, struct amdgpu_irq_src *irq_src, - unsigned irq_type) + unsigned int irq_type) { struct amdgpu_device *adev = ring->adev; uint64_t index; @@ -582,7 +578,8 @@ void amdgpu_fence_driver_hw_fini(struct amdgpu_device *adev) if (r) amdgpu_fence_driver_force_completion(ring); - if (ring->fence_drv.irq_src) + if (!drm_dev_is_unplugged(adev_to_drm(adev)) && + ring->fence_drv.irq_src) amdgpu_irq_put(adev, ring->fence_drv.irq_src, ring->fence_drv.irq_type); @@ -653,6 +650,7 @@ void amdgpu_fence_driver_hw_init(struct amdgpu_device *adev) for (i = 0; i < AMDGPU_MAX_RINGS; i++) { struct amdgpu_ring *ring = adev->rings[i]; + if (!ring || !ring->fence_drv.initialized) continue; @@ -694,6 +692,30 @@ void amdgpu_fence_driver_clear_job_fences(struct amdgpu_ring *ring) } /** + * amdgpu_fence_driver_set_error - set error code on fences + * @ring: the ring which contains the fences + * @error: the error code to set + * + * Set an error code to all the fences pending on the ring. + */ +void amdgpu_fence_driver_set_error(struct amdgpu_ring *ring, int error) +{ + struct amdgpu_fence_driver *drv = &ring->fence_drv; + unsigned long flags; + + spin_lock_irqsave(&drv->lock, flags); + for (unsigned int i = 0; i <= drv->num_fences_mask; ++i) { + struct dma_fence *fence; + + fence = rcu_dereference_protected(drv->fences[i], + lockdep_is_held(&drv->lock)); + if (fence && !dma_fence_is_signaled_locked(fence)) + dma_fence_set_error(fence, error); + } + spin_unlock_irqrestore(&drv->lock, flags); +} + +/** * amdgpu_fence_driver_force_completion - force signal latest fence of ring * * @ring: fence of the ring to signal @@ -701,6 +723,7 @@ void amdgpu_fence_driver_clear_job_fences(struct amdgpu_ring *ring) */ void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring) { + amdgpu_fence_driver_set_error(ring, -ECANCELED); amdgpu_fence_write(ring, ring->fence_drv.sync_seq); amdgpu_fence_process(ring); } @@ -835,11 +858,12 @@ static const struct dma_fence_ops amdgpu_job_fence_ops = { #if defined(CONFIG_DEBUG_FS) static int amdgpu_debugfs_fence_info_show(struct seq_file *m, void *unused) { - struct amdgpu_device *adev = (struct amdgpu_device *)m->private; + struct amdgpu_device *adev = m->private; int i; for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = adev->rings[i]; + if (!ring || !ring->fence_drv.initialized) continue; @@ -913,6 +937,7 @@ static void amdgpu_debugfs_reset_work(struct work_struct *work) reset_work); struct amdgpu_reset_context reset_context; + memset(&reset_context, 0, sizeof(reset_context)); reset_context.method = AMD_RESET_METHOD_NONE; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c index 01cb89ffbd56..73b8cca35bab 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c @@ -35,6 +35,7 @@ #endif #include "amdgpu.h" #include <drm/drm_drv.h> +#include <drm/ttm/ttm_tt.h> /* * GART @@ -103,6 +104,142 @@ void amdgpu_gart_dummy_page_fini(struct amdgpu_device *adev) } /** + * amdgpu_gart_table_ram_alloc - allocate system ram for gart page table + * + * @adev: amdgpu_device pointer + * + * Allocate system memory for GART page table for ASICs that don't have + * dedicated VRAM. + * Returns 0 for success, error for failure. + */ +int amdgpu_gart_table_ram_alloc(struct amdgpu_device *adev) +{ + unsigned int order = get_order(adev->gart.table_size); + gfp_t gfp_flags = GFP_KERNEL | __GFP_ZERO; + struct amdgpu_bo *bo = NULL; + struct sg_table *sg = NULL; + struct amdgpu_bo_param bp; + dma_addr_t dma_addr; + struct page *p; + int ret; + + if (adev->gart.bo != NULL) + return 0; + + p = alloc_pages(gfp_flags, order); + if (!p) + return -ENOMEM; + + /* If the hardware does not support UTCL2 snooping of the CPU caches + * then set_memory_wc() could be used as a workaround to mark the pages + * as write combine memory. + */ + dma_addr = dma_map_page(&adev->pdev->dev, p, 0, adev->gart.table_size, + DMA_BIDIRECTIONAL); + if (dma_mapping_error(&adev->pdev->dev, dma_addr)) { + dev_err(&adev->pdev->dev, "Failed to DMA MAP the GART BO page\n"); + __free_pages(p, order); + p = NULL; + return -EFAULT; + } + + dev_info(adev->dev, "%s dma_addr:%pad\n", __func__, &dma_addr); + /* Create SG table */ + sg = kmalloc(sizeof(*sg), GFP_KERNEL); + if (!sg) { + ret = -ENOMEM; + goto error; + } + ret = sg_alloc_table(sg, 1, GFP_KERNEL); + if (ret) + goto error; + + sg_dma_address(sg->sgl) = dma_addr; + sg->sgl->length = adev->gart.table_size; +#ifdef CONFIG_NEED_SG_DMA_LENGTH + sg->sgl->dma_length = adev->gart.table_size; +#endif + /* Create SG BO */ + memset(&bp, 0, sizeof(bp)); + bp.size = adev->gart.table_size; + bp.byte_align = PAGE_SIZE; + bp.domain = AMDGPU_GEM_DOMAIN_CPU; + bp.type = ttm_bo_type_sg; + bp.resv = NULL; + bp.bo_ptr_size = sizeof(struct amdgpu_bo); + bp.flags = 0; + ret = amdgpu_bo_create(adev, &bp, &bo); + if (ret) + goto error; + + bo->tbo.sg = sg; + bo->tbo.ttm->sg = sg; + bo->allowed_domains = AMDGPU_GEM_DOMAIN_GTT; + bo->preferred_domains = AMDGPU_GEM_DOMAIN_GTT; + + ret = amdgpu_bo_reserve(bo, true); + if (ret) { + dev_err(adev->dev, "(%d) failed to reserve bo for GART system bo\n", ret); + goto error; + } + + ret = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT); + WARN(ret, "Pinning the GART table failed"); + if (ret) + goto error_resv; + + adev->gart.bo = bo; + adev->gart.ptr = page_to_virt(p); + /* Make GART table accessible in VMID0 */ + ret = amdgpu_ttm_alloc_gart(&adev->gart.bo->tbo); + if (ret) + amdgpu_gart_table_ram_free(adev); + amdgpu_bo_unreserve(bo); + + return 0; + +error_resv: + amdgpu_bo_unreserve(bo); +error: + amdgpu_bo_unref(&bo); + if (sg) { + sg_free_table(sg); + kfree(sg); + } + __free_pages(p, order); + return ret; +} + +/** + * amdgpu_gart_table_ram_free - free gart page table system ram + * + * @adev: amdgpu_device pointer + * + * Free the system memory used for the GART page tableon ASICs that don't + * have dedicated VRAM. + */ +void amdgpu_gart_table_ram_free(struct amdgpu_device *adev) +{ + unsigned int order = get_order(adev->gart.table_size); + struct sg_table *sg = adev->gart.bo->tbo.sg; + struct page *p; + int ret; + + ret = amdgpu_bo_reserve(adev->gart.bo, false); + if (!ret) { + amdgpu_bo_unpin(adev->gart.bo); + amdgpu_bo_unreserve(adev->gart.bo); + } + amdgpu_bo_unref(&adev->gart.bo); + sg_free_table(sg); + kfree(sg); + p = virt_to_page(adev->gart.ptr); + __free_pages(p, order); + + adev->gart.ptr = NULL; +} + +/** * amdgpu_gart_table_vram_alloc - allocate vram for gart page table * * @adev: amdgpu_device pointer @@ -182,7 +319,7 @@ void amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset, } mb(); amdgpu_device_flush_hdp(adev, NULL); - for (i = 0; i < adev->num_vmhubs; i++) + for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) amdgpu_gmc_flush_gpu_tlb(adev, 0, i, 0); drm_dev_exit(idx); @@ -264,7 +401,7 @@ void amdgpu_gart_invalidate_tlb(struct amdgpu_device *adev) mb(); amdgpu_device_flush_hdp(adev, NULL); - for (i = 0; i < adev->num_vmhubs; i++) + for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) amdgpu_gmc_flush_gpu_tlb(adev, 0, i, 0); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h index 8fea3e04e411..8283d682f543 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h @@ -51,6 +51,8 @@ struct amdgpu_gart { uint64_t gart_pte_flags; }; +int amdgpu_gart_table_ram_alloc(struct amdgpu_device *adev); +void amdgpu_gart_table_ram_free(struct amdgpu_device *adev); int amdgpu_gart_table_vram_alloc(struct amdgpu_device *adev); void amdgpu_gart_table_vram_free(struct amdgpu_device *adev); int amdgpu_gart_table_vram_pin(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index 863cb668e000..74055cba3dc9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -98,7 +98,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size, int alignment, u32 initial_domain, u64 flags, enum ttm_bo_type type, struct dma_resv *resv, - struct drm_gem_object **obj) + struct drm_gem_object **obj, int8_t xcp_id_plus1) { struct amdgpu_bo *bo; struct amdgpu_bo_user *ubo; @@ -116,6 +116,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size, bp.flags = flags; bp.domain = initial_domain; bp.bo_ptr_size = sizeof(struct amdgpu_bo); + bp.xcp_id_plus1 = xcp_id_plus1; r = amdgpu_bo_create_user(adev, &bp, &ubo); if (r) @@ -336,7 +337,7 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data, retry: r = amdgpu_gem_object_create(adev, size, args->in.alignment, initial_domain, - flags, ttm_bo_type_device, resv, &gobj); + flags, ttm_bo_type_device, resv, &gobj, fpriv->xcp_id + 1); if (r && r != -ERESTARTSYS) { if (flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED) { flags &= ~AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED; @@ -379,6 +380,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data, struct ttm_operation_ctx ctx = { true, false }; struct amdgpu_device *adev = drm_to_adev(dev); struct drm_amdgpu_gem_userptr *args = data; + struct amdgpu_fpriv *fpriv = filp->driver_priv; struct drm_gem_object *gobj; struct hmm_range *range; struct amdgpu_bo *bo; @@ -405,7 +407,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data, /* create a gem object to contain this object in */ r = amdgpu_gem_object_create(adev, args->size, 0, AMDGPU_GEM_DOMAIN_CPU, - 0, ttm_bo_type_device, NULL, &gobj); + 0, ttm_bo_type_device, NULL, &gobj, fpriv->xcp_id + 1); if (r) return r; @@ -908,6 +910,7 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv, struct drm_mode_create_dumb *args) { struct amdgpu_device *adev = drm_to_adev(dev); + struct amdgpu_fpriv *fpriv = file_priv->driver_priv; struct drm_gem_object *gobj; uint32_t handle; u64 flags = AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED | @@ -931,7 +934,7 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv, domain = amdgpu_bo_get_preferred_domain(adev, amdgpu_display_supported_domains(adev, flags)); r = amdgpu_gem_object_create(adev, args->size, 0, domain, flags, - ttm_bo_type_device, NULL, &gobj); + ttm_bo_type_device, NULL, &gobj, fpriv->xcp_id + 1); if (r) return -ENOMEM; @@ -948,7 +951,7 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv, #if defined(CONFIG_DEBUG_FS) static int amdgpu_debugfs_gem_info_show(struct seq_file *m, void *unused) { - struct amdgpu_device *adev = (struct amdgpu_device *)m->private; + struct amdgpu_device *adev = m->private; struct drm_device *dev = adev_to_drm(adev); struct drm_file *file; int r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.h index 637bf51dbf06..f30264782ba2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.h @@ -43,8 +43,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size, int alignment, u32 initial_domain, u64 flags, enum ttm_bo_type type, struct dma_resv *resv, - struct drm_gem_object **obj); - + struct drm_gem_object **obj, int8_t xcp_id_plus1); int amdgpu_mode_dumb_create(struct drm_file *file_priv, struct drm_device *dev, struct drm_mode_create_dumb *args); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c index f3f541ba0aca..a33d4bc34cee 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c @@ -28,6 +28,7 @@ #include "amdgpu_gfx.h" #include "amdgpu_rlc.h" #include "amdgpu_ras.h" +#include "amdgpu_xcp.h" /* delay 0.1 second to enable gfx off feature */ #define GFX_OFF_DELAY_ENABLE msecs_to_jiffies(100) @@ -63,10 +64,10 @@ void amdgpu_queue_mask_bit_to_mec_queue(struct amdgpu_device *adev, int bit, } bool amdgpu_gfx_is_mec_queue_enabled(struct amdgpu_device *adev, - int mec, int pipe, int queue) + int xcc_id, int mec, int pipe, int queue) { return test_bit(amdgpu_gfx_mec_queue_to_bit(adev, mec, pipe, queue), - adev->gfx.mec.queue_bitmap); + adev->gfx.mec_bitmap[xcc_id].queue_bitmap); } int amdgpu_gfx_me_queue_to_bit(struct amdgpu_device *adev, @@ -204,29 +205,38 @@ bool amdgpu_gfx_is_high_priority_compute_queue(struct amdgpu_device *adev, void amdgpu_gfx_compute_queue_acquire(struct amdgpu_device *adev) { - int i, queue, pipe; + int i, j, queue, pipe; bool multipipe_policy = amdgpu_gfx_is_compute_multipipe_capable(adev); int max_queues_per_mec = min(adev->gfx.mec.num_pipe_per_mec * adev->gfx.mec.num_queue_per_pipe, adev->gfx.num_compute_rings); + int num_xcc = adev->gfx.xcc_mask ? NUM_XCC(adev->gfx.xcc_mask) : 1; if (multipipe_policy) { - /* policy: make queues evenly cross all pipes on MEC1 only */ - for (i = 0; i < max_queues_per_mec; i++) { - pipe = i % adev->gfx.mec.num_pipe_per_mec; - queue = (i / adev->gfx.mec.num_pipe_per_mec) % - adev->gfx.mec.num_queue_per_pipe; - - set_bit(pipe * adev->gfx.mec.num_queue_per_pipe + queue, - adev->gfx.mec.queue_bitmap); + /* policy: make queues evenly cross all pipes on MEC1 only + * for multiple xcc, just use the original policy for simplicity */ + for (j = 0; j < num_xcc; j++) { + for (i = 0; i < max_queues_per_mec; i++) { + pipe = i % adev->gfx.mec.num_pipe_per_mec; + queue = (i / adev->gfx.mec.num_pipe_per_mec) % + adev->gfx.mec.num_queue_per_pipe; + + set_bit(pipe * adev->gfx.mec.num_queue_per_pipe + queue, + adev->gfx.mec_bitmap[j].queue_bitmap); + } } } else { /* policy: amdgpu owns all queues in the given pipe */ - for (i = 0; i < max_queues_per_mec; ++i) - set_bit(i, adev->gfx.mec.queue_bitmap); + for (j = 0; j < num_xcc; j++) { + for (i = 0; i < max_queues_per_mec; ++i) + set_bit(i, adev->gfx.mec_bitmap[j].queue_bitmap); + } } - dev_dbg(adev->dev, "mec queue bitmap weight=%d\n", bitmap_weight(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES)); + for (j = 0; j < num_xcc; j++) { + dev_dbg(adev->dev, "mec queue bitmap weight=%d\n", + bitmap_weight(adev->gfx.mec_bitmap[j].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES)); + } } void amdgpu_gfx_graphics_queue_acquire(struct amdgpu_device *adev) @@ -258,7 +268,7 @@ void amdgpu_gfx_graphics_queue_acquire(struct amdgpu_device *adev) } static int amdgpu_gfx_kiq_acquire(struct amdgpu_device *adev, - struct amdgpu_ring *ring) + struct amdgpu_ring *ring, int xcc_id) { int queue_bit; int mec, pipe, queue; @@ -268,7 +278,7 @@ static int amdgpu_gfx_kiq_acquire(struct amdgpu_device *adev, * adev->gfx.mec.num_queue_per_pipe; while (--queue_bit >= 0) { - if (test_bit(queue_bit, adev->gfx.mec.queue_bitmap)) + if (test_bit(queue_bit, adev->gfx.mec_bitmap[xcc_id].queue_bitmap)) continue; amdgpu_queue_mask_bit_to_mec_queue(adev, queue_bit, &mec, &pipe, &queue); @@ -294,9 +304,9 @@ static int amdgpu_gfx_kiq_acquire(struct amdgpu_device *adev, int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev, struct amdgpu_ring *ring, - struct amdgpu_irq_src *irq) + struct amdgpu_irq_src *irq, int xcc_id) { - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id]; int r = 0; spin_lock_init(&kiq->ring_lock); @@ -304,16 +314,20 @@ int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev, ring->adev = NULL; ring->ring_obj = NULL; ring->use_doorbell = true; - ring->doorbell_index = adev->doorbell_index.kiq; - ring->vm_hub = AMDGPU_GFXHUB_0; - - r = amdgpu_gfx_kiq_acquire(adev, ring); + ring->xcc_id = xcc_id; + ring->vm_hub = AMDGPU_GFXHUB(xcc_id); + ring->doorbell_index = + (adev->doorbell_index.kiq + + xcc_id * adev->doorbell_index.xcc_doorbell_range) + << 1; + + r = amdgpu_gfx_kiq_acquire(adev, ring, xcc_id); if (r) return r; ring->eop_gpu_addr = kiq->eop_gpu_addr; ring->no_scheduler = true; - sprintf(ring->name, "kiq_%d.%d.%d", ring->me, ring->pipe, ring->queue); + sprintf(ring->name, "kiq_%d.%d.%d.%d", xcc_id, ring->me, ring->pipe, ring->queue); r = amdgpu_ring_init(adev, ring, 1024, irq, AMDGPU_CP_KIQ_IRQ_DRIVER0, AMDGPU_RING_PRIO_DEFAULT, NULL); if (r) @@ -327,19 +341,19 @@ void amdgpu_gfx_kiq_free_ring(struct amdgpu_ring *ring) amdgpu_ring_fini(ring); } -void amdgpu_gfx_kiq_fini(struct amdgpu_device *adev) +void amdgpu_gfx_kiq_fini(struct amdgpu_device *adev, int xcc_id) { - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id]; amdgpu_bo_free_kernel(&kiq->eop_obj, &kiq->eop_gpu_addr, NULL); } int amdgpu_gfx_kiq_init(struct amdgpu_device *adev, - unsigned hpd_size) + unsigned hpd_size, int xcc_id) { int r; u32 *hpd; - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id]; r = amdgpu_bo_create_kernel(adev, hpd_size, PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT, &kiq->eop_obj, @@ -362,13 +376,18 @@ int amdgpu_gfx_kiq_init(struct amdgpu_device *adev, /* create MQD for each compute/gfx queue */ int amdgpu_gfx_mqd_sw_init(struct amdgpu_device *adev, - unsigned mqd_size) + unsigned mqd_size, int xcc_id) { - struct amdgpu_ring *ring = NULL; - int r, i; + int r, i, j; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id]; + struct amdgpu_ring *ring = &kiq->ring; + u32 domain = AMDGPU_GEM_DOMAIN_GTT; + + /* Only enable on gfx10 and 11 for now to avoid changing behavior on older chips */ + if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 0, 0)) + domain |= AMDGPU_GEM_DOMAIN_VRAM; /* create MQD for KIQ */ - ring = &adev->gfx.kiq.ring; if (!adev->enable_mes_kiq && !ring->mqd_obj) { /* originaly the KIQ MQD is put in GTT domain, but for SRIOV VRAM domain is a must * otherwise hypervisor trigger SAVE_VF fail after driver unloaded which mean MQD @@ -387,8 +406,8 @@ int amdgpu_gfx_mqd_sw_init(struct amdgpu_device *adev, } /* prepare MQD backup */ - adev->gfx.mec.mqd_backup[AMDGPU_MAX_COMPUTE_RINGS] = kmalloc(mqd_size, GFP_KERNEL); - if (!adev->gfx.mec.mqd_backup[AMDGPU_MAX_COMPUTE_RINGS]) + kiq->mqd_backup = kmalloc(mqd_size, GFP_KERNEL); + if (!kiq->mqd_backup) dev_warn(adev->dev, "no memory to create MQD backup for ring %s\n", ring->name); } @@ -398,13 +417,14 @@ int amdgpu_gfx_mqd_sw_init(struct amdgpu_device *adev, ring = &adev->gfx.gfx_ring[i]; if (!ring->mqd_obj) { r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_GTT, &ring->mqd_obj, + domain, &ring->mqd_obj, &ring->mqd_gpu_addr, &ring->mqd_ptr); if (r) { dev_warn(adev->dev, "failed to create ring mqd bo (%d)", r); return r; } + ring->mqd_size = mqd_size; /* prepare MQD backup */ adev->gfx.me.mqd_backup[i] = kmalloc(mqd_size, GFP_KERNEL); if (!adev->gfx.me.mqd_backup[i]) @@ -415,19 +435,21 @@ int amdgpu_gfx_mqd_sw_init(struct amdgpu_device *adev, /* create MQD for each KCQ */ for (i = 0; i < adev->gfx.num_compute_rings; i++) { - ring = &adev->gfx.compute_ring[i]; + j = i + xcc_id * adev->gfx.num_compute_rings; + ring = &adev->gfx.compute_ring[j]; if (!ring->mqd_obj) { r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_GTT, &ring->mqd_obj, + domain, &ring->mqd_obj, &ring->mqd_gpu_addr, &ring->mqd_ptr); if (r) { dev_warn(adev->dev, "failed to create ring mqd bo (%d)", r); return r; } + ring->mqd_size = mqd_size; /* prepare MQD backup */ - adev->gfx.mec.mqd_backup[i] = kmalloc(mqd_size, GFP_KERNEL); - if (!adev->gfx.mec.mqd_backup[i]) + adev->gfx.mec.mqd_backup[j] = kmalloc(mqd_size, GFP_KERNEL); + if (!adev->gfx.mec.mqd_backup[j]) dev_warn(adev->dev, "no memory to create MQD backup for ring %s\n", ring->name); } } @@ -435,10 +457,11 @@ int amdgpu_gfx_mqd_sw_init(struct amdgpu_device *adev, return 0; } -void amdgpu_gfx_mqd_sw_fini(struct amdgpu_device *adev) +void amdgpu_gfx_mqd_sw_fini(struct amdgpu_device *adev, int xcc_id) { struct amdgpu_ring *ring = NULL; - int i; + int i, j; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id]; if (adev->asic_type >= CHIP_NAVI10 && amdgpu_async_gfx_ring) { for (i = 0; i < adev->gfx.num_gfx_rings; i++) { @@ -451,43 +474,81 @@ void amdgpu_gfx_mqd_sw_fini(struct amdgpu_device *adev) } for (i = 0; i < adev->gfx.num_compute_rings; i++) { - ring = &adev->gfx.compute_ring[i]; - kfree(adev->gfx.mec.mqd_backup[i]); + j = i + xcc_id * adev->gfx.num_compute_rings; + ring = &adev->gfx.compute_ring[j]; + kfree(adev->gfx.mec.mqd_backup[j]); amdgpu_bo_free_kernel(&ring->mqd_obj, &ring->mqd_gpu_addr, &ring->mqd_ptr); } - ring = &adev->gfx.kiq.ring; - kfree(adev->gfx.mec.mqd_backup[AMDGPU_MAX_COMPUTE_RINGS]); + ring = &kiq->ring; + kfree(kiq->mqd_backup); amdgpu_bo_free_kernel(&ring->mqd_obj, &ring->mqd_gpu_addr, &ring->mqd_ptr); } -int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev) +int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev, int xcc_id) { - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id]; struct amdgpu_ring *kiq_ring = &kiq->ring; int i, r = 0; + int j; if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues) return -EINVAL; - spin_lock(&adev->gfx.kiq.ring_lock); + spin_lock(&kiq->ring_lock); if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size * adev->gfx.num_compute_rings)) { - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&kiq->ring_lock); return -ENOMEM; } - for (i = 0; i < adev->gfx.num_compute_rings; i++) - kiq->pmf->kiq_unmap_queues(kiq_ring, &adev->gfx.compute_ring[i], + for (i = 0; i < adev->gfx.num_compute_rings; i++) { + j = i + xcc_id * adev->gfx.num_compute_rings; + kiq->pmf->kiq_unmap_queues(kiq_ring, + &adev->gfx.compute_ring[j], RESET_QUEUES, 0, 0); + } - if (adev->gfx.kiq.ring.sched.ready && !adev->job_hang) + if (kiq_ring->sched.ready && !adev->job_hang) r = amdgpu_ring_test_helper(kiq_ring); - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&kiq->ring_lock); + + return r; +} + +int amdgpu_gfx_disable_kgq(struct amdgpu_device *adev, int xcc_id) +{ + struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id]; + struct amdgpu_ring *kiq_ring = &kiq->ring; + int i, r = 0; + int j; + + if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues) + return -EINVAL; + + spin_lock(&kiq->ring_lock); + if (amdgpu_gfx_is_master_xcc(adev, xcc_id)) { + if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size * + adev->gfx.num_gfx_rings)) { + spin_unlock(&kiq->ring_lock); + return -ENOMEM; + } + + for (i = 0; i < adev->gfx.num_gfx_rings; i++) { + j = i + xcc_id * adev->gfx.num_gfx_rings; + kiq->pmf->kiq_unmap_queues(kiq_ring, + &adev->gfx.gfx_ring[j], + PREEMPT_QUEUES, 0, 0); + } + } + + if (adev->gfx.kiq[0].ring.sched.ready && !adev->job_hang) + r = amdgpu_ring_test_helper(kiq_ring); + spin_unlock(&kiq->ring_lock); return r; } @@ -505,18 +566,18 @@ int amdgpu_queue_mask_bit_to_set_resource_bit(struct amdgpu_device *adev, return set_resource_bit; } -int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev) +int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev, int xcc_id) { - struct amdgpu_kiq *kiq = &adev->gfx.kiq; - struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id]; + struct amdgpu_ring *kiq_ring = &kiq->ring; uint64_t queue_mask = 0; - int r, i; + int r, i, j; if (!kiq->pmf || !kiq->pmf->kiq_map_queues || !kiq->pmf->kiq_set_resources) return -EINVAL; for (i = 0; i < AMDGPU_MAX_COMPUTE_QUEUES; ++i) { - if (!test_bit(i, adev->gfx.mec.queue_bitmap)) + if (!test_bit(i, adev->gfx.mec_bitmap[xcc_id].queue_bitmap)) continue; /* This situation may be hit in the future if a new HW @@ -532,13 +593,15 @@ int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev) DRM_INFO("kiq ring mec %d pipe %d q %d\n", kiq_ring->me, kiq_ring->pipe, kiq_ring->queue); - spin_lock(&adev->gfx.kiq.ring_lock); + amdgpu_device_flush_hdp(adev, NULL); + + spin_lock(&kiq->ring_lock); r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size * adev->gfx.num_compute_rings + kiq->pmf->set_resources_size); if (r) { DRM_ERROR("Failed to lock KIQ (%d).\n", r); - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&kiq->ring_lock); return r; } @@ -546,11 +609,51 @@ int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev) queue_mask = ~0ULL; kiq->pmf->kiq_set_resources(kiq_ring, queue_mask); - for (i = 0; i < adev->gfx.num_compute_rings; i++) - kiq->pmf->kiq_map_queues(kiq_ring, &adev->gfx.compute_ring[i]); + for (i = 0; i < adev->gfx.num_compute_rings; i++) { + j = i + xcc_id * adev->gfx.num_compute_rings; + kiq->pmf->kiq_map_queues(kiq_ring, + &adev->gfx.compute_ring[j]); + } r = amdgpu_ring_test_helper(kiq_ring); - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&kiq->ring_lock); + if (r) + DRM_ERROR("KCQ enable failed\n"); + + return r; +} + +int amdgpu_gfx_enable_kgq(struct amdgpu_device *adev, int xcc_id) +{ + struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id]; + struct amdgpu_ring *kiq_ring = &kiq->ring; + int r, i, j; + + if (!kiq->pmf || !kiq->pmf->kiq_map_queues) + return -EINVAL; + + amdgpu_device_flush_hdp(adev, NULL); + + spin_lock(&kiq->ring_lock); + /* No need to map kcq on the slave */ + if (amdgpu_gfx_is_master_xcc(adev, xcc_id)) { + r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size * + adev->gfx.num_gfx_rings); + if (r) { + DRM_ERROR("Failed to lock KIQ (%d).\n", r); + spin_unlock(&kiq->ring_lock); + return r; + } + + for (i = 0; i < adev->gfx.num_gfx_rings; i++) { + j = i + xcc_id * adev->gfx.num_gfx_rings; + kiq->pmf->kiq_map_queues(kiq_ring, + &adev->gfx.gfx_ring[j]); + } + } + + r = amdgpu_ring_test_helper(kiq_ring); + spin_unlock(&kiq->ring_lock); if (r) DRM_ERROR("KCQ enable failed\n"); @@ -785,12 +888,31 @@ int amdgpu_gfx_cp_ecc_error_irq(struct amdgpu_device *adev, return 0; } +void amdgpu_gfx_ras_error_func(struct amdgpu_device *adev, + void *ras_error_status, + void (*func)(struct amdgpu_device *adev, void *ras_error_status, + int xcc_id)) +{ + int i; + int num_xcc = adev->gfx.xcc_mask ? NUM_XCC(adev->gfx.xcc_mask) : 1; + uint32_t xcc_mask = GENMASK(num_xcc - 1, 0); + struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status; + + if (err_data) { + err_data->ue_count = 0; + err_data->ce_count = 0; + } + + for_each_inst(i, xcc_mask) + func(adev, ras_error_status, i); +} + uint32_t amdgpu_kiq_rreg(struct amdgpu_device *adev, uint32_t reg) { signed long r, cnt = 0; unsigned long flags; uint32_t seq, reg_val_offs = 0, value = 0; - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; struct amdgpu_ring *ring = &kiq->ring; if (amdgpu_device_skip_hw_access(adev)) @@ -858,7 +980,7 @@ void amdgpu_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v) signed long r, cnt = 0; unsigned long flags; uint32_t seq; - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; struct amdgpu_ring *ring = &kiq->ring; BUG_ON(!ring->funcs->emit_wreg); @@ -1062,3 +1184,125 @@ void amdgpu_gfx_cp_init_microcode(struct amdgpu_device *adev, adev->firmware.fw_size += ALIGN(fw_size, PAGE_SIZE); } } + +bool amdgpu_gfx_is_master_xcc(struct amdgpu_device *adev, int xcc_id) +{ + return !(xcc_id % (adev->gfx.num_xcc_per_xcp ? + adev->gfx.num_xcc_per_xcp : 1)); +} + +static ssize_t amdgpu_gfx_get_current_compute_partition(struct device *dev, + struct device_attribute *addr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = drm_to_adev(ddev); + int mode; + + mode = amdgpu_xcp_query_partition_mode(adev->xcp_mgr, + AMDGPU_XCP_FL_NONE); + + return sysfs_emit(buf, "%s\n", amdgpu_gfx_compute_mode_desc(mode)); +} + +static ssize_t amdgpu_gfx_set_compute_partition(struct device *dev, + struct device_attribute *addr, + const char *buf, size_t count) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = drm_to_adev(ddev); + enum amdgpu_gfx_partition mode; + int ret = 0, num_xcc; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + if (num_xcc % 2 != 0) + return -EINVAL; + + if (!strncasecmp("SPX", buf, strlen("SPX"))) { + mode = AMDGPU_SPX_PARTITION_MODE; + } else if (!strncasecmp("DPX", buf, strlen("DPX"))) { + /* + * DPX mode needs AIDs to be in multiple of 2. + * Each AID connects 2 XCCs. + */ + if (num_xcc%4) + return -EINVAL; + mode = AMDGPU_DPX_PARTITION_MODE; + } else if (!strncasecmp("TPX", buf, strlen("TPX"))) { + if (num_xcc != 6) + return -EINVAL; + mode = AMDGPU_TPX_PARTITION_MODE; + } else if (!strncasecmp("QPX", buf, strlen("QPX"))) { + if (num_xcc != 8) + return -EINVAL; + mode = AMDGPU_QPX_PARTITION_MODE; + } else if (!strncasecmp("CPX", buf, strlen("CPX"))) { + mode = AMDGPU_CPX_PARTITION_MODE; + } else { + return -EINVAL; + } + + ret = amdgpu_xcp_switch_partition_mode(adev->xcp_mgr, mode); + + if (ret) + return ret; + + return count; +} + +static ssize_t amdgpu_gfx_get_available_compute_partition(struct device *dev, + struct device_attribute *addr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = drm_to_adev(ddev); + char *supported_partition; + + /* TBD */ + switch (NUM_XCC(adev->gfx.xcc_mask)) { + case 8: + supported_partition = "SPX, DPX, QPX, CPX"; + break; + case 6: + supported_partition = "SPX, TPX, CPX"; + break; + case 4: + supported_partition = "SPX, DPX, CPX"; + break; + /* this seems only existing in emulation phase */ + case 2: + supported_partition = "SPX, CPX"; + break; + default: + supported_partition = "Not supported"; + break; + } + + return sysfs_emit(buf, "%s\n", supported_partition); +} + +static DEVICE_ATTR(current_compute_partition, S_IRUGO | S_IWUSR, + amdgpu_gfx_get_current_compute_partition, + amdgpu_gfx_set_compute_partition); + +static DEVICE_ATTR(available_compute_partition, S_IRUGO, + amdgpu_gfx_get_available_compute_partition, NULL); + +int amdgpu_gfx_sysfs_init(struct amdgpu_device *adev) +{ + int r; + + r = device_create_file(adev->dev, &dev_attr_current_compute_partition); + if (r) + return r; + + r = device_create_file(adev->dev, &dev_attr_available_compute_partition); + + return r; +} + +void amdgpu_gfx_sysfs_fini(struct amdgpu_device *adev) +{ + device_remove_file(adev->dev, &dev_attr_current_compute_partition); + device_remove_file(adev->dev, &dev_attr_available_compute_partition); +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h index bfabea76d166..ce0f7a8ad4b8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h @@ -61,7 +61,42 @@ enum amdgpu_gfx_partition { AMDGPU_TPX_PARTITION_MODE = 2, AMDGPU_QPX_PARTITION_MODE = 3, AMDGPU_CPX_PARTITION_MODE = 4, - AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE, + AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE = -1, + /* Automatically choose the right mode */ + AMDGPU_AUTO_COMPUTE_PARTITION_MODE = -2, +}; + +#define NUM_XCC(x) hweight16(x) + +enum amdgpu_pkg_type { + AMDGPU_PKG_TYPE_APU = 2, + AMDGPU_PKG_TYPE_UNKNOWN, +}; + +enum amdgpu_gfx_ras_mem_id_type { + AMDGPU_GFX_CP_MEM = 0, + AMDGPU_GFX_GCEA_MEM, + AMDGPU_GFX_GC_CANE_MEM, + AMDGPU_GFX_GCUTCL2_MEM, + AMDGPU_GFX_GDS_MEM, + AMDGPU_GFX_LDS_MEM, + AMDGPU_GFX_RLC_MEM, + AMDGPU_GFX_SP_MEM, + AMDGPU_GFX_SPI_MEM, + AMDGPU_GFX_SQC_MEM, + AMDGPU_GFX_SQ_MEM, + AMDGPU_GFX_TA_MEM, + AMDGPU_GFX_TCC_MEM, + AMDGPU_GFX_TCA_MEM, + AMDGPU_GFX_TCI_MEM, + AMDGPU_GFX_TCP_MEM, + AMDGPU_GFX_TD_MEM, + AMDGPU_GFX_TCX_MEM, + AMDGPU_GFX_ATC_L2_MEM, + AMDGPU_GFX_UTCL2_MEM, + AMDGPU_GFX_VML2_MEM, + AMDGPU_GFX_VML2_WALKER_MEM, + AMDGPU_GFX_MEM_TYPE_NUM }; struct amdgpu_mec { @@ -75,8 +110,10 @@ struct amdgpu_mec { u32 num_mec; u32 num_pipe_per_mec; u32 num_queue_per_pipe; - void *mqd_backup[AMDGPU_MAX_COMPUTE_RINGS + 1]; + void *mqd_backup[AMDGPU_MAX_COMPUTE_RINGS * AMDGPU_MAX_GC_INSTANCES]; +}; +struct amdgpu_mec_bitmap { /* These are the resources for which amdgpu takes ownership */ DECLARE_BITMAP(queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES); }; @@ -120,6 +157,7 @@ struct amdgpu_kiq { struct amdgpu_ring ring; struct amdgpu_irq_src irq; const struct kiq_pm4_funcs *pmf; + void *mqd_backup; }; /* @@ -230,23 +268,37 @@ struct amdgpu_gfx_ras { struct amdgpu_iv_entry *entry); }; +struct amdgpu_gfx_shadow_info { + u32 shadow_size; + u32 shadow_alignment; + u32 csa_size; + u32 csa_alignment; +}; + struct amdgpu_gfx_funcs { /* get the gpu clock counter */ uint64_t (*get_gpu_clock_counter)(struct amdgpu_device *adev); void (*select_se_sh)(struct amdgpu_device *adev, u32 se_num, - u32 sh_num, u32 instance); - void (*read_wave_data)(struct amdgpu_device *adev, uint32_t simd, + u32 sh_num, u32 instance, int xcc_id); + void (*read_wave_data)(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields); - void (*read_wave_vgprs)(struct amdgpu_device *adev, uint32_t simd, + void (*read_wave_vgprs)(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t thread, uint32_t start, uint32_t size, uint32_t *dst); - void (*read_wave_sgprs)(struct amdgpu_device *adev, uint32_t simd, + void (*read_wave_sgprs)(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t start, uint32_t size, uint32_t *dst); void (*select_me_pipe_q)(struct amdgpu_device *adev, u32 me, u32 pipe, - u32 queue, u32 vmid); + u32 queue, u32 vmid, u32 xcc_id); void (*init_spm_golden)(struct amdgpu_device *adev); void (*update_perfmon_mgcg)(struct amdgpu_device *adev, bool enable); + int (*get_gfx_shadow_info)(struct amdgpu_device *adev, + struct amdgpu_gfx_shadow_info *shadow_info); + enum amdgpu_gfx_partition + (*query_partition_mode)(struct amdgpu_device *adev); + int (*switch_partition_mode)(struct amdgpu_device *adev, + int num_xccs_per_xcp); + int (*ih_node_to_logical_xcc)(struct amdgpu_device *adev, int ih_node); }; struct sq_work { @@ -296,7 +348,8 @@ struct amdgpu_gfx { struct amdgpu_ce ce; struct amdgpu_me me; struct amdgpu_mec mec; - struct amdgpu_kiq kiq; + struct amdgpu_mec_bitmap mec_bitmap[AMDGPU_MAX_GC_INSTANCES]; + struct amdgpu_kiq kiq[AMDGPU_MAX_GC_INSTANCES]; struct amdgpu_imu imu; bool rs64_enable; /* firmware format */ const struct firmware *me_fw; /* ME firmware */ @@ -376,15 +429,31 @@ struct amdgpu_gfx { struct amdgpu_ring sw_gfx_ring[AMDGPU_MAX_SW_GFX_RINGS]; struct amdgpu_ring_mux muxer; - enum amdgpu_gfx_partition partition_mode; - uint32_t num_xcd; + bool cp_gfx_shadow; /* for gfx11 */ + + uint16_t xcc_mask; uint32_t num_xcc_per_xcp; + struct mutex partition_mutex; }; +struct amdgpu_gfx_ras_reg_entry { + struct amdgpu_ras_err_status_reg_entry reg_entry; + enum amdgpu_gfx_ras_mem_id_type mem_id_type; + uint32_t se_num; +}; + +struct amdgpu_gfx_ras_mem_id_entry { + const struct amdgpu_ras_memory_id_entry *mem_id_ent; + uint32_t size; +}; + +#define AMDGPU_GFX_MEMID_ENT(x) {(x), ARRAY_SIZE(x)}, + #define amdgpu_gfx_get_gpu_clock_counter(adev) (adev)->gfx.funcs->get_gpu_clock_counter((adev)) -#define amdgpu_gfx_select_se_sh(adev, se, sh, instance) (adev)->gfx.funcs->select_se_sh((adev), (se), (sh), (instance)) -#define amdgpu_gfx_select_me_pipe_q(adev, me, pipe, q, vmid) (adev)->gfx.funcs->select_me_pipe_q((adev), (me), (pipe), (q), (vmid)) +#define amdgpu_gfx_select_se_sh(adev, se, sh, instance, xcc_id) ((adev)->gfx.funcs->select_se_sh((adev), (se), (sh), (instance), (xcc_id))) +#define amdgpu_gfx_select_me_pipe_q(adev, me, pipe, q, vmid, xcc_id) ((adev)->gfx.funcs->select_me_pipe_q((adev), (me), (pipe), (q), (vmid), (xcc_id))) #define amdgpu_gfx_init_spm_golden(adev) (adev)->gfx.funcs->init_spm_golden((adev)) +#define amdgpu_gfx_get_gfx_shadow_info(adev, si) ((adev)->gfx.funcs->get_gfx_shadow_info((adev), (si))) /** * amdgpu_gfx_create_bitmask - create a bitmask @@ -404,19 +473,21 @@ void amdgpu_gfx_parse_disable_cu(unsigned *mask, unsigned max_se, int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev, struct amdgpu_ring *ring, - struct amdgpu_irq_src *irq); + struct amdgpu_irq_src *irq, int xcc_id); void amdgpu_gfx_kiq_free_ring(struct amdgpu_ring *ring); -void amdgpu_gfx_kiq_fini(struct amdgpu_device *adev); +void amdgpu_gfx_kiq_fini(struct amdgpu_device *adev, int xcc_id); int amdgpu_gfx_kiq_init(struct amdgpu_device *adev, - unsigned hpd_size); + unsigned hpd_size, int xcc_id); int amdgpu_gfx_mqd_sw_init(struct amdgpu_device *adev, - unsigned mqd_size); -void amdgpu_gfx_mqd_sw_fini(struct amdgpu_device *adev); -int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev); -int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev); + unsigned mqd_size, int xcc_id); +void amdgpu_gfx_mqd_sw_fini(struct amdgpu_device *adev, int xcc_id); +int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev, int xcc_id); +int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev, int xcc_id); +int amdgpu_gfx_disable_kgq(struct amdgpu_device *adev, int xcc_id); +int amdgpu_gfx_enable_kgq(struct amdgpu_device *adev, int xcc_id); void amdgpu_gfx_compute_queue_acquire(struct amdgpu_device *adev); void amdgpu_gfx_graphics_queue_acquire(struct amdgpu_device *adev); @@ -425,8 +496,8 @@ int amdgpu_gfx_mec_queue_to_bit(struct amdgpu_device *adev, int mec, int pipe, int queue); void amdgpu_queue_mask_bit_to_mec_queue(struct amdgpu_device *adev, int bit, int *mec, int *pipe, int *queue); -bool amdgpu_gfx_is_mec_queue_enabled(struct amdgpu_device *adev, int mec, - int pipe, int queue); +bool amdgpu_gfx_is_mec_queue_enabled(struct amdgpu_device *adev, int xcc_id, + int mec, int pipe, int queue); bool amdgpu_gfx_is_high_priority_compute_queue(struct amdgpu_device *adev, struct amdgpu_ring *ring); bool amdgpu_gfx_is_high_priority_graphics_queue(struct amdgpu_device *adev, @@ -458,4 +529,33 @@ void amdgpu_gfx_cp_init_microcode(struct amdgpu_device *adev, uint32_t ucode_id) int amdgpu_gfx_ras_sw_init(struct amdgpu_device *adev); int amdgpu_gfx_poison_consumption_handler(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry); + +bool amdgpu_gfx_is_master_xcc(struct amdgpu_device *adev, int xcc_id); +int amdgpu_gfx_sysfs_init(struct amdgpu_device *adev); +void amdgpu_gfx_sysfs_fini(struct amdgpu_device *adev); +void amdgpu_gfx_ras_error_func(struct amdgpu_device *adev, + void *ras_error_status, + void (*func)(struct amdgpu_device *adev, void *ras_error_status, + int xcc_id)); + +static inline const char *amdgpu_gfx_compute_mode_desc(int mode) +{ + switch (mode) { + case AMDGPU_SPX_PARTITION_MODE: + return "SPX"; + case AMDGPU_DPX_PARTITION_MODE: + return "DPX"; + case AMDGPU_TPX_PARTITION_MODE: + return "TPX"; + case AMDGPU_QPX_PARTITION_MODE: + return "QPX"; + case AMDGPU_CPX_PARTITION_MODE: + return "CPX"; + default: + return "UNKNOWN"; + } + + return "UNKNOWN"; +} + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c index 4e2531758866..d78bd9732543 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c @@ -534,22 +534,21 @@ void amdgpu_gmc_ras_fini(struct amdgpu_device *adev) * subject to change when ring number changes * Engine 17: Gart flushes */ -#define GFXHUB_FREE_VM_INV_ENGS_BITMAP 0x1FFF3 -#define MMHUB_FREE_VM_INV_ENGS_BITMAP 0x1FFF3 +#define AMDGPU_VMHUB_INV_ENG_BITMAP 0x1FFF3 int amdgpu_gmc_allocate_vm_inv_eng(struct amdgpu_device *adev) { struct amdgpu_ring *ring; - unsigned vm_inv_engs[AMDGPU_MAX_VMHUBS] = - {GFXHUB_FREE_VM_INV_ENGS_BITMAP, MMHUB_FREE_VM_INV_ENGS_BITMAP, - GFXHUB_FREE_VM_INV_ENGS_BITMAP}; + unsigned vm_inv_engs[AMDGPU_MAX_VMHUBS] = {0}; unsigned i; unsigned vmhub, inv_eng; - if (adev->enable_mes) { + /* init the vm inv eng for all vmhubs */ + for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) { + vm_inv_engs[i] = AMDGPU_VMHUB_INV_ENG_BITMAP; /* reserve engine 5 for firmware */ - for (vmhub = 0; vmhub < AMDGPU_MAX_VMHUBS; vmhub++) - vm_inv_engs[vmhub] &= ~(1 << 5); + if (adev->enable_mes) + vm_inv_engs[i] &= ~(1 << 5); } for (i = 0; i < adev->num_rings; ++i) { @@ -593,6 +592,8 @@ void amdgpu_gmc_tmz_set(struct amdgpu_device *adev) case IP_VERSION(9, 3, 0): /* GC 10.3.7 */ case IP_VERSION(10, 3, 7): + /* GC 11.0.1 */ + case IP_VERSION(11, 0, 1): if (amdgpu_tmz == 0) { adev->gmc.tmz_enabled = false; dev_info(adev->dev, @@ -616,7 +617,6 @@ void amdgpu_gmc_tmz_set(struct amdgpu_device *adev) case IP_VERSION(10, 3, 1): /* YELLOW_CARP*/ case IP_VERSION(10, 3, 3): - case IP_VERSION(11, 0, 1): case IP_VERSION(11, 0, 4): /* Don't enable it by default yet. */ @@ -670,7 +670,7 @@ void amdgpu_gmc_set_vm_fault_masks(struct amdgpu_device *adev, int hub_type, for (i = 0; i < 16; i++) { reg = hub->vm_context0_cntl + hub->ctx_distance * i; - tmp = (hub_type == AMDGPU_GFXHUB_0) ? + tmp = (hub_type == AMDGPU_GFXHUB(0)) ? RREG32_SOC15_IP(GC, reg) : RREG32_SOC15_IP(MMHUB, reg); @@ -679,7 +679,7 @@ void amdgpu_gmc_set_vm_fault_masks(struct amdgpu_device *adev, int hub_type, else tmp &= ~hub->vm_cntx_cntl_vm_fault; - (hub_type == AMDGPU_GFXHUB_0) ? + (hub_type == AMDGPU_GFXHUB(0)) ? WREG32_SOC15_IP(GC, reg, tmp) : WREG32_SOC15_IP(MMHUB, reg, tmp); } @@ -892,3 +892,47 @@ int amdgpu_gmc_vram_checking(struct amdgpu_device *adev) return 0; } + +static ssize_t current_memory_partition_show( + struct device *dev, struct device_attribute *addr, char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = drm_to_adev(ddev); + enum amdgpu_memory_partition mode; + + mode = adev->gmc.gmc_funcs->query_mem_partition_mode(adev); + switch (mode) { + case AMDGPU_NPS1_PARTITION_MODE: + return sysfs_emit(buf, "NPS1\n"); + case AMDGPU_NPS2_PARTITION_MODE: + return sysfs_emit(buf, "NPS2\n"); + case AMDGPU_NPS3_PARTITION_MODE: + return sysfs_emit(buf, "NPS3\n"); + case AMDGPU_NPS4_PARTITION_MODE: + return sysfs_emit(buf, "NPS4\n"); + case AMDGPU_NPS6_PARTITION_MODE: + return sysfs_emit(buf, "NPS6\n"); + case AMDGPU_NPS8_PARTITION_MODE: + return sysfs_emit(buf, "NPS8\n"); + default: + return sysfs_emit(buf, "UNKNOWN\n"); + } + + return sysfs_emit(buf, "UNKNOWN\n"); +} + +static DEVICE_ATTR_RO(current_memory_partition); + +int amdgpu_gmc_sysfs_init(struct amdgpu_device *adev) +{ + if (!adev->gmc.gmc_funcs->query_mem_partition_mode) + return 0; + + return device_create_file(adev->dev, + &dev_attr_current_memory_partition); +} + +void amdgpu_gmc_sysfs_fini(struct amdgpu_device *adev) +{ + device_remove_file(adev->dev, &dev_attr_current_memory_partition); +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h index 6d105d7fb98b..56d73fade568 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h @@ -63,6 +63,16 @@ struct firmware; +enum amdgpu_memory_partition { + UNKNOWN_MEMORY_PARTITION_MODE = 0, + AMDGPU_NPS1_PARTITION_MODE = 1, + AMDGPU_NPS2_PARTITION_MODE = 2, + AMDGPU_NPS3_PARTITION_MODE = 3, + AMDGPU_NPS4_PARTITION_MODE = 4, + AMDGPU_NPS6_PARTITION_MODE = 6, + AMDGPU_NPS8_PARTITION_MODE = 8, +}; + /* * GMC page fault information */ @@ -119,7 +129,8 @@ struct amdgpu_gmc_funcs { uint32_t vmhub, uint32_t flush_type); /* flush the vm tlb via pasid */ int (*flush_gpu_tlb_pasid)(struct amdgpu_device *adev, uint16_t pasid, - uint32_t flush_type, bool all_hub); + uint32_t flush_type, bool all_hub, + uint32_t inst); /* flush the vm tlb via ring */ uint64_t (*emit_flush_gpu_tlb)(struct amdgpu_ring *ring, unsigned vmid, uint64_t pd_addr); @@ -137,8 +148,15 @@ struct amdgpu_gmc_funcs { void (*get_vm_pte)(struct amdgpu_device *adev, struct amdgpu_bo_va_mapping *mapping, uint64_t *flags); + /* override per-page pte flags */ + void (*override_vm_pte_flags)(struct amdgpu_device *dev, + struct amdgpu_vm *vm, + uint64_t addr, uint64_t *flags); /* get the amount of memory used by the vbios for pre-OS console */ unsigned int (*get_vbios_fb_size)(struct amdgpu_device *adev); + + enum amdgpu_memory_partition (*query_mem_partition_mode)( + struct amdgpu_device *adev); }; struct amdgpu_xgmi_ras { @@ -164,6 +182,21 @@ struct amdgpu_xgmi { struct amdgpu_xgmi_ras *ras; }; +struct amdgpu_mem_partition_info { + union { + struct { + uint32_t fpfn; + uint32_t lpfn; + } range; + struct { + int node; + } numa; + }; + uint64_t size; +}; + +#define INVALID_PFN -1 + struct amdgpu_gmc { /* FB's physical address in MMIO space (for CPU to * map FB). This is different compared to the agp/ @@ -250,7 +283,10 @@ struct amdgpu_gmc { uint64_t last_fault:AMDGPU_GMC_FAULT_RING_ORDER; bool tmz_enabled; + bool is_app_apu; + struct amdgpu_mem_partition_info *mem_partitions; + uint8_t num_mem_partitions; const struct amdgpu_gmc_funcs *gmc_funcs; struct amdgpu_xgmi xgmi; @@ -265,6 +301,8 @@ struct amdgpu_gmc { /* MALL size */ u64 mall_size; + uint32_t m_half_use; + /* number of UMC instances */ int num_umc; /* mode2 save restore */ @@ -296,14 +334,17 @@ struct amdgpu_gmc { }; #define amdgpu_gmc_flush_gpu_tlb(adev, vmid, vmhub, type) ((adev)->gmc.gmc_funcs->flush_gpu_tlb((adev), (vmid), (vmhub), (type))) -#define amdgpu_gmc_flush_gpu_tlb_pasid(adev, pasid, type, allhub) \ +#define amdgpu_gmc_flush_gpu_tlb_pasid(adev, pasid, type, allhub, inst) \ ((adev)->gmc.gmc_funcs->flush_gpu_tlb_pasid \ - ((adev), (pasid), (type), (allhub))) + ((adev), (pasid), (type), (allhub), (inst))) #define amdgpu_gmc_emit_flush_gpu_tlb(r, vmid, addr) (r)->adev->gmc.gmc_funcs->emit_flush_gpu_tlb((r), (vmid), (addr)) #define amdgpu_gmc_emit_pasid_mapping(r, vmid, pasid) (r)->adev->gmc.gmc_funcs->emit_pasid_mapping((r), (vmid), (pasid)) #define amdgpu_gmc_map_mtype(adev, flags) (adev)->gmc.gmc_funcs->map_mtype((adev),(flags)) #define amdgpu_gmc_get_vm_pde(adev, level, dst, flags) (adev)->gmc.gmc_funcs->get_vm_pde((adev), (level), (dst), (flags)) #define amdgpu_gmc_get_vm_pte(adev, mapping, flags) (adev)->gmc.gmc_funcs->get_vm_pte((adev), (mapping), (flags)) +#define amdgpu_gmc_override_vm_pte_flags(adev, vm, addr, pte_flags) \ + (adev)->gmc.gmc_funcs->override_vm_pte_flags \ + ((adev), (vm), (addr), (pte_flags)) #define amdgpu_gmc_get_vbios_fb_size(adev) (adev)->gmc.gmc_funcs->get_vbios_fb_size((adev)) /** @@ -373,4 +414,7 @@ uint64_t amdgpu_gmc_vram_mc2pa(struct amdgpu_device *adev, uint64_t mc_addr); uint64_t amdgpu_gmc_vram_pa(struct amdgpu_device *adev, struct amdgpu_bo *bo); uint64_t amdgpu_gmc_vram_cpu_pa(struct amdgpu_device *adev, struct amdgpu_bo *bo); int amdgpu_gmc_vram_checking(struct amdgpu_device *adev); +int amdgpu_gmc_sysfs_init(struct amdgpu_device *adev); +void amdgpu_gmc_sysfs_fini(struct amdgpu_device *adev); + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c index 2dadcfe43d03..081267161d40 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c @@ -190,8 +190,8 @@ int amdgpu_hmm_range_get_pages(struct mmu_interval_notifier *notifier, pr_debug("hmm range: start = 0x%lx, end = 0x%lx", hmm_range->start, hmm_range->end); - /* Assuming 512MB takes maxmium 1 second to fault page address */ - timeout = max((hmm_range->end - hmm_range->start) >> 29, 1UL); + /* Assuming 128MB takes maximum 1 second to fault page address */ + timeout = max((hmm_range->end - hmm_range->start) >> 27, 1UL); timeout *= HMM_RANGE_DEFAULT_TIMEOUT; timeout = jiffies + msecs_to_jiffies(timeout); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c index 4ff348e10e4d..ebeddc9a37e9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c @@ -136,7 +136,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, uint64_t fence_ctx; uint32_t status = 0, alloc_size; unsigned fence_flags = 0; - bool secure; + bool secure, init_shadow; + u64 shadow_va, csa_va, gds_va; + int vmid = AMDGPU_JOB_GET_VMID(job); unsigned i; int r = 0; @@ -150,9 +152,17 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, vm = job->vm; fence_ctx = job->base.s_fence ? job->base.s_fence->scheduled.context : 0; + shadow_va = job->shadow_va; + csa_va = job->csa_va; + gds_va = job->gds_va; + init_shadow = job->init_shadow; } else { vm = NULL; fence_ctx = 0; + shadow_va = 0; + csa_va = 0; + gds_va = 0; + init_shadow = false; } if (!ring->sched.ready && !ring->is_mes_queue) { @@ -212,7 +222,12 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, } amdgpu_ring_ib_begin(ring); - if (job && ring->funcs->init_cond_exec) + + if (ring->funcs->emit_gfx_shadow) + amdgpu_ring_emit_gfx_shadow(ring, shadow_va, csa_va, gds_va, + init_shadow, vmid); + + if (ring->funcs->init_cond_exec) patch_offset = amdgpu_ring_init_cond_exec(ring); amdgpu_device_flush_hdp(adev, ring); @@ -263,6 +278,18 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, fence_flags | AMDGPU_FENCE_FLAG_64BIT); } + if (ring->funcs->emit_gfx_shadow) { + amdgpu_ring_emit_gfx_shadow(ring, 0, 0, 0, false, 0); + + if (ring->funcs->init_cond_exec) { + unsigned ce_offset = ~0; + + ce_offset = amdgpu_ring_init_cond_exec(ring); + if (ce_offset != ~0 && ring->funcs->patch_cond_exec) + amdgpu_ring_patch_cond_exec(ring, ce_offset); + } + } + r = amdgpu_fence_emit(ring, f, job, fence_flags); if (r) { dev_err(adev->dev, "failed to emit fence (%d)\n", r); @@ -436,7 +463,7 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev) static int amdgpu_debugfs_sa_info_show(struct seq_file *m, void *unused) { - struct amdgpu_device *adev = (struct amdgpu_device *)m->private; + struct amdgpu_device *adev = m->private; seq_printf(m, "--------------------- DELAYED --------------------- \n"); amdgpu_sa_bo_dump_debug_info(&adev->ib_pools[AMDGPU_IB_POOL_DELAYED], diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c index c991ca0b7a1c..ff1ea99292fb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c @@ -409,7 +409,7 @@ int amdgpu_vmid_grab(struct amdgpu_vm *vm, struct amdgpu_ring *ring, if (r || !idle) goto error; - if (vm->reserved_vmid[vmhub]) { + if (vm->reserved_vmid[vmhub] || (enforce_isolation && (vmhub == AMDGPU_GFXHUB(0)))) { r = amdgpu_vmid_grab_reserved(vm, ring, job, &id, fence); if (r || !id) goto error; @@ -460,14 +460,11 @@ error: } int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev, - struct amdgpu_vm *vm, unsigned vmhub) { struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub]; mutex_lock(&id_mgr->lock); - if (vm->reserved_vmid[vmhub]) - goto unlock; ++id_mgr->reserved_use_count; if (!id_mgr->reserved) { @@ -479,27 +476,23 @@ int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev, list_del_init(&id->list); id_mgr->reserved = id; } - vm->reserved_vmid[vmhub] = true; -unlock: mutex_unlock(&id_mgr->lock); return 0; } void amdgpu_vmid_free_reserved(struct amdgpu_device *adev, - struct amdgpu_vm *vm, unsigned vmhub) { struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub]; mutex_lock(&id_mgr->lock); - if (vm->reserved_vmid[vmhub] && - !--id_mgr->reserved_use_count) { + if (!--id_mgr->reserved_use_count) { /* give the reserved ID back to normal round robin */ list_add(&id_mgr->reserved->list, &id_mgr->ids_lru); id_mgr->reserved = NULL; } - vm->reserved_vmid[vmhub] = false; + mutex_unlock(&id_mgr->lock); } @@ -578,6 +571,10 @@ void amdgpu_vmid_mgr_init(struct amdgpu_device *adev) list_add_tail(&id_mgr->ids[j].list, &id_mgr->ids_lru); } } + /* alloc a default reserved vmid to enforce isolation */ + if (enforce_isolation) + amdgpu_vmid_alloc_reserved(adev, AMDGPU_GFXHUB(0)); + } /** diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h index d1cc09b45da4..fa8c42c83d5d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h @@ -79,11 +79,9 @@ void amdgpu_pasid_free_delayed(struct dma_resv *resv, bool amdgpu_vmid_had_gpu_reset(struct amdgpu_device *adev, struct amdgpu_vmid *id); int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev, - struct amdgpu_vm *vm, - unsigned vmhub); + unsigned vmhub); void amdgpu_vmid_free_reserved(struct amdgpu_device *adev, - struct amdgpu_vm *vm, - unsigned vmhub); + unsigned vmhub); int amdgpu_vmid_grab(struct amdgpu_vm *vm, struct amdgpu_ring *ring, struct amdgpu_job *job, struct dma_fence **fence); void amdgpu_vmid_reset(struct amdgpu_device *adev, unsigned vmhub, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c index 1d5af50331e4..fceb3b384955 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c @@ -270,7 +270,7 @@ void amdgpu_ih_decode_iv_helper(struct amdgpu_device *adev, entry->timestamp = dw[1] | ((u64)(dw[2] & 0xffff) << 32); entry->timestamp_src = dw[2] >> 31; entry->pasid = dw[3] & 0xffff; - entry->pasid_src = dw[3] >> 31; + entry->node_id = (dw[3] >> 16) & 0xff; entry->src_data[0] = dw[4]; entry->src_data[1] = dw[5]; entry->src_data[2] = dw[6]; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c index fafebec5b7b6..5273decc5753 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c @@ -99,6 +99,21 @@ const char *soc15_ih_clientid_name[] = { "MP1" }; +const int node_id_to_phys_map[NODEID_MAX] = { + [AID0_NODEID] = 0, + [XCD0_NODEID] = 0, + [XCD1_NODEID] = 1, + [AID1_NODEID] = 1, + [XCD2_NODEID] = 2, + [XCD3_NODEID] = 3, + [AID2_NODEID] = 2, + [XCD4_NODEID] = 4, + [XCD5_NODEID] = 5, + [AID3_NODEID] = 3, + [XCD6_NODEID] = 6, + [XCD7_NODEID] = 7, +}; + /** * amdgpu_irq_disable_all - disable *all* interrupts * @@ -109,7 +124,7 @@ const char *soc15_ih_clientid_name[] = { void amdgpu_irq_disable_all(struct amdgpu_device *adev) { unsigned long irqflags; - unsigned i, j, k; + unsigned int i, j, k; int r; spin_lock_irqsave(&adev->irq.lock, irqflags); @@ -124,7 +139,6 @@ void amdgpu_irq_disable_all(struct amdgpu_device *adev) continue; for (k = 0; k < src->num_types; ++k) { - atomic_set(&src->enabled_types[k], 0); r = src->funcs->set(adev, src, k, AMDGPU_IRQ_STATE_DISABLE); if (r) @@ -268,11 +282,11 @@ int amdgpu_irq_init(struct amdgpu_device *adev) int nvec = pci_msix_vec_count(adev->pdev); unsigned int flags; - if (nvec <= 0) { + if (nvec <= 0) flags = PCI_IRQ_MSI; - } else { + else flags = PCI_IRQ_MSI | PCI_IRQ_MSIX; - } + /* we only need one vector */ nvec = pci_alloc_irq_vectors(adev->pdev, 1, 1, flags); if (nvec > 0) { @@ -331,7 +345,7 @@ void amdgpu_irq_fini_hw(struct amdgpu_device *adev) */ void amdgpu_irq_fini_sw(struct amdgpu_device *adev) { - unsigned i, j; + unsigned int i, j; for (i = 0; i < AMDGPU_IRQ_CLIENTID_MAX; ++i) { if (!adev->irq.client[i].sources) @@ -365,7 +379,7 @@ void amdgpu_irq_fini_sw(struct amdgpu_device *adev) * 0 on success or error code otherwise */ int amdgpu_irq_add_id(struct amdgpu_device *adev, - unsigned client_id, unsigned src_id, + unsigned int client_id, unsigned int src_id, struct amdgpu_irq_src *source) { if (client_id >= AMDGPU_IRQ_CLIENTID_MAX) @@ -417,7 +431,7 @@ void amdgpu_irq_dispatch(struct amdgpu_device *adev, { u32 ring_index = ih->rptr >> 2; struct amdgpu_iv_entry entry; - unsigned client_id, src_id; + unsigned int client_id, src_id; struct amdgpu_irq_src *src; bool handled = false; int r; @@ -453,7 +467,8 @@ void amdgpu_irq_dispatch(struct amdgpu_device *adev, handled = true; } else { - DRM_DEBUG("Unhandled interrupt src_id: %d\n", src_id); + DRM_DEBUG("Unregistered interrupt src_id: %d of client_id:%d\n", + src_id, client_id); } /* Send it to amdkfd as well if it isn't already handled */ @@ -492,7 +507,7 @@ void amdgpu_irq_delegate(struct amdgpu_device *adev, * Updates interrupt state for the specific source (all ASICs). */ int amdgpu_irq_update(struct amdgpu_device *adev, - struct amdgpu_irq_src *src, unsigned type) + struct amdgpu_irq_src *src, unsigned int type) { unsigned long irqflags; enum amdgpu_interrupt_state state; @@ -501,7 +516,8 @@ int amdgpu_irq_update(struct amdgpu_device *adev, spin_lock_irqsave(&adev->irq.lock, irqflags); /* We need to determine after taking the lock, otherwise - we might disable just enabled interrupts again */ + * we might disable just enabled interrupts again + */ if (amdgpu_irq_enabled(adev, src, type)) state = AMDGPU_IRQ_STATE_ENABLE; else @@ -555,7 +571,7 @@ void amdgpu_irq_gpu_reset_resume_helper(struct amdgpu_device *adev) * 0 on success or error code otherwise */ int amdgpu_irq_get(struct amdgpu_device *adev, struct amdgpu_irq_src *src, - unsigned type) + unsigned int type) { if (!adev->irq.installed) return -ENOENT; @@ -585,7 +601,7 @@ int amdgpu_irq_get(struct amdgpu_device *adev, struct amdgpu_irq_src *src, * 0 on success or error code otherwise */ int amdgpu_irq_put(struct amdgpu_device *adev, struct amdgpu_irq_src *src, - unsigned type) + unsigned int type) { if (!adev->irq.installed) return -ENOENT; @@ -619,7 +635,7 @@ int amdgpu_irq_put(struct amdgpu_device *adev, struct amdgpu_irq_src *src, * invalid parameters */ bool amdgpu_irq_enabled(struct amdgpu_device *adev, struct amdgpu_irq_src *src, - unsigned type) + unsigned int type) { if (!adev->irq.installed) return false; @@ -732,7 +748,7 @@ void amdgpu_irq_remove_domain(struct amdgpu_device *adev) * Returns: * Linux IRQ */ -unsigned amdgpu_irq_create_mapping(struct amdgpu_device *adev, unsigned src_id) +unsigned int amdgpu_irq_create_mapping(struct amdgpu_device *adev, unsigned int src_id) { adev->irq.virq[src_id] = irq_create_mapping(adev->irq.domain, src_id); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h index be243adf3e65..04c0b4fa17a4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h @@ -53,7 +53,7 @@ struct amdgpu_iv_entry { uint64_t timestamp; unsigned timestamp_src; unsigned pasid; - unsigned pasid_src; + unsigned node_id; unsigned src_data[AMDGPU_IRQ_SRC_DATA_MAX_SIZE_DW]; const uint32_t *iv_entry; }; @@ -102,6 +102,24 @@ struct amdgpu_irq { bool retry_cam_enabled; }; +enum interrupt_node_id_per_aid { + AID0_NODEID = 0, + XCD0_NODEID = 1, + XCD1_NODEID = 2, + AID1_NODEID = 4, + XCD2_NODEID = 5, + XCD3_NODEID = 6, + AID2_NODEID = 8, + XCD4_NODEID = 9, + XCD5_NODEID = 10, + AID3_NODEID = 12, + XCD6_NODEID = 13, + XCD7_NODEID = 14, + NODEID_MAX, +}; + +extern const int node_id_to_phys_map[NODEID_MAX]; + void amdgpu_irq_disable_all(struct amdgpu_device *adev); int amdgpu_irq_init(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c index c3d9d75143f4..78476bc75b4e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c @@ -65,6 +65,8 @@ static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job) DRM_ERROR("Process information: process %s pid %d thread %s pid %d\n", ti.process_name, ti.tgid, ti.task_name, ti.pid); + dma_fence_set_error(&s_job->s_fence->finished, -ETIME); + if (amdgpu_device_should_recover_gpu(ring->adev)) { struct amdgpu_reset_context reset_context; memset(&reset_context, 0, sizeof(reset_context)); @@ -107,7 +109,7 @@ int amdgpu_job_alloc(struct amdgpu_device *adev, struct amdgpu_vm *vm, (*job)->vm = vm; amdgpu_sync_create(&(*job)->explicit_sync); - (*job)->vram_lost_counter = atomic_read(&adev->vram_lost_counter); + (*job)->generation = amdgpu_vm_generation(adev, vm); (*job)->vm_pd_addr = AMDGPU_BO_INVALID_OFFSET; if (!entity) @@ -256,16 +258,27 @@ amdgpu_job_prepare_job(struct drm_sched_job *sched_job, struct dma_fence *fence = NULL; int r; + /* Ignore soft recovered fences here */ + r = drm_sched_entity_error(s_entity); + if (r && r != -ENODATA) + goto error; + if (!fence && job->gang_submit) fence = amdgpu_device_switch_gang(ring->adev, job->gang_submit); while (!fence && job->vm && !job->vmid) { r = amdgpu_vmid_grab(job->vm, ring, job, &fence); - if (r) + if (r) { DRM_ERROR("Error getting VM ID (%d)\n", r); + goto error; + } } return fence; + +error: + dma_fence_set_error(&job->base.s_fence->finished, r); + return NULL; } static struct dma_fence *amdgpu_job_run(struct drm_sched_job *sched_job) @@ -282,7 +295,7 @@ static struct dma_fence *amdgpu_job_run(struct drm_sched_job *sched_job) trace_amdgpu_sched_run_job(job); /* Skip job if VRAM is lost and never resubmit gangs */ - if (job->vram_lost_counter != atomic_read(&adev->vram_lost_counter) || + if (job->generation != amdgpu_vm_generation(adev, job->vm) || (job->job_run_counter && job->gang_submit)) dma_fence_set_error(finished, -ECANCELED); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h index 52f2e313ea17..a963a25ddd62 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h @@ -61,12 +61,18 @@ struct amdgpu_job { uint32_t gds_base, gds_size; uint32_t gws_base, gws_size; uint32_t oa_base, oa_size; - uint32_t vram_lost_counter; + uint64_t generation; /* user fence handling */ uint64_t uf_addr; uint64_t uf_sequence; + /* virtual addresses for shadow/GDS/CSA */ + uint64_t shadow_va; + uint64_t csa_va; + uint64_t gds_va; + bool init_shadow; + /* job_run_counter >= 1 means a resubmit job */ uint32_t job_run_counter; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c index b07c000fc8ba..3add4b4f0667 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c @@ -45,13 +45,14 @@ int amdgpu_jpeg_sw_init(struct amdgpu_device *adev) int amdgpu_jpeg_sw_fini(struct amdgpu_device *adev) { - int i; + int i, j; for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { if (adev->jpeg.harvest_config & (1 << i)) continue; - amdgpu_ring_fini(&adev->jpeg.inst[i].ring_dec); + for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) + amdgpu_ring_fini(&adev->jpeg.inst[i].ring_dec[j]); } mutex_destroy(&adev->jpeg.jpeg_pg_lock); @@ -76,13 +77,14 @@ static void amdgpu_jpeg_idle_work_handler(struct work_struct *work) struct amdgpu_device *adev = container_of(work, struct amdgpu_device, jpeg.idle_work.work); unsigned int fences = 0; - unsigned int i; + unsigned int i, j; for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { if (adev->jpeg.harvest_config & (1 << i)) continue; - fences += amdgpu_fence_count_emitted(&adev->jpeg.inst[i].ring_dec); + for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) + fences += amdgpu_fence_count_emitted(&adev->jpeg.inst[i].ring_dec[j]); } if (!fences && !atomic_read(&adev->jpeg.total_submission_cnt)) @@ -122,18 +124,21 @@ int amdgpu_jpeg_dec_ring_test_ring(struct amdgpu_ring *ring) if (amdgpu_sriov_vf(adev)) return 0; - WREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch, 0xCAFEDEAD); r = amdgpu_ring_alloc(ring, 3); if (r) return r; - amdgpu_ring_write(ring, PACKET0(adev->jpeg.internal.jpeg_pitch, 0)); - amdgpu_ring_write(ring, 0xDEADBEEF); + WREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch[ring->pipe], 0xCAFEDEAD); + /* Add a read register to make sure the write register is executed. */ + RREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch[ring->pipe]); + + amdgpu_ring_write(ring, PACKET0(adev->jpeg.internal.jpeg_pitch[ring->pipe], 0)); + amdgpu_ring_write(ring, 0xABADCAFE); amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch); - if (tmp == 0xDEADBEEF) + tmp = RREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch[ring->pipe]); + if (tmp == 0xABADCAFE) break; udelay(1); } @@ -161,8 +166,7 @@ static int amdgpu_jpeg_dec_set_reg(struct amdgpu_ring *ring, uint32_t handle, ib = &job->ibs[0]; - ib->ptr[0] = PACKETJ(adev->jpeg.internal.jpeg_pitch, 0, 0, - PACKETJ_TYPE0); + ib->ptr[0] = PACKETJ(adev->jpeg.internal.jpeg_pitch[ring->pipe], 0, 0, PACKETJ_TYPE0); ib->ptr[1] = 0xDEADBEEF; for (i = 2; i < 16; i += 2) { ib->ptr[i] = PACKETJ(0, 0, 0, PACKETJ_TYPE6); @@ -208,7 +212,7 @@ int amdgpu_jpeg_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout) } if (!amdgpu_sriov_vf(adev)) { for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch); + tmp = RREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch[ring->pipe]); if (tmp == 0xDEADBEEF) break; udelay(1); @@ -241,6 +245,31 @@ int amdgpu_jpeg_process_poison_irq(struct amdgpu_device *adev, return 0; } +int amdgpu_jpeg_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block) +{ + int r, i; + + r = amdgpu_ras_block_late_init(adev, ras_block); + if (r) + return r; + + if (amdgpu_ras_is_supported(adev, ras_block->block)) { + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + r = amdgpu_irq_get(adev, &adev->jpeg.inst[i].ras_poison_irq, 0); + if (r) + goto late_fini; + } + } + return 0; + +late_fini: + amdgpu_ras_block_late_fini(adev, ras_block); + return r; +} + int amdgpu_jpeg_ras_sw_init(struct amdgpu_device *adev) { int err; @@ -262,7 +291,7 @@ int amdgpu_jpeg_ras_sw_init(struct amdgpu_device *adev) adev->jpeg.ras_if = &ras->ras_block.ras_comm; if (!ras->ras_block.ras_late_init) - ras->ras_block.ras_late_init = amdgpu_ras_block_late_init; + ras->ras_block.ras_late_init = amdgpu_jpeg_ras_late_init; return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h index 0ca76f0f23e9..ffe47e9f5bf2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h @@ -26,19 +26,22 @@ #include "amdgpu_ras.h" -#define AMDGPU_MAX_JPEG_INSTANCES 2 +#define AMDGPU_MAX_JPEG_INSTANCES 4 +#define AMDGPU_MAX_JPEG_RINGS 8 #define AMDGPU_JPEG_HARVEST_JPEG0 (1 << 0) #define AMDGPU_JPEG_HARVEST_JPEG1 (1 << 1) struct amdgpu_jpeg_reg{ - unsigned jpeg_pitch; + unsigned jpeg_pitch[AMDGPU_MAX_JPEG_RINGS]; }; struct amdgpu_jpeg_inst { - struct amdgpu_ring ring_dec; + struct amdgpu_ring ring_dec[AMDGPU_MAX_JPEG_RINGS]; struct amdgpu_irq_src irq; + struct amdgpu_irq_src ras_poison_irq; struct amdgpu_jpeg_reg external; + uint8_t aid_id; }; struct amdgpu_jpeg_ras { @@ -48,6 +51,7 @@ struct amdgpu_jpeg_ras { struct amdgpu_jpeg { uint8_t num_jpeg_inst; struct amdgpu_jpeg_inst inst[AMDGPU_MAX_JPEG_INSTANCES]; + unsigned num_jpeg_rings; struct amdgpu_jpeg_reg internal; unsigned harvest_config; struct delayed_work idle_work; @@ -56,6 +60,9 @@ struct amdgpu_jpeg { atomic_t total_submission_cnt; struct ras_common_if *ras_if; struct amdgpu_jpeg_ras *ras; + + uint16_t inst_mask; + uint8_t num_inst_per_aid; }; int amdgpu_jpeg_sw_init(struct amdgpu_device *adev); @@ -72,6 +79,8 @@ int amdgpu_jpeg_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout); int amdgpu_jpeg_process_poison_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry); +int amdgpu_jpeg_ras_late_init(struct amdgpu_device *adev, + struct ras_common_if *ras_block); int amdgpu_jpeg_ras_sw_init(struct amdgpu_device *adev); #endif /*__AMDGPU_JPEG_H__*/ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index 0efb38539d70..e3531aa3c8bd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -462,8 +462,9 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, if (adev->jpeg.harvest_config & (1 << i)) continue; - if (adev->jpeg.inst[i].ring_dec.sched.ready) - ++num_rings; + for (j = 0; j < adev->jpeg.num_jpeg_rings; j++) + if (adev->jpeg.inst[i].ring_dec[j].sched.ready) + ++num_rings; } ib_start_alignment = 16; ib_size_alignment = 16; @@ -876,6 +877,19 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) dev_info->gl2c_cache_size = adev->gfx.config.gc_gl2c_per_gpu; dev_info->mall_size = adev->gmc.mall_size; + + if (adev->gfx.funcs->get_gfx_shadow_info) { + struct amdgpu_gfx_shadow_info shadow_info; + + ret = amdgpu_gfx_get_gfx_shadow_info(adev, &shadow_info); + if (!ret) { + dev_info->shadow_size = shadow_info.shadow_size; + dev_info->shadow_alignment = shadow_info.shadow_alignment; + dev_info->csa_size = shadow_info.csa_size; + dev_info->csa_alignment = shadow_info.csa_alignment; + } + } + ret = copy_to_user(out, dev_info, min((size_t)size, sizeof(*dev_info))) ? -EFAULT : 0; kfree(dev_info); @@ -1140,6 +1154,15 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) kfree(caps); return r; } + case AMDGPU_INFO_MAX_IBS: { + uint32_t max_ibs[AMDGPU_HW_IP_NUM]; + + for (i = 0; i < AMDGPU_HW_IP_NUM; ++i) + max_ibs[i] = amdgpu_ring_max_ibs(i); + + return copy_to_user(out, max_ibs, + min((size_t)size, sizeof(max_ibs))) ? -EFAULT : 0; + } default: DRM_DEBUG_KMS("Invalid request %d\n", info->query); return -EINVAL; @@ -1210,6 +1233,10 @@ int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv) if (r) goto error_pasid; + r = amdgpu_xcp_open_device(adev, fpriv, file_priv); + if (r) + goto error_vm; + r = amdgpu_vm_set_pasid(adev, &fpriv->vm, pasid); if (r) goto error_vm; @@ -1284,12 +1311,12 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev, if (amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_VCE) != NULL) amdgpu_vce_free_handles(adev, file_priv); - if (amdgpu_mcbp) { - /* TODO: how to handle reserve failure */ - BUG_ON(amdgpu_bo_reserve(adev->virt.csa_obj, true)); - amdgpu_vm_bo_del(adev, fpriv->csa_va); + if (fpriv->csa_va) { + uint64_t csa_addr = amdgpu_csa_vaddr(adev) & AMDGPU_GMC_HOLE_MASK; + + WARN_ON(amdgpu_unmap_static_csa(adev, &fpriv->vm, adev->virt.csa_obj, + fpriv->csa_va, csa_addr)); fpriv->csa_va = NULL; - amdgpu_bo_unreserve(adev->virt.csa_obj); } pasid = fpriv->vm.pasid; @@ -1441,7 +1468,7 @@ void amdgpu_disable_vblank_kms(struct drm_crtc *crtc) static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused) { - struct amdgpu_device *adev = (struct amdgpu_device *)m->private; + struct amdgpu_device *adev = m->private; struct drm_amdgpu_info_firmware fw_info; struct drm_amdgpu_query_fw query_fw; struct atom_context *ctx = adev->mode_info.atom_context; @@ -1449,7 +1476,7 @@ static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused) int ret, i; static const char *ta_fw_name[TA_FW_TYPE_MAX_INDEX] = { -#define TA_FW_NAME(type) [TA_FW_TYPE_PSP_##type] = #type +#define TA_FW_NAME(type)[TA_FW_TYPE_PSP_##type] = #type TA_FW_NAME(XGMI), TA_FW_NAME(RAS), TA_FW_NAME(HDCP), @@ -1548,7 +1575,7 @@ static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused) fw_info.feature, fw_info.ver); /* RLCV */ - query_fw.fw_type = AMDGPU_INFO_FW_GFX_RLCV; + query_fw.fw_type = AMDGPU_INFO_FW_GFX_RLCV; ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); if (ret) return ret; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c index f0f00466b59f..e9091ebfe230 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c @@ -924,6 +924,43 @@ error: return r; } +int amdgpu_mes_set_shader_debugger(struct amdgpu_device *adev, + uint64_t process_context_addr, + uint32_t spi_gdbg_per_vmid_cntl, + const uint32_t *tcp_watch_cntl, + uint32_t flags, + bool trap_en) +{ + struct mes_misc_op_input op_input = {0}; + int r; + + if (!adev->mes.funcs->misc_op) { + DRM_ERROR("mes set shader debugger is not supported!\n"); + return -EINVAL; + } + + op_input.op = MES_MISC_OP_SET_SHADER_DEBUGGER; + op_input.set_shader_debugger.process_context_addr = process_context_addr; + op_input.set_shader_debugger.flags.u32all = flags; + op_input.set_shader_debugger.spi_gdbg_per_vmid_cntl = spi_gdbg_per_vmid_cntl; + memcpy(op_input.set_shader_debugger.tcp_watch_cntl, tcp_watch_cntl, + sizeof(op_input.set_shader_debugger.tcp_watch_cntl)); + + if (((adev->mes.sched_version & AMDGPU_MES_API_VERSION_MASK) >> + AMDGPU_MES_API_VERSION_SHIFT) >= 14) + op_input.set_shader_debugger.trap_en = trap_en; + + amdgpu_mes_lock(&adev->mes); + + r = adev->mes.funcs->misc_op(&adev->mes, &op_input); + if (r) + DRM_ERROR("failed to set_shader_debugger\n"); + + amdgpu_mes_unlock(&adev->mes); + + return r; +} + static void amdgpu_mes_ring_to_queue_props(struct amdgpu_device *adev, struct amdgpu_ring *ring, @@ -1305,14 +1342,9 @@ static int amdgpu_mes_test_queues(struct amdgpu_ring **added_rings) if (!ring) continue; - r = amdgpu_ring_test_ring(ring); - if (r) { - DRM_DEV_ERROR(ring->adev->dev, - "ring %s test failed (%d)\n", - ring->name, r); + r = amdgpu_ring_test_helper(ring); + if (r) return r; - } else - DRM_INFO("ring %s test pass\n", ring->name); r = amdgpu_ring_test_ib(ring, 1000 * 10); if (r) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h index 547ec35691fa..2d6ac30b7135 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h @@ -219,6 +219,8 @@ struct mes_add_queue_input { uint32_t gws_size; uint64_t tba_addr; uint64_t tma_addr; + uint32_t trap_en; + uint32_t skip_process_ctx_clear; uint32_t is_kfd_process; uint32_t is_aql_queue; uint32_t queue_size; @@ -256,6 +258,7 @@ enum mes_misc_opcode { MES_MISC_OP_READ_REG, MES_MISC_OP_WRM_REG_WAIT, MES_MISC_OP_WRM_REG_WR_WAIT, + MES_MISC_OP_SET_SHADER_DEBUGGER, }; struct mes_misc_op_input { @@ -278,6 +281,21 @@ struct mes_misc_op_input { uint32_t reg0; uint32_t reg1; } wrm_reg; + + struct { + uint64_t process_context_addr; + union { + struct { + uint64_t single_memop : 1; + uint64_t single_alu_op : 1; + uint64_t reserved: 30; + }; + uint32_t u32all; + } flags; + uint32_t spi_gdbg_per_vmid_cntl; + uint32_t tcp_watch_cntl[4]; + uint32_t trap_en; + } set_shader_debugger; }; }; @@ -340,6 +358,12 @@ int amdgpu_mes_reg_wait(struct amdgpu_device *adev, uint32_t reg, int amdgpu_mes_reg_write_reg_wait(struct amdgpu_device *adev, uint32_t reg0, uint32_t reg1, uint32_t ref, uint32_t mask); +int amdgpu_mes_set_shader_debugger(struct amdgpu_device *adev, + uint64_t process_context_addr, + uint32_t spi_gdbg_per_vmid_cntl, + const uint32_t *tcp_watch_cntl, + uint32_t flags, + bool trap_en); int amdgpu_mes_add_ring(struct amdgpu_device *adev, int gang_id, int queue_type, int idx, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.h index d21bb6dae56e..1ca9d4ed8063 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.h @@ -21,6 +21,29 @@ #ifndef __AMDGPU_MMHUB_H__ #define __AMDGPU_MMHUB_H__ +enum amdgpu_mmhub_ras_memory_id { + AMDGPU_MMHUB_WGMI_PAGEMEM = 0, + AMDGPU_MMHUB_RGMI_PAGEMEM = 1, + AMDGPU_MMHUB_WDRAM_PAGEMEM = 2, + AMDGPU_MMHUB_RDRAM_PAGEMEM = 3, + AMDGPU_MMHUB_WIO_CMDMEM = 4, + AMDGPU_MMHUB_RIO_CMDMEM = 5, + AMDGPU_MMHUB_WGMI_CMDMEM = 6, + AMDGPU_MMHUB_RGMI_CMDMEM = 7, + AMDGPU_MMHUB_WDRAM_CMDMEM = 8, + AMDGPU_MMHUB_RDRAM_CMDMEM = 9, + AMDGPU_MMHUB_MAM_DMEM0 = 10, + AMDGPU_MMHUB_MAM_DMEM1 = 11, + AMDGPU_MMHUB_MAM_DMEM2 = 12, + AMDGPU_MMHUB_MAM_DMEM3 = 13, + AMDGPU_MMHUB_WRET_TAGMEM = 19, + AMDGPU_MMHUB_RRET_TAGMEM = 20, + AMDGPU_MMHUB_WIO_DATAMEM = 21, + AMDGPU_MMHUB_WGMI_DATAMEM = 22, + AMDGPU_MMHUB_WDRAM_DATAMEM = 23, + AMDGPU_MMHUB_MEMORY_BLOCK_LAST, +}; + struct amdgpu_mmhub_ras { struct amdgpu_ras_block_object ras_block; }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h index c686ff4bcc39..8ab8ae01f87c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h @@ -61,6 +61,7 @@ struct amdgpu_nbio_funcs { u32 (*get_hdp_flush_done_offset)(struct amdgpu_device *adev); u32 (*get_pcie_index_offset)(struct amdgpu_device *adev); u32 (*get_pcie_data_offset)(struct amdgpu_device *adev); + u32 (*get_pcie_index_hi_offset)(struct amdgpu_device *adev); u32 (*get_pcie_port_index_offset)(struct amdgpu_device *adev); u32 (*get_pcie_port_data_offset)(struct amdgpu_device *adev); u32 (*get_rev_id)(struct amdgpu_device *adev); @@ -95,6 +96,9 @@ struct amdgpu_nbio_funcs { void (*apply_l1_link_width_reconfig_wa)(struct amdgpu_device *adev); void (*clear_doorbell_interrupt)(struct amdgpu_device *adev); u32 (*get_rom_offset)(struct amdgpu_device *adev); + int (*get_compute_partition_mode)(struct amdgpu_device *adev); + u32 (*get_memory_partition_mode)(struct amdgpu_device *adev, + u32 *supp_modes); }; struct amdgpu_nbio { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 2bd1a54ee866..f7905bce0de1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -79,9 +79,10 @@ static void amdgpu_bo_user_destroy(struct ttm_buffer_object *tbo) static void amdgpu_bo_vm_destroy(struct ttm_buffer_object *tbo) { struct amdgpu_device *adev = amdgpu_ttm_adev(tbo->bdev); - struct amdgpu_bo *bo = ttm_to_amdgpu_bo(tbo); + struct amdgpu_bo *shadow_bo = ttm_to_amdgpu_bo(tbo), *bo; struct amdgpu_bo_vm *vmbo; + bo = shadow_bo->parent; vmbo = to_amdgpu_bo_vm(bo); /* in case amdgpu_device_recover_vram got NULL of bo->parent */ if (!list_empty(&vmbo->shadow_list)) { @@ -130,16 +131,26 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain) u32 c = 0; if (domain & AMDGPU_GEM_DOMAIN_VRAM) { - unsigned visible_pfn = adev->gmc.visible_vram_size >> PAGE_SHIFT; - - places[c].fpfn = 0; - places[c].lpfn = 0; + unsigned int visible_pfn = adev->gmc.visible_vram_size >> PAGE_SHIFT; + int8_t mem_id = KFD_XCP_MEM_ID(adev, abo->xcp_id); + + if (adev->gmc.mem_partitions && mem_id >= 0) { + places[c].fpfn = adev->gmc.mem_partitions[mem_id].range.fpfn; + /* + * memory partition range lpfn is inclusive start + size - 1 + * TTM place lpfn is exclusive start + size + */ + places[c].lpfn = adev->gmc.mem_partitions[mem_id].range.lpfn + 1; + } else { + places[c].fpfn = 0; + places[c].lpfn = 0; + } places[c].mem_type = TTM_PL_VRAM; places[c].flags = 0; if (flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED) - places[c].lpfn = visible_pfn; - else if (adev->gmc.real_vram_size != adev->gmc.visible_vram_size) + places[c].lpfn = min_not_zero(places[c].lpfn, visible_pfn); + else places[c].flags |= TTM_PL_FLAG_TOPDOWN; if (flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS) @@ -574,6 +585,13 @@ int amdgpu_bo_create(struct amdgpu_device *adev, bo->flags = bp->flags; + if (adev->gmc.mem_partitions) + /* For GPUs with spatial partitioning, bo->xcp_id=-1 means any partition */ + bo->xcp_id = bp->xcp_id_plus1 - 1; + else + /* For GPUs without spatial partitioning */ + bo->xcp_id = 0; + if (!amdgpu_bo_support_uswc(bo->flags)) bo->flags &= ~AMDGPU_GEM_CREATE_CPU_GTT_USWC; @@ -610,7 +628,7 @@ int amdgpu_bo_create(struct amdgpu_device *adev, bo->tbo.resource->mem_type == TTM_PL_VRAM) { struct dma_fence *fence; - r = amdgpu_fill_buffer(bo, 0, bo->tbo.base.resv, &fence); + r = amdgpu_fill_buffer(bo, 0, bo->tbo.base.resv, &fence, true); if (unlikely(r)) goto fail_unreserve; @@ -694,11 +712,6 @@ int amdgpu_bo_create_vm(struct amdgpu_device *adev, return r; *vmbo_ptr = to_amdgpu_bo_vm(bo_ptr); - INIT_LIST_HEAD(&(*vmbo_ptr)->shadow_list); - /* Set destroy callback to amdgpu_bo_vm_destroy after vmbo->shadow_list - * is initialized. - */ - bo_ptr->tbo.destroy = &amdgpu_bo_vm_destroy; return r; } @@ -715,6 +728,8 @@ void amdgpu_bo_add_to_shadow_list(struct amdgpu_bo_vm *vmbo) mutex_lock(&adev->shadow_list_lock); list_add_tail(&vmbo->shadow_list, &adev->shadow_list); + vmbo->shadow->parent = amdgpu_bo_ref(&vmbo->bo); + vmbo->shadow->tbo.destroy = &amdgpu_bo_vm_destroy; mutex_unlock(&adev->shadow_list_lock); } @@ -935,7 +950,7 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain, bo->flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED; amdgpu_bo_placement_from_domain(bo, domain); for (i = 0; i < bo->placement.num_placement; i++) { - unsigned fpfn, lpfn; + unsigned int fpfn, lpfn; fpfn = min_offset >> PAGE_SHIFT; lpfn = max_offset >> PAGE_SHIFT; @@ -1016,7 +1031,7 @@ void amdgpu_bo_unpin(struct amdgpu_bo *bo) } } -static const char *amdgpu_vram_names[] = { +static const char * const amdgpu_vram_names[] = { "UNKNOWN", "GDDR1", "DDR2", @@ -1044,7 +1059,7 @@ static const char *amdgpu_vram_names[] = { int amdgpu_bo_init(struct amdgpu_device *adev) { /* On A+A platform, VRAM can be mapped as WB */ - if (!adev->gmc.xgmi.connected_to_cpu) { + if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu) { /* reserve PAT memory space to WC for VRAM */ int r = arch_io_reserve_memtype_wc(adev->gmc.aper_base, adev->gmc.aper_size); @@ -1080,8 +1095,7 @@ void amdgpu_bo_fini(struct amdgpu_device *adev) amdgpu_ttm_fini(adev); if (drm_dev_enter(adev_to_drm(adev), &idx)) { - - if (!adev->gmc.xgmi.connected_to_cpu) { + if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu) { arch_phys_wc_del(adev->gmc.vram_mtrr); arch_io_free_memtype_wc(adev->gmc.aper_base, adev->gmc.aper_size); } @@ -1148,8 +1162,8 @@ void amdgpu_bo_get_tiling_flags(struct amdgpu_bo *bo, u64 *tiling_flags) * Returns: * 0 for success or a negative error code on failure. */ -int amdgpu_bo_set_metadata (struct amdgpu_bo *bo, void *metadata, - uint32_t metadata_size, uint64_t flags) +int amdgpu_bo_set_metadata(struct amdgpu_bo *bo, void *metadata, + u32 metadata_size, uint64_t flags) { struct amdgpu_bo_user *ubo; void *buffer; @@ -1268,8 +1282,12 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo, void amdgpu_bo_get_memory(struct amdgpu_bo *bo, struct amdgpu_mem_stats *stats) { - unsigned int domain; uint64_t size = amdgpu_bo_size(bo); + unsigned int domain; + + /* Abort if the BO doesn't currently have a backing store */ + if (!bo->tbo.resource) + return; domain = amdgpu_mem_type_to_domain(bo->tbo.resource->mem_type); switch (domain) { @@ -1338,7 +1356,7 @@ void amdgpu_bo_release_notify(struct ttm_buffer_object *bo) if (WARN_ON_ONCE(!dma_resv_trylock(bo->base.resv))) return; - r = amdgpu_fill_buffer(abo, AMDGPU_POISON, bo->base.resv, &fence); + r = amdgpu_fill_buffer(abo, AMDGPU_POISON, bo->base.resv, &fence, true); if (!WARN_ON(r)) { amdgpu_bo_fence(abo, fence, false); dma_fence_put(fence); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h index 35b8106816a1..05496b97ef93 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h @@ -56,6 +56,8 @@ struct amdgpu_bo_param { bool no_wait_gpu; struct dma_resv *resv; void (*destroy)(struct ttm_buffer_object *bo); + /* xcp partition number plus 1, 0 means any partition */ + int8_t xcp_id_plus1; }; /* bo virtual addresses in a vm */ @@ -108,6 +110,13 @@ struct amdgpu_bo { struct mmu_interval_notifier notifier; #endif struct kgd_mem *kfd_bo; + + /* + * For GPUs with spatial partitioning, xcp partition number, -1 means + * any partition. For other ASICs without spatial partition, always 0 + * for memory accounting. + */ + int8_t xcp_id; }; struct amdgpu_bo_user { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index 9d7e6e0e73ed..e15c27e05564 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c @@ -146,6 +146,9 @@ static int psp_init_sriov_microcode(struct psp_context *psp) case IP_VERSION(13, 0, 0): adev->virt.autoload_ucode_id = 0; break; + case IP_VERSION(13, 0, 6): + ret = psp_init_cap_microcode(psp, ucode_prefix); + break; case IP_VERSION(13, 0, 10): adev->virt.autoload_ucode_id = AMDGPU_UCODE_ID_CP_MES1_DATA; ret = psp_init_cap_microcode(psp, ucode_prefix); @@ -329,6 +332,9 @@ static bool psp_get_runtime_db_entry(struct amdgpu_device *adev, bool ret = false; int i; + if (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(13, 0, 6)) + return false; + db_header_pos = adev->gmc.mc_vram_size - PSP_RUNTIME_DB_OFFSET; db_dir_pos = db_header_pos + sizeof(struct psp_runtime_data_header); @@ -411,7 +417,7 @@ static int psp_sw_init(void *handle) if ((psp_get_runtime_db_entry(adev, PSP_RUNTIME_ENTRY_TYPE_PPTABLE_ERR_STATUS, &scpm_entry)) && - (SCPM_DISABLE != scpm_entry.scpm_status)) { + (scpm_entry.scpm_status != SCPM_DISABLE)) { adev->scpm_enabled = true; adev->scpm_status = scpm_entry.scpm_status; } else { @@ -458,10 +464,9 @@ static int psp_sw_init(void *handle) if (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 0) || adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 7)) { - ret= psp_sysfs_init(adev); - if (ret) { + ret = psp_sysfs_init(adev); + if (ret) return ret; - } } ret = amdgpu_bo_create_kernel(adev, PSP_1_MEG, PSP_1_MEG, @@ -474,7 +479,8 @@ static int psp_sw_init(void *handle) return ret; ret = amdgpu_bo_create_kernel(adev, PSP_FENCE_BUFFER_SIZE, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_DOMAIN_VRAM | + AMDGPU_GEM_DOMAIN_GTT, &psp->fence_buf_bo, &psp->fence_buf_mc_addr, &psp->fence_buf); @@ -482,7 +488,8 @@ static int psp_sw_init(void *handle) goto failed1; ret = amdgpu_bo_create_kernel(adev, PSP_CMD_BUFFER_SIZE, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_DOMAIN_VRAM | + AMDGPU_GEM_DOMAIN_GTT, &psp->cmd_buf_bo, &psp->cmd_buf_mc_addr, (void **)&psp->cmd_buf_mem); if (ret) @@ -520,6 +527,8 @@ static int psp_sw_fini(void *handle) kfree(cmd); cmd = NULL; + psp_free_shared_bufs(psp); + if (psp->km_ring.ring_mem) amdgpu_bo_free_kernel(&adev->firmware.rbuf, &psp->km_ring.ring_mem_mc_addr, @@ -560,6 +569,26 @@ int psp_wait_for(struct psp_context *psp, uint32_t reg_index, return -ETIME; } +int psp_wait_for_spirom_update(struct psp_context *psp, uint32_t reg_index, + uint32_t reg_val, uint32_t mask, uint32_t msec_timeout) +{ + uint32_t val; + int i; + struct amdgpu_device *adev = psp->adev; + + if (psp->adev->no_hw_access) + return 0; + + for (i = 0; i < msec_timeout; i++) { + val = RREG32(reg_index); + if ((val & mask) == reg_val) + return 0; + msleep(1); + } + + return -ETIME; +} + static const char *psp_gfx_cmd_name(enum psp_gfx_cmd_id cmd_id) { switch (cmd_id) { @@ -643,7 +672,7 @@ psp_cmd_submit_buf(struct psp_context *psp, skip_unsupport = (psp->cmd_buf_mem->resp.status == TEE_ERROR_NOT_SUPPORTED || psp->cmd_buf_mem->resp.status == PSP_ERR_UNKNOWN_COMMAND) && amdgpu_sriov_vf(psp->adev); - memcpy((void*)&cmd->resp, (void*)&psp->cmd_buf_mem->resp, sizeof(struct psp_gfx_resp)); + memcpy(&cmd->resp, &psp->cmd_buf_mem->resp, sizeof(struct psp_gfx_resp)); /* In some cases, psp response status is not 0 even there is no * problem while the command is submitted. Some version of PSP FW @@ -699,8 +728,13 @@ static void psp_prep_tmr_cmd_buf(struct psp_context *psp, uint64_t tmr_mc, struct amdgpu_bo *tmr_bo) { struct amdgpu_device *adev = psp->adev; - uint32_t size = amdgpu_bo_size(tmr_bo); - uint64_t tmr_pa = amdgpu_gmc_vram_pa(adev, tmr_bo); + uint32_t size = 0; + uint64_t tmr_pa = 0; + + if (tmr_bo) { + size = amdgpu_bo_size(tmr_bo); + tmr_pa = amdgpu_gmc_vram_pa(adev, tmr_bo); + } if (amdgpu_sriov_vf(psp->adev)) cmd->cmd_id = GFX_CMD_ID_SETUP_VMR; @@ -745,6 +779,16 @@ static int psp_load_toc(struct psp_context *psp, return ret; } +static bool psp_boottime_tmr(struct psp_context *psp) +{ + switch (psp->adev->ip_versions[MP0_HWIP][0]) { + case IP_VERSION(13, 0, 6): + return true; + default: + return false; + } +} + /* Set up Trusted Memory Region */ static int psp_tmr_init(struct psp_context *psp) { @@ -816,8 +860,9 @@ static int psp_tmr_load(struct psp_context *psp) cmd = acquire_psp_cmd_buf(psp); psp_prep_tmr_cmd_buf(psp, cmd, psp->tmr_mc_addr, psp->tmr_bo); - DRM_INFO("reserve 0x%lx from 0x%llx for PSP TMR\n", - amdgpu_bo_size(psp->tmr_bo), psp->tmr_mc_addr); + if (psp->tmr_bo) + DRM_INFO("reserve 0x%lx from 0x%llx for PSP TMR\n", + amdgpu_bo_size(psp->tmr_bo), psp->tmr_mc_addr); ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); @@ -828,7 +873,7 @@ static int psp_tmr_load(struct psp_context *psp) } static void psp_prep_tmr_unload_cmd_buf(struct psp_context *psp, - struct psp_gfx_cmd_resp *cmd) + struct psp_gfx_cmd_resp *cmd) { if (amdgpu_sriov_vf(psp->adev)) cmd->cmd_id = GFX_CMD_ID_DESTROY_VMR; @@ -969,6 +1014,27 @@ static int psp_rl_load(struct amdgpu_device *adev) return ret; } +int psp_spatial_partition(struct psp_context *psp, int mode) +{ + struct psp_gfx_cmd_resp *cmd; + int ret; + + if (amdgpu_sriov_vf(psp->adev)) + return 0; + + cmd = acquire_psp_cmd_buf(psp); + + cmd->cmd_id = GFX_CMD_ID_SRIOV_SPATIAL_PART; + cmd->cmd.cmd_spatial_part.mode = mode; + + dev_info(psp->adev->dev, "Requesting %d partitions through PSP", mode); + ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr); + + release_psp_cmd_buf(psp); + + return ret; +} + static int psp_asd_initialize(struct psp_context *psp) { int ret; @@ -1065,7 +1131,7 @@ static void psp_prep_ta_load_cmd_buf(struct psp_gfx_cmd_resp *cmd, struct ta_context *context) { cmd->cmd_id = context->ta_load_type; - cmd->cmd.cmd_load_ta.app_phy_addr_lo = lower_32_bits(ta_bin_mc); + cmd->cmd.cmd_load_ta.app_phy_addr_lo = lower_32_bits(ta_bin_mc); cmd->cmd.cmd_load_ta.app_phy_addr_hi = upper_32_bits(ta_bin_mc); cmd->cmd.cmd_load_ta.app_len = context->bin_desc.size_bytes; @@ -1136,9 +1202,8 @@ int psp_ta_load(struct psp_context *psp, struct ta_context *context) context->resp_status = cmd->resp.status; - if (!ret) { + if (!ret) context->session_id = cmd->resp.session_id; - } release_psp_cmd_buf(psp); @@ -1254,8 +1319,9 @@ int psp_xgmi_get_node_id(struct psp_context *psp, uint64_t *node_id) static bool psp_xgmi_peer_link_info_supported(struct psp_context *psp) { - return psp->adev->ip_versions[MP0_HWIP][0] == IP_VERSION(13, 0, 2) && - psp->xgmi_context.context.bin_desc.fw_version >= 0x2000000b; + return (psp->adev->ip_versions[MP0_HWIP][0] == IP_VERSION(13, 0, 2) && + psp->xgmi_context.context.bin_desc.fw_version >= 0x2000000b) || + psp->adev->ip_versions[MP0_HWIP][0] >= IP_VERSION(13, 0, 6); } /* @@ -1363,6 +1429,9 @@ int psp_xgmi_get_topology_info(struct psp_context *psp, /* Invoke xgmi ta again to get the link information */ if (psp_xgmi_peer_link_info_supported(psp)) { struct ta_xgmi_cmd_get_peer_link_info_output *link_info_output; + bool requires_reflection = + (psp->xgmi_context.supports_extended_data && get_extended_data) || + psp->adev->ip_versions[MP0_HWIP][0] == IP_VERSION(13, 0, 6); xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_PEER_LINKS; @@ -1377,11 +1446,11 @@ int psp_xgmi_get_topology_info(struct psp_context *psp, topology->nodes[i].num_links = get_extended_data ? topology->nodes[i].num_links + link_info_output->nodes[i].num_links : - link_info_output->nodes[i].num_links; + ((requires_reflection && topology->nodes[i].num_links) ? topology->nodes[i].num_links : + link_info_output->nodes[i].num_links); /* reflect the topology information for bi-directionality */ - if (psp->xgmi_context.supports_extended_data && - get_extended_data && topology->nodes[i].num_hops) + if (requires_reflection && topology->nodes[i].num_hops) psp_xgmi_reflect_topology_info(psp, topology->nodes[i]); } } @@ -1465,8 +1534,7 @@ int psp_ras_invoke(struct psp_context *psp, uint32_t ta_cmd_id) if (amdgpu_ras_intr_triggered()) return ret; - if (ras_cmd->if_version > RAS_TA_HOST_IF_VER) - { + if (ras_cmd->if_version > RAS_TA_HOST_IF_VER) { DRM_WARN("RAS: Unsupported Interface"); return -EINVAL; } @@ -1476,8 +1544,7 @@ int psp_ras_invoke(struct psp_context *psp, uint32_t ta_cmd_id) dev_warn(psp->adev->dev, "ECC switch disabled\n"); ras_cmd->ras_status = TA_RAS_STATUS__ERROR_RAS_NOT_AVAILABLE; - } - else if (ras_cmd->ras_out_message.flags.reg_access_failure_flag) + } else if (ras_cmd->ras_out_message.flags.reg_access_failure_flag) dev_warn(psp->adev->dev, "RAS internal register access blocked\n"); @@ -1573,11 +1640,10 @@ int psp_ras_initialize(struct psp_context *psp) if (ret) dev_warn(adev->dev, "PSP set boot config failed\n"); else - dev_warn(adev->dev, "GECC will be disabled in next boot cycle " - "if set amdgpu_ras_enable and/or amdgpu_ras_mask to 0x0\n"); + dev_warn(adev->dev, "GECC will be disabled in next boot cycle if set amdgpu_ras_enable and/or amdgpu_ras_mask to 0x0\n"); } } else { - if (1 == boot_cfg) { + if (boot_cfg == 1) { dev_info(adev->dev, "GECC is enabled\n"); } else { /* enable GECC in next boot cycle if it is disabled @@ -1607,8 +1673,11 @@ int psp_ras_initialize(struct psp_context *psp) if (amdgpu_ras_is_poison_mode_supported(adev)) ras_cmd->ras_in_message.init_flags.poison_mode_en = 1; - if (!adev->gmc.xgmi.connected_to_cpu) + if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu) ras_cmd->ras_in_message.init_flags.dgpu_mode = 1; + ras_cmd->ras_in_message.init_flags.xcc_mask = + adev->gfx.xcc_mask; + ras_cmd->ras_in_message.init_flags.channel_dis_num = hweight32(adev->gmc.m_half_use) * 2; ret = psp_ta_load(psp, &psp->ras_context.context); @@ -1626,14 +1695,37 @@ int psp_ras_initialize(struct psp_context *psp) } int psp_ras_trigger_error(struct psp_context *psp, - struct ta_ras_trigger_error_input *info) + struct ta_ras_trigger_error_input *info, uint32_t instance_mask) { struct ta_ras_shared_memory *ras_cmd; + struct amdgpu_device *adev = psp->adev; int ret; + uint32_t dev_mask; if (!psp->ras_context.context.initialized) return -EINVAL; + switch (info->block_id) { + case TA_RAS_BLOCK__GFX: + dev_mask = GET_MASK(GC, instance_mask); + break; + case TA_RAS_BLOCK__SDMA: + dev_mask = GET_MASK(SDMA0, instance_mask); + break; + case TA_RAS_BLOCK__VCN: + case TA_RAS_BLOCK__JPEG: + dev_mask = GET_MASK(VCN, instance_mask); + break; + default: + dev_mask = instance_mask; + break; + } + + /* reuse sub_block_index for backward compatibility */ + dev_mask <<= AMDGPU_RAS_INST_SHIFT; + dev_mask &= AMDGPU_RAS_INST_MASK; + info->sub_block_index |= dev_mask; + ras_cmd = (struct ta_ras_shared_memory *)psp->ras_context.context.mem_context.shared_buf; memset(ras_cmd, 0, sizeof(struct ta_ras_shared_memory)); @@ -2077,10 +2169,12 @@ static int psp_hw_start(struct psp_context *psp) if (amdgpu_sriov_vf(adev) && amdgpu_in_reset(adev)) goto skip_pin_bo; - ret = psp_tmr_init(psp); - if (ret) { - DRM_ERROR("PSP tmr init failed!\n"); - return ret; + if (!psp_boottime_tmr(psp)) { + ret = psp_tmr_init(psp); + if (ret) { + DRM_ERROR("PSP tmr init failed!\n"); + return ret; + } } skip_pin_bo: @@ -2363,7 +2457,7 @@ static int psp_prep_load_ip_fw_cmd_buf(struct amdgpu_firmware_info *ucode, } static int psp_execute_non_psp_fw_load(struct psp_context *psp, - struct amdgpu_firmware_info *ucode) + struct amdgpu_firmware_info *ucode) { int ret = 0; struct psp_gfx_cmd_resp *cmd = acquire_psp_cmd_buf(psp); @@ -2402,9 +2496,8 @@ static int psp_load_smu_fw(struct psp_context *psp) (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 4) || adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 2)))) { ret = amdgpu_dpm_set_mp1_state(adev, PP_MP1_STATE_UNLOAD); - if (ret) { + if (ret) DRM_WARN("Failed to set MP1 state prepare for reload\n"); - } } ret = psp_execute_non_psp_fw_load(psp, ucode); @@ -2655,8 +2748,6 @@ static int psp_hw_fini(void *handle) psp_ring_destroy(psp, PSP_RING_TYPE__KM); - psp_free_shared_bufs(psp); - return 0; } @@ -2716,9 +2807,8 @@ static int psp_suspend(void *handle) } ret = psp_ring_stop(psp, PSP_RING_TYPE__KM); - if (ret) { + if (ret) DRM_ERROR("PSP ring stop failed\n"); - } out: return ret; @@ -2967,7 +3057,7 @@ static int parse_sos_bin_descriptor(struct psp_context *psp, psp->sos.fw_version = le32_to_cpu(desc->fw_version); psp->sos.feature_version = le32_to_cpu(desc->fw_version); psp->sos.size_bytes = le32_to_cpu(desc->size_bytes); - psp->sos.start_addr = ucode_start_addr; + psp->sos.start_addr = ucode_start_addr; break; case PSP_FW_TYPE_PSP_SYS_DRV: psp->sys.fw_version = le32_to_cpu(desc->fw_version); @@ -3491,7 +3581,7 @@ void psp_copy_fw(struct psp_context *psp, uint8_t *start_addr, uint32_t bin_size drm_dev_exit(idx); } -static DEVICE_ATTR(usbc_pd_fw, S_IRUGO | S_IWUSR, +static DEVICE_ATTR(usbc_pd_fw, 0644, psp_usbc_pd_fw_sysfs_read, psp_usbc_pd_fw_sysfs_write); @@ -3548,6 +3638,9 @@ static ssize_t amdgpu_psp_vbflash_read(struct file *filp, struct kobject *kobj, void *fw_pri_cpu_addr; int ret; + if (adev->psp.vbflash_image_size == 0) + return -EINVAL; + dev_info(adev->dev, "VBIOS flash to PSP started"); ret = amdgpu_bo_create_kernel(adev, adev->psp.vbflash_image_size, @@ -3599,13 +3692,13 @@ static ssize_t amdgpu_psp_vbflash_status(struct device *dev, } static const struct bin_attribute psp_vbflash_bin_attr = { - .attr = {.name = "psp_vbflash", .mode = 0664}, + .attr = {.name = "psp_vbflash", .mode = 0660}, .size = 0, .write = amdgpu_psp_vbflash_write, .read = amdgpu_psp_vbflash_read, }; -static DEVICE_ATTR(psp_vbflash_status, 0444, amdgpu_psp_vbflash_status, NULL); +static DEVICE_ATTR(psp_vbflash_status, 0440, amdgpu_psp_vbflash_status, NULL); int amdgpu_psp_sysfs_init(struct amdgpu_device *adev) { @@ -3618,6 +3711,7 @@ int amdgpu_psp_sysfs_init(struct amdgpu_device *adev) switch (adev->ip_versions[MP0_HWIP][0]) { case IP_VERSION(13, 0, 0): case IP_VERSION(13, 0, 7): + case IP_VERSION(13, 0, 10): if (!psp->adev) { psp->adev = adev; psp_v13_0_set_psp_funcs(psp); @@ -3673,8 +3767,7 @@ static void psp_sysfs_fini(struct amdgpu_device *adev) device_remove_file(adev->dev, &dev_attr_usbc_pd_fw); } -const struct amdgpu_ip_block_version psp_v3_1_ip_block = -{ +const struct amdgpu_ip_block_version psp_v3_1_ip_block = { .type = AMD_IP_BLOCK_TYPE_PSP, .major = 3, .minor = 1, @@ -3682,8 +3775,7 @@ const struct amdgpu_ip_block_version psp_v3_1_ip_block = .funcs = &psp_ip_funcs, }; -const struct amdgpu_ip_block_version psp_v10_0_ip_block = -{ +const struct amdgpu_ip_block_version psp_v10_0_ip_block = { .type = AMD_IP_BLOCK_TYPE_PSP, .major = 10, .minor = 0, @@ -3691,8 +3783,7 @@ const struct amdgpu_ip_block_version psp_v10_0_ip_block = .funcs = &psp_ip_funcs, }; -const struct amdgpu_ip_block_version psp_v11_0_ip_block = -{ +const struct amdgpu_ip_block_version psp_v11_0_ip_block = { .type = AMD_IP_BLOCK_TYPE_PSP, .major = 11, .minor = 0, @@ -3708,8 +3799,7 @@ const struct amdgpu_ip_block_version psp_v11_0_8_ip_block = { .funcs = &psp_ip_funcs, }; -const struct amdgpu_ip_block_version psp_v12_0_ip_block = -{ +const struct amdgpu_ip_block_version psp_v12_0_ip_block = { .type = AMD_IP_BLOCK_TYPE_PSP, .major = 12, .minor = 0, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h index cf4f60c66122..2cae0b1a0b8a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h @@ -455,6 +455,8 @@ extern const struct amdgpu_ip_block_version psp_v13_0_4_ip_block; extern int psp_wait_for(struct psp_context *psp, uint32_t reg_index, uint32_t field_val, uint32_t mask, bool check_changed); +extern int psp_wait_for_spirom_update(struct psp_context *psp, uint32_t reg_index, + uint32_t field_val, uint32_t mask, uint32_t msec_timeout); int psp_gpu_reset(struct amdgpu_device *adev); int psp_update_vcn_sram(struct amdgpu_device *adev, int inst_idx, @@ -486,7 +488,7 @@ int psp_ras_invoke(struct psp_context *psp, uint32_t ta_cmd_id); int psp_ras_enable_features(struct psp_context *psp, union ta_ras_cmd_input *info, bool enable); int psp_ras_trigger_error(struct psp_context *psp, - struct ta_ras_trigger_error_input *info); + struct ta_ras_trigger_error_input *info, uint32_t instance_mask); int psp_ras_terminate(struct psp_context *psp); int psp_hdcp_invoke(struct psp_context *psp, uint32_t ta_cmd_id); @@ -519,6 +521,8 @@ int psp_load_fw_list(struct psp_context *psp, struct amdgpu_firmware_info **ucode_list, int ucode_count); void psp_copy_fw(struct psp_context *psp, uint8_t *start_addr, uint32_t bin_size); +int psp_spatial_partition(struct psp_context *psp, int mode); + int is_psp_fw_valid(struct psp_bin_desc bin); int amdgpu_psp_sysfs_init(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 3ab8a88789c8..4769a18304d7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -171,8 +171,7 @@ static int amdgpu_reserve_page_direct(struct amdgpu_device *adev, uint64_t addre memset(&err_rec, 0x0, sizeof(struct eeprom_table_record)); err_data.err_addr = &err_rec; - amdgpu_umc_fill_error_record(&err_data, address, - (address >> AMDGPU_GPU_PAGE_SHIFT), 0, 0); + amdgpu_umc_fill_error_record(&err_data, address, address, 0, 0); if (amdgpu_bad_page_threshold != 0) { amdgpu_ras_add_bad_pages(adev, err_data.err_addr, @@ -256,6 +255,8 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, int block_id; uint32_t sub_block; u64 address, value; + /* default value is 0 if the mask is not set by user */ + u32 instance_mask = 0; if (*pos) return -EINVAL; @@ -306,7 +307,11 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, data->op = op; if (op == 2) { - if (sscanf(str, "%*s %*s %*s 0x%x 0x%llx 0x%llx", + if (sscanf(str, "%*s %*s %*s 0x%x 0x%llx 0x%llx 0x%x", + &sub_block, &address, &value, &instance_mask) != 4 && + sscanf(str, "%*s %*s %*s %u %llu %llu %u", + &sub_block, &address, &value, &instance_mask) != 4 && + sscanf(str, "%*s %*s %*s 0x%x 0x%llx 0x%llx", &sub_block, &address, &value) != 3 && sscanf(str, "%*s %*s %*s %u %llu %llu", &sub_block, &address, &value) != 3) @@ -314,6 +319,7 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, data->head.sub_block_index = sub_block; data->inject.address = address; data->inject.value = value; + data->inject.instance_mask = instance_mask; } } else { if (size < sizeof(*data)) @@ -326,6 +332,46 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, return 0; } +static void amdgpu_ras_instance_mask_check(struct amdgpu_device *adev, + struct ras_debug_if *data) +{ + int num_xcc = adev->gfx.xcc_mask ? NUM_XCC(adev->gfx.xcc_mask) : 1; + uint32_t mask, inst_mask = data->inject.instance_mask; + + /* no need to set instance mask if there is only one instance */ + if (num_xcc <= 1 && inst_mask) { + data->inject.instance_mask = 0; + dev_dbg(adev->dev, + "RAS inject mask(0x%x) isn't supported and force it to 0.\n", + inst_mask); + + return; + } + + switch (data->head.block) { + case AMDGPU_RAS_BLOCK__GFX: + mask = GENMASK(num_xcc - 1, 0); + break; + case AMDGPU_RAS_BLOCK__SDMA: + mask = GENMASK(adev->sdma.num_instances - 1, 0); + break; + case AMDGPU_RAS_BLOCK__VCN: + case AMDGPU_RAS_BLOCK__JPEG: + mask = GENMASK(adev->vcn.num_vcn_inst - 1, 0); + break; + default: + mask = inst_mask; + break; + } + + /* remove invalid bits in instance mask */ + data->inject.instance_mask &= mask; + if (inst_mask != data->inject.instance_mask) + dev_dbg(adev->dev, + "Adjust RAS inject mask 0x%x to 0x%x\n", + inst_mask, data->inject.instance_mask); +} + /** * DOC: AMDGPU RAS debugfs control interface * @@ -341,7 +387,7 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, * sub_block_index: some IPs have subcomponets. say, GFX, sDMA. * name: the name of IP. * - * inject has two more members than head, they are address, value. + * inject has three more members than head, they are address, value and mask. * As their names indicate, inject operation will write the * value to the address. * @@ -365,7 +411,7 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, * * echo "disable <block>" > /sys/kernel/debug/dri/<N>/ras/ras_ctrl * echo "enable <block> <error>" > /sys/kernel/debug/dri/<N>/ras/ras_ctrl - * echo "inject <block> <error> <sub-block> <address> <value> > /sys/kernel/debug/dri/<N>/ras/ras_ctrl + * echo "inject <block> <error> <sub-block> <address> <value> <mask>" > /sys/kernel/debug/dri/<N>/ras/ras_ctrl * * Where N, is the card which you want to affect. * @@ -382,13 +428,14 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, * * The sub-block is a the sub-block index, pass 0 if there is no sub-block. * The address and value are hexadecimal numbers, leading 0x is optional. + * The mask means instance mask, is optional, default value is 0x1. * * For instance, * * .. code-block:: bash * * echo inject umc ue 0x0 0x0 0x0 > /sys/kernel/debug/dri/0/ras/ras_ctrl - * echo inject umc ce 0 0 0 > /sys/kernel/debug/dri/0/ras/ras_ctrl + * echo inject umc ce 0 0 0 3 > /sys/kernel/debug/dri/0/ras/ras_ctrl * echo disable umc > /sys/kernel/debug/dri/0/ras/ras_ctrl * * How to check the result of the operation? @@ -442,7 +489,8 @@ static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f, ret = amdgpu_ras_feature_enable(adev, &data.head, 1); break; case 2: - if ((data.inject.address >= adev->gmc.mc_vram_size) || + if ((data.inject.address >= adev->gmc.mc_vram_size && + adev->gmc.mc_vram_size) || (data.inject.address >= RAS_UMC_INJECT_ADDR_LIMIT)) { dev_warn(adev->dev, "RAS WARN: input address " "0x%llx is invalid.", @@ -460,6 +508,8 @@ static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f, break; } + amdgpu_ras_instance_mask_check(adev, &data); + /* data.inject.address is offset instead of absolute gpu address */ ret = amdgpu_ras_error_inject(adev, &data.inject); break; @@ -1115,15 +1165,15 @@ int amdgpu_ras_error_inject(struct amdgpu_device *adev, block_info.address); } - if (info->head.block == AMDGPU_RAS_BLOCK__GFX) { - if (block_obj->hw_ops->ras_error_inject) - ret = block_obj->hw_ops->ras_error_inject(adev, info); + if (block_obj->hw_ops->ras_error_inject) { + if (info->head.block == AMDGPU_RAS_BLOCK__GFX) + ret = block_obj->hw_ops->ras_error_inject(adev, info, info->instance_mask); + else /* Special ras_error_inject is defined (e.g: xgmi) */ + ret = block_obj->hw_ops->ras_error_inject(adev, &block_info, + info->instance_mask); } else { - /* If defined special ras_error_inject(e.g: xgmi), implement special ras_error_inject */ - if (block_obj->hw_ops->ras_error_inject) - ret = block_obj->hw_ops->ras_error_inject(adev, &block_info); - else /*If not defined .ras_error_inject, use default ras_error_inject*/ - ret = psp_ras_trigger_error(&adev->psp, &block_info); + /* default path */ + ret = psp_ras_trigger_error(&adev->psp, &block_info, info->instance_mask); } if (ret) @@ -1441,6 +1491,7 @@ static int amdgpu_ras_sysfs_remove_all(struct amdgpu_device *adev) static struct dentry *amdgpu_ras_debugfs_create_ctrl_node(struct amdgpu_device *adev) { struct amdgpu_ras *con = amdgpu_ras_get_context(adev); + struct amdgpu_ras_eeprom_control *eeprom = &con->eeprom_control; struct drm_minor *minor = adev_to_drm(adev)->primary; struct dentry *dir; @@ -1451,6 +1502,7 @@ static struct dentry *amdgpu_ras_debugfs_create_ctrl_node(struct amdgpu_device * &amdgpu_ras_debugfs_eeprom_ops); debugfs_create_u32("bad_page_cnt_threshold", 0444, dir, &con->bad_page_cnt_threshold); + debugfs_create_u32("ras_num_recs", 0444, dir, &eeprom->ras_num_recs); debugfs_create_x32("ras_hw_enabled", 0444, dir, &adev->ras_hw_enabled); debugfs_create_x32("ras_enabled", 0444, dir, &adev->ras_enabled); debugfs_create_file("ras_eeprom_size", S_IRUGO, dir, adev, @@ -1597,8 +1649,7 @@ static int amdgpu_ras_fs_fini(struct amdgpu_device *adev) void amdgpu_ras_interrupt_fatal_error_handler(struct amdgpu_device *adev) { /* Fatal error events are handled on host side */ - if (amdgpu_sriov_vf(adev) || - !amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__PCIE_BIF)) + if (amdgpu_sriov_vf(adev)) return; if (adev->nbio.ras && @@ -1636,8 +1687,7 @@ static void amdgpu_ras_interrupt_poison_consumption_handler(struct ras_manager * } } - if (!adev->gmc.xgmi.connected_to_cpu) - amdgpu_umc_poison_handler(adev, false); + amdgpu_umc_poison_handler(adev, false); if (block_obj->hw_ops && block_obj->hw_ops->handle_poison_consumption) poison_stat = block_obj->hw_ops->handle_poison_consumption(adev); @@ -2008,9 +2058,15 @@ static void amdgpu_ras_do_recovery(struct work_struct *work) /* Perform full reset in fatal error mode */ if (!amdgpu_ras_is_poison_mode_supported(ras->adev)) set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); - else + else { clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); + if (ras->gpu_reset_flags & AMDGPU_RAS_GPU_RESET_MODE2_RESET) { + ras->gpu_reset_flags &= ~AMDGPU_RAS_GPU_RESET_MODE2_RESET; + reset_context.method = AMD_RESET_METHOD_MODE2; + } + } + amdgpu_device_gpu_recover(ras->adev, NULL, &reset_context); } atomic_set(&ras->in_recovery, 0); @@ -2259,7 +2315,7 @@ int amdgpu_ras_recovery_init(struct amdgpu_device *adev) atomic_set(&con->in_recovery, 0); con->eeprom_control.bad_channel_bitmap = 0; - max_eeprom_records_count = amdgpu_ras_eeprom_max_record_count(); + max_eeprom_records_count = amdgpu_ras_eeprom_max_record_count(&con->eeprom_control); amdgpu_ras_validate_threshold(adev, max_eeprom_records_count); /* Todo: During test the SMU might fail to read the eeprom through I2C @@ -2396,11 +2452,10 @@ static void amdgpu_ras_check_supported(struct amdgpu_device *adev) { adev->ras_hw_enabled = adev->ras_enabled = 0; - if (!adev->is_atom_fw || - !amdgpu_ras_asic_supported(adev)) + if (!amdgpu_ras_asic_supported(adev)) return; - if (!adev->gmc.xgmi.connected_to_cpu) { + if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu) { if (amdgpu_atomfirmware_mem_ecc_supported(adev)) { dev_info(adev->dev, "MEM ECC is active.\n"); adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__UMC | @@ -2625,7 +2680,8 @@ release_con: int amdgpu_persistent_edc_harvesting_supported(struct amdgpu_device *adev) { - if (adev->gmc.xgmi.connected_to_cpu) + if (adev->gmc.xgmi.connected_to_cpu || + adev->gmc.is_app_apu) return 1; return 0; } @@ -3104,3 +3160,143 @@ int amdgpu_ras_register_ras_block(struct amdgpu_device *adev, return 0; } + +void amdgpu_ras_get_error_type_name(uint32_t err_type, char *err_type_name) +{ + if (!err_type_name) + return; + + switch (err_type) { + case AMDGPU_RAS_ERROR__SINGLE_CORRECTABLE: + sprintf(err_type_name, "correctable"); + break; + case AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE: + sprintf(err_type_name, "uncorrectable"); + break; + default: + sprintf(err_type_name, "unknown"); + break; + } +} + +bool amdgpu_ras_inst_get_memory_id_field(struct amdgpu_device *adev, + const struct amdgpu_ras_err_status_reg_entry *reg_entry, + uint32_t instance, + uint32_t *memory_id) +{ + uint32_t err_status_lo_data, err_status_lo_offset; + + if (!reg_entry) + return false; + + err_status_lo_offset = + AMDGPU_RAS_REG_ENTRY_OFFSET(reg_entry->hwip, instance, + reg_entry->seg_lo, reg_entry->reg_lo); + err_status_lo_data = RREG32(err_status_lo_offset); + + if ((reg_entry->flags & AMDGPU_RAS_ERR_STATUS_VALID) && + !REG_GET_FIELD(err_status_lo_data, ERR_STATUS_LO, ERR_STATUS_VALID_FLAG)) + return false; + + *memory_id = REG_GET_FIELD(err_status_lo_data, ERR_STATUS_LO, MEMORY_ID); + + return true; +} + +bool amdgpu_ras_inst_get_err_cnt_field(struct amdgpu_device *adev, + const struct amdgpu_ras_err_status_reg_entry *reg_entry, + uint32_t instance, + unsigned long *err_cnt) +{ + uint32_t err_status_hi_data, err_status_hi_offset; + + if (!reg_entry) + return false; + + err_status_hi_offset = + AMDGPU_RAS_REG_ENTRY_OFFSET(reg_entry->hwip, instance, + reg_entry->seg_hi, reg_entry->reg_hi); + err_status_hi_data = RREG32(err_status_hi_offset); + + if ((reg_entry->flags & AMDGPU_RAS_ERR_INFO_VALID) && + !REG_GET_FIELD(err_status_hi_data, ERR_STATUS_HI, ERR_INFO_VALID_FLAG)) + /* keep the check here in case we need to refer to the result later */ + dev_dbg(adev->dev, "Invalid err_info field\n"); + + /* read err count */ + *err_cnt = REG_GET_FIELD(err_status_hi_data, ERR_STATUS, ERR_CNT); + + return true; +} + +void amdgpu_ras_inst_query_ras_error_count(struct amdgpu_device *adev, + const struct amdgpu_ras_err_status_reg_entry *reg_list, + uint32_t reg_list_size, + const struct amdgpu_ras_memory_id_entry *mem_list, + uint32_t mem_list_size, + uint32_t instance, + uint32_t err_type, + unsigned long *err_count) +{ + uint32_t memory_id; + unsigned long err_cnt; + char err_type_name[16]; + uint32_t i, j; + + for (i = 0; i < reg_list_size; i++) { + /* query memory_id from err_status_lo */ + if (!amdgpu_ras_inst_get_memory_id_field(adev, ®_list[i], + instance, &memory_id)) + continue; + + /* query err_cnt from err_status_hi */ + if (!amdgpu_ras_inst_get_err_cnt_field(adev, ®_list[i], + instance, &err_cnt) || + !err_cnt) + continue; + + *err_count += err_cnt; + + /* log the errors */ + amdgpu_ras_get_error_type_name(err_type, err_type_name); + if (!mem_list) { + /* memory_list is not supported */ + dev_info(adev->dev, + "%ld %s hardware errors detected in %s, instance: %d, memory_id: %d\n", + err_cnt, err_type_name, + reg_list[i].block_name, + instance, memory_id); + } else { + for (j = 0; j < mem_list_size; j++) { + if (memory_id == mem_list[j].memory_id) { + dev_info(adev->dev, + "%ld %s hardware errors detected in %s, instance: %d, memory block: %s\n", + err_cnt, err_type_name, + reg_list[i].block_name, + instance, mem_list[j].name); + break; + } + } + } + } +} + +void amdgpu_ras_inst_reset_ras_error_count(struct amdgpu_device *adev, + const struct amdgpu_ras_err_status_reg_entry *reg_list, + uint32_t reg_list_size, + uint32_t instance) +{ + uint32_t err_status_lo_offset, err_status_hi_offset; + uint32_t i; + + for (i = 0; i < reg_list_size; i++) { + err_status_lo_offset = + AMDGPU_RAS_REG_ENTRY_OFFSET(reg_list[i].hwip, instance, + reg_list[i].seg_lo, reg_list[i].reg_lo); + err_status_hi_offset = + AMDGPU_RAS_REG_ENTRY_OFFSET(reg_list[i].hwip, instance, + reg_list[i].seg_hi, reg_list[i].reg_hi); + WREG32(err_status_lo_offset, 0); + WREG32(err_status_hi_offset, 0); + } +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h index 17b3d1992e80..46bf1889a9d7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h @@ -32,6 +32,11 @@ struct amdgpu_iv_entry; #define AMDGPU_RAS_FLAG_INIT_BY_VBIOS (0x1 << 0) +/* position of instance value in sub_block_index of + * ta_ras_trigger_error_input, the sub block uses lower 12 bits + */ +#define AMDGPU_RAS_INST_MASK 0xfffff000 +#define AMDGPU_RAS_INST_SHIFT 0xc enum amdgpu_ras_block { AMDGPU_RAS_BLOCK__UMC = 0, @@ -314,6 +319,45 @@ enum amdgpu_ras_ret { AMDGPU_RAS_PT, }; +/* ras error status reisger fields */ +#define ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +#define ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define ERR_STATUS__ERR_CNT__SHIFT 0x17 +#define ERR_STATUS__ERR_CNT_MASK 0x03800000L + +#define AMDGPU_RAS_REG_ENTRY(ip, inst, reg_lo, reg_hi) \ + ip##_HWIP, inst, reg_lo##_BASE_IDX, reg_lo, reg_hi##_BASE_IDX, reg_hi + +#define AMDGPU_RAS_REG_ENTRY_OFFSET(hwip, ip_inst, segment, reg) \ + (adev->reg_offset[hwip][ip_inst][segment] + (reg)) + +#define AMDGPU_RAS_ERR_INFO_VALID (1 << 0) +#define AMDGPU_RAS_ERR_STATUS_VALID (1 << 1) +#define AMDGPU_RAS_ERR_ADDRESS_VALID (1 << 2) + +#define AMDGPU_RAS_GPU_RESET_MODE2_RESET (0x1 << 0) + +struct amdgpu_ras_err_status_reg_entry { + uint32_t hwip; + uint32_t ip_inst; + uint32_t seg_lo; + uint32_t reg_lo; + uint32_t seg_hi; + uint32_t reg_hi; + uint32_t reg_inst; + uint32_t flags; + const char *block_name; +}; + +struct amdgpu_ras_memory_id_entry { + uint32_t memory_id; + const char *name; +}; + struct ras_common_if { enum amdgpu_ras_block block; enum amdgpu_ras_error_type type; @@ -385,6 +429,9 @@ struct amdgpu_ras { /* Indicates smu whether need update bad channel info */ bool update_channel_flag; + + /* Record special requirements of gpu reset caller */ + uint32_t gpu_reset_flags; }; struct ras_fs_data { @@ -471,6 +518,7 @@ struct ras_inject_if { struct ras_common_if head; uint64_t address; uint64_t value; + uint32_t instance_mask; }; struct ras_cure_if { @@ -508,7 +556,8 @@ struct amdgpu_ras_block_object { }; struct amdgpu_ras_block_hw_ops { - int (*ras_error_inject)(struct amdgpu_device *adev, void *inject_if); + int (*ras_error_inject)(struct amdgpu_device *adev, + void *inject_if, uint32_t instance_mask); void (*query_ras_error_count)(struct amdgpu_device *adev, void *ras_error_status); void (*query_ras_error_status)(struct amdgpu_device *adev); void (*query_ras_error_address)(struct amdgpu_device *adev, void *ras_error_status); @@ -696,4 +745,25 @@ int amdgpu_ras_set_context(struct amdgpu_device *adev, struct amdgpu_ras *ras_co int amdgpu_ras_register_ras_block(struct amdgpu_device *adev, struct amdgpu_ras_block_object *ras_block_obj); void amdgpu_ras_interrupt_fatal_error_handler(struct amdgpu_device *adev); +void amdgpu_ras_get_error_type_name(uint32_t err_type, char *err_type_name); +bool amdgpu_ras_inst_get_memory_id_field(struct amdgpu_device *adev, + const struct amdgpu_ras_err_status_reg_entry *reg_entry, + uint32_t instance, + uint32_t *memory_id); +bool amdgpu_ras_inst_get_err_cnt_field(struct amdgpu_device *adev, + const struct amdgpu_ras_err_status_reg_entry *reg_entry, + uint32_t instance, + unsigned long *err_cnt); +void amdgpu_ras_inst_query_ras_error_count(struct amdgpu_device *adev, + const struct amdgpu_ras_err_status_reg_entry *reg_list, + uint32_t reg_list_size, + const struct amdgpu_ras_memory_id_entry *mem_list, + uint32_t mem_list_size, + uint32_t instance, + uint32_t err_type, + unsigned long *err_count); +void amdgpu_ras_inst_reset_ras_error_count(struct amdgpu_device *adev, + const struct amdgpu_ras_err_status_reg_entry *reg_list, + uint32_t reg_list_size, + uint32_t instance); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c index c2c2a7718613..0648dfe559af 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c @@ -68,11 +68,24 @@ /* Table hdr is 'AMDR' */ #define RAS_TABLE_HDR_VAL 0x414d4452 -#define RAS_TABLE_VER 0x00010000 /* Bad GPU tag ‘BADG’ */ #define RAS_TABLE_HDR_BAD 0x42414447 +/* + * EEPROM Table structure v1 + * --------------------------------- + * | | + * | EEPROM TABLE HEADER | + * | ( size 20 Bytes ) | + * | | + * --------------------------------- + * | | + * | BAD PAGE RECORD AREA | + * | | + * --------------------------------- + */ + /* Assume 2-Mbit size EEPROM and take up the whole space. */ #define RAS_TBL_SIZE_BYTES (256 * 1024) #define RAS_TABLE_START 0 @@ -81,6 +94,35 @@ #define RAS_MAX_RECORD_COUNT ((RAS_TBL_SIZE_BYTES - RAS_TABLE_HEADER_SIZE) \ / RAS_TABLE_RECORD_SIZE) +/* + * EEPROM Table structrue v2.1 + * --------------------------------- + * | | + * | EEPROM TABLE HEADER | + * | ( size 20 Bytes ) | + * | | + * --------------------------------- + * | | + * | EEPROM TABLE RAS INFO | + * | (available info size 4 Bytes) | + * | ( reserved size 252 Bytes ) | + * | | + * --------------------------------- + * | | + * | BAD PAGE RECORD AREA | + * | | + * --------------------------------- + */ + +/* EEPROM Table V2_1 */ +#define RAS_TABLE_V2_1_INFO_SIZE 256 +#define RAS_TABLE_V2_1_INFO_START RAS_TABLE_HEADER_SIZE +#define RAS_RECORD_START_V2_1 (RAS_HDR_START + RAS_TABLE_HEADER_SIZE + \ + RAS_TABLE_V2_1_INFO_SIZE) +#define RAS_MAX_RECORD_COUNT_V2_1 ((RAS_TBL_SIZE_BYTES - RAS_TABLE_HEADER_SIZE - \ + RAS_TABLE_V2_1_INFO_SIZE) \ + / RAS_TABLE_RECORD_SIZE) + /* Given a zero-based index of an EEPROM RAS record, yields the EEPROM * offset off of RAS_TABLE_START. That is, this is something you can * add to control->i2c_address, and then tell I2C layer to read @@ -103,6 +145,10 @@ #define RAS_NUM_RECS(_tbl_hdr) (((_tbl_hdr)->tbl_size - \ RAS_TABLE_HEADER_SIZE) / RAS_TABLE_RECORD_SIZE) +#define RAS_NUM_RECS_V2_1(_tbl_hdr) (((_tbl_hdr)->tbl_size - \ + RAS_TABLE_HEADER_SIZE - \ + RAS_TABLE_V2_1_INFO_SIZE) / RAS_TABLE_RECORD_SIZE) + #define to_amdgpu_device(x) (container_of(x, struct amdgpu_ras, eeprom_control))->adev static bool __is_ras_eeprom_supported(struct amdgpu_device *adev) @@ -230,6 +276,69 @@ static int __write_table_header(struct amdgpu_ras_eeprom_control *control) return res; } +static void +__encode_table_ras_info_to_buf(struct amdgpu_ras_eeprom_table_ras_info *rai, + unsigned char *buf) +{ + u32 *pp = (uint32_t *)buf; + u32 tmp; + + tmp = ((uint32_t)(rai->rma_status) & 0xFF) | + (((uint32_t)(rai->health_percent) << 8) & 0xFF00) | + (((uint32_t)(rai->ecc_page_threshold) << 16) & 0xFFFF0000); + pp[0] = cpu_to_le32(tmp); +} + +static void +__decode_table_ras_info_from_buf(struct amdgpu_ras_eeprom_table_ras_info *rai, + unsigned char *buf) +{ + u32 *pp = (uint32_t *)buf; + u32 tmp; + + tmp = le32_to_cpu(pp[0]); + rai->rma_status = tmp & 0xFF; + rai->health_percent = (tmp >> 8) & 0xFF; + rai->ecc_page_threshold = (tmp >> 16) & 0xFFFF; +} + +static int __write_table_ras_info(struct amdgpu_ras_eeprom_control *control) +{ + struct amdgpu_device *adev = to_amdgpu_device(control); + u8 *buf; + int res; + + buf = kzalloc(RAS_TABLE_V2_1_INFO_SIZE, GFP_KERNEL); + if (!buf) { + DRM_ERROR("Failed to alloc buf to write table ras info\n"); + return -ENOMEM; + } + + __encode_table_ras_info_to_buf(&control->tbl_rai, buf); + + /* i2c may be unstable in gpu reset */ + down_read(&adev->reset_domain->sem); + res = amdgpu_eeprom_write(adev->pm.ras_eeprom_i2c_bus, + control->i2c_address + + control->ras_info_offset, + buf, RAS_TABLE_V2_1_INFO_SIZE); + up_read(&adev->reset_domain->sem); + + if (res < 0) { + DRM_ERROR("Failed to write EEPROM table ras info:%d", res); + } else if (res < RAS_TABLE_V2_1_INFO_SIZE) { + DRM_ERROR("Short write:%d out of %d\n", + res, RAS_TABLE_V2_1_INFO_SIZE); + res = -EIO; + } else { + res = 0; + } + + kfree(buf); + + return res; +} + static u8 __calc_hdr_byte_sum(const struct amdgpu_ras_eeprom_control *control) { int ii; @@ -246,6 +355,21 @@ static u8 __calc_hdr_byte_sum(const struct amdgpu_ras_eeprom_control *control) return csum; } +static u8 __calc_ras_info_byte_sum(const struct amdgpu_ras_eeprom_control *control) +{ + int ii; + u8 *pp, csum; + size_t sz; + + sz = sizeof(control->tbl_rai); + pp = (u8 *) &control->tbl_rai; + csum = 0; + for (ii = 0; ii < sz; ii++, pp++) + csum += *pp; + + return csum; +} + static int amdgpu_ras_eeprom_correct_header_tag( struct amdgpu_ras_eeprom_control *control, uint32_t header) @@ -282,6 +406,7 @@ int amdgpu_ras_eeprom_reset_table(struct amdgpu_ras_eeprom_control *control) { struct amdgpu_device *adev = to_amdgpu_device(control); struct amdgpu_ras_eeprom_table_header *hdr = &control->tbl_hdr; + struct amdgpu_ras_eeprom_table_ras_info *rai = &control->tbl_rai; struct amdgpu_ras *con = amdgpu_ras_get_context(adev); u8 csum; int res; @@ -289,14 +414,37 @@ int amdgpu_ras_eeprom_reset_table(struct amdgpu_ras_eeprom_control *control) mutex_lock(&control->ras_tbl_mutex); hdr->header = RAS_TABLE_HDR_VAL; - hdr->version = RAS_TABLE_VER; - hdr->first_rec_offset = RAS_RECORD_START; - hdr->tbl_size = RAS_TABLE_HEADER_SIZE; + if (adev->umc.ras && + adev->umc.ras->set_eeprom_table_version) + adev->umc.ras->set_eeprom_table_version(hdr); + else + hdr->version = RAS_TABLE_VER_V1; + + if (hdr->version == RAS_TABLE_VER_V2_1) { + hdr->first_rec_offset = RAS_RECORD_START_V2_1; + hdr->tbl_size = RAS_TABLE_HEADER_SIZE + + RAS_TABLE_V2_1_INFO_SIZE; + rai->rma_status = GPU_HEALTH_USABLE; + /** + * GPU health represented as a percentage. + * 0 means worst health, 100 means fully health. + */ + rai->health_percent = 100; + /* ecc_page_threshold = 0 means disable bad page retirement */ + rai->ecc_page_threshold = con->bad_page_cnt_threshold; + } else { + hdr->first_rec_offset = RAS_RECORD_START; + hdr->tbl_size = RAS_TABLE_HEADER_SIZE; + } csum = __calc_hdr_byte_sum(control); + if (hdr->version == RAS_TABLE_VER_V2_1) + csum += __calc_ras_info_byte_sum(control); csum = -csum; hdr->checksum = csum; res = __write_table_header(control); + if (!res && hdr->version > RAS_TABLE_VER_V1) + res = __write_table_ras_info(control); control->ras_num_recs = 0; control->ras_fri = 0; @@ -573,11 +721,19 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control) "Saved bad pages %d reaches threshold value %d\n", control->ras_num_recs, ras->bad_page_cnt_threshold); control->tbl_hdr.header = RAS_TABLE_HDR_BAD; + if (control->tbl_hdr.version == RAS_TABLE_VER_V2_1) { + control->tbl_rai.rma_status = GPU_RETIRED__ECC_REACH_THRESHOLD; + control->tbl_rai.health_percent = 0; + } } - control->tbl_hdr.version = RAS_TABLE_VER; - control->tbl_hdr.first_rec_offset = RAS_INDEX_TO_OFFSET(control, control->ras_fri); - control->tbl_hdr.tbl_size = RAS_TABLE_HEADER_SIZE + control->ras_num_recs * RAS_TABLE_RECORD_SIZE; + if (control->tbl_hdr.version == RAS_TABLE_VER_V2_1) + control->tbl_hdr.tbl_size = RAS_TABLE_HEADER_SIZE + + RAS_TABLE_V2_1_INFO_SIZE + + control->ras_num_recs * RAS_TABLE_RECORD_SIZE; + else + control->tbl_hdr.tbl_size = RAS_TABLE_HEADER_SIZE + + control->ras_num_recs * RAS_TABLE_RECORD_SIZE; control->tbl_hdr.checksum = 0; buf_size = control->ras_num_recs * RAS_TABLE_RECORD_SIZE; @@ -606,6 +762,17 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control) goto Out; } + /** + * bad page records have been stored in eeprom, + * now calculate gpu health percent + */ + if (amdgpu_bad_page_threshold != 0 && + control->tbl_hdr.version == RAS_TABLE_VER_V2_1 && + control->ras_num_recs < ras->bad_page_cnt_threshold) + control->tbl_rai.health_percent = ((ras->bad_page_cnt_threshold - + control->ras_num_recs) * 100) / + ras->bad_page_cnt_threshold; + /* Recalc the checksum. */ csum = 0; @@ -613,10 +780,14 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control) csum += *pp; csum += __calc_hdr_byte_sum(control); + if (control->tbl_hdr.version == RAS_TABLE_VER_V2_1) + csum += __calc_ras_info_byte_sum(control); /* avoid sign extension when assigning to "checksum" */ csum = -csum; control->tbl_hdr.checksum = csum; res = __write_table_header(control); + if (!res && control->tbl_hdr.version > RAS_TABLE_VER_V1) + res = __write_table_ras_info(control); Out: kfree(buf); return res; @@ -807,9 +978,12 @@ Out: return res; } -uint32_t amdgpu_ras_eeprom_max_record_count(void) +uint32_t amdgpu_ras_eeprom_max_record_count(struct amdgpu_ras_eeprom_control *control) { - return RAS_MAX_RECORD_COUNT; + if (control->tbl_hdr.version == RAS_TABLE_VER_V2_1) + return RAS_MAX_RECORD_COUNT_V2_1; + else + return RAS_MAX_RECORD_COUNT; } static ssize_t @@ -1051,8 +1225,14 @@ static int __verify_ras_table_checksum(struct amdgpu_ras_eeprom_control *control int buf_size, res; u8 csum, *buf, *pp; - buf_size = RAS_TABLE_HEADER_SIZE + - control->ras_num_recs * RAS_TABLE_RECORD_SIZE; + if (control->tbl_hdr.version == RAS_TABLE_VER_V2_1) + buf_size = RAS_TABLE_HEADER_SIZE + + RAS_TABLE_V2_1_INFO_SIZE + + control->ras_num_recs * RAS_TABLE_RECORD_SIZE; + else + buf_size = RAS_TABLE_HEADER_SIZE + + control->ras_num_recs * RAS_TABLE_RECORD_SIZE; + buf = kzalloc(buf_size, GFP_KERNEL); if (!buf) { DRM_ERROR("Out of memory checking RAS table checksum.\n"); @@ -1080,6 +1260,39 @@ Out: return res < 0 ? res : csum; } +static int __read_table_ras_info(struct amdgpu_ras_eeprom_control *control) +{ + struct amdgpu_ras_eeprom_table_ras_info *rai = &control->tbl_rai; + struct amdgpu_device *adev = to_amdgpu_device(control); + unsigned char *buf; + int res; + + buf = kzalloc(RAS_TABLE_V2_1_INFO_SIZE, GFP_KERNEL); + if (!buf) { + DRM_ERROR("Failed to alloc buf to read EEPROM table ras info\n"); + return -ENOMEM; + } + + /** + * EEPROM table V2_1 supports ras info, + * read EEPROM table ras info + */ + res = amdgpu_eeprom_read(adev->pm.ras_eeprom_i2c_bus, + control->i2c_address + control->ras_info_offset, + buf, RAS_TABLE_V2_1_INFO_SIZE); + if (res < RAS_TABLE_V2_1_INFO_SIZE) { + DRM_ERROR("Failed to read EEPROM table ras info, res:%d", res); + res = res >= 0 ? -EIO : res; + goto Out; + } + + __decode_table_ras_info_from_buf(rai, buf); + +Out: + kfree(buf); + return res == RAS_TABLE_V2_1_INFO_SIZE ? 0 : res; +} + int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control, bool *exceed_err_limit) { @@ -1102,8 +1315,7 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control, return -EINVAL; control->ras_header_offset = RAS_HDR_START; - control->ras_record_offset = RAS_RECORD_START; - control->ras_max_record_count = RAS_MAX_RECORD_COUNT; + control->ras_info_offset = RAS_TABLE_V2_1_INFO_START; mutex_init(&control->ras_tbl_mutex); /* Read the table header from EEPROM address */ @@ -1117,12 +1329,27 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control, __decode_table_header_from_buf(hdr, buf); - control->ras_num_recs = RAS_NUM_RECS(hdr); + if (hdr->version == RAS_TABLE_VER_V2_1) { + control->ras_num_recs = RAS_NUM_RECS_V2_1(hdr); + control->ras_record_offset = RAS_RECORD_START_V2_1; + control->ras_max_record_count = RAS_MAX_RECORD_COUNT_V2_1; + } else { + control->ras_num_recs = RAS_NUM_RECS(hdr); + control->ras_record_offset = RAS_RECORD_START; + control->ras_max_record_count = RAS_MAX_RECORD_COUNT; + } control->ras_fri = RAS_OFFSET_TO_INDEX(control, hdr->first_rec_offset); if (hdr->header == RAS_TABLE_HDR_VAL) { DRM_DEBUG_DRIVER("Found existing EEPROM table with %d records", control->ras_num_recs); + + if (hdr->version == RAS_TABLE_VER_V2_1) { + res = __read_table_ras_info(control); + if (res) + return res; + } + res = __verify_ras_table_checksum(control); if (res) DRM_ERROR("RAS table incorrect checksum or error:%d\n", @@ -1136,6 +1363,12 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control, ras->bad_page_cnt_threshold); } else if (hdr->header == RAS_TABLE_HDR_BAD && amdgpu_bad_page_threshold != 0) { + if (hdr->version == RAS_TABLE_VER_V2_1) { + res = __read_table_ras_info(control); + if (res) + return res; + } + res = __verify_ras_table_checksum(control); if (res) DRM_ERROR("RAS Table incorrect checksum or error:%d\n", diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h index 54d9bfe0881d..6dfd667f3013 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h @@ -26,8 +26,16 @@ #include <linux/i2c.h> +#define RAS_TABLE_VER_V1 0x00010000 +#define RAS_TABLE_VER_V2_1 0x00021000 + struct amdgpu_device; +enum amdgpu_ras_gpu_health_status { + GPU_HEALTH_USABLE = 0, + GPU_RETIRED__ECC_REACH_THRESHOLD = 2, +}; + enum amdgpu_ras_eeprom_err_type { AMDGPU_RAS_EEPROM_ERR_NA, AMDGPU_RAS_EEPROM_ERR_RECOVERABLE, @@ -43,9 +51,18 @@ struct amdgpu_ras_eeprom_table_header { uint32_t checksum; } __packed; +struct amdgpu_ras_eeprom_table_ras_info { + u8 rma_status; + u8 health_percent; + u16 ecc_page_threshold; + u32 padding[64 - 1]; +} __packed; + struct amdgpu_ras_eeprom_control { struct amdgpu_ras_eeprom_table_header tbl_hdr; + struct amdgpu_ras_eeprom_table_ras_info tbl_rai; + /* Base I2C EEPPROM 19-bit memory address, * where the table is located. For more information, * see top of amdgpu_eeprom.c. @@ -58,6 +75,7 @@ struct amdgpu_ras_eeprom_control { * right after the header. */ u32 ras_header_offset; + u32 ras_info_offset; u32 ras_record_offset; /* Number of records in the table. @@ -124,7 +142,7 @@ int amdgpu_ras_eeprom_read(struct amdgpu_ras_eeprom_control *control, int amdgpu_ras_eeprom_append(struct amdgpu_ras_eeprom_control *control, struct eeprom_table_record *records, const u32 num); -uint32_t amdgpu_ras_eeprom_max_record_count(void); +uint32_t amdgpu_ras_eeprom_max_record_count(struct amdgpu_ras_eeprom_control *control); void amdgpu_ras_debugfs_set_ret_size(struct amdgpu_ras_eeprom_control *control); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c index 6437ead87e5f..eec41ad30406 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c @@ -40,6 +40,7 @@ int amdgpu_reset_init(struct amdgpu_device *adev) switch (adev->ip_versions[MP1_HWIP][0]) { case IP_VERSION(13, 0, 2): + case IP_VERSION(13, 0, 6): ret = aldebaran_reset_init(adev); break; case IP_VERSION(11, 0, 7): @@ -61,6 +62,7 @@ int amdgpu_reset_fini(struct amdgpu_device *adev) switch (adev->ip_versions[MP1_HWIP][0]) { case IP_VERSION(13, 0, 2): + case IP_VERSION(13, 0, 6): ret = aldebaran_reset_fini(adev); break; case IP_VERSION(11, 0, 7): diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c index dc474b809604..80d6e132e409 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c @@ -50,6 +50,26 @@ */ /** + * amdgpu_ring_max_ibs - Return max IBs that fit in a single submission. + * + * @type: ring type for which to return the limit. + */ +unsigned int amdgpu_ring_max_ibs(enum amdgpu_ring_type type) +{ + switch (type) { + case AMDGPU_RING_TYPE_GFX: + /* Need to keep at least 192 on GFX7+ for old radv. */ + return 192; + case AMDGPU_RING_TYPE_COMPUTE: + return 125; + case AMDGPU_RING_TYPE_VCN_JPEG: + return 16; + default: + return 49; + } +} + +/** * amdgpu_ring_alloc - allocate space on the ring buffer * * @ring: amdgpu_ring structure holding ring information @@ -58,7 +78,7 @@ * Allocate @ndw dwords in the ring buffer (all asics). * Returns 0 on success, error on failure. */ -int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned ndw) +int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned int ndw) { /* Align requested size with padding so unlock_commit can * pad safely */ @@ -182,6 +202,7 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring, int sched_hw_submission = amdgpu_sched_hw_submission; u32 *num_sched; u32 hw_ip; + unsigned int max_ibs_dw; /* Set the hw submission limit higher for KIQ because * it's used for a number of gfx/compute tasks by both @@ -290,6 +311,13 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring, return r; } + max_ibs_dw = ring->funcs->emit_frame_size + + amdgpu_ring_max_ibs(ring->funcs->type) * ring->funcs->emit_ib_size; + max_ibs_dw = (max_ibs_dw + ring->funcs->align_mask) & ~ring->funcs->align_mask; + + if (WARN_ON(max_ibs_dw > max_dw)) + max_dw = max_ibs_dw; + ring->ring_size = roundup_pow_of_two(max_dw * 4 * sched_hw_submission); ring->buf_mask = (ring->ring_size / 4) - 1; @@ -361,6 +389,8 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring) amdgpu_bo_free_kernel(&ring->ring_obj, &ring->gpu_addr, (void **)&ring->ring); + } else { + kfree(ring->fence_drv.fences); } dma_fence_put(ring->vmid_wait); @@ -403,11 +433,18 @@ void amdgpu_ring_emit_reg_write_reg_wait_helper(struct amdgpu_ring *ring, bool amdgpu_ring_soft_recovery(struct amdgpu_ring *ring, unsigned int vmid, struct dma_fence *fence) { + unsigned long flags; + ktime_t deadline = ktime_add_us(ktime_get(), 10000); if (amdgpu_sriov_vf(ring->adev) || !ring->funcs->soft_recovery || !fence) return false; + spin_lock_irqsave(fence->lock, flags); + if (!dma_fence_is_signaled_locked(fence)) + dma_fence_set_error(fence, -ENODATA); + spin_unlock_irqrestore(fence->lock, flags); + atomic_inc(&ring->adev->gpu_reset_counter); while (!dma_fence_is_signaled(fence) && ktime_to_ns(ktime_sub(deadline, ktime_get())) > 0) @@ -478,6 +515,70 @@ static const struct file_operations amdgpu_debugfs_ring_fops = { .llseek = default_llseek }; +static ssize_t amdgpu_debugfs_mqd_read(struct file *f, char __user *buf, + size_t size, loff_t *pos) +{ + struct amdgpu_ring *ring = file_inode(f)->i_private; + volatile u32 *mqd; + int r; + uint32_t value, result; + + if (*pos & 3 || size & 3) + return -EINVAL; + + result = 0; + + r = amdgpu_bo_reserve(ring->mqd_obj, false); + if (unlikely(r != 0)) + return r; + + r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&mqd); + if (r) { + amdgpu_bo_unreserve(ring->mqd_obj); + return r; + } + + while (size) { + if (*pos >= ring->mqd_size) + goto done; + + value = mqd[*pos/4]; + r = put_user(value, (uint32_t *)buf); + if (r) + goto done; + buf += 4; + result += 4; + size -= 4; + *pos += 4; + } + +done: + amdgpu_bo_kunmap(ring->mqd_obj); + mqd = NULL; + amdgpu_bo_unreserve(ring->mqd_obj); + if (r) + return r; + + return result; +} + +static const struct file_operations amdgpu_debugfs_mqd_fops = { + .owner = THIS_MODULE, + .read = amdgpu_debugfs_mqd_read, + .llseek = default_llseek +}; + +static int amdgpu_debugfs_ring_error(void *data, u64 val) +{ + struct amdgpu_ring *ring = data; + + amdgpu_fence_driver_set_error(ring, val); + return 0; +} + +DEFINE_DEBUGFS_ATTRIBUTE_SIGNED(amdgpu_debugfs_error_fops, NULL, + amdgpu_debugfs_ring_error, "%lld\n"); + #endif void amdgpu_debugfs_ring_init(struct amdgpu_device *adev, @@ -489,10 +590,21 @@ void amdgpu_debugfs_ring_init(struct amdgpu_device *adev, char name[32]; sprintf(name, "amdgpu_ring_%s", ring->name); - debugfs_create_file_size(name, S_IFREG | S_IRUGO, root, ring, + debugfs_create_file_size(name, S_IFREG | 0444, root, ring, &amdgpu_debugfs_ring_fops, ring->ring_size + 12); + if (ring->mqd_obj) { + sprintf(name, "amdgpu_mqd_%s", ring->name); + debugfs_create_file_size(name, S_IFREG | 0444, root, ring, + &amdgpu_debugfs_mqd_fops, + ring->mqd_size); + } + + sprintf(name, "amdgpu_error_%s", ring->name); + debugfs_create_file(name, 0200, root, ring, + &amdgpu_debugfs_error_fops); + #endif } @@ -581,3 +693,21 @@ void amdgpu_ring_ib_end(struct amdgpu_ring *ring) if (ring->is_sw_ring) amdgpu_sw_ring_ib_end(ring); } + +void amdgpu_ring_ib_on_emit_cntl(struct amdgpu_ring *ring) +{ + if (ring->is_sw_ring) + amdgpu_sw_ring_ib_mark_offset(ring, AMDGPU_MUX_OFFSET_TYPE_CONTROL); +} + +void amdgpu_ring_ib_on_emit_ce(struct amdgpu_ring *ring) +{ + if (ring->is_sw_ring) + amdgpu_sw_ring_ib_mark_offset(ring, AMDGPU_MUX_OFFSET_TYPE_CE); +} + +void amdgpu_ring_ib_on_emit_de(struct amdgpu_ring *ring) +{ + if (ring->is_sw_ring) + amdgpu_sw_ring_ib_mark_offset(ring, AMDGPU_MUX_OFFSET_TYPE_DE); +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index d8749444b689..028ff075db51 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -37,8 +37,8 @@ struct amdgpu_job; struct amdgpu_vm; /* max number of rings */ -#define AMDGPU_MAX_RINGS 28 -#define AMDGPU_MAX_HWIP_RINGS 8 +#define AMDGPU_MAX_RINGS 124 +#define AMDGPU_MAX_HWIP_RINGS 64 #define AMDGPU_MAX_GFX_RINGS 2 #define AMDGPU_MAX_SW_GFX_RINGS 2 #define AMDGPU_MAX_COMPUTE_RINGS 8 @@ -126,6 +126,7 @@ struct amdgpu_fence_driver { extern const struct drm_sched_backend_ops amdgpu_sched_ops; void amdgpu_fence_driver_clear_job_fences(struct amdgpu_ring *ring); +void amdgpu_fence_driver_set_error(struct amdgpu_ring *ring, int error); void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring); int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring); @@ -212,6 +213,8 @@ struct amdgpu_ring_funcs { void (*end_use)(struct amdgpu_ring *ring); void (*emit_switch_buffer) (struct amdgpu_ring *ring); void (*emit_cntxcntl) (struct amdgpu_ring *ring, uint32_t flags); + void (*emit_gfx_shadow)(struct amdgpu_ring *ring, u64 shadow_va, u64 csa_va, + u64 gds_va, bool init_shadow, int vmid); void (*emit_rreg)(struct amdgpu_ring *ring, uint32_t reg, uint32_t reg_val_offs); void (*emit_wreg)(struct amdgpu_ring *ring, uint32_t reg, uint32_t val); @@ -227,6 +230,9 @@ struct amdgpu_ring_funcs { int (*preempt_ib)(struct amdgpu_ring *ring); void (*emit_mem_sync)(struct amdgpu_ring *ring); void (*emit_wave_limit)(struct amdgpu_ring *ring, bool enable); + void (*patch_cntl)(struct amdgpu_ring *ring, unsigned offset); + void (*patch_ce)(struct amdgpu_ring *ring, unsigned offset); + void (*patch_de)(struct amdgpu_ring *ring, unsigned offset); }; struct amdgpu_ring { @@ -250,12 +256,14 @@ struct amdgpu_ring { uint32_t buf_mask; u32 idx; u32 xcc_id; + u32 xcp_id; u32 me; u32 pipe; u32 queue; struct amdgpu_bo *mqd_obj; uint64_t mqd_gpu_addr; void *mqd_ptr; + unsigned mqd_size; uint64_t eop_gpu_addr; u32 doorbell_index; bool use_doorbell; @@ -309,6 +317,7 @@ struct amdgpu_ring { #define amdgpu_ring_emit_hdp_flush(r) (r)->funcs->emit_hdp_flush((r)) #define amdgpu_ring_emit_switch_buffer(r) (r)->funcs->emit_switch_buffer((r)) #define amdgpu_ring_emit_cntxcntl(r, d) (r)->funcs->emit_cntxcntl((r), (d)) +#define amdgpu_ring_emit_gfx_shadow(r, s, c, g, i, v) ((r)->funcs->emit_gfx_shadow((r), (s), (c), (g), (i), (v))) #define amdgpu_ring_emit_rreg(r, d, o) (r)->funcs->emit_rreg((r), (d), (o)) #define amdgpu_ring_emit_wreg(r, d, v) (r)->funcs->emit_wreg((r), (d), (v)) #define amdgpu_ring_emit_reg_wait(r, d, v, m) (r)->funcs->emit_reg_wait((r), (d), (v), (m)) @@ -318,10 +327,17 @@ struct amdgpu_ring { #define amdgpu_ring_init_cond_exec(r) (r)->funcs->init_cond_exec((r)) #define amdgpu_ring_patch_cond_exec(r,o) (r)->funcs->patch_cond_exec((r),(o)) #define amdgpu_ring_preempt_ib(r) (r)->funcs->preempt_ib(r) +#define amdgpu_ring_patch_cntl(r, o) ((r)->funcs->patch_cntl((r), (o))) +#define amdgpu_ring_patch_ce(r, o) ((r)->funcs->patch_ce((r), (o))) +#define amdgpu_ring_patch_de(r, o) ((r)->funcs->patch_de((r), (o))) +unsigned int amdgpu_ring_max_ibs(enum amdgpu_ring_type type); int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned ndw); void amdgpu_ring_ib_begin(struct amdgpu_ring *ring); void amdgpu_ring_ib_end(struct amdgpu_ring *ring); +void amdgpu_ring_ib_on_emit_cntl(struct amdgpu_ring *ring); +void amdgpu_ring_ib_on_emit_ce(struct amdgpu_ring *ring); +void amdgpu_ring_ib_on_emit_de(struct amdgpu_ring *ring); void amdgpu_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count); void amdgpu_ring_generic_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.c index 62079f0e3ee8..73516abef662 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.c @@ -105,6 +105,16 @@ static void amdgpu_mux_resubmit_chunks(struct amdgpu_ring_mux *mux) amdgpu_fence_update_start_timestamp(e->ring, chunk->sync_seq, ktime_get()); + if (chunk->sync_seq == + le32_to_cpu(*(e->ring->fence_drv.cpu_addr + 2))) { + if (chunk->cntl_offset <= e->ring->buf_mask) + amdgpu_ring_patch_cntl(e->ring, + chunk->cntl_offset); + if (chunk->ce_offset <= e->ring->buf_mask) + amdgpu_ring_patch_ce(e->ring, chunk->ce_offset); + if (chunk->de_offset <= e->ring->buf_mask) + amdgpu_ring_patch_de(e->ring, chunk->de_offset); + } amdgpu_ring_mux_copy_pkt_from_sw_ring(mux, e->ring, chunk->start, chunk->end); @@ -407,6 +417,17 @@ void amdgpu_sw_ring_ib_end(struct amdgpu_ring *ring) amdgpu_ring_mux_end_ib(mux, ring); } +void amdgpu_sw_ring_ib_mark_offset(struct amdgpu_ring *ring, enum amdgpu_ring_mux_offset_type type) +{ + struct amdgpu_device *adev = ring->adev; + struct amdgpu_ring_mux *mux = &adev->gfx.muxer; + unsigned offset; + + offset = ring->wptr & ring->buf_mask; + + amdgpu_ring_mux_ib_mark_offset(mux, ring, offset, type); +} + void amdgpu_ring_mux_start_ib(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring) { struct amdgpu_mux_entry *e; @@ -429,6 +450,10 @@ void amdgpu_ring_mux_start_ib(struct amdgpu_ring_mux *mux, struct amdgpu_ring *r } chunk->start = ring->wptr; + /* the initialized value used to check if they are set by the ib submission*/ + chunk->cntl_offset = ring->buf_mask + 1; + chunk->de_offset = ring->buf_mask + 1; + chunk->ce_offset = ring->buf_mask + 1; list_add_tail(&chunk->entry, &e->list); } @@ -454,6 +479,41 @@ static void scan_and_remove_signaled_chunk(struct amdgpu_ring_mux *mux, struct a } } +void amdgpu_ring_mux_ib_mark_offset(struct amdgpu_ring_mux *mux, + struct amdgpu_ring *ring, u64 offset, + enum amdgpu_ring_mux_offset_type type) +{ + struct amdgpu_mux_entry *e; + struct amdgpu_mux_chunk *chunk; + + e = amdgpu_ring_mux_sw_entry(mux, ring); + if (!e) { + DRM_ERROR("cannot find entry!\n"); + return; + } + + chunk = list_last_entry(&e->list, struct amdgpu_mux_chunk, entry); + if (!chunk) { + DRM_ERROR("cannot find chunk!\n"); + return; + } + + switch (type) { + case AMDGPU_MUX_OFFSET_TYPE_CONTROL: + chunk->cntl_offset = offset; + break; + case AMDGPU_MUX_OFFSET_TYPE_DE: + chunk->de_offset = offset; + break; + case AMDGPU_MUX_OFFSET_TYPE_CE: + chunk->ce_offset = offset; + break; + default: + DRM_ERROR("invalid type (%d)\n", type); + break; + } +} + void amdgpu_ring_mux_end_ib(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring) { struct amdgpu_mux_entry *e; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.h index 4be45fc14954..b22d4fb2a847 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.h @@ -50,6 +50,12 @@ struct amdgpu_mux_entry { struct list_head list; }; +enum amdgpu_ring_mux_offset_type { + AMDGPU_MUX_OFFSET_TYPE_CONTROL, + AMDGPU_MUX_OFFSET_TYPE_DE, + AMDGPU_MUX_OFFSET_TYPE_CE, +}; + struct amdgpu_ring_mux { struct amdgpu_ring *real_ring; @@ -72,12 +78,18 @@ struct amdgpu_ring_mux { * @sync_seq: the fence seqno related with the saved IB. * @start:- start location on the software ring. * @end:- end location on the software ring. + * @control_offset:- the PRE_RESUME bit position used for resubmission. + * @de_offset:- the anchor in write_data for de meta of resubmission. + * @ce_offset:- the anchor in write_data for ce meta of resubmission. */ struct amdgpu_mux_chunk { struct list_head entry; uint32_t sync_seq; u64 start; u64 end; + u64 cntl_offset; + u64 de_offset; + u64 ce_offset; }; int amdgpu_ring_mux_init(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring, @@ -89,6 +101,8 @@ u64 amdgpu_ring_mux_get_wptr(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ri u64 amdgpu_ring_mux_get_rptr(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring); void amdgpu_ring_mux_start_ib(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring); void amdgpu_ring_mux_end_ib(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring); +void amdgpu_ring_mux_ib_mark_offset(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring, + u64 offset, enum amdgpu_ring_mux_offset_type type); bool amdgpu_mcbp_handle_trailing_fence_irq(struct amdgpu_ring_mux *mux); u64 amdgpu_sw_ring_get_rptr_gfx(struct amdgpu_ring *ring); @@ -97,6 +111,7 @@ void amdgpu_sw_ring_set_wptr_gfx(struct amdgpu_ring *ring); void amdgpu_sw_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count); void amdgpu_sw_ring_ib_begin(struct amdgpu_ring *ring); void amdgpu_sw_ring_ib_end(struct amdgpu_ring *ring); +void amdgpu_sw_ring_ib_mark_offset(struct amdgpu_ring *ring, enum amdgpu_ring_mux_offset_type type); const char *amdgpu_sw_ring_name(int idx); unsigned int amdgpu_sw_ring_priority(int idx); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c index 85fb730d9fc8..35e0ae9acadc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c @@ -31,12 +31,13 @@ * amdgpu_gfx_rlc_enter_safe_mode - Set RLC into safe mode * * @adev: amdgpu_device pointer + * @xcc_id: xcc accelerated compute core id * * Set RLC enter into safe mode if RLC is enabled and haven't in safe mode. */ -void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev) +void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev, int xcc_id) { - if (adev->gfx.rlc.in_safe_mode) + if (adev->gfx.rlc.in_safe_mode[xcc_id]) return; /* if RLC is not enabled, do nothing */ @@ -46,8 +47,8 @@ void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev) if (adev->cg_flags & (AMD_CG_SUPPORT_GFX_CGCG | AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_3D_CGCG)) { - adev->gfx.rlc.funcs->set_safe_mode(adev); - adev->gfx.rlc.in_safe_mode = true; + adev->gfx.rlc.funcs->set_safe_mode(adev, xcc_id); + adev->gfx.rlc.in_safe_mode[xcc_id] = true; } } @@ -55,12 +56,13 @@ void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev) * amdgpu_gfx_rlc_exit_safe_mode - Set RLC out of safe mode * * @adev: amdgpu_device pointer + * @xcc_id: xcc accelerated compute core id * * Set RLC exit safe mode if RLC is enabled and have entered into safe mode. */ -void amdgpu_gfx_rlc_exit_safe_mode(struct amdgpu_device *adev) +void amdgpu_gfx_rlc_exit_safe_mode(struct amdgpu_device *adev, int xcc_id) { - if (!(adev->gfx.rlc.in_safe_mode)) + if (!(adev->gfx.rlc.in_safe_mode[xcc_id])) return; /* if RLC is not enabled, do nothing */ @@ -70,8 +72,8 @@ void amdgpu_gfx_rlc_exit_safe_mode(struct amdgpu_device *adev) if (adev->cg_flags & (AMD_CG_SUPPORT_GFX_CGCG | AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_3D_CGCG)) { - adev->gfx.rlc.funcs->unset_safe_mode(adev); - adev->gfx.rlc.in_safe_mode = false; + adev->gfx.rlc.funcs->unset_safe_mode(adev, xcc_id); + adev->gfx.rlc.in_safe_mode[xcc_id] = false; } } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h index 23f060db9255..80b263646966 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h @@ -157,8 +157,8 @@ typedef struct _RLC_TABLE_OF_CONTENT { struct amdgpu_rlc_funcs { bool (*is_rlc_enabled)(struct amdgpu_device *adev); - void (*set_safe_mode)(struct amdgpu_device *adev); - void (*unset_safe_mode)(struct amdgpu_device *adev); + void (*set_safe_mode)(struct amdgpu_device *adev, int xcc_id); + void (*unset_safe_mode)(struct amdgpu_device *adev, int xcc_id); int (*init)(struct amdgpu_device *adev); u32 (*get_csb_size)(struct amdgpu_device *adev); void (*get_csb_buffer)(struct amdgpu_device *adev, volatile u32 *buffer); @@ -201,7 +201,7 @@ struct amdgpu_rlc { u32 cp_table_size; /* safe mode for updating CG/PG state */ - bool in_safe_mode; + bool in_safe_mode[8]; const struct amdgpu_rlc_funcs *funcs; /* for firmware data */ @@ -260,8 +260,8 @@ struct amdgpu_rlc { struct amdgpu_rlcg_reg_access_ctrl reg_access_ctrl; }; -void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev); -void amdgpu_gfx_rlc_exit_safe_mode(struct amdgpu_device *adev); +void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev, int xcc_id); +void amdgpu_gfx_rlc_exit_safe_mode(struct amdgpu_device *adev, int xcc_id); int amdgpu_gfx_rlc_init_sr(struct amdgpu_device *adev, u32 dws); int amdgpu_gfx_rlc_init_csb(struct amdgpu_device *adev); int amdgpu_gfx_rlc_init_cpt(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c index 231ca06bc9c7..78ec3420ef85 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c @@ -64,7 +64,7 @@ int amdgpu_sdma_get_index_from_ring(struct amdgpu_ring *ring, uint32_t *index) } uint64_t amdgpu_sdma_get_csa_mc_addr(struct amdgpu_ring *ring, - unsigned vmid) + unsigned int vmid) { struct amdgpu_device *adev = ring->adev; uint64_t csa_mc_addr; @@ -252,6 +252,13 @@ int amdgpu_sdma_init_microcode(struct amdgpu_device *adev, if (!duplicate && (instance != i)) continue; else { + /* Use a single copy per SDMA firmware type. PSP uses the same instance for all + * groups of SDMAs */ + if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 4, 2) && + adev->firmware.load_type == AMDGPU_FW_LOAD_PSP && + adev->sdma.num_inst_per_aid == i) { + break; + } info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA0 + i]; info->ucode_id = AMDGPU_UCODE_ID_SDMA0 + i; info->fw = adev->sdma.instance[i].fw; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h index fc8528812598..513ac22120c1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h @@ -26,7 +26,7 @@ #include "amdgpu_ras.h" /* max number of IP instances */ -#define AMDGPU_MAX_SDMA_INSTANCES 8 +#define AMDGPU_MAX_SDMA_INSTANCES 16 enum amdgpu_sdma_irq { AMDGPU_SDMA_IRQ_INSTANCE0 = 0, @@ -37,9 +37,19 @@ enum amdgpu_sdma_irq { AMDGPU_SDMA_IRQ_INSTANCE5, AMDGPU_SDMA_IRQ_INSTANCE6, AMDGPU_SDMA_IRQ_INSTANCE7, + AMDGPU_SDMA_IRQ_INSTANCE8, + AMDGPU_SDMA_IRQ_INSTANCE9, + AMDGPU_SDMA_IRQ_INSTANCE10, + AMDGPU_SDMA_IRQ_INSTANCE11, + AMDGPU_SDMA_IRQ_INSTANCE12, + AMDGPU_SDMA_IRQ_INSTANCE13, + AMDGPU_SDMA_IRQ_INSTANCE14, + AMDGPU_SDMA_IRQ_INSTANCE15, AMDGPU_SDMA_IRQ_LAST }; +#define NUM_SDMA(x) hweight32(x) + struct amdgpu_sdma_instance { /* SDMA firmware */ const struct firmware *fw; @@ -49,6 +59,35 @@ struct amdgpu_sdma_instance { struct amdgpu_ring ring; struct amdgpu_ring page; bool burst_nop; + uint32_t aid_id; +}; + +enum amdgpu_sdma_ras_memory_id { + AMDGPU_SDMA_MBANK_DATA_BUF0 = 1, + AMDGPU_SDMA_MBANK_DATA_BUF1 = 2, + AMDGPU_SDMA_MBANK_DATA_BUF2 = 3, + AMDGPU_SDMA_MBANK_DATA_BUF3 = 4, + AMDGPU_SDMA_MBANK_DATA_BUF4 = 5, + AMDGPU_SDMA_MBANK_DATA_BUF5 = 6, + AMDGPU_SDMA_MBANK_DATA_BUF6 = 7, + AMDGPU_SDMA_MBANK_DATA_BUF7 = 8, + AMDGPU_SDMA_MBANK_DATA_BUF8 = 9, + AMDGPU_SDMA_MBANK_DATA_BUF9 = 10, + AMDGPU_SDMA_MBANK_DATA_BUF10 = 11, + AMDGPU_SDMA_MBANK_DATA_BUF11 = 12, + AMDGPU_SDMA_MBANK_DATA_BUF12 = 13, + AMDGPU_SDMA_MBANK_DATA_BUF13 = 14, + AMDGPU_SDMA_MBANK_DATA_BUF14 = 15, + AMDGPU_SDMA_MBANK_DATA_BUF15 = 16, + AMDGPU_SDMA_UCODE_BUF = 17, + AMDGPU_SDMA_RB_CMD_BUF = 18, + AMDGPU_SDMA_IB_CMD_BUF = 19, + AMDGPU_SDMA_UTCL1_RD_FIFO = 20, + AMDGPU_SDMA_UTCL1_RDBST_FIFO = 21, + AMDGPU_SDMA_UTCL1_WR_FIFO = 22, + AMDGPU_SDMA_DATA_LUT_FIFO = 23, + AMDGPU_SDMA_SPLIT_DAT_BUF = 24, + AMDGPU_SDMA_MEMORY_BLOCK_LAST, }; struct amdgpu_sdma_ras { @@ -66,6 +105,8 @@ struct amdgpu_sdma { struct amdgpu_irq_src srbm_write_irq; int num_instances; + uint32_t sdma_mask; + int num_inst_per_aid; uint32_t srbm_soft_reset; bool has_page_queue; struct ras_common_if *ras_if; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_smuio.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_smuio.h index c7a823f3f2c5..89c38d864471 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_smuio.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_smuio.h @@ -30,6 +30,7 @@ struct amdgpu_smuio_funcs { void (*get_clock_gating_state)(struct amdgpu_device *adev, u64 *flags); u32 (*get_die_id)(struct amdgpu_device *adev); u32 (*get_socket_id)(struct amdgpu_device *adev); + enum amdgpu_pkg_type (*get_pkg_type)(struct amdgpu_device *adev); bool (*is_host_gpu_xgmi_supported)(struct amdgpu_device *adev); }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index 2cd081cbf706..0534ab716809 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -38,7 +38,6 @@ #include <linux/seq_file.h> #include <linux/slab.h> #include <linux/swap.h> -#include <linux/swiotlb.h> #include <linux/dma-buf.h> #include <linux/sizes.h> #include <linux/module.h> @@ -65,7 +64,7 @@ MODULE_IMPORT_NS(DMA_BUF); -#define AMDGPU_TTM_VRAM_MAX_DW_READ (size_t)128 +#define AMDGPU_TTM_VRAM_MAX_DW_READ ((size_t)128) static int amdgpu_ttm_backend_bind(struct ttm_device *bdev, struct ttm_tt *ttm, @@ -184,11 +183,11 @@ static void amdgpu_evict_flags(struct ttm_buffer_object *bo, static int amdgpu_ttm_map_buffer(struct ttm_buffer_object *bo, struct ttm_resource *mem, struct amdgpu_res_cursor *mm_cur, - unsigned window, struct amdgpu_ring *ring, + unsigned int window, struct amdgpu_ring *ring, bool tmz, uint64_t *size, uint64_t *addr) { struct amdgpu_device *adev = ring->adev; - unsigned offset, num_pages, num_dw, num_bytes; + unsigned int offset, num_pages, num_dw, num_bytes; uint64_t src_addr, dst_addr; struct amdgpu_job *job; void *cpu_addr; @@ -229,7 +228,7 @@ static int amdgpu_ttm_map_buffer(struct ttm_buffer_object *bo, num_dw = ALIGN(adev->mman.buffer_funcs->copy_num_dw, 8); num_bytes = num_pages * 8 * AMDGPU_GPU_PAGES_IN_CPU_PAGE; - r = amdgpu_job_alloc_with_ib(adev, &adev->mman.entity, + r = amdgpu_job_alloc_with_ib(adev, &adev->mman.high_pr, AMDGPU_FENCE_OWNER_UNDEFINED, num_dw * 4 + num_bytes, AMDGPU_IB_POOL_DELAYED, &job); @@ -384,7 +383,8 @@ static int amdgpu_move_blit(struct ttm_buffer_object *bo, (abo->flags & AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE)) { struct dma_fence *wipe_fence = NULL; - r = amdgpu_fill_buffer(abo, AMDGPU_POISON, NULL, &wipe_fence); + r = amdgpu_fill_buffer(abo, AMDGPU_POISON, NULL, &wipe_fence, + false); if (r) { goto error; } else if (wipe_fence) { @@ -631,6 +631,7 @@ struct amdgpu_ttm_tt { struct task_struct *usertask; uint32_t userflags; bool bound; + int32_t pool_id; }; #define ttm_to_amdgpu_ttm_tt(ptr) container_of(ptr, struct amdgpu_ttm_tt, ttm) @@ -800,6 +801,44 @@ static void amdgpu_ttm_tt_unpin_userptr(struct ttm_device *bdev, sg_free_table(ttm->sg); } +/* + * total_pages is constructed as MQD0+CtrlStack0 + MQD1+CtrlStack1 + ... + * MQDn+CtrlStackn where n is the number of XCCs per partition. + * pages_per_xcc is the size of one MQD+CtrlStack. The first page is MQD + * and uses memory type default, UC. The rest of pages_per_xcc are + * Ctrl stack and modify their memory type to NC. + */ +static void amdgpu_ttm_gart_bind_gfx9_mqd(struct amdgpu_device *adev, + struct ttm_tt *ttm, uint64_t flags) +{ + struct amdgpu_ttm_tt *gtt = (void *)ttm; + uint64_t total_pages = ttm->num_pages; + int num_xcc = max(1U, adev->gfx.num_xcc_per_xcp); + uint64_t page_idx, pages_per_xcc; + int i; + uint64_t ctrl_flags = (flags & ~AMDGPU_PTE_MTYPE_VG10_MASK) | + AMDGPU_PTE_MTYPE_VG10(AMDGPU_MTYPE_NC); + + pages_per_xcc = total_pages; + do_div(pages_per_xcc, num_xcc); + + for (i = 0, page_idx = 0; i < num_xcc; i++, page_idx += pages_per_xcc) { + /* MQD page: use default flags */ + amdgpu_gart_bind(adev, + gtt->offset + (page_idx << PAGE_SHIFT), + 1, >t->ttm.dma_address[page_idx], flags); + /* + * Ctrl pages - modify the memory type to NC (ctrl_flags) from + * the second page of the BO onward. + */ + amdgpu_gart_bind(adev, + gtt->offset + ((page_idx + 1) << PAGE_SHIFT), + pages_per_xcc - 1, + >t->ttm.dma_address[page_idx + 1], + ctrl_flags); + } +} + static void amdgpu_ttm_gart_bind(struct amdgpu_device *adev, struct ttm_buffer_object *tbo, uint64_t flags) @@ -812,21 +851,7 @@ static void amdgpu_ttm_gart_bind(struct amdgpu_device *adev, flags |= AMDGPU_PTE_TMZ; if (abo->flags & AMDGPU_GEM_CREATE_CP_MQD_GFX9) { - uint64_t page_idx = 1; - - amdgpu_gart_bind(adev, gtt->offset, page_idx, - gtt->ttm.dma_address, flags); - - /* The memory type of the first page defaults to UC. Now - * modify the memory type to NC from the second page of - * the BO onward. - */ - flags &= ~AMDGPU_PTE_MTYPE_VG10_MASK; - flags |= AMDGPU_PTE_MTYPE_VG10(AMDGPU_MTYPE_NC); - - amdgpu_gart_bind(adev, gtt->offset + (page_idx << PAGE_SHIFT), - ttm->num_pages - page_idx, - &(gtt->ttm.dma_address[page_idx]), flags); + amdgpu_ttm_gart_bind_gfx9_mqd(adev, ttm, flags); } else { amdgpu_gart_bind(adev, gtt->offset, ttm->num_pages, gtt->ttm.dma_address, flags); @@ -1029,15 +1054,20 @@ static void amdgpu_ttm_backend_destroy(struct ttm_device *bdev, static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo, uint32_t page_flags) { + struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev); struct amdgpu_bo *abo = ttm_to_amdgpu_bo(bo); struct amdgpu_ttm_tt *gtt; enum ttm_caching caching; gtt = kzalloc(sizeof(struct amdgpu_ttm_tt), GFP_KERNEL); - if (gtt == NULL) { + if (!gtt) return NULL; - } + gtt->gobj = &bo->base; + if (adev->gmc.mem_partitions && abo->xcp_id >= 0) + gtt->pool_id = KFD_XCP_MEM_ID(adev, abo->xcp_id); + else + gtt->pool_id = abo->xcp_id; if (abo->flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC) caching = ttm_write_combined; @@ -1064,6 +1094,7 @@ static int amdgpu_ttm_tt_populate(struct ttm_device *bdev, { struct amdgpu_device *adev = amdgpu_ttm_adev(bdev); struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm); + struct ttm_pool *pool; pgoff_t i; int ret; @@ -1078,7 +1109,11 @@ static int amdgpu_ttm_tt_populate(struct ttm_device *bdev, if (ttm->page_flags & TTM_TT_FLAG_EXTERNAL) return 0; - ret = ttm_pool_alloc(&adev->mman.bdev.pool, ttm, ctx); + if (adev->mman.ttm_pools && gtt->pool_id >= 0) + pool = &adev->mman.ttm_pools[gtt->pool_id]; + else + pool = &adev->mman.bdev.pool; + ret = ttm_pool_alloc(pool, ttm, ctx); if (ret) return ret; @@ -1099,6 +1134,7 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_device *bdev, { struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm); struct amdgpu_device *adev; + struct ttm_pool *pool; pgoff_t i; amdgpu_ttm_backend_unbind(bdev, ttm); @@ -1117,7 +1153,13 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_device *bdev, ttm->pages[i]->mapping = NULL; adev = amdgpu_ttm_adev(bdev); - return ttm_pool_free(&adev->mman.bdev.pool, ttm); + + if (adev->mman.ttm_pools && gtt->pool_id >= 0) + pool = &adev->mman.ttm_pools[gtt->pool_id]; + else + pool = &adev->mman.bdev.pool; + + return ttm_pool_free(pool, ttm); } /** @@ -1414,7 +1456,7 @@ static int amdgpu_ttm_access_memory_sdma(struct ttm_buffer_object *bo, memcpy(adev->mman.sdma_access_ptr, buf, len); num_dw = ALIGN(adev->mman.buffer_funcs->copy_num_dw, 8); - r = amdgpu_job_alloc_with_ib(adev, &adev->mman.entity, + r = amdgpu_job_alloc_with_ib(adev, &adev->mman.high_pr, AMDGPU_FENCE_OWNER_UNDEFINED, num_dw * 4, AMDGPU_IB_POOL_DELAYED, &job); @@ -1623,14 +1665,15 @@ static int amdgpu_ttm_training_reserve_vram_fini(struct amdgpu_device *adev) return 0; } -static void amdgpu_ttm_training_data_block_init(struct amdgpu_device *adev) +static void amdgpu_ttm_training_data_block_init(struct amdgpu_device *adev, + uint32_t reserve_size) { struct psp_memory_training_context *ctx = &adev->psp.mem_train_ctx; memset(ctx, 0, sizeof(*ctx)); ctx->c2p_train_data_offset = - ALIGN((adev->gmc.mc_vram_size - adev->mman.discovery_tmr_size - SZ_1M), SZ_1M); + ALIGN((adev->gmc.mc_vram_size - reserve_size - SZ_1M), SZ_1M); ctx->p2c_train_data_offset = (adev->gmc.mc_vram_size - GDDR6_MEM_TRAINING_OFFSET); ctx->train_data_size = @@ -1648,11 +1691,12 @@ static void amdgpu_ttm_training_data_block_init(struct amdgpu_device *adev) */ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev) { - int ret; struct psp_memory_training_context *ctx = &adev->psp.mem_train_ctx; bool mem_train_support = false; + uint32_t reserve_size = 0; + int ret; - if (!amdgpu_sriov_vf(adev)) { + if (adev->bios && !amdgpu_sriov_vf(adev)) { if (amdgpu_atomfirmware_mem_training_supported(adev)) mem_train_support = true; else @@ -1666,14 +1710,18 @@ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev) * Otherwise, fallback to legacy approach to check and reserve tmr block for ip * discovery data and G6 memory training data respectively */ - adev->mman.discovery_tmr_size = - amdgpu_atomfirmware_get_fw_reserved_fb_size(adev); - if (!adev->mman.discovery_tmr_size) - adev->mman.discovery_tmr_size = DISCOVERY_TMR_OFFSET; + if (adev->bios) + reserve_size = + amdgpu_atomfirmware_get_fw_reserved_fb_size(adev); + + if (!adev->bios && adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3)) + reserve_size = max(reserve_size, (uint32_t)280 << 20); + else if (!reserve_size) + reserve_size = DISCOVERY_TMR_OFFSET; if (mem_train_support) { /* reserve vram for mem train according to TMR location */ - amdgpu_ttm_training_data_block_init(adev); + amdgpu_ttm_training_data_block_init(adev, reserve_size); ret = amdgpu_bo_create_kernel_at(adev, ctx->c2p_train_data_offset, ctx->train_data_size, @@ -1687,20 +1735,58 @@ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev) ctx->init = PSP_MEM_TRAIN_RESERVE_SUCCESS; } - ret = amdgpu_bo_create_kernel_at(adev, - adev->gmc.real_vram_size - adev->mman.discovery_tmr_size, - adev->mman.discovery_tmr_size, - &adev->mman.discovery_memory, - NULL); - if (ret) { - DRM_ERROR("alloc tmr failed(%d)!\n", ret); - amdgpu_bo_free_kernel(&adev->mman.discovery_memory, NULL, NULL); - return ret; + if (!adev->gmc.is_app_apu) { + ret = amdgpu_bo_create_kernel_at( + adev, adev->gmc.real_vram_size - reserve_size, + reserve_size, &adev->mman.fw_reserved_memory, NULL); + if (ret) { + DRM_ERROR("alloc tmr failed(%d)!\n", ret); + amdgpu_bo_free_kernel(&adev->mman.fw_reserved_memory, + NULL, NULL); + return ret; + } + } else { + DRM_DEBUG_DRIVER("backdoor fw loading path for PSP TMR, no reservation needed\n"); } return 0; } +static int amdgpu_ttm_pools_init(struct amdgpu_device *adev) +{ + int i; + + if (!adev->gmc.is_app_apu || !adev->gmc.num_mem_partitions) + return 0; + + adev->mman.ttm_pools = kcalloc(adev->gmc.num_mem_partitions, + sizeof(*adev->mman.ttm_pools), + GFP_KERNEL); + if (!adev->mman.ttm_pools) + return -ENOMEM; + + for (i = 0; i < adev->gmc.num_mem_partitions; i++) { + ttm_pool_init(&adev->mman.ttm_pools[i], adev->dev, + adev->gmc.mem_partitions[i].numa.node, + false, false); + } + return 0; +} + +static void amdgpu_ttm_pools_fini(struct amdgpu_device *adev) +{ + int i; + + if (!adev->gmc.is_app_apu || !adev->mman.ttm_pools) + return; + + for (i = 0; i < adev->gmc.num_mem_partitions; i++) + ttm_pool_fini(&adev->mman.ttm_pools[i]); + + kfree(adev->mman.ttm_pools); + adev->mman.ttm_pools = NULL; +} + /* * amdgpu_ttm_init - Init the memory management (ttm) as well as various * gtt/vram related fields. @@ -1727,6 +1813,12 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) DRM_ERROR("failed initializing buffer object driver(%d).\n", r); return r; } + + r = amdgpu_ttm_pools_init(adev); + if (r) { + DRM_ERROR("failed to init ttm pools(%d).\n", r); + return r; + } adev->mman.initialized = true; /* Initialize VRAM pool with all of VRAM divided into pages */ @@ -1744,6 +1836,9 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) adev->mman.aper_base_kaddr = ioremap_cache(adev->gmc.aper_base, adev->gmc.visible_vram_size); + else if (adev->gmc.is_app_apu) + DRM_DEBUG_DRIVER( + "No need to ioremap when real vram size is 0\n"); else #endif adev->mman.aper_base_kaddr = ioremap_wc(adev->gmc.aper_base, @@ -1755,9 +1850,8 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) *place on the VRAM, so reserve it early. */ r = amdgpu_ttm_fw_reserve_vram_init(adev); - if (r) { + if (r) return r; - } /* *The reserved vram for driver must be pinned to the specified @@ -1781,49 +1875,46 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) /* allocate memory as required for VGA * This is used for VGA emulation and pre-OS scanout buffers to * avoid display artifacts while transitioning between pre-OS - * and driver. */ - r = amdgpu_bo_create_kernel_at(adev, 0, adev->mman.stolen_vga_size, - &adev->mman.stolen_vga_memory, - NULL); - if (r) - return r; - r = amdgpu_bo_create_kernel_at(adev, adev->mman.stolen_vga_size, - adev->mman.stolen_extended_size, - &adev->mman.stolen_extended_memory, - NULL); - if (r) - return r; - r = amdgpu_bo_create_kernel_at(adev, adev->mman.stolen_reserved_offset, - adev->mman.stolen_reserved_size, - &adev->mman.stolen_reserved_memory, - NULL); - if (r) - return r; + * and driver. + */ + if (!adev->gmc.is_app_apu) { + r = amdgpu_bo_create_kernel_at(adev, 0, + adev->mman.stolen_vga_size, + &adev->mman.stolen_vga_memory, + NULL); + if (r) + return r; - DRM_INFO("amdgpu: %uM of VRAM memory ready\n", - (unsigned) (adev->gmc.real_vram_size / (1024 * 1024))); - - /* Compute GTT size, either based on 1/2 the size of RAM size - * or whatever the user passed on module init */ - if (amdgpu_gtt_size == -1) { - struct sysinfo si; - - si_meminfo(&si); - /* Certain GL unit tests for large textures can cause problems - * with the OOM killer since there is no way to link this memory - * to a process. This was originally mitigated (but not necessarily - * eliminated) by limiting the GTT size. The problem is this limit - * is often too low for many modern games so just make the limit 1/2 - * of system memory which aligns with TTM. The OOM accounting needs - * to be addressed, but we shouldn't prevent common 3D applications - * from being usable just to potentially mitigate that corner case. - */ - gtt_size = max((AMDGPU_DEFAULT_GTT_SIZE_MB << 20), - (u64)si.totalram * si.mem_unit / 2); + r = amdgpu_bo_create_kernel_at(adev, adev->mman.stolen_vga_size, + adev->mman.stolen_extended_size, + &adev->mman.stolen_extended_memory, + NULL); + + if (r) + return r; + + r = amdgpu_bo_create_kernel_at(adev, + adev->mman.stolen_reserved_offset, + adev->mman.stolen_reserved_size, + &adev->mman.stolen_reserved_memory, + NULL); + if (r) + return r; } else { - gtt_size = (uint64_t)amdgpu_gtt_size << 20; + DRM_DEBUG_DRIVER("Skipped stolen memory reservation\n"); } + DRM_INFO("amdgpu: %uM of VRAM memory ready\n", + (unsigned int)(adev->gmc.real_vram_size / (1024 * 1024))); + + /* Compute GTT size, either based on TTM limit + * or whatever the user passed on module init. + */ + if (amdgpu_gtt_size == -1) + gtt_size = ttm_tt_pages_limit() << PAGE_SHIFT; + else + gtt_size = (uint64_t)amdgpu_gtt_size << 20; + /* Initialize GTT memory pool */ r = amdgpu_gtt_mgr_init(adev, gtt_size); if (r) { @@ -1831,7 +1922,7 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) return r; } DRM_INFO("amdgpu: %uM of GTT memory ready.\n", - (unsigned)(gtt_size / (1024 * 1024))); + (unsigned int)(gtt_size / (1024 * 1024))); /* Initialize preemptible memory pool */ r = amdgpu_preempt_mgr_init(adev); @@ -1858,7 +1949,6 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) DRM_ERROR("Failed initializing oa heap.\n"); return r; } - if (amdgpu_bo_create_kernel(adev, PAGE_SIZE, PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT, &adev->mman.sdma_access_bo, NULL, @@ -1874,18 +1964,24 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) void amdgpu_ttm_fini(struct amdgpu_device *adev) { int idx; + if (!adev->mman.initialized) return; + amdgpu_ttm_pools_fini(adev); + amdgpu_ttm_training_reserve_vram_fini(adev); /* return the stolen vga memory back to VRAM */ - amdgpu_bo_free_kernel(&adev->mman.stolen_vga_memory, NULL, NULL); - amdgpu_bo_free_kernel(&adev->mman.stolen_extended_memory, NULL, NULL); - /* return the IP Discovery TMR memory back to VRAM */ - amdgpu_bo_free_kernel(&adev->mman.discovery_memory, NULL, NULL); - if (adev->mman.stolen_reserved_size) - amdgpu_bo_free_kernel(&adev->mman.stolen_reserved_memory, - NULL, NULL); + if (!adev->gmc.is_app_apu) { + amdgpu_bo_free_kernel(&adev->mman.stolen_vga_memory, NULL, NULL); + amdgpu_bo_free_kernel(&adev->mman.stolen_extended_memory, NULL, NULL); + /* return the FW reserved memory back to VRAM */ + amdgpu_bo_free_kernel(&adev->mman.fw_reserved_memory, NULL, + NULL); + if (adev->mman.stolen_reserved_size) + amdgpu_bo_free_kernel(&adev->mman.stolen_reserved_memory, + NULL, NULL); + } amdgpu_bo_free_kernel(&adev->mman.sdma_access_bo, NULL, &adev->mman.sdma_access_ptr); amdgpu_ttm_fw_reserve_vram_fini(adev); @@ -1927,7 +2023,7 @@ void amdgpu_ttm_set_buffer_funcs_status(struct amdgpu_device *adev, bool enable) int r; if (!adev->mman.initialized || amdgpu_in_reset(adev) || - adev->mman.buffer_funcs_enabled == enable) + adev->mman.buffer_funcs_enabled == enable || adev->gmc.is_app_apu) return; if (enable) { @@ -1936,7 +2032,7 @@ void amdgpu_ttm_set_buffer_funcs_status(struct amdgpu_device *adev, bool enable) ring = adev->mman.buffer_funcs_ring; sched = &ring->sched; - r = drm_sched_entity_init(&adev->mman.entity, + r = drm_sched_entity_init(&adev->mman.high_pr, DRM_SCHED_PRIORITY_KERNEL, &sched, 1, NULL); if (r) { @@ -1944,8 +2040,18 @@ void amdgpu_ttm_set_buffer_funcs_status(struct amdgpu_device *adev, bool enable) r); return; } + + r = drm_sched_entity_init(&adev->mman.low_pr, + DRM_SCHED_PRIORITY_NORMAL, &sched, + 1, NULL); + if (r) { + DRM_ERROR("Failed setting up TTM BO move entity (%d)\n", + r); + goto error_free_entity; + } } else { - drm_sched_entity_destroy(&adev->mman.entity); + drm_sched_entity_destroy(&adev->mman.high_pr); + drm_sched_entity_destroy(&adev->mman.low_pr); dma_fence_put(man->move); man->move = NULL; } @@ -1957,6 +2063,11 @@ void amdgpu_ttm_set_buffer_funcs_status(struct amdgpu_device *adev, bool enable) size = adev->gmc.visible_vram_size; man->size = size; adev->mman.buffer_funcs_enabled = enable; + + return; + +error_free_entity: + drm_sched_entity_destroy(&adev->mman.high_pr); } static int amdgpu_ttm_prepare_job(struct amdgpu_device *adev, @@ -1964,14 +2075,16 @@ static int amdgpu_ttm_prepare_job(struct amdgpu_device *adev, unsigned int num_dw, struct dma_resv *resv, bool vm_needs_flush, - struct amdgpu_job **job) + struct amdgpu_job **job, + bool delayed) { enum amdgpu_ib_pool_type pool = direct_submit ? AMDGPU_IB_POOL_DIRECT : AMDGPU_IB_POOL_DELAYED; int r; - - r = amdgpu_job_alloc_with_ib(adev, &adev->mman.entity, + struct drm_sched_entity *entity = delayed ? &adev->mman.low_pr : + &adev->mman.high_pr; + r = amdgpu_job_alloc_with_ib(adev, entity, AMDGPU_FENCE_OWNER_UNDEFINED, num_dw * 4, pool, job); if (r) @@ -1997,10 +2110,10 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring, uint64_t src_offset, bool vm_needs_flush, bool tmz) { struct amdgpu_device *adev = ring->adev; - unsigned num_loops, num_dw; + unsigned int num_loops, num_dw; struct amdgpu_job *job; uint32_t max_bytes; - unsigned i; + unsigned int i; int r; if (!direct_submit && !ring->sched.ready) { @@ -2012,7 +2125,7 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring, uint64_t src_offset, num_loops = DIV_ROUND_UP(byte_count, max_bytes); num_dw = ALIGN(num_loops * adev->mman.buffer_funcs->copy_num_dw, 8); r = amdgpu_ttm_prepare_job(adev, direct_submit, num_dw, - resv, vm_needs_flush, &job); + resv, vm_needs_flush, &job, false); if (r) return r; @@ -2048,7 +2161,7 @@ static int amdgpu_ttm_fill_mem(struct amdgpu_ring *ring, uint32_t src_data, uint64_t dst_addr, uint32_t byte_count, struct dma_resv *resv, struct dma_fence **fence, - bool vm_needs_flush) + bool vm_needs_flush, bool delayed) { struct amdgpu_device *adev = ring->adev; unsigned int num_loops, num_dw; @@ -2061,7 +2174,7 @@ static int amdgpu_ttm_fill_mem(struct amdgpu_ring *ring, uint32_t src_data, num_loops = DIV_ROUND_UP_ULL(byte_count, max_bytes); num_dw = ALIGN(num_loops * adev->mman.buffer_funcs->fill_num_dw, 8); r = amdgpu_ttm_prepare_job(adev, false, num_dw, resv, vm_needs_flush, - &job); + &job, delayed); if (r) return r; @@ -2084,7 +2197,8 @@ static int amdgpu_ttm_fill_mem(struct amdgpu_ring *ring, uint32_t src_data, int amdgpu_fill_buffer(struct amdgpu_bo *bo, uint32_t src_data, struct dma_resv *resv, - struct dma_fence **f) + struct dma_fence **f, + bool delayed) { struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring; @@ -2113,7 +2227,7 @@ int amdgpu_fill_buffer(struct amdgpu_bo *bo, goto error; r = amdgpu_ttm_fill_mem(ring, src_data, to, cur_size, resv, - &next, true); + &next, true, delayed); if (r) goto error; @@ -2164,7 +2278,7 @@ int amdgpu_ttm_evict_resources(struct amdgpu_device *adev, int mem_type) static int amdgpu_ttm_page_pool_show(struct seq_file *m, void *unused) { - struct amdgpu_device *adev = (struct amdgpu_device *)m->private; + struct amdgpu_device *adev = m->private; return ttm_pool_debugfs(&adev->mman.bdev.pool, m); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h index e2cd5894afc9..6d0d66e40db9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h @@ -49,6 +49,7 @@ struct amdgpu_gtt_mgr { struct amdgpu_mman { struct ttm_device bdev; + struct ttm_pool *ttm_pools; bool initialized; void __iomem *aper_base_kaddr; @@ -58,8 +59,10 @@ struct amdgpu_mman { bool buffer_funcs_enabled; struct mutex gtt_window_lock; - /* Scheduler entity for buffer moves */ - struct drm_sched_entity entity; + /* High priority scheduler entity for buffer moves */ + struct drm_sched_entity high_pr; + /* Low priority scheduler entity for VRAM clearing */ + struct drm_sched_entity low_pr; struct amdgpu_vram_mgr vram_mgr; struct amdgpu_gtt_mgr gtt_mgr; @@ -78,7 +81,8 @@ struct amdgpu_mman { /* discovery */ uint8_t *discovery_bin; uint32_t discovery_tmr_size; - struct amdgpu_bo *discovery_memory; + /* fw reserved memory */ + struct amdgpu_bo *fw_reserved_memory; /* firmware VRAM reservation */ u64 fw_vram_usage_start_offset; @@ -150,7 +154,8 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev, int amdgpu_fill_buffer(struct amdgpu_bo *bo, uint32_t src_data, struct dma_resv *resv, - struct dma_fence **fence); + struct dma_fence **fence, + bool delayed); int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo); void amdgpu_ttm_recover_gart(struct ttm_buffer_object *tbo); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c index f76b1cb8baf8..16807ff96dc9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c @@ -748,7 +748,7 @@ static int amdgpu_ucode_init_single_fw(struct amdgpu_device *adev, const struct imu_firmware_header_v1_0 *imu_hdr = NULL; u8 *ucode_addr; - if (NULL == ucode->fw) + if (!ucode->fw) return 0; ucode->mc_addr = mc_addr; @@ -972,7 +972,7 @@ static int amdgpu_ucode_patch_jt(struct amdgpu_firmware_info *ucode, uint8_t *src_addr = NULL; uint8_t *dst_addr = NULL; - if (NULL == ucode->fw) + if (!ucode->fw) return 0; comm_hdr = (const struct common_firmware_header *)ucode->fw->data; @@ -1043,6 +1043,7 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev) if (i == AMDGPU_UCODE_ID_CP_MEC1 && adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) { const struct gfx_firmware_header_v1_0 *cp_hdr; + cp_hdr = (const struct gfx_firmware_header_v1_0 *)ucode->fw->data; amdgpu_ucode_patch_jt(ucode, adev->firmware.fw_buf_mc + fw_offset, adev->firmware.fw_buf_ptr + fw_offset); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c index 1edf8e6aeb16..db0d94ca4ffc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c @@ -169,27 +169,31 @@ int amdgpu_umc_poison_handler(struct amdgpu_device *adev, bool reset) { int ret = AMDGPU_RAS_SUCCESS; - if (!amdgpu_sriov_vf(adev)) { - if (!adev->gmc.xgmi.connected_to_cpu) { - struct ras_err_data err_data = {0, 0, 0, NULL}; - struct ras_common_if head = { - .block = AMDGPU_RAS_BLOCK__UMC, - }; - struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head); - - ret = amdgpu_umc_do_page_retirement(adev, &err_data, NULL, reset); - - if (ret == AMDGPU_RAS_SUCCESS && obj) { - obj->err_data.ue_count += err_data.ue_count; - obj->err_data.ce_count += err_data.ce_count; - } - } else if (reset) { + if (adev->gmc.xgmi.connected_to_cpu || + adev->gmc.is_app_apu) { + if (reset) { /* MCA poison handler is only responsible for GPU reset, * let MCA notifier do page retirement. */ kgd2kfd_set_sram_ecc_flag(adev->kfd.dev); amdgpu_ras_reset_gpu(adev); } + return ret; + } + + if (!amdgpu_sriov_vf(adev)) { + struct ras_err_data err_data = {0, 0, 0, NULL}; + struct ras_common_if head = { + .block = AMDGPU_RAS_BLOCK__UMC, + }; + struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head); + + ret = amdgpu_umc_do_page_retirement(adev, &err_data, NULL, reset); + + if (ret == AMDGPU_RAS_SUCCESS && obj) { + obj->err_data.ue_count += err_data.ue_count; + obj->err_data.ce_count += err_data.ce_count; + } } else { if (adev->virt.ops && adev->virt.ops->ras_poison_handler) adev->virt.ops->ras_poison_handler(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h index 86133f77a9a4..43321f57f557 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h @@ -59,6 +59,8 @@ struct amdgpu_umc_ras { void *ras_error_status); void (*ecc_info_query_ras_error_address)(struct amdgpu_device *adev, void *ras_error_status); + /* support different eeprom table version for different asic */ + void (*set_eeprom_table_version)(struct amdgpu_ras_eeprom_table_header *hdr); }; struct amdgpu_umc_funcs { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h index 919d9d401750..107f9bb0e24f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h @@ -35,17 +35,51 @@ struct amdgpu_debugfs_regs2_iocdata { } srbm; }; +struct amdgpu_debugfs_regs2_iocdata_v2 { + __u32 use_srbm, use_grbm, pg_lock; + struct { + __u32 se, sh, instance; + } grbm; + struct { + __u32 me, pipe, queue, vmid; + } srbm; + u32 xcc_id; +}; + +struct amdgpu_debugfs_gprwave_iocdata { + u32 gpr_or_wave, se, sh, cu, wave, simd, xcc_id; + struct { + u32 thread, vpgr_or_sgpr; + } gpr; +}; + /* * MMIO debugfs state data (per file* handle) */ struct amdgpu_debugfs_regs2_data { struct amdgpu_device *adev; struct mutex lock; - struct amdgpu_debugfs_regs2_iocdata id; + struct amdgpu_debugfs_regs2_iocdata_v2 id; +}; + +struct amdgpu_debugfs_gprwave_data { + struct amdgpu_device *adev; + struct mutex lock; + struct amdgpu_debugfs_gprwave_iocdata id; }; enum AMDGPU_DEBUGFS_REGS2_CMDS { AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE=0, + AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE_V2, +}; + +enum AMDGPU_DEBUGFS_GPRWAVE_CMDS { + AMDGPU_DEBUGFS_GPRWAVE_CMD_SET_STATE=0, }; +//reg2 interface #define AMDGPU_DEBUGFS_REGS2_IOC_SET_STATE _IOWR(0x20, AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE, struct amdgpu_debugfs_regs2_iocdata) +#define AMDGPU_DEBUGFS_REGS2_IOC_SET_STATE_V2 _IOWR(0x20, AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE_V2, struct amdgpu_debugfs_regs2_iocdata_v2) + +//gprwave interface +#define AMDGPU_DEBUGFS_GPRWAVE_IOC_SET_STATE _IOWR(0x20, AMDGPU_DEBUGFS_GPRWAVE_CMD_SET_STATE, struct amdgpu_debugfs_gprwave_iocdata) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c index 6887109abb13..b7441654e6fa 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c @@ -96,16 +96,16 @@ */ struct amdgpu_uvd_cs_ctx { struct amdgpu_cs_parser *parser; - unsigned reg, count; - unsigned data0, data1; - unsigned idx; + unsigned int reg, count; + unsigned int data0, data1; + unsigned int idx; struct amdgpu_ib *ib; /* does the IB has a msg command */ bool has_msg_cmd; /* minimum buffer sizes */ - unsigned *buf_sizes; + unsigned int *buf_sizes; }; #ifdef CONFIG_DRM_AMDGPU_SI @@ -186,7 +186,7 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev) unsigned long bo_size; const char *fw_name; const struct common_firmware_header *hdr; - unsigned family_id; + unsigned int family_id; int i, j, r; INIT_DELAYED_WORK(&adev->uvd.idle_work, amdgpu_uvd_idle_work_handler); @@ -275,7 +275,7 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev) family_id = le32_to_cpu(hdr->ucode_version) & 0xff; if (adev->asic_type < CHIP_VEGA20) { - unsigned version_major, version_minor; + unsigned int version_major, version_minor; version_major = (le32_to_cpu(hdr->ucode_version) >> 24) & 0xff; version_minor = (le32_to_cpu(hdr->ucode_version) >> 8) & 0xff; @@ -420,7 +420,7 @@ int amdgpu_uvd_entity_init(struct amdgpu_device *adev) int amdgpu_uvd_suspend(struct amdgpu_device *adev) { - unsigned size; + unsigned int size; void *ptr; int i, j, idx; bool in_ras_intr = amdgpu_ras_intr_triggered(); @@ -469,7 +469,7 @@ int amdgpu_uvd_suspend(struct amdgpu_device *adev) int amdgpu_uvd_resume(struct amdgpu_device *adev) { - unsigned size; + unsigned int size; void *ptr; int i, idx; @@ -491,7 +491,7 @@ int amdgpu_uvd_resume(struct amdgpu_device *adev) adev->uvd.inst[i].saved_bo = NULL; } else { const struct common_firmware_header *hdr; - unsigned offset; + unsigned int offset; hdr = (const struct common_firmware_header *)adev->uvd.fw->data; if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) { @@ -542,6 +542,7 @@ void amdgpu_uvd_free_handles(struct amdgpu_device *adev, struct drm_file *filp) static void amdgpu_uvd_force_into_uvd_segment(struct amdgpu_bo *abo) { int i; + for (i = 0; i < abo->placement.num_placement; ++i) { abo->placements[i].fpfn = 0 >> PAGE_SHIFT; abo->placements[i].lpfn = (256 * 1024 * 1024) >> PAGE_SHIFT; @@ -579,7 +580,7 @@ static int amdgpu_uvd_cs_pass1(struct amdgpu_uvd_cs_ctx *ctx) r = amdgpu_cs_find_mapping(ctx->parser, addr, &bo, &mapping); if (r) { - DRM_ERROR("Can't find BO for addr 0x%08Lx\n", addr); + DRM_ERROR("Can't find BO for addr 0x%08llx\n", addr); return r; } @@ -589,6 +590,7 @@ static int amdgpu_uvd_cs_pass1(struct amdgpu_uvd_cs_ctx *ctx) if (cmd == 0x0 || cmd == 0x3) { /* yes, force it into VRAM */ uint32_t domain = AMDGPU_GEM_DOMAIN_VRAM; + amdgpu_bo_placement_from_domain(bo, domain); } amdgpu_uvd_force_into_uvd_segment(bo); @@ -609,21 +611,21 @@ static int amdgpu_uvd_cs_pass1(struct amdgpu_uvd_cs_ctx *ctx) * Peek into the decode message and calculate the necessary buffer sizes. */ static int amdgpu_uvd_cs_msg_decode(struct amdgpu_device *adev, uint32_t *msg, - unsigned buf_sizes[]) + unsigned int buf_sizes[]) { - unsigned stream_type = msg[4]; - unsigned width = msg[6]; - unsigned height = msg[7]; - unsigned dpb_size = msg[9]; - unsigned pitch = msg[28]; - unsigned level = msg[57]; + unsigned int stream_type = msg[4]; + unsigned int width = msg[6]; + unsigned int height = msg[7]; + unsigned int dpb_size = msg[9]; + unsigned int pitch = msg[28]; + unsigned int level = msg[57]; - unsigned width_in_mb = width / 16; - unsigned height_in_mb = ALIGN(height / 16, 2); - unsigned fs_in_mb = width_in_mb * height_in_mb; + unsigned int width_in_mb = width / 16; + unsigned int height_in_mb = ALIGN(height / 16, 2); + unsigned int fs_in_mb = width_in_mb * height_in_mb; - unsigned image_size, tmp, min_dpb_size, num_dpb_buffer; - unsigned min_ctx_size = ~0; + unsigned int image_size, tmp, min_dpb_size, num_dpb_buffer; + unsigned int min_ctx_size = ~0; image_size = width * height; image_size += image_size / 2; @@ -631,7 +633,7 @@ static int amdgpu_uvd_cs_msg_decode(struct amdgpu_device *adev, uint32_t *msg, switch (stream_type) { case 0: /* H264 */ - switch(level) { + switch (level) { case 30: num_dpb_buffer = 8100 / fs_in_mb; break; @@ -709,7 +711,7 @@ static int amdgpu_uvd_cs_msg_decode(struct amdgpu_device *adev, uint32_t *msg, break; case 7: /* H264 Perf */ - switch(level) { + switch (level) { case 30: num_dpb_buffer = 8100 / fs_in_mb; break; @@ -742,7 +744,7 @@ static int amdgpu_uvd_cs_msg_decode(struct amdgpu_device *adev, uint32_t *msg, /* reference picture buffer */ min_dpb_size = image_size * num_dpb_buffer; - if (!adev->uvd.use_ctx_buf){ + if (!adev->uvd.use_ctx_buf) { /* macroblock context buffer */ min_dpb_size += width_in_mb * height_in_mb * num_dpb_buffer * 192; @@ -805,7 +807,7 @@ static int amdgpu_uvd_cs_msg_decode(struct amdgpu_device *adev, uint32_t *msg, * Make sure that we don't open up to many sessions. */ static int amdgpu_uvd_cs_msg(struct amdgpu_uvd_cs_ctx *ctx, - struct amdgpu_bo *bo, unsigned offset) + struct amdgpu_bo *bo, unsigned int offset) { struct amdgpu_device *adev = ctx->parser->adev; int32_t *msg, msg_type, handle; @@ -911,7 +913,7 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx) r = amdgpu_cs_find_mapping(ctx->parser, addr, &bo, &mapping); if (r) { - DRM_ERROR("Can't find BO for addr 0x%08Lx\n", addr); + DRM_ERROR("Can't find BO for addr 0x%08llx\n", addr); return r; } @@ -930,7 +932,7 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx) if (cmd < 0x4) { if ((end - start) < ctx->buf_sizes[cmd]) { DRM_ERROR("buffer (%d) to small (%d / %d)!\n", cmd, - (unsigned)(end - start), + (unsigned int)(end - start), ctx->buf_sizes[cmd]); return -EINVAL; } @@ -938,7 +940,7 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx) } else if (cmd == 0x206) { if ((end - start) < ctx->buf_sizes[4]) { DRM_ERROR("buffer (%d) to small (%d / %d)!\n", cmd, - (unsigned)(end - start), + (unsigned int)(end - start), ctx->buf_sizes[4]); return -EINVAL; } @@ -949,14 +951,14 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx) if (!ctx->parser->adev->uvd.address_64_bit) { if ((start >> 28) != ((end - 1) >> 28)) { - DRM_ERROR("reloc %LX-%LX crossing 256MB boundary!\n", + DRM_ERROR("reloc %llx-%llx crossing 256MB boundary!\n", start, end); return -EINVAL; } if ((cmd == 0 || cmd == 0x3) && (start >> 28) != (ctx->parser->adev->uvd.inst->gpu_addr >> 28)) { - DRM_ERROR("msg/fb buffer %LX-%LX out of 256MB segment!\n", + DRM_ERROR("msg/fb buffer %llx-%llx out of 256MB segment!\n", start, end); return -EINVAL; } @@ -990,7 +992,7 @@ static int amdgpu_uvd_cs_reg(struct amdgpu_uvd_cs_ctx *ctx, ctx->idx++; for (i = 0; i <= ctx->count; ++i) { - unsigned reg = ctx->reg + i; + unsigned int reg = ctx->reg + i; if (ctx->idx >= ctx->ib->length_dw) { DRM_ERROR("Register command after end of CS!\n"); @@ -1036,7 +1038,8 @@ static int amdgpu_uvd_cs_packets(struct amdgpu_uvd_cs_ctx *ctx, for (ctx->idx = 0 ; ctx->idx < ctx->ib->length_dw; ) { uint32_t cmd = amdgpu_ib_get_value(ctx->ib, ctx->idx); - unsigned type = CP_PACKET_GET_TYPE(cmd); + unsigned int type = CP_PACKET_GET_TYPE(cmd); + switch (type) { case PACKET_TYPE0: ctx->reg = CP_PACKET0_GET_REG(cmd); @@ -1070,7 +1073,7 @@ int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, struct amdgpu_ib *ib) { struct amdgpu_uvd_cs_ctx ctx = {}; - unsigned buf_sizes[] = { + unsigned int buf_sizes[] = { [0x00000000] = 2048, [0x00000001] = 0xFFFFFFFF, [0x00000002] = 0xFFFFFFFF, @@ -1185,8 +1188,9 @@ err_free: } /* multiple fence commands without any stream commands in between can - crash the vcpu so just try to emmit a dummy create/destroy msg to - avoid this */ + * crash the vcpu so just try to emmit a dummy create/destroy msg to + * avoid this + */ int amdgpu_uvd_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, struct dma_fence **fence) { @@ -1252,15 +1256,14 @@ static void amdgpu_uvd_idle_work_handler(struct work_struct *work) { struct amdgpu_device *adev = container_of(work, struct amdgpu_device, uvd.idle_work.work); - unsigned fences = 0, i, j; + unsigned int fences = 0, i, j; for (i = 0; i < adev->uvd.num_uvd_inst; ++i) { if (adev->uvd.harvest_config & (1 << i)) continue; fences += amdgpu_fence_count_emitted(&adev->uvd.inst[i].ring); - for (j = 0; j < adev->uvd.num_enc_rings; ++j) { + for (j = 0; j < adev->uvd.num_enc_rings; ++j) fences += amdgpu_fence_count_emitted(&adev->uvd.inst[i].ring_enc[j]); - } } if (fences == 0) { @@ -1356,7 +1359,7 @@ error: */ uint32_t amdgpu_uvd_used_handles(struct amdgpu_device *adev) { - unsigned i; + unsigned int i; uint32_t used_handles = 0; for (i = 0; i < adev->uvd.max_handles; ++i) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c index e2b7324a70cb..1904edf68407 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c @@ -99,7 +99,7 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size) { const char *fw_name; const struct common_firmware_header *hdr; - unsigned ucode_version, version_major, version_minor, binary_id; + unsigned int ucode_version, version_major, version_minor, binary_id; int i, r; switch (adev->asic_type) { @@ -207,7 +207,7 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size) */ int amdgpu_vce_sw_fini(struct amdgpu_device *adev) { - unsigned i; + unsigned int i; if (adev->vce.vcpu_bo == NULL) return 0; @@ -286,7 +286,7 @@ int amdgpu_vce_resume(struct amdgpu_device *adev) { void *cpu_addr; const struct common_firmware_header *hdr; - unsigned offset; + unsigned int offset; int r, idx; if (adev->vce.vcpu_bo == NULL) @@ -332,7 +332,7 @@ static void amdgpu_vce_idle_work_handler(struct work_struct *work) { struct amdgpu_device *adev = container_of(work, struct amdgpu_device, vce.idle_work.work); - unsigned i, count = 0; + unsigned int i, count = 0; for (i = 0; i < adev->vce.num_rings; i++) count += amdgpu_fence_count_emitted(&adev->vce.ring[i]); @@ -409,6 +409,7 @@ void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp) { struct amdgpu_ring *ring = &adev->vce.ring[0]; int i, r; + for (i = 0; i < AMDGPU_MAX_VCE_HANDLES; ++i) { uint32_t handle = atomic_read(&adev->vce.handles[i]); @@ -436,7 +437,7 @@ void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp) static int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, struct dma_fence **fence) { - const unsigned ib_size_dw = 1024; + const unsigned int ib_size_dw = 1024; struct amdgpu_job *job; struct amdgpu_ib *ib; struct amdgpu_ib ib_msg; @@ -528,7 +529,7 @@ err: static int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, bool direct, struct dma_fence **fence) { - const unsigned ib_size_dw = 1024; + const unsigned int ib_size_dw = 1024; struct amdgpu_job *job; struct amdgpu_ib *ib; struct dma_fence *f = NULL; @@ -596,12 +597,12 @@ err: */ static int amdgpu_vce_validate_bo(struct amdgpu_cs_parser *p, struct amdgpu_ib *ib, int lo, int hi, - unsigned size, int32_t index) + unsigned int size, int32_t index) { int64_t offset = ((uint64_t)size) * ((int64_t)index); struct ttm_operation_ctx ctx = { false, false }; struct amdgpu_bo_va_mapping *mapping; - unsigned i, fpfn, lpfn; + unsigned int i, fpfn, lpfn; struct amdgpu_bo *bo; uint64_t addr; int r; @@ -619,7 +620,7 @@ static int amdgpu_vce_validate_bo(struct amdgpu_cs_parser *p, r = amdgpu_cs_find_mapping(p, addr, &bo, &mapping); if (r) { - DRM_ERROR("Can't find BO for addr 0x%010Lx %d %d %d %d\n", + DRM_ERROR("Can't find BO for addr 0x%010llx %d %d %d %d\n", addr, lo, hi, size, index); return r; } @@ -646,7 +647,7 @@ static int amdgpu_vce_validate_bo(struct amdgpu_cs_parser *p, * Patch relocation inside command stream with real buffer address */ static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, struct amdgpu_ib *ib, - int lo, int hi, unsigned size, uint32_t index) + int lo, int hi, unsigned int size, uint32_t index) { struct amdgpu_bo_va_mapping *mapping; struct amdgpu_bo *bo; @@ -662,14 +663,14 @@ static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, struct amdgpu_ib *ib, r = amdgpu_cs_find_mapping(p, addr, &bo, &mapping); if (r) { - DRM_ERROR("Can't find BO for addr 0x%010Lx %d %d %d %d\n", + DRM_ERROR("Can't find BO for addr 0x%010llx %d %d %d %d\n", addr, lo, hi, size, index); return r; } if ((addr + (uint64_t)size) > (mapping->last + 1) * AMDGPU_GPU_PAGE_SIZE) { - DRM_ERROR("BO too small for addr 0x%010Lx %d %d\n", + DRM_ERROR("BO too small for addr 0x%010llx %d %d\n", addr, lo, hi); return -EINVAL; } @@ -692,12 +693,12 @@ static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, struct amdgpu_ib *ib, * @allocated: allocated a new handle? * * Validates the handle and return the found session index or -EINVAL - * we we don't have another free session index. + * we don't have another free session index. */ static int amdgpu_vce_validate_handle(struct amdgpu_cs_parser *p, uint32_t handle, uint32_t *allocated) { - unsigned i; + unsigned int i; /* validate the handle */ for (i = 0; i < AMDGPU_MAX_VCE_HANDLES; ++i) { @@ -735,14 +736,14 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, struct amdgpu_job *job, struct amdgpu_ib *ib) { - unsigned fb_idx = 0, bs_idx = 0; + unsigned int fb_idx = 0, bs_idx = 0; int session_idx = -1; uint32_t destroyed = 0; uint32_t created = 0; uint32_t allocated = 0; uint32_t tmp, handle = 0; uint32_t *size = &tmp; - unsigned idx; + unsigned int idx; int i, r = 0; job->vm = NULL; @@ -1084,7 +1085,7 @@ void amdgpu_vce_ring_emit_ib(struct amdgpu_ring *ring, * */ void amdgpu_vce_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq, - unsigned flags) + unsigned int flags) { WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT); @@ -1106,7 +1107,7 @@ int amdgpu_vce_ring_test_ring(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; uint32_t rptr; - unsigned i; + unsigned int i; int r, timeout = adev->usec_timeout; /* skip ring test for sriov*/ @@ -1171,7 +1172,7 @@ error: enum amdgpu_ring_priority_level amdgpu_vce_get_ring_prio(int ring) { - switch(ring) { + switch (ring) { case 0: return AMDGPU_RING_PRIO_0; case 1: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index e63fcc58e8e0..acbef1a24b9c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -56,6 +56,7 @@ #define FIRMWARE_VCN_3_1_2 "amdgpu/vcn_3_1_2.bin" #define FIRMWARE_VCN4_0_0 "amdgpu/vcn_4_0_0.bin" #define FIRMWARE_VCN4_0_2 "amdgpu/vcn_4_0_2.bin" +#define FIRMWARE_VCN4_0_3 "amdgpu/vcn_4_0_3.bin" #define FIRMWARE_VCN4_0_4 "amdgpu/vcn_4_0_4.bin" MODULE_FIRMWARE(FIRMWARE_RAVEN); @@ -77,6 +78,7 @@ MODULE_FIRMWARE(FIRMWARE_YELLOW_CARP); MODULE_FIRMWARE(FIRMWARE_VCN_3_1_2); MODULE_FIRMWARE(FIRMWARE_VCN4_0_0); MODULE_FIRMWARE(FIRMWARE_VCN4_0_2); +MODULE_FIRMWARE(FIRMWARE_VCN4_0_3); MODULE_FIRMWARE(FIRMWARE_VCN4_0_4); static void amdgpu_vcn_idle_work_handler(struct work_struct *work); @@ -167,7 +169,7 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev) if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) bo_size += AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8); - if (adev->ip_versions[UVD_HWIP][0] >= IP_VERSION(4, 0, 0)){ + if (adev->ip_versions[UVD_HWIP][0] >= IP_VERSION(4, 0, 0)) { fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)); log_offset = offsetof(struct amdgpu_vcn4_fw_shared, fw_log); } else { @@ -233,11 +235,11 @@ int amdgpu_vcn_sw_fini(struct amdgpu_device *adev) if (adev->vcn.harvest_config & (1 << j)) continue; - if (adev->vcn.indirect_sram) { - amdgpu_bo_free_kernel(&adev->vcn.inst[j].dpg_sram_bo, - &adev->vcn.inst[j].dpg_sram_gpu_addr, - (void **)&adev->vcn.inst[j].dpg_sram_cpu_addr); - } + amdgpu_bo_free_kernel( + &adev->vcn.inst[j].dpg_sram_bo, + &adev->vcn.inst[j].dpg_sram_gpu_addr, + (void **)&adev->vcn.inst[j].dpg_sram_cpu_addr); + kvfree(adev->vcn.inst[j].saved_bo); amdgpu_bo_free_kernel(&adev->vcn.inst[j].vcpu_bo, @@ -274,20 +276,19 @@ bool amdgpu_vcn_is_disabled_vcn(struct amdgpu_device *adev, enum vcn_ring_type t bool ret = false; int vcn_config = adev->vcn.vcn_config[vcn_instance]; - if ((type == VCN_ENCODE_RING) && (vcn_config & VCN_BLOCK_ENCODE_DISABLE_MASK)) { + if ((type == VCN_ENCODE_RING) && (vcn_config & VCN_BLOCK_ENCODE_DISABLE_MASK)) ret = true; - } else if ((type == VCN_DECODE_RING) && (vcn_config & VCN_BLOCK_DECODE_DISABLE_MASK)) { + else if ((type == VCN_DECODE_RING) && (vcn_config & VCN_BLOCK_DECODE_DISABLE_MASK)) ret = true; - } else if ((type == VCN_UNIFIED_RING) && (vcn_config & VCN_BLOCK_QUEUE_DISABLE_MASK)) { + else if ((type == VCN_UNIFIED_RING) && (vcn_config & VCN_BLOCK_QUEUE_DISABLE_MASK)) ret = true; - } return ret; } int amdgpu_vcn_suspend(struct amdgpu_device *adev) { - unsigned size; + unsigned int size; void *ptr; int i, idx; @@ -316,7 +317,7 @@ int amdgpu_vcn_suspend(struct amdgpu_device *adev) int amdgpu_vcn_resume(struct amdgpu_device *adev) { - unsigned size; + unsigned int size; void *ptr; int i, idx; @@ -338,7 +339,7 @@ int amdgpu_vcn_resume(struct amdgpu_device *adev) adev->vcn.inst[i].saved_bo = NULL; } else { const struct common_firmware_header *hdr; - unsigned offset; + unsigned int offset; hdr = (const struct common_firmware_header *)adev->vcn.fw->data; if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) { @@ -369,9 +370,8 @@ static void amdgpu_vcn_idle_work_handler(struct work_struct *work) if (adev->vcn.harvest_config & (1 << j)) continue; - for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + for (i = 0; i < adev->vcn.num_enc_rings; ++i) fence[j] += amdgpu_fence_count_emitted(&adev->vcn.inst[j].ring_enc[i]); - } if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { struct dpg_pause_state new_state; @@ -458,7 +458,7 @@ int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; uint32_t tmp = 0; - unsigned i; + unsigned int i; int r; /* VCN in SRIOV does not support direct register read/write */ @@ -795,7 +795,7 @@ int amdgpu_vcn_enc_ring_test_ring(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; uint32_t rptr; - unsigned i; + unsigned int i; int r; if (amdgpu_sriov_vf(adev)) @@ -993,11 +993,14 @@ error: int amdgpu_vcn_unified_ring_test_ib(struct amdgpu_ring *ring, long timeout) { + struct amdgpu_device *adev = ring->adev; long r; - r = amdgpu_vcn_enc_ring_test_ib(ring, timeout); - if (r) - goto error; + if (adev->ip_versions[UVD_HWIP][0] != IP_VERSION(4, 0, 3)) { + r = amdgpu_vcn_enc_ring_test_ib(ring, timeout); + if (r) + goto error; + } r = amdgpu_vcn_dec_sw_ring_test_ib(ring, timeout); @@ -1007,7 +1010,7 @@ error: enum amdgpu_ring_priority_level amdgpu_vcn_get_enc_ring_prio(int ring) { - switch(ring) { + switch (ring) { case 0: return AMDGPU_RING_PRIO_0; case 1: @@ -1026,6 +1029,7 @@ void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev) if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { const struct common_firmware_header *hdr; + hdr = (const struct common_firmware_header *)adev->vcn.fw->data; for (i = 0; i < adev->vcn.num_vcn_inst; i++) { @@ -1041,6 +1045,9 @@ void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev) adev->firmware.ucode[idx].fw = adev->vcn.fw; adev->firmware.fw_size += ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE); + + if (adev->ip_versions[UVD_HWIP][0] == IP_VERSION(4, 0, 3)) + break; } dev_info(adev->dev, "Will use PSP to load VCN firmware\n"); } @@ -1051,7 +1058,7 @@ void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev) */ #if defined(CONFIG_DEBUG_FS) static ssize_t amdgpu_debugfs_vcn_fwlog_read(struct file *f, char __user *buf, - size_t size, loff_t *pos) + size_t size, loff_t *pos) { struct amdgpu_vcn_inst *vcn; void *log_buf; @@ -1097,7 +1104,7 @@ static ssize_t amdgpu_debugfs_vcn_fwlog_read(struct file *f, char __user *buf, if (read_pos == AMDGPU_VCNFW_LOG_SIZE) read_pos = plog->header_size; if (read_num[i] == copy_to_user((buf + read_bytes), - (log_buf + read_pos), read_num[i])) + (log_buf + read_pos), read_num[i])) return -EFAULT; read_bytes += read_num[i]; @@ -1118,7 +1125,7 @@ static const struct file_operations amdgpu_debugfs_vcnfwlog_fops = { #endif void amdgpu_debugfs_vcn_fwlog_init(struct amdgpu_device *adev, uint8_t i, - struct amdgpu_vcn_inst *vcn) + struct amdgpu_vcn_inst *vcn) { #if defined(CONFIG_DEBUG_FS) struct drm_minor *minor = adev_to_drm(adev)->primary; @@ -1126,7 +1133,7 @@ void amdgpu_debugfs_vcn_fwlog_init(struct amdgpu_device *adev, uint8_t i, char name[32]; sprintf(name, "amdgpu_vcn_%d_fwlog", i); - debugfs_create_file_size(name, S_IFREG | S_IRUGO, root, vcn, + debugfs_create_file_size(name, S_IFREG | 0444, root, vcn, &amdgpu_debugfs_vcnfwlog_fops, AMDGPU_VCNFW_LOG_SIZE); #endif @@ -1140,7 +1147,7 @@ void amdgpu_vcn_fwlog_init(struct amdgpu_vcn_inst *vcn) uint64_t fw_log_gpu_addr = vcn->fw_shared.gpu_addr + vcn->fw_shared.mem_size; volatile struct amdgpu_vcn_fwlog *log_buf = fw_log_cpu_addr; volatile struct amdgpu_fw_shared_fw_logging *fw_log = vcn->fw_shared.cpu_addr - + vcn->fw_shared.log_offset; + + vcn->fw_shared.log_offset; *flag |= cpu_to_le32(AMDGPU_VCN_FW_LOGGING_FLAG); fw_log->is_enabled = 1; fw_log->addr_lo = cpu_to_le32(fw_log_gpu_addr & 0xFFFFFFFF); @@ -1181,6 +1188,31 @@ int amdgpu_vcn_process_poison_irq(struct amdgpu_device *adev, return 0; } +int amdgpu_vcn_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block) +{ + int r, i; + + r = amdgpu_ras_block_late_init(adev, ras_block); + if (r) + return r; + + if (amdgpu_ras_is_supported(adev, ras_block->block)) { + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + if (adev->vcn.harvest_config & (1 << i)) + continue; + + r = amdgpu_irq_get(adev, &adev->vcn.inst[i].ras_poison_irq, 0); + if (r) + goto late_fini; + } + } + return 0; + +late_fini: + amdgpu_ras_block_late_fini(adev, ras_block); + return r; +} + int amdgpu_vcn_ras_sw_init(struct amdgpu_device *adev) { int err; @@ -1202,7 +1234,7 @@ int amdgpu_vcn_ras_sw_init(struct amdgpu_device *adev) adev->vcn.ras_if = &ras->ras_block.ras_comm; if (!ras->ras_block.ras_late_init) - ras->ras_block.ras_late_init = amdgpu_ras_block_late_init; + ras->ras_block.ras_late_init = amdgpu_vcn_ras_late_init; return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index c730949ece7d..92d5534df5f4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -32,7 +32,7 @@ #define AMDGPU_VCN_FIRMWARE_OFFSET 256 #define AMDGPU_VCN_MAX_ENC_RINGS 3 -#define AMDGPU_MAX_VCN_INSTANCES 2 +#define AMDGPU_MAX_VCN_INSTANCES 4 #define AMDGPU_MAX_VCN_ENC_RINGS AMDGPU_VCN_MAX_ENC_RINGS * AMDGPU_MAX_VCN_INSTANCES #define AMDGPU_VCN_HARVEST_VCN0 (1 << 0) @@ -141,18 +141,23 @@ RREG32_SOC15(VCN, inst_idx, mmUVD_DPG_LMA_DATA); \ }) -#define WREG32_SOC15_DPG_MODE(inst_idx, offset, value, mask_en, indirect) \ - do { \ - if (!indirect) { \ - WREG32_SOC15(VCN, inst_idx, mmUVD_DPG_LMA_DATA, value); \ - WREG32_SOC15(VCN, inst_idx, mmUVD_DPG_LMA_CTL, \ - (0x1 << UVD_DPG_LMA_CTL__READ_WRITE__SHIFT | \ - mask_en << UVD_DPG_LMA_CTL__MASK_EN__SHIFT | \ - offset << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT)); \ - } else { \ - *adev->vcn.inst[inst_idx].dpg_sram_curr_addr++ = offset; \ - *adev->vcn.inst[inst_idx].dpg_sram_curr_addr++ = value; \ - } \ +#define WREG32_SOC15_DPG_MODE(inst_idx, offset, value, mask_en, indirect) \ + do { \ + if (!indirect) { \ + WREG32_SOC15(VCN, GET_INST(VCN, inst_idx), \ + mmUVD_DPG_LMA_DATA, value); \ + WREG32_SOC15( \ + VCN, GET_INST(VCN, inst_idx), \ + mmUVD_DPG_LMA_CTL, \ + (0x1 << UVD_DPG_LMA_CTL__READ_WRITE__SHIFT | \ + mask_en << UVD_DPG_LMA_CTL__MASK_EN__SHIFT | \ + offset << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT)); \ + } else { \ + *adev->vcn.inst[inst_idx].dpg_sram_curr_addr++ = \ + offset; \ + *adev->vcn.inst[inst_idx].dpg_sram_curr_addr++ = \ + value; \ + } \ } while (0) #define AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE (1 << 2) @@ -234,6 +239,7 @@ struct amdgpu_vcn_inst { struct amdgpu_ring ring_enc[AMDGPU_VCN_MAX_ENC_RINGS]; atomic_t sched_score; struct amdgpu_irq_src irq; + struct amdgpu_irq_src ras_poison_irq; struct amdgpu_vcn_reg external; struct amdgpu_bo *dpg_sram_bo; struct dpg_pause_state pause_state; @@ -242,6 +248,7 @@ struct amdgpu_vcn_inst { uint32_t *dpg_sram_curr_addr; atomic_t dpg_enc_submission_cnt; struct amdgpu_vcn_fw_shared fw_shared; + uint8_t aid_id; }; struct amdgpu_vcn_ras { @@ -271,6 +278,9 @@ struct amdgpu_vcn { struct ras_common_if *ras_if; struct amdgpu_vcn_ras *ras; + + uint16_t inst_mask; + uint8_t num_inst_per_aid; }; struct amdgpu_fw_shared_rb_ptrs_struct { @@ -400,6 +410,8 @@ void amdgpu_debugfs_vcn_fwlog_init(struct amdgpu_device *adev, int amdgpu_vcn_process_poison_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry); +int amdgpu_vcn_ras_late_init(struct amdgpu_device *adev, + struct ras_common_if *ras_block); int amdgpu_vcn_ras_sw_init(struct amdgpu_device *adev); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index f2e2cbaa7fde..25b4d7f0bd35 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -56,7 +56,8 @@ void amdgpu_virt_init_setting(struct amdgpu_device *adev) /* enable virtual display */ if (adev->asic_type != CHIP_ALDEBARAN && - adev->asic_type != CHIP_ARCTURUS) { + adev->asic_type != CHIP_ARCTURUS && + ((adev->pdev->class >> 8) != PCI_CLASS_ACCELERATOR_PROCESSING)) { if (adev->mode_info.num_crtc == 0) adev->mode_info.num_crtc = 1; adev->enable_virtual_display = true; @@ -65,16 +66,19 @@ void amdgpu_virt_init_setting(struct amdgpu_device *adev) adev->cg_flags = 0; adev->pg_flags = 0; - /* enable mcbp for sriov asic_type before soc21 */ - amdgpu_mcbp = (adev->asic_type < CHIP_IP_DISCOVERY) ? 1 : 0; + /* enable mcbp for sriov */ + amdgpu_mcbp = 1; + /* Reduce kcq number to 2 to reduce latency */ + if (amdgpu_num_kcq == -1) + amdgpu_num_kcq = 2; } void amdgpu_virt_kiq_reg_write_reg_wait(struct amdgpu_device *adev, uint32_t reg0, uint32_t reg1, uint32_t ref, uint32_t mask) { - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; struct amdgpu_ring *ring = &kiq->ring; signed long r, cnt = 0; unsigned long flags; @@ -557,7 +561,6 @@ static void amdgpu_virt_populate_vf2pf_ucode_info(struct amdgpu_device *adev) POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_RLC_SRLS, adev->gfx.rlc_srls_fw_version); POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_MEC, adev->gfx.mec_fw_version); POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_MEC2, adev->gfx.mec2_fw_version); - POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_IMU, adev->gfx.imu_fw_version); POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_SOS, adev->psp.sos.fw_version); POPULATE_UCODE_INFO(vf2pf_info, AMD_SRIOV_UCODE_ID_ASD, adev->psp.asd_context.bin_desc.fw_version); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 3c0310576b3b..143d11afe0e5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -267,6 +267,32 @@ static void amdgpu_vm_bo_done(struct amdgpu_vm_bo_base *vm_bo) } /** + * amdgpu_vm_bo_reset_state_machine - reset the vm_bo state machine + * @vm: the VM which state machine to reset + * + * Move all vm_bo object in the VM into a state where they will be updated + * again during validation. + */ +static void amdgpu_vm_bo_reset_state_machine(struct amdgpu_vm *vm) +{ + struct amdgpu_vm_bo_base *vm_bo, *tmp; + + spin_lock(&vm->status_lock); + list_splice_init(&vm->done, &vm->invalidated); + list_for_each_entry(vm_bo, &vm->invalidated, vm_status) + vm_bo->moved = true; + list_for_each_entry_safe(vm_bo, tmp, &vm->idle, vm_status) { + struct amdgpu_bo *bo = vm_bo->bo; + + if (!bo || bo->tbo.type != ttm_bo_type_kernel) + list_move(&vm_bo->vm_status, &vm_bo->vm->moved); + else if (bo->parent) + list_move(&vm_bo->vm_status, &vm_bo->vm->relocated); + } + spin_unlock(&vm->status_lock); +} + +/** * amdgpu_vm_bo_base_init - Adds bo to the list of bos associated with the vm * * @base: base structure for tracking BO usage in a VM @@ -351,6 +377,58 @@ void amdgpu_vm_move_to_lru_tail(struct amdgpu_device *adev, spin_unlock(&adev->mman.bdev.lru_lock); } +/* Create scheduler entities for page table updates */ +static int amdgpu_vm_init_entities(struct amdgpu_device *adev, + struct amdgpu_vm *vm) +{ + int r; + + r = drm_sched_entity_init(&vm->immediate, DRM_SCHED_PRIORITY_NORMAL, + adev->vm_manager.vm_pte_scheds, + adev->vm_manager.vm_pte_num_scheds, NULL); + if (r) + goto error; + + return drm_sched_entity_init(&vm->delayed, DRM_SCHED_PRIORITY_NORMAL, + adev->vm_manager.vm_pte_scheds, + adev->vm_manager.vm_pte_num_scheds, NULL); + +error: + drm_sched_entity_destroy(&vm->immediate); + return r; +} + +/* Destroy the entities for page table updates again */ +static void amdgpu_vm_fini_entities(struct amdgpu_vm *vm) +{ + drm_sched_entity_destroy(&vm->immediate); + drm_sched_entity_destroy(&vm->delayed); +} + +/** + * amdgpu_vm_generation - return the page table re-generation counter + * @adev: the amdgpu_device + * @vm: optional VM to check, might be NULL + * + * Returns a page table re-generation token to allow checking if submissions + * are still valid to use this VM. The VM parameter might be NULL in which case + * just the VRAM lost counter will be used. + */ +uint64_t amdgpu_vm_generation(struct amdgpu_device *adev, struct amdgpu_vm *vm) +{ + uint64_t result = (u64)atomic_read(&adev->vram_lost_counter) << 32; + + if (!vm) + return result; + + result += vm->generation; + /* Add one if the page tables will be re-generated on next CS */ + if (drm_sched_entity_error(&vm->delayed)) + ++result; + + return result; +} + /** * amdgpu_vm_validate_pt_bos - validate the page table BOs * @@ -373,6 +451,15 @@ int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm, struct amdgpu_bo *bo; int r; + if (drm_sched_entity_error(&vm->delayed)) { + ++vm->generation; + amdgpu_vm_bo_reset_state_machine(vm); + amdgpu_vm_fini_entities(vm); + r = amdgpu_vm_init_entities(adev, vm); + if (r) + return r; + } + spin_lock(&vm->status_lock); while (!list_empty(&vm->evicted)) { bo_base = list_first_entry(&vm->evicted, @@ -920,42 +1007,51 @@ error_unlock: return r; } +static void amdgpu_vm_bo_get_memory(struct amdgpu_bo_va *bo_va, + struct amdgpu_mem_stats *stats) +{ + struct amdgpu_vm *vm = bo_va->base.vm; + struct amdgpu_bo *bo = bo_va->base.bo; + + if (!bo) + return; + + /* + * For now ignore BOs which are currently locked and potentially + * changing their location. + */ + if (bo->tbo.base.resv != vm->root.bo->tbo.base.resv && + !dma_resv_trylock(bo->tbo.base.resv)) + return; + + amdgpu_bo_get_memory(bo, stats); + if (bo->tbo.base.resv != vm->root.bo->tbo.base.resv) + dma_resv_unlock(bo->tbo.base.resv); +} + void amdgpu_vm_get_memory(struct amdgpu_vm *vm, struct amdgpu_mem_stats *stats) { struct amdgpu_bo_va *bo_va, *tmp; spin_lock(&vm->status_lock); - list_for_each_entry_safe(bo_va, tmp, &vm->idle, base.vm_status) { - if (!bo_va->base.bo) - continue; - amdgpu_bo_get_memory(bo_va->base.bo, stats); - } - list_for_each_entry_safe(bo_va, tmp, &vm->evicted, base.vm_status) { - if (!bo_va->base.bo) - continue; - amdgpu_bo_get_memory(bo_va->base.bo, stats); - } - list_for_each_entry_safe(bo_va, tmp, &vm->relocated, base.vm_status) { - if (!bo_va->base.bo) - continue; - amdgpu_bo_get_memory(bo_va->base.bo, stats); - } - list_for_each_entry_safe(bo_va, tmp, &vm->moved, base.vm_status) { - if (!bo_va->base.bo) - continue; - amdgpu_bo_get_memory(bo_va->base.bo, stats); - } - list_for_each_entry_safe(bo_va, tmp, &vm->invalidated, base.vm_status) { - if (!bo_va->base.bo) - continue; - amdgpu_bo_get_memory(bo_va->base.bo, stats); - } - list_for_each_entry_safe(bo_va, tmp, &vm->done, base.vm_status) { - if (!bo_va->base.bo) - continue; - amdgpu_bo_get_memory(bo_va->base.bo, stats); - } + list_for_each_entry_safe(bo_va, tmp, &vm->idle, base.vm_status) + amdgpu_vm_bo_get_memory(bo_va, stats); + + list_for_each_entry_safe(bo_va, tmp, &vm->evicted, base.vm_status) + amdgpu_vm_bo_get_memory(bo_va, stats); + + list_for_each_entry_safe(bo_va, tmp, &vm->relocated, base.vm_status) + amdgpu_vm_bo_get_memory(bo_va, stats); + + list_for_each_entry_safe(bo_va, tmp, &vm->moved, base.vm_status) + amdgpu_vm_bo_get_memory(bo_va, stats); + + list_for_each_entry_safe(bo_va, tmp, &vm->invalidated, base.vm_status) + amdgpu_vm_bo_get_memory(bo_va, stats); + + list_for_each_entry_safe(bo_va, tmp, &vm->done, base.vm_status) + amdgpu_vm_bo_get_memory(bo_va, stats); spin_unlock(&vm->status_lock); } @@ -1358,6 +1454,7 @@ struct amdgpu_bo_va *amdgpu_vm_bo_add(struct amdgpu_device *adev, amdgpu_vm_bo_base_init(&bo_va->base, vm, bo); bo_va->ref_count = 1; + bo_va->last_pt_update = dma_fence_get_stub(); INIT_LIST_HEAD(&bo_va->valids); INIT_LIST_HEAD(&bo_va->invalids); @@ -1433,14 +1530,14 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev, uint64_t eaddr; /* validate the parameters */ - if (saddr & ~PAGE_MASK || offset & ~PAGE_MASK || - size == 0 || size & ~PAGE_MASK) + if (saddr & ~PAGE_MASK || offset & ~PAGE_MASK || size & ~PAGE_MASK) + return -EINVAL; + if (saddr + size <= saddr || offset + size <= offset) return -EINVAL; /* make sure object fit at this offset */ eaddr = saddr + size - 1; - if (saddr >= eaddr || - (bo && offset + size > amdgpu_bo_size(bo)) || + if ((bo && offset + size > amdgpu_bo_size(bo)) || (eaddr >= adev->vm_manager.max_pfn << AMDGPU_GPU_PAGE_SHIFT)) return -EINVAL; @@ -1499,14 +1596,14 @@ int amdgpu_vm_bo_replace_map(struct amdgpu_device *adev, int r; /* validate the parameters */ - if (saddr & ~PAGE_MASK || offset & ~PAGE_MASK || - size == 0 || size & ~PAGE_MASK) + if (saddr & ~PAGE_MASK || offset & ~PAGE_MASK || size & ~PAGE_MASK) + return -EINVAL; + if (saddr + size <= saddr || offset + size <= offset) return -EINVAL; /* make sure object fit at this offset */ eaddr = saddr + size - 1; - if (saddr >= eaddr || - (bo && offset + size > amdgpu_bo_size(bo)) || + if ((bo && offset + size > amdgpu_bo_size(bo)) || (eaddr >= adev->vm_manager.max_pfn << AMDGPU_GPU_PAGE_SHIFT)) return -EINVAL; @@ -2038,19 +2135,10 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm) INIT_LIST_HEAD(&vm->pt_freed); INIT_WORK(&vm->pt_free_work, amdgpu_vm_pt_free_work); - /* create scheduler entities for page table updates */ - r = drm_sched_entity_init(&vm->immediate, DRM_SCHED_PRIORITY_NORMAL, - adev->vm_manager.vm_pte_scheds, - adev->vm_manager.vm_pte_num_scheds, NULL); + r = amdgpu_vm_init_entities(adev, vm); if (r) return r; - r = drm_sched_entity_init(&vm->delayed, DRM_SCHED_PRIORITY_NORMAL, - adev->vm_manager.vm_pte_scheds, - adev->vm_manager.vm_pte_num_scheds, NULL); - if (r) - goto error_free_immediate; - vm->pte_support_ats = false; vm->is_compute_context = false; @@ -2067,9 +2155,11 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm) vm->update_funcs = &amdgpu_vm_cpu_funcs; else vm->update_funcs = &amdgpu_vm_sdma_funcs; - vm->last_update = NULL; + + vm->last_update = dma_fence_get_stub(); vm->last_unlocked = dma_fence_get_stub(); vm->last_tlb_flush = dma_fence_get_stub(); + vm->generation = 0; mutex_init(&vm->eviction_lock); vm->evicting = false; @@ -2110,10 +2200,7 @@ error_free_root: error_free_delayed: dma_fence_put(vm->last_tlb_flush); dma_fence_put(vm->last_unlocked); - drm_sched_entity_destroy(&vm->delayed); - -error_free_immediate: - drm_sched_entity_destroy(&vm->immediate); + amdgpu_vm_fini_entities(vm); return r; } @@ -2192,7 +2279,7 @@ int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm) goto unreserve_bo; dma_fence_put(vm->last_update); - vm->last_update = NULL; + vm->last_update = dma_fence_get_stub(); vm->is_compute_context = true; /* Free the shadow bo for compute VM */ @@ -2266,8 +2353,7 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) amdgpu_bo_unref(&root); WARN_ON(vm->root.bo); - drm_sched_entity_destroy(&vm->immediate); - drm_sched_entity_destroy(&vm->delayed); + amdgpu_vm_fini_entities(vm); if (!RB_EMPTY_ROOT(&vm->va.rb_root)) { dev_err(adev->dev, "still active bo inside vm\n"); @@ -2282,8 +2368,14 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) } dma_fence_put(vm->last_update); - for (i = 0; i < AMDGPU_MAX_VMHUBS; i++) - amdgpu_vmid_free_reserved(adev, vm, i); + + for (i = 0; i < AMDGPU_MAX_VMHUBS; i++) { + if (vm->reserved_vmid[i]) { + amdgpu_vmid_free_reserved(adev, i); + vm->reserved_vmid[i] = false; + } + } + } /** @@ -2366,18 +2458,25 @@ int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) union drm_amdgpu_vm *args = data; struct amdgpu_device *adev = drm_to_adev(dev); struct amdgpu_fpriv *fpriv = filp->driver_priv; - int r; + + /* No valid flags defined yet */ + if (args->in.flags) + return -EINVAL; switch (args->in.op) { case AMDGPU_VM_OP_RESERVE_VMID: /* We only have requirement to reserve vmid from gfxhub */ - r = amdgpu_vmid_alloc_reserved(adev, &fpriv->vm, - AMDGPU_GFXHUB_0); - if (r) - return r; + if (!fpriv->vm.reserved_vmid[AMDGPU_GFXHUB(0)]) { + amdgpu_vmid_alloc_reserved(adev, AMDGPU_GFXHUB(0)); + fpriv->vm.reserved_vmid[AMDGPU_GFXHUB(0)] = true; + } + break; case AMDGPU_VM_OP_UNRESERVE_VMID: - amdgpu_vmid_free_reserved(adev, &fpriv->vm, AMDGPU_GFXHUB_0); + if (fpriv->vm.reserved_vmid[AMDGPU_GFXHUB(0)]) { + amdgpu_vmid_free_reserved(adev, AMDGPU_GFXHUB(0)); + fpriv->vm.reserved_vmid[AMDGPU_GFXHUB(0)] = false; + } break; default: return -EINVAL; @@ -2432,6 +2531,9 @@ void amdgpu_vm_set_task_info(struct amdgpu_vm *vm) * amdgpu_vm_handle_fault - graceful handling of VM faults. * @adev: amdgpu device pointer * @pasid: PASID of the VM + * @vmid: VMID, only used for GFX 9.4.3. + * @node_id: Node_id received in IH cookie. Only applicable for + * GFX 9.4.3. * @addr: Address of the fault * @write_fault: true is write fault, false is read fault * @@ -2439,7 +2541,8 @@ void amdgpu_vm_set_task_info(struct amdgpu_vm *vm) * shouldn't be reported any more. */ bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid, - uint64_t addr, bool write_fault) + u32 vmid, u32 node_id, uint64_t addr, + bool write_fault) { bool is_compute_context = false; struct amdgpu_bo *root; @@ -2463,8 +2566,8 @@ bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid, addr /= AMDGPU_GPU_PAGE_SIZE; - if (is_compute_context && - !svm_range_restore_pages(adev, pasid, addr, write_fault)) { + if (is_compute_context && !svm_range_restore_pages(adev, pasid, vmid, + node_id, addr, write_fault)) { amdgpu_bo_unref(&root); return true; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h index 6f085f0b4ef3..9c85d494f2a2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h @@ -111,11 +111,14 @@ struct amdgpu_mem_stats; /* Reserve 4MB VRAM for page tables */ #define AMDGPU_VM_RESERVED_VRAM (8ULL << 20) -/* max number of VMHUB */ -#define AMDGPU_MAX_VMHUBS 3 -#define AMDGPU_GFXHUB_0 0 -#define AMDGPU_MMHUB_0 1 -#define AMDGPU_MMHUB_1 2 +/* + * max number of VMHUB + * layout: max 8 GFXHUB + 4 MMHUB0 + 1 MMHUB1 + */ +#define AMDGPU_MAX_VMHUBS 13 +#define AMDGPU_GFXHUB(x) (x) +#define AMDGPU_MMHUB0(x) (8 + x) +#define AMDGPU_MMHUB1(x) (8 + 4 + x) /* Reserve 2MB at top/bottom of address space for kernel use */ #define AMDGPU_VA_RESERVED_SIZE (2ULL << 20) @@ -292,6 +295,9 @@ struct amdgpu_vm { atomic64_t tlb_seq; struct dma_fence *last_tlb_flush; + /* How many times we had to re-generate the page tables */ + uint64_t generation; + /* Last unlocked submission to the scheduler entities */ struct dma_fence *last_unlocked; @@ -326,6 +332,9 @@ struct amdgpu_vm { struct ttm_lru_bulk_move lru_bulk_move; /* Flag to indicate if VM is used for compute */ bool is_compute_context; + + /* Memory partition number, -1 means any partition */ + int8_t mem_id; }; struct amdgpu_vm_manager { @@ -391,6 +400,7 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, struct list_head *validated, struct amdgpu_bo_list_entry *entry); bool amdgpu_vm_ready(struct amdgpu_vm *vm); +uint64_t amdgpu_vm_generation(struct amdgpu_device *adev, struct amdgpu_vm *vm); int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm, int (*callback)(void *p, struct amdgpu_bo *bo), void *param); @@ -452,7 +462,8 @@ void amdgpu_vm_check_compute_bug(struct amdgpu_device *adev); void amdgpu_vm_get_task_info(struct amdgpu_device *adev, u32 pasid, struct amdgpu_task_info *task_info); bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid, - uint64_t addr, bool write_fault); + u32 vmid, u32 node_id, uint64_t addr, + bool write_fault); void amdgpu_vm_set_task_info(struct amdgpu_vm *vm); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c index df63dc3bca18..dea1a64be44d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c @@ -502,6 +502,7 @@ exit: int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm, int level, bool immediate, struct amdgpu_bo_vm **vmbo) { + struct amdgpu_fpriv *fpriv = container_of(vm, struct amdgpu_fpriv, vm); struct amdgpu_bo_param bp; struct amdgpu_bo *bo; struct dma_resv *resv; @@ -512,7 +513,12 @@ int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm, bp.size = amdgpu_vm_pt_size(adev, level); bp.byte_align = AMDGPU_GPU_PAGE_SIZE; - bp.domain = AMDGPU_GEM_DOMAIN_VRAM; + + if (!adev->gmc.is_app_apu) + bp.domain = AMDGPU_GEM_DOMAIN_VRAM; + else + bp.domain = AMDGPU_GEM_DOMAIN_GTT; + bp.domain = amdgpu_bo_get_preferred_domain(adev, bp.domain); bp.flags = AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS | AMDGPU_GEM_CREATE_CPU_GTT_USWC; @@ -529,6 +535,8 @@ int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm, bp.type = ttm_bo_type_kernel; bp.no_wait_gpu = immediate; + bp.xcp_id_plus1 = fpriv->xcp_id == ~0 ? 0 : fpriv->xcp_id + 1; + if (vm->root.bo) bp.resv = vm->root.bo->tbo.base.resv; @@ -553,6 +561,7 @@ int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm, bp.type = ttm_bo_type_kernel; bp.resv = bo->tbo.base.resv; bp.bo_ptr_size = sizeof(struct amdgpu_bo); + bp.xcp_id_plus1 = fpriv->xcp_id == ~0 ? 0 : fpriv->xcp_id + 1; r = amdgpu_bo_create(adev, &bp, &(*vmbo)->shadow); @@ -564,7 +573,6 @@ int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm, return r; } - (*vmbo)->shadow->parent = amdgpu_bo_ref(bo); amdgpu_bo_add_to_shadow_list(*vmbo); return 0; @@ -781,13 +789,14 @@ static void amdgpu_vm_pte_update_flags(struct amdgpu_vm_update_params *params, uint64_t pe, uint64_t addr, unsigned int count, uint32_t incr, uint64_t flags) - { + struct amdgpu_device *adev = params->adev; + if (level != AMDGPU_VM_PTB) { flags |= AMDGPU_PDE_PTE; - amdgpu_gmc_get_vm_pde(params->adev, level, &addr, &flags); + amdgpu_gmc_get_vm_pde(adev, level, &addr, &flags); - } else if (params->adev->asic_type >= CHIP_VEGA10 && + } else if (adev->asic_type >= CHIP_VEGA10 && !(flags & AMDGPU_PTE_VALID) && !(flags & AMDGPU_PTE_PRT)) { @@ -795,6 +804,21 @@ static void amdgpu_vm_pte_update_flags(struct amdgpu_vm_update_params *params, flags |= AMDGPU_PTE_EXECUTABLE; } + /* APUs mapping system memory may need different MTYPEs on different + * NUMA nodes. Only do this for contiguous ranges that can be assumed + * to be on the same NUMA node. + */ + if ((flags & AMDGPU_PTE_SYSTEM) && (adev->flags & AMD_IS_APU) && + adev->gmc.gmc_funcs->override_vm_pte_flags && + num_possible_nodes() > 1) { + if (!params->pages_addr) + amdgpu_gmc_override_vm_pte_flags(adev, params->vm, + addr, &flags); + else + dev_dbg(adev->dev, + "override_vm_pte_flags skipped: non-contiguous\n"); + } + params->vm->update_funcs->update(params, pt, pe, addr, count, incr, flags); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c index 43d6a9d6a538..c7085a747b03 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c @@ -370,6 +370,45 @@ out: return ret; } +static void amdgpu_dummy_vram_mgr_debug(struct ttm_resource_manager *man, + struct drm_printer *printer) +{ + DRM_DEBUG_DRIVER("Dummy vram mgr debug\n"); +} + +static bool amdgpu_dummy_vram_mgr_compatible(struct ttm_resource_manager *man, + struct ttm_resource *res, + const struct ttm_place *place, + size_t size) +{ + DRM_DEBUG_DRIVER("Dummy vram mgr compatible\n"); + return false; +} + +static bool amdgpu_dummy_vram_mgr_intersects(struct ttm_resource_manager *man, + struct ttm_resource *res, + const struct ttm_place *place, + size_t size) +{ + DRM_DEBUG_DRIVER("Dummy vram mgr intersects\n"); + return true; +} + +static void amdgpu_dummy_vram_mgr_del(struct ttm_resource_manager *man, + struct ttm_resource *res) +{ + DRM_DEBUG_DRIVER("Dummy vram mgr deleted\n"); +} + +static int amdgpu_dummy_vram_mgr_new(struct ttm_resource_manager *man, + struct ttm_buffer_object *tbo, + const struct ttm_place *place, + struct ttm_resource **res) +{ + DRM_DEBUG_DRIVER("Dummy vram mgr new\n"); + return -ENOSPC; +} + /** * amdgpu_vram_mgr_new - allocate new ranges * @@ -800,7 +839,7 @@ static void amdgpu_vram_mgr_debug(struct ttm_resource_manager *man, { struct amdgpu_vram_mgr *mgr = to_vram_mgr(man); struct drm_buddy *mm = &mgr->mm; - struct drm_buddy_block *block; + struct amdgpu_vram_reservation *rsv; drm_printf(printer, " vis usage:%llu\n", amdgpu_vram_mgr_vis_usage(mgr)); @@ -812,11 +851,20 @@ static void amdgpu_vram_mgr_debug(struct ttm_resource_manager *man, drm_buddy_print(mm, printer); drm_printf(printer, "reserved:\n"); - list_for_each_entry(block, &mgr->reserved_pages, link) - drm_buddy_block_print(mm, block, printer); + list_for_each_entry(rsv, &mgr->reserved_pages, blocks) + drm_printf(printer, "%#018llx-%#018llx: %llu\n", + rsv->start, rsv->start + rsv->size, rsv->size); mutex_unlock(&mgr->lock); } +static const struct ttm_resource_manager_func amdgpu_dummy_vram_mgr_func = { + .alloc = amdgpu_dummy_vram_mgr_new, + .free = amdgpu_dummy_vram_mgr_del, + .intersects = amdgpu_dummy_vram_mgr_intersects, + .compatible = amdgpu_dummy_vram_mgr_compatible, + .debug = amdgpu_dummy_vram_mgr_debug +}; + static const struct ttm_resource_manager_func amdgpu_vram_mgr_func = { .alloc = amdgpu_vram_mgr_new, .free = amdgpu_vram_mgr_del, @@ -841,17 +889,22 @@ int amdgpu_vram_mgr_init(struct amdgpu_device *adev) ttm_resource_manager_init(man, &adev->mman.bdev, adev->gmc.real_vram_size); - man->func = &amdgpu_vram_mgr_func; - - err = drm_buddy_init(&mgr->mm, man->size, PAGE_SIZE); - if (err) - return err; - mutex_init(&mgr->lock); INIT_LIST_HEAD(&mgr->reservations_pending); INIT_LIST_HEAD(&mgr->reserved_pages); mgr->default_page_size = PAGE_SIZE; + if (!adev->gmc.is_app_apu) { + man->func = &amdgpu_vram_mgr_func; + + err = drm_buddy_init(&mgr->mm, man->size, PAGE_SIZE); + if (err) + return err; + } else { + man->func = &amdgpu_dummy_vram_mgr_func; + DRM_INFO("Setup dummy vram mgr\n"); + } + ttm_set_driver_manager(&adev->mman.bdev, TTM_PL_VRAM, &mgr->manager); ttm_resource_manager_set_used(man, true); return 0; @@ -886,7 +939,8 @@ void amdgpu_vram_mgr_fini(struct amdgpu_device *adev) drm_buddy_free_list(&mgr->mm, &rsv->allocated); kfree(rsv); } - drm_buddy_fini(&mgr->mm); + if (!adev->gmc.is_app_apu) + drm_buddy_fini(&mgr->mm); mutex_unlock(&mgr->lock); ttm_resource_manager_cleanup(man); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c new file mode 100644 index 000000000000..d733fa6e7477 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c @@ -0,0 +1,399 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#include "amdgpu.h" +#include "amdgpu_xcp.h" +#include "amdgpu_drv.h" + +#include <drm/drm_drv.h> +#include "../amdxcp/amdgpu_xcp_drv.h" + +static int __amdgpu_xcp_run(struct amdgpu_xcp_mgr *xcp_mgr, + struct amdgpu_xcp_ip *xcp_ip, int xcp_state) +{ + int (*run_func)(void *handle, uint32_t inst_mask); + int ret = 0; + + if (!xcp_ip || !xcp_ip->valid || !xcp_ip->ip_funcs) + return 0; + + run_func = NULL; + + switch (xcp_state) { + case AMDGPU_XCP_PREPARE_SUSPEND: + run_func = xcp_ip->ip_funcs->prepare_suspend; + break; + case AMDGPU_XCP_SUSPEND: + run_func = xcp_ip->ip_funcs->suspend; + break; + case AMDGPU_XCP_PREPARE_RESUME: + run_func = xcp_ip->ip_funcs->prepare_resume; + break; + case AMDGPU_XCP_RESUME: + run_func = xcp_ip->ip_funcs->resume; + break; + } + + if (run_func) + ret = run_func(xcp_mgr->adev, xcp_ip->inst_mask); + + return ret; +} + +static int amdgpu_xcp_run_transition(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id, + int state) +{ + struct amdgpu_xcp_ip *xcp_ip; + struct amdgpu_xcp *xcp; + int i, ret; + + if (xcp_id >= MAX_XCP || !xcp_mgr->xcp[xcp_id].valid) + return -EINVAL; + + xcp = &xcp_mgr->xcp[xcp_id]; + for (i = 0; i < AMDGPU_XCP_MAX_BLOCKS; ++i) { + xcp_ip = &xcp->ip[i]; + ret = __amdgpu_xcp_run(xcp_mgr, xcp_ip, state); + if (ret) + break; + } + + return ret; +} + +int amdgpu_xcp_prepare_suspend(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id) +{ + return amdgpu_xcp_run_transition(xcp_mgr, xcp_id, + AMDGPU_XCP_PREPARE_SUSPEND); +} + +int amdgpu_xcp_suspend(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id) +{ + return amdgpu_xcp_run_transition(xcp_mgr, xcp_id, AMDGPU_XCP_SUSPEND); +} + +int amdgpu_xcp_prepare_resume(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id) +{ + return amdgpu_xcp_run_transition(xcp_mgr, xcp_id, + AMDGPU_XCP_PREPARE_RESUME); +} + +int amdgpu_xcp_resume(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id) +{ + return amdgpu_xcp_run_transition(xcp_mgr, xcp_id, AMDGPU_XCP_RESUME); +} + +static void __amdgpu_xcp_add_block(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id, + struct amdgpu_xcp_ip *ip) +{ + struct amdgpu_xcp *xcp; + + if (!ip) + return; + + xcp = &xcp_mgr->xcp[xcp_id]; + xcp->ip[ip->ip_id] = *ip; + xcp->ip[ip->ip_id].valid = true; + + xcp->valid = true; +} + +int amdgpu_xcp_init(struct amdgpu_xcp_mgr *xcp_mgr, int num_xcps, int mode) +{ + struct amdgpu_device *adev = xcp_mgr->adev; + struct amdgpu_xcp_ip ip; + uint8_t mem_id; + int i, j, ret; + + if (!num_xcps || num_xcps > MAX_XCP) + return -EINVAL; + + xcp_mgr->mode = mode; + + for (i = 0; i < MAX_XCP; ++i) + xcp_mgr->xcp[i].valid = false; + + for (i = 0; i < num_xcps; ++i) { + for (j = AMDGPU_XCP_GFXHUB; j < AMDGPU_XCP_MAX_BLOCKS; ++j) { + ret = xcp_mgr->funcs->get_ip_details(xcp_mgr, i, j, + &ip); + if (ret) + continue; + + __amdgpu_xcp_add_block(xcp_mgr, i, &ip); + } + + xcp_mgr->xcp[i].id = i; + + if (xcp_mgr->funcs->get_xcp_mem_id) { + ret = xcp_mgr->funcs->get_xcp_mem_id( + xcp_mgr, &xcp_mgr->xcp[i], &mem_id); + if (ret) + continue; + else + xcp_mgr->xcp[i].mem_id = mem_id; + } + } + + xcp_mgr->num_xcps = num_xcps; + amdgpu_xcp_update_partition_sched_list(adev); + + xcp_mgr->num_xcp_per_mem_partition = num_xcps / xcp_mgr->adev->gmc.num_mem_partitions; + return 0; +} + +int amdgpu_xcp_switch_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr, int mode) +{ + int ret, curr_mode, num_xcps = 0; + + if (!xcp_mgr || mode == AMDGPU_XCP_MODE_NONE) + return -EINVAL; + + if (xcp_mgr->mode == mode) + return 0; + + if (!xcp_mgr->funcs || !xcp_mgr->funcs->switch_partition_mode) + return 0; + + mutex_lock(&xcp_mgr->xcp_lock); + + curr_mode = xcp_mgr->mode; + /* State set to transient mode */ + xcp_mgr->mode = AMDGPU_XCP_MODE_TRANS; + + ret = xcp_mgr->funcs->switch_partition_mode(xcp_mgr, mode, &num_xcps); + + if (ret) { + /* Failed, get whatever mode it's at now */ + if (xcp_mgr->funcs->query_partition_mode) + xcp_mgr->mode = amdgpu_xcp_query_partition_mode( + xcp_mgr, AMDGPU_XCP_FL_LOCKED); + else + xcp_mgr->mode = curr_mode; + + goto out; + } + +out: + mutex_unlock(&xcp_mgr->xcp_lock); + + return ret; +} + +int amdgpu_xcp_query_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr, u32 flags) +{ + int mode; + + if (xcp_mgr->mode == AMDGPU_XCP_MODE_NONE) + return xcp_mgr->mode; + + if (!xcp_mgr->funcs || !xcp_mgr->funcs->query_partition_mode) + return xcp_mgr->mode; + + if (!(flags & AMDGPU_XCP_FL_LOCKED)) + mutex_lock(&xcp_mgr->xcp_lock); + mode = xcp_mgr->funcs->query_partition_mode(xcp_mgr); + if (xcp_mgr->mode != AMDGPU_XCP_MODE_TRANS && mode != xcp_mgr->mode) + dev_WARN( + xcp_mgr->adev->dev, + "Cached partition mode %d not matching with device mode %d", + xcp_mgr->mode, mode); + + if (!(flags & AMDGPU_XCP_FL_LOCKED)) + mutex_unlock(&xcp_mgr->xcp_lock); + + return mode; +} + +static int amdgpu_xcp_dev_alloc(struct amdgpu_device *adev) +{ + struct drm_device *p_ddev; + struct drm_device *ddev; + int i, ret; + + ddev = adev_to_drm(adev); + + for (i = 0; i < MAX_XCP; i++) { + ret = amdgpu_xcp_drm_dev_alloc(&p_ddev); + if (ret) + return ret; + + /* Redirect all IOCTLs to the primary device */ + adev->xcp_mgr->xcp[i].rdev = p_ddev->render->dev; + adev->xcp_mgr->xcp[i].pdev = p_ddev->primary->dev; + adev->xcp_mgr->xcp[i].driver = (struct drm_driver *)p_ddev->driver; + adev->xcp_mgr->xcp[i].vma_offset_manager = p_ddev->vma_offset_manager; + p_ddev->render->dev = ddev; + p_ddev->primary->dev = ddev; + p_ddev->vma_offset_manager = ddev->vma_offset_manager; + p_ddev->driver = &amdgpu_partition_driver; + adev->xcp_mgr->xcp[i].ddev = p_ddev; + } + + return 0; +} + +int amdgpu_xcp_mgr_init(struct amdgpu_device *adev, int init_mode, + int init_num_xcps, + struct amdgpu_xcp_mgr_funcs *xcp_funcs) +{ + struct amdgpu_xcp_mgr *xcp_mgr; + + if (!xcp_funcs || !xcp_funcs->switch_partition_mode || + !xcp_funcs->get_ip_details) + return -EINVAL; + + xcp_mgr = kzalloc(sizeof(*xcp_mgr), GFP_KERNEL); + + if (!xcp_mgr) + return -ENOMEM; + + xcp_mgr->adev = adev; + xcp_mgr->funcs = xcp_funcs; + xcp_mgr->mode = init_mode; + mutex_init(&xcp_mgr->xcp_lock); + + if (init_mode != AMDGPU_XCP_MODE_NONE) + amdgpu_xcp_init(xcp_mgr, init_num_xcps, init_mode); + + adev->xcp_mgr = xcp_mgr; + + return amdgpu_xcp_dev_alloc(adev); +} + +int amdgpu_xcp_get_partition(struct amdgpu_xcp_mgr *xcp_mgr, + enum AMDGPU_XCP_IP_BLOCK ip, int instance) +{ + struct amdgpu_xcp *xcp; + int i, id_mask = 0; + + if (ip >= AMDGPU_XCP_MAX_BLOCKS) + return -EINVAL; + + for (i = 0; i < xcp_mgr->num_xcps; ++i) { + xcp = &xcp_mgr->xcp[i]; + if ((xcp->valid) && (xcp->ip[ip].valid) && + (xcp->ip[ip].inst_mask & BIT(instance))) + id_mask |= BIT(i); + } + + if (!id_mask) + id_mask = -ENXIO; + + return id_mask; +} + +int amdgpu_xcp_get_inst_details(struct amdgpu_xcp *xcp, + enum AMDGPU_XCP_IP_BLOCK ip, + uint32_t *inst_mask) +{ + if (!xcp->valid || !inst_mask || !(xcp->ip[ip].valid)) + return -EINVAL; + + *inst_mask = xcp->ip[ip].inst_mask; + + return 0; +} + +int amdgpu_xcp_dev_register(struct amdgpu_device *adev, + const struct pci_device_id *ent) +{ + int i, ret; + + if (!adev->xcp_mgr) + return 0; + + for (i = 0; i < MAX_XCP; i++) { + ret = drm_dev_register(adev->xcp_mgr->xcp[i].ddev, ent->driver_data); + if (ret) + return ret; + } + + return 0; +} + +void amdgpu_xcp_dev_unplug(struct amdgpu_device *adev) +{ + struct drm_device *p_ddev; + int i; + + if (!adev->xcp_mgr) + return; + + for (i = 0; i < MAX_XCP; i++) { + p_ddev = adev->xcp_mgr->xcp[i].ddev; + drm_dev_unplug(p_ddev); + p_ddev->render->dev = adev->xcp_mgr->xcp[i].rdev; + p_ddev->primary->dev = adev->xcp_mgr->xcp[i].pdev; + p_ddev->driver = adev->xcp_mgr->xcp[i].driver; + p_ddev->vma_offset_manager = adev->xcp_mgr->xcp[i].vma_offset_manager; + } +} + +int amdgpu_xcp_open_device(struct amdgpu_device *adev, + struct amdgpu_fpriv *fpriv, + struct drm_file *file_priv) +{ + int i; + + if (!adev->xcp_mgr) + return 0; + + fpriv->xcp_id = ~0; + for (i = 0; i < MAX_XCP; ++i) { + if (!adev->xcp_mgr->xcp[i].ddev) + break; + + if (file_priv->minor == adev->xcp_mgr->xcp[i].ddev->render) { + if (adev->xcp_mgr->xcp[i].valid == FALSE) { + dev_err(adev->dev, "renderD%d partition %d not valid!", + file_priv->minor->index, i); + return -ENOENT; + } + dev_dbg(adev->dev, "renderD%d partition %d opened!", + file_priv->minor->index, i); + fpriv->xcp_id = i; + break; + } + } + + fpriv->vm.mem_id = fpriv->xcp_id == ~0 ? -1 : + adev->xcp_mgr->xcp[fpriv->xcp_id].mem_id; + return 0; +} + +void amdgpu_xcp_release_sched(struct amdgpu_device *adev, + struct amdgpu_ctx_entity *entity) +{ + struct drm_gpu_scheduler *sched; + struct amdgpu_ring *ring; + + if (!adev->xcp_mgr) + return; + + sched = entity->entity.rq->sched; + if (sched->ready) { + ring = to_amdgpu_ring(entity->entity.rq->sched); + atomic_dec(&adev->xcp_mgr->xcp[ring->xcp_id].ref_cnt); + } +} + diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.h new file mode 100644 index 000000000000..0f8026d64ea5 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.h @@ -0,0 +1,182 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef AMDGPU_XCP_H +#define AMDGPU_XCP_H + +#include <linux/pci.h> +#include <linux/xarray.h> + +#include "amdgpu_ctx.h" + +#define MAX_XCP 8 + +#define AMDGPU_XCP_MODE_NONE -1 +#define AMDGPU_XCP_MODE_TRANS -2 + +#define AMDGPU_XCP_FL_NONE 0 +#define AMDGPU_XCP_FL_LOCKED (1 << 0) + +struct amdgpu_fpriv; + +enum AMDGPU_XCP_IP_BLOCK { + AMDGPU_XCP_GFXHUB, + AMDGPU_XCP_GFX, + AMDGPU_XCP_SDMA, + AMDGPU_XCP_VCN, + AMDGPU_XCP_MAX_BLOCKS +}; + +enum AMDGPU_XCP_STATE { + AMDGPU_XCP_PREPARE_SUSPEND, + AMDGPU_XCP_SUSPEND, + AMDGPU_XCP_PREPARE_RESUME, + AMDGPU_XCP_RESUME, +}; + +struct amdgpu_xcp_ip_funcs { + int (*prepare_suspend)(void *handle, uint32_t inst_mask); + int (*suspend)(void *handle, uint32_t inst_mask); + int (*prepare_resume)(void *handle, uint32_t inst_mask); + int (*resume)(void *handle, uint32_t inst_mask); +}; + +struct amdgpu_xcp_ip { + struct amdgpu_xcp_ip_funcs *ip_funcs; + uint32_t inst_mask; + + enum AMDGPU_XCP_IP_BLOCK ip_id; + bool valid; +}; + +struct amdgpu_xcp { + struct amdgpu_xcp_ip ip[AMDGPU_XCP_MAX_BLOCKS]; + + uint8_t id; + uint8_t mem_id; + bool valid; + atomic_t ref_cnt; + struct drm_device *ddev; + struct drm_device *rdev; + struct drm_device *pdev; + struct drm_driver *driver; + struct drm_vma_offset_manager *vma_offset_manager; + struct amdgpu_sched gpu_sched[AMDGPU_HW_IP_NUM][AMDGPU_RING_PRIO_MAX]; +}; + +struct amdgpu_xcp_mgr { + struct amdgpu_device *adev; + struct mutex xcp_lock; + struct amdgpu_xcp_mgr_funcs *funcs; + + struct amdgpu_xcp xcp[MAX_XCP]; + uint8_t num_xcps; + int8_t mode; + + /* Used to determine KFD memory size limits per XCP */ + unsigned int num_xcp_per_mem_partition; +}; + +struct amdgpu_xcp_mgr_funcs { + int (*switch_partition_mode)(struct amdgpu_xcp_mgr *xcp_mgr, int mode, + int *num_xcps); + int (*query_partition_mode)(struct amdgpu_xcp_mgr *xcp_mgr); + int (*get_ip_details)(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id, + enum AMDGPU_XCP_IP_BLOCK ip_id, + struct amdgpu_xcp_ip *ip); + int (*get_xcp_mem_id)(struct amdgpu_xcp_mgr *xcp_mgr, + struct amdgpu_xcp *xcp, uint8_t *mem_id); + + int (*prepare_suspend)(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id); + int (*suspend)(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id); + int (*prepare_resume)(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id); + int (*resume)(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id); + int (*select_scheds)(struct amdgpu_device *adev, + u32 hw_ip, u32 hw_prio, struct amdgpu_fpriv *fpriv, + unsigned int *num_scheds, struct drm_gpu_scheduler ***scheds); + int (*update_partition_sched_list)(struct amdgpu_device *adev); +}; + +int amdgpu_xcp_prepare_suspend(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id); +int amdgpu_xcp_suspend(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id); +int amdgpu_xcp_prepare_resume(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id); +int amdgpu_xcp_resume(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id); + +int amdgpu_xcp_mgr_init(struct amdgpu_device *adev, int init_mode, + int init_xcps, struct amdgpu_xcp_mgr_funcs *xcp_funcs); +int amdgpu_xcp_init(struct amdgpu_xcp_mgr *xcp_mgr, int num_xcps, int mode); +int amdgpu_xcp_query_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr, u32 flags); +int amdgpu_xcp_switch_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr, int mode); +int amdgpu_xcp_get_partition(struct amdgpu_xcp_mgr *xcp_mgr, + enum AMDGPU_XCP_IP_BLOCK ip, int instance); + +int amdgpu_xcp_get_inst_details(struct amdgpu_xcp *xcp, + enum AMDGPU_XCP_IP_BLOCK ip, + uint32_t *inst_mask); + +int amdgpu_xcp_dev_register(struct amdgpu_device *adev, + const struct pci_device_id *ent); +void amdgpu_xcp_dev_unplug(struct amdgpu_device *adev); +int amdgpu_xcp_open_device(struct amdgpu_device *adev, + struct amdgpu_fpriv *fpriv, + struct drm_file *file_priv); +void amdgpu_xcp_release_sched(struct amdgpu_device *adev, + struct amdgpu_ctx_entity *entity); + +#define amdgpu_xcp_select_scheds(adev, e, c, d, x, y) \ + ((adev)->xcp_mgr && (adev)->xcp_mgr->funcs && \ + (adev)->xcp_mgr->funcs->select_scheds ? \ + (adev)->xcp_mgr->funcs->select_scheds((adev), (e), (c), (d), (x), (y)) : -ENOENT) +#define amdgpu_xcp_update_partition_sched_list(adev) \ + ((adev)->xcp_mgr && (adev)->xcp_mgr->funcs && \ + (adev)->xcp_mgr->funcs->update_partition_sched_list ? \ + (adev)->xcp_mgr->funcs->update_partition_sched_list(adev) : 0) + +static inline int amdgpu_xcp_get_num_xcp(struct amdgpu_xcp_mgr *xcp_mgr) +{ + if (!xcp_mgr) + return 1; + else + return xcp_mgr->num_xcps; +} + +static inline struct amdgpu_xcp * +amdgpu_get_next_xcp(struct amdgpu_xcp_mgr *xcp_mgr, int *from) +{ + if (!xcp_mgr) + return NULL; + + while (*from < MAX_XCP) { + if (xcp_mgr->xcp[*from].valid) + return &xcp_mgr->xcp[*from]; + ++(*from); + } + + return NULL; +} + +#define for_each_xcp(xcp_mgr, xcp, i) \ + for (i = 0, xcp = amdgpu_get_next_xcp(xcp_mgr, &i); xcp; \ + xcp = amdgpu_get_next_xcp(xcp_mgr, &i)) + +#endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c index 439925477fb8..03dc59cbe8aa 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c @@ -325,6 +325,36 @@ static ssize_t amdgpu_xgmi_show_device_id(struct device *dev, } +static ssize_t amdgpu_xgmi_show_num_hops(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = drm_to_adev(ddev); + struct psp_xgmi_topology_info *top = &adev->psp.xgmi_context.top_info; + int i; + + for (i = 0; i < top->num_nodes; i++) + sprintf(buf + 3 * i, "%02x ", top->nodes[i].num_hops); + + return sysfs_emit(buf, "%s\n", buf); +} + +static ssize_t amdgpu_xgmi_show_num_links(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = drm_to_adev(ddev); + struct psp_xgmi_topology_info *top = &adev->psp.xgmi_context.top_info; + int i; + + for (i = 0; i < top->num_nodes; i++) + sprintf(buf + 3 * i, "%02x ", top->nodes[i].num_links); + + return sysfs_emit(buf, "%s\n", buf); +} + #define AMDGPU_XGMI_SET_FICAA(o) ((o) | 0x456801) static ssize_t amdgpu_xgmi_show_error(struct device *dev, struct device_attribute *attr, @@ -361,6 +391,8 @@ static ssize_t amdgpu_xgmi_show_error(struct device *dev, static DEVICE_ATTR(xgmi_device_id, S_IRUGO, amdgpu_xgmi_show_device_id, NULL); static DEVICE_ATTR(xgmi_error, S_IRUGO, amdgpu_xgmi_show_error, NULL); +static DEVICE_ATTR(xgmi_num_hops, S_IRUGO, amdgpu_xgmi_show_num_hops, NULL); +static DEVICE_ATTR(xgmi_num_links, S_IRUGO, amdgpu_xgmi_show_num_links, NULL); static int amdgpu_xgmi_sysfs_add_dev_info(struct amdgpu_device *adev, struct amdgpu_hive_info *hive) @@ -380,6 +412,15 @@ static int amdgpu_xgmi_sysfs_add_dev_info(struct amdgpu_device *adev, if (ret) pr_err("failed to create xgmi_error\n"); + /* Create xgmi num hops file */ + ret = device_create_file(adev->dev, &dev_attr_xgmi_num_hops); + if (ret) + pr_err("failed to create xgmi_num_hops\n"); + + /* Create xgmi num links file */ + ret = device_create_file(adev->dev, &dev_attr_xgmi_num_links); + if (ret) + pr_err("failed to create xgmi_num_links\n"); /* Create sysfs link to hive info folder on the first device */ if (hive->kobj.parent != (&adev->dev->kobj)) { @@ -407,6 +448,9 @@ remove_link: remove_file: device_remove_file(adev->dev, &dev_attr_xgmi_device_id); + device_remove_file(adev->dev, &dev_attr_xgmi_error); + device_remove_file(adev->dev, &dev_attr_xgmi_num_hops); + device_remove_file(adev->dev, &dev_attr_xgmi_num_links); success: return ret; @@ -420,6 +464,8 @@ static void amdgpu_xgmi_sysfs_rem_dev_info(struct amdgpu_device *adev, device_remove_file(adev->dev, &dev_attr_xgmi_device_id); device_remove_file(adev->dev, &dev_attr_xgmi_error); + device_remove_file(adev->dev, &dev_attr_xgmi_num_hops); + device_remove_file(adev->dev, &dev_attr_xgmi_num_links); if (hive->kobj.parent != (&adev->dev->kobj)) sysfs_remove_link(&adev->dev->kobj,"xgmi_hive_info"); @@ -1014,7 +1060,8 @@ static void amdgpu_xgmi_query_ras_error_count(struct amdgpu_device *adev, } /* Trigger XGMI/WAFL error */ -static int amdgpu_ras_error_inject_xgmi(struct amdgpu_device *adev, void *inject_if) +static int amdgpu_ras_error_inject_xgmi(struct amdgpu_device *adev, + void *inject_if, uint32_t instance_mask) { int ret = 0; struct ta_ras_trigger_error_input *block_info = @@ -1026,7 +1073,7 @@ static int amdgpu_ras_error_inject_xgmi(struct amdgpu_device *adev, void *injec if (amdgpu_dpm_allow_xgmi_power_down(adev, false)) dev_warn(adev->dev, "Failed to disallow XGMI power down"); - ret = psp_ras_trigger_error(&adev->psp, block_info); + ret = psp_ras_trigger_error(&adev->psp, block_info, instance_mask); if (amdgpu_ras_intr_triggered()) return ret; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h b/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h index 24d42d24e6a0..104a5ad8397d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h @@ -70,7 +70,6 @@ enum amd_sriov_ucode_engine_id { AMD_SRIOV_UCODE_ID_RLC_SRLS, AMD_SRIOV_UCODE_ID_MEC, AMD_SRIOV_UCODE_ID_MEC2, - AMD_SRIOV_UCODE_ID_IMU, AMD_SRIOV_UCODE_ID_SOS, AMD_SRIOV_UCODE_ID_ASD, AMD_SRIOV_UCODE_ID_TA_RAS, diff --git a/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c new file mode 100644 index 000000000000..16471b81a1f5 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c @@ -0,0 +1,658 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#include "amdgpu.h" +#include "soc15.h" + +#include "soc15_common.h" +#include "amdgpu_xcp.h" +#include "gfx_v9_4_3.h" +#include "gfxhub_v1_2.h" +#include "sdma_v4_4_2.h" + +#define XCP_INST_MASK(num_inst, xcp_id) \ + (num_inst ? GENMASK(num_inst - 1, 0) << (xcp_id * num_inst) : 0) + +#define AMDGPU_XCP_OPS_KFD (1 << 0) + +void aqua_vanjaram_doorbell_index_init(struct amdgpu_device *adev) +{ + int i; + + adev->doorbell_index.kiq = AMDGPU_DOORBELL_LAYOUT1_KIQ_START; + + adev->doorbell_index.mec_ring0 = AMDGPU_DOORBELL_LAYOUT1_MEC_RING_START; + + adev->doorbell_index.userqueue_start = AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_START; + adev->doorbell_index.userqueue_end = AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_END; + adev->doorbell_index.xcc_doorbell_range = AMDGPU_DOORBELL_LAYOUT1_XCC_RANGE; + + adev->doorbell_index.sdma_doorbell_range = 20; + for (i = 0; i < adev->sdma.num_instances; i++) + adev->doorbell_index.sdma_engine[i] = + AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START + + i * (adev->doorbell_index.sdma_doorbell_range >> 1); + + adev->doorbell_index.ih = AMDGPU_DOORBELL_LAYOUT1_IH; + adev->doorbell_index.vcn.vcn_ring0_1 = AMDGPU_DOORBELL_LAYOUT1_VCN_START; + + adev->doorbell_index.first_non_cp = AMDGPU_DOORBELL_LAYOUT1_FIRST_NON_CP; + adev->doorbell_index.last_non_cp = AMDGPU_DOORBELL_LAYOUT1_LAST_NON_CP; + + adev->doorbell_index.max_assignment = AMDGPU_DOORBELL_LAYOUT1_MAX_ASSIGNMENT << 1; +} + +static void aqua_vanjaram_set_xcp_id(struct amdgpu_device *adev, + uint32_t inst_idx, struct amdgpu_ring *ring) +{ + int xcp_id; + enum AMDGPU_XCP_IP_BLOCK ip_blk; + uint32_t inst_mask; + + ring->xcp_id = ~0; + if (adev->xcp_mgr->mode == AMDGPU_XCP_MODE_NONE) + return; + + inst_mask = 1 << inst_idx; + + switch (ring->funcs->type) { + case AMDGPU_HW_IP_GFX: + case AMDGPU_RING_TYPE_COMPUTE: + case AMDGPU_RING_TYPE_KIQ: + ip_blk = AMDGPU_XCP_GFX; + break; + case AMDGPU_RING_TYPE_SDMA: + ip_blk = AMDGPU_XCP_SDMA; + break; + case AMDGPU_RING_TYPE_VCN_ENC: + case AMDGPU_RING_TYPE_VCN_JPEG: + ip_blk = AMDGPU_XCP_VCN; + if (adev->xcp_mgr->mode == AMDGPU_CPX_PARTITION_MODE) + inst_mask = 1 << (inst_idx * 2); + break; + default: + DRM_ERROR("Not support ring type %d!", ring->funcs->type); + return; + } + + for (xcp_id = 0; xcp_id < adev->xcp_mgr->num_xcps; xcp_id++) { + if (adev->xcp_mgr->xcp[xcp_id].ip[ip_blk].inst_mask & inst_mask) { + ring->xcp_id = xcp_id; + break; + } + } +} + +static void aqua_vanjaram_xcp_gpu_sched_update( + struct amdgpu_device *adev, + struct amdgpu_ring *ring, + unsigned int sel_xcp_id) +{ + unsigned int *num_gpu_sched; + + num_gpu_sched = &adev->xcp_mgr->xcp[sel_xcp_id] + .gpu_sched[ring->funcs->type][ring->hw_prio].num_scheds; + adev->xcp_mgr->xcp[sel_xcp_id].gpu_sched[ring->funcs->type][ring->hw_prio] + .sched[(*num_gpu_sched)++] = &ring->sched; + DRM_DEBUG("%s :[%d] gpu_sched[%d][%d] = %d", ring->name, + sel_xcp_id, ring->funcs->type, + ring->hw_prio, *num_gpu_sched); +} + +static int aqua_vanjaram_xcp_sched_list_update( + struct amdgpu_device *adev) +{ + struct amdgpu_ring *ring; + int i; + + for (i = 0; i < MAX_XCP; i++) { + atomic_set(&adev->xcp_mgr->xcp[i].ref_cnt, 0); + memset(adev->xcp_mgr->xcp[i].gpu_sched, 0, sizeof(adev->xcp_mgr->xcp->gpu_sched)); + } + + if (adev->xcp_mgr->mode == AMDGPU_XCP_MODE_NONE) + return 0; + + for (i = 0; i < AMDGPU_MAX_RINGS; i++) { + ring = adev->rings[i]; + if (!ring || !ring->sched.ready) + continue; + + aqua_vanjaram_xcp_gpu_sched_update(adev, ring, ring->xcp_id); + + /* VCN is shared by two partitions under CPX MODE */ + if ((ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC || + ring->funcs->type == AMDGPU_RING_TYPE_VCN_JPEG) && + adev->xcp_mgr->mode == AMDGPU_CPX_PARTITION_MODE) + aqua_vanjaram_xcp_gpu_sched_update(adev, ring, ring->xcp_id + 1); + } + + return 0; +} + +static int aqua_vanjaram_update_partition_sched_list(struct amdgpu_device *adev) +{ + int i; + + for (i = 0; i < adev->num_rings; i++) { + struct amdgpu_ring *ring = adev->rings[i]; + + if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE || + ring->funcs->type == AMDGPU_RING_TYPE_KIQ) + aqua_vanjaram_set_xcp_id(adev, ring->xcc_id, ring); + else + aqua_vanjaram_set_xcp_id(adev, ring->me, ring); + } + + return aqua_vanjaram_xcp_sched_list_update(adev); +} + +static int aqua_vanjaram_select_scheds( + struct amdgpu_device *adev, + u32 hw_ip, + u32 hw_prio, + struct amdgpu_fpriv *fpriv, + unsigned int *num_scheds, + struct drm_gpu_scheduler ***scheds) +{ + u32 sel_xcp_id; + int i; + + if (fpriv->xcp_id == ~0) { + u32 least_ref_cnt = ~0; + + fpriv->xcp_id = 0; + for (i = 0; i < adev->xcp_mgr->num_xcps; i++) { + u32 total_ref_cnt; + + total_ref_cnt = atomic_read(&adev->xcp_mgr->xcp[i].ref_cnt); + if (total_ref_cnt < least_ref_cnt) { + fpriv->xcp_id = i; + least_ref_cnt = total_ref_cnt; + } + } + } + sel_xcp_id = fpriv->xcp_id; + + if (adev->xcp_mgr->xcp[sel_xcp_id].gpu_sched[hw_ip][hw_prio].num_scheds) { + *num_scheds = adev->xcp_mgr->xcp[fpriv->xcp_id].gpu_sched[hw_ip][hw_prio].num_scheds; + *scheds = adev->xcp_mgr->xcp[fpriv->xcp_id].gpu_sched[hw_ip][hw_prio].sched; + atomic_inc(&adev->xcp_mgr->xcp[sel_xcp_id].ref_cnt); + DRM_DEBUG("Selected partition #%d", sel_xcp_id); + } else { + DRM_ERROR("Failed to schedule partition #%d.", sel_xcp_id); + return -ENOENT; + } + + return 0; +} + +static int8_t aqua_vanjaram_logical_to_dev_inst(struct amdgpu_device *adev, + enum amd_hw_ip_block_type block, + int8_t inst) +{ + int8_t dev_inst; + + switch (block) { + case GC_HWIP: + case SDMA0_HWIP: + /* Both JPEG and VCN as JPEG is only alias of VCN */ + case VCN_HWIP: + dev_inst = adev->ip_map.dev_inst[block][inst]; + break; + default: + /* For rest of the IPs, no look up required. + * Assume 'logical instance == physical instance' for all configs. */ + dev_inst = inst; + break; + } + + return dev_inst; +} + +static uint32_t aqua_vanjaram_logical_to_dev_mask(struct amdgpu_device *adev, + enum amd_hw_ip_block_type block, + uint32_t mask) +{ + uint32_t dev_mask = 0; + int8_t log_inst, dev_inst; + + while (mask) { + log_inst = ffs(mask) - 1; + dev_inst = aqua_vanjaram_logical_to_dev_inst(adev, block, log_inst); + dev_mask |= (1 << dev_inst); + mask &= ~(1 << log_inst); + } + + return dev_mask; +} + +static void aqua_vanjaram_populate_ip_map(struct amdgpu_device *adev, + enum amd_hw_ip_block_type ip_block, + uint32_t inst_mask) +{ + int l = 0, i; + + while (inst_mask) { + i = ffs(inst_mask) - 1; + adev->ip_map.dev_inst[ip_block][l++] = i; + inst_mask &= ~(1 << i); + } + for (; l < HWIP_MAX_INSTANCE; l++) + adev->ip_map.dev_inst[ip_block][l] = -1; +} + +void aqua_vanjaram_ip_map_init(struct amdgpu_device *adev) +{ + u32 ip_map[][2] = { + { GC_HWIP, adev->gfx.xcc_mask }, + { SDMA0_HWIP, adev->sdma.sdma_mask }, + { VCN_HWIP, adev->vcn.inst_mask }, + }; + int i; + + for (i = 0; i < ARRAY_SIZE(ip_map); ++i) + aqua_vanjaram_populate_ip_map(adev, ip_map[i][0], ip_map[i][1]); + + adev->ip_map.logical_to_dev_inst = aqua_vanjaram_logical_to_dev_inst; + adev->ip_map.logical_to_dev_mask = aqua_vanjaram_logical_to_dev_mask; +} + +/* Fixed pattern for smn addressing on different AIDs: + * bit[34]: indicate cross AID access + * bit[33:32]: indicate target AID id + * AID id range is 0 ~ 3 as maximum AID number is 4. + */ +u64 aqua_vanjaram_encode_ext_smn_addressing(int ext_id) +{ + u64 ext_offset; + + /* local routing and bit[34:32] will be zeros */ + if (ext_id == 0) + return 0; + + /* Initiated from host, accessing to all non-zero aids are cross traffic */ + ext_offset = ((u64)(ext_id & 0x3) << 32) | (1ULL << 34); + + return ext_offset; +} + +static int aqua_vanjaram_query_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr) +{ + enum amdgpu_gfx_partition mode = AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE; + struct amdgpu_device *adev = xcp_mgr->adev; + + if (adev->nbio.funcs->get_compute_partition_mode) + mode = adev->nbio.funcs->get_compute_partition_mode(adev); + + return mode; +} + +static int __aqua_vanjaram_get_xcc_per_xcp(struct amdgpu_xcp_mgr *xcp_mgr, int mode) +{ + int num_xcc, num_xcc_per_xcp = 0; + + num_xcc = NUM_XCC(xcp_mgr->adev->gfx.xcc_mask); + + switch (mode) { + case AMDGPU_SPX_PARTITION_MODE: + num_xcc_per_xcp = num_xcc; + break; + case AMDGPU_DPX_PARTITION_MODE: + num_xcc_per_xcp = num_xcc / 2; + break; + case AMDGPU_TPX_PARTITION_MODE: + num_xcc_per_xcp = num_xcc / 3; + break; + case AMDGPU_QPX_PARTITION_MODE: + num_xcc_per_xcp = num_xcc / 4; + break; + case AMDGPU_CPX_PARTITION_MODE: + num_xcc_per_xcp = 1; + break; + } + + return num_xcc_per_xcp; +} + +static int __aqua_vanjaram_get_xcp_ip_info(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id, + enum AMDGPU_XCP_IP_BLOCK ip_id, + struct amdgpu_xcp_ip *ip) +{ + struct amdgpu_device *adev = xcp_mgr->adev; + int num_xcc_xcp, num_sdma_xcp, num_vcn_xcp; + int num_sdma, num_vcn; + + num_sdma = adev->sdma.num_instances; + num_vcn = adev->vcn.num_vcn_inst; + + switch (xcp_mgr->mode) { + case AMDGPU_SPX_PARTITION_MODE: + num_sdma_xcp = num_sdma; + num_vcn_xcp = num_vcn; + break; + case AMDGPU_DPX_PARTITION_MODE: + num_sdma_xcp = num_sdma / 2; + num_vcn_xcp = num_vcn / 2; + break; + case AMDGPU_TPX_PARTITION_MODE: + num_sdma_xcp = num_sdma / 3; + num_vcn_xcp = num_vcn / 3; + break; + case AMDGPU_QPX_PARTITION_MODE: + num_sdma_xcp = num_sdma / 4; + num_vcn_xcp = num_vcn / 4; + break; + case AMDGPU_CPX_PARTITION_MODE: + num_sdma_xcp = 2; + num_vcn_xcp = num_vcn ? 1 : 0; + break; + default: + return -EINVAL; + } + + num_xcc_xcp = adev->gfx.num_xcc_per_xcp; + + switch (ip_id) { + case AMDGPU_XCP_GFXHUB: + ip->inst_mask = XCP_INST_MASK(num_xcc_xcp, xcp_id); + ip->ip_funcs = &gfxhub_v1_2_xcp_funcs; + break; + case AMDGPU_XCP_GFX: + ip->inst_mask = XCP_INST_MASK(num_xcc_xcp, xcp_id); + ip->ip_funcs = &gfx_v9_4_3_xcp_funcs; + break; + case AMDGPU_XCP_SDMA: + ip->inst_mask = XCP_INST_MASK(num_sdma_xcp, xcp_id); + ip->ip_funcs = &sdma_v4_4_2_xcp_funcs; + break; + case AMDGPU_XCP_VCN: + ip->inst_mask = XCP_INST_MASK(num_vcn_xcp, xcp_id); + /* TODO : Assign IP funcs */ + break; + default: + return -EINVAL; + } + + ip->ip_id = ip_id; + + return 0; +} + +static enum amdgpu_gfx_partition +__aqua_vanjaram_get_auto_mode(struct amdgpu_xcp_mgr *xcp_mgr) +{ + struct amdgpu_device *adev = xcp_mgr->adev; + int num_xcc; + + num_xcc = NUM_XCC(xcp_mgr->adev->gfx.xcc_mask); + + if (adev->gmc.num_mem_partitions == 1) + return AMDGPU_SPX_PARTITION_MODE; + + if (adev->gmc.num_mem_partitions == num_xcc) + return AMDGPU_CPX_PARTITION_MODE; + + if (adev->gmc.num_mem_partitions == num_xcc / 2) + return (adev->flags & AMD_IS_APU) ? AMDGPU_TPX_PARTITION_MODE : + AMDGPU_QPX_PARTITION_MODE; + + if (adev->gmc.num_mem_partitions == 2 && !(adev->flags & AMD_IS_APU)) + return AMDGPU_DPX_PARTITION_MODE; + + return AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE; +} + +static bool __aqua_vanjaram_is_valid_mode(struct amdgpu_xcp_mgr *xcp_mgr, + enum amdgpu_gfx_partition mode) +{ + struct amdgpu_device *adev = xcp_mgr->adev; + int num_xcc, num_xccs_per_xcp; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + switch (mode) { + case AMDGPU_SPX_PARTITION_MODE: + return adev->gmc.num_mem_partitions == 1 && num_xcc > 0; + case AMDGPU_DPX_PARTITION_MODE: + return adev->gmc.num_mem_partitions != 8 && (num_xcc % 4) == 0; + case AMDGPU_TPX_PARTITION_MODE: + return (adev->gmc.num_mem_partitions == 1 || + adev->gmc.num_mem_partitions == 3) && + ((num_xcc % 3) == 0); + case AMDGPU_QPX_PARTITION_MODE: + num_xccs_per_xcp = num_xcc / 4; + return (adev->gmc.num_mem_partitions == 1 || + adev->gmc.num_mem_partitions == 4) && + (num_xccs_per_xcp >= 2); + case AMDGPU_CPX_PARTITION_MODE: + return ((num_xcc > 1) && + (adev->gmc.num_mem_partitions == 1 || adev->gmc.num_mem_partitions == 4) && + (num_xcc % adev->gmc.num_mem_partitions) == 0); + default: + return false; + } + + return false; +} + +static int __aqua_vanjaram_pre_partition_switch(struct amdgpu_xcp_mgr *xcp_mgr, u32 flags) +{ + /* TODO: + * Stop user queues and threads, and make sure GPU is empty of work. + */ + + if (flags & AMDGPU_XCP_OPS_KFD) + amdgpu_amdkfd_device_fini_sw(xcp_mgr->adev); + + return 0; +} + +static int __aqua_vanjaram_post_partition_switch(struct amdgpu_xcp_mgr *xcp_mgr, u32 flags) +{ + int ret = 0; + + if (flags & AMDGPU_XCP_OPS_KFD) { + amdgpu_amdkfd_device_probe(xcp_mgr->adev); + amdgpu_amdkfd_device_init(xcp_mgr->adev); + /* If KFD init failed, return failure */ + if (!xcp_mgr->adev->kfd.init_complete) + ret = -EIO; + } + + return ret; +} + +static int aqua_vanjaram_switch_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr, + int mode, int *num_xcps) +{ + int num_xcc_per_xcp, num_xcc, ret; + struct amdgpu_device *adev; + u32 flags = 0; + + adev = xcp_mgr->adev; + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + + if (mode == AMDGPU_AUTO_COMPUTE_PARTITION_MODE) { + mode = __aqua_vanjaram_get_auto_mode(xcp_mgr); + } else if (!__aqua_vanjaram_is_valid_mode(xcp_mgr, mode)) { + dev_err(adev->dev, + "Invalid compute partition mode requested, requested: %s, available memory partitions: %d", + amdgpu_gfx_compute_mode_desc(mode), adev->gmc.num_mem_partitions); + return -EINVAL; + } + + if (adev->kfd.init_complete) + flags |= AMDGPU_XCP_OPS_KFD; + + if (flags & AMDGPU_XCP_OPS_KFD) { + ret = amdgpu_amdkfd_check_and_lock_kfd(adev); + if (ret) + goto out; + } + + ret = __aqua_vanjaram_pre_partition_switch(xcp_mgr, flags); + if (ret) + goto unlock; + + num_xcc_per_xcp = __aqua_vanjaram_get_xcc_per_xcp(xcp_mgr, mode); + if (adev->gfx.funcs->switch_partition_mode) + adev->gfx.funcs->switch_partition_mode(xcp_mgr->adev, + num_xcc_per_xcp); + + /* Init info about new xcps */ + *num_xcps = num_xcc / num_xcc_per_xcp; + amdgpu_xcp_init(xcp_mgr, *num_xcps, mode); + + ret = __aqua_vanjaram_post_partition_switch(xcp_mgr, flags); +unlock: + if (flags & AMDGPU_XCP_OPS_KFD) + amdgpu_amdkfd_unlock_kfd(adev); +out: + return ret; +} + +static int __aqua_vanjaram_get_xcp_mem_id(struct amdgpu_device *adev, + int xcc_id, uint8_t *mem_id) +{ + /* memory/spatial modes validation check is already done */ + *mem_id = xcc_id / adev->gfx.num_xcc_per_xcp; + *mem_id /= adev->xcp_mgr->num_xcp_per_mem_partition; + + return 0; +} + +static int aqua_vanjaram_get_xcp_mem_id(struct amdgpu_xcp_mgr *xcp_mgr, + struct amdgpu_xcp *xcp, uint8_t *mem_id) +{ + struct amdgpu_numa_info numa_info; + struct amdgpu_device *adev; + uint32_t xcc_mask; + int r, i, xcc_id; + + adev = xcp_mgr->adev; + /* TODO: BIOS is not returning the right info now + * Check on this later + */ + /* + if (adev->gmc.gmc_funcs->query_mem_partition_mode) + mode = adev->gmc.gmc_funcs->query_mem_partition_mode(adev); + */ + if (adev->gmc.num_mem_partitions == 1) { + /* Only one range */ + *mem_id = 0; + return 0; + } + + r = amdgpu_xcp_get_inst_details(xcp, AMDGPU_XCP_GFX, &xcc_mask); + if (r || !xcc_mask) + return -EINVAL; + + xcc_id = ffs(xcc_mask) - 1; + if (!adev->gmc.is_app_apu) + return __aqua_vanjaram_get_xcp_mem_id(adev, xcc_id, mem_id); + + r = amdgpu_acpi_get_mem_info(adev, xcc_id, &numa_info); + + if (r) + return r; + + r = -EINVAL; + for (i = 0; i < adev->gmc.num_mem_partitions; ++i) { + if (adev->gmc.mem_partitions[i].numa.node == numa_info.nid) { + *mem_id = i; + r = 0; + break; + } + } + + return r; +} + +static int aqua_vanjaram_get_xcp_ip_details(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id, + enum AMDGPU_XCP_IP_BLOCK ip_id, + struct amdgpu_xcp_ip *ip) +{ + if (!ip) + return -EINVAL; + + return __aqua_vanjaram_get_xcp_ip_info(xcp_mgr, xcp_id, ip_id, ip); +} + +struct amdgpu_xcp_mgr_funcs aqua_vanjaram_xcp_funcs = { + .switch_partition_mode = &aqua_vanjaram_switch_partition_mode, + .query_partition_mode = &aqua_vanjaram_query_partition_mode, + .get_ip_details = &aqua_vanjaram_get_xcp_ip_details, + .get_xcp_mem_id = &aqua_vanjaram_get_xcp_mem_id, + .select_scheds = &aqua_vanjaram_select_scheds, + .update_partition_sched_list = &aqua_vanjaram_update_partition_sched_list +}; + +static int aqua_vanjaram_xcp_mgr_init(struct amdgpu_device *adev) +{ + int ret; + + ret = amdgpu_xcp_mgr_init(adev, AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE, 1, + &aqua_vanjaram_xcp_funcs); + if (ret) + return ret; + + /* TODO: Default memory node affinity init */ + + return ret; +} + +int aqua_vanjaram_init_soc_config(struct amdgpu_device *adev) +{ + u32 mask, inst_mask = adev->sdma.sdma_mask; + int ret, i; + + /* generally 1 AID supports 4 instances */ + adev->sdma.num_inst_per_aid = 4; + adev->sdma.num_instances = NUM_SDMA(adev->sdma.sdma_mask); + + adev->aid_mask = i = 1; + inst_mask >>= adev->sdma.num_inst_per_aid; + + for (mask = (1 << adev->sdma.num_inst_per_aid) - 1; inst_mask; + inst_mask >>= adev->sdma.num_inst_per_aid, ++i) { + if ((inst_mask & mask) == mask) + adev->aid_mask |= (1 << i); + } + + /* Harvest config is not used for aqua vanjaram. VCN and JPEGs will be + * addressed based on logical instance ids. + */ + adev->vcn.harvest_config = 0; + adev->vcn.num_inst_per_aid = 1; + adev->vcn.num_vcn_inst = hweight32(adev->vcn.inst_mask); + adev->jpeg.harvest_config = 0; + adev->jpeg.num_inst_per_aid = 1; + adev->jpeg.num_jpeg_inst = hweight32(adev->jpeg.inst_mask); + + ret = aqua_vanjaram_xcp_mgr_init(adev); + if (ret) + return ret; + + aqua_vanjaram_ip_map_init(adev); + + return 0; +} diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c index de6d10390ab2..5641cf05d856 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik.c +++ b/drivers/gpu/drm/amd/amdgpu/cik.c @@ -1141,12 +1141,12 @@ static uint32_t cik_get_register_value(struct amdgpu_device *adev, mutex_lock(&adev->grbm_idx_mutex); if (se_num != 0xffffffff || sh_num != 0xffffffff) - amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0); val = RREG32(reg_offset); if (se_num != 0xffffffff || sh_num != 0xffffffff) - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); return val; } else { diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c index 67d16236b216..52598fbc9b39 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c @@ -489,8 +489,6 @@ static int cik_sdma_gfx_resume(struct amdgpu_device *adev) #endif /* enable DMA IBs */ WREG32(mmSDMA0_GFX_IB_CNTL + sdma_offsets[i], ib_cntl); - - ring->sched.ready = true; } cik_sdma_enable(adev, true); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c index f5b5ce1051a2..be984f8c71c7 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c @@ -3490,7 +3490,7 @@ static int gfx_v10_0_get_cu_info(struct amdgpu_device *adev, struct amdgpu_cu_info *cu_info); static uint64_t gfx_v10_0_get_gpu_clock_counter(struct amdgpu_device *adev); static void gfx_v10_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, - u32 sh_num, u32 instance); + u32 sh_num, u32 instance, int xcc_id); static u32 gfx_v10_0_get_wgp_active_bitmap_per_sh(struct amdgpu_device *adev); static int gfx_v10_0_rlc_backdoor_autoload_buffer_init(struct amdgpu_device *adev); @@ -3568,7 +3568,7 @@ static void gfx10_kiq_unmap_queues(struct amdgpu_ring *kiq_ring, struct amdgpu_device *adev = kiq_ring->adev; uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0; - if (adev->enable_mes && !adev->gfx.kiq.ring.sched.ready) { + if (adev->enable_mes && !adev->gfx.kiq[0].ring.sched.ready) { amdgpu_mes_unmap_legacy_queue(adev, ring, action, gpu_addr, seq); return; } @@ -3636,7 +3636,7 @@ static const struct kiq_pm4_funcs gfx_v10_0_kiq_pm4_funcs = { static void gfx_v10_0_set_kiq_pm4_funcs(struct amdgpu_device *adev) { - adev->gfx.kiq.pmf = &gfx_v10_0_kiq_pm4_funcs; + adev->gfx.kiq[0].pmf = &gfx_v10_0_kiq_pm4_funcs; } static void gfx_v10_0_init_spm_golden_registers(struct amdgpu_device *adev) @@ -4219,7 +4219,7 @@ static int gfx_v10_0_mec_init(struct amdgpu_device *adev) const struct gfx_firmware_header_v1_0 *mec_hdr = NULL; - bitmap_zero(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES); + bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES); /* take ownership of the relevant compute queues */ amdgpu_gfx_compute_queue_acquire(adev); @@ -4291,7 +4291,7 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t wave, *(out++) = RREG32_SOC15(GC, 0, mmSQ_IND_DATA); } -static void gfx_v10_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) +static void gfx_v10_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) { /* in gfx10 the SIMD_ID is specified as part of the INSTANCE * field when performing a select_se_sh so it should be @@ -4318,7 +4318,7 @@ static void gfx_v10_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_MODE); } -static void gfx_v10_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd, +static void gfx_v10_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t start, uint32_t size, uint32_t *dst) { @@ -4329,7 +4329,7 @@ static void gfx_v10_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd, dst); } -static void gfx_v10_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd, +static void gfx_v10_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t thread, uint32_t start, uint32_t size, uint32_t *dst) @@ -4340,7 +4340,7 @@ static void gfx_v10_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd, } static void gfx_v10_0_select_me_pipe_q(struct amdgpu_device *adev, - u32 me, u32 pipe, u32 q, u32 vm) + u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id) { nv_grbm_select(adev, me, pipe, q, vm); } @@ -4461,7 +4461,7 @@ static int gfx_v10_0_gfx_ring_init(struct amdgpu_device *adev, int ring_id, ring->doorbell_index = adev->doorbell_index.gfx_ring0 << 1; else ring->doorbell_index = adev->doorbell_index.gfx_ring1 << 1; - ring->vm_hub = AMDGPU_GFXHUB_0; + ring->vm_hub = AMDGPU_GFXHUB(0); sprintf(ring->name, "gfx_%d.%d.%d", ring->me, ring->pipe, ring->queue); irq_type = AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP + ring->pipe; @@ -4490,7 +4490,7 @@ static int gfx_v10_0_compute_ring_init(struct amdgpu_device *adev, int ring_id, ring->doorbell_index = (adev->doorbell_index.mec_ring0 + ring_id) << 1; ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr + (ring_id * GFX10_MEC_HPD_SIZE); - ring->vm_hub = AMDGPU_GFXHUB_0; + ring->vm_hub = AMDGPU_GFXHUB(0); sprintf(ring->name, "comp_%d.%d.%d", ring->me, ring->pipe, ring->queue); irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP @@ -4550,7 +4550,7 @@ static int gfx_v10_0_sw_init(void *handle) /* KIQ event */ r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP, GFX_10_1__SRCID__CP_IB2_INTERRUPT_PKT, - &adev->gfx.kiq.irq); + &adev->gfx.kiq[0].irq); if (r) return r; @@ -4614,8 +4614,8 @@ static int gfx_v10_0_sw_init(void *handle) for (i = 0; i < adev->gfx.mec.num_mec; ++i) { for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) { for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) { - if (!amdgpu_gfx_is_mec_queue_enabled(adev, i, k, - j)) + if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i, + k, j)) continue; r = gfx_v10_0_compute_ring_init(adev, ring_id, @@ -4629,19 +4629,19 @@ static int gfx_v10_0_sw_init(void *handle) } if (!adev->enable_mes_kiq) { - r = amdgpu_gfx_kiq_init(adev, GFX10_MEC_HPD_SIZE); + r = amdgpu_gfx_kiq_init(adev, GFX10_MEC_HPD_SIZE, 0); if (r) { DRM_ERROR("Failed to init KIQ BOs!\n"); return r; } - kiq = &adev->gfx.kiq; - r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq); + kiq = &adev->gfx.kiq[0]; + r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq, 0); if (r) return r; } - r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v10_compute_mqd)); + r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v10_compute_mqd), 0); if (r) return r; @@ -4690,11 +4690,11 @@ static int gfx_v10_0_sw_fini(void *handle) for (i = 0; i < adev->gfx.num_compute_rings; i++) amdgpu_ring_fini(&adev->gfx.compute_ring[i]); - amdgpu_gfx_mqd_sw_fini(adev); + amdgpu_gfx_mqd_sw_fini(adev, 0); if (!adev->enable_mes_kiq) { - amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq.ring); - amdgpu_gfx_kiq_fini(adev); + amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq[0].ring); + amdgpu_gfx_kiq_fini(adev, 0); } gfx_v10_0_pfp_fini(adev); @@ -4712,7 +4712,7 @@ static int gfx_v10_0_sw_fini(void *handle) } static void gfx_v10_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, - u32 sh_num, u32 instance) + u32 sh_num, u32 instance, int xcc_id) { u32 data; @@ -4772,13 +4772,13 @@ static void gfx_v10_0_setup_rb(struct amdgpu_device *adev) (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 6))) && ((gfx_v10_3_get_disabled_sa(adev) >> bitmap) & 1)) continue; - gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff, 0); data = gfx_v10_0_get_rb_active_bitmap(adev); active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) * rb_bitmap_width_per_sh); } } - gfx_v10_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v10_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); adev->gfx.config.backend_enable_mask = active_rbs; @@ -4825,6 +4825,29 @@ static u32 gfx_v10_0_init_pa_sc_tile_steering_override(struct amdgpu_device *ade #define DEFAULT_SH_MEM_BASES (0x6000) +static void gfx_v10_0_debug_trap_config_init(struct amdgpu_device *adev, + uint32_t first_vmid, + uint32_t last_vmid) +{ + uint32_t data; + uint32_t trap_config_vmid_mask = 0; + int i; + + /* Calculate trap config vmid mask */ + for (i = first_vmid; i < last_vmid; i++) + trap_config_vmid_mask |= (1 << i); + + data = REG_SET_FIELD(0, SPI_GDBG_TRAP_CONFIG, + VMID_SEL, trap_config_vmid_mask); + data = REG_SET_FIELD(data, SPI_GDBG_TRAP_CONFIG, + TRAP_EN, 1); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_CONFIG), data); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0); + + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA0), 0); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA1), 0); +} + static void gfx_v10_0_init_compute_vmid(struct amdgpu_device *adev) { int i; @@ -4856,6 +4879,9 @@ static void gfx_v10_0_init_compute_vmid(struct amdgpu_device *adev) WREG32_SOC15_OFFSET(GC, 0, mmGDS_GWS_VMID0, i, 0); WREG32_SOC15_OFFSET(GC, 0, mmGDS_OA_VMID0, i, 0); } + + gfx_v10_0_debug_trap_config_init(adev, adev->vm_manager.first_kfd_vmid, + AMDGPU_NUM_VMID); } static void gfx_v10_0_init_gds_vmid(struct amdgpu_device *adev) @@ -4907,7 +4933,7 @@ static void gfx_v10_0_tcp_harvest(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff, 0); wgp_active_bitmap = gfx_v10_0_get_wgp_active_bitmap_per_sh(adev); /* * Set corresponding TCP bits for the inactive WGPs in @@ -4940,7 +4966,7 @@ static void gfx_v10_0_tcp_harvest(struct amdgpu_device *adev) } } - gfx_v10_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v10_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); } @@ -4978,7 +5004,7 @@ static void gfx_v10_0_constants_init(struct amdgpu_device *adev) /* XXX SH_MEM regs */ /* where to put LDS, scratch, GPUVM in FSA64 space */ mutex_lock(&adev->srbm_mutex); - for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB_0].num_ids; i++) { + for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB(0)].num_ids; i++) { nv_grbm_select(adev, 0, 0, 0, i); /* CP and shaders */ WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, DEFAULT_SH_MEM_CONFIG); @@ -6073,7 +6099,6 @@ static int gfx_v10_0_cp_gfx_resume(struct amdgpu_device *adev) u32 tmp; u32 rb_bufsz; u64 rb_addr, rptr_addr, wptr_gpu_addr; - u32 i; /* Set the write pointer delay */ WREG32_SOC15(GC, 0, mmCP_RB_WPTR_DELAY, 0); @@ -6168,11 +6193,6 @@ static int gfx_v10_0_cp_gfx_resume(struct amdgpu_device *adev) /* start the ring */ gfx_v10_0_cp_gfx_start(adev); - for (i = 0; i < adev->gfx.num_gfx_rings; i++) { - ring = &adev->gfx.gfx_ring[i]; - ring->sched.ready = true; - } - return 0; } @@ -6214,7 +6234,7 @@ static void gfx_v10_0_cp_compute_enable(struct amdgpu_device *adev, bool enable) CP_MEC_CNTL__MEC_ME2_HALT_MASK)); break; } - adev->gfx.kiq.ring.sched.ready = false; + adev->gfx.kiq[0].ring.sched.ready = false; } udelay(50); } @@ -6423,55 +6443,6 @@ static int gfx_v10_0_gfx_mqd_init(struct amdgpu_device *adev, void *m, return 0; } -#ifdef BRING_UP_DEBUG -static int gfx_v10_0_gfx_queue_init_register(struct amdgpu_ring *ring) -{ - struct amdgpu_device *adev = ring->adev; - struct v10_gfx_mqd *mqd = ring->mqd_ptr; - - /* set mmCP_GFX_HQD_WPTR/_HI to 0 */ - WREG32_SOC15(GC, 0, mmCP_GFX_HQD_WPTR, mqd->cp_gfx_hqd_wptr); - WREG32_SOC15(GC, 0, mmCP_GFX_HQD_WPTR_HI, mqd->cp_gfx_hqd_wptr_hi); - - /* set GFX_MQD_BASE */ - WREG32_SOC15(GC, 0, mmCP_MQD_BASE_ADDR, mqd->cp_mqd_base_addr); - WREG32_SOC15(GC, 0, mmCP_MQD_BASE_ADDR_HI, mqd->cp_mqd_base_addr_hi); - - /* set GFX_MQD_CONTROL */ - WREG32_SOC15(GC, 0, mmCP_GFX_MQD_CONTROL, mqd->cp_gfx_mqd_control); - - /* set GFX_HQD_VMID to 0 */ - WREG32_SOC15(GC, 0, mmCP_GFX_HQD_VMID, mqd->cp_gfx_hqd_vmid); - - WREG32_SOC15(GC, 0, mmCP_GFX_HQD_QUEUE_PRIORITY, - mqd->cp_gfx_hqd_queue_priority); - WREG32_SOC15(GC, 0, mmCP_GFX_HQD_QUANTUM, mqd->cp_gfx_hqd_quantum); - - /* set GFX_HQD_BASE, similar as CP_RB_BASE */ - WREG32_SOC15(GC, 0, mmCP_GFX_HQD_BASE, mqd->cp_gfx_hqd_base); - WREG32_SOC15(GC, 0, mmCP_GFX_HQD_BASE_HI, mqd->cp_gfx_hqd_base_hi); - - /* set GFX_HQD_RPTR_ADDR, similar as CP_RB_RPTR */ - WREG32_SOC15(GC, 0, mmCP_GFX_HQD_RPTR_ADDR, mqd->cp_gfx_hqd_rptr_addr); - WREG32_SOC15(GC, 0, mmCP_GFX_HQD_RPTR_ADDR_HI, mqd->cp_gfx_hqd_rptr_addr_hi); - - /* set GFX_HQD_CNTL, similar as CP_RB_CNTL */ - WREG32_SOC15(GC, 0, mmCP_GFX_HQD_CNTL, mqd->cp_gfx_hqd_cntl); - - /* set RB_WPTR_POLL_ADDR */ - WREG32_SOC15(GC, 0, mmCP_RB_WPTR_POLL_ADDR_LO, mqd->cp_rb_wptr_poll_addr_lo); - WREG32_SOC15(GC, 0, mmCP_RB_WPTR_POLL_ADDR_HI, mqd->cp_rb_wptr_poll_addr_hi); - - /* set RB_DOORBELL_CONTROL */ - WREG32_SOC15(GC, 0, mmCP_RB_DOORBELL_CONTROL, mqd->cp_rb_doorbell_control); - - /* active the queue */ - WREG32_SOC15(GC, 0, mmCP_GFX_HQD_ACTIVE, mqd->cp_gfx_hqd_active); - - return 0; -} -#endif - static int gfx_v10_0_gfx_init_queue(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; @@ -6492,59 +6463,23 @@ static int gfx_v10_0_gfx_init_queue(struct amdgpu_ring *ring) if (ring->doorbell_index == adev->doorbell_index.gfx_ring0 << 1) gfx_v10_0_cp_gfx_set_doorbell(adev, ring); -#ifdef BRING_UP_DEBUG - gfx_v10_0_gfx_queue_init_register(ring); -#endif nv_grbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); if (adev->gfx.me.mqd_backup[mqd_idx]) memcpy(adev->gfx.me.mqd_backup[mqd_idx], mqd, sizeof(*mqd)); - } else if (amdgpu_in_reset(adev)) { - /* reset mqd with the backup copy */ + } else { + /* restore mqd with the backup copy */ if (adev->gfx.me.mqd_backup[mqd_idx]) memcpy(mqd, adev->gfx.me.mqd_backup[mqd_idx], sizeof(*mqd)); /* reset the ring */ ring->wptr = 0; *ring->wptr_cpu_addr = 0; amdgpu_ring_clear_ring(ring); -#ifdef BRING_UP_DEBUG - mutex_lock(&adev->srbm_mutex); - nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); - gfx_v10_0_gfx_queue_init_register(ring); - nv_grbm_select(adev, 0, 0, 0, 0); - mutex_unlock(&adev->srbm_mutex); -#endif - } else { - amdgpu_ring_clear_ring(ring); } return 0; } -#ifndef BRING_UP_DEBUG -static int gfx_v10_0_kiq_enable_kgq(struct amdgpu_device *adev) -{ - struct amdgpu_kiq *kiq = &adev->gfx.kiq; - struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; - int r, i; - - if (!kiq->pmf || !kiq->pmf->kiq_map_queues) - return -EINVAL; - - r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size * - adev->gfx.num_gfx_rings); - if (r) { - DRM_ERROR("Failed to lock KIQ (%d).\n", r); - return r; - } - - for (i = 0; i < adev->gfx.num_gfx_rings; i++) - kiq->pmf->kiq_map_queues(kiq_ring, &adev->gfx.gfx_ring[i]); - - return amdgpu_ring_test_helper(kiq_ring); -} -#endif - static int gfx_v10_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev) { int r, i; @@ -6555,7 +6490,7 @@ static int gfx_v10_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev) r = amdgpu_bo_reserve(ring->mqd_obj, false); if (unlikely(r != 0)) - goto done; + return r; r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); if (!r) { @@ -6565,23 +6500,14 @@ static int gfx_v10_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev) } amdgpu_bo_unreserve(ring->mqd_obj); if (r) - goto done; + return r; } -#ifndef BRING_UP_DEBUG - r = gfx_v10_0_kiq_enable_kgq(adev); - if (r) - goto done; -#endif - r = gfx_v10_0_cp_gfx_start(adev); + + r = amdgpu_gfx_enable_kgq(adev, 0); if (r) - goto done; + return r; - for (i = 0; i < adev->gfx.num_gfx_rings; i++) { - ring = &adev->gfx.gfx_ring[i]; - ring->sched.ready = true; - } -done: - return r; + return gfx_v10_0_cp_gfx_start(adev); } static int gfx_v10_0_compute_mqd_init(struct amdgpu_device *adev, void *m, @@ -6812,14 +6738,13 @@ static int gfx_v10_0_kiq_init_queue(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; struct v10_compute_mqd *mqd = ring->mqd_ptr; - int mqd_idx = AMDGPU_MAX_COMPUTE_RINGS; gfx_v10_0_kiq_setting(ring); if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */ /* reset MQD to a clean status */ - if (adev->gfx.mec.mqd_backup[mqd_idx]) - memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd)); + if (adev->gfx.kiq[0].mqd_backup) + memcpy(mqd, adev->gfx.kiq[0].mqd_backup, sizeof(*mqd)); /* reset ring buffer */ ring->wptr = 0; @@ -6841,8 +6766,8 @@ static int gfx_v10_0_kiq_init_queue(struct amdgpu_ring *ring) nv_grbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); - if (adev->gfx.mec.mqd_backup[mqd_idx]) - memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd)); + if (adev->gfx.kiq[0].mqd_backup) + memcpy(adev->gfx.kiq[0].mqd_backup, mqd, sizeof(*mqd)); } return 0; @@ -6864,17 +6789,14 @@ static int gfx_v10_0_kcq_init_queue(struct amdgpu_ring *ring) if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd)); - } else if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */ - /* reset MQD to a clean status */ + } else { + /* restore MQD to a clean status */ if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd)); - /* reset ring buffer */ ring->wptr = 0; atomic64_set((atomic64_t *)ring->wptr_cpu_addr, 0); amdgpu_ring_clear_ring(ring); - } else { - amdgpu_ring_clear_ring(ring); } return 0; @@ -6885,21 +6807,22 @@ static int gfx_v10_0_kiq_resume(struct amdgpu_device *adev) struct amdgpu_ring *ring; int r; - ring = &adev->gfx.kiq.ring; + ring = &adev->gfx.kiq[0].ring; r = amdgpu_bo_reserve(ring->mqd_obj, false); if (unlikely(r != 0)) return r; r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); - if (unlikely(r != 0)) + if (unlikely(r != 0)) { + amdgpu_bo_unreserve(ring->mqd_obj); return r; + } gfx_v10_0_kiq_init_queue(ring); amdgpu_bo_kunmap(ring->mqd_obj); ring->mqd_ptr = NULL; amdgpu_bo_unreserve(ring->mqd_obj); - ring->sched.ready = true; return 0; } @@ -6927,7 +6850,7 @@ static int gfx_v10_0_kcq_resume(struct amdgpu_device *adev) goto done; } - r = amdgpu_gfx_enable_kcq(adev); + r = amdgpu_gfx_enable_kcq(adev, 0); done: return r; } @@ -7240,47 +7163,20 @@ static int gfx_v10_0_hw_init(void *handle) return r; } -#ifndef BRING_UP_DEBUG -static int gfx_v10_0_kiq_disable_kgq(struct amdgpu_device *adev) -{ - struct amdgpu_kiq *kiq = &adev->gfx.kiq; - struct amdgpu_ring *kiq_ring = &kiq->ring; - int i; - - if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues) - return -EINVAL; - - if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size * - adev->gfx.num_gfx_rings)) - return -ENOMEM; - - for (i = 0; i < adev->gfx.num_gfx_rings; i++) - kiq->pmf->kiq_unmap_queues(kiq_ring, &adev->gfx.gfx_ring[i], - PREEMPT_QUEUES, 0, 0); - if (!adev->job_hang) - return amdgpu_ring_test_helper(kiq_ring); - else - return 0; -} -#endif - static int gfx_v10_0_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - int r; amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0); amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0); if (!adev->no_hw_access) { -#ifndef BRING_UP_DEBUG if (amdgpu_async_gfx_ring) { - r = gfx_v10_0_kiq_disable_kgq(adev); - if (r) + if (amdgpu_gfx_disable_kgq(adev, 0)) DRM_ERROR("KGQ disable failed\n"); } -#endif - if (amdgpu_gfx_disable_kcq(adev)) + + if (amdgpu_gfx_disable_kcq(adev, 0)) DRM_ERROR("KCQ disable failed\n"); } @@ -7572,7 +7468,7 @@ static bool gfx_v10_0_is_rlc_enabled(struct amdgpu_device *adev) return (REG_GET_FIELD(rlc_cntl, RLC_CNTL, RLC_ENABLE_F32)) ? true : false; } -static void gfx_v10_0_set_safe_mode(struct amdgpu_device *adev) +static void gfx_v10_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id) { uint32_t data; unsigned i; @@ -7613,7 +7509,7 @@ static void gfx_v10_0_set_safe_mode(struct amdgpu_device *adev) } } -static void gfx_v10_0_unset_safe_mode(struct amdgpu_device *adev) +static void gfx_v10_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id) { uint32_t data; @@ -7960,7 +7856,7 @@ static void gfx_v10_0_apply_medium_grain_clock_gating_workaround(struct amdgpu_d static int gfx_v10_0_update_gfx_clock_gating(struct amdgpu_device *adev, bool enable) { - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); if (enable) { /* enable FGCG firstly*/ @@ -7999,7 +7895,7 @@ static int gfx_v10_0_update_gfx_clock_gating(struct amdgpu_device *adev, AMD_CG_SUPPORT_GFX_3D_CGLS)) gfx_v10_0_enable_gui_idle_interrupt(adev, enable); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return 0; } @@ -8093,11 +7989,11 @@ static void gfx_v10_cntl_power_gating(struct amdgpu_device *adev, bool enable) static void gfx_v10_cntl_pg(struct amdgpu_device *adev, bool enable) { - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); gfx_v10_cntl_power_gating(adev, enable); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } static const struct amdgpu_rlc_funcs gfx_v10_0_rlc_funcs = { @@ -8152,8 +8048,14 @@ static int gfx_v10_0_set_powergating_state(void *handle, case IP_VERSION(10, 3, 3): case IP_VERSION(10, 3, 6): case IP_VERSION(10, 3, 7): + if (!enable) + amdgpu_gfx_off_ctrl(adev, false); + gfx_v10_cntl_pg(adev, enable); - amdgpu_gfx_off_ctrl(adev, enable); + + if (enable) + amdgpu_gfx_off_ctrl(adev, true); + break; default: break; @@ -8640,7 +8542,7 @@ static int gfx_v10_0_ring_preempt_ib(struct amdgpu_ring *ring) { int i, r = 0; struct amdgpu_device *adev = ring->adev; - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; struct amdgpu_ring *kiq_ring = &kiq->ring; unsigned long flags; @@ -9148,7 +9050,7 @@ static int gfx_v10_0_kiq_set_interrupt_state(struct amdgpu_device *adev, enum amdgpu_interrupt_state state) { uint32_t tmp, target; - struct amdgpu_ring *ring = &(adev->gfx.kiq.ring); + struct amdgpu_ring *ring = &(adev->gfx.kiq[0].ring); if (ring->me == 1) target = SOC15_REG_OFFSET(GC, 0, mmCP_ME1_PIPE0_INT_CNTL); @@ -9192,7 +9094,7 @@ static int gfx_v10_0_kiq_irq(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry) { u8 me_id, pipe_id, queue_id; - struct amdgpu_ring *ring = &(adev->gfx.kiq.ring); + struct amdgpu_ring *ring = &(adev->gfx.kiq[0].ring); me_id = (entry->ring_id & 0x0c) >> 2; pipe_id = (entry->ring_id & 0x03) >> 0; @@ -9369,7 +9271,7 @@ static void gfx_v10_0_set_ring_funcs(struct amdgpu_device *adev) { int i; - adev->gfx.kiq.ring.funcs = &gfx_v10_0_ring_funcs_kiq; + adev->gfx.kiq[0].ring.funcs = &gfx_v10_0_ring_funcs_kiq; for (i = 0; i < adev->gfx.num_gfx_rings; i++) adev->gfx.gfx_ring[i].funcs = &gfx_v10_0_ring_funcs_gfx; @@ -9403,8 +9305,8 @@ static void gfx_v10_0_set_irq_funcs(struct amdgpu_device *adev) adev->gfx.eop_irq.num_types = AMDGPU_CP_IRQ_LAST; adev->gfx.eop_irq.funcs = &gfx_v10_0_eop_irq_funcs; - adev->gfx.kiq.irq.num_types = AMDGPU_CP_KIQ_IRQ_LAST; - adev->gfx.kiq.irq.funcs = &gfx_v10_0_kiq_irq_funcs; + adev->gfx.kiq[0].irq.num_types = AMDGPU_CP_KIQ_IRQ_LAST; + adev->gfx.kiq[0].irq.funcs = &gfx_v10_0_kiq_irq_funcs; adev->gfx.priv_reg_irq.num_types = 1; adev->gfx.priv_reg_irq.funcs = &gfx_v10_0_priv_reg_irq_funcs; @@ -9541,7 +9443,7 @@ static int gfx_v10_0_get_cu_info(struct amdgpu_device *adev, mask = 1; ao_bitmap = 0; counter = 0; - gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff, 0); if (i < 4 && j < 2) gfx_v10_0_set_user_wgp_inactive_bitmap_per_sh( adev, disable_masks[i * 2 + j]); @@ -9562,7 +9464,7 @@ static int gfx_v10_0_get_cu_info(struct amdgpu_device *adev, cu_info->ao_cu_bitmap[i][j] = ao_bitmap; } } - gfx_v10_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v10_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); cu_info->number = active_cu_number; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c index f5c376276984..690e121d9dda 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c @@ -112,7 +112,7 @@ static int gfx_v11_0_get_cu_info(struct amdgpu_device *adev, struct amdgpu_cu_info *cu_info); static uint64_t gfx_v11_0_get_gpu_clock_counter(struct amdgpu_device *adev); static void gfx_v11_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, - u32 sh_num, u32 instance); + u32 sh_num, u32 instance, int xcc_id); static u32 gfx_v11_0_get_wgp_active_bitmap_per_sh(struct amdgpu_device *adev); static void gfx_v11_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume); @@ -123,8 +123,8 @@ static int gfx_v11_0_wait_for_rlc_autoload_complete(struct amdgpu_device *adev); static void gfx_v11_0_ring_invalidate_tlbs(struct amdgpu_ring *ring, uint16_t pasid, uint32_t flush_type, bool all_hub, uint8_t dst_sel); -static void gfx_v11_0_set_safe_mode(struct amdgpu_device *adev); -static void gfx_v11_0_unset_safe_mode(struct amdgpu_device *adev); +static void gfx_v11_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id); +static void gfx_v11_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id); static void gfx_v11_0_update_perf_clk(struct amdgpu_device *adev, bool enable); @@ -192,7 +192,7 @@ static void gfx11_kiq_unmap_queues(struct amdgpu_ring *kiq_ring, struct amdgpu_device *adev = kiq_ring->adev; uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0; - if (adev->enable_mes && !adev->gfx.kiq.ring.sched.ready) { + if (adev->enable_mes && !adev->gfx.kiq[0].ring.sched.ready) { amdgpu_mes_unmap_legacy_queue(adev, ring, action, gpu_addr, seq); return; } @@ -260,7 +260,7 @@ static const struct kiq_pm4_funcs gfx_v11_0_kiq_pm4_funcs = { static void gfx_v11_0_set_kiq_pm4_funcs(struct amdgpu_device *adev) { - adev->gfx.kiq.pmf = &gfx_v11_0_kiq_pm4_funcs; + adev->gfx.kiq[0].pmf = &gfx_v11_0_kiq_pm4_funcs; } static void gfx_v11_0_init_golden_registers(struct amdgpu_device *adev) @@ -463,6 +463,23 @@ out: return err; } +static void gfx_v11_0_check_fw_cp_gfx_shadow(struct amdgpu_device *adev) +{ + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 2): + case IP_VERSION(11, 0, 3): + if ((adev->gfx.me_fw_version >= 1505) && + (adev->gfx.pfp_fw_version >= 1600) && + (adev->gfx.mec_fw_version >= 512)) + adev->gfx.cp_gfx_shadow = true; + break; + default: + adev->gfx.cp_gfx_shadow = false; + break; + } +} + static int gfx_v11_0_init_microcode(struct amdgpu_device *adev) { char fw_name[40]; @@ -539,6 +556,7 @@ static int gfx_v11_0_init_microcode(struct amdgpu_device *adev) /* only one MEC for gfx 11.0.0. */ adev->gfx.mec2_fw = NULL; + gfx_v11_0_check_fw_cp_gfx_shadow(adev); out: if (err) { amdgpu_ucode_release(&adev->gfx.pfp_fw); @@ -699,7 +717,7 @@ static int gfx_v11_0_mec_init(struct amdgpu_device *adev) u32 *hpd; size_t mec_hpd_size; - bitmap_zero(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES); + bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES); /* take ownership of the relevant compute queues */ amdgpu_gfx_compute_queue_acquire(adev); @@ -747,7 +765,7 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t wave, *(out++) = RREG32_SOC15(GC, 0, regSQ_IND_DATA); } -static void gfx_v11_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) +static void gfx_v11_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) { /* in gfx11 the SIMD_ID is specified as part of the INSTANCE * field when performing a select_se_sh so it should be @@ -773,7 +791,7 @@ static void gfx_v11_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_MODE); } -static void gfx_v11_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd, +static void gfx_v11_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t start, uint32_t size, uint32_t *dst) { @@ -784,7 +802,7 @@ static void gfx_v11_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd, dst); } -static void gfx_v11_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd, +static void gfx_v11_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t thread, uint32_t start, uint32_t size, uint32_t *dst) @@ -795,11 +813,32 @@ static void gfx_v11_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd, } static void gfx_v11_0_select_me_pipe_q(struct amdgpu_device *adev, - u32 me, u32 pipe, u32 q, u32 vm) + u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id) { soc21_grbm_select(adev, me, pipe, q, vm); } +/* all sizes are in bytes */ +#define MQD_SHADOW_BASE_SIZE 73728 +#define MQD_SHADOW_BASE_ALIGNMENT 256 +#define MQD_FWWORKAREA_SIZE 484 +#define MQD_FWWORKAREA_ALIGNMENT 256 + +static int gfx_v11_0_get_gfx_shadow_info(struct amdgpu_device *adev, + struct amdgpu_gfx_shadow_info *shadow_info) +{ + if (adev->gfx.cp_gfx_shadow) { + shadow_info->shadow_size = MQD_SHADOW_BASE_SIZE; + shadow_info->shadow_alignment = MQD_SHADOW_BASE_ALIGNMENT; + shadow_info->csa_size = MQD_FWWORKAREA_SIZE; + shadow_info->csa_alignment = MQD_FWWORKAREA_ALIGNMENT; + return 0; + } else { + memset(shadow_info, 0, sizeof(struct amdgpu_gfx_shadow_info)); + return -ENOTSUPP; + } +} + static const struct amdgpu_gfx_funcs gfx_v11_0_gfx_funcs = { .get_gpu_clock_counter = &gfx_v11_0_get_gpu_clock_counter, .select_se_sh = &gfx_v11_0_select_se_sh, @@ -808,6 +847,7 @@ static const struct amdgpu_gfx_funcs gfx_v11_0_gfx_funcs = { .read_wave_vgprs = &gfx_v11_0_read_wave_vgprs, .select_me_pipe_q = &gfx_v11_0_select_me_pipe_q, .update_perfmon_mgcg = &gfx_v11_0_update_perf_clk, + .get_gfx_shadow_info = &gfx_v11_0_get_gfx_shadow_info, }; static int gfx_v11_0_gpu_early_init(struct amdgpu_device *adev) @@ -866,7 +906,7 @@ static int gfx_v11_0_gfx_ring_init(struct amdgpu_device *adev, int ring_id, ring->doorbell_index = adev->doorbell_index.gfx_ring0 << 1; else ring->doorbell_index = adev->doorbell_index.gfx_ring1 << 1; - ring->vm_hub = AMDGPU_GFXHUB_0; + ring->vm_hub = AMDGPU_GFXHUB(0); sprintf(ring->name, "gfx_%d.%d.%d", ring->me, ring->pipe, ring->queue); irq_type = AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP + ring->pipe; @@ -897,7 +937,7 @@ static int gfx_v11_0_compute_ring_init(struct amdgpu_device *adev, int ring_id, ring->doorbell_index = (adev->doorbell_index.mec_ring0 + ring_id) << 1; ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr + (ring_id * GFX11_MEC_HPD_SIZE); - ring->vm_hub = AMDGPU_GFXHUB_0; + ring->vm_hub = AMDGPU_GFXHUB(0); sprintf(ring->name, "comp_%d.%d.%d", ring->me, ring->pipe, ring->queue); irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP @@ -1367,8 +1407,8 @@ static int gfx_v11_0_sw_init(void *handle) for (i = 0; i < adev->gfx.mec.num_mec; ++i) { for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) { for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) { - if (!amdgpu_gfx_is_mec_queue_enabled(adev, i, k, - j)) + if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i, + k, j)) continue; r = gfx_v11_0_compute_ring_init(adev, ring_id, @@ -1382,19 +1422,19 @@ static int gfx_v11_0_sw_init(void *handle) } if (!adev->enable_mes_kiq) { - r = amdgpu_gfx_kiq_init(adev, GFX11_MEC_HPD_SIZE); + r = amdgpu_gfx_kiq_init(adev, GFX11_MEC_HPD_SIZE, 0); if (r) { DRM_ERROR("Failed to init KIQ BOs!\n"); return r; } - kiq = &adev->gfx.kiq; - r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq); + kiq = &adev->gfx.kiq[0]; + r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq, 0); if (r) return r; } - r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v11_compute_mqd)); + r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v11_compute_mqd), 0); if (r) return r; @@ -1456,11 +1496,11 @@ static int gfx_v11_0_sw_fini(void *handle) for (i = 0; i < adev->gfx.num_compute_rings; i++) amdgpu_ring_fini(&adev->gfx.compute_ring[i]); - amdgpu_gfx_mqd_sw_fini(adev); + amdgpu_gfx_mqd_sw_fini(adev, 0); if (!adev->enable_mes_kiq) { - amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq.ring); - amdgpu_gfx_kiq_fini(adev); + amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq[0].ring); + amdgpu_gfx_kiq_fini(adev, 0); } gfx_v11_0_pfp_fini(adev); @@ -1477,7 +1517,7 @@ static int gfx_v11_0_sw_fini(void *handle) } static void gfx_v11_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, - u32 sh_num, u32 instance) + u32 sh_num, u32 instance, int xcc_id) { u32 data; @@ -1598,6 +1638,7 @@ static void gfx_v11_0_init_compute_vmid(struct amdgpu_device *adev) /* Enable trap for each kfd vmid. */ data = RREG32_SOC15(GC, 0, regSPI_GDBG_PER_VMID_CNTL); data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1); + WREG32_SOC15(GC, 0, regSPI_GDBG_PER_VMID_CNTL, data); } soc21_grbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); @@ -1667,7 +1708,7 @@ static void gfx_v11_0_constants_init(struct amdgpu_device *adev) /* XXX SH_MEM regs */ /* where to put LDS, scratch, GPUVM in FSA64 space */ mutex_lock(&adev->srbm_mutex); - for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB_0].num_ids; i++) { + for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB(0)].num_ids; i++) { soc21_grbm_select(adev, 0, 0, 0, i); /* CP and shaders */ WREG32_SOC15(GC, 0, regSH_MEM_CONFIG, DEFAULT_SH_MEM_CONFIG); @@ -3188,7 +3229,6 @@ static int gfx_v11_0_cp_gfx_resume(struct amdgpu_device *adev) u32 tmp; u32 rb_bufsz; u64 rb_addr, rptr_addr, wptr_gpu_addr; - u32 i; /* Set the write pointer delay */ WREG32_SOC15(GC, 0, regCP_RB_WPTR_DELAY, 0); @@ -3280,11 +3320,6 @@ static int gfx_v11_0_cp_gfx_resume(struct amdgpu_device *adev) /* start the ring */ gfx_v11_0_cp_gfx_start(adev); - for (i = 0; i < adev->gfx.num_gfx_rings; i++) { - ring = &adev->gfx.gfx_ring[i]; - ring->sched.ready = true; - } - return 0; } @@ -3330,8 +3365,6 @@ static void gfx_v11_0_cp_compute_enable(struct amdgpu_device *adev, bool enable) WREG32_SOC15(GC, 0, regCP_MEC_CNTL, data); } - adev->gfx.kiq.ring.sched.ready = enable; - udelay(50); } @@ -3633,55 +3666,6 @@ static int gfx_v11_0_gfx_mqd_init(struct amdgpu_device *adev, void *m, return 0; } -#ifdef BRING_UP_DEBUG -static int gfx_v11_0_gfx_queue_init_register(struct amdgpu_ring *ring) -{ - struct amdgpu_device *adev = ring->adev; - struct v11_gfx_mqd *mqd = ring->mqd_ptr; - - /* set mmCP_GFX_HQD_WPTR/_HI to 0 */ - WREG32_SOC15(GC, 0, regCP_GFX_HQD_WPTR, mqd->cp_gfx_hqd_wptr); - WREG32_SOC15(GC, 0, regCP_GFX_HQD_WPTR_HI, mqd->cp_gfx_hqd_wptr_hi); - - /* set GFX_MQD_BASE */ - WREG32_SOC15(GC, 0, regCP_MQD_BASE_ADDR, mqd->cp_mqd_base_addr); - WREG32_SOC15(GC, 0, regCP_MQD_BASE_ADDR_HI, mqd->cp_mqd_base_addr_hi); - - /* set GFX_MQD_CONTROL */ - WREG32_SOC15(GC, 0, regCP_GFX_MQD_CONTROL, mqd->cp_gfx_mqd_control); - - /* set GFX_HQD_VMID to 0 */ - WREG32_SOC15(GC, 0, regCP_GFX_HQD_VMID, mqd->cp_gfx_hqd_vmid); - - WREG32_SOC15(GC, 0, regCP_GFX_HQD_QUEUE_PRIORITY, - mqd->cp_gfx_hqd_queue_priority); - WREG32_SOC15(GC, 0, regCP_GFX_HQD_QUANTUM, mqd->cp_gfx_hqd_quantum); - - /* set GFX_HQD_BASE, similar as CP_RB_BASE */ - WREG32_SOC15(GC, 0, regCP_GFX_HQD_BASE, mqd->cp_gfx_hqd_base); - WREG32_SOC15(GC, 0, regCP_GFX_HQD_BASE_HI, mqd->cp_gfx_hqd_base_hi); - - /* set GFX_HQD_RPTR_ADDR, similar as CP_RB_RPTR */ - WREG32_SOC15(GC, 0, regCP_GFX_HQD_RPTR_ADDR, mqd->cp_gfx_hqd_rptr_addr); - WREG32_SOC15(GC, 0, regCP_GFX_HQD_RPTR_ADDR_HI, mqd->cp_gfx_hqd_rptr_addr_hi); - - /* set GFX_HQD_CNTL, similar as CP_RB_CNTL */ - WREG32_SOC15(GC, 0, regCP_GFX_HQD_CNTL, mqd->cp_gfx_hqd_cntl); - - /* set RB_WPTR_POLL_ADDR */ - WREG32_SOC15(GC, 0, regCP_RB_WPTR_POLL_ADDR_LO, mqd->cp_rb_wptr_poll_addr_lo); - WREG32_SOC15(GC, 0, regCP_RB_WPTR_POLL_ADDR_HI, mqd->cp_rb_wptr_poll_addr_hi); - - /* set RB_DOORBELL_CONTROL */ - WREG32_SOC15(GC, 0, regCP_RB_DOORBELL_CONTROL, mqd->cp_rb_doorbell_control); - - /* active the queue */ - WREG32_SOC15(GC, 0, regCP_GFX_HQD_ACTIVE, mqd->cp_gfx_hqd_active); - - return 0; -} -#endif - static int gfx_v11_0_gfx_init_queue(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; @@ -3693,59 +3677,23 @@ static int gfx_v11_0_gfx_init_queue(struct amdgpu_ring *ring) mutex_lock(&adev->srbm_mutex); soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); amdgpu_ring_init_mqd(ring); -#ifdef BRING_UP_DEBUG - gfx_v11_0_gfx_queue_init_register(ring); -#endif soc21_grbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); if (adev->gfx.me.mqd_backup[mqd_idx]) memcpy(adev->gfx.me.mqd_backup[mqd_idx], mqd, sizeof(*mqd)); - } else if (amdgpu_in_reset(adev)) { - /* reset mqd with the backup copy */ + } else { + /* restore mqd with the backup copy */ if (adev->gfx.me.mqd_backup[mqd_idx]) memcpy(mqd, adev->gfx.me.mqd_backup[mqd_idx], sizeof(*mqd)); /* reset the ring */ ring->wptr = 0; *ring->wptr_cpu_addr = 0; amdgpu_ring_clear_ring(ring); -#ifdef BRING_UP_DEBUG - mutex_lock(&adev->srbm_mutex); - soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); - gfx_v11_0_gfx_queue_init_register(ring); - soc21_grbm_select(adev, 0, 0, 0, 0); - mutex_unlock(&adev->srbm_mutex); -#endif - } else { - amdgpu_ring_clear_ring(ring); } return 0; } -#ifndef BRING_UP_DEBUG -static int gfx_v11_0_kiq_enable_kgq(struct amdgpu_device *adev) -{ - struct amdgpu_kiq *kiq = &adev->gfx.kiq; - struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; - int r, i; - - if (!kiq->pmf || !kiq->pmf->kiq_map_queues) - return -EINVAL; - - r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size * - adev->gfx.num_gfx_rings); - if (r) { - DRM_ERROR("Failed to lock KIQ (%d).\n", r); - return r; - } - - for (i = 0; i < adev->gfx.num_gfx_rings; i++) - kiq->pmf->kiq_map_queues(kiq_ring, &adev->gfx.gfx_ring[i]); - - return amdgpu_ring_test_helper(kiq_ring); -} -#endif - static int gfx_v11_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev) { int r, i; @@ -3756,7 +3704,7 @@ static int gfx_v11_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev) r = amdgpu_bo_reserve(ring->mqd_obj, false); if (unlikely(r != 0)) - goto done; + return r; r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); if (!r) { @@ -3766,23 +3714,14 @@ static int gfx_v11_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev) } amdgpu_bo_unreserve(ring->mqd_obj); if (r) - goto done; + return r; } -#ifndef BRING_UP_DEBUG - r = gfx_v11_0_kiq_enable_kgq(adev); - if (r) - goto done; -#endif - r = gfx_v11_0_cp_gfx_start(adev); + + r = amdgpu_gfx_enable_kgq(adev, 0); if (r) - goto done; + return r; - for (i = 0; i < adev->gfx.num_gfx_rings; i++) { - ring = &adev->gfx.gfx_ring[i]; - ring->sched.ready = true; - } -done: - return r; + return gfx_v11_0_cp_gfx_start(adev); } static int gfx_v11_0_compute_mqd_init(struct amdgpu_device *adev, void *m, @@ -4028,14 +3967,13 @@ static int gfx_v11_0_kiq_init_queue(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; struct v11_compute_mqd *mqd = ring->mqd_ptr; - int mqd_idx = AMDGPU_MAX_COMPUTE_RINGS; gfx_v11_0_kiq_setting(ring); if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */ /* reset MQD to a clean status */ - if (adev->gfx.mec.mqd_backup[mqd_idx]) - memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd)); + if (adev->gfx.kiq[0].mqd_backup) + memcpy(mqd, adev->gfx.kiq[0].mqd_backup, sizeof(*mqd)); /* reset ring buffer */ ring->wptr = 0; @@ -4057,8 +3995,8 @@ static int gfx_v11_0_kiq_init_queue(struct amdgpu_ring *ring) soc21_grbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); - if (adev->gfx.mec.mqd_backup[mqd_idx]) - memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd)); + if (adev->gfx.kiq[0].mqd_backup) + memcpy(adev->gfx.kiq[0].mqd_backup, mqd, sizeof(*mqd)); } return 0; @@ -4080,17 +4018,14 @@ static int gfx_v11_0_kcq_init_queue(struct amdgpu_ring *ring) if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd)); - } else if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */ - /* reset MQD to a clean status */ + } else { + /* restore MQD to a clean status */ if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd)); - /* reset ring buffer */ ring->wptr = 0; atomic64_set((atomic64_t *)ring->wptr_cpu_addr, 0); amdgpu_ring_clear_ring(ring); - } else { - amdgpu_ring_clear_ring(ring); } return 0; @@ -4101,7 +4036,7 @@ static int gfx_v11_0_kiq_resume(struct amdgpu_device *adev) struct amdgpu_ring *ring; int r; - ring = &adev->gfx.kiq.ring; + ring = &adev->gfx.kiq[0].ring; r = amdgpu_bo_reserve(ring->mqd_obj, false); if (unlikely(r != 0)) @@ -4146,7 +4081,7 @@ static int gfx_v11_0_kcq_resume(struct amdgpu_device *adev) goto done; } - r = amdgpu_gfx_enable_kcq(adev); + r = amdgpu_gfx_enable_kcq(adev, 0); done: return r; } @@ -4239,7 +4174,7 @@ static int gfx_v11_0_gfxhub_enable(struct amdgpu_device *adev) false : true; adev->gfxhub.funcs->set_fault_enable_default(adev, value); - amdgpu_gmc_flush_gpu_tlb(adev, 0, AMDGPU_GFXHUB_0, 0); + amdgpu_gmc_flush_gpu_tlb(adev, 0, AMDGPU_GFXHUB(0), 0); return 0; } @@ -4407,48 +4342,20 @@ static int gfx_v11_0_hw_init(void *handle) return r; } -#ifndef BRING_UP_DEBUG -static int gfx_v11_0_kiq_disable_kgq(struct amdgpu_device *adev) -{ - struct amdgpu_kiq *kiq = &adev->gfx.kiq; - struct amdgpu_ring *kiq_ring = &kiq->ring; - int i, r = 0; - - if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues) - return -EINVAL; - - if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size * - adev->gfx.num_gfx_rings)) - return -ENOMEM; - - for (i = 0; i < adev->gfx.num_gfx_rings; i++) - kiq->pmf->kiq_unmap_queues(kiq_ring, &adev->gfx.gfx_ring[i], - PREEMPT_QUEUES, 0, 0); - - if (adev->gfx.kiq.ring.sched.ready) - r = amdgpu_ring_test_helper(kiq_ring); - - return r; -} -#endif - static int gfx_v11_0_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - int r; amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0); amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0); if (!adev->no_hw_access) { -#ifndef BRING_UP_DEBUG if (amdgpu_async_gfx_ring) { - r = gfx_v11_0_kiq_disable_kgq(adev); - if (r) + if (amdgpu_gfx_disable_kgq(adev, 0)) DRM_ERROR("KGQ disable failed\n"); } -#endif - if (amdgpu_gfx_disable_kcq(adev)) + + if (amdgpu_gfx_disable_kcq(adev, 0)) DRM_ERROR("KCQ disable failed\n"); amdgpu_mes_kiq_hw_fini(adev); @@ -4525,7 +4432,7 @@ static int gfx_v11_0_soft_reset(void *handle) tmp = REG_SET_FIELD(tmp, CP_INT_CNTL, GFX_IDLE_INT_ENABLE, 0); WREG32_SOC15(GC, 0, regCP_INT_CNTL, tmp); - gfx_v11_0_set_safe_mode(adev); + gfx_v11_0_set_safe_mode(adev, 0); for (i = 0; i < adev->gfx.mec.num_mec; ++i) { for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) { @@ -4625,7 +4532,7 @@ static int gfx_v11_0_soft_reset(void *handle) tmp = REG_SET_FIELD(tmp, CP_INT_CNTL, GFX_IDLE_INT_ENABLE, 1); WREG32_SOC15(GC, 0, regCP_INT_CNTL, tmp); - gfx_v11_0_unset_safe_mode(adev); + gfx_v11_0_unset_safe_mode(adev, 0); return gfx_v11_0_cp_resume(adev); } @@ -4667,24 +4574,27 @@ static uint64_t gfx_v11_0_get_gpu_clock_counter(struct amdgpu_device *adev) uint64_t clock; uint64_t clock_counter_lo, clock_counter_hi_pre, clock_counter_hi_after; - amdgpu_gfx_off_ctrl(adev, false); - mutex_lock(&adev->gfx.gpu_clock_mutex); if (amdgpu_sriov_vf(adev)) { + amdgpu_gfx_off_ctrl(adev, false); + mutex_lock(&adev->gfx.gpu_clock_mutex); clock_counter_hi_pre = (uint64_t)RREG32_SOC15(GC, 0, regCP_MES_MTIME_HI); clock_counter_lo = (uint64_t)RREG32_SOC15(GC, 0, regCP_MES_MTIME_LO); clock_counter_hi_after = (uint64_t)RREG32_SOC15(GC, 0, regCP_MES_MTIME_HI); if (clock_counter_hi_pre != clock_counter_hi_after) clock_counter_lo = (uint64_t)RREG32_SOC15(GC, 0, regCP_MES_MTIME_LO); + mutex_unlock(&adev->gfx.gpu_clock_mutex); + amdgpu_gfx_off_ctrl(adev, true); } else { + preempt_disable(); clock_counter_hi_pre = (uint64_t)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_UPPER); clock_counter_lo = (uint64_t)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_LOWER); clock_counter_hi_after = (uint64_t)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_UPPER); if (clock_counter_hi_pre != clock_counter_hi_after) clock_counter_lo = (uint64_t)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_LOWER); + preempt_enable(); } clock = clock_counter_lo | (clock_counter_hi_after << 32ULL); - mutex_unlock(&adev->gfx.gpu_clock_mutex); - amdgpu_gfx_off_ctrl(adev, true); + return clock; } @@ -4791,7 +4701,7 @@ static bool gfx_v11_0_is_rlc_enabled(struct amdgpu_device *adev) return (REG_GET_FIELD(rlc_cntl, RLC_CNTL, RLC_ENABLE_F32)) ? true : false; } -static void gfx_v11_0_set_safe_mode(struct amdgpu_device *adev) +static void gfx_v11_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id) { uint32_t data; unsigned i; @@ -4810,7 +4720,7 @@ static void gfx_v11_0_set_safe_mode(struct amdgpu_device *adev) } } -static void gfx_v11_0_unset_safe_mode(struct amdgpu_device *adev) +static void gfx_v11_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id) { WREG32_SOC15(GC, 0, regRLC_SAFE_MODE, RLC_SAFE_MODE__CMD_MASK); } @@ -5038,7 +4948,7 @@ static void gfx_v11_0_update_coarse_grain_clock_gating(struct amdgpu_device *ade static int gfx_v11_0_update_gfx_clock_gating(struct amdgpu_device *adev, bool enable) { - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); gfx_v11_0_update_coarse_grain_clock_gating(adev, enable); @@ -5058,7 +4968,7 @@ static int gfx_v11_0_update_gfx_clock_gating(struct amdgpu_device *adev, AMD_CG_SUPPORT_GFX_3D_CGLS)) gfx_v11_0_enable_gui_idle_interrupt(adev, enable); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return 0; } @@ -5126,11 +5036,11 @@ static void gfx_v11_cntl_power_gating(struct amdgpu_device *adev, bool enable) static void gfx_v11_cntl_pg(struct amdgpu_device *adev, bool enable) { - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); gfx_v11_cntl_power_gating(adev, enable); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } static int gfx_v11_0_set_powergating_state(void *handle, @@ -5150,8 +5060,14 @@ static int gfx_v11_0_set_powergating_state(void *handle, break; case IP_VERSION(11, 0, 1): case IP_VERSION(11, 0, 4): + if (!enable) + amdgpu_gfx_off_ctrl(adev, false); + gfx_v11_cntl_pg(adev, enable); - amdgpu_gfx_off_ctrl(adev, enable); + + if (enable) + amdgpu_gfx_off_ctrl(adev, true); + break; default: break; @@ -5583,6 +5499,29 @@ static void gfx_v11_0_ring_emit_cntxcntl(struct amdgpu_ring *ring, amdgpu_ring_write(ring, 0); } +static void gfx_v11_0_ring_emit_gfx_shadow(struct amdgpu_ring *ring, + u64 shadow_va, u64 csa_va, + u64 gds_va, bool init_shadow, + int vmid) +{ + struct amdgpu_device *adev = ring->adev; + + if (!adev->gfx.cp_gfx_shadow) + return; + + amdgpu_ring_write(ring, PACKET3(PACKET3_SET_Q_PREEMPTION_MODE, 7)); + amdgpu_ring_write(ring, lower_32_bits(shadow_va)); + amdgpu_ring_write(ring, upper_32_bits(shadow_va)); + amdgpu_ring_write(ring, lower_32_bits(gds_va)); + amdgpu_ring_write(ring, upper_32_bits(gds_va)); + amdgpu_ring_write(ring, lower_32_bits(csa_va)); + amdgpu_ring_write(ring, upper_32_bits(csa_va)); + amdgpu_ring_write(ring, shadow_va ? + PACKET3_SET_Q_PREEMPTION_MODE_IB_VMID(vmid) : 0); + amdgpu_ring_write(ring, init_shadow ? + PACKET3_SET_Q_PREEMPTION_MODE_INIT_SHADOW_MEM : 0); +} + static unsigned gfx_v11_0_ring_emit_init_cond_exec(struct amdgpu_ring *ring) { unsigned ret; @@ -5614,7 +5553,7 @@ static int gfx_v11_0_ring_preempt_ib(struct amdgpu_ring *ring) { int i, r = 0; struct amdgpu_device *adev = ring->adev; - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; struct amdgpu_ring *kiq_ring = &kiq->ring; unsigned long flags; @@ -6082,7 +6021,7 @@ static int gfx_v11_0_kiq_set_interrupt_state(struct amdgpu_device *adev, enum amdgpu_interrupt_state state) { uint32_t tmp, target; - struct amdgpu_ring *ring = &(adev->gfx.kiq.ring); + struct amdgpu_ring *ring = &(adev->gfx.kiq[0].ring); target = SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE0_INT_CNTL); target += ring->pipe; @@ -6173,6 +6112,7 @@ static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_gfx = { .set_wptr = gfx_v11_0_ring_set_wptr_gfx, .emit_frame_size = /* totally 242 maximum if 16 IBs */ 5 + /* COND_EXEC */ + 9 + /* SET_Q_PREEMPTION_MODE */ 7 + /* PIPELINE_SYNC */ SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 + @@ -6199,6 +6139,7 @@ static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_gfx = { .insert_nop = amdgpu_ring_insert_nop, .pad_ib = amdgpu_ring_generic_pad_ib, .emit_cntxcntl = gfx_v11_0_ring_emit_cntxcntl, + .emit_gfx_shadow = gfx_v11_0_ring_emit_gfx_shadow, .init_cond_exec = gfx_v11_0_ring_emit_init_cond_exec, .patch_cond_exec = gfx_v11_0_ring_emit_patch_cond_exec, .preempt_ib = gfx_v11_0_ring_preempt_ib, @@ -6279,7 +6220,7 @@ static void gfx_v11_0_set_ring_funcs(struct amdgpu_device *adev) { int i; - adev->gfx.kiq.ring.funcs = &gfx_v11_0_ring_funcs_kiq; + adev->gfx.kiq[0].ring.funcs = &gfx_v11_0_ring_funcs_kiq; for (i = 0; i < adev->gfx.num_gfx_rings; i++) adev->gfx.gfx_ring[i].funcs = &gfx_v11_0_ring_funcs_gfx; @@ -6428,7 +6369,7 @@ static int gfx_v11_0_get_cu_info(struct amdgpu_device *adev, for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { mask = 1; counter = 0; - gfx_v11_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v11_0_select_se_sh(adev, i, j, 0xffffffff, 0); if (i < 8 && j < 2) gfx_v11_0_set_user_wgp_inactive_bitmap_per_sh( adev, disable_masks[i * 2 + j]); @@ -6460,7 +6401,7 @@ static int gfx_v11_0_get_cu_info(struct amdgpu_device *adev, active_cu_number += counter; } } - gfx_v11_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v11_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); cu_info->number = active_cu_number; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.c index 068b9586a223..26d6286d86c9 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.c @@ -84,8 +84,20 @@ static int gfx_v11_0_3_poison_consumption_handler(struct amdgpu_device *adev, /* Workaround: when vmid and pasid are both zero, trigger gpu reset in KGD. */ if (entry && (entry->client_id == SOC21_IH_CLIENTID_GFX) && (entry->src_id == GFX_11_0_0__SRCID__RLC_GC_FED_INTERRUPT) && - !entry->vmid && !entry->pasid) + !entry->vmid && !entry->pasid) { + uint32_t rlc_status0 = 0; + + rlc_status0 = RREG32_SOC15(GC, 0, regRLC_RLCS_FED_STATUS_0); + + if (REG_GET_FIELD(rlc_status0, RLC_RLCS_FED_STATUS_0, SDMA0_FED_ERR) || + REG_GET_FIELD(rlc_status0, RLC_RLCS_FED_STATUS_0, SDMA1_FED_ERR)) { + struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); + + ras->gpu_reset_flags |= AMDGPU_RAS_GPU_RESET_MODE2_RESET; + } + amdgpu_ras_reset_gpu(adev); + } return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c index c41219e23151..da6caff78c22 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c @@ -1285,7 +1285,7 @@ static void gfx_v6_0_tiling_mode_table_init(struct amdgpu_device *adev) } static void gfx_v6_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, - u32 sh_num, u32 instance) + u32 sh_num, u32 instance, int xcc_id) { u32 data; @@ -1438,12 +1438,12 @@ static void gfx_v6_0_write_harvested_raster_configs(struct amdgpu_device *adev, } /* GRBM_GFX_INDEX has a different offset on SI */ - gfx_v6_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff); + gfx_v6_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff, 0); WREG32(mmPA_SC_RASTER_CONFIG, raster_config_se); } /* GRBM_GFX_INDEX has a different offset on SI */ - gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); } static void gfx_v6_0_setup_rb(struct amdgpu_device *adev) @@ -1459,14 +1459,14 @@ static void gfx_v6_0_setup_rb(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff, 0); data = gfx_v6_0_get_rb_active_bitmap(adev); active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) * rb_bitmap_width_per_sh); } } - gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); adev->gfx.config.backend_enable_mask = active_rbs; adev->gfx.config.num_rbs = hweight32(active_rbs); @@ -1487,7 +1487,7 @@ static void gfx_v6_0_setup_rb(struct amdgpu_device *adev) /* cache the values for userspace */ for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff, 0); adev->gfx.config.rb_config[i][j].rb_backend_disable = RREG32(mmCC_RB_BACKEND_DISABLE); adev->gfx.config.rb_config[i][j].user_rb_backend_disable = @@ -1496,7 +1496,7 @@ static void gfx_v6_0_setup_rb(struct amdgpu_device *adev) RREG32(mmPA_SC_RASTER_CONFIG); } } - gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); } @@ -1535,7 +1535,7 @@ static void gfx_v6_0_setup_spi(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff, 0); data = RREG32(mmSPI_STATIC_THREAD_MGMT_3); active_cu = gfx_v6_0_get_cu_enabled(adev); @@ -1550,7 +1550,7 @@ static void gfx_v6_0_setup_spi(struct amdgpu_device *adev) } } } - gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); } @@ -2391,7 +2391,7 @@ static void gfx_v6_0_enable_lbpw(struct amdgpu_device *adev, bool enable) WREG32_FIELD(RLC_LB_CNTL, LOAD_BALANCE_ENABLE, enable ? 1 : 0); if (!enable) { - gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); WREG32(mmSPI_LB_CU_MASK, 0x00ff); } } @@ -2968,7 +2968,7 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t simd, *(out++) = RREG32(mmSQ_IND_DATA); } -static void gfx_v6_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) +static void gfx_v6_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) { /* type 0 wave data */ dst[(*no_fields)++] = 0; @@ -2993,7 +2993,7 @@ static void gfx_v6_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, u dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_MODE); } -static void gfx_v6_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd, +static void gfx_v6_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t start, uint32_t size, uint32_t *dst) { @@ -3003,7 +3003,7 @@ static void gfx_v6_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd, } static void gfx_v6_0_select_me_pipe_q(struct amdgpu_device *adev, - u32 me, u32 pipe, u32 q, u32 vm) + u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id) { DRM_INFO("Not implemented\n"); } @@ -3028,6 +3028,7 @@ static int gfx_v6_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + adev->gfx.xcc_mask = 1; adev->gfx.num_gfx_rings = GFX6_NUM_GFX_RINGS; adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev), GFX6_NUM_COMPUTE_RINGS); @@ -3073,7 +3074,7 @@ static int gfx_v6_0_sw_init(void *handle) ring = &adev->gfx.gfx_ring[i]; ring->ring_obj = NULL; sprintf(ring->name, "gfx"); - r = amdgpu_ring_init(adev, ring, 1024, + r = amdgpu_ring_init(adev, ring, 2048, &adev->gfx.eop_irq, AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP, AMDGPU_RING_PRIO_DEFAULT, NULL); @@ -3571,7 +3572,7 @@ static void gfx_v6_0_get_cu_info(struct amdgpu_device *adev) mask = 1; ao_bitmap = 0; counter = 0; - gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff, 0); if (i < 4 && j < 2) gfx_v6_0_set_user_cu_inactive_bitmap( adev, disable_masks[i * 2 + j]); @@ -3593,7 +3594,7 @@ static void gfx_v6_0_get_cu_info(struct amdgpu_device *adev) } } - gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); cu_info->number = active_cu_number; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c index 9d5c1e29b4a3..8c174c11eaee 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c @@ -1548,11 +1548,12 @@ static void gfx_v7_0_tiling_mode_table_init(struct amdgpu_device *adev) * @sh_num: sh block to address * @instance: Certain registers are instanced per SE or SH. * 0xffffffff means broadcast to all SEs or SHs (CIK). - * + * @xcc_id: xcc accelerated compute core id * Select which SE, SH combinations to address. */ static void gfx_v7_0_select_se_sh(struct amdgpu_device *adev, - u32 se_num, u32 sh_num, u32 instance) + u32 se_num, u32 sh_num, u32 instance, + int xcc_id) { u32 data; @@ -1732,13 +1733,13 @@ gfx_v7_0_write_harvested_raster_configs(struct amdgpu_device *adev, } /* GRBM_GFX_INDEX has a different offset on CI+ */ - gfx_v7_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff, 0); WREG32(mmPA_SC_RASTER_CONFIG, raster_config_se); WREG32(mmPA_SC_RASTER_CONFIG_1, raster_config_1); } /* GRBM_GFX_INDEX has a different offset on CI+ */ - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); } /** @@ -1761,13 +1762,13 @@ static void gfx_v7_0_setup_rb(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff, 0); data = gfx_v7_0_get_rb_active_bitmap(adev); active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) * rb_bitmap_width_per_sh); } } - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); adev->gfx.config.backend_enable_mask = active_rbs; adev->gfx.config.num_rbs = hweight32(active_rbs); @@ -1790,7 +1791,7 @@ static void gfx_v7_0_setup_rb(struct amdgpu_device *adev) /* cache the values for userspace */ for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff, 0); adev->gfx.config.rb_config[i][j].rb_backend_disable = RREG32(mmCC_RB_BACKEND_DISABLE); adev->gfx.config.rb_config[i][j].user_rb_backend_disable = @@ -1801,7 +1802,7 @@ static void gfx_v7_0_setup_rb(struct amdgpu_device *adev) RREG32(mmPA_SC_RASTER_CONFIG_1); } } - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); } @@ -1911,7 +1912,7 @@ static void gfx_v7_0_constants_init(struct amdgpu_device *adev) * making sure that the following register writes will be broadcasted * to all the shaders */ - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); /* XXX SH_MEM regs */ /* where to put LDS, scratch, GPUVM in FSA64 space */ @@ -2728,7 +2729,7 @@ static int gfx_v7_0_mec_init(struct amdgpu_device *adev) u32 *hpd; size_t mec_hpd_size; - bitmap_zero(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES); + bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES); /* take ownership of the relevant compute queues */ amdgpu_gfx_compute_queue_acquire(adev); @@ -3301,7 +3302,7 @@ static void gfx_v7_0_wait_for_rlc_serdes(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff, 0); for (k = 0; k < adev->usec_timeout; k++) { if (RREG32(mmRLC_SERDES_CU_MASTER_BUSY) == 0) break; @@ -3309,7 +3310,7 @@ static void gfx_v7_0_wait_for_rlc_serdes(struct amdgpu_device *adev) } } } - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK | @@ -3361,7 +3362,7 @@ static bool gfx_v7_0_is_rlc_enabled(struct amdgpu_device *adev) return true; } -static void gfx_v7_0_set_safe_mode(struct amdgpu_device *adev) +static void gfx_v7_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id) { u32 tmp, i, mask; @@ -3383,7 +3384,7 @@ static void gfx_v7_0_set_safe_mode(struct amdgpu_device *adev) } } -static void gfx_v7_0_unset_safe_mode(struct amdgpu_device *adev) +static void gfx_v7_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id) { u32 tmp; @@ -3474,7 +3475,7 @@ static int gfx_v7_0_rlc_resume(struct amdgpu_device *adev) WREG32(mmRLC_LB_CNTR_MAX, 0x00008000); mutex_lock(&adev->grbm_idx_mutex); - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); WREG32(mmRLC_LB_INIT_CU_MASK, 0xffffffff); WREG32(mmRLC_LB_PARAMS, 0x00600408); WREG32(mmRLC_LB_CNTL, 0x80000004); @@ -3530,7 +3531,7 @@ static void gfx_v7_0_enable_cgcg(struct amdgpu_device *adev, bool enable) tmp = gfx_v7_0_halt_rlc(adev); mutex_lock(&adev->grbm_idx_mutex); - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff); WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff); tmp2 = RLC_SERDES_WR_CTRL__BPM_ADDR_MASK | @@ -3584,7 +3585,7 @@ static void gfx_v7_0_enable_mgcg(struct amdgpu_device *adev, bool enable) tmp = gfx_v7_0_halt_rlc(adev); mutex_lock(&adev->grbm_idx_mutex); - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff); WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff); data = RLC_SERDES_WR_CTRL__BPM_ADDR_MASK | @@ -3635,7 +3636,7 @@ static void gfx_v7_0_enable_mgcg(struct amdgpu_device *adev, bool enable) tmp = gfx_v7_0_halt_rlc(adev); mutex_lock(&adev->grbm_idx_mutex); - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff); WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff); data = RLC_SERDES_WR_CTRL__BPM_ADDR_MASK | RLC_SERDES_WR_CTRL__MGCG_OVERRIDE_1_MASK; @@ -4111,7 +4112,7 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t simd, *(out++) = RREG32(mmSQ_IND_DATA); } -static void gfx_v7_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) +static void gfx_v7_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) { /* type 0 wave data */ dst[(*no_fields)++] = 0; @@ -4136,7 +4137,7 @@ static void gfx_v7_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, u dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_MODE); } -static void gfx_v7_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd, +static void gfx_v7_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t start, uint32_t size, uint32_t *dst) { @@ -4146,7 +4147,7 @@ static void gfx_v7_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd, } static void gfx_v7_0_select_me_pipe_q(struct amdgpu_device *adev, - u32 me, u32 pipe, u32 q, u32 vm) + u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id) { cik_srbm_select(adev, me, pipe, q, vm); } @@ -4178,6 +4179,7 @@ static int gfx_v7_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + adev->gfx.xcc_mask = 1; adev->gfx.num_gfx_rings = GFX7_NUM_GFX_RINGS; adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev), AMDGPU_MAX_COMPUTE_RINGS); @@ -4456,7 +4458,8 @@ static int gfx_v7_0_sw_init(void *handle) for (i = 0; i < adev->gfx.mec.num_mec; ++i) { for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) { for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) { - if (!amdgpu_gfx_is_mec_queue_enabled(adev, i, k, j)) + if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i, + k, j)) continue; r = gfx_v7_0_compute_ring_init(adev, @@ -5114,7 +5117,7 @@ static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev) mask = 1; ao_bitmap = 0; counter = 0; - gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff, 0); if (i < 4 && j < 2) gfx_v7_0_set_user_cu_inactive_bitmap( adev, disable_masks[i * 2 + j]); @@ -5135,7 +5138,7 @@ static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev) cu_info->ao_cu_bitmap[i][j] = ao_bitmap; } } - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); cu_info->number = active_cu_number; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index b1f2684d854a..51c1745c8369 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -1304,7 +1304,7 @@ static int gfx_v8_0_mec_init(struct amdgpu_device *adev) u32 *hpd; size_t mec_hpd_size; - bitmap_zero(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES); + bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES); /* take ownership of the relevant compute queues */ amdgpu_gfx_compute_queue_acquire(adev); @@ -2001,7 +2001,8 @@ static int gfx_v8_0_sw_init(void *handle) for (i = 0; i < adev->gfx.mec.num_mec; ++i) { for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) { for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) { - if (!amdgpu_gfx_is_mec_queue_enabled(adev, i, k, j)) + if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i, + k, j)) continue; r = gfx_v8_0_compute_ring_init(adev, @@ -2015,19 +2016,19 @@ static int gfx_v8_0_sw_init(void *handle) } } - r = amdgpu_gfx_kiq_init(adev, GFX8_MEC_HPD_SIZE); + r = amdgpu_gfx_kiq_init(adev, GFX8_MEC_HPD_SIZE, 0); if (r) { DRM_ERROR("Failed to init KIQ BOs!\n"); return r; } - kiq = &adev->gfx.kiq; - r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq); + kiq = &adev->gfx.kiq[0]; + r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq, 0); if (r) return r; /* create MQD for all compute queues as well as KIQ for SRIOV case */ - r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct vi_mqd_allocation)); + r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct vi_mqd_allocation), 0); if (r) return r; @@ -2050,9 +2051,9 @@ static int gfx_v8_0_sw_fini(void *handle) for (i = 0; i < adev->gfx.num_compute_rings; i++) amdgpu_ring_fini(&adev->gfx.compute_ring[i]); - amdgpu_gfx_mqd_sw_fini(adev); - amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq.ring); - amdgpu_gfx_kiq_fini(adev); + amdgpu_gfx_mqd_sw_fini(adev, 0); + amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq[0].ring); + amdgpu_gfx_kiq_fini(adev, 0); gfx_v8_0_mec_fini(adev); amdgpu_gfx_rlc_fini(adev); @@ -3394,7 +3395,8 @@ static void gfx_v8_0_tiling_mode_table_init(struct amdgpu_device *adev) } static void gfx_v8_0_select_se_sh(struct amdgpu_device *adev, - u32 se_num, u32 sh_num, u32 instance) + u32 se_num, u32 sh_num, u32 instance, + int xcc_id) { u32 data; @@ -3417,7 +3419,7 @@ static void gfx_v8_0_select_se_sh(struct amdgpu_device *adev, } static void gfx_v8_0_select_me_pipe_q(struct amdgpu_device *adev, - u32 me, u32 pipe, u32 q, u32 vm) + u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id) { vi_srbm_select(adev, me, pipe, q, vm); } @@ -3578,13 +3580,13 @@ gfx_v8_0_write_harvested_raster_configs(struct amdgpu_device *adev, } /* GRBM_GFX_INDEX has a different offset on VI */ - gfx_v8_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff, 0); WREG32(mmPA_SC_RASTER_CONFIG, raster_config_se); WREG32(mmPA_SC_RASTER_CONFIG_1, raster_config_1); } /* GRBM_GFX_INDEX has a different offset on VI */ - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); } static void gfx_v8_0_setup_rb(struct amdgpu_device *adev) @@ -3600,13 +3602,13 @@ static void gfx_v8_0_setup_rb(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff, 0); data = gfx_v8_0_get_rb_active_bitmap(adev); active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) * rb_bitmap_width_per_sh); } } - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); adev->gfx.config.backend_enable_mask = active_rbs; adev->gfx.config.num_rbs = hweight32(active_rbs); @@ -3629,7 +3631,7 @@ static void gfx_v8_0_setup_rb(struct amdgpu_device *adev) /* cache the values for userspace */ for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff, 0); adev->gfx.config.rb_config[i][j].rb_backend_disable = RREG32(mmCC_RB_BACKEND_DISABLE); adev->gfx.config.rb_config[i][j].user_rb_backend_disable = @@ -3640,7 +3642,7 @@ static void gfx_v8_0_setup_rb(struct amdgpu_device *adev) RREG32(mmPA_SC_RASTER_CONFIG_1); } } - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); } @@ -3787,7 +3789,7 @@ static void gfx_v8_0_constants_init(struct amdgpu_device *adev) * making sure that the following register writes will be broadcasted * to all the shaders */ - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); WREG32(mmPA_SC_FIFO_SIZE, (adev->gfx.config.sc_prim_fifo_size_frontend << @@ -3818,7 +3820,7 @@ static void gfx_v8_0_wait_for_rlc_serdes(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff, 0); for (k = 0; k < adev->usec_timeout; k++) { if (RREG32(mmRLC_SERDES_CU_MASTER_BUSY) == 0) break; @@ -3826,7 +3828,7 @@ static void gfx_v8_0_wait_for_rlc_serdes(struct amdgpu_device *adev) } if (k == adev->usec_timeout) { gfx_v8_0_select_se_sh(adev, 0xffffffff, - 0xffffffff, 0xffffffff); + 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); DRM_INFO("Timeout wait for RLC serdes %u,%u\n", i, j); @@ -3834,7 +3836,7 @@ static void gfx_v8_0_wait_for_rlc_serdes(struct amdgpu_device *adev) } } } - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK | @@ -4281,7 +4283,6 @@ static int gfx_v8_0_cp_gfx_resume(struct amdgpu_device *adev) /* start the ring */ amdgpu_ring_clear_ring(ring); gfx_v8_0_cp_gfx_start(adev); - ring->sched.ready = true; return 0; } @@ -4292,7 +4293,7 @@ static void gfx_v8_0_cp_compute_enable(struct amdgpu_device *adev, bool enable) WREG32(mmCP_MEC_CNTL, 0); } else { WREG32(mmCP_MEC_CNTL, (CP_MEC_CNTL__MEC_ME1_HALT_MASK | CP_MEC_CNTL__MEC_ME2_HALT_MASK)); - adev->gfx.kiq.ring.sched.ready = false; + adev->gfx.kiq[0].ring.sched.ready = false; } udelay(50); } @@ -4314,12 +4315,12 @@ static void gfx_v8_0_kiq_setting(struct amdgpu_ring *ring) static int gfx_v8_0_kiq_kcq_enable(struct amdgpu_device *adev) { - struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; + struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring; uint64_t queue_mask = 0; int r, i; for (i = 0; i < AMDGPU_MAX_COMPUTE_QUEUES; ++i) { - if (!test_bit(i, adev->gfx.mec.queue_bitmap)) + if (!test_bit(i, adev->gfx.mec_bitmap[0].queue_bitmap)) continue; /* This situation may be hit in the future if a new HW @@ -4595,14 +4596,13 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; struct vi_mqd *mqd = ring->mqd_ptr; - int mqd_idx = AMDGPU_MAX_COMPUTE_RINGS; gfx_v8_0_kiq_setting(ring); if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */ /* reset MQD to a clean status */ - if (adev->gfx.mec.mqd_backup[mqd_idx]) - memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation)); + if (adev->gfx.kiq[0].mqd_backup) + memcpy(mqd, adev->gfx.kiq[0].mqd_backup, sizeof(struct vi_mqd_allocation)); /* reset ring buffer */ ring->wptr = 0; @@ -4625,8 +4625,8 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring) vi_srbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); - if (adev->gfx.mec.mqd_backup[mqd_idx]) - memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct vi_mqd_allocation)); + if (adev->gfx.kiq[0].mqd_backup) + memcpy(adev->gfx.kiq[0].mqd_backup, mqd, sizeof(struct vi_mqd_allocation)); } return 0; @@ -4650,15 +4650,13 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring) if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct vi_mqd_allocation)); - } else if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */ - /* reset MQD to a clean status */ + } else { + /* restore MQD to a clean status */ if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation)); /* reset ring buffer */ ring->wptr = 0; amdgpu_ring_clear_ring(ring); - } else { - amdgpu_ring_clear_ring(ring); } return 0; } @@ -4678,21 +4676,22 @@ static int gfx_v8_0_kiq_resume(struct amdgpu_device *adev) struct amdgpu_ring *ring; int r; - ring = &adev->gfx.kiq.ring; + ring = &adev->gfx.kiq[0].ring; r = amdgpu_bo_reserve(ring->mqd_obj, false); if (unlikely(r != 0)) return r; r = amdgpu_bo_kmap(ring->mqd_obj, &ring->mqd_ptr); - if (unlikely(r != 0)) + if (unlikely(r != 0)) { + amdgpu_bo_unreserve(ring->mqd_obj); return r; + } gfx_v8_0_kiq_init_queue(ring); amdgpu_bo_kunmap(ring->mqd_obj); ring->mqd_ptr = NULL; amdgpu_bo_unreserve(ring->mqd_obj); - ring->sched.ready = true; return 0; } @@ -4741,7 +4740,7 @@ static int gfx_v8_0_cp_test_all_rings(struct amdgpu_device *adev) if (r) return r; - ring = &adev->gfx.kiq.ring; + ring = &adev->gfx.kiq[0].ring; r = amdgpu_ring_test_helper(ring); if (r) return r; @@ -4808,7 +4807,7 @@ static int gfx_v8_0_hw_init(void *handle) static int gfx_v8_0_kcq_disable(struct amdgpu_device *adev) { int r, i; - struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; + struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring; r = amdgpu_ring_alloc(kiq_ring, 6 * adev->gfx.num_compute_rings); if (r) @@ -4902,7 +4901,7 @@ static int gfx_v8_0_hw_fini(void *handle) pr_debug("For SRIOV client, shouldn't do anything.\n"); return 0; } - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); if (!gfx_v8_0_wait_for_idle(adev)) gfx_v8_0_cp_enable(adev, false); else @@ -4911,7 +4910,7 @@ static int gfx_v8_0_hw_fini(void *handle) adev->gfx.rlc.funcs->stop(adev); else pr_err("rlc is busy, skip halt rlc\n"); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return 0; } @@ -5216,7 +5215,7 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t simd, *(out++) = RREG32(mmSQ_IND_DATA); } -static void gfx_v8_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) +static void gfx_v8_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) { /* type 0 wave data */ dst[(*no_fields)++] = 0; @@ -5241,7 +5240,7 @@ static void gfx_v8_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, u dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_MODE); } -static void gfx_v8_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd, +static void gfx_v8_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t start, uint32_t size, uint32_t *dst) { @@ -5263,6 +5262,7 @@ static int gfx_v8_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + adev->gfx.xcc_mask = 1; adev->gfx.num_gfx_rings = GFX8_NUM_GFX_RINGS; adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev), AMDGPU_MAX_COMPUTE_RINGS); @@ -5376,7 +5376,7 @@ static int gfx_v8_0_set_powergating_state(void *handle, AMD_PG_SUPPORT_RLC_SMU_HS | AMD_PG_SUPPORT_CP | AMD_PG_SUPPORT_GFX_DMG)) - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); switch (adev->asic_type) { case CHIP_CARRIZO: case CHIP_STONEY: @@ -5430,7 +5430,7 @@ static int gfx_v8_0_set_powergating_state(void *handle, AMD_PG_SUPPORT_RLC_SMU_HS | AMD_PG_SUPPORT_CP | AMD_PG_SUPPORT_GFX_DMG)) - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return 0; } @@ -5481,7 +5481,7 @@ static void gfx_v8_0_send_serdes_cmd(struct amdgpu_device *adev, { uint32_t data; - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff); WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff); @@ -5535,7 +5535,7 @@ static bool gfx_v8_0_is_rlc_enabled(struct amdgpu_device *adev) return true; } -static void gfx_v8_0_set_safe_mode(struct amdgpu_device *adev) +static void gfx_v8_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id) { uint32_t data; unsigned i; @@ -5562,7 +5562,7 @@ static void gfx_v8_0_set_safe_mode(struct amdgpu_device *adev) } } -static void gfx_v8_0_unset_safe_mode(struct amdgpu_device *adev) +static void gfx_v8_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id) { uint32_t data; unsigned i; @@ -5621,7 +5621,7 @@ static void gfx_v8_0_update_medium_grain_clock_gating(struct amdgpu_device *adev { uint32_t temp, data; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); /* It is disabled by HW by default */ if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) { @@ -5717,7 +5717,7 @@ static void gfx_v8_0_update_medium_grain_clock_gating(struct amdgpu_device *adev gfx_v8_0_wait_for_rlc_serdes(adev); } - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev, @@ -5727,7 +5727,7 @@ static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev temp = data = RREG32(mmRLC_CGCG_CGLS_CTRL); - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)) { temp1 = data1 = RREG32(mmRLC_CGTT_MGCG_OVERRIDE); @@ -5810,7 +5810,7 @@ static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev gfx_v8_0_wait_for_rlc_serdes(adev); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } static int gfx_v8_0_update_gfx_clock_gating(struct amdgpu_device *adev, bool enable) @@ -6723,11 +6723,11 @@ static void gfx_v8_0_parse_sq_irq(struct amdgpu_device *adev, unsigned ih_data, */ if (from_wq) { mutex_lock(&adev->grbm_idx_mutex); - gfx_v8_0_select_se_sh(adev, se_id, sh_id, cu_id); + gfx_v8_0_select_se_sh(adev, se_id, sh_id, cu_id, 0); sq_edc_source = REG_GET_FIELD(RREG32(mmSQ_EDC_INFO), SQ_EDC_INFO, SOURCE); - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); } @@ -7001,7 +7001,7 @@ static void gfx_v8_0_set_ring_funcs(struct amdgpu_device *adev) { int i; - adev->gfx.kiq.ring.funcs = &gfx_v8_0_ring_funcs_kiq; + adev->gfx.kiq[0].ring.funcs = &gfx_v8_0_ring_funcs_kiq; for (i = 0; i < adev->gfx.num_gfx_rings; i++) adev->gfx.gfx_ring[i].funcs = &gfx_v8_0_ring_funcs_gfx; @@ -7116,7 +7116,7 @@ static void gfx_v8_0_get_cu_info(struct amdgpu_device *adev) mask = 1; ao_bitmap = 0; counter = 0; - gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff); + gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff, 0); if (i < 4 && j < 2) gfx_v8_0_set_user_cu_inactive_bitmap( adev, disable_masks[i * 2 + j]); @@ -7137,7 +7137,7 @@ static void gfx_v8_0_get_cu_info(struct amdgpu_device *adev) cu_info->ao_cu_bitmap[i][j] = ao_bitmap; } } - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); cu_info->number = active_cu_number; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index f46d4b18a3fa..65577eca58f1 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -149,16 +149,6 @@ MODULE_FIRMWARE("amdgpu/aldebaran_sjt_mec2.bin"); #define mmGOLDEN_TSC_COUNT_LOWER_Renoir 0x0026 #define mmGOLDEN_TSC_COUNT_LOWER_Renoir_BASE_IDX 1 -#define mmGOLDEN_TSC_COUNT_UPPER_Raven 0x007a -#define mmGOLDEN_TSC_COUNT_UPPER_Raven_BASE_IDX 0 -#define mmGOLDEN_TSC_COUNT_LOWER_Raven 0x007b -#define mmGOLDEN_TSC_COUNT_LOWER_Raven_BASE_IDX 0 - -#define mmGOLDEN_TSC_COUNT_UPPER_Raven2 0x0068 -#define mmGOLDEN_TSC_COUNT_UPPER_Raven2_BASE_IDX 0 -#define mmGOLDEN_TSC_COUNT_LOWER_Raven2 0x0069 -#define mmGOLDEN_TSC_COUNT_LOWER_Raven2_BASE_IDX 0 - enum ta_ras_gfx_subblock { /*CPC*/ TA_RAS_BLOCK__GFX_CPC_INDEX_START = 0, @@ -765,12 +755,12 @@ static void gfx_v9_0_set_rlc_funcs(struct amdgpu_device *adev); static int gfx_v9_0_get_cu_info(struct amdgpu_device *adev, struct amdgpu_cu_info *cu_info); static uint64_t gfx_v9_0_get_gpu_clock_counter(struct amdgpu_device *adev); -static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume); +static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume, bool usegds); static u64 gfx_v9_0_ring_get_rptr_compute(struct amdgpu_ring *ring); static void gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev, void *ras_error_status); static int gfx_v9_0_ras_error_inject(struct amdgpu_device *adev, - void *inject_if); + void *inject_if, uint32_t instance_mask); static void gfx_v9_0_reset_ras_error_count(struct amdgpu_device *adev); static void gfx_v9_0_kiq_set_resources(struct amdgpu_ring *kiq_ring, @@ -898,7 +888,7 @@ static const struct kiq_pm4_funcs gfx_v9_0_kiq_pm4_funcs = { static void gfx_v9_0_set_kiq_pm4_funcs(struct amdgpu_device *adev) { - adev->gfx.kiq.pmf = &gfx_v9_0_kiq_pm4_funcs; + adev->gfx.kiq[0].pmf = &gfx_v9_0_kiq_pm4_funcs; } static void gfx_v9_0_init_golden_registers(struct amdgpu_device *adev) @@ -1504,7 +1494,7 @@ static void gfx_v9_0_init_always_on_cu_mask(struct amdgpu_device *adev) mask = 1; cu_bitmap = 0; counter = 0; - amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff, 0); for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) { if (cu_info->bitmap[i][j] & mask) { @@ -1523,7 +1513,7 @@ static void gfx_v9_0_init_always_on_cu_mask(struct amdgpu_device *adev) cu_info->ao_cu_bitmap[i][j] = cu_bitmap; } } - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); } @@ -1545,7 +1535,7 @@ static void gfx_v9_0_init_lbpw(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); /* set mmRLC_LB_INIT_CU_MASK thru broadcast mode to enable all SE/SH*/ - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); WREG32_SOC15(GC, 0, mmRLC_LB_INIT_CU_MASK, 0xffffffff); /* set mmRLC_LB_PARAMS = 0x003F_1006 */ @@ -1594,7 +1584,7 @@ static void gfx_v9_4_init_lbpw(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); /* set mmRLC_LB_INIT_CU_MASK thru broadcast mode to enable all SE/SH*/ - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); WREG32_SOC15(GC, 0, mmRLC_LB_INIT_CU_MASK, 0xffffffff); /* set mmRLC_LB_PARAMS = 0x003F_1006 */ @@ -1713,7 +1703,7 @@ static int gfx_v9_0_mec_init(struct amdgpu_device *adev) const struct gfx_firmware_header_v1_0 *mec_hdr; - bitmap_zero(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES); + bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES); /* take ownership of the relevant compute queues */ amdgpu_gfx_compute_queue_acquire(adev); @@ -1788,7 +1778,7 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t simd, *(out++) = RREG32_SOC15(GC, 0, mmSQ_IND_DATA); } -static void gfx_v9_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) +static void gfx_v9_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) { /* type 1 wave data */ dst[(*no_fields)++] = 1; @@ -1809,7 +1799,7 @@ static void gfx_v9_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, u dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_MODE); } -static void gfx_v9_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd, +static void gfx_v9_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t start, uint32_t size, uint32_t *dst) { @@ -1818,7 +1808,7 @@ static void gfx_v9_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd, start + SQIND_WAVE_SGPRS_OFFSET, size, dst); } -static void gfx_v9_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd, +static void gfx_v9_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t thread, uint32_t start, uint32_t size, uint32_t *dst) @@ -1829,9 +1819,9 @@ static void gfx_v9_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd, } static void gfx_v9_0_select_me_pipe_q(struct amdgpu_device *adev, - u32 me, u32 pipe, u32 q, u32 vm) + u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id) { - soc15_grbm_select(adev, me, pipe, q, vm); + soc15_grbm_select(adev, me, pipe, q, vm, 0); } static const struct amdgpu_gfx_funcs gfx_v9_0_gfx_funcs = { @@ -2005,7 +1995,7 @@ static int gfx_v9_0_compute_ring_init(struct amdgpu_device *adev, int ring_id, ring->doorbell_index = (adev->doorbell_index.mec_ring0 + ring_id) << 1; ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr + (ring_id * GFX9_MEC_HPD_SIZE); - ring->vm_hub = AMDGPU_GFXHUB_0; + ring->vm_hub = AMDGPU_GFXHUB(0); sprintf(ring->name, "comp_%d.%d.%d", ring->me, ring->pipe, ring->queue); irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP @@ -2105,7 +2095,7 @@ static int gfx_v9_0_sw_init(void *handle) /* disable scheduler on the real ring */ ring->no_scheduler = true; - ring->vm_hub = AMDGPU_GFXHUB_0; + ring->vm_hub = AMDGPU_GFXHUB(0); r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP, AMDGPU_RING_PRIO_DEFAULT, NULL); @@ -2123,7 +2113,7 @@ static int gfx_v9_0_sw_init(void *handle) ring->doorbell_index = adev->doorbell_index.gfx_ring0 << 1; ring->is_sw_ring = true; hw_prio = amdgpu_sw_ring_priority(i); - ring->vm_hub = AMDGPU_GFXHUB_0; + ring->vm_hub = AMDGPU_GFXHUB(0); r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP, hw_prio, NULL); @@ -2154,7 +2144,8 @@ static int gfx_v9_0_sw_init(void *handle) for (i = 0; i < adev->gfx.mec.num_mec; ++i) { for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) { for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) { - if (!amdgpu_gfx_is_mec_queue_enabled(adev, i, k, j)) + if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i, + k, j)) continue; r = gfx_v9_0_compute_ring_init(adev, @@ -2168,19 +2159,19 @@ static int gfx_v9_0_sw_init(void *handle) } } - r = amdgpu_gfx_kiq_init(adev, GFX9_MEC_HPD_SIZE); + r = amdgpu_gfx_kiq_init(adev, GFX9_MEC_HPD_SIZE, 0); if (r) { DRM_ERROR("Failed to init KIQ BOs!\n"); return r; } - kiq = &adev->gfx.kiq; - r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq); + kiq = &adev->gfx.kiq[0]; + r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq, 0); if (r) return r; /* create MQD for all compute queues as wel as KIQ for SRIOV case */ - r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v9_mqd_allocation)); + r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v9_mqd_allocation), 0); if (r) return r; @@ -2215,9 +2206,9 @@ static int gfx_v9_0_sw_fini(void *handle) for (i = 0; i < adev->gfx.num_compute_rings; i++) amdgpu_ring_fini(&adev->gfx.compute_ring[i]); - amdgpu_gfx_mqd_sw_fini(adev); - amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq.ring); - amdgpu_gfx_kiq_fini(adev); + amdgpu_gfx_mqd_sw_fini(adev, 0); + amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq[0].ring); + amdgpu_gfx_kiq_fini(adev, 0); gfx_v9_0_mec_fini(adev); amdgpu_bo_free_kernel(&adev->gfx.rlc.clear_state_obj, @@ -2240,7 +2231,7 @@ static void gfx_v9_0_tiling_mode_table_init(struct amdgpu_device *adev) } void gfx_v9_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num, - u32 instance) + u32 instance, int xcc_id) { u32 data; @@ -2289,19 +2280,42 @@ static void gfx_v9_0_setup_rb(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff, 0); data = gfx_v9_0_get_rb_active_bitmap(adev); active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) * rb_bitmap_width_per_sh); } } - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); adev->gfx.config.backend_enable_mask = active_rbs; adev->gfx.config.num_rbs = hweight32(active_rbs); } +static void gfx_v9_0_debug_trap_config_init(struct amdgpu_device *adev, + uint32_t first_vmid, + uint32_t last_vmid) +{ + uint32_t data; + uint32_t trap_config_vmid_mask = 0; + int i; + + /* Calculate trap config vmid mask */ + for (i = first_vmid; i < last_vmid; i++) + trap_config_vmid_mask |= (1 << i); + + data = REG_SET_FIELD(0, SPI_GDBG_TRAP_CONFIG, + VMID_SEL, trap_config_vmid_mask); + data = REG_SET_FIELD(data, SPI_GDBG_TRAP_CONFIG, + TRAP_EN, 1); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_CONFIG), data); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0); + + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA0), 0); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA1), 0); +} + #define DEFAULT_SH_MEM_BASES (0x6000) static void gfx_v9_0_init_compute_vmid(struct amdgpu_device *adev) { @@ -2323,12 +2337,12 @@ static void gfx_v9_0_init_compute_vmid(struct amdgpu_device *adev) mutex_lock(&adev->srbm_mutex); for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) { - soc15_grbm_select(adev, 0, 0, 0, i); + soc15_grbm_select(adev, 0, 0, 0, i, 0); /* CP and shaders */ WREG32_SOC15_RLC(GC, 0, mmSH_MEM_CONFIG, sh_mem_config); WREG32_SOC15_RLC(GC, 0, mmSH_MEM_BASES, sh_mem_bases); } - soc15_grbm_select(adev, 0, 0, 0, 0); + soc15_grbm_select(adev, 0, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); /* Initialize all compute VMIDs to have no GDS, GWS, or OA @@ -2366,8 +2380,8 @@ static void gfx_v9_0_init_sq_config(struct amdgpu_device *adev) switch (adev->ip_versions[GC_HWIP][0]) { case IP_VERSION(9, 4, 1): tmp = RREG32_SOC15(GC, 0, mmSQ_CONFIG); - tmp = REG_SET_FIELD(tmp, SQ_CONFIG, - DISABLE_BARRIER_WAITCNT, 1); + tmp = REG_SET_FIELD(tmp, SQ_CONFIG, DISABLE_BARRIER_WAITCNT, + !READ_ONCE(adev->barrier_has_auto_waitcnt)); WREG32_SOC15(GC, 0, mmSQ_CONFIG, tmp); break; default: @@ -2392,8 +2406,8 @@ static void gfx_v9_0_constants_init(struct amdgpu_device *adev) /* XXX SH_MEM regs */ /* where to put LDS, scratch, GPUVM in FSA64 space */ mutex_lock(&adev->srbm_mutex); - for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB_0].num_ids; i++) { - soc15_grbm_select(adev, 0, 0, 0, i); + for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB(0)].num_ids; i++) { + soc15_grbm_select(adev, 0, 0, 0, i, 0); /* CP and shaders */ if (i == 0) { tmp = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE, @@ -2415,7 +2429,7 @@ static void gfx_v9_0_constants_init(struct amdgpu_device *adev) WREG32_SOC15_RLC(GC, 0, mmSH_MEM_BASES, tmp); } } - soc15_grbm_select(adev, 0, 0, 0, 0); + soc15_grbm_select(adev, 0, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); @@ -2432,7 +2446,7 @@ static void gfx_v9_0_wait_for_rlc_serdes(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff, 0); for (k = 0; k < adev->usec_timeout; k++) { if (RREG32_SOC15(GC, 0, mmRLC_SERDES_CU_MASTER_BUSY) == 0) break; @@ -2440,7 +2454,7 @@ static void gfx_v9_0_wait_for_rlc_serdes(struct amdgpu_device *adev) } if (k == adev->usec_timeout) { amdgpu_gfx_select_se_sh(adev, 0xffffffff, - 0xffffffff, 0xffffffff); + 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); DRM_INFO("Timeout wait for RLC serdes %u,%u\n", i, j); @@ -2448,7 +2462,7 @@ static void gfx_v9_0_wait_for_rlc_serdes(struct amdgpu_device *adev) } } } - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK | @@ -3143,7 +3157,6 @@ static int gfx_v9_0_cp_gfx_resume(struct amdgpu_device *adev) /* start the ring */ gfx_v9_0_cp_gfx_start(adev); - ring->sched.ready = true; return 0; } @@ -3155,7 +3168,7 @@ static void gfx_v9_0_cp_compute_enable(struct amdgpu_device *adev, bool enable) } else { WREG32_SOC15_RLC(GC, 0, mmCP_MEC_CNTL, (CP_MEC_CNTL__MEC_ME1_HALT_MASK | CP_MEC_CNTL__MEC_ME2_HALT_MASK)); - adev->gfx.kiq.ring.sched.ready = false; + adev->gfx.kiq[0].ring.sched.ready = false; } udelay(50); } @@ -3519,7 +3532,6 @@ static int gfx_v9_0_kiq_init_queue(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; struct v9_mqd *mqd = ring->mqd_ptr; - int mqd_idx = AMDGPU_MAX_COMPUTE_RINGS; struct v9_mqd *tmp_mqd; gfx_v9_0_kiq_setting(ring); @@ -3529,20 +3541,20 @@ static int gfx_v9_0_kiq_init_queue(struct amdgpu_ring *ring) * driver need to re-init the mqd. * check mqd->cp_hqd_pq_control since this value should not be 0 */ - tmp_mqd = (struct v9_mqd *)adev->gfx.mec.mqd_backup[mqd_idx]; + tmp_mqd = (struct v9_mqd *)adev->gfx.kiq[0].mqd_backup; if (amdgpu_in_reset(adev) && tmp_mqd->cp_hqd_pq_control){ /* for GPU_RESET case , reset MQD to a clean status */ - if (adev->gfx.mec.mqd_backup[mqd_idx]) - memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation)); + if (adev->gfx.kiq[0].mqd_backup) + memcpy(mqd, adev->gfx.kiq[0].mqd_backup, sizeof(struct v9_mqd_allocation)); /* reset ring buffer */ ring->wptr = 0; amdgpu_ring_clear_ring(ring); mutex_lock(&adev->srbm_mutex); - soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); + soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, 0); gfx_v9_0_kiq_init_register(ring); - soc15_grbm_select(adev, 0, 0, 0, 0); + soc15_grbm_select(adev, 0, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); } else { memset((void *)mqd, 0, sizeof(struct v9_mqd_allocation)); @@ -3551,14 +3563,14 @@ static int gfx_v9_0_kiq_init_queue(struct amdgpu_ring *ring) if (amdgpu_sriov_vf(adev) && adev->in_suspend) amdgpu_ring_clear_ring(ring); mutex_lock(&adev->srbm_mutex); - soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); + soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, 0); gfx_v9_0_mqd_init(ring); gfx_v9_0_kiq_init_register(ring); - soc15_grbm_select(adev, 0, 0, 0, 0); + soc15_grbm_select(adev, 0, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); - if (adev->gfx.mec.mqd_backup[mqd_idx]) - memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct v9_mqd_allocation)); + if (adev->gfx.kiq[0].mqd_backup) + memcpy(adev->gfx.kiq[0].mqd_backup, mqd, sizeof(struct v9_mqd_allocation)); } return 0; @@ -3582,24 +3594,21 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring) ((struct v9_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF; ((struct v9_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF; mutex_lock(&adev->srbm_mutex); - soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); + soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, 0); gfx_v9_0_mqd_init(ring); - soc15_grbm_select(adev, 0, 0, 0, 0); + soc15_grbm_select(adev, 0, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct v9_mqd_allocation)); - } else if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */ - /* reset MQD to a clean status */ + } else { + /* restore MQD to a clean status */ if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation)); - /* reset ring buffer */ ring->wptr = 0; atomic64_set((atomic64_t *)ring->wptr_cpu_addr, 0); amdgpu_ring_clear_ring(ring); - } else { - amdgpu_ring_clear_ring(ring); } return 0; @@ -3610,21 +3619,22 @@ static int gfx_v9_0_kiq_resume(struct amdgpu_device *adev) struct amdgpu_ring *ring; int r; - ring = &adev->gfx.kiq.ring; + ring = &adev->gfx.kiq[0].ring; r = amdgpu_bo_reserve(ring->mqd_obj, false); if (unlikely(r != 0)) return r; r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); - if (unlikely(r != 0)) + if (unlikely(r != 0)) { + amdgpu_bo_unreserve(ring->mqd_obj); return r; + } gfx_v9_0_kiq_init_queue(ring); amdgpu_bo_kunmap(ring->mqd_obj); ring->mqd_ptr = NULL; amdgpu_bo_unreserve(ring->mqd_obj); - ring->sched.ready = true; return 0; } @@ -3652,7 +3662,7 @@ static int gfx_v9_0_kcq_resume(struct amdgpu_device *adev) goto done; } - r = amdgpu_gfx_enable_kcq(adev); + r = amdgpu_gfx_enable_kcq(adev, 0); done: return r; } @@ -3772,7 +3782,7 @@ static int gfx_v9_0_hw_fini(void *handle) /* DF freeze and kcq disable will fail */ if (!amdgpu_ras_intr_triggered()) /* disable KCQ to avoid CPC touch memory not valid anymore */ - amdgpu_gfx_disable_kcq(adev); + amdgpu_gfx_disable_kcq(adev, 0); if (amdgpu_sriov_vf(adev)) { gfx_v9_0_cp_gfx_enable(adev, false); @@ -3790,11 +3800,11 @@ static int gfx_v9_0_hw_fini(void *handle) */ if (!amdgpu_in_reset(adev) && !adev->in_suspend) { mutex_lock(&adev->srbm_mutex); - soc15_grbm_select(adev, adev->gfx.kiq.ring.me, - adev->gfx.kiq.ring.pipe, - adev->gfx.kiq.ring.queue, 0); - gfx_v9_0_kiq_fini_register(&adev->gfx.kiq.ring); - soc15_grbm_select(adev, 0, 0, 0, 0); + soc15_grbm_select(adev, adev->gfx.kiq[0].ring.me, + adev->gfx.kiq[0].ring.pipe, + adev->gfx.kiq[0].ring.queue, 0, 0); + gfx_v9_0_kiq_fini_register(&adev->gfx.kiq[0].ring); + soc15_grbm_select(adev, 0, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); } @@ -3914,7 +3924,7 @@ static uint64_t gfx_v9_0_kiq_read_clock(struct amdgpu_device *adev) unsigned long flags; uint32_t seq, reg_val_offs = 0; uint64_t value = 0; - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; struct amdgpu_ring *ring = &kiq->ring; BUG_ON(!ring->funcs->emit_rreg); @@ -4002,36 +4012,6 @@ static uint64_t gfx_v9_0_get_gpu_clock_counter(struct amdgpu_device *adev) preempt_enable(); clock = clock_lo | (clock_hi << 32ULL); break; - case IP_VERSION(9, 1, 0): - preempt_disable(); - clock_hi = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_UPPER_Raven); - clock_lo = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_LOWER_Raven); - hi_check = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_UPPER_Raven); - /* The PWR TSC clock frequency is 100MHz, which sets 32-bit carry over - * roughly every 42 seconds. - */ - if (hi_check != clock_hi) { - clock_lo = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_LOWER_Raven); - clock_hi = hi_check; - } - preempt_enable(); - clock = clock_lo | (clock_hi << 32ULL); - break; - case IP_VERSION(9, 2, 2): - preempt_disable(); - clock_hi = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_UPPER_Raven2); - clock_lo = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_LOWER_Raven2); - hi_check = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_UPPER_Raven2); - /* The PWR TSC clock frequency is 100MHz, which sets 32-bit carry over - * roughly every 42 seconds. - */ - if (hi_check != clock_hi) { - clock_lo = RREG32_SOC15_NO_KIQ(PWR, 0, mmGOLDEN_TSC_COUNT_LOWER_Raven2); - clock_hi = hi_check; - } - preempt_enable(); - clock = clock_lo | (clock_hi << 32ULL); - break; default: amdgpu_gfx_off_ctrl(adev, false); mutex_lock(&adev->gfx.gpu_clock_mutex); @@ -4544,6 +4524,7 @@ static int gfx_v9_0_early_init(void *handle) adev->gfx.num_gfx_rings = 0; else adev->gfx.num_gfx_rings = GFX9_NUM_GFX_RINGS; + adev->gfx.xcc_mask = 1; adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev), AMDGPU_MAX_COMPUTE_RINGS); gfx_v9_0_set_kiq_pm4_funcs(adev); @@ -4609,6 +4590,13 @@ static int gfx_v9_0_late_init(void *handle) if (r) return r; + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) + gfx_v9_4_2_debug_trap_config_init(adev, + adev->vm_manager.first_kfd_vmid, AMDGPU_NUM_VMID); + else + gfx_v9_0_debug_trap_config_init(adev, + adev->vm_manager.first_kfd_vmid, AMDGPU_NUM_VMID); + return 0; } @@ -4624,7 +4612,7 @@ static bool gfx_v9_0_is_rlc_enabled(struct amdgpu_device *adev) return true; } -static void gfx_v9_0_set_safe_mode(struct amdgpu_device *adev) +static void gfx_v9_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id) { uint32_t data; unsigned i; @@ -4641,7 +4629,7 @@ static void gfx_v9_0_set_safe_mode(struct amdgpu_device *adev) } } -static void gfx_v9_0_unset_safe_mode(struct amdgpu_device *adev) +static void gfx_v9_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id) { uint32_t data; @@ -4652,7 +4640,7 @@ static void gfx_v9_0_unset_safe_mode(struct amdgpu_device *adev) static void gfx_v9_0_update_gfx_cg_power_gating(struct amdgpu_device *adev, bool enable) { - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) && enable) { gfx_v9_0_enable_gfx_cg_power_gating(adev, true); @@ -4664,7 +4652,7 @@ static void gfx_v9_0_update_gfx_cg_power_gating(struct amdgpu_device *adev, gfx_v9_0_enable_gfx_pipeline_powergating(adev, false); } - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } static void gfx_v9_0_update_gfx_mg_power_gating(struct amdgpu_device *adev, @@ -4691,7 +4679,7 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev { uint32_t data, def; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); /* It is disabled by HW by default */ if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) { @@ -4758,7 +4746,7 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev } } - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev, @@ -4769,7 +4757,7 @@ static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev, if (!adev->gfx.num_gfx_rings) return; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); /* Enable 3D CGCG/CGLS */ if (enable) { @@ -4813,7 +4801,7 @@ static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev, WREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL_3D, data); } - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev, @@ -4821,7 +4809,7 @@ static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev { uint32_t def, data; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)) { def = data = RREG32_SOC15(GC, 0, mmRLC_CGTT_MGCG_OVERRIDE); @@ -4865,7 +4853,7 @@ static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev WREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL, data); } - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } static int gfx_v9_0_update_gfx_clock_gating(struct amdgpu_device *adev, @@ -5165,7 +5153,8 @@ static void gfx_v9_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, gfx_v9_0_ring_emit_de_meta(ring, (!amdgpu_sriov_vf(ring->adev) && flags & AMDGPU_IB_PREEMPTED) ? - true : false); + true : false, + job->gds_size > 0 && job->gds_base != 0); } amdgpu_ring_write(ring, header); @@ -5176,9 +5165,83 @@ static void gfx_v9_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, #endif lower_32_bits(ib->gpu_addr)); amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr)); + amdgpu_ring_ib_on_emit_cntl(ring); amdgpu_ring_write(ring, control); } +static void gfx_v9_0_ring_patch_cntl(struct amdgpu_ring *ring, + unsigned offset) +{ + u32 control = ring->ring[offset]; + + control |= INDIRECT_BUFFER_PRE_RESUME(1); + ring->ring[offset] = control; +} + +static void gfx_v9_0_ring_patch_ce_meta(struct amdgpu_ring *ring, + unsigned offset) +{ + struct amdgpu_device *adev = ring->adev; + void *ce_payload_cpu_addr; + uint64_t payload_offset, payload_size; + + payload_size = sizeof(struct v9_ce_ib_state); + + if (ring->is_mes_queue) { + payload_offset = offsetof(struct amdgpu_mes_ctx_meta_data, + gfx[0].gfx_meta_data) + + offsetof(struct v9_gfx_meta_data, ce_payload); + ce_payload_cpu_addr = + amdgpu_mes_ctx_get_offs_cpu_addr(ring, payload_offset); + } else { + payload_offset = offsetof(struct v9_gfx_meta_data, ce_payload); + ce_payload_cpu_addr = adev->virt.csa_cpu_addr + payload_offset; + } + + if (offset + (payload_size >> 2) <= ring->buf_mask + 1) { + memcpy((void *)&ring->ring[offset], ce_payload_cpu_addr, payload_size); + } else { + memcpy((void *)&ring->ring[offset], ce_payload_cpu_addr, + (ring->buf_mask + 1 - offset) << 2); + payload_size -= (ring->buf_mask + 1 - offset) << 2; + memcpy((void *)&ring->ring[0], + ce_payload_cpu_addr + ((ring->buf_mask + 1 - offset) << 2), + payload_size); + } +} + +static void gfx_v9_0_ring_patch_de_meta(struct amdgpu_ring *ring, + unsigned offset) +{ + struct amdgpu_device *adev = ring->adev; + void *de_payload_cpu_addr; + uint64_t payload_offset, payload_size; + + payload_size = sizeof(struct v9_de_ib_state); + + if (ring->is_mes_queue) { + payload_offset = offsetof(struct amdgpu_mes_ctx_meta_data, + gfx[0].gfx_meta_data) + + offsetof(struct v9_gfx_meta_data, de_payload); + de_payload_cpu_addr = + amdgpu_mes_ctx_get_offs_cpu_addr(ring, payload_offset); + } else { + payload_offset = offsetof(struct v9_gfx_meta_data, de_payload); + de_payload_cpu_addr = adev->virt.csa_cpu_addr + payload_offset; + } + + if (offset + (payload_size >> 2) <= ring->buf_mask + 1) { + memcpy((void *)&ring->ring[offset], de_payload_cpu_addr, payload_size); + } else { + memcpy((void *)&ring->ring[offset], de_payload_cpu_addr, + (ring->buf_mask + 1 - offset) << 2); + payload_size -= (ring->buf_mask + 1 - offset) << 2; + memcpy((void *)&ring->ring[0], + de_payload_cpu_addr + ((ring->buf_mask + 1 - offset) << 2), + payload_size); + } +} + static void gfx_v9_0_ring_emit_ib_compute(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, @@ -5374,6 +5437,8 @@ static void gfx_v9_0_ring_emit_ce_meta(struct amdgpu_ring *ring, bool resume) amdgpu_ring_write(ring, lower_32_bits(ce_payload_gpu_addr)); amdgpu_ring_write(ring, upper_32_bits(ce_payload_gpu_addr)); + amdgpu_ring_ib_on_emit_ce(ring); + if (resume) amdgpu_ring_write_multiple(ring, ce_payload_cpu_addr, sizeof(ce_payload) >> 2); @@ -5386,7 +5451,7 @@ static int gfx_v9_0_ring_preempt_ib(struct amdgpu_ring *ring) { int i, r = 0; struct amdgpu_device *adev = ring->adev; - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; struct amdgpu_ring *kiq_ring = &kiq->ring; unsigned long flags; @@ -5407,10 +5472,6 @@ static int gfx_v9_0_ring_preempt_ib(struct amdgpu_ring *ring) amdgpu_ring_alloc(ring, 13); gfx_v9_0_ring_emit_fence(ring, ring->trail_fence_gpu_addr, ring->trail_seq, AMDGPU_FENCE_FLAG_EXEC | AMDGPU_FENCE_FLAG_INT); - /*reset the CP_VMID_PREEMPT after trailing fence*/ - amdgpu_ring_emit_wreg(ring, - SOC15_REG_OFFSET(GC, 0, mmCP_VMID_PREEMPT), - 0x0); /* assert IB preemption, emit the trailing fence */ kiq->pmf->kiq_unmap_queues(kiq_ring, ring, PREEMPT_QUEUES_NO_UNMAP, @@ -5433,6 +5494,10 @@ static int gfx_v9_0_ring_preempt_ib(struct amdgpu_ring *ring) DRM_WARN("ring %d timeout to preempt ib\n", ring->idx); } + /*reset the CP_VMID_PREEMPT after trailing fence*/ + amdgpu_ring_emit_wreg(ring, + SOC15_REG_OFFSET(GC, 0, mmCP_VMID_PREEMPT), + 0x0); amdgpu_ring_commit(ring); /* deassert preemption condition */ @@ -5440,7 +5505,7 @@ static int gfx_v9_0_ring_preempt_ib(struct amdgpu_ring *ring) return r; } -static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume) +static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume, bool usegds) { struct amdgpu_device *adev = ring->adev; struct v9_de_ib_state de_payload = {0}; @@ -5471,8 +5536,10 @@ static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume) PAGE_SIZE); } - de_payload.gds_backup_addrlo = lower_32_bits(gds_addr); - de_payload.gds_backup_addrhi = upper_32_bits(gds_addr); + if (usegds) { + de_payload.gds_backup_addrlo = lower_32_bits(gds_addr); + de_payload.gds_backup_addrhi = upper_32_bits(gds_addr); + } cnt = (sizeof(de_payload) >> 2) + 4 - 2; amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, cnt)); @@ -5483,6 +5550,7 @@ static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume) amdgpu_ring_write(ring, lower_32_bits(de_payload_gpu_addr)); amdgpu_ring_write(ring, upper_32_bits(de_payload_gpu_addr)); + amdgpu_ring_ib_on_emit_de(ring); if (resume) amdgpu_ring_write_multiple(ring, de_payload_cpu_addr, sizeof(de_payload) >> 2); @@ -6342,7 +6410,7 @@ static const struct soc15_ras_field_entry gfx_v9_0_ras_fields[] = { }; static int gfx_v9_0_ras_error_inject(struct amdgpu_device *adev, - void *inject_if) + void *inject_if, uint32_t instance_mask) { struct ras_inject_if *info = (struct ras_inject_if *)inject_if; int ret; @@ -6381,7 +6449,7 @@ static int gfx_v9_0_ras_error_inject(struct amdgpu_device *adev, block_info.value = info->value; mutex_lock(&adev->grbm_idx_mutex); - ret = psp_ras_trigger_error(&adev->psp, &block_info); + ret = psp_ras_trigger_error(&adev->psp, &block_info, instance_mask); mutex_unlock(&adev->grbm_idx_mutex); return ret; @@ -6609,7 +6677,7 @@ static void gfx_v9_0_reset_ras_error_count(struct amdgpu_device *adev) for (i = 0; i < ARRAY_SIZE(gfx_v9_0_edc_counter_regs); i++) { for (j = 0; j < gfx_v9_0_edc_counter_regs[i].se_num; j++) { for (k = 0; k < gfx_v9_0_edc_counter_regs[i].instance; k++) { - amdgpu_gfx_select_se_sh(adev, j, 0x0, k); + amdgpu_gfx_select_se_sh(adev, j, 0x0, k, 0); RREG32(SOC15_REG_ENTRY_OFFSET(gfx_v9_0_edc_counter_regs[i])); } } @@ -6671,7 +6739,7 @@ static void gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev, for (i = 0; i < ARRAY_SIZE(gfx_v9_0_edc_counter_regs); i++) { for (j = 0; j < gfx_v9_0_edc_counter_regs[i].se_num; j++) { for (k = 0; k < gfx_v9_0_edc_counter_regs[i].instance; k++) { - amdgpu_gfx_select_se_sh(adev, j, 0, k); + amdgpu_gfx_select_se_sh(adev, j, 0, k, 0); reg_value = RREG32(SOC15_REG_ENTRY_OFFSET(gfx_v9_0_edc_counter_regs[i])); if (reg_value) @@ -6686,7 +6754,7 @@ static void gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev, err_data->ce_count += sec_count; err_data->ue_count += ded_count; - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); gfx_v9_0_query_utc_edc_status(adev, err_data); @@ -6893,6 +6961,9 @@ static const struct amdgpu_ring_funcs gfx_v9_0_sw_ring_funcs_gfx = { .emit_reg_write_reg_wait = gfx_v9_0_ring_emit_reg_write_reg_wait, .soft_recovery = gfx_v9_0_ring_soft_recovery, .emit_mem_sync = gfx_v9_0_emit_mem_sync, + .patch_cntl = gfx_v9_0_ring_patch_cntl, + .patch_de = gfx_v9_0_ring_patch_de_meta, + .patch_ce = gfx_v9_0_ring_patch_ce_meta, }; static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = { @@ -6965,7 +7036,7 @@ static void gfx_v9_0_set_ring_funcs(struct amdgpu_device *adev) { int i; - adev->gfx.kiq.ring.funcs = &gfx_v9_0_ring_funcs_kiq; + adev->gfx.kiq[0].ring.funcs = &gfx_v9_0_ring_funcs_kiq; for (i = 0; i < adev->gfx.num_gfx_rings; i++) adev->gfx.gfx_ring[i].funcs = &gfx_v9_0_ring_funcs_gfx; @@ -7146,7 +7217,7 @@ static int gfx_v9_0_get_cu_info(struct amdgpu_device *adev, mask = 1; ao_bitmap = 0; counter = 0; - amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff, 0); gfx_v9_0_set_user_cu_inactive_bitmap( adev, disable_masks[i * adev->gfx.config.max_sh_per_se + j]); bitmap = gfx_v9_0_get_cu_active_bitmap(adev); @@ -7179,7 +7250,7 @@ static int gfx_v9_0_get_cu_info(struct amdgpu_device *adev, cu_info->ao_cu_bitmap[i % 4][j + i / 4] = ao_bitmap; } } - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); cu_info->number = active_cu_number; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.h b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.h index dfe8d4841f58..f9f6edc5e558 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.h +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.h @@ -27,6 +27,6 @@ extern const struct amdgpu_ip_block_version gfx_v9_0_ip_block; void gfx_v9_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num, - u32 instance); + u32 instance, int xcc_id); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c index c67e387a97f5..bc8416afb62c 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c @@ -970,29 +970,6 @@ static void gfx_v9_4_reset_ras_error_count(struct amdgpu_device *adev) WREG32_SOC15(GC, 0, mmATC_L2_CACHE_4K_DSM_INDEX, 255); } -static int gfx_v9_4_ras_error_inject(struct amdgpu_device *adev, - void *inject_if) -{ - struct ras_inject_if *info = (struct ras_inject_if *)inject_if; - int ret; - struct ta_ras_trigger_error_input block_info = { 0 }; - - if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX)) - return -EINVAL; - - block_info.block_id = amdgpu_ras_block_to_ta(info->head.block); - block_info.sub_block_index = info->head.sub_block_index; - block_info.inject_error_type = amdgpu_ras_error_to_ta(info->head.type); - block_info.address = info->address; - block_info.value = info->value; - - mutex_lock(&adev->grbm_idx_mutex); - ret = psp_ras_trigger_error(&adev->psp, &block_info); - mutex_unlock(&adev->grbm_idx_mutex); - - return ret; -} - static const struct soc15_reg_entry gfx_v9_4_ea_err_status_regs = { SOC15_REG_ENTRY(GC, 0, mmGCEA_ERR_STATUS), 0, 1, 32 }; @@ -1030,7 +1007,6 @@ static void gfx_v9_4_query_ras_error_status(struct amdgpu_device *adev) const struct amdgpu_ras_block_hw_ops gfx_v9_4_ras_ops = { - .ras_error_inject = &gfx_v9_4_ras_error_inject, .query_ras_error_count = &gfx_v9_4_query_ras_error_count, .reset_ras_error_count = &gfx_v9_4_reset_ras_error_count, .query_ras_error_status = &gfx_v9_4_query_ras_error_status, diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c index 3a797424579c..63f6843a069e 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c @@ -761,7 +761,7 @@ void gfx_v9_4_2_debug_trap_config_init(struct amdgpu_device *adev, for (i = first_vmid; i < last_vmid; i++) { data = 0; - soc15_grbm_select(adev, 0, 0, 0, i); + soc15_grbm_select(adev, 0, 0, 0, i, 0); data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1); data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0); data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, @@ -769,15 +769,18 @@ void gfx_v9_4_2_debug_trap_config_init(struct amdgpu_device *adev, WREG32(SOC15_REG_OFFSET(GC, 0, regSPI_GDBG_PER_VMID_CNTL), data); } - soc15_grbm_select(adev, 0, 0, 0, 0); + soc15_grbm_select(adev, 0, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); + + WREG32(SOC15_REG_OFFSET(GC, 0, regSPI_GDBG_TRAP_DATA0), 0); + WREG32(SOC15_REG_OFFSET(GC, 0, regSPI_GDBG_TRAP_DATA1), 0); } void gfx_v9_4_2_set_power_brake_sequence(struct amdgpu_device *adev) { u32 tmp; - gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); tmp = 0; tmp = REG_SET_FIELD(tmp, GC_THROTTLE_CTRL, PATTERN_MODE, 1); @@ -1699,28 +1702,6 @@ static void gfx_v9_4_2_reset_ras_error_count(struct amdgpu_device *adev) gfx_v9_4_2_query_utc_edc_count(adev, NULL, NULL); } -static int gfx_v9_4_2_ras_error_inject(struct amdgpu_device *adev, void *inject_if) -{ - struct ras_inject_if *info = (struct ras_inject_if *)inject_if; - int ret; - struct ta_ras_trigger_error_input block_info = { 0 }; - - if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX)) - return -EINVAL; - - block_info.block_id = amdgpu_ras_block_to_ta(info->head.block); - block_info.sub_block_index = info->head.sub_block_index; - block_info.inject_error_type = amdgpu_ras_error_to_ta(info->head.type); - block_info.address = info->address; - block_info.value = info->value; - - mutex_lock(&adev->grbm_idx_mutex); - ret = psp_ras_trigger_error(&adev->psp, &block_info); - mutex_unlock(&adev->grbm_idx_mutex); - - return ret; -} - static void gfx_v9_4_2_query_ea_err_status(struct amdgpu_device *adev) { uint32_t i, j; @@ -1935,7 +1916,7 @@ static bool gfx_v9_4_2_query_uctl2_poison_status(struct amdgpu_device *adev) u32 status = 0; struct amdgpu_vmhub *hub; - hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; status = RREG32(hub->vm_l2_pro_fault_status); /* reset page fault status */ WREG32_P(hub->vm_l2_pro_fault_cntl, 1, ~1); @@ -1944,7 +1925,6 @@ static bool gfx_v9_4_2_query_uctl2_poison_status(struct amdgpu_device *adev) } struct amdgpu_ras_block_hw_ops gfx_v9_4_2_ras_ops = { - .ras_error_inject = &gfx_v9_4_2_ras_error_inject, .query_ras_error_count = &gfx_v9_4_2_query_ras_error_count, .reset_ras_error_count = &gfx_v9_4_2_reset_ras_error_count, .query_ras_error_status = &gfx_v9_4_2_query_ras_error_status, diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c index 5f8500577c02..c1ee54d4c3d3 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c @@ -25,35 +25,504 @@ #include "amdgpu.h" #include "amdgpu_gfx.h" #include "soc15.h" +#include "soc15d.h" #include "soc15_common.h" #include "vega10_enum.h" +#include "v9_structs.h" + +#include "ivsrcid/gfx/irqsrcs_gfx_9_0.h" + #include "gc/gc_9_4_3_offset.h" #include "gc/gc_9_4_3_sh_mask.h" #include "gfx_v9_4_3.h" +#include "amdgpu_xcp.h" + +MODULE_FIRMWARE("amdgpu/gc_9_4_3_mec.bin"); +MODULE_FIRMWARE("amdgpu/gc_9_4_3_rlc.bin"); +#define GFX9_MEC_HPD_SIZE 4096 #define RLCG_UCODE_LOADING_START_ADDRESS 0x00002000L +#define GOLDEN_GB_ADDR_CONFIG 0x2a114042 + +struct amdgpu_gfx_ras gfx_v9_4_3_ras; + +static void gfx_v9_4_3_set_ring_funcs(struct amdgpu_device *adev); +static void gfx_v9_4_3_set_irq_funcs(struct amdgpu_device *adev); +static void gfx_v9_4_3_set_gds_init(struct amdgpu_device *adev); +static void gfx_v9_4_3_set_rlc_funcs(struct amdgpu_device *adev); +static int gfx_v9_4_3_get_cu_info(struct amdgpu_device *adev, + struct amdgpu_cu_info *cu_info); + +static void gfx_v9_4_3_kiq_set_resources(struct amdgpu_ring *kiq_ring, + uint64_t queue_mask) +{ + amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 6)); + amdgpu_ring_write(kiq_ring, + PACKET3_SET_RESOURCES_VMID_MASK(0) | + /* vmid_mask:0* queue_type:0 (KIQ) */ + PACKET3_SET_RESOURCES_QUEUE_TYPE(0)); + amdgpu_ring_write(kiq_ring, + lower_32_bits(queue_mask)); /* queue mask lo */ + amdgpu_ring_write(kiq_ring, + upper_32_bits(queue_mask)); /* queue mask hi */ + amdgpu_ring_write(kiq_ring, 0); /* gws mask lo */ + amdgpu_ring_write(kiq_ring, 0); /* gws mask hi */ + amdgpu_ring_write(kiq_ring, 0); /* oac mask */ + amdgpu_ring_write(kiq_ring, 0); /* gds heap base:0, gds heap size:0 */ +} + +static void gfx_v9_4_3_kiq_map_queues(struct amdgpu_ring *kiq_ring, + struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = kiq_ring->adev; + uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj); + uint64_t wptr_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4); + uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0; + + amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_MAP_QUEUES, 5)); + /* Q_sel:0, vmid:0, vidmem: 1, engine:0, num_Q:1*/ + amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */ + PACKET3_MAP_QUEUES_QUEUE_SEL(0) | /* Queue_Sel */ + PACKET3_MAP_QUEUES_VMID(0) | /* VMID */ + PACKET3_MAP_QUEUES_QUEUE(ring->queue) | + PACKET3_MAP_QUEUES_PIPE(ring->pipe) | + PACKET3_MAP_QUEUES_ME((ring->me == 1 ? 0 : 1)) | + /*queue_type: normal compute queue */ + PACKET3_MAP_QUEUES_QUEUE_TYPE(0) | + /* alloc format: all_on_one_pipe */ + PACKET3_MAP_QUEUES_ALLOC_FORMAT(0) | + PACKET3_MAP_QUEUES_ENGINE_SEL(eng_sel) | + /* num_queues: must be 1 */ + PACKET3_MAP_QUEUES_NUM_QUEUES(1)); + amdgpu_ring_write(kiq_ring, + PACKET3_MAP_QUEUES_DOORBELL_OFFSET(ring->doorbell_index)); + amdgpu_ring_write(kiq_ring, lower_32_bits(mqd_addr)); + amdgpu_ring_write(kiq_ring, upper_32_bits(mqd_addr)); + amdgpu_ring_write(kiq_ring, lower_32_bits(wptr_addr)); + amdgpu_ring_write(kiq_ring, upper_32_bits(wptr_addr)); +} + +static void gfx_v9_4_3_kiq_unmap_queues(struct amdgpu_ring *kiq_ring, + struct amdgpu_ring *ring, + enum amdgpu_unmap_queues_action action, + u64 gpu_addr, u64 seq) +{ + uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0; + + amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_UNMAP_QUEUES, 4)); + amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */ + PACKET3_UNMAP_QUEUES_ACTION(action) | + PACKET3_UNMAP_QUEUES_QUEUE_SEL(0) | + PACKET3_UNMAP_QUEUES_ENGINE_SEL(eng_sel) | + PACKET3_UNMAP_QUEUES_NUM_QUEUES(1)); + amdgpu_ring_write(kiq_ring, + PACKET3_UNMAP_QUEUES_DOORBELL_OFFSET0(ring->doorbell_index)); + + if (action == PREEMPT_QUEUES_NO_UNMAP) { + amdgpu_ring_write(kiq_ring, lower_32_bits(gpu_addr)); + amdgpu_ring_write(kiq_ring, upper_32_bits(gpu_addr)); + amdgpu_ring_write(kiq_ring, seq); + } else { + amdgpu_ring_write(kiq_ring, 0); + amdgpu_ring_write(kiq_ring, 0); + amdgpu_ring_write(kiq_ring, 0); + } +} + +static void gfx_v9_4_3_kiq_query_status(struct amdgpu_ring *kiq_ring, + struct amdgpu_ring *ring, + u64 addr, + u64 seq) +{ + uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0; + + amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_QUERY_STATUS, 5)); + amdgpu_ring_write(kiq_ring, + PACKET3_QUERY_STATUS_CONTEXT_ID(0) | + PACKET3_QUERY_STATUS_INTERRUPT_SEL(0) | + PACKET3_QUERY_STATUS_COMMAND(2)); + /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */ + amdgpu_ring_write(kiq_ring, + PACKET3_QUERY_STATUS_DOORBELL_OFFSET(ring->doorbell_index) | + PACKET3_QUERY_STATUS_ENG_SEL(eng_sel)); + amdgpu_ring_write(kiq_ring, lower_32_bits(addr)); + amdgpu_ring_write(kiq_ring, upper_32_bits(addr)); + amdgpu_ring_write(kiq_ring, lower_32_bits(seq)); + amdgpu_ring_write(kiq_ring, upper_32_bits(seq)); +} + +static void gfx_v9_4_3_kiq_invalidate_tlbs(struct amdgpu_ring *kiq_ring, + uint16_t pasid, uint32_t flush_type, + bool all_hub) +{ + amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_INVALIDATE_TLBS, 0)); + amdgpu_ring_write(kiq_ring, + PACKET3_INVALIDATE_TLBS_DST_SEL(1) | + PACKET3_INVALIDATE_TLBS_ALL_HUB(all_hub) | + PACKET3_INVALIDATE_TLBS_PASID(pasid) | + PACKET3_INVALIDATE_TLBS_FLUSH_TYPE(flush_type)); +} + +static const struct kiq_pm4_funcs gfx_v9_4_3_kiq_pm4_funcs = { + .kiq_set_resources = gfx_v9_4_3_kiq_set_resources, + .kiq_map_queues = gfx_v9_4_3_kiq_map_queues, + .kiq_unmap_queues = gfx_v9_4_3_kiq_unmap_queues, + .kiq_query_status = gfx_v9_4_3_kiq_query_status, + .kiq_invalidate_tlbs = gfx_v9_4_3_kiq_invalidate_tlbs, + .set_resources_size = 8, + .map_queues_size = 7, + .unmap_queues_size = 6, + .query_status_size = 7, + .invalidate_tlbs_size = 2, +}; + +static void gfx_v9_4_3_set_kiq_pm4_funcs(struct amdgpu_device *adev) +{ + int i, num_xcc; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < num_xcc; i++) + adev->gfx.kiq[i].pmf = &gfx_v9_4_3_kiq_pm4_funcs; +} + +static void gfx_v9_4_3_init_golden_registers(struct amdgpu_device *adev) +{ + int i, num_xcc, dev_inst; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < num_xcc; i++) { + dev_inst = GET_INST(GC, i); + if (dev_inst >= 2) + WREG32_SOC15(GC, dev_inst, regGRBM_MCM_ADDR, 0x4); + + /* Golden settings applied by driver for ASIC with rev_id 0 */ + if (adev->rev_id == 0) { + WREG32_SOC15(GC, dev_inst, regGB_ADDR_CONFIG, + GOLDEN_GB_ADDR_CONFIG); + + WREG32_FIELD15_PREREG(GC, dev_inst, TCP_UTCL1_CNTL1, + REDUCE_FIFO_DEPTH_BY_2, 2); + } + } +} + +static void gfx_v9_4_3_write_data_to_reg(struct amdgpu_ring *ring, int eng_sel, + bool wc, uint32_t reg, uint32_t val) +{ + amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); + amdgpu_ring_write(ring, WRITE_DATA_ENGINE_SEL(eng_sel) | + WRITE_DATA_DST_SEL(0) | + (wc ? WR_CONFIRM : 0)); + amdgpu_ring_write(ring, reg); + amdgpu_ring_write(ring, 0); + amdgpu_ring_write(ring, val); +} + +static void gfx_v9_4_3_wait_reg_mem(struct amdgpu_ring *ring, int eng_sel, + int mem_space, int opt, uint32_t addr0, + uint32_t addr1, uint32_t ref, uint32_t mask, + uint32_t inv) +{ + amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5)); + amdgpu_ring_write(ring, + /* memory (1) or register (0) */ + (WAIT_REG_MEM_MEM_SPACE(mem_space) | + WAIT_REG_MEM_OPERATION(opt) | /* wait */ + WAIT_REG_MEM_FUNCTION(3) | /* equal */ + WAIT_REG_MEM_ENGINE(eng_sel))); + + if (mem_space) + BUG_ON(addr0 & 0x3); /* Dword align */ + amdgpu_ring_write(ring, addr0); + amdgpu_ring_write(ring, addr1); + amdgpu_ring_write(ring, ref); + amdgpu_ring_write(ring, mask); + amdgpu_ring_write(ring, inv); /* poll interval */ +} + +static int gfx_v9_4_3_ring_test_ring(struct amdgpu_ring *ring) +{ + uint32_t scratch_reg0_offset, xcc_offset; + struct amdgpu_device *adev = ring->adev; + uint32_t tmp = 0; + unsigned i; + int r; + + /* Use register offset which is local to XCC in the packet */ + xcc_offset = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG0); + scratch_reg0_offset = SOC15_REG_OFFSET(GC, GET_INST(GC, ring->xcc_id), regSCRATCH_REG0); + WREG32(scratch_reg0_offset, 0xCAFEDEAD); + + r = amdgpu_ring_alloc(ring, 3); + if (r) + return r; + + amdgpu_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1)); + amdgpu_ring_write(ring, xcc_offset - PACKET3_SET_UCONFIG_REG_START); + amdgpu_ring_write(ring, 0xDEADBEEF); + amdgpu_ring_commit(ring); + + for (i = 0; i < adev->usec_timeout; i++) { + tmp = RREG32(scratch_reg0_offset); + if (tmp == 0xDEADBEEF) + break; + udelay(1); + } + + if (i >= adev->usec_timeout) + r = -ETIMEDOUT; + return r; +} + +static int gfx_v9_4_3_ring_test_ib(struct amdgpu_ring *ring, long timeout) +{ + struct amdgpu_device *adev = ring->adev; + struct amdgpu_ib ib; + struct dma_fence *f = NULL; + + unsigned index; + uint64_t gpu_addr; + uint32_t tmp; + long r; + + r = amdgpu_device_wb_get(adev, &index); + if (r) + return r; + + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD); + memset(&ib, 0, sizeof(ib)); + r = amdgpu_ib_get(adev, NULL, 16, + AMDGPU_IB_POOL_DIRECT, &ib); + if (r) + goto err1; + + ib.ptr[0] = PACKET3(PACKET3_WRITE_DATA, 3); + ib.ptr[1] = WRITE_DATA_DST_SEL(5) | WR_CONFIRM; + ib.ptr[2] = lower_32_bits(gpu_addr); + ib.ptr[3] = upper_32_bits(gpu_addr); + ib.ptr[4] = 0xDEADBEEF; + ib.length_dw = 5; + + r = amdgpu_ib_schedule(ring, 1, &ib, NULL, &f); + if (r) + goto err2; + + r = dma_fence_wait_timeout(f, false, timeout); + if (r == 0) { + r = -ETIMEDOUT; + goto err2; + } else if (r < 0) { + goto err2; + } + + tmp = adev->wb.wb[index]; + if (tmp == 0xDEADBEEF) + r = 0; + else + r = -EINVAL; + +err2: + amdgpu_ib_free(adev, &ib, NULL); + dma_fence_put(f); +err1: + amdgpu_device_wb_free(adev, index); + return r; +} + + +/* This value might differs per partition */ static uint64_t gfx_v9_4_3_get_gpu_clock_counter(struct amdgpu_device *adev) { uint64_t clock; amdgpu_gfx_off_ctrl(adev, false); mutex_lock(&adev->gfx.gpu_clock_mutex); - WREG32_SOC15(GC, 0, regRLC_CAPTURE_GPU_CLOCK_COUNT, 1); - clock = (uint64_t)RREG32_SOC15(GC, 0, regRLC_GPU_CLOCK_COUNT_LSB) | - ((uint64_t)RREG32_SOC15(GC, 0, regRLC_GPU_CLOCK_COUNT_MSB) << 32ULL); + WREG32_SOC15(GC, GET_INST(GC, 0), regRLC_CAPTURE_GPU_CLOCK_COUNT, 1); + clock = (uint64_t)RREG32_SOC15(GC, GET_INST(GC, 0), regRLC_GPU_CLOCK_COUNT_LSB) | + ((uint64_t)RREG32_SOC15(GC, GET_INST(GC, 0), regRLC_GPU_CLOCK_COUNT_MSB) << 32ULL); mutex_unlock(&adev->gfx.gpu_clock_mutex); amdgpu_gfx_off_ctrl(adev, true); return clock; } -static void gfx_v9_4_3_select_se_sh(struct amdgpu_device *adev, - u32 se_num, - u32 sh_num, - u32 instance) +static void gfx_v9_4_3_free_microcode(struct amdgpu_device *adev) +{ + amdgpu_ucode_release(&adev->gfx.pfp_fw); + amdgpu_ucode_release(&adev->gfx.me_fw); + amdgpu_ucode_release(&adev->gfx.ce_fw); + amdgpu_ucode_release(&adev->gfx.rlc_fw); + amdgpu_ucode_release(&adev->gfx.mec_fw); + amdgpu_ucode_release(&adev->gfx.mec2_fw); + + kfree(adev->gfx.rlc.register_list_format); +} + +static int gfx_v9_4_3_init_rlc_microcode(struct amdgpu_device *adev, + const char *chip_name) +{ + char fw_name[30]; + int err; + const struct rlc_firmware_header_v2_0 *rlc_hdr; + uint16_t version_major; + uint16_t version_minor; + + snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_rlc.bin", chip_name); + + err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw, fw_name); + if (err) + goto out; + rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data; + + version_major = le16_to_cpu(rlc_hdr->header.header_version_major); + version_minor = le16_to_cpu(rlc_hdr->header.header_version_minor); + err = amdgpu_gfx_rlc_init_microcode(adev, version_major, version_minor); +out: + if (err) + amdgpu_ucode_release(&adev->gfx.rlc_fw); + + return err; +} + +static bool gfx_v9_4_3_should_disable_gfxoff(struct pci_dev *pdev) +{ + return true; +} + +static void gfx_v9_4_3_check_if_need_gfxoff(struct amdgpu_device *adev) +{ + if (gfx_v9_4_3_should_disable_gfxoff(adev->pdev)) + adev->pm.pp_feature &= ~PP_GFXOFF_MASK; +} + +static int gfx_v9_4_3_init_cp_compute_microcode(struct amdgpu_device *adev, + const char *chip_name) +{ + char fw_name[30]; + int err; + + snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name); + + err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw, fw_name); + if (err) + goto out; + amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC1); + amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC1_JT); + + adev->gfx.mec2_fw_version = adev->gfx.mec_fw_version; + adev->gfx.mec2_feature_version = adev->gfx.mec_feature_version; + + gfx_v9_4_3_check_if_need_gfxoff(adev); + +out: + if (err) + amdgpu_ucode_release(&adev->gfx.mec_fw); + return err; +} + +static int gfx_v9_4_3_init_microcode(struct amdgpu_device *adev) +{ + const char *chip_name; + int r; + + chip_name = "gc_9_4_3"; + + r = gfx_v9_4_3_init_rlc_microcode(adev, chip_name); + if (r) + return r; + + r = gfx_v9_4_3_init_cp_compute_microcode(adev, chip_name); + if (r) + return r; + + return r; +} + +static void gfx_v9_4_3_mec_fini(struct amdgpu_device *adev) +{ + amdgpu_bo_free_kernel(&adev->gfx.mec.hpd_eop_obj, NULL, NULL); + amdgpu_bo_free_kernel(&adev->gfx.mec.mec_fw_obj, NULL, NULL); +} + +static int gfx_v9_4_3_mec_init(struct amdgpu_device *adev) +{ + int r, i, num_xcc; + u32 *hpd; + const __le32 *fw_data; + unsigned fw_size; + u32 *fw; + size_t mec_hpd_size; + + const struct gfx_firmware_header_v1_0 *mec_hdr; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < num_xcc; i++) + bitmap_zero(adev->gfx.mec_bitmap[i].queue_bitmap, + AMDGPU_MAX_COMPUTE_QUEUES); + + /* take ownership of the relevant compute queues */ + amdgpu_gfx_compute_queue_acquire(adev); + mec_hpd_size = + adev->gfx.num_compute_rings * num_xcc * GFX9_MEC_HPD_SIZE; + if (mec_hpd_size) { + r = amdgpu_bo_create_reserved(adev, mec_hpd_size, PAGE_SIZE, + AMDGPU_GEM_DOMAIN_VRAM | + AMDGPU_GEM_DOMAIN_GTT, + &adev->gfx.mec.hpd_eop_obj, + &adev->gfx.mec.hpd_eop_gpu_addr, + (void **)&hpd); + if (r) { + dev_warn(adev->dev, "(%d) create HDP EOP bo failed\n", r); + gfx_v9_4_3_mec_fini(adev); + return r; + } + + if (amdgpu_emu_mode == 1) { + for (i = 0; i < mec_hpd_size / 4; i++) { + memset((void *)(hpd + i), 0, 4); + if (i % 50 == 0) + msleep(1); + } + } else { + memset(hpd, 0, mec_hpd_size); + } + + amdgpu_bo_kunmap(adev->gfx.mec.hpd_eop_obj); + amdgpu_bo_unreserve(adev->gfx.mec.hpd_eop_obj); + } + + mec_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data; + + fw_data = (const __le32 *) + (adev->gfx.mec_fw->data + + le32_to_cpu(mec_hdr->header.ucode_array_offset_bytes)); + fw_size = le32_to_cpu(mec_hdr->header.ucode_size_bytes); + + r = amdgpu_bo_create_reserved(adev, mec_hdr->header.ucode_size_bytes, + PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT, + &adev->gfx.mec.mec_fw_obj, + &adev->gfx.mec.mec_fw_gpu_addr, + (void **)&fw); + if (r) { + dev_warn(adev->dev, "(%d) create mec firmware bo failed\n", r); + gfx_v9_4_3_mec_fini(adev); + return r; + } + + memcpy(fw, fw_data, fw_size); + + amdgpu_bo_kunmap(adev->gfx.mec.mec_fw_obj); + amdgpu_bo_unreserve(adev->gfx.mec.mec_fw_obj); + + return 0; +} + +static void gfx_v9_4_3_xcc_select_se_sh(struct amdgpu_device *adev, u32 se_num, + u32 sh_num, u32 instance, int xcc_id) { u32 data; @@ -76,24 +545,24 @@ static void gfx_v9_4_3_select_se_sh(struct amdgpu_device *adev, else data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SH_INDEX, sh_num); - WREG32_SOC15_RLC_SHADOW_EX(reg, GC, 0, regGRBM_GFX_INDEX, data); + WREG32_SOC15_RLC_SHADOW_EX(reg, GC, GET_INST(GC, xcc_id), regGRBM_GFX_INDEX, data); } -static uint32_t wave_read_ind(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t address) +static uint32_t wave_read_ind(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t address) { - WREG32_SOC15_RLC(GC, 0, regSQ_IND_INDEX, + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regSQ_IND_INDEX, (wave << SQ_IND_INDEX__WAVE_ID__SHIFT) | (simd << SQ_IND_INDEX__SIMD_ID__SHIFT) | (address << SQ_IND_INDEX__INDEX__SHIFT) | (SQ_IND_INDEX__FORCE_READ_MASK)); - return RREG32_SOC15(GC, 0, regSQ_IND_DATA); + return RREG32_SOC15(GC, GET_INST(GC, xcc_id), regSQ_IND_DATA); } -static void wave_read_regs(struct amdgpu_device *adev, uint32_t simd, +static void wave_read_regs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t thread, uint32_t regno, uint32_t num, uint32_t *out) { - WREG32_SOC15_RLC(GC, 0, regSQ_IND_INDEX, + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regSQ_IND_INDEX, (wave << SQ_IND_INDEX__WAVE_ID__SHIFT) | (simd << SQ_IND_INDEX__SIMD_ID__SHIFT) | (regno << SQ_IND_INDEX__INDEX__SHIFT) | @@ -101,53 +570,481 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t simd, (SQ_IND_INDEX__FORCE_READ_MASK) | (SQ_IND_INDEX__AUTO_INCR_MASK)); while (num--) - *(out++) = RREG32_SOC15(GC, 0, regSQ_IND_DATA); + *(out++) = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regSQ_IND_DATA); } static void gfx_v9_4_3_read_wave_data(struct amdgpu_device *adev, - uint32_t simd, uint32_t wave, + uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields) { /* type 1 wave data */ dst[(*no_fields)++] = 1; - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_STATUS); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_PC_LO); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_PC_HI); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_EXEC_LO); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_EXEC_HI); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_HW_ID); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_INST_DW0); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_INST_DW1); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_GPR_ALLOC); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_LDS_ALLOC); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_TRAPSTS); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_IB_STS); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_IB_DBG0); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_M0); - dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_MODE); -} - -static void gfx_v9_4_3_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd, + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_STATUS); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_PC_LO); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_PC_HI); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_EXEC_LO); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_EXEC_HI); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_HW_ID); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_INST_DW0); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_INST_DW1); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_GPR_ALLOC); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_LDS_ALLOC); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_TRAPSTS); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_IB_STS); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_IB_DBG0); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_M0); + dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_MODE); +} + +static void gfx_v9_4_3_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t start, uint32_t size, uint32_t *dst) { - wave_read_regs(adev, simd, wave, 0, + wave_read_regs(adev, xcc_id, simd, wave, 0, start + SQIND_WAVE_SGPRS_OFFSET, size, dst); } -static void gfx_v9_4_3_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd, +static void gfx_v9_4_3_read_wave_vgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t thread, uint32_t start, uint32_t size, uint32_t *dst) { - wave_read_regs(adev, simd, wave, thread, + wave_read_regs(adev, xcc_id, simd, wave, thread, start + SQIND_WAVE_VGPRS_OFFSET, size, dst); } static void gfx_v9_4_3_select_me_pipe_q(struct amdgpu_device *adev, - u32 me, u32 pipe, u32 q, u32 vm) + u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id) +{ + soc15_grbm_select(adev, me, pipe, q, vm, GET_INST(GC, xcc_id)); +} + + +static int gfx_v9_4_3_switch_compute_partition(struct amdgpu_device *adev, + int num_xccs_per_xcp) +{ + int ret; + + ret = psp_spatial_partition(&adev->psp, NUM_XCC(adev->gfx.xcc_mask) / + num_xccs_per_xcp); + if (ret) + return ret; + + adev->gfx.num_xcc_per_xcp = num_xccs_per_xcp; + + return ret; +} + +static int gfx_v9_4_3_ih_to_xcc_inst(struct amdgpu_device *adev, int ih_node) +{ + int xcc; + + xcc = hweight8(adev->gfx.xcc_mask & GENMASK(ih_node / 2, 0)); + if (!xcc) { + dev_err(adev->dev, "Couldn't find xcc mapping from IH node"); + return -EINVAL; + } + + return xcc - 1; +} + +static const struct amdgpu_gfx_funcs gfx_v9_4_3_gfx_funcs = { + .get_gpu_clock_counter = &gfx_v9_4_3_get_gpu_clock_counter, + .select_se_sh = &gfx_v9_4_3_xcc_select_se_sh, + .read_wave_data = &gfx_v9_4_3_read_wave_data, + .read_wave_sgprs = &gfx_v9_4_3_read_wave_sgprs, + .read_wave_vgprs = &gfx_v9_4_3_read_wave_vgprs, + .select_me_pipe_q = &gfx_v9_4_3_select_me_pipe_q, + .switch_partition_mode = &gfx_v9_4_3_switch_compute_partition, + .ih_node_to_logical_xcc = &gfx_v9_4_3_ih_to_xcc_inst, +}; + +static int gfx_v9_4_3_gpu_early_init(struct amdgpu_device *adev) +{ + u32 gb_addr_config; + + adev->gfx.funcs = &gfx_v9_4_3_gfx_funcs; + adev->gfx.ras = &gfx_v9_4_3_ras; + + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 4, 3): + adev->gfx.config.max_hw_contexts = 8; + adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; + adev->gfx.config.sc_prim_fifo_size_backend = 0x100; + adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; + adev->gfx.config.sc_earlyz_tile_fifo_size = 0x4C0; + gb_addr_config = RREG32_SOC15(GC, GET_INST(GC, 0), regGB_ADDR_CONFIG); + break; + default: + BUG(); + break; + } + + adev->gfx.config.gb_addr_config = gb_addr_config; + + adev->gfx.config.gb_addr_config_fields.num_pipes = 1 << + REG_GET_FIELD( + adev->gfx.config.gb_addr_config, + GB_ADDR_CONFIG, + NUM_PIPES); + + adev->gfx.config.max_tile_pipes = + adev->gfx.config.gb_addr_config_fields.num_pipes; + + adev->gfx.config.gb_addr_config_fields.num_banks = 1 << + REG_GET_FIELD( + adev->gfx.config.gb_addr_config, + GB_ADDR_CONFIG, + NUM_BANKS); + adev->gfx.config.gb_addr_config_fields.max_compress_frags = 1 << + REG_GET_FIELD( + adev->gfx.config.gb_addr_config, + GB_ADDR_CONFIG, + MAX_COMPRESSED_FRAGS); + adev->gfx.config.gb_addr_config_fields.num_rb_per_se = 1 << + REG_GET_FIELD( + adev->gfx.config.gb_addr_config, + GB_ADDR_CONFIG, + NUM_RB_PER_SE); + adev->gfx.config.gb_addr_config_fields.num_se = 1 << + REG_GET_FIELD( + adev->gfx.config.gb_addr_config, + GB_ADDR_CONFIG, + NUM_SHADER_ENGINES); + adev->gfx.config.gb_addr_config_fields.pipe_interleave_size = 1 << (8 + + REG_GET_FIELD( + adev->gfx.config.gb_addr_config, + GB_ADDR_CONFIG, + PIPE_INTERLEAVE_SIZE)); + + return 0; +} + +static int gfx_v9_4_3_compute_ring_init(struct amdgpu_device *adev, int ring_id, + int xcc_id, int mec, int pipe, int queue) +{ + unsigned irq_type; + struct amdgpu_ring *ring = &adev->gfx.compute_ring[ring_id]; + unsigned int hw_prio; + uint32_t xcc_doorbell_start; + + ring = &adev->gfx.compute_ring[xcc_id * adev->gfx.num_compute_rings + + ring_id]; + + /* mec0 is me1 */ + ring->xcc_id = xcc_id; + ring->me = mec + 1; + ring->pipe = pipe; + ring->queue = queue; + + ring->ring_obj = NULL; + ring->use_doorbell = true; + xcc_doorbell_start = adev->doorbell_index.mec_ring0 + + xcc_id * adev->doorbell_index.xcc_doorbell_range; + ring->doorbell_index = (xcc_doorbell_start + ring_id) << 1; + ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr + + (ring_id + xcc_id * adev->gfx.num_compute_rings) * + GFX9_MEC_HPD_SIZE; + ring->vm_hub = AMDGPU_GFXHUB(xcc_id); + sprintf(ring->name, "comp_%d.%d.%d.%d", + ring->xcc_id, ring->me, ring->pipe, ring->queue); + + irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP + + ((ring->me - 1) * adev->gfx.mec.num_pipe_per_mec) + + ring->pipe; + hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring) ? + AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_GFX_PIPE_PRIO_NORMAL; + /* type-2 packets are deprecated on MEC, use type-3 instead */ + return amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type, + hw_prio, NULL); +} + +static int gfx_v9_4_3_sw_init(void *handle) +{ + int i, j, k, r, ring_id, xcc_id, num_xcc; + struct amdgpu_kiq *kiq; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + adev->gfx.mec.num_mec = 2; + adev->gfx.mec.num_pipe_per_mec = 4; + adev->gfx.mec.num_queue_per_pipe = 8; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + + /* EOP Event */ + r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP, GFX_9_0__SRCID__CP_EOP_INTERRUPT, &adev->gfx.eop_irq); + if (r) + return r; + + /* Privileged reg */ + r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP, GFX_9_0__SRCID__CP_PRIV_REG_FAULT, + &adev->gfx.priv_reg_irq); + if (r) + return r; + + /* Privileged inst */ + r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP, GFX_9_0__SRCID__CP_PRIV_INSTR_FAULT, + &adev->gfx.priv_inst_irq); + if (r) + return r; + + adev->gfx.gfx_current_status = AMDGPU_GFX_NORMAL_MODE; + + r = adev->gfx.rlc.funcs->init(adev); + if (r) { + DRM_ERROR("Failed to init rlc BOs!\n"); + return r; + } + + r = gfx_v9_4_3_mec_init(adev); + if (r) { + DRM_ERROR("Failed to init MEC BOs!\n"); + return r; + } + + /* set up the compute queues - allocate horizontally across pipes */ + for (xcc_id = 0; xcc_id < num_xcc; xcc_id++) { + ring_id = 0; + for (i = 0; i < adev->gfx.mec.num_mec; ++i) { + for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) { + for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; + k++) { + if (!amdgpu_gfx_is_mec_queue_enabled( + adev, xcc_id, i, k, j)) + continue; + + r = gfx_v9_4_3_compute_ring_init(adev, + ring_id, + xcc_id, + i, k, j); + if (r) + return r; + + ring_id++; + } + } + } + + r = amdgpu_gfx_kiq_init(adev, GFX9_MEC_HPD_SIZE, xcc_id); + if (r) { + DRM_ERROR("Failed to init KIQ BOs!\n"); + return r; + } + + kiq = &adev->gfx.kiq[xcc_id]; + r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq, xcc_id); + if (r) + return r; + + /* create MQD for all compute queues as wel as KIQ for SRIOV case */ + r = amdgpu_gfx_mqd_sw_init(adev, + sizeof(struct v9_mqd_allocation), xcc_id); + if (r) + return r; + } + + r = gfx_v9_4_3_gpu_early_init(adev); + if (r) + return r; + + r = amdgpu_gfx_sysfs_init(adev); + if (r) + return r; + + return amdgpu_gfx_ras_sw_init(adev); +} + +static int gfx_v9_4_3_sw_fini(void *handle) { - soc15_grbm_select(adev, me, pipe, q, vm); + int i, num_xcc; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < adev->gfx.num_compute_rings * num_xcc; i++) + amdgpu_ring_fini(&adev->gfx.compute_ring[i]); + + for (i = 0; i < num_xcc; i++) { + amdgpu_gfx_mqd_sw_fini(adev, i); + amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq[i].ring); + amdgpu_gfx_kiq_fini(adev, i); + } + + gfx_v9_4_3_mec_fini(adev); + amdgpu_bo_unref(&adev->gfx.rlc.clear_state_obj); + gfx_v9_4_3_free_microcode(adev); + amdgpu_gfx_sysfs_fini(adev); + + return 0; +} + +#define DEFAULT_SH_MEM_BASES (0x6000) +static void gfx_v9_4_3_xcc_init_compute_vmid(struct amdgpu_device *adev, + int xcc_id) +{ + int i; + uint32_t sh_mem_config; + uint32_t sh_mem_bases; + + /* + * Configure apertures: + * LDS: 0x60000000'00000000 - 0x60000001'00000000 (4GB) + * Scratch: 0x60000001'00000000 - 0x60000002'00000000 (4GB) + * GPUVM: 0x60010000'00000000 - 0x60020000'00000000 (1TB) + */ + sh_mem_bases = DEFAULT_SH_MEM_BASES | (DEFAULT_SH_MEM_BASES << 16); + + sh_mem_config = SH_MEM_ADDRESS_MODE_64 | + SH_MEM_ALIGNMENT_MODE_UNALIGNED << + SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT; + + mutex_lock(&adev->srbm_mutex); + for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) { + soc15_grbm_select(adev, 0, 0, 0, i, GET_INST(GC, xcc_id)); + /* CP and shaders */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regSH_MEM_CONFIG, sh_mem_config); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regSH_MEM_BASES, sh_mem_bases); + } + soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id)); + mutex_unlock(&adev->srbm_mutex); + + /* Initialize all compute VMIDs to have no GDS, GWS, or OA + acccess. These should be enabled by FW for target VMIDs. */ + for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) { + WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_VMID0_BASE, 2 * i, 0); + WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_VMID0_SIZE, 2 * i, 0); + WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_GWS_VMID0, i, 0); + WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_OA_VMID0, i, 0); + } +} + +static void gfx_v9_4_3_xcc_init_gds_vmid(struct amdgpu_device *adev, int xcc_id) +{ + int vmid; + + /* + * Initialize all compute and user-gfx VMIDs to have no GDS, GWS, or OA + * access. Compute VMIDs should be enabled by FW for target VMIDs, + * the driver can enable them for graphics. VMID0 should maintain + * access so that HWS firmware can save/restore entries. + */ + for (vmid = 1; vmid < AMDGPU_NUM_VMID; vmid++) { + WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_VMID0_BASE, 2 * vmid, 0); + WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_VMID0_SIZE, 2 * vmid, 0); + WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_GWS_VMID0, vmid, 0); + WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_OA_VMID0, vmid, 0); + } +} + +static void gfx_v9_4_3_xcc_constants_init(struct amdgpu_device *adev, + int xcc_id) +{ + u32 tmp; + int i; + + /* XXX SH_MEM regs */ + /* where to put LDS, scratch, GPUVM in FSA64 space */ + mutex_lock(&adev->srbm_mutex); + for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB(0)].num_ids; i++) { + soc15_grbm_select(adev, 0, 0, 0, i, GET_INST(GC, xcc_id)); + /* CP and shaders */ + if (i == 0) { + tmp = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE, + SH_MEM_ALIGNMENT_MODE_UNALIGNED); + tmp = REG_SET_FIELD(tmp, SH_MEM_CONFIG, RETRY_DISABLE, + !!adev->gmc.noretry); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), + regSH_MEM_CONFIG, tmp); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), + regSH_MEM_BASES, 0); + } else { + tmp = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE, + SH_MEM_ALIGNMENT_MODE_UNALIGNED); + tmp = REG_SET_FIELD(tmp, SH_MEM_CONFIG, RETRY_DISABLE, + !!adev->gmc.noretry); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), + regSH_MEM_CONFIG, tmp); + tmp = REG_SET_FIELD(0, SH_MEM_BASES, PRIVATE_BASE, + (adev->gmc.private_aperture_start >> + 48)); + tmp = REG_SET_FIELD(tmp, SH_MEM_BASES, SHARED_BASE, + (adev->gmc.shared_aperture_start >> + 48)); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), + regSH_MEM_BASES, tmp); + } + } + soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, 0)); + + mutex_unlock(&adev->srbm_mutex); + + gfx_v9_4_3_xcc_init_compute_vmid(adev, xcc_id); + gfx_v9_4_3_xcc_init_gds_vmid(adev, xcc_id); +} + +static void gfx_v9_4_3_constants_init(struct amdgpu_device *adev) +{ + int i, num_xcc; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + + gfx_v9_4_3_get_cu_info(adev, &adev->gfx.cu_info); + adev->gfx.config.db_debug2 = + RREG32_SOC15(GC, GET_INST(GC, 0), regDB_DEBUG2); + + for (i = 0; i < num_xcc; i++) + gfx_v9_4_3_xcc_constants_init(adev, i); +} + +static void +gfx_v9_4_3_xcc_enable_save_restore_machine(struct amdgpu_device *adev, + int xcc_id) +{ + WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), RLC_SRM_CNTL, SRM_ENABLE, 1); +} + +static void gfx_v9_4_3_xcc_init_pg(struct amdgpu_device *adev, int xcc_id) +{ + /* + * Rlc save restore list is workable since v2_1. + * And it's needed by gfxoff feature. + */ + if (adev->gfx.rlc.is_rlc_v2_1) + gfx_v9_4_3_xcc_enable_save_restore_machine(adev, xcc_id); +} + +static void gfx_v9_4_3_xcc_disable_gpa_mode(struct amdgpu_device *adev, int xcc_id) +{ + uint32_t data; + + data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCPC_PSP_DEBUG); + data |= CPC_PSP_DEBUG__UTCL2IUGPAOVERRIDE_MASK; + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCPC_PSP_DEBUG, data); +} + +static void gfx_v9_4_3_xcc_program_xcc_id(struct amdgpu_device *adev, + int xcc_id) +{ + uint32_t tmp = 0; + int num_xcc; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + switch (num_xcc) { + /* directly config VIRTUAL_XCC_ID to 0 for 1-XCC */ + case 1: + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HYP_XCP_CTL, 0x8); + break; + case 2: + case 4: + case 6: + case 8: + tmp = (xcc_id % adev->gfx.num_xcc_per_xcp) << REG_FIELD_SHIFT(CP_HYP_XCP_CTL, VIRTUAL_XCC_ID); + tmp = tmp | (adev->gfx.num_xcc_per_xcp << REG_FIELD_SHIFT(CP_HYP_XCP_CTL, NUM_XCC_IN_XCP)); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HYP_XCP_CTL, tmp); + + break; + default: + break; + } } static bool gfx_v9_4_3_is_rlc_enabled(struct amdgpu_device *adev) @@ -155,36 +1052,37 @@ static bool gfx_v9_4_3_is_rlc_enabled(struct amdgpu_device *adev) uint32_t rlc_setting; /* if RLC is not enabled, do nothing */ - rlc_setting = RREG32_SOC15(GC, 0, regRLC_CNTL); + rlc_setting = RREG32_SOC15(GC, GET_INST(GC, 0), regRLC_CNTL); if (!(rlc_setting & RLC_CNTL__RLC_ENABLE_F32_MASK)) return false; return true; } -static void gfx_v9_4_3_set_safe_mode(struct amdgpu_device *adev) +static void gfx_v9_4_3_xcc_set_safe_mode(struct amdgpu_device *adev, int xcc_id) { uint32_t data; unsigned i; data = RLC_SAFE_MODE__CMD_MASK; data |= (1 << RLC_SAFE_MODE__MESSAGE__SHIFT); - WREG32_SOC15(GC, 0, regRLC_SAFE_MODE, data); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_SAFE_MODE, data); /* wait for RLC_SAFE_MODE */ for (i = 0; i < adev->usec_timeout; i++) { - if (!REG_GET_FIELD(RREG32_SOC15(GC, 0, regRLC_SAFE_MODE), RLC_SAFE_MODE, CMD)) + if (!REG_GET_FIELD(RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_SAFE_MODE), RLC_SAFE_MODE, CMD)) break; udelay(1); } } -static void gfx_v9_4_3_unset_safe_mode(struct amdgpu_device *adev) +static void gfx_v9_4_3_xcc_unset_safe_mode(struct amdgpu_device *adev, + int xcc_id) { uint32_t data; data = RLC_SAFE_MODE__CMD_MASK; - WREG32_SOC15(GC, 0, regRLC_SAFE_MODE, data); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_SAFE_MODE, data); } static int gfx_v9_4_3_rlc_init(struct amdgpu_device *adev) @@ -196,7 +1094,8 @@ static int gfx_v9_4_3_rlc_init(struct amdgpu_device *adev) return 0; } -static void gfx_v9_4_3_wait_for_rlc_serdes(struct amdgpu_device *adev) +static void gfx_v9_4_3_xcc_wait_for_rlc_serdes(struct amdgpu_device *adev, + int xcc_id) { u32 i, j, k; u32 mask; @@ -204,15 +1103,17 @@ static void gfx_v9_4_3_wait_for_rlc_serdes(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v9_4_3_select_se_sh(adev, i, j, 0xffffffff); + gfx_v9_4_3_xcc_select_se_sh(adev, i, j, 0xffffffff, + xcc_id); for (k = 0; k < adev->usec_timeout; k++) { - if (RREG32_SOC15(GC, 0, regRLC_SERDES_CU_MASTER_BUSY) == 0) + if (RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_SERDES_CU_MASTER_BUSY) == 0) break; udelay(1); } if (k == adev->usec_timeout) { - gfx_v9_4_3_select_se_sh(adev, 0xffffffff, - 0xffffffff, 0xffffffff); + gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, + 0xffffffff, + 0xffffffff, xcc_id); mutex_unlock(&adev->grbm_idx_mutex); DRM_INFO("Timeout wait for RLC serdes %u,%u\n", i, j); @@ -220,7 +1121,8 @@ static void gfx_v9_4_3_wait_for_rlc_serdes(struct amdgpu_device *adev) } } } - gfx_v9_4_3_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, + xcc_id); mutex_unlock(&adev->grbm_idx_mutex); mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK | @@ -228,79 +1130,108 @@ static void gfx_v9_4_3_wait_for_rlc_serdes(struct amdgpu_device *adev) RLC_SERDES_NONCU_MASTER_BUSY__TC0_MASTER_BUSY_MASK | RLC_SERDES_NONCU_MASTER_BUSY__TC1_MASTER_BUSY_MASK; for (k = 0; k < adev->usec_timeout; k++) { - if ((RREG32_SOC15(GC, 0, regRLC_SERDES_NONCU_MASTER_BUSY) & mask) == 0) + if ((RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_SERDES_NONCU_MASTER_BUSY) & mask) == 0) break; udelay(1); } } -static void gfx_v9_4_3_enable_gui_idle_interrupt(struct amdgpu_device *adev, - bool enable) +static void gfx_v9_4_3_xcc_enable_gui_idle_interrupt(struct amdgpu_device *adev, + bool enable, int xcc_id) { u32 tmp; /* These interrupts should be enabled to drive DS clock */ - tmp = RREG32_SOC15(GC, 0, regCP_INT_CNTL_RING0); + tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_INT_CNTL_RING0); tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE, enable ? 1 : 0); tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_EMPTY_INT_ENABLE, enable ? 1 : 0); tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CMP_BUSY_INT_ENABLE, enable ? 1 : 0); - if (adev->gfx.num_gfx_rings) - tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, GFX_IDLE_INT_ENABLE, enable ? 1 : 0); - WREG32_SOC15(GC, 0, regCP_INT_CNTL_RING0, tmp); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_INT_CNTL_RING0, tmp); +} + +static void gfx_v9_4_3_xcc_rlc_stop(struct amdgpu_device *adev, int xcc_id) +{ + WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), RLC_CNTL, + RLC_ENABLE_F32, 0); + gfx_v9_4_3_xcc_enable_gui_idle_interrupt(adev, false, xcc_id); + gfx_v9_4_3_xcc_wait_for_rlc_serdes(adev, xcc_id); } static void gfx_v9_4_3_rlc_stop(struct amdgpu_device *adev) { - WREG32_FIELD15_PREREG(GC, 0, RLC_CNTL, RLC_ENABLE_F32, 0); - gfx_v9_4_3_enable_gui_idle_interrupt(adev, false); - gfx_v9_4_3_wait_for_rlc_serdes(adev); + int i, num_xcc; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < num_xcc; i++) + gfx_v9_4_3_xcc_rlc_stop(adev, i); } -static void gfx_v9_4_3_rlc_reset(struct amdgpu_device *adev) +static void gfx_v9_4_3_xcc_rlc_reset(struct amdgpu_device *adev, int xcc_id) { - WREG32_FIELD15_PREREG(GC, 0, GRBM_SOFT_RESET, SOFT_RESET_RLC, 1); + WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), GRBM_SOFT_RESET, + SOFT_RESET_RLC, 1); udelay(50); - WREG32_FIELD15_PREREG(GC, 0, GRBM_SOFT_RESET, SOFT_RESET_RLC, 0); + WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), GRBM_SOFT_RESET, + SOFT_RESET_RLC, 0); udelay(50); } -static void gfx_v9_4_3_rlc_start(struct amdgpu_device *adev) +static void gfx_v9_4_3_rlc_reset(struct amdgpu_device *adev) { -#ifdef AMDGPU_RLC_DEBUG_RETRY - u32 rlc_ucode_ver; -#endif + int i, num_xcc; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < num_xcc; i++) + gfx_v9_4_3_xcc_rlc_reset(adev, i); +} - WREG32_FIELD15_PREREG(GC, 0, RLC_CNTL, RLC_ENABLE_F32, 1); +static void gfx_v9_4_3_xcc_rlc_start(struct amdgpu_device *adev, int xcc_id) +{ + WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), RLC_CNTL, + RLC_ENABLE_F32, 1); udelay(50); /* carrizo do enable cp interrupt after cp inited */ if (!(adev->flags & AMD_IS_APU)) { - gfx_v9_4_3_enable_gui_idle_interrupt(adev, true); + gfx_v9_4_3_xcc_enable_gui_idle_interrupt(adev, true, xcc_id); udelay(50); } +} +static void gfx_v9_4_3_rlc_start(struct amdgpu_device *adev) +{ #ifdef AMDGPU_RLC_DEBUG_RETRY - /* RLC_GPM_GENERAL_6 : RLC Ucode version */ - rlc_ucode_ver = RREG32_SOC15(GC, 0, regRLC_GPM_GENERAL_6); - if (rlc_ucode_ver == 0x108) { - dev_info(adev->dev, - "Using rlc debug ucode. regRLC_GPM_GENERAL_6 ==0x08%x / fw_ver == %i \n", - rlc_ucode_ver, adev->gfx.rlc_fw_version); - /* RLC_GPM_TIMER_INT_3 : Timer interval in RefCLK cycles, - * default is 0x9C4 to create a 100us interval */ - WREG32_SOC15(GC, 0, regRLC_GPM_TIMER_INT_3, 0x9C4); - /* RLC_GPM_GENERAL_12 : Minimum gap between wptr and rptr - * to disable the page fault retry interrupts, default is - * 0x100 (256) */ - WREG32_SOC15(GC, 0, regRLC_GPM_GENERAL_12, 0x100); - } + u32 rlc_ucode_ver; +#endif + int i, num_xcc; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < num_xcc; i++) { + gfx_v9_4_3_xcc_rlc_start(adev, i); +#ifdef AMDGPU_RLC_DEBUG_RETRY + /* RLC_GPM_GENERAL_6 : RLC Ucode version */ + rlc_ucode_ver = RREG32_SOC15(GC, GET_INST(GC, i), regRLC_GPM_GENERAL_6); + if (rlc_ucode_ver == 0x108) { + dev_info(adev->dev, + "Using rlc debug ucode. regRLC_GPM_GENERAL_6 ==0x08%x / fw_ver == %i \n", + rlc_ucode_ver, adev->gfx.rlc_fw_version); + /* RLC_GPM_TIMER_INT_3 : Timer interval in RefCLK cycles, + * default is 0x9C4 to create a 100us interval */ + WREG32_SOC15(GC, GET_INST(GC, i), regRLC_GPM_TIMER_INT_3, 0x9C4); + /* RLC_GPM_GENERAL_12 : Minimum gap between wptr and rptr + * to disable the page fault retry interrupts, default is + * 0x100 (256) */ + WREG32_SOC15(GC, GET_INST(GC, i), regRLC_GPM_GENERAL_12, 0x100); + } #endif + } } -static int gfx_v9_4_3_rlc_load_microcode(struct amdgpu_device *adev) +static int gfx_v9_4_3_xcc_rlc_load_microcode(struct amdgpu_device *adev, + int xcc_id) { const struct rlc_firmware_header_v2_0 *hdr; const __le32 *fw_data; @@ -316,49 +1247,65 @@ static int gfx_v9_4_3_rlc_load_microcode(struct amdgpu_device *adev) le32_to_cpu(hdr->header.ucode_array_offset_bytes)); fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4; - WREG32_SOC15(GC, 0, regRLC_GPM_UCODE_ADDR, + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_GPM_UCODE_ADDR, RLCG_UCODE_LOADING_START_ADDRESS); for (i = 0; i < fw_size; i++) { if (amdgpu_emu_mode == 1 && i % 100 == 0) { dev_info(adev->dev, "Write RLC ucode data %u DWs\n", i); msleep(1); } - WREG32_SOC15(GC, 0, regRLC_GPM_UCODE_DATA, le32_to_cpup(fw_data++)); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_GPM_UCODE_DATA, le32_to_cpup(fw_data++)); } - WREG32_SOC15(GC, 0, regRLC_GPM_UCODE_ADDR, adev->gfx.rlc_fw_version); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_GPM_UCODE_ADDR, adev->gfx.rlc_fw_version); return 0; } -static int gfx_v9_4_3_rlc_resume(struct amdgpu_device *adev) +static int gfx_v9_4_3_xcc_rlc_resume(struct amdgpu_device *adev, int xcc_id) { int r; - adev->gfx.rlc.funcs->stop(adev); + if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) { + gfx_v9_4_3_xcc_rlc_stop(adev, xcc_id); + /* legacy rlc firmware loading */ + r = gfx_v9_4_3_xcc_rlc_load_microcode(adev, xcc_id); + if (r) + return r; + gfx_v9_4_3_xcc_rlc_start(adev, xcc_id); + } + amdgpu_gfx_rlc_enter_safe_mode(adev, xcc_id); /* disable CG */ - WREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL, 0); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGCG_CGLS_CTRL, 0); + gfx_v9_4_3_xcc_init_pg(adev, xcc_id); + amdgpu_gfx_rlc_exit_safe_mode(adev, xcc_id); - /* TODO: revisit pg function */ - /* gfx_v9_4_3_init_pg(adev);*/ + return 0; +} - if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) { - /* legacy rlc firmware loading */ - r = gfx_v9_4_3_rlc_load_microcode(adev); +static int gfx_v9_4_3_rlc_resume(struct amdgpu_device *adev) +{ + int r, i, num_xcc; + + if (amdgpu_sriov_vf(adev)) + return 0; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < num_xcc; i++) { + r = gfx_v9_4_3_xcc_rlc_resume(adev, i); if (r) return r; } - adev->gfx.rlc.funcs->start(adev); - return 0; } -static void gfx_v9_4_3_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid) +static void gfx_v9_4_3_update_spm_vmid(struct amdgpu_device *adev, + unsigned vmid) { u32 reg, data; - reg = SOC15_REG_OFFSET(GC, 0, regRLC_SPM_MC_CNTL); + reg = SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regRLC_SPM_MC_CNTL); if (amdgpu_sriov_is_pp_one_vf(adev)) data = RREG32_NO_KIQ(reg); else @@ -368,9 +1315,9 @@ static void gfx_v9_4_3_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid data |= (vmid & RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK) << RLC_SPM_MC_CNTL__RLC_SPM_VMID__SHIFT; if (amdgpu_sriov_is_pp_one_vf(adev)) - WREG32_SOC15_NO_KIQ(GC, 0, regRLC_SPM_MC_CNTL, data); + WREG32_SOC15_NO_KIQ(GC, GET_INST(GC, 0), regRLC_SPM_MC_CNTL, data); else - WREG32_SOC15(GC, 0, regRLC_SPM_MC_CNTL, data); + WREG32_SOC15(GC, GET_INST(GC, 0), regRLC_SPM_MC_CNTL, data); } static const struct soc15_reg_rlcg rlcg_access_gc_9_4_3[] = { @@ -382,7 +1329,7 @@ static bool gfx_v9_4_3_check_rlcg_range(struct amdgpu_device *adev, uint32_t offset, struct soc15_reg_rlcg *entries, int arr_size) { - int i; + int i, inst; uint32_t reg; if (!entries) @@ -392,7 +1339,12 @@ static bool gfx_v9_4_3_check_rlcg_range(struct amdgpu_device *adev, const struct soc15_reg_rlcg *entry; entry = &entries[i]; - reg = adev->reg_offset[entry->hwip][entry->instance][entry->segment] + entry->reg; + inst = adev->ip_map.logical_to_dev_inst ? + adev->ip_map.logical_to_dev_inst( + adev, entry->hwip, entry->instance) : + entry->instance; + reg = adev->reg_offset[entry->hwip][inst][entry->segment] + + entry->reg; if (offset == reg) return true; } @@ -407,19 +1359,1025 @@ static bool gfx_v9_4_3_is_rlcg_access_range(struct amdgpu_device *adev, u32 offs ARRAY_SIZE(rlcg_access_gc_9_4_3)); } -const struct amdgpu_gfx_funcs gfx_v9_4_3_gfx_funcs = { - .get_gpu_clock_counter = &gfx_v9_4_3_get_gpu_clock_counter, - .select_se_sh = &gfx_v9_4_3_select_se_sh, - .read_wave_data = &gfx_v9_4_3_read_wave_data, - .read_wave_sgprs = &gfx_v9_4_3_read_wave_sgprs, - .read_wave_vgprs = &gfx_v9_4_3_read_wave_vgprs, - .select_me_pipe_q = &gfx_v9_4_3_select_me_pipe_q, -}; +static void gfx_v9_4_3_xcc_cp_compute_enable(struct amdgpu_device *adev, + bool enable, int xcc_id) +{ + if (enable) { + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_MEC_CNTL, 0); + } else { + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_MEC_CNTL, + (CP_MEC_CNTL__MEC_ME1_HALT_MASK | CP_MEC_CNTL__MEC_ME2_HALT_MASK)); + adev->gfx.kiq[xcc_id].ring.sched.ready = false; + } + udelay(50); +} + +static int gfx_v9_4_3_xcc_cp_compute_load_microcode(struct amdgpu_device *adev, + int xcc_id) +{ + const struct gfx_firmware_header_v1_0 *mec_hdr; + const __le32 *fw_data; + unsigned i; + u32 tmp; + u32 mec_ucode_addr_offset; + u32 mec_ucode_data_offset; + + if (!adev->gfx.mec_fw) + return -EINVAL; + + gfx_v9_4_3_xcc_cp_compute_enable(adev, false, xcc_id); + + mec_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data; + amdgpu_ucode_print_gfx_hdr(&mec_hdr->header); + + fw_data = (const __le32 *) + (adev->gfx.mec_fw->data + + le32_to_cpu(mec_hdr->header.ucode_array_offset_bytes)); + tmp = 0; + tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, VMID, 0); + tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, CACHE_POLICY, 0); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_CPC_IC_BASE_CNTL, tmp); + + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_CPC_IC_BASE_LO, + adev->gfx.mec.mec_fw_gpu_addr & 0xFFFFF000); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_CPC_IC_BASE_HI, + upper_32_bits(adev->gfx.mec.mec_fw_gpu_addr)); + + mec_ucode_addr_offset = + SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_MEC_ME1_UCODE_ADDR); + mec_ucode_data_offset = + SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_MEC_ME1_UCODE_DATA); + + /* MEC1 */ + WREG32(mec_ucode_addr_offset, mec_hdr->jt_offset); + for (i = 0; i < mec_hdr->jt_size; i++) + WREG32(mec_ucode_data_offset, + le32_to_cpup(fw_data + mec_hdr->jt_offset + i)); + + WREG32(mec_ucode_addr_offset, adev->gfx.mec_fw_version); + /* Todo : Loading MEC2 firmware is only necessary if MEC2 should run different microcode than MEC1. */ + + return 0; +} + +/* KIQ functions */ +static void gfx_v9_4_3_xcc_kiq_setting(struct amdgpu_ring *ring, int xcc_id) +{ + uint32_t tmp; + struct amdgpu_device *adev = ring->adev; + + /* tell RLC which is KIQ queue */ + tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS); + tmp &= 0xffffff00; + tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS, tmp); + tmp |= 0x80; + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS, tmp); +} + +static void gfx_v9_4_3_mqd_set_priority(struct amdgpu_ring *ring, struct v9_mqd *mqd) +{ + struct amdgpu_device *adev = ring->adev; + + if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE) { + if (amdgpu_gfx_is_high_priority_compute_queue(adev, ring)) { + mqd->cp_hqd_pipe_priority = AMDGPU_GFX_PIPE_PRIO_HIGH; + mqd->cp_hqd_queue_priority = + AMDGPU_GFX_QUEUE_PRIORITY_MAXIMUM; + } + } +} + +static int gfx_v9_4_3_xcc_mqd_init(struct amdgpu_ring *ring, int xcc_id) +{ + struct amdgpu_device *adev = ring->adev; + struct v9_mqd *mqd = ring->mqd_ptr; + uint64_t hqd_gpu_addr, wb_gpu_addr, eop_base_addr; + uint32_t tmp; + + mqd->header = 0xC0310800; + mqd->compute_pipelinestat_enable = 0x00000001; + mqd->compute_static_thread_mgmt_se0 = 0xffffffff; + mqd->compute_static_thread_mgmt_se1 = 0xffffffff; + mqd->compute_static_thread_mgmt_se2 = 0xffffffff; + mqd->compute_static_thread_mgmt_se3 = 0xffffffff; + mqd->compute_misc_reserved = 0x00000003; + + mqd->dynamic_cu_mask_addr_lo = + lower_32_bits(ring->mqd_gpu_addr + + offsetof(struct v9_mqd_allocation, dynamic_cu_mask)); + mqd->dynamic_cu_mask_addr_hi = + upper_32_bits(ring->mqd_gpu_addr + + offsetof(struct v9_mqd_allocation, dynamic_cu_mask)); + + eop_base_addr = ring->eop_gpu_addr >> 8; + mqd->cp_hqd_eop_base_addr_lo = eop_base_addr; + mqd->cp_hqd_eop_base_addr_hi = upper_32_bits(eop_base_addr); + + /* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */ + tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_EOP_CONTROL); + tmp = REG_SET_FIELD(tmp, CP_HQD_EOP_CONTROL, EOP_SIZE, + (order_base_2(GFX9_MEC_HPD_SIZE / 4) - 1)); + + mqd->cp_hqd_eop_control = tmp; + + /* enable doorbell? */ + tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_DOORBELL_CONTROL); + + if (ring->use_doorbell) { + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL, + DOORBELL_OFFSET, ring->doorbell_index); + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL, + DOORBELL_EN, 1); + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL, + DOORBELL_SOURCE, 0); + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL, + DOORBELL_HIT, 0); + } else { + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL, + DOORBELL_EN, 0); + } + + mqd->cp_hqd_pq_doorbell_control = tmp; + + /* disable the queue if it's active */ + ring->wptr = 0; + mqd->cp_hqd_dequeue_request = 0; + mqd->cp_hqd_pq_rptr = 0; + mqd->cp_hqd_pq_wptr_lo = 0; + mqd->cp_hqd_pq_wptr_hi = 0; + + /* set the pointer to the MQD */ + mqd->cp_mqd_base_addr_lo = ring->mqd_gpu_addr & 0xfffffffc; + mqd->cp_mqd_base_addr_hi = upper_32_bits(ring->mqd_gpu_addr); + + /* set MQD vmid to 0 */ + tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_MQD_CONTROL); + tmp = REG_SET_FIELD(tmp, CP_MQD_CONTROL, VMID, 0); + mqd->cp_mqd_control = tmp; + + /* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */ + hqd_gpu_addr = ring->gpu_addr >> 8; + mqd->cp_hqd_pq_base_lo = hqd_gpu_addr; + mqd->cp_hqd_pq_base_hi = upper_32_bits(hqd_gpu_addr); + + /* set up the HQD, this is similar to CP_RB0_CNTL */ + tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_CONTROL); + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, QUEUE_SIZE, + (order_base_2(ring->ring_size / 4) - 1)); + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE, + ((order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1) << 8)); +#ifdef __BIG_ENDIAN + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, ENDIAN_SWAP, 1); +#endif + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 0); + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, ROQ_PQ_IB_FLIP, 0); + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, PRIV_STATE, 1); + tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, KMD_QUEUE, 1); + mqd->cp_hqd_pq_control = tmp; + + /* set the wb address whether it's enabled or not */ + wb_gpu_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4); + mqd->cp_hqd_pq_rptr_report_addr_lo = wb_gpu_addr & 0xfffffffc; + mqd->cp_hqd_pq_rptr_report_addr_hi = + upper_32_bits(wb_gpu_addr) & 0xffff; + + /* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */ + wb_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4); + mqd->cp_hqd_pq_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffffc; + mqd->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff; + + /* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */ + ring->wptr = 0; + mqd->cp_hqd_pq_rptr = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_RPTR); + + /* set the vmid for the queue */ + mqd->cp_hqd_vmid = 0; + + tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_PERSISTENT_STATE); + tmp = REG_SET_FIELD(tmp, CP_HQD_PERSISTENT_STATE, PRELOAD_SIZE, 0x53); + mqd->cp_hqd_persistent_state = tmp; + + /* set MIN_IB_AVAIL_SIZE */ + tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_IB_CONTROL); + tmp = REG_SET_FIELD(tmp, CP_HQD_IB_CONTROL, MIN_IB_AVAIL_SIZE, 3); + mqd->cp_hqd_ib_control = tmp; + + /* set static priority for a queue/ring */ + gfx_v9_4_3_mqd_set_priority(ring, mqd); + mqd->cp_hqd_quantum = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_QUANTUM); + + /* map_queues packet doesn't need activate the queue, + * so only kiq need set this field. + */ + if (ring->funcs->type == AMDGPU_RING_TYPE_KIQ) + mqd->cp_hqd_active = 1; + + return 0; +} + +static int gfx_v9_4_3_xcc_kiq_init_register(struct amdgpu_ring *ring, + int xcc_id) +{ + struct amdgpu_device *adev = ring->adev; + struct v9_mqd *mqd = ring->mqd_ptr; + int j; + + /* disable wptr polling */ + WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), CP_PQ_WPTR_POLL_CNTL, EN, 0); + + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_EOP_BASE_ADDR, + mqd->cp_hqd_eop_base_addr_lo); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_EOP_BASE_ADDR_HI, + mqd->cp_hqd_eop_base_addr_hi); + + /* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_EOP_CONTROL, + mqd->cp_hqd_eop_control); + + /* enable doorbell? */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_DOORBELL_CONTROL, + mqd->cp_hqd_pq_doorbell_control); + + /* disable the queue if it's active */ + if (RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE) & 1) { + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_DEQUEUE_REQUEST, 1); + for (j = 0; j < adev->usec_timeout; j++) { + if (!(RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE) & 1)) + break; + udelay(1); + } + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_DEQUEUE_REQUEST, + mqd->cp_hqd_dequeue_request); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_RPTR, + mqd->cp_hqd_pq_rptr); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_LO, + mqd->cp_hqd_pq_wptr_lo); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_HI, + mqd->cp_hqd_pq_wptr_hi); + } + + /* set the pointer to the MQD */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_MQD_BASE_ADDR, + mqd->cp_mqd_base_addr_lo); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_MQD_BASE_ADDR_HI, + mqd->cp_mqd_base_addr_hi); + + /* set MQD vmid to 0 */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_MQD_CONTROL, + mqd->cp_mqd_control); + + /* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_BASE, + mqd->cp_hqd_pq_base_lo); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_BASE_HI, + mqd->cp_hqd_pq_base_hi); + + /* set up the HQD, this is similar to CP_RB0_CNTL */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_CONTROL, + mqd->cp_hqd_pq_control); + + /* set the wb address whether it's enabled or not */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_RPTR_REPORT_ADDR, + mqd->cp_hqd_pq_rptr_report_addr_lo); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_RPTR_REPORT_ADDR_HI, + mqd->cp_hqd_pq_rptr_report_addr_hi); + + /* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_POLL_ADDR, + mqd->cp_hqd_pq_wptr_poll_addr_lo); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_POLL_ADDR_HI, + mqd->cp_hqd_pq_wptr_poll_addr_hi); + + /* enable the doorbell if requested */ + if (ring->use_doorbell) { + WREG32_SOC15( + GC, GET_INST(GC, xcc_id), + regCP_MEC_DOORBELL_RANGE_LOWER, + ((adev->doorbell_index.kiq + + xcc_id * adev->doorbell_index.xcc_doorbell_range) * + 2) << 2); + WREG32_SOC15( + GC, GET_INST(GC, xcc_id), + regCP_MEC_DOORBELL_RANGE_UPPER, + ((adev->doorbell_index.userqueue_end + + xcc_id * adev->doorbell_index.xcc_doorbell_range) * + 2) << 2); + } + + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_DOORBELL_CONTROL, + mqd->cp_hqd_pq_doorbell_control); + + /* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_LO, + mqd->cp_hqd_pq_wptr_lo); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_HI, + mqd->cp_hqd_pq_wptr_hi); + + /* set the vmid for the queue */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_VMID, mqd->cp_hqd_vmid); + + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PERSISTENT_STATE, + mqd->cp_hqd_persistent_state); + + /* activate the queue */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE, + mqd->cp_hqd_active); + + if (ring->use_doorbell) + WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), CP_PQ_STATUS, DOORBELL_ENABLE, 1); + + return 0; +} + +static int gfx_v9_4_3_xcc_q_fini_register(struct amdgpu_ring *ring, + int xcc_id) +{ + struct amdgpu_device *adev = ring->adev; + int j; + + /* disable the queue if it's active */ + if (RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE) & 1) { + + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_DEQUEUE_REQUEST, 1); + + for (j = 0; j < adev->usec_timeout; j++) { + if (!(RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE) & 1)) + break; + udelay(1); + } + + if (j == AMDGPU_MAX_USEC_TIMEOUT) { + DRM_DEBUG("%s dequeue request failed.\n", ring->name); + + /* Manual disable if dequeue request times out */ + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE, 0); + } + + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_DEQUEUE_REQUEST, + 0); + } + + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_IQ_TIMER, 0); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_IB_CONTROL, 0); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PERSISTENT_STATE, 0); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_DOORBELL_CONTROL, 0x40000000); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_DOORBELL_CONTROL, 0); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_RPTR, 0); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_HI, 0); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_LO, 0); + + return 0; +} + +static int gfx_v9_4_3_xcc_kiq_init_queue(struct amdgpu_ring *ring, int xcc_id) +{ + struct amdgpu_device *adev = ring->adev; + struct v9_mqd *mqd = ring->mqd_ptr; + struct v9_mqd *tmp_mqd; + + gfx_v9_4_3_xcc_kiq_setting(ring, xcc_id); + + /* GPU could be in bad state during probe, driver trigger the reset + * after load the SMU, in this case , the mqd is not be initialized. + * driver need to re-init the mqd. + * check mqd->cp_hqd_pq_control since this value should not be 0 + */ + tmp_mqd = (struct v9_mqd *)adev->gfx.kiq[xcc_id].mqd_backup; + if (amdgpu_in_reset(adev) && tmp_mqd->cp_hqd_pq_control) { + /* for GPU_RESET case , reset MQD to a clean status */ + if (adev->gfx.kiq[xcc_id].mqd_backup) + memcpy(mqd, adev->gfx.kiq[xcc_id].mqd_backup, sizeof(struct v9_mqd_allocation)); + + /* reset ring buffer */ + ring->wptr = 0; + amdgpu_ring_clear_ring(ring); + mutex_lock(&adev->srbm_mutex); + soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, GET_INST(GC, xcc_id)); + gfx_v9_4_3_xcc_kiq_init_register(ring, xcc_id); + soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id)); + mutex_unlock(&adev->srbm_mutex); + } else { + memset((void *)mqd, 0, sizeof(struct v9_mqd_allocation)); + ((struct v9_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF; + ((struct v9_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF; + mutex_lock(&adev->srbm_mutex); + soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, GET_INST(GC, xcc_id)); + gfx_v9_4_3_xcc_mqd_init(ring, xcc_id); + gfx_v9_4_3_xcc_kiq_init_register(ring, xcc_id); + soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id)); + mutex_unlock(&adev->srbm_mutex); + + if (adev->gfx.kiq[xcc_id].mqd_backup) + memcpy(adev->gfx.kiq[xcc_id].mqd_backup, mqd, sizeof(struct v9_mqd_allocation)); + } + + return 0; +} + +static int gfx_v9_4_3_xcc_kcq_init_queue(struct amdgpu_ring *ring, int xcc_id) +{ + struct amdgpu_device *adev = ring->adev; + struct v9_mqd *mqd = ring->mqd_ptr; + int mqd_idx = ring - &adev->gfx.compute_ring[0]; + struct v9_mqd *tmp_mqd; + + /* Same as above kiq init, driver need to re-init the mqd if mqd->cp_hqd_pq_control + * is not be initialized before + */ + tmp_mqd = (struct v9_mqd *)adev->gfx.mec.mqd_backup[mqd_idx]; + + if (!tmp_mqd->cp_hqd_pq_control || + (!amdgpu_in_reset(adev) && !adev->in_suspend)) { + memset((void *)mqd, 0, sizeof(struct v9_mqd_allocation)); + ((struct v9_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF; + ((struct v9_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF; + mutex_lock(&adev->srbm_mutex); + soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, GET_INST(GC, xcc_id)); + gfx_v9_4_3_xcc_mqd_init(ring, xcc_id); + soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id)); + mutex_unlock(&adev->srbm_mutex); + + if (adev->gfx.mec.mqd_backup[mqd_idx]) + memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct v9_mqd_allocation)); + } else { + /* restore MQD to a clean status */ + if (adev->gfx.mec.mqd_backup[mqd_idx]) + memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation)); + /* reset ring buffer */ + ring->wptr = 0; + atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs], 0); + amdgpu_ring_clear_ring(ring); + } + + return 0; +} + +static int gfx_v9_4_3_xcc_kcq_fini_register(struct amdgpu_device *adev, int xcc_id) +{ + struct amdgpu_ring *ring; + int j; + + for (j = 0; j < adev->gfx.num_compute_rings; j++) { + ring = &adev->gfx.compute_ring[j + xcc_id * adev->gfx.num_compute_rings]; + if (!amdgpu_in_reset(adev) && !adev->in_suspend) { + mutex_lock(&adev->srbm_mutex); + soc15_grbm_select(adev, ring->me, + ring->pipe, + ring->queue, 0, GET_INST(GC, xcc_id)); + gfx_v9_4_3_xcc_q_fini_register(ring, xcc_id); + soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id)); + mutex_unlock(&adev->srbm_mutex); + } + } + + return 0; +} + +static int gfx_v9_4_3_xcc_kiq_resume(struct amdgpu_device *adev, int xcc_id) +{ + struct amdgpu_ring *ring; + int r; + + ring = &adev->gfx.kiq[xcc_id].ring; + + r = amdgpu_bo_reserve(ring->mqd_obj, false); + if (unlikely(r != 0)) + return r; + + r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); + if (unlikely(r != 0)) { + amdgpu_bo_unreserve(ring->mqd_obj); + return r; + } + + gfx_v9_4_3_xcc_kiq_init_queue(ring, xcc_id); + amdgpu_bo_kunmap(ring->mqd_obj); + ring->mqd_ptr = NULL; + amdgpu_bo_unreserve(ring->mqd_obj); + return 0; +} + +static int gfx_v9_4_3_xcc_kcq_resume(struct amdgpu_device *adev, int xcc_id) +{ + struct amdgpu_ring *ring = NULL; + int r = 0, i; + + gfx_v9_4_3_xcc_cp_compute_enable(adev, true, xcc_id); + + for (i = 0; i < adev->gfx.num_compute_rings; i++) { + ring = &adev->gfx.compute_ring[i + xcc_id * adev->gfx.num_compute_rings]; + + r = amdgpu_bo_reserve(ring->mqd_obj, false); + if (unlikely(r != 0)) + goto done; + r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); + if (!r) { + r = gfx_v9_4_3_xcc_kcq_init_queue(ring, xcc_id); + amdgpu_bo_kunmap(ring->mqd_obj); + ring->mqd_ptr = NULL; + } + amdgpu_bo_unreserve(ring->mqd_obj); + if (r) + goto done; + } + + r = amdgpu_gfx_enable_kcq(adev, xcc_id); +done: + return r; +} + +static int gfx_v9_4_3_xcc_cp_resume(struct amdgpu_device *adev, int xcc_id) +{ + struct amdgpu_ring *ring; + int r, j; + + gfx_v9_4_3_xcc_enable_gui_idle_interrupt(adev, false, xcc_id); + + if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) { + gfx_v9_4_3_xcc_disable_gpa_mode(adev, xcc_id); + + r = gfx_v9_4_3_xcc_cp_compute_load_microcode(adev, xcc_id); + if (r) + return r; + } + + /* set the virtual and physical id based on partition_mode */ + gfx_v9_4_3_xcc_program_xcc_id(adev, xcc_id); + + r = gfx_v9_4_3_xcc_kiq_resume(adev, xcc_id); + if (r) + return r; + + r = gfx_v9_4_3_xcc_kcq_resume(adev, xcc_id); + if (r) + return r; + + for (j = 0; j < adev->gfx.num_compute_rings; j++) { + ring = &adev->gfx.compute_ring + [j + xcc_id * adev->gfx.num_compute_rings]; + r = amdgpu_ring_test_helper(ring); + if (r) + return r; + } + + gfx_v9_4_3_xcc_enable_gui_idle_interrupt(adev, true, xcc_id); + + return 0; +} + +static int gfx_v9_4_3_cp_resume(struct amdgpu_device *adev) +{ + int r = 0, i, num_xcc; + + if (amdgpu_xcp_query_partition_mode(adev->xcp_mgr, + AMDGPU_XCP_FL_NONE) == + AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE) + r = amdgpu_xcp_switch_partition_mode(adev->xcp_mgr, + amdgpu_user_partt_mode); + + if (r) + return r; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < num_xcc; i++) { + r = gfx_v9_4_3_xcc_cp_resume(adev, i); + if (r) + return r; + } + + return 0; +} + +static void gfx_v9_4_3_xcc_cp_enable(struct amdgpu_device *adev, bool enable, + int xcc_id) +{ + gfx_v9_4_3_xcc_cp_compute_enable(adev, enable, xcc_id); +} + +static void gfx_v9_4_3_xcc_fini(struct amdgpu_device *adev, int xcc_id) +{ + if (amdgpu_gfx_disable_kcq(adev, xcc_id)) + DRM_ERROR("XCD %d KCQ disable failed\n", xcc_id); + + /* Use deinitialize sequence from CAIL when unbinding device + * from driver, otherwise KIQ is hanging when binding back + */ + if (!amdgpu_in_reset(adev) && !adev->in_suspend) { + mutex_lock(&adev->srbm_mutex); + soc15_grbm_select(adev, adev->gfx.kiq[xcc_id].ring.me, + adev->gfx.kiq[xcc_id].ring.pipe, + adev->gfx.kiq[xcc_id].ring.queue, 0, + GET_INST(GC, xcc_id)); + gfx_v9_4_3_xcc_q_fini_register(&adev->gfx.kiq[xcc_id].ring, + xcc_id); + soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id)); + mutex_unlock(&adev->srbm_mutex); + } + + gfx_v9_4_3_xcc_kcq_fini_register(adev, xcc_id); + gfx_v9_4_3_xcc_cp_enable(adev, false, xcc_id); +} + +static int gfx_v9_4_3_hw_init(void *handle) +{ + int r; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + gfx_v9_4_3_init_golden_registers(adev); + + gfx_v9_4_3_constants_init(adev); + + r = adev->gfx.rlc.funcs->resume(adev); + if (r) + return r; + + r = gfx_v9_4_3_cp_resume(adev); + if (r) + return r; + + return r; +} + +static int gfx_v9_4_3_hw_fini(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int i, num_xcc; + + amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0); + amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0); + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < num_xcc; i++) { + gfx_v9_4_3_xcc_fini(adev, i); + } + + return 0; +} + +static int gfx_v9_4_3_suspend(void *handle) +{ + return gfx_v9_4_3_hw_fini(handle); +} + +static int gfx_v9_4_3_resume(void *handle) +{ + return gfx_v9_4_3_hw_init(handle); +} + +static bool gfx_v9_4_3_is_idle(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int i, num_xcc; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < num_xcc; i++) { + if (REG_GET_FIELD(RREG32_SOC15(GC, GET_INST(GC, i), regGRBM_STATUS), + GRBM_STATUS, GUI_ACTIVE)) + return false; + } + return true; +} + +static int gfx_v9_4_3_wait_for_idle(void *handle) +{ + unsigned i; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + for (i = 0; i < adev->usec_timeout; i++) { + if (gfx_v9_4_3_is_idle(handle)) + return 0; + udelay(1); + } + return -ETIMEDOUT; +} + +static int gfx_v9_4_3_soft_reset(void *handle) +{ + u32 grbm_soft_reset = 0; + u32 tmp; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + /* GRBM_STATUS */ + tmp = RREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_STATUS); + if (tmp & (GRBM_STATUS__PA_BUSY_MASK | GRBM_STATUS__SC_BUSY_MASK | + GRBM_STATUS__BCI_BUSY_MASK | GRBM_STATUS__SX_BUSY_MASK | + GRBM_STATUS__TA_BUSY_MASK | GRBM_STATUS__VGT_BUSY_MASK | + GRBM_STATUS__DB_BUSY_MASK | GRBM_STATUS__CB_BUSY_MASK | + GRBM_STATUS__GDS_BUSY_MASK | GRBM_STATUS__SPI_BUSY_MASK | + GRBM_STATUS__IA_BUSY_MASK | GRBM_STATUS__IA_BUSY_NO_DMA_MASK)) { + grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, + GRBM_SOFT_RESET, SOFT_RESET_CP, 1); + grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, + GRBM_SOFT_RESET, SOFT_RESET_GFX, 1); + } + + if (tmp & (GRBM_STATUS__CP_BUSY_MASK | GRBM_STATUS__CP_COHERENCY_BUSY_MASK)) { + grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, + GRBM_SOFT_RESET, SOFT_RESET_CP, 1); + } + + /* GRBM_STATUS2 */ + tmp = RREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_STATUS2); + if (REG_GET_FIELD(tmp, GRBM_STATUS2, RLC_BUSY)) + grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, + GRBM_SOFT_RESET, SOFT_RESET_RLC, 1); + + + if (grbm_soft_reset) { + /* stop the rlc */ + adev->gfx.rlc.funcs->stop(adev); + + /* Disable MEC parsing/prefetching */ + gfx_v9_4_3_xcc_cp_compute_enable(adev, false, 0); + + if (grbm_soft_reset) { + tmp = RREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_SOFT_RESET); + tmp |= grbm_soft_reset; + dev_info(adev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp); + WREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_SOFT_RESET, tmp); + tmp = RREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_SOFT_RESET); + + udelay(50); + + tmp &= ~grbm_soft_reset; + WREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_SOFT_RESET, tmp); + tmp = RREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_SOFT_RESET); + } + + /* Wait a little for things to settle down */ + udelay(50); + } + return 0; +} + +static void gfx_v9_4_3_ring_emit_gds_switch(struct amdgpu_ring *ring, + uint32_t vmid, + uint32_t gds_base, uint32_t gds_size, + uint32_t gws_base, uint32_t gws_size, + uint32_t oa_base, uint32_t oa_size) +{ + struct amdgpu_device *adev = ring->adev; + + /* GDS Base */ + gfx_v9_4_3_write_data_to_reg(ring, 0, false, + SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regGDS_VMID0_BASE) + 2 * vmid, + gds_base); + + /* GDS Size */ + gfx_v9_4_3_write_data_to_reg(ring, 0, false, + SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regGDS_VMID0_SIZE) + 2 * vmid, + gds_size); + + /* GWS */ + gfx_v9_4_3_write_data_to_reg(ring, 0, false, + SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regGDS_GWS_VMID0) + vmid, + gws_size << GDS_GWS_VMID0__SIZE__SHIFT | gws_base); + + /* OA */ + gfx_v9_4_3_write_data_to_reg(ring, 0, false, + SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regGDS_OA_VMID0) + vmid, + (1 << (oa_size + oa_base)) - (1 << oa_base)); +} + +static int gfx_v9_4_3_early_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev), + AMDGPU_MAX_COMPUTE_RINGS); + gfx_v9_4_3_set_kiq_pm4_funcs(adev); + gfx_v9_4_3_set_ring_funcs(adev); + gfx_v9_4_3_set_irq_funcs(adev); + gfx_v9_4_3_set_gds_init(adev); + gfx_v9_4_3_set_rlc_funcs(adev); + + return gfx_v9_4_3_init_microcode(adev); +} + +static int gfx_v9_4_3_late_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int r; + + r = amdgpu_irq_get(adev, &adev->gfx.priv_reg_irq, 0); + if (r) + return r; + + r = amdgpu_irq_get(adev, &adev->gfx.priv_inst_irq, 0); + if (r) + return r; + + return 0; +} + +static void gfx_v9_4_3_xcc_update_sram_fgcg(struct amdgpu_device *adev, + bool enable, int xcc_id) +{ + uint32_t def, data; + + if (!(adev->cg_flags & AMD_CG_SUPPORT_GFX_FGCG)) + return; + + def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), + regRLC_CGTT_MGCG_OVERRIDE); + + if (enable) + data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_FGCG_OVERRIDE_MASK; + else + data |= RLC_CGTT_MGCG_OVERRIDE__GFXIP_FGCG_OVERRIDE_MASK; + + if (def != data) + WREG32_SOC15(GC, GET_INST(GC, xcc_id), + regRLC_CGTT_MGCG_OVERRIDE, data); + + def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CLK_CNTL); + + if (enable) + data &= ~RLC_CLK_CNTL__RLC_SRAM_CLK_GATER_OVERRIDE_MASK; + else + data |= RLC_CLK_CNTL__RLC_SRAM_CLK_GATER_OVERRIDE_MASK; + + if (def != data) + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CLK_CNTL, data); +} + +static void gfx_v9_4_3_xcc_update_repeater_fgcg(struct amdgpu_device *adev, + bool enable, int xcc_id) +{ + uint32_t def, data; + + if (!(adev->cg_flags & AMD_CG_SUPPORT_REPEATER_FGCG)) + return; + + def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), + regRLC_CGTT_MGCG_OVERRIDE); + + if (enable) + data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_REP_FGCG_OVERRIDE_MASK; + else + data |= RLC_CGTT_MGCG_OVERRIDE__GFXIP_REP_FGCG_OVERRIDE_MASK; + + if (def != data) + WREG32_SOC15(GC, GET_INST(GC, xcc_id), + regRLC_CGTT_MGCG_OVERRIDE, data); +} + +static void +gfx_v9_4_3_xcc_update_medium_grain_clock_gating(struct amdgpu_device *adev, + bool enable, int xcc_id) +{ + uint32_t data, def; + + /* It is disabled by HW by default */ + if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) { + /* 1 - RLC_CGTT_MGCG_OVERRIDE */ + def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGTT_MGCG_OVERRIDE); + + data &= ~(RLC_CGTT_MGCG_OVERRIDE__GRBM_CGTT_SCLK_OVERRIDE_MASK | + RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGCG_OVERRIDE_MASK | + RLC_CGTT_MGCG_OVERRIDE__RLC_CGTT_SCLK_OVERRIDE_MASK | + RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGLS_OVERRIDE_MASK); + + if (def != data) + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGTT_MGCG_OVERRIDE, data); + + /* MGLS is a global flag to control all MGLS in GFX */ + if (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGLS) { + /* 2 - RLC memory Light sleep */ + if (adev->cg_flags & AMD_CG_SUPPORT_GFX_RLC_LS) { + def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_MEM_SLP_CNTL); + data |= RLC_MEM_SLP_CNTL__RLC_MEM_LS_EN_MASK; + if (def != data) + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_MEM_SLP_CNTL, data); + } + /* 3 - CP memory Light sleep */ + if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CP_LS) { + def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_MEM_SLP_CNTL); + data |= CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK; + if (def != data) + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_MEM_SLP_CNTL, data); + } + } + } else { + /* 1 - MGCG_OVERRIDE */ + def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGTT_MGCG_OVERRIDE); + + data |= (RLC_CGTT_MGCG_OVERRIDE__RLC_CGTT_SCLK_OVERRIDE_MASK | + RLC_CGTT_MGCG_OVERRIDE__GRBM_CGTT_SCLK_OVERRIDE_MASK | + RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGCG_OVERRIDE_MASK | + RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGLS_OVERRIDE_MASK); + + if (def != data) + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGTT_MGCG_OVERRIDE, data); + + /* 2 - disable MGLS in RLC */ + data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_MEM_SLP_CNTL); + if (data & RLC_MEM_SLP_CNTL__RLC_MEM_LS_EN_MASK) { + data &= ~RLC_MEM_SLP_CNTL__RLC_MEM_LS_EN_MASK; + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_MEM_SLP_CNTL, data); + } + + /* 3 - disable MGLS in CP */ + data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_MEM_SLP_CNTL); + if (data & CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK) { + data &= ~CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK; + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_MEM_SLP_CNTL, data); + } + } + +} + +static void +gfx_v9_4_3_xcc_update_coarse_grain_clock_gating(struct amdgpu_device *adev, + bool enable, int xcc_id) +{ + uint32_t def, data; + + if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)) { + + def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGTT_MGCG_OVERRIDE); + /* unset CGCG override */ + data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_CGCG_OVERRIDE_MASK; + if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGLS) + data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_CGLS_OVERRIDE_MASK; + else + data |= RLC_CGTT_MGCG_OVERRIDE__GFXIP_CGLS_OVERRIDE_MASK; + /* update CGCG and CGLS override bits */ + if (def != data) + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGTT_MGCG_OVERRIDE, data); + + /* enable cgcg FSM(0x0000363F) */ + def = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGCG_CGLS_CTRL); + + data = (0x36 + << RLC_CGCG_CGLS_CTRL__CGCG_GFX_IDLE_THRESHOLD__SHIFT) | + RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK; + if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGLS) + data |= (0x000F << RLC_CGCG_CGLS_CTRL__CGLS_REP_COMPANSAT_DELAY__SHIFT) | + RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK; + if (def != data) + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGCG_CGLS_CTRL, data); + + /* set IDLE_POLL_COUNT(0x00900100) */ + def = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_RB_WPTR_POLL_CNTL); + data = (0x0100 << CP_RB_WPTR_POLL_CNTL__POLL_FREQUENCY__SHIFT) | + (0x0090 << CP_RB_WPTR_POLL_CNTL__IDLE_POLL_COUNT__SHIFT); + if (def != data) + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_RB_WPTR_POLL_CNTL, data); + } else { + def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGCG_CGLS_CTRL); + /* reset CGCG/CGLS bits */ + data &= ~(RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK | RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK); + /* disable cgcg and cgls in FSM */ + if (def != data) + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGCG_CGLS_CTRL, data); + } + +} -const struct amdgpu_rlc_funcs gfx_v9_4_3_rlc_funcs = { +static int gfx_v9_4_3_xcc_update_gfx_clock_gating(struct amdgpu_device *adev, + bool enable, int xcc_id) +{ + amdgpu_gfx_rlc_enter_safe_mode(adev, xcc_id); + + if (enable) { + /* FGCG */ + gfx_v9_4_3_xcc_update_sram_fgcg(adev, enable, xcc_id); + gfx_v9_4_3_xcc_update_repeater_fgcg(adev, enable, xcc_id); + + /* CGCG/CGLS should be enabled after MGCG/MGLS + * === MGCG + MGLS === + */ + gfx_v9_4_3_xcc_update_medium_grain_clock_gating(adev, enable, + xcc_id); + /* === CGCG + CGLS === */ + gfx_v9_4_3_xcc_update_coarse_grain_clock_gating(adev, enable, + xcc_id); + } else { + /* CGCG/CGLS should be disabled before MGCG/MGLS + * === CGCG + CGLS === + */ + gfx_v9_4_3_xcc_update_coarse_grain_clock_gating(adev, enable, + xcc_id); + /* === MGCG + MGLS === */ + gfx_v9_4_3_xcc_update_medium_grain_clock_gating(adev, enable, + xcc_id); + + /* FGCG */ + gfx_v9_4_3_xcc_update_sram_fgcg(adev, enable, xcc_id); + gfx_v9_4_3_xcc_update_repeater_fgcg(adev, enable, xcc_id); + } + + amdgpu_gfx_rlc_exit_safe_mode(adev, xcc_id); + + return 0; +} + +static const struct amdgpu_rlc_funcs gfx_v9_4_3_rlc_funcs = { .is_rlc_enabled = gfx_v9_4_3_is_rlc_enabled, - .set_safe_mode = gfx_v9_4_3_set_safe_mode, - .unset_safe_mode = gfx_v9_4_3_unset_safe_mode, + .set_safe_mode = gfx_v9_4_3_xcc_set_safe_mode, + .unset_safe_mode = gfx_v9_4_3_xcc_unset_safe_mode, .init = gfx_v9_4_3_rlc_init, .resume = gfx_v9_4_3_rlc_resume, .stop = gfx_v9_4_3_rlc_stop, @@ -428,3 +2386,1982 @@ const struct amdgpu_rlc_funcs gfx_v9_4_3_rlc_funcs = { .update_spm_vmid = gfx_v9_4_3_update_spm_vmid, .is_rlcg_access_range = gfx_v9_4_3_is_rlcg_access_range, }; + +static int gfx_v9_4_3_set_powergating_state(void *handle, + enum amd_powergating_state state) +{ + return 0; +} + +static int gfx_v9_4_3_set_clockgating_state(void *handle, + enum amd_clockgating_state state) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int i, num_xcc; + + if (amdgpu_sriov_vf(adev)) + return 0; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 4, 3): + for (i = 0; i < num_xcc; i++) + gfx_v9_4_3_xcc_update_gfx_clock_gating( + adev, state == AMD_CG_STATE_GATE, i); + break; + default: + break; + } + return 0; +} + +static void gfx_v9_4_3_get_clockgating_state(void *handle, u64 *flags) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int data; + + if (amdgpu_sriov_vf(adev)) + *flags = 0; + + /* AMD_CG_SUPPORT_GFX_MGCG */ + data = RREG32_KIQ(SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regRLC_CGTT_MGCG_OVERRIDE)); + if (!(data & RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGCG_OVERRIDE_MASK)) + *flags |= AMD_CG_SUPPORT_GFX_MGCG; + + /* AMD_CG_SUPPORT_GFX_CGCG */ + data = RREG32_KIQ(SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regRLC_CGCG_CGLS_CTRL)); + if (data & RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK) + *flags |= AMD_CG_SUPPORT_GFX_CGCG; + + /* AMD_CG_SUPPORT_GFX_CGLS */ + if (data & RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK) + *flags |= AMD_CG_SUPPORT_GFX_CGLS; + + /* AMD_CG_SUPPORT_GFX_RLC_LS */ + data = RREG32_KIQ(SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regRLC_MEM_SLP_CNTL)); + if (data & RLC_MEM_SLP_CNTL__RLC_MEM_LS_EN_MASK) + *flags |= AMD_CG_SUPPORT_GFX_RLC_LS | AMD_CG_SUPPORT_GFX_MGLS; + + /* AMD_CG_SUPPORT_GFX_CP_LS */ + data = RREG32_KIQ(SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regCP_MEM_SLP_CNTL)); + if (data & CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK) + *flags |= AMD_CG_SUPPORT_GFX_CP_LS | AMD_CG_SUPPORT_GFX_MGLS; +} + +static void gfx_v9_4_3_ring_emit_hdp_flush(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + u32 ref_and_mask, reg_mem_engine; + const struct nbio_hdp_flush_reg *nbio_hf_reg = adev->nbio.hdp_flush_reg; + + if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE) { + switch (ring->me) { + case 1: + ref_and_mask = nbio_hf_reg->ref_and_mask_cp2 << ring->pipe; + break; + case 2: + ref_and_mask = nbio_hf_reg->ref_and_mask_cp6 << ring->pipe; + break; + default: + return; + } + reg_mem_engine = 0; + } else { + ref_and_mask = nbio_hf_reg->ref_and_mask_cp0; + reg_mem_engine = 1; /* pfp */ + } + + gfx_v9_4_3_wait_reg_mem(ring, reg_mem_engine, 0, 1, + adev->nbio.funcs->get_hdp_flush_req_offset(adev), + adev->nbio.funcs->get_hdp_flush_done_offset(adev), + ref_and_mask, ref_and_mask, 0x20); +} + +static void gfx_v9_4_3_ring_emit_ib_compute(struct amdgpu_ring *ring, + struct amdgpu_job *job, + struct amdgpu_ib *ib, + uint32_t flags) +{ + unsigned vmid = AMDGPU_JOB_GET_VMID(job); + u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24); + + /* Currently, there is a high possibility to get wave ID mismatch + * between ME and GDS, leading to a hw deadlock, because ME generates + * different wave IDs than the GDS expects. This situation happens + * randomly when at least 5 compute pipes use GDS ordered append. + * The wave IDs generated by ME are also wrong after suspend/resume. + * Those are probably bugs somewhere else in the kernel driver. + * + * Writing GDS_COMPUTE_MAX_WAVE_ID resets wave ID counters in ME and + * GDS to 0 for this ring (me/pipe). + */ + if (ib->flags & AMDGPU_IB_FLAG_RESET_GDS_MAX_WAVE_ID) { + amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + amdgpu_ring_write(ring, regGDS_COMPUTE_MAX_WAVE_ID); + amdgpu_ring_write(ring, ring->adev->gds.gds_compute_max_wave_id); + } + + amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); + BUG_ON(ib->gpu_addr & 0x3); /* Dword align */ + amdgpu_ring_write(ring, +#ifdef __BIG_ENDIAN + (2 << 0) | +#endif + lower_32_bits(ib->gpu_addr)); + amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr)); + amdgpu_ring_write(ring, control); +} + +static void gfx_v9_4_3_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, + u64 seq, unsigned flags) +{ + bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT; + bool int_sel = flags & AMDGPU_FENCE_FLAG_INT; + bool writeback = flags & AMDGPU_FENCE_FLAG_TC_WB_ONLY; + + /* RELEASE_MEM - flush caches, send int */ + amdgpu_ring_write(ring, PACKET3(PACKET3_RELEASE_MEM, 6)); + amdgpu_ring_write(ring, ((writeback ? (EOP_TC_WB_ACTION_EN | + EOP_TC_NC_ACTION_EN) : + (EOP_TCL1_ACTION_EN | + EOP_TC_ACTION_EN | + EOP_TC_WB_ACTION_EN | + EOP_TC_MD_ACTION_EN)) | + EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) | + EVENT_INDEX(5))); + amdgpu_ring_write(ring, DATA_SEL(write64bit ? 2 : 1) | INT_SEL(int_sel ? 2 : 0)); + + /* + * the address should be Qword aligned if 64bit write, Dword + * aligned if only send 32bit data low (discard data high) + */ + if (write64bit) + BUG_ON(addr & 0x7); + else + BUG_ON(addr & 0x3); + amdgpu_ring_write(ring, lower_32_bits(addr)); + amdgpu_ring_write(ring, upper_32_bits(addr)); + amdgpu_ring_write(ring, lower_32_bits(seq)); + amdgpu_ring_write(ring, upper_32_bits(seq)); + amdgpu_ring_write(ring, 0); +} + +static void gfx_v9_4_3_ring_emit_pipeline_sync(struct amdgpu_ring *ring) +{ + int usepfp = (ring->funcs->type == AMDGPU_RING_TYPE_GFX); + uint32_t seq = ring->fence_drv.sync_seq; + uint64_t addr = ring->fence_drv.gpu_addr; + + gfx_v9_4_3_wait_reg_mem(ring, usepfp, 1, 0, + lower_32_bits(addr), upper_32_bits(addr), + seq, 0xffffffff, 4); +} + +static void gfx_v9_4_3_ring_emit_vm_flush(struct amdgpu_ring *ring, + unsigned vmid, uint64_t pd_addr) +{ + amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr); +} + +static u64 gfx_v9_4_3_ring_get_rptr_compute(struct amdgpu_ring *ring) +{ + return ring->adev->wb.wb[ring->rptr_offs]; /* gfx9 hardware is 32bit rptr */ +} + +static u64 gfx_v9_4_3_ring_get_wptr_compute(struct amdgpu_ring *ring) +{ + u64 wptr; + + /* XXX check if swapping is necessary on BE */ + if (ring->use_doorbell) + wptr = atomic64_read((atomic64_t *)&ring->adev->wb.wb[ring->wptr_offs]); + else + BUG(); + return wptr; +} + +static void gfx_v9_4_3_ring_set_wptr_compute(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + + /* XXX check if swapping is necessary on BE */ + if (ring->use_doorbell) { + atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs], ring->wptr); + WDOORBELL64(ring->doorbell_index, ring->wptr); + } else { + BUG(); /* only DOORBELL method supported on gfx9 now */ + } +} + +static void gfx_v9_4_3_ring_emit_fence_kiq(struct amdgpu_ring *ring, u64 addr, + u64 seq, unsigned int flags) +{ + struct amdgpu_device *adev = ring->adev; + + /* we only allocate 32bit for each seq wb address */ + BUG_ON(flags & AMDGPU_FENCE_FLAG_64BIT); + + /* write fence seq to the "addr" */ + amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); + amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | + WRITE_DATA_DST_SEL(5) | WR_CONFIRM)); + amdgpu_ring_write(ring, lower_32_bits(addr)); + amdgpu_ring_write(ring, upper_32_bits(addr)); + amdgpu_ring_write(ring, lower_32_bits(seq)); + + if (flags & AMDGPU_FENCE_FLAG_INT) { + /* set register to trigger INT */ + amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); + amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | + WRITE_DATA_DST_SEL(0) | WR_CONFIRM)); + amdgpu_ring_write(ring, SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regCPC_INT_STATUS)); + amdgpu_ring_write(ring, 0); + amdgpu_ring_write(ring, 0x20000000); /* src_id is 178 */ + } +} + +static void gfx_v9_4_3_ring_emit_rreg(struct amdgpu_ring *ring, uint32_t reg, + uint32_t reg_val_offs) +{ + struct amdgpu_device *adev = ring->adev; + + amdgpu_ring_write(ring, PACKET3(PACKET3_COPY_DATA, 4)); + amdgpu_ring_write(ring, 0 | /* src: register*/ + (5 << 8) | /* dst: memory */ + (1 << 20)); /* write confirm */ + amdgpu_ring_write(ring, reg); + amdgpu_ring_write(ring, 0); + amdgpu_ring_write(ring, lower_32_bits(adev->wb.gpu_addr + + reg_val_offs * 4)); + amdgpu_ring_write(ring, upper_32_bits(adev->wb.gpu_addr + + reg_val_offs * 4)); +} + +static void gfx_v9_4_3_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, + uint32_t val) +{ + uint32_t cmd = 0; + + switch (ring->funcs->type) { + case AMDGPU_RING_TYPE_GFX: + cmd = WRITE_DATA_ENGINE_SEL(1) | WR_CONFIRM; + break; + case AMDGPU_RING_TYPE_KIQ: + cmd = (1 << 16); /* no inc addr */ + break; + default: + cmd = WR_CONFIRM; + break; + } + amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); + amdgpu_ring_write(ring, cmd); + amdgpu_ring_write(ring, reg); + amdgpu_ring_write(ring, 0); + amdgpu_ring_write(ring, val); +} + +static void gfx_v9_4_3_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg, + uint32_t val, uint32_t mask) +{ + gfx_v9_4_3_wait_reg_mem(ring, 0, 0, 0, reg, 0, val, mask, 0x20); +} + +static void gfx_v9_4_3_ring_emit_reg_write_reg_wait(struct amdgpu_ring *ring, + uint32_t reg0, uint32_t reg1, + uint32_t ref, uint32_t mask) +{ + amdgpu_ring_emit_reg_write_reg_wait_helper(ring, reg0, reg1, + ref, mask); +} + +static void gfx_v9_4_3_xcc_set_compute_eop_interrupt_state( + struct amdgpu_device *adev, int me, int pipe, + enum amdgpu_interrupt_state state, int xcc_id) +{ + u32 mec_int_cntl, mec_int_cntl_reg; + + /* + * amdgpu controls only the first MEC. That's why this function only + * handles the setting of interrupts for this specific MEC. All other + * pipes' interrupts are set by amdkfd. + */ + + if (me == 1) { + switch (pipe) { + case 0: + mec_int_cntl_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_ME1_PIPE0_INT_CNTL); + break; + case 1: + mec_int_cntl_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_ME1_PIPE1_INT_CNTL); + break; + case 2: + mec_int_cntl_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_ME1_PIPE2_INT_CNTL); + break; + case 3: + mec_int_cntl_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_ME1_PIPE3_INT_CNTL); + break; + default: + DRM_DEBUG("invalid pipe %d\n", pipe); + return; + } + } else { + DRM_DEBUG("invalid me %d\n", me); + return; + } + + switch (state) { + case AMDGPU_IRQ_STATE_DISABLE: + mec_int_cntl = RREG32(mec_int_cntl_reg); + mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL, + TIME_STAMP_INT_ENABLE, 0); + WREG32(mec_int_cntl_reg, mec_int_cntl); + break; + case AMDGPU_IRQ_STATE_ENABLE: + mec_int_cntl = RREG32(mec_int_cntl_reg); + mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL, + TIME_STAMP_INT_ENABLE, 1); + WREG32(mec_int_cntl_reg, mec_int_cntl); + break; + default: + break; + } +} + +static int gfx_v9_4_3_set_priv_reg_fault_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned type, + enum amdgpu_interrupt_state state) +{ + int i, num_xcc; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + switch (state) { + case AMDGPU_IRQ_STATE_DISABLE: + case AMDGPU_IRQ_STATE_ENABLE: + for (i = 0; i < num_xcc; i++) + WREG32_FIELD15_PREREG(GC, GET_INST(GC, i), CP_INT_CNTL_RING0, + PRIV_REG_INT_ENABLE, + state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0); + break; + default: + break; + } + + return 0; +} + +static int gfx_v9_4_3_set_priv_inst_fault_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned type, + enum amdgpu_interrupt_state state) +{ + int i, num_xcc; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + switch (state) { + case AMDGPU_IRQ_STATE_DISABLE: + case AMDGPU_IRQ_STATE_ENABLE: + for (i = 0; i < num_xcc; i++) + WREG32_FIELD15_PREREG(GC, GET_INST(GC, i), CP_INT_CNTL_RING0, + PRIV_INSTR_INT_ENABLE, + state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0); + break; + default: + break; + } + + return 0; +} + +static int gfx_v9_4_3_set_eop_interrupt_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *src, + unsigned type, + enum amdgpu_interrupt_state state) +{ + int i, num_xcc; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < num_xcc; i++) { + switch (type) { + case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP: + gfx_v9_4_3_xcc_set_compute_eop_interrupt_state( + adev, 1, 0, state, i); + break; + case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE1_EOP: + gfx_v9_4_3_xcc_set_compute_eop_interrupt_state( + adev, 1, 1, state, i); + break; + case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE2_EOP: + gfx_v9_4_3_xcc_set_compute_eop_interrupt_state( + adev, 1, 2, state, i); + break; + case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE3_EOP: + gfx_v9_4_3_xcc_set_compute_eop_interrupt_state( + adev, 1, 3, state, i); + break; + case AMDGPU_CP_IRQ_COMPUTE_MEC2_PIPE0_EOP: + gfx_v9_4_3_xcc_set_compute_eop_interrupt_state( + adev, 2, 0, state, i); + break; + case AMDGPU_CP_IRQ_COMPUTE_MEC2_PIPE1_EOP: + gfx_v9_4_3_xcc_set_compute_eop_interrupt_state( + adev, 2, 1, state, i); + break; + case AMDGPU_CP_IRQ_COMPUTE_MEC2_PIPE2_EOP: + gfx_v9_4_3_xcc_set_compute_eop_interrupt_state( + adev, 2, 2, state, i); + break; + case AMDGPU_CP_IRQ_COMPUTE_MEC2_PIPE3_EOP: + gfx_v9_4_3_xcc_set_compute_eop_interrupt_state( + adev, 2, 3, state, i); + break; + default: + break; + } + } + + return 0; +} + +static int gfx_v9_4_3_eop_irq(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + struct amdgpu_iv_entry *entry) +{ + int i, xcc_id; + u8 me_id, pipe_id, queue_id; + struct amdgpu_ring *ring; + + DRM_DEBUG("IH: CP EOP\n"); + me_id = (entry->ring_id & 0x0c) >> 2; + pipe_id = (entry->ring_id & 0x03) >> 0; + queue_id = (entry->ring_id & 0x70) >> 4; + + xcc_id = gfx_v9_4_3_ih_to_xcc_inst(adev, entry->node_id); + + if (xcc_id == -EINVAL) + return -EINVAL; + + switch (me_id) { + case 0: + case 1: + case 2: + for (i = 0; i < adev->gfx.num_compute_rings; i++) { + ring = &adev->gfx.compute_ring + [i + + xcc_id * adev->gfx.num_compute_rings]; + /* Per-queue interrupt is supported for MEC starting from VI. + * The interrupt can only be enabled/disabled per pipe instead of per queue. + */ + + if ((ring->me == me_id) && (ring->pipe == pipe_id) && (ring->queue == queue_id)) + amdgpu_fence_process(ring); + } + break; + } + return 0; +} + +static void gfx_v9_4_3_fault(struct amdgpu_device *adev, + struct amdgpu_iv_entry *entry) +{ + u8 me_id, pipe_id, queue_id; + struct amdgpu_ring *ring; + int i, xcc_id; + + me_id = (entry->ring_id & 0x0c) >> 2; + pipe_id = (entry->ring_id & 0x03) >> 0; + queue_id = (entry->ring_id & 0x70) >> 4; + + xcc_id = gfx_v9_4_3_ih_to_xcc_inst(adev, entry->node_id); + + if (xcc_id == -EINVAL) + return; + + switch (me_id) { + case 0: + case 1: + case 2: + for (i = 0; i < adev->gfx.num_compute_rings; i++) { + ring = &adev->gfx.compute_ring + [i + + xcc_id * adev->gfx.num_compute_rings]; + if (ring->me == me_id && ring->pipe == pipe_id && + ring->queue == queue_id) + drm_sched_fault(&ring->sched); + } + break; + } +} + +static int gfx_v9_4_3_priv_reg_irq(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + struct amdgpu_iv_entry *entry) +{ + DRM_ERROR("Illegal register access in command stream\n"); + gfx_v9_4_3_fault(adev, entry); + return 0; +} + +static int gfx_v9_4_3_priv_inst_irq(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + struct amdgpu_iv_entry *entry) +{ + DRM_ERROR("Illegal instruction in command stream\n"); + gfx_v9_4_3_fault(adev, entry); + return 0; +} + +static void gfx_v9_4_3_emit_mem_sync(struct amdgpu_ring *ring) +{ + const unsigned int cp_coher_cntl = + PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_ICACHE_ACTION_ENA(1) | + PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_KCACHE_ACTION_ENA(1) | + PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TC_ACTION_ENA(1) | + PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TCL1_ACTION_ENA(1) | + PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TC_WB_ACTION_ENA(1); + + /* ACQUIRE_MEM -make one or more surfaces valid for use by the subsequent operations */ + amdgpu_ring_write(ring, PACKET3(PACKET3_ACQUIRE_MEM, 5)); + amdgpu_ring_write(ring, cp_coher_cntl); /* CP_COHER_CNTL */ + amdgpu_ring_write(ring, 0xffffffff); /* CP_COHER_SIZE */ + amdgpu_ring_write(ring, 0xffffff); /* CP_COHER_SIZE_HI */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE_HI */ + amdgpu_ring_write(ring, 0x0000000A); /* POLL_INTERVAL */ +} + +static void gfx_v9_4_3_emit_wave_limit_cs(struct amdgpu_ring *ring, + uint32_t pipe, bool enable) +{ + struct amdgpu_device *adev = ring->adev; + uint32_t val; + uint32_t wcl_cs_reg; + + /* regSPI_WCL_PIPE_PERCENT_CS[0-7]_DEFAULT values are same */ + val = enable ? 0x1 : 0x7f; + + switch (pipe) { + case 0: + wcl_cs_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regSPI_WCL_PIPE_PERCENT_CS0); + break; + case 1: + wcl_cs_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regSPI_WCL_PIPE_PERCENT_CS1); + break; + case 2: + wcl_cs_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regSPI_WCL_PIPE_PERCENT_CS2); + break; + case 3: + wcl_cs_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regSPI_WCL_PIPE_PERCENT_CS3); + break; + default: + DRM_DEBUG("invalid pipe %d\n", pipe); + return; + } + + amdgpu_ring_emit_wreg(ring, wcl_cs_reg, val); + +} +static void gfx_v9_4_3_emit_wave_limit(struct amdgpu_ring *ring, bool enable) +{ + struct amdgpu_device *adev = ring->adev; + uint32_t val; + int i; + + /* regSPI_WCL_PIPE_PERCENT_GFX is 7 bit multiplier register to limit + * number of gfx waves. Setting 5 bit will make sure gfx only gets + * around 25% of gpu resources. + */ + val = enable ? 0x1f : 0x07ffffff; + amdgpu_ring_emit_wreg(ring, + SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regSPI_WCL_PIPE_PERCENT_GFX), + val); + + /* Restrict waves for normal/low priority compute queues as well + * to get best QoS for high priority compute jobs. + * + * amdgpu controls only 1st ME(0-3 CS pipes). + */ + for (i = 0; i < adev->gfx.mec.num_pipe_per_mec; i++) { + if (i != ring->pipe) + gfx_v9_4_3_emit_wave_limit_cs(ring, i, enable); + + } +} + +enum amdgpu_gfx_cp_ras_mem_id { + AMDGPU_GFX_CP_MEM1 = 1, + AMDGPU_GFX_CP_MEM2, + AMDGPU_GFX_CP_MEM3, + AMDGPU_GFX_CP_MEM4, + AMDGPU_GFX_CP_MEM5, +}; + +enum amdgpu_gfx_gcea_ras_mem_id { + AMDGPU_GFX_GCEA_IOWR_CMDMEM = 4, + AMDGPU_GFX_GCEA_IORD_CMDMEM, + AMDGPU_GFX_GCEA_GMIWR_CMDMEM, + AMDGPU_GFX_GCEA_GMIRD_CMDMEM, + AMDGPU_GFX_GCEA_DRAMWR_CMDMEM, + AMDGPU_GFX_GCEA_DRAMRD_CMDMEM, + AMDGPU_GFX_GCEA_MAM_DMEM0, + AMDGPU_GFX_GCEA_MAM_DMEM1, + AMDGPU_GFX_GCEA_MAM_DMEM2, + AMDGPU_GFX_GCEA_MAM_DMEM3, + AMDGPU_GFX_GCEA_MAM_AMEM0, + AMDGPU_GFX_GCEA_MAM_AMEM1, + AMDGPU_GFX_GCEA_MAM_AMEM2, + AMDGPU_GFX_GCEA_MAM_AMEM3, + AMDGPU_GFX_GCEA_MAM_AFLUSH_BUFFER, + AMDGPU_GFX_GCEA_WRET_TAGMEM, + AMDGPU_GFX_GCEA_RRET_TAGMEM, + AMDGPU_GFX_GCEA_IOWR_DATAMEM, + AMDGPU_GFX_GCEA_GMIWR_DATAMEM, + AMDGPU_GFX_GCEA_DRAM_DATAMEM, +}; + +enum amdgpu_gfx_gc_cane_ras_mem_id { + AMDGPU_GFX_GC_CANE_MEM0 = 0, +}; + +enum amdgpu_gfx_gcutcl2_ras_mem_id { + AMDGPU_GFX_GCUTCL2_MEM2P512X95 = 160, +}; + +enum amdgpu_gfx_gds_ras_mem_id { + AMDGPU_GFX_GDS_MEM0 = 0, +}; + +enum amdgpu_gfx_lds_ras_mem_id { + AMDGPU_GFX_LDS_BANK0 = 0, + AMDGPU_GFX_LDS_BANK1, + AMDGPU_GFX_LDS_BANK2, + AMDGPU_GFX_LDS_BANK3, + AMDGPU_GFX_LDS_BANK4, + AMDGPU_GFX_LDS_BANK5, + AMDGPU_GFX_LDS_BANK6, + AMDGPU_GFX_LDS_BANK7, + AMDGPU_GFX_LDS_BANK8, + AMDGPU_GFX_LDS_BANK9, + AMDGPU_GFX_LDS_BANK10, + AMDGPU_GFX_LDS_BANK11, + AMDGPU_GFX_LDS_BANK12, + AMDGPU_GFX_LDS_BANK13, + AMDGPU_GFX_LDS_BANK14, + AMDGPU_GFX_LDS_BANK15, + AMDGPU_GFX_LDS_BANK16, + AMDGPU_GFX_LDS_BANK17, + AMDGPU_GFX_LDS_BANK18, + AMDGPU_GFX_LDS_BANK19, + AMDGPU_GFX_LDS_BANK20, + AMDGPU_GFX_LDS_BANK21, + AMDGPU_GFX_LDS_BANK22, + AMDGPU_GFX_LDS_BANK23, + AMDGPU_GFX_LDS_BANK24, + AMDGPU_GFX_LDS_BANK25, + AMDGPU_GFX_LDS_BANK26, + AMDGPU_GFX_LDS_BANK27, + AMDGPU_GFX_LDS_BANK28, + AMDGPU_GFX_LDS_BANK29, + AMDGPU_GFX_LDS_BANK30, + AMDGPU_GFX_LDS_BANK31, + AMDGPU_GFX_LDS_SP_BUFFER_A, + AMDGPU_GFX_LDS_SP_BUFFER_B, +}; + +enum amdgpu_gfx_rlc_ras_mem_id { + AMDGPU_GFX_RLC_GPMF32 = 1, + AMDGPU_GFX_RLC_RLCVF32, + AMDGPU_GFX_RLC_SCRATCH, + AMDGPU_GFX_RLC_SRM_ARAM, + AMDGPU_GFX_RLC_SRM_DRAM, + AMDGPU_GFX_RLC_TCTAG, + AMDGPU_GFX_RLC_SPM_SE, + AMDGPU_GFX_RLC_SPM_GRBMT, +}; + +enum amdgpu_gfx_sp_ras_mem_id { + AMDGPU_GFX_SP_SIMDID0 = 0, +}; + +enum amdgpu_gfx_spi_ras_mem_id { + AMDGPU_GFX_SPI_MEM0 = 0, + AMDGPU_GFX_SPI_MEM1, + AMDGPU_GFX_SPI_MEM2, + AMDGPU_GFX_SPI_MEM3, +}; + +enum amdgpu_gfx_sqc_ras_mem_id { + AMDGPU_GFX_SQC_INST_CACHE_A = 100, + AMDGPU_GFX_SQC_INST_CACHE_B = 101, + AMDGPU_GFX_SQC_INST_CACHE_TAG_A = 102, + AMDGPU_GFX_SQC_INST_CACHE_TAG_B = 103, + AMDGPU_GFX_SQC_INST_CACHE_MISS_FIFO_A = 104, + AMDGPU_GFX_SQC_INST_CACHE_MISS_FIFO_B = 105, + AMDGPU_GFX_SQC_INST_CACHE_GATCL1_MISS_FIFO_A = 106, + AMDGPU_GFX_SQC_INST_CACHE_GATCL1_MISS_FIFO_B = 107, + AMDGPU_GFX_SQC_DATA_CACHE_A = 200, + AMDGPU_GFX_SQC_DATA_CACHE_B = 201, + AMDGPU_GFX_SQC_DATA_CACHE_TAG_A = 202, + AMDGPU_GFX_SQC_DATA_CACHE_TAG_B = 203, + AMDGPU_GFX_SQC_DATA_CACHE_MISS_FIFO_A = 204, + AMDGPU_GFX_SQC_DATA_CACHE_MISS_FIFO_B = 205, + AMDGPU_GFX_SQC_DATA_CACHE_HIT_FIFO_A = 206, + AMDGPU_GFX_SQC_DATA_CACHE_HIT_FIFO_B = 207, + AMDGPU_GFX_SQC_DIRTY_BIT_A = 208, + AMDGPU_GFX_SQC_DIRTY_BIT_B = 209, + AMDGPU_GFX_SQC_WRITE_DATA_BUFFER_CU0 = 210, + AMDGPU_GFX_SQC_WRITE_DATA_BUFFER_CU1 = 211, + AMDGPU_GFX_SQC_UTCL1_MISS_LFIFO_DATA_CACHE_A = 212, + AMDGPU_GFX_SQC_UTCL1_MISS_LFIFO_DATA_CACHE_B = 213, + AMDGPU_GFX_SQC_UTCL1_MISS_LFIFO_INST_CACHE = 108, +}; + +enum amdgpu_gfx_sq_ras_mem_id { + AMDGPU_GFX_SQ_SGPR_MEM0 = 0, + AMDGPU_GFX_SQ_SGPR_MEM1, + AMDGPU_GFX_SQ_SGPR_MEM2, + AMDGPU_GFX_SQ_SGPR_MEM3, +}; + +enum amdgpu_gfx_ta_ras_mem_id { + AMDGPU_GFX_TA_FS_AFIFO_RAM_LO = 1, + AMDGPU_GFX_TA_FS_AFIFO_RAM_HI, + AMDGPU_GFX_TA_FS_CFIFO_RAM, + AMDGPU_GFX_TA_FSX_LFIFO, + AMDGPU_GFX_TA_FS_DFIFO_RAM, +}; + +enum amdgpu_gfx_tcc_ras_mem_id { + AMDGPU_GFX_TCC_MEM1 = 1, +}; + +enum amdgpu_gfx_tca_ras_mem_id { + AMDGPU_GFX_TCA_MEM1 = 1, +}; + +enum amdgpu_gfx_tci_ras_mem_id { + AMDGPU_GFX_TCIW_MEM = 1, +}; + +enum amdgpu_gfx_tcp_ras_mem_id { + AMDGPU_GFX_TCP_LFIFO0 = 1, + AMDGPU_GFX_TCP_SET0BANK0_RAM, + AMDGPU_GFX_TCP_SET0BANK1_RAM, + AMDGPU_GFX_TCP_SET0BANK2_RAM, + AMDGPU_GFX_TCP_SET0BANK3_RAM, + AMDGPU_GFX_TCP_SET1BANK0_RAM, + AMDGPU_GFX_TCP_SET1BANK1_RAM, + AMDGPU_GFX_TCP_SET1BANK2_RAM, + AMDGPU_GFX_TCP_SET1BANK3_RAM, + AMDGPU_GFX_TCP_SET2BANK0_RAM, + AMDGPU_GFX_TCP_SET2BANK1_RAM, + AMDGPU_GFX_TCP_SET2BANK2_RAM, + AMDGPU_GFX_TCP_SET2BANK3_RAM, + AMDGPU_GFX_TCP_SET3BANK0_RAM, + AMDGPU_GFX_TCP_SET3BANK1_RAM, + AMDGPU_GFX_TCP_SET3BANK2_RAM, + AMDGPU_GFX_TCP_SET3BANK3_RAM, + AMDGPU_GFX_TCP_VM_FIFO, + AMDGPU_GFX_TCP_DB_TAGRAM0, + AMDGPU_GFX_TCP_DB_TAGRAM1, + AMDGPU_GFX_TCP_DB_TAGRAM2, + AMDGPU_GFX_TCP_DB_TAGRAM3, + AMDGPU_GFX_TCP_UTCL1_LFIFO_PROBE0, + AMDGPU_GFX_TCP_UTCL1_LFIFO_PROBE1, + AMDGPU_GFX_TCP_CMD_FIFO, +}; + +enum amdgpu_gfx_td_ras_mem_id { + AMDGPU_GFX_TD_UTD_CS_FIFO_MEM = 1, + AMDGPU_GFX_TD_UTD_SS_FIFO_LO_MEM, + AMDGPU_GFX_TD_UTD_SS_FIFO_HI_MEM, +}; + +enum amdgpu_gfx_tcx_ras_mem_id { + AMDGPU_GFX_TCX_FIFOD0 = 0, + AMDGPU_GFX_TCX_FIFOD1, + AMDGPU_GFX_TCX_FIFOD2, + AMDGPU_GFX_TCX_FIFOD3, + AMDGPU_GFX_TCX_FIFOD4, + AMDGPU_GFX_TCX_FIFOD5, + AMDGPU_GFX_TCX_FIFOD6, + AMDGPU_GFX_TCX_FIFOD7, + AMDGPU_GFX_TCX_FIFOB0, + AMDGPU_GFX_TCX_FIFOB1, + AMDGPU_GFX_TCX_FIFOB2, + AMDGPU_GFX_TCX_FIFOB3, + AMDGPU_GFX_TCX_FIFOB4, + AMDGPU_GFX_TCX_FIFOB5, + AMDGPU_GFX_TCX_FIFOB6, + AMDGPU_GFX_TCX_FIFOB7, + AMDGPU_GFX_TCX_FIFOA0, + AMDGPU_GFX_TCX_FIFOA1, + AMDGPU_GFX_TCX_FIFOA2, + AMDGPU_GFX_TCX_FIFOA3, + AMDGPU_GFX_TCX_FIFOA4, + AMDGPU_GFX_TCX_FIFOA5, + AMDGPU_GFX_TCX_FIFOA6, + AMDGPU_GFX_TCX_FIFOA7, + AMDGPU_GFX_TCX_CFIFO0, + AMDGPU_GFX_TCX_CFIFO1, + AMDGPU_GFX_TCX_CFIFO2, + AMDGPU_GFX_TCX_CFIFO3, + AMDGPU_GFX_TCX_CFIFO4, + AMDGPU_GFX_TCX_CFIFO5, + AMDGPU_GFX_TCX_CFIFO6, + AMDGPU_GFX_TCX_CFIFO7, + AMDGPU_GFX_TCX_FIFO_ACKB0, + AMDGPU_GFX_TCX_FIFO_ACKB1, + AMDGPU_GFX_TCX_FIFO_ACKB2, + AMDGPU_GFX_TCX_FIFO_ACKB3, + AMDGPU_GFX_TCX_FIFO_ACKB4, + AMDGPU_GFX_TCX_FIFO_ACKB5, + AMDGPU_GFX_TCX_FIFO_ACKB6, + AMDGPU_GFX_TCX_FIFO_ACKB7, + AMDGPU_GFX_TCX_FIFO_ACKD0, + AMDGPU_GFX_TCX_FIFO_ACKD1, + AMDGPU_GFX_TCX_FIFO_ACKD2, + AMDGPU_GFX_TCX_FIFO_ACKD3, + AMDGPU_GFX_TCX_FIFO_ACKD4, + AMDGPU_GFX_TCX_FIFO_ACKD5, + AMDGPU_GFX_TCX_FIFO_ACKD6, + AMDGPU_GFX_TCX_FIFO_ACKD7, + AMDGPU_GFX_TCX_DST_FIFOA0, + AMDGPU_GFX_TCX_DST_FIFOA1, + AMDGPU_GFX_TCX_DST_FIFOA2, + AMDGPU_GFX_TCX_DST_FIFOA3, + AMDGPU_GFX_TCX_DST_FIFOA4, + AMDGPU_GFX_TCX_DST_FIFOA5, + AMDGPU_GFX_TCX_DST_FIFOA6, + AMDGPU_GFX_TCX_DST_FIFOA7, + AMDGPU_GFX_TCX_DST_FIFOB0, + AMDGPU_GFX_TCX_DST_FIFOB1, + AMDGPU_GFX_TCX_DST_FIFOB2, + AMDGPU_GFX_TCX_DST_FIFOB3, + AMDGPU_GFX_TCX_DST_FIFOB4, + AMDGPU_GFX_TCX_DST_FIFOB5, + AMDGPU_GFX_TCX_DST_FIFOB6, + AMDGPU_GFX_TCX_DST_FIFOB7, + AMDGPU_GFX_TCX_DST_FIFOD0, + AMDGPU_GFX_TCX_DST_FIFOD1, + AMDGPU_GFX_TCX_DST_FIFOD2, + AMDGPU_GFX_TCX_DST_FIFOD3, + AMDGPU_GFX_TCX_DST_FIFOD4, + AMDGPU_GFX_TCX_DST_FIFOD5, + AMDGPU_GFX_TCX_DST_FIFOD6, + AMDGPU_GFX_TCX_DST_FIFOD7, + AMDGPU_GFX_TCX_DST_FIFO_ACKB0, + AMDGPU_GFX_TCX_DST_FIFO_ACKB1, + AMDGPU_GFX_TCX_DST_FIFO_ACKB2, + AMDGPU_GFX_TCX_DST_FIFO_ACKB3, + AMDGPU_GFX_TCX_DST_FIFO_ACKB4, + AMDGPU_GFX_TCX_DST_FIFO_ACKB5, + AMDGPU_GFX_TCX_DST_FIFO_ACKB6, + AMDGPU_GFX_TCX_DST_FIFO_ACKB7, + AMDGPU_GFX_TCX_DST_FIFO_ACKD0, + AMDGPU_GFX_TCX_DST_FIFO_ACKD1, + AMDGPU_GFX_TCX_DST_FIFO_ACKD2, + AMDGPU_GFX_TCX_DST_FIFO_ACKD3, + AMDGPU_GFX_TCX_DST_FIFO_ACKD4, + AMDGPU_GFX_TCX_DST_FIFO_ACKD5, + AMDGPU_GFX_TCX_DST_FIFO_ACKD6, + AMDGPU_GFX_TCX_DST_FIFO_ACKD7, +}; + +enum amdgpu_gfx_atc_l2_ras_mem_id { + AMDGPU_GFX_ATC_L2_MEM0 = 0, +}; + +enum amdgpu_gfx_utcl2_ras_mem_id { + AMDGPU_GFX_UTCL2_MEM0 = 0, +}; + +enum amdgpu_gfx_vml2_ras_mem_id { + AMDGPU_GFX_VML2_MEM0 = 0, +}; + +enum amdgpu_gfx_vml2_walker_ras_mem_id { + AMDGPU_GFX_VML2_WALKER_MEM0 = 0, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_cp_mem_list[] = { + {AMDGPU_GFX_CP_MEM1, "CP_MEM1"}, + {AMDGPU_GFX_CP_MEM2, "CP_MEM2"}, + {AMDGPU_GFX_CP_MEM3, "CP_MEM3"}, + {AMDGPU_GFX_CP_MEM4, "CP_MEM4"}, + {AMDGPU_GFX_CP_MEM5, "CP_MEM5"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_gcea_mem_list[] = { + {AMDGPU_GFX_GCEA_IOWR_CMDMEM, "GCEA_IOWR_CMDMEM"}, + {AMDGPU_GFX_GCEA_IORD_CMDMEM, "GCEA_IORD_CMDMEM"}, + {AMDGPU_GFX_GCEA_GMIWR_CMDMEM, "GCEA_GMIWR_CMDMEM"}, + {AMDGPU_GFX_GCEA_GMIRD_CMDMEM, "GCEA_GMIRD_CMDMEM"}, + {AMDGPU_GFX_GCEA_DRAMWR_CMDMEM, "GCEA_DRAMWR_CMDMEM"}, + {AMDGPU_GFX_GCEA_DRAMRD_CMDMEM, "GCEA_DRAMRD_CMDMEM"}, + {AMDGPU_GFX_GCEA_MAM_DMEM0, "GCEA_MAM_DMEM0"}, + {AMDGPU_GFX_GCEA_MAM_DMEM1, "GCEA_MAM_DMEM1"}, + {AMDGPU_GFX_GCEA_MAM_DMEM2, "GCEA_MAM_DMEM2"}, + {AMDGPU_GFX_GCEA_MAM_DMEM3, "GCEA_MAM_DMEM3"}, + {AMDGPU_GFX_GCEA_MAM_AMEM0, "GCEA_MAM_AMEM0"}, + {AMDGPU_GFX_GCEA_MAM_AMEM1, "GCEA_MAM_AMEM1"}, + {AMDGPU_GFX_GCEA_MAM_AMEM2, "GCEA_MAM_AMEM2"}, + {AMDGPU_GFX_GCEA_MAM_AMEM3, "GCEA_MAM_AMEM3"}, + {AMDGPU_GFX_GCEA_MAM_AFLUSH_BUFFER, "GCEA_MAM_AFLUSH_BUFFER"}, + {AMDGPU_GFX_GCEA_WRET_TAGMEM, "GCEA_WRET_TAGMEM"}, + {AMDGPU_GFX_GCEA_RRET_TAGMEM, "GCEA_RRET_TAGMEM"}, + {AMDGPU_GFX_GCEA_IOWR_DATAMEM, "GCEA_IOWR_DATAMEM"}, + {AMDGPU_GFX_GCEA_GMIWR_DATAMEM, "GCEA_GMIWR_DATAMEM"}, + {AMDGPU_GFX_GCEA_DRAM_DATAMEM, "GCEA_DRAM_DATAMEM"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_gc_cane_mem_list[] = { + {AMDGPU_GFX_GC_CANE_MEM0, "GC_CANE_MEM0"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_gcutcl2_mem_list[] = { + {AMDGPU_GFX_GCUTCL2_MEM2P512X95, "GCUTCL2_MEM2P512X95"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_gds_mem_list[] = { + {AMDGPU_GFX_GDS_MEM0, "GDS_MEM"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_lds_mem_list[] = { + {AMDGPU_GFX_LDS_BANK0, "LDS_BANK0"}, + {AMDGPU_GFX_LDS_BANK1, "LDS_BANK1"}, + {AMDGPU_GFX_LDS_BANK2, "LDS_BANK2"}, + {AMDGPU_GFX_LDS_BANK3, "LDS_BANK3"}, + {AMDGPU_GFX_LDS_BANK4, "LDS_BANK4"}, + {AMDGPU_GFX_LDS_BANK5, "LDS_BANK5"}, + {AMDGPU_GFX_LDS_BANK6, "LDS_BANK6"}, + {AMDGPU_GFX_LDS_BANK7, "LDS_BANK7"}, + {AMDGPU_GFX_LDS_BANK8, "LDS_BANK8"}, + {AMDGPU_GFX_LDS_BANK9, "LDS_BANK9"}, + {AMDGPU_GFX_LDS_BANK10, "LDS_BANK10"}, + {AMDGPU_GFX_LDS_BANK11, "LDS_BANK11"}, + {AMDGPU_GFX_LDS_BANK12, "LDS_BANK12"}, + {AMDGPU_GFX_LDS_BANK13, "LDS_BANK13"}, + {AMDGPU_GFX_LDS_BANK14, "LDS_BANK14"}, + {AMDGPU_GFX_LDS_BANK15, "LDS_BANK15"}, + {AMDGPU_GFX_LDS_BANK16, "LDS_BANK16"}, + {AMDGPU_GFX_LDS_BANK17, "LDS_BANK17"}, + {AMDGPU_GFX_LDS_BANK18, "LDS_BANK18"}, + {AMDGPU_GFX_LDS_BANK19, "LDS_BANK19"}, + {AMDGPU_GFX_LDS_BANK20, "LDS_BANK20"}, + {AMDGPU_GFX_LDS_BANK21, "LDS_BANK21"}, + {AMDGPU_GFX_LDS_BANK22, "LDS_BANK22"}, + {AMDGPU_GFX_LDS_BANK23, "LDS_BANK23"}, + {AMDGPU_GFX_LDS_BANK24, "LDS_BANK24"}, + {AMDGPU_GFX_LDS_BANK25, "LDS_BANK25"}, + {AMDGPU_GFX_LDS_BANK26, "LDS_BANK26"}, + {AMDGPU_GFX_LDS_BANK27, "LDS_BANK27"}, + {AMDGPU_GFX_LDS_BANK28, "LDS_BANK28"}, + {AMDGPU_GFX_LDS_BANK29, "LDS_BANK29"}, + {AMDGPU_GFX_LDS_BANK30, "LDS_BANK30"}, + {AMDGPU_GFX_LDS_BANK31, "LDS_BANK31"}, + {AMDGPU_GFX_LDS_SP_BUFFER_A, "LDS_SP_BUFFER_A"}, + {AMDGPU_GFX_LDS_SP_BUFFER_B, "LDS_SP_BUFFER_B"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_rlc_mem_list[] = { + {AMDGPU_GFX_RLC_GPMF32, "RLC_GPMF32"}, + {AMDGPU_GFX_RLC_RLCVF32, "RLC_RLCVF32"}, + {AMDGPU_GFX_RLC_SCRATCH, "RLC_SCRATCH"}, + {AMDGPU_GFX_RLC_SRM_ARAM, "RLC_SRM_ARAM"}, + {AMDGPU_GFX_RLC_SRM_DRAM, "RLC_SRM_DRAM"}, + {AMDGPU_GFX_RLC_TCTAG, "RLC_TCTAG"}, + {AMDGPU_GFX_RLC_SPM_SE, "RLC_SPM_SE"}, + {AMDGPU_GFX_RLC_SPM_GRBMT, "RLC_SPM_GRBMT"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_sp_mem_list[] = { + {AMDGPU_GFX_SP_SIMDID0, "SP_SIMDID0"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_spi_mem_list[] = { + {AMDGPU_GFX_SPI_MEM0, "SPI_MEM0"}, + {AMDGPU_GFX_SPI_MEM1, "SPI_MEM1"}, + {AMDGPU_GFX_SPI_MEM2, "SPI_MEM2"}, + {AMDGPU_GFX_SPI_MEM3, "SPI_MEM3"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_sqc_mem_list[] = { + {AMDGPU_GFX_SQC_INST_CACHE_A, "SQC_INST_CACHE_A"}, + {AMDGPU_GFX_SQC_INST_CACHE_B, "SQC_INST_CACHE_B"}, + {AMDGPU_GFX_SQC_INST_CACHE_TAG_A, "SQC_INST_CACHE_TAG_A"}, + {AMDGPU_GFX_SQC_INST_CACHE_TAG_B, "SQC_INST_CACHE_TAG_B"}, + {AMDGPU_GFX_SQC_INST_CACHE_MISS_FIFO_A, "SQC_INST_CACHE_MISS_FIFO_A"}, + {AMDGPU_GFX_SQC_INST_CACHE_MISS_FIFO_B, "SQC_INST_CACHE_MISS_FIFO_B"}, + {AMDGPU_GFX_SQC_INST_CACHE_GATCL1_MISS_FIFO_A, "SQC_INST_CACHE_GATCL1_MISS_FIFO_A"}, + {AMDGPU_GFX_SQC_INST_CACHE_GATCL1_MISS_FIFO_B, "SQC_INST_CACHE_GATCL1_MISS_FIFO_B"}, + {AMDGPU_GFX_SQC_DATA_CACHE_A, "SQC_DATA_CACHE_A"}, + {AMDGPU_GFX_SQC_DATA_CACHE_B, "SQC_DATA_CACHE_B"}, + {AMDGPU_GFX_SQC_DATA_CACHE_TAG_A, "SQC_DATA_CACHE_TAG_A"}, + {AMDGPU_GFX_SQC_DATA_CACHE_TAG_B, "SQC_DATA_CACHE_TAG_B"}, + {AMDGPU_GFX_SQC_DATA_CACHE_MISS_FIFO_A, "SQC_DATA_CACHE_MISS_FIFO_A"}, + {AMDGPU_GFX_SQC_DATA_CACHE_MISS_FIFO_B, "SQC_DATA_CACHE_MISS_FIFO_B"}, + {AMDGPU_GFX_SQC_DATA_CACHE_HIT_FIFO_A, "SQC_DATA_CACHE_HIT_FIFO_A"}, + {AMDGPU_GFX_SQC_DATA_CACHE_HIT_FIFO_B, "SQC_DATA_CACHE_HIT_FIFO_B"}, + {AMDGPU_GFX_SQC_DIRTY_BIT_A, "SQC_DIRTY_BIT_A"}, + {AMDGPU_GFX_SQC_DIRTY_BIT_B, "SQC_DIRTY_BIT_B"}, + {AMDGPU_GFX_SQC_WRITE_DATA_BUFFER_CU0, "SQC_WRITE_DATA_BUFFER_CU0"}, + {AMDGPU_GFX_SQC_WRITE_DATA_BUFFER_CU1, "SQC_WRITE_DATA_BUFFER_CU1"}, + {AMDGPU_GFX_SQC_UTCL1_MISS_LFIFO_DATA_CACHE_A, "SQC_UTCL1_MISS_LFIFO_DATA_CACHE_A"}, + {AMDGPU_GFX_SQC_UTCL1_MISS_LFIFO_DATA_CACHE_B, "SQC_UTCL1_MISS_LFIFO_DATA_CACHE_B"}, + {AMDGPU_GFX_SQC_UTCL1_MISS_LFIFO_INST_CACHE, "SQC_UTCL1_MISS_LFIFO_INST_CACHE"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_sq_mem_list[] = { + {AMDGPU_GFX_SQ_SGPR_MEM0, "SQ_SGPR_MEM0"}, + {AMDGPU_GFX_SQ_SGPR_MEM1, "SQ_SGPR_MEM1"}, + {AMDGPU_GFX_SQ_SGPR_MEM2, "SQ_SGPR_MEM2"}, + {AMDGPU_GFX_SQ_SGPR_MEM3, "SQ_SGPR_MEM3"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_ta_mem_list[] = { + {AMDGPU_GFX_TA_FS_AFIFO_RAM_LO, "TA_FS_AFIFO_RAM_LO"}, + {AMDGPU_GFX_TA_FS_AFIFO_RAM_HI, "TA_FS_AFIFO_RAM_HI"}, + {AMDGPU_GFX_TA_FS_CFIFO_RAM, "TA_FS_CFIFO_RAM"}, + {AMDGPU_GFX_TA_FSX_LFIFO, "TA_FSX_LFIFO"}, + {AMDGPU_GFX_TA_FS_DFIFO_RAM, "TA_FS_DFIFO_RAM"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_tcc_mem_list[] = { + {AMDGPU_GFX_TCC_MEM1, "TCC_MEM1"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_tca_mem_list[] = { + {AMDGPU_GFX_TCA_MEM1, "TCA_MEM1"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_tci_mem_list[] = { + {AMDGPU_GFX_TCIW_MEM, "TCIW_MEM"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_tcp_mem_list[] = { + {AMDGPU_GFX_TCP_LFIFO0, "TCP_LFIFO0"}, + {AMDGPU_GFX_TCP_SET0BANK0_RAM, "TCP_SET0BANK0_RAM"}, + {AMDGPU_GFX_TCP_SET0BANK1_RAM, "TCP_SET0BANK1_RAM"}, + {AMDGPU_GFX_TCP_SET0BANK2_RAM, "TCP_SET0BANK2_RAM"}, + {AMDGPU_GFX_TCP_SET0BANK3_RAM, "TCP_SET0BANK3_RAM"}, + {AMDGPU_GFX_TCP_SET1BANK0_RAM, "TCP_SET1BANK0_RAM"}, + {AMDGPU_GFX_TCP_SET1BANK1_RAM, "TCP_SET1BANK1_RAM"}, + {AMDGPU_GFX_TCP_SET1BANK2_RAM, "TCP_SET1BANK2_RAM"}, + {AMDGPU_GFX_TCP_SET1BANK3_RAM, "TCP_SET1BANK3_RAM"}, + {AMDGPU_GFX_TCP_SET2BANK0_RAM, "TCP_SET2BANK0_RAM"}, + {AMDGPU_GFX_TCP_SET2BANK1_RAM, "TCP_SET2BANK1_RAM"}, + {AMDGPU_GFX_TCP_SET2BANK2_RAM, "TCP_SET2BANK2_RAM"}, + {AMDGPU_GFX_TCP_SET2BANK3_RAM, "TCP_SET2BANK3_RAM"}, + {AMDGPU_GFX_TCP_SET3BANK0_RAM, "TCP_SET3BANK0_RAM"}, + {AMDGPU_GFX_TCP_SET3BANK1_RAM, "TCP_SET3BANK1_RAM"}, + {AMDGPU_GFX_TCP_SET3BANK2_RAM, "TCP_SET3BANK2_RAM"}, + {AMDGPU_GFX_TCP_SET3BANK3_RAM, "TCP_SET3BANK3_RAM"}, + {AMDGPU_GFX_TCP_VM_FIFO, "TCP_VM_FIFO"}, + {AMDGPU_GFX_TCP_DB_TAGRAM0, "TCP_DB_TAGRAM0"}, + {AMDGPU_GFX_TCP_DB_TAGRAM1, "TCP_DB_TAGRAM1"}, + {AMDGPU_GFX_TCP_DB_TAGRAM2, "TCP_DB_TAGRAM2"}, + {AMDGPU_GFX_TCP_DB_TAGRAM3, "TCP_DB_TAGRAM3"}, + {AMDGPU_GFX_TCP_UTCL1_LFIFO_PROBE0, "TCP_UTCL1_LFIFO_PROBE0"}, + {AMDGPU_GFX_TCP_UTCL1_LFIFO_PROBE1, "TCP_UTCL1_LFIFO_PROBE1"}, + {AMDGPU_GFX_TCP_CMD_FIFO, "TCP_CMD_FIFO"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_td_mem_list[] = { + {AMDGPU_GFX_TD_UTD_CS_FIFO_MEM, "TD_UTD_CS_FIFO_MEM"}, + {AMDGPU_GFX_TD_UTD_SS_FIFO_LO_MEM, "TD_UTD_SS_FIFO_LO_MEM"}, + {AMDGPU_GFX_TD_UTD_SS_FIFO_HI_MEM, "TD_UTD_SS_FIFO_HI_MEM"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_tcx_mem_list[] = { + {AMDGPU_GFX_TCX_FIFOD0, "TCX_FIFOD0"}, + {AMDGPU_GFX_TCX_FIFOD1, "TCX_FIFOD1"}, + {AMDGPU_GFX_TCX_FIFOD2, "TCX_FIFOD2"}, + {AMDGPU_GFX_TCX_FIFOD3, "TCX_FIFOD3"}, + {AMDGPU_GFX_TCX_FIFOD4, "TCX_FIFOD4"}, + {AMDGPU_GFX_TCX_FIFOD5, "TCX_FIFOD5"}, + {AMDGPU_GFX_TCX_FIFOD6, "TCX_FIFOD6"}, + {AMDGPU_GFX_TCX_FIFOD7, "TCX_FIFOD7"}, + {AMDGPU_GFX_TCX_FIFOB0, "TCX_FIFOB0"}, + {AMDGPU_GFX_TCX_FIFOB1, "TCX_FIFOB1"}, + {AMDGPU_GFX_TCX_FIFOB2, "TCX_FIFOB2"}, + {AMDGPU_GFX_TCX_FIFOB3, "TCX_FIFOB3"}, + {AMDGPU_GFX_TCX_FIFOB4, "TCX_FIFOB4"}, + {AMDGPU_GFX_TCX_FIFOB5, "TCX_FIFOB5"}, + {AMDGPU_GFX_TCX_FIFOB6, "TCX_FIFOB6"}, + {AMDGPU_GFX_TCX_FIFOB7, "TCX_FIFOB7"}, + {AMDGPU_GFX_TCX_FIFOA0, "TCX_FIFOA0"}, + {AMDGPU_GFX_TCX_FIFOA1, "TCX_FIFOA1"}, + {AMDGPU_GFX_TCX_FIFOA2, "TCX_FIFOA2"}, + {AMDGPU_GFX_TCX_FIFOA3, "TCX_FIFOA3"}, + {AMDGPU_GFX_TCX_FIFOA4, "TCX_FIFOA4"}, + {AMDGPU_GFX_TCX_FIFOA5, "TCX_FIFOA5"}, + {AMDGPU_GFX_TCX_FIFOA6, "TCX_FIFOA6"}, + {AMDGPU_GFX_TCX_FIFOA7, "TCX_FIFOA7"}, + {AMDGPU_GFX_TCX_CFIFO0, "TCX_CFIFO0"}, + {AMDGPU_GFX_TCX_CFIFO1, "TCX_CFIFO1"}, + {AMDGPU_GFX_TCX_CFIFO2, "TCX_CFIFO2"}, + {AMDGPU_GFX_TCX_CFIFO3, "TCX_CFIFO3"}, + {AMDGPU_GFX_TCX_CFIFO4, "TCX_CFIFO4"}, + {AMDGPU_GFX_TCX_CFIFO5, "TCX_CFIFO5"}, + {AMDGPU_GFX_TCX_CFIFO6, "TCX_CFIFO6"}, + {AMDGPU_GFX_TCX_CFIFO7, "TCX_CFIFO7"}, + {AMDGPU_GFX_TCX_FIFO_ACKB0, "TCX_FIFO_ACKB0"}, + {AMDGPU_GFX_TCX_FIFO_ACKB1, "TCX_FIFO_ACKB1"}, + {AMDGPU_GFX_TCX_FIFO_ACKB2, "TCX_FIFO_ACKB2"}, + {AMDGPU_GFX_TCX_FIFO_ACKB3, "TCX_FIFO_ACKB3"}, + {AMDGPU_GFX_TCX_FIFO_ACKB4, "TCX_FIFO_ACKB4"}, + {AMDGPU_GFX_TCX_FIFO_ACKB5, "TCX_FIFO_ACKB5"}, + {AMDGPU_GFX_TCX_FIFO_ACKB6, "TCX_FIFO_ACKB6"}, + {AMDGPU_GFX_TCX_FIFO_ACKB7, "TCX_FIFO_ACKB7"}, + {AMDGPU_GFX_TCX_FIFO_ACKD0, "TCX_FIFO_ACKD0"}, + {AMDGPU_GFX_TCX_FIFO_ACKD1, "TCX_FIFO_ACKD1"}, + {AMDGPU_GFX_TCX_FIFO_ACKD2, "TCX_FIFO_ACKD2"}, + {AMDGPU_GFX_TCX_FIFO_ACKD3, "TCX_FIFO_ACKD3"}, + {AMDGPU_GFX_TCX_FIFO_ACKD4, "TCX_FIFO_ACKD4"}, + {AMDGPU_GFX_TCX_FIFO_ACKD5, "TCX_FIFO_ACKD5"}, + {AMDGPU_GFX_TCX_FIFO_ACKD6, "TCX_FIFO_ACKD6"}, + {AMDGPU_GFX_TCX_FIFO_ACKD7, "TCX_FIFO_ACKD7"}, + {AMDGPU_GFX_TCX_DST_FIFOA0, "TCX_DST_FIFOA0"}, + {AMDGPU_GFX_TCX_DST_FIFOA1, "TCX_DST_FIFOA1"}, + {AMDGPU_GFX_TCX_DST_FIFOA2, "TCX_DST_FIFOA2"}, + {AMDGPU_GFX_TCX_DST_FIFOA3, "TCX_DST_FIFOA3"}, + {AMDGPU_GFX_TCX_DST_FIFOA4, "TCX_DST_FIFOA4"}, + {AMDGPU_GFX_TCX_DST_FIFOA5, "TCX_DST_FIFOA5"}, + {AMDGPU_GFX_TCX_DST_FIFOA6, "TCX_DST_FIFOA6"}, + {AMDGPU_GFX_TCX_DST_FIFOA7, "TCX_DST_FIFOA7"}, + {AMDGPU_GFX_TCX_DST_FIFOB0, "TCX_DST_FIFOB0"}, + {AMDGPU_GFX_TCX_DST_FIFOB1, "TCX_DST_FIFOB1"}, + {AMDGPU_GFX_TCX_DST_FIFOB2, "TCX_DST_FIFOB2"}, + {AMDGPU_GFX_TCX_DST_FIFOB3, "TCX_DST_FIFOB3"}, + {AMDGPU_GFX_TCX_DST_FIFOB4, "TCX_DST_FIFOB4"}, + {AMDGPU_GFX_TCX_DST_FIFOB5, "TCX_DST_FIFOB5"}, + {AMDGPU_GFX_TCX_DST_FIFOB6, "TCX_DST_FIFOB6"}, + {AMDGPU_GFX_TCX_DST_FIFOB7, "TCX_DST_FIFOB7"}, + {AMDGPU_GFX_TCX_DST_FIFOD0, "TCX_DST_FIFOD0"}, + {AMDGPU_GFX_TCX_DST_FIFOD1, "TCX_DST_FIFOD1"}, + {AMDGPU_GFX_TCX_DST_FIFOD2, "TCX_DST_FIFOD2"}, + {AMDGPU_GFX_TCX_DST_FIFOD3, "TCX_DST_FIFOD3"}, + {AMDGPU_GFX_TCX_DST_FIFOD4, "TCX_DST_FIFOD4"}, + {AMDGPU_GFX_TCX_DST_FIFOD5, "TCX_DST_FIFOD5"}, + {AMDGPU_GFX_TCX_DST_FIFOD6, "TCX_DST_FIFOD6"}, + {AMDGPU_GFX_TCX_DST_FIFOD7, "TCX_DST_FIFOD7"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKB0, "TCX_DST_FIFO_ACKB0"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKB1, "TCX_DST_FIFO_ACKB1"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKB2, "TCX_DST_FIFO_ACKB2"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKB3, "TCX_DST_FIFO_ACKB3"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKB4, "TCX_DST_FIFO_ACKB4"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKB5, "TCX_DST_FIFO_ACKB5"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKB6, "TCX_DST_FIFO_ACKB6"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKB7, "TCX_DST_FIFO_ACKB7"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKD0, "TCX_DST_FIFO_ACKD0"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKD1, "TCX_DST_FIFO_ACKD1"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKD2, "TCX_DST_FIFO_ACKD2"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKD3, "TCX_DST_FIFO_ACKD3"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKD4, "TCX_DST_FIFO_ACKD4"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKD5, "TCX_DST_FIFO_ACKD5"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKD6, "TCX_DST_FIFO_ACKD6"}, + {AMDGPU_GFX_TCX_DST_FIFO_ACKD7, "TCX_DST_FIFO_ACKD7"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_atc_l2_mem_list[] = { + {AMDGPU_GFX_ATC_L2_MEM, "ATC_L2_MEM"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_utcl2_mem_list[] = { + {AMDGPU_GFX_UTCL2_MEM, "UTCL2_MEM"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_vml2_mem_list[] = { + {AMDGPU_GFX_VML2_MEM, "VML2_MEM"}, +}; + +static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_vml2_walker_mem_list[] = { + {AMDGPU_GFX_VML2_WALKER_MEM, "VML2_WALKER_MEM"}, +}; + +static const struct amdgpu_gfx_ras_mem_id_entry gfx_v9_4_3_ras_mem_list_array[AMDGPU_GFX_MEM_TYPE_NUM] = { + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_cp_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_gcea_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_gc_cane_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_gcutcl2_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_gds_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_lds_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_rlc_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_sp_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_spi_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_sqc_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_sq_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_ta_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_tcc_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_tca_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_tci_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_tcp_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_td_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_tcx_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_atc_l2_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_utcl2_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_vml2_mem_list) + AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_vml2_walker_mem_list) +}; + +static const struct amdgpu_gfx_ras_reg_entry gfx_v9_4_3_ce_reg_list[] = { + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regRLC_CE_ERR_STATUS_LOW, regRLC_CE_ERR_STATUS_HIGH), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "RLC"}, + AMDGPU_GFX_RLC_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regCPC_CE_ERR_STATUS_LO, regCPC_CE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CPC"}, + AMDGPU_GFX_CP_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regCPF_CE_ERR_STATUS_LO, regCPF_CE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CPF"}, + AMDGPU_GFX_CP_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regCPG_CE_ERR_STATUS_LO, regCPG_CE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CPG"}, + AMDGPU_GFX_CP_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regGDS_CE_ERR_STATUS_LO, regGDS_CE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "GDS"}, + AMDGPU_GFX_GDS_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regGC_CANE_CE_ERR_STATUS_LO, regGC_CANE_CE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CANE"}, + AMDGPU_GFX_GC_CANE_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regSPI_CE_ERR_STATUS_LO, regSPI_CE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SPI"}, + AMDGPU_GFX_SPI_MEM, 8}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regSP0_CE_ERR_STATUS_LO, regSP0_CE_ERR_STATUS_HI), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SP0"}, + AMDGPU_GFX_SP_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regSP1_CE_ERR_STATUS_LO, regSP1_CE_ERR_STATUS_HI), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SP1"}, + AMDGPU_GFX_SP_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regSQ_CE_ERR_STATUS_LO, regSQ_CE_ERR_STATUS_HI), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SQ"}, + AMDGPU_GFX_SQ_MEM, 8}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regSQC_CE_EDC_LO, regSQC_CE_EDC_HI), + 5, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SQC"}, + AMDGPU_GFX_SQC_MEM, 8}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCX_CE_ERR_STATUS_LO, regTCX_CE_ERR_STATUS_HI), + 2, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCX"}, + AMDGPU_GFX_TCX_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCC_CE_ERR_STATUS_LO, regTCC_CE_ERR_STATUS_HI), + 16, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCC"}, + AMDGPU_GFX_TCC_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTA_CE_EDC_LO, regTA_CE_EDC_HI), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TA"}, + AMDGPU_GFX_TA_MEM, 8}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCI_CE_EDC_LO_REG, regTCI_CE_EDC_HI_REG), + 31, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCI"}, + AMDGPU_GFX_TCI_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCP_CE_EDC_LO_REG, regTCP_CE_EDC_HI_REG), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCP"}, + AMDGPU_GFX_TCP_MEM, 8}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTD_CE_EDC_LO, regTD_CE_EDC_HI), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TD"}, + AMDGPU_GFX_TD_MEM, 8}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regGCEA_CE_ERR_STATUS_LO, regGCEA_CE_ERR_STATUS_HI), + 16, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "GCEA"}, + AMDGPU_GFX_GCEA_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regLDS_CE_ERR_STATUS_LO, regLDS_CE_ERR_STATUS_HI), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "LDS"}, + AMDGPU_GFX_LDS_MEM, 1}, +}; + +static const struct amdgpu_gfx_ras_reg_entry gfx_v9_4_3_ue_reg_list[] = { + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regRLC_UE_ERR_STATUS_LOW, regRLC_UE_ERR_STATUS_HIGH), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "RLC"}, + AMDGPU_GFX_RLC_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regCPC_UE_ERR_STATUS_LO, regCPC_UE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CPC"}, + AMDGPU_GFX_CP_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regCPF_UE_ERR_STATUS_LO, regCPF_UE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CPF"}, + AMDGPU_GFX_CP_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regCPG_UE_ERR_STATUS_LO, regCPG_UE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CPG"}, + AMDGPU_GFX_CP_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regGDS_UE_ERR_STATUS_LO, regGDS_UE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "GDS"}, + AMDGPU_GFX_GDS_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regGC_CANE_UE_ERR_STATUS_LO, regGC_CANE_UE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CANE"}, + AMDGPU_GFX_GC_CANE_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regSPI_UE_ERR_STATUS_LO, regSPI_UE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SPI"}, + AMDGPU_GFX_SPI_MEM, 8}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regSP0_UE_ERR_STATUS_LO, regSP0_UE_ERR_STATUS_HI), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SP0"}, + AMDGPU_GFX_SP_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regSP1_UE_ERR_STATUS_LO, regSP1_UE_ERR_STATUS_HI), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SP1"}, + AMDGPU_GFX_SP_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regSQ_UE_ERR_STATUS_LO, regSQ_UE_ERR_STATUS_HI), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SQ"}, + AMDGPU_GFX_SQ_MEM, 8}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regSQC_UE_EDC_LO, regSQC_UE_EDC_HI), + 5, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SQC"}, + AMDGPU_GFX_SQC_MEM, 8}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCX_UE_ERR_STATUS_LO, regTCX_UE_ERR_STATUS_HI), + 2, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCX"}, + AMDGPU_GFX_TCX_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCC_UE_ERR_STATUS_LO, regTCC_UE_ERR_STATUS_HI), + 16, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCC"}, + AMDGPU_GFX_TCC_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTA_UE_EDC_LO, regTA_UE_EDC_HI), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TA"}, + AMDGPU_GFX_TA_MEM, 8}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCI_UE_EDC_LO_REG, regTCI_UE_EDC_HI_REG), + 31, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCI"}, + AMDGPU_GFX_TCI_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCP_UE_EDC_LO_REG, regTCP_UE_EDC_HI_REG), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCP"}, + AMDGPU_GFX_TCP_MEM, 8}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTD_UE_EDC_LO, regTD_UE_EDC_HI), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TD"}, + AMDGPU_GFX_TD_MEM, 8}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCA_UE_ERR_STATUS_LO, regTCA_UE_ERR_STATUS_HI), + 2, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCA"}, + AMDGPU_GFX_TCA_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regGCEA_UE_ERR_STATUS_LO, regGCEA_UE_ERR_STATUS_HI), + 16, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "GCEA"}, + AMDGPU_GFX_GCEA_MEM, 1}, + {{AMDGPU_RAS_REG_ENTRY(GC, 0, regLDS_UE_ERR_STATUS_LO, regLDS_UE_ERR_STATUS_HI), + 10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "LDS"}, + AMDGPU_GFX_LDS_MEM, 1}, +}; + +static const struct soc15_reg_entry gfx_v9_4_3_ea_err_status_regs = { + SOC15_REG_ENTRY(GC, 0, regGCEA_ERR_STATUS), 0, 1, 16 +}; + +static void gfx_v9_4_3_inst_query_ras_err_count(struct amdgpu_device *adev, + void *ras_error_status, int xcc_id) +{ + struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status; + unsigned long ce_count = 0, ue_count = 0; + uint32_t i, j, k; + + mutex_lock(&adev->grbm_idx_mutex); + + for (i = 0; i < ARRAY_SIZE(gfx_v9_4_3_ce_reg_list); i++) { + for (j = 0; j < gfx_v9_4_3_ce_reg_list[i].se_num; j++) { + for (k = 0; k < gfx_v9_4_3_ce_reg_list[i].reg_entry.reg_inst; k++) { + /* no need to select if instance number is 1 */ + if (gfx_v9_4_3_ce_reg_list[i].se_num > 1 || + gfx_v9_4_3_ce_reg_list[i].reg_entry.reg_inst > 1) + gfx_v9_4_3_xcc_select_se_sh(adev, j, 0, k, xcc_id); + + amdgpu_ras_inst_query_ras_error_count(adev, + &(gfx_v9_4_3_ce_reg_list[i].reg_entry), + 1, + gfx_v9_4_3_ras_mem_list_array[gfx_v9_4_3_ce_reg_list[i].mem_id_type].mem_id_ent, + gfx_v9_4_3_ras_mem_list_array[gfx_v9_4_3_ce_reg_list[i].mem_id_type].size, + GET_INST(GC, xcc_id), + AMDGPU_RAS_ERROR__SINGLE_CORRECTABLE, + &ce_count); + + amdgpu_ras_inst_query_ras_error_count(adev, + &(gfx_v9_4_3_ue_reg_list[i].reg_entry), + 1, + gfx_v9_4_3_ras_mem_list_array[gfx_v9_4_3_ue_reg_list[i].mem_id_type].mem_id_ent, + gfx_v9_4_3_ras_mem_list_array[gfx_v9_4_3_ue_reg_list[i].mem_id_type].size, + GET_INST(GC, xcc_id), + AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE, + &ue_count); + } + } + } + + gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, + xcc_id); + mutex_unlock(&adev->grbm_idx_mutex); + + /* the caller should make sure initialize value of + * err_data->ue_count and err_data->ce_count + */ + err_data->ce_count += ce_count; + err_data->ue_count += ue_count; +} + +static void gfx_v9_4_3_inst_reset_ras_err_count(struct amdgpu_device *adev, + void *ras_error_status, int xcc_id) +{ + uint32_t i, j, k; + + mutex_lock(&adev->grbm_idx_mutex); + + for (i = 0; i < ARRAY_SIZE(gfx_v9_4_3_ce_reg_list); i++) { + for (j = 0; j < gfx_v9_4_3_ce_reg_list[i].se_num; j++) { + for (k = 0; k < gfx_v9_4_3_ce_reg_list[i].reg_entry.reg_inst; k++) { + /* no need to select if instance number is 1 */ + if (gfx_v9_4_3_ce_reg_list[i].se_num > 1 || + gfx_v9_4_3_ce_reg_list[i].reg_entry.reg_inst > 1) + gfx_v9_4_3_xcc_select_se_sh(adev, j, 0, k, xcc_id); + + amdgpu_ras_inst_reset_ras_error_count(adev, + &(gfx_v9_4_3_ce_reg_list[i].reg_entry), + 1, + GET_INST(GC, xcc_id)); + + amdgpu_ras_inst_reset_ras_error_count(adev, + &(gfx_v9_4_3_ue_reg_list[i].reg_entry), + 1, + GET_INST(GC, xcc_id)); + } + } + } + + gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, + xcc_id); + mutex_unlock(&adev->grbm_idx_mutex); +} + +static void gfx_v9_4_3_inst_query_ea_err_status(struct amdgpu_device *adev, + int xcc_id) +{ + uint32_t i, j; + uint32_t reg_value; + + mutex_lock(&adev->grbm_idx_mutex); + + for (i = 0; i < gfx_v9_4_3_ea_err_status_regs.se_num; i++) { + for (j = 0; j < gfx_v9_4_3_ea_err_status_regs.instance; j++) { + gfx_v9_4_3_xcc_select_se_sh(adev, i, 0, j, xcc_id); + reg_value = RREG32_SOC15(GC, GET_INST(GC, xcc_id), + regGCEA_ERR_STATUS); + if (REG_GET_FIELD(reg_value, GCEA_ERR_STATUS, SDP_RDRSP_STATUS) || + REG_GET_FIELD(reg_value, GCEA_ERR_STATUS, SDP_WRRSP_STATUS) || + REG_GET_FIELD(reg_value, GCEA_ERR_STATUS, SDP_RDRSP_DATAPARITY_ERROR)) { + dev_warn(adev->dev, + "GCEA err detected at instance: %d, status: 0x%x!\n", + j, reg_value); + } + /* clear after read */ + reg_value = REG_SET_FIELD(reg_value, GCEA_ERR_STATUS, + CLEAR_ERROR_STATUS, 0x1); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regGCEA_ERR_STATUS, + reg_value); + } + } + + gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, + xcc_id); + mutex_unlock(&adev->grbm_idx_mutex); +} + +static void gfx_v9_4_3_inst_query_utc_err_status(struct amdgpu_device *adev, + int xcc_id) +{ + uint32_t data; + + data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regUTCL2_MEM_ECC_STATUS); + if (data) { + dev_warn(adev->dev, "GFX UTCL2 Mem Ecc Status: 0x%x!\n", data); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regUTCL2_MEM_ECC_STATUS, 0x3); + } + + data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regVML2_MEM_ECC_STATUS); + if (data) { + dev_warn(adev->dev, "GFX VML2 Mem Ecc Status: 0x%x!\n", data); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regVML2_MEM_ECC_STATUS, 0x3); + } + + data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), + regVML2_WALKER_MEM_ECC_STATUS); + if (data) { + dev_warn(adev->dev, "GFX VML2 Walker Mem Ecc Status: 0x%x!\n", data); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regVML2_WALKER_MEM_ECC_STATUS, + 0x3); + } +} + +static void gfx_v9_4_3_log_cu_timeout_status(struct amdgpu_device *adev, + uint32_t status, int xcc_id) +{ + struct amdgpu_cu_info *cu_info = &adev->gfx.cu_info; + uint32_t i, simd, wave; + uint32_t wave_status; + uint32_t wave_pc_lo, wave_pc_hi; + uint32_t wave_exec_lo, wave_exec_hi; + uint32_t wave_inst_dw0, wave_inst_dw1; + uint32_t wave_ib_sts; + + for (i = 0; i < 32; i++) { + if (!((i << 1) & status)) + continue; + + simd = i / cu_info->max_waves_per_simd; + wave = i % cu_info->max_waves_per_simd; + + wave_status = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_STATUS); + wave_pc_lo = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_PC_LO); + wave_pc_hi = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_PC_HI); + wave_exec_lo = + wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_EXEC_LO); + wave_exec_hi = + wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_EXEC_HI); + wave_inst_dw0 = + wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_INST_DW0); + wave_inst_dw1 = + wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_INST_DW1); + wave_ib_sts = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_IB_STS); + + dev_info( + adev->dev, + "\t SIMD %d, Wave %d: status 0x%x, pc 0x%llx, exec 0x%llx, inst 0x%llx, ib_sts 0x%x\n", + simd, wave, wave_status, + ((uint64_t)wave_pc_hi << 32 | wave_pc_lo), + ((uint64_t)wave_exec_hi << 32 | wave_exec_lo), + ((uint64_t)wave_inst_dw1 << 32 | wave_inst_dw0), + wave_ib_sts); + } +} + +static void gfx_v9_4_3_inst_query_sq_timeout_status(struct amdgpu_device *adev, + int xcc_id) +{ + uint32_t se_idx, sh_idx, cu_idx; + uint32_t status; + + mutex_lock(&adev->grbm_idx_mutex); + for (se_idx = 0; se_idx < adev->gfx.config.max_shader_engines; se_idx++) { + for (sh_idx = 0; sh_idx < adev->gfx.config.max_sh_per_se; sh_idx++) { + for (cu_idx = 0; cu_idx < adev->gfx.config.max_cu_per_sh; cu_idx++) { + gfx_v9_4_3_xcc_select_se_sh(adev, se_idx, sh_idx, + cu_idx, xcc_id); + status = RREG32_SOC15(GC, GET_INST(GC, xcc_id), + regSQ_TIMEOUT_STATUS); + if (status != 0) { + dev_info( + adev->dev, + "GFX Watchdog Timeout: SE %d, SH %d, CU %d\n", + se_idx, sh_idx, cu_idx); + gfx_v9_4_3_log_cu_timeout_status( + adev, status, xcc_id); + } + /* clear old status */ + WREG32_SOC15(GC, GET_INST(GC, xcc_id), + regSQ_TIMEOUT_STATUS, 0); + } + } + } + gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, + xcc_id); + mutex_unlock(&adev->grbm_idx_mutex); +} + +static void gfx_v9_4_3_inst_query_ras_err_status(struct amdgpu_device *adev, + void *ras_error_status, int xcc_id) +{ + gfx_v9_4_3_inst_query_ea_err_status(adev, xcc_id); + gfx_v9_4_3_inst_query_utc_err_status(adev, xcc_id); + gfx_v9_4_3_inst_query_sq_timeout_status(adev, xcc_id); +} + +static void gfx_v9_4_3_inst_reset_utc_err_status(struct amdgpu_device *adev, + int xcc_id) +{ + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regUTCL2_MEM_ECC_STATUS, 0x3); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regVML2_MEM_ECC_STATUS, 0x3); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regVML2_WALKER_MEM_ECC_STATUS, 0x3); +} + +static void gfx_v9_4_3_inst_reset_ea_err_status(struct amdgpu_device *adev, + int xcc_id) +{ + uint32_t i, j; + uint32_t value; + + mutex_lock(&adev->grbm_idx_mutex); + for (i = 0; i < gfx_v9_4_3_ea_err_status_regs.se_num; i++) { + for (j = 0; j < gfx_v9_4_3_ea_err_status_regs.instance; j++) { + gfx_v9_4_3_xcc_select_se_sh(adev, i, 0, j, xcc_id); + value = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regGCEA_ERR_STATUS); + value = REG_SET_FIELD(value, GCEA_ERR_STATUS, + CLEAR_ERROR_STATUS, 0x1); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), regGCEA_ERR_STATUS, value); + } + } + gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, + xcc_id); + mutex_unlock(&adev->grbm_idx_mutex); +} + +static void gfx_v9_4_3_inst_reset_sq_timeout_status(struct amdgpu_device *adev, + int xcc_id) +{ + uint32_t se_idx, sh_idx, cu_idx; + + mutex_lock(&adev->grbm_idx_mutex); + for (se_idx = 0; se_idx < adev->gfx.config.max_shader_engines; se_idx++) { + for (sh_idx = 0; sh_idx < adev->gfx.config.max_sh_per_se; sh_idx++) { + for (cu_idx = 0; cu_idx < adev->gfx.config.max_cu_per_sh; cu_idx++) { + gfx_v9_4_3_xcc_select_se_sh(adev, se_idx, sh_idx, + cu_idx, xcc_id); + WREG32_SOC15(GC, GET_INST(GC, xcc_id), + regSQ_TIMEOUT_STATUS, 0); + } + } + } + gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, + xcc_id); + mutex_unlock(&adev->grbm_idx_mutex); +} + +static void gfx_v9_4_3_inst_reset_ras_err_status(struct amdgpu_device *adev, + void *ras_error_status, int xcc_id) +{ + gfx_v9_4_3_inst_reset_utc_err_status(adev, xcc_id); + gfx_v9_4_3_inst_reset_ea_err_status(adev, xcc_id); + gfx_v9_4_3_inst_reset_sq_timeout_status(adev, xcc_id); +} + +static void gfx_v9_4_3_query_ras_error_count(struct amdgpu_device *adev, + void *ras_error_status) +{ + amdgpu_gfx_ras_error_func(adev, ras_error_status, + gfx_v9_4_3_inst_query_ras_err_count); +} + +static void gfx_v9_4_3_reset_ras_error_count(struct amdgpu_device *adev) +{ + amdgpu_gfx_ras_error_func(adev, NULL, gfx_v9_4_3_inst_reset_ras_err_count); +} + +static void gfx_v9_4_3_query_ras_error_status(struct amdgpu_device *adev) +{ + amdgpu_gfx_ras_error_func(adev, NULL, gfx_v9_4_3_inst_query_ras_err_status); +} + +static void gfx_v9_4_3_reset_ras_error_status(struct amdgpu_device *adev) +{ + amdgpu_gfx_ras_error_func(adev, NULL, gfx_v9_4_3_inst_reset_ras_err_status); +} + +static const struct amd_ip_funcs gfx_v9_4_3_ip_funcs = { + .name = "gfx_v9_4_3", + .early_init = gfx_v9_4_3_early_init, + .late_init = gfx_v9_4_3_late_init, + .sw_init = gfx_v9_4_3_sw_init, + .sw_fini = gfx_v9_4_3_sw_fini, + .hw_init = gfx_v9_4_3_hw_init, + .hw_fini = gfx_v9_4_3_hw_fini, + .suspend = gfx_v9_4_3_suspend, + .resume = gfx_v9_4_3_resume, + .is_idle = gfx_v9_4_3_is_idle, + .wait_for_idle = gfx_v9_4_3_wait_for_idle, + .soft_reset = gfx_v9_4_3_soft_reset, + .set_clockgating_state = gfx_v9_4_3_set_clockgating_state, + .set_powergating_state = gfx_v9_4_3_set_powergating_state, + .get_clockgating_state = gfx_v9_4_3_get_clockgating_state, +}; + +static const struct amdgpu_ring_funcs gfx_v9_4_3_ring_funcs_compute = { + .type = AMDGPU_RING_TYPE_COMPUTE, + .align_mask = 0xff, + .nop = PACKET3(PACKET3_NOP, 0x3FFF), + .support_64bit_ptrs = true, + .get_rptr = gfx_v9_4_3_ring_get_rptr_compute, + .get_wptr = gfx_v9_4_3_ring_get_wptr_compute, + .set_wptr = gfx_v9_4_3_ring_set_wptr_compute, + .emit_frame_size = + 20 + /* gfx_v9_4_3_ring_emit_gds_switch */ + 7 + /* gfx_v9_4_3_ring_emit_hdp_flush */ + 5 + /* hdp invalidate */ + 7 + /* gfx_v9_4_3_ring_emit_pipeline_sync */ + SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + + SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 + + 2 + /* gfx_v9_4_3_ring_emit_vm_flush */ + 8 + 8 + 8 + /* gfx_v9_4_3_ring_emit_fence x3 for user fence, vm fence */ + 7 + /* gfx_v9_4_3_emit_mem_sync */ + 5 + /* gfx_v9_4_3_emit_wave_limit for updating regSPI_WCL_PIPE_PERCENT_GFX register */ + 15, /* for updating 3 regSPI_WCL_PIPE_PERCENT_CS registers */ + .emit_ib_size = 7, /* gfx_v9_4_3_ring_emit_ib_compute */ + .emit_ib = gfx_v9_4_3_ring_emit_ib_compute, + .emit_fence = gfx_v9_4_3_ring_emit_fence, + .emit_pipeline_sync = gfx_v9_4_3_ring_emit_pipeline_sync, + .emit_vm_flush = gfx_v9_4_3_ring_emit_vm_flush, + .emit_gds_switch = gfx_v9_4_3_ring_emit_gds_switch, + .emit_hdp_flush = gfx_v9_4_3_ring_emit_hdp_flush, + .test_ring = gfx_v9_4_3_ring_test_ring, + .test_ib = gfx_v9_4_3_ring_test_ib, + .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, + .emit_wreg = gfx_v9_4_3_ring_emit_wreg, + .emit_reg_wait = gfx_v9_4_3_ring_emit_reg_wait, + .emit_reg_write_reg_wait = gfx_v9_4_3_ring_emit_reg_write_reg_wait, + .emit_mem_sync = gfx_v9_4_3_emit_mem_sync, + .emit_wave_limit = gfx_v9_4_3_emit_wave_limit, +}; + +static const struct amdgpu_ring_funcs gfx_v9_4_3_ring_funcs_kiq = { + .type = AMDGPU_RING_TYPE_KIQ, + .align_mask = 0xff, + .nop = PACKET3(PACKET3_NOP, 0x3FFF), + .support_64bit_ptrs = true, + .get_rptr = gfx_v9_4_3_ring_get_rptr_compute, + .get_wptr = gfx_v9_4_3_ring_get_wptr_compute, + .set_wptr = gfx_v9_4_3_ring_set_wptr_compute, + .emit_frame_size = + 20 + /* gfx_v9_4_3_ring_emit_gds_switch */ + 7 + /* gfx_v9_4_3_ring_emit_hdp_flush */ + 5 + /* hdp invalidate */ + 7 + /* gfx_v9_4_3_ring_emit_pipeline_sync */ + SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + + SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 + + 2 + /* gfx_v9_4_3_ring_emit_vm_flush */ + 8 + 8 + 8, /* gfx_v9_4_3_ring_emit_fence_kiq x3 for user fence, vm fence */ + .emit_ib_size = 7, /* gfx_v9_4_3_ring_emit_ib_compute */ + .emit_fence = gfx_v9_4_3_ring_emit_fence_kiq, + .test_ring = gfx_v9_4_3_ring_test_ring, + .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, + .emit_rreg = gfx_v9_4_3_ring_emit_rreg, + .emit_wreg = gfx_v9_4_3_ring_emit_wreg, + .emit_reg_wait = gfx_v9_4_3_ring_emit_reg_wait, + .emit_reg_write_reg_wait = gfx_v9_4_3_ring_emit_reg_write_reg_wait, +}; + +static void gfx_v9_4_3_set_ring_funcs(struct amdgpu_device *adev) +{ + int i, j, num_xcc; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + for (i = 0; i < num_xcc; i++) { + adev->gfx.kiq[i].ring.funcs = &gfx_v9_4_3_ring_funcs_kiq; + + for (j = 0; j < adev->gfx.num_compute_rings; j++) + adev->gfx.compute_ring[j + i * adev->gfx.num_compute_rings].funcs + = &gfx_v9_4_3_ring_funcs_compute; + } +} + +static const struct amdgpu_irq_src_funcs gfx_v9_4_3_eop_irq_funcs = { + .set = gfx_v9_4_3_set_eop_interrupt_state, + .process = gfx_v9_4_3_eop_irq, +}; + +static const struct amdgpu_irq_src_funcs gfx_v9_4_3_priv_reg_irq_funcs = { + .set = gfx_v9_4_3_set_priv_reg_fault_state, + .process = gfx_v9_4_3_priv_reg_irq, +}; + +static const struct amdgpu_irq_src_funcs gfx_v9_4_3_priv_inst_irq_funcs = { + .set = gfx_v9_4_3_set_priv_inst_fault_state, + .process = gfx_v9_4_3_priv_inst_irq, +}; + +static void gfx_v9_4_3_set_irq_funcs(struct amdgpu_device *adev) +{ + adev->gfx.eop_irq.num_types = AMDGPU_CP_IRQ_LAST; + adev->gfx.eop_irq.funcs = &gfx_v9_4_3_eop_irq_funcs; + + adev->gfx.priv_reg_irq.num_types = 1; + adev->gfx.priv_reg_irq.funcs = &gfx_v9_4_3_priv_reg_irq_funcs; + + adev->gfx.priv_inst_irq.num_types = 1; + adev->gfx.priv_inst_irq.funcs = &gfx_v9_4_3_priv_inst_irq_funcs; +} + +static void gfx_v9_4_3_set_rlc_funcs(struct amdgpu_device *adev) +{ + adev->gfx.rlc.funcs = &gfx_v9_4_3_rlc_funcs; +} + + +static void gfx_v9_4_3_set_gds_init(struct amdgpu_device *adev) +{ + /* init asci gds info */ + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 4, 3): + /* 9.4.3 removed all the GDS internal memory, + * only support GWS opcode in kernel, like barrier + * semaphore.etc */ + adev->gds.gds_size = 0; + break; + default: + adev->gds.gds_size = 0x10000; + break; + } + + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 4, 3): + /* deprecated for 9.4.3, no usage at all */ + adev->gds.gds_compute_max_wave_id = 0; + break; + default: + /* this really depends on the chip */ + adev->gds.gds_compute_max_wave_id = 0x7ff; + break; + } + + adev->gds.gws_size = 64; + adev->gds.oa_size = 16; +} + +static void gfx_v9_4_3_set_user_cu_inactive_bitmap(struct amdgpu_device *adev, + u32 bitmap) +{ + u32 data; + + if (!bitmap) + return; + + data = bitmap << GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT; + data &= GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK; + + WREG32_SOC15(GC, GET_INST(GC, 0), regGC_USER_SHADER_ARRAY_CONFIG, data); +} + +static u32 gfx_v9_4_3_get_cu_active_bitmap(struct amdgpu_device *adev) +{ + u32 data, mask; + + data = RREG32_SOC15(GC, GET_INST(GC, 0), regCC_GC_SHADER_ARRAY_CONFIG); + data |= RREG32_SOC15(GC, GET_INST(GC, 0), regGC_USER_SHADER_ARRAY_CONFIG); + + data &= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK; + data >>= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT; + + mask = amdgpu_gfx_create_bitmask(adev->gfx.config.max_cu_per_sh); + + return (~data) & mask; +} + +static int gfx_v9_4_3_get_cu_info(struct amdgpu_device *adev, + struct amdgpu_cu_info *cu_info) +{ + int i, j, k, counter, active_cu_number = 0; + u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0; + unsigned disable_masks[4 * 4]; + + if (!adev || !cu_info) + return -EINVAL; + + /* + * 16 comes from bitmap array size 4*4, and it can cover all gfx9 ASICs + */ + if (adev->gfx.config.max_shader_engines * + adev->gfx.config.max_sh_per_se > 16) + return -EINVAL; + + amdgpu_gfx_parse_disable_cu(disable_masks, + adev->gfx.config.max_shader_engines, + adev->gfx.config.max_sh_per_se); + + mutex_lock(&adev->grbm_idx_mutex); + for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { + for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { + mask = 1; + ao_bitmap = 0; + counter = 0; + gfx_v9_4_3_xcc_select_se_sh(adev, i, j, 0xffffffff, 0); + gfx_v9_4_3_set_user_cu_inactive_bitmap( + adev, disable_masks[i * adev->gfx.config.max_sh_per_se + j]); + bitmap = gfx_v9_4_3_get_cu_active_bitmap(adev); + + /* + * The bitmap(and ao_cu_bitmap) in cu_info structure is + * 4x4 size array, and it's usually suitable for Vega + * ASICs which has 4*2 SE/SH layout. + * But for Arcturus, SE/SH layout is changed to 8*1. + * To mostly reduce the impact, we make it compatible + * with current bitmap array as below: + * SE4,SH0 --> bitmap[0][1] + * SE5,SH0 --> bitmap[1][1] + * SE6,SH0 --> bitmap[2][1] + * SE7,SH0 --> bitmap[3][1] + */ + cu_info->bitmap[i % 4][j + i / 4] = bitmap; + + for (k = 0; k < adev->gfx.config.max_cu_per_sh; k++) { + if (bitmap & mask) { + if (counter < adev->gfx.config.max_cu_per_sh) + ao_bitmap |= mask; + counter++; + } + mask <<= 1; + } + active_cu_number += counter; + if (i < 2 && j < 2) + ao_cu_mask |= (ao_bitmap << (i * 16 + j * 8)); + cu_info->ao_cu_bitmap[i % 4][j + i / 4] = ao_bitmap; + } + } + gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, + 0); + mutex_unlock(&adev->grbm_idx_mutex); + + cu_info->number = active_cu_number; + cu_info->ao_cu_mask = ao_cu_mask; + cu_info->simd_per_cu = NUM_SIMD_PER_CU; + + return 0; +} + +const struct amdgpu_ip_block_version gfx_v9_4_3_ip_block = { + .type = AMD_IP_BLOCK_TYPE_GFX, + .major = 9, + .minor = 4, + .rev = 0, + .funcs = &gfx_v9_4_3_ip_funcs, +}; + +static int gfx_v9_4_3_xcp_resume(void *handle, uint32_t inst_mask) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + uint32_t tmp_mask; + int i, r; + + /* TODO : Initialize golden regs */ + /* gfx_v9_4_3_init_golden_registers(adev); */ + + tmp_mask = inst_mask; + for_each_inst(i, tmp_mask) + gfx_v9_4_3_xcc_constants_init(adev, i); + + if (!amdgpu_sriov_vf(adev)) { + tmp_mask = inst_mask; + for_each_inst(i, tmp_mask) { + r = gfx_v9_4_3_xcc_rlc_resume(adev, i); + if (r) + return r; + } + } + + tmp_mask = inst_mask; + for_each_inst(i, tmp_mask) { + r = gfx_v9_4_3_xcc_cp_resume(adev, i); + if (r) + return r; + } + + return 0; +} + +static int gfx_v9_4_3_xcp_suspend(void *handle, uint32_t inst_mask) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int i; + + for_each_inst(i, inst_mask) + gfx_v9_4_3_xcc_fini(adev, i); + + return 0; +} + +struct amdgpu_xcp_ip_funcs gfx_v9_4_3_xcp_funcs = { + .suspend = &gfx_v9_4_3_xcp_suspend, + .resume = &gfx_v9_4_3_xcp_resume +}; + +struct amdgpu_ras_block_hw_ops gfx_v9_4_3_ras_ops = { + .query_ras_error_count = &gfx_v9_4_3_query_ras_error_count, + .reset_ras_error_count = &gfx_v9_4_3_reset_ras_error_count, + .query_ras_error_status = &gfx_v9_4_3_query_ras_error_status, + .reset_ras_error_status = &gfx_v9_4_3_reset_ras_error_status, +}; + +struct amdgpu_gfx_ras gfx_v9_4_3_ras = { + .ras_block = { + .hw_ops = &gfx_v9_4_3_ras_ops, + }, +}; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.h b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.h index 84e69701b81a..42d67ee0e7ef 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.h +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.h @@ -24,7 +24,8 @@ #ifndef __GFX_V9_4_3_H__ #define __GFX_V9_4_3_H__ -extern const struct amdgpu_gfx_funcs gfx_v9_4_3_gfx_funcs; -extern const struct amdgpu_rlc_funcs gfx_v9_4_3_rlc_funcs; +extern const struct amdgpu_ip_block_version gfx_v9_4_3_ip_block; + +extern struct amdgpu_xcp_ip_funcs gfx_v9_4_3_xcp_funcs; #endif /* __GFX_V9_4_3_H__ */ diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c index ab2325f6c7ac..d94cc1ec7242 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c @@ -40,7 +40,7 @@ static void gfxhub_v1_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; WREG32_SOC15_OFFSET(GC, 0, mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, hub->ctx_addr_distance * vmid, @@ -247,7 +247,7 @@ static void gfxhub_v1_0_disable_identity_aperture(struct amdgpu_device *adev) static void gfxhub_v1_0_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; unsigned num_level, block_size; uint32_t tmp; int i; @@ -307,7 +307,7 @@ static void gfxhub_v1_0_setup_vmid_config(struct amdgpu_device *adev) static void gfxhub_v1_0_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; unsigned i; for (i = 0 ; i < 18; ++i) { @@ -338,7 +338,7 @@ static int gfxhub_v1_0_gart_enable(struct amdgpu_device *adev) static void gfxhub_v1_0_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; u32 tmp; u32 i; @@ -411,7 +411,7 @@ static void gfxhub_v1_0_set_fault_enable_default(struct amdgpu_device *adev, static void gfxhub_v1_0_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(GC, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c index c59c6c85fbff..4dabf910334b 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c @@ -21,6 +21,7 @@ * */ #include "amdgpu.h" +#include "amdgpu_xcp.h" #include "gfxhub_v1_2.h" #include "gfxhub_v1_1.h" @@ -35,227 +36,288 @@ static u64 gfxhub_v1_2_get_mc_fb_offset(struct amdgpu_device *adev) { - return (u64)RREG32_SOC15(GC, 0, regMC_VM_FB_OFFSET) << 24; + return (u64)RREG32_SOC15(GC, GET_INST(GC, 0), regMC_VM_FB_OFFSET) << 24; +} + +static void gfxhub_v1_2_xcc_setup_vm_pt_regs(struct amdgpu_device *adev, + uint32_t vmid, + uint64_t page_table_base, + uint32_t xcc_mask) +{ + struct amdgpu_vmhub *hub; + int i; + + for_each_inst(i, xcc_mask) { + hub = &adev->vmhub[AMDGPU_GFXHUB(i)]; + WREG32_SOC15_OFFSET(GC, GET_INST(GC, i), + regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, + hub->ctx_addr_distance * vmid, + lower_32_bits(page_table_base)); + + WREG32_SOC15_OFFSET(GC, GET_INST(GC, i), + regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32, + hub->ctx_addr_distance * vmid, + upper_32_bits(page_table_base)); + } } static void gfxhub_v1_2_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; - - WREG32_SOC15_OFFSET(GC, 0, regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, - hub->ctx_addr_distance * vmid, - lower_32_bits(page_table_base)); + uint32_t xcc_mask; - WREG32_SOC15_OFFSET(GC, 0, regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32, - hub->ctx_addr_distance * vmid, - upper_32_bits(page_table_base)); + xcc_mask = GENMASK(NUM_XCC(adev->gfx.xcc_mask) - 1, 0); + gfxhub_v1_2_xcc_setup_vm_pt_regs(adev, vmid, page_table_base, xcc_mask); } -static void gfxhub_v1_2_init_gart_aperture_regs(struct amdgpu_device *adev) +static void gfxhub_v1_2_xcc_init_gart_aperture_regs(struct amdgpu_device *adev, + uint32_t xcc_mask) { uint64_t pt_base; + int i; if (adev->gmc.pdb0_bo) pt_base = amdgpu_gmc_pd_addr(adev->gmc.pdb0_bo); else pt_base = amdgpu_gmc_pd_addr(adev->gart.bo); - gfxhub_v1_2_setup_vm_pt_regs(adev, 0, pt_base); + gfxhub_v1_2_xcc_setup_vm_pt_regs(adev, 0, pt_base, xcc_mask); /* If use GART for FB translation, vmid0 page table covers both * vram and system memory (gart) */ - if (adev->gmc.pdb0_bo) { - WREG32_SOC15(GC, 0, regVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32, - (u32)(adev->gmc.fb_start >> 12)); - WREG32_SOC15(GC, 0, regVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32, - (u32)(adev->gmc.fb_start >> 44)); - - WREG32_SOC15(GC, 0, regVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32, - (u32)(adev->gmc.gart_end >> 12)); - WREG32_SOC15(GC, 0, regVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32, - (u32)(adev->gmc.gart_end >> 44)); - } else { - WREG32_SOC15(GC, 0, regVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32, - (u32)(adev->gmc.gart_start >> 12)); - WREG32_SOC15(GC, 0, regVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32, - (u32)(adev->gmc.gart_start >> 44)); - - WREG32_SOC15(GC, 0, regVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32, - (u32)(adev->gmc.gart_end >> 12)); - WREG32_SOC15(GC, 0, regVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32, - (u32)(adev->gmc.gart_end >> 44)); + for_each_inst(i, xcc_mask) { + if (adev->gmc.pdb0_bo) { + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32, + (u32)(adev->gmc.fb_start >> 12)); + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32, + (u32)(adev->gmc.fb_start >> 44)); + + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32, + (u32)(adev->gmc.gart_end >> 12)); + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32, + (u32)(adev->gmc.gart_end >> 44)); + } else { + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32, + (u32)(adev->gmc.gart_start >> 12)); + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32, + (u32)(adev->gmc.gart_start >> 44)); + + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32, + (u32)(adev->gmc.gart_end >> 12)); + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32, + (u32)(adev->gmc.gart_end >> 44)); + } } } -static void gfxhub_v1_2_init_system_aperture_regs(struct amdgpu_device *adev) +static void +gfxhub_v1_2_xcc_init_system_aperture_regs(struct amdgpu_device *adev, + uint32_t xcc_mask) { uint64_t value; uint32_t tmp; + int i; - /* Program the AGP BAR */ - WREG32_SOC15_RLC(GC, 0, regMC_VM_AGP_BASE, 0); - WREG32_SOC15_RLC(GC, 0, regMC_VM_AGP_BOT, adev->gmc.agp_start >> 24); - WREG32_SOC15_RLC(GC, 0, regMC_VM_AGP_TOP, adev->gmc.agp_end >> 24); - - if (!amdgpu_sriov_vf(adev) || adev->asic_type <= CHIP_VEGA10) { - /* Program the system aperture low logical page number. */ - WREG32_SOC15_RLC(GC, 0, regMC_VM_SYSTEM_APERTURE_LOW_ADDR, - min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18); - - if (adev->apu_flags & AMD_APU_IS_RAVEN2) - /* - * Raven2 has a HW issue that it is unable to use the - * vram which is out of MC_VM_SYSTEM_APERTURE_HIGH_ADDR. - * So here is the workaround that increase system - * aperture high address (add 1) to get rid of the VM - * fault and hardware hang. - */ - WREG32_SOC15_RLC(GC, 0, - regMC_VM_SYSTEM_APERTURE_HIGH_ADDR, - max((adev->gmc.fb_end >> 18) + 0x1, - adev->gmc.agp_end >> 18)); - else - WREG32_SOC15_RLC(GC, 0, - regMC_VM_SYSTEM_APERTURE_HIGH_ADDR, - max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18); - - /* Set default page address. */ - value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr); - WREG32_SOC15(GC, 0, regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB, - (u32)(value >> 12)); - WREG32_SOC15(GC, 0, regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB, - (u32)(value >> 44)); - - /* Program "protection fault". */ - WREG32_SOC15(GC, 0, regVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32, - (u32)(adev->dummy_page_addr >> 12)); - WREG32_SOC15(GC, 0, regVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32, - (u32)((u64)adev->dummy_page_addr >> 44)); - - tmp = RREG32_SOC15(GC, 0, regVM_L2_PROTECTION_FAULT_CNTL2); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL2, - ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1); - WREG32_SOC15(GC, 0, regVM_L2_PROTECTION_FAULT_CNTL2, tmp); - } - - /* In the case squeezing vram into GART aperture, we don't use - * FB aperture and AGP aperture. Disable them. - */ - if (adev->gmc.pdb0_bo) { - WREG32_SOC15(GC, 0, regMC_VM_FB_LOCATION_TOP, 0); - WREG32_SOC15(GC, 0, regMC_VM_FB_LOCATION_BASE, 0x00FFFFFF); - WREG32_SOC15(GC, 0, regMC_VM_AGP_TOP, 0); - WREG32_SOC15(GC, 0, regMC_VM_AGP_BOT, 0xFFFFFF); - WREG32_SOC15(GC, 0, regMC_VM_SYSTEM_APERTURE_LOW_ADDR, 0x3FFFFFFF); - WREG32_SOC15(GC, 0, regMC_VM_SYSTEM_APERTURE_HIGH_ADDR, 0); + for_each_inst(i, xcc_mask) { + /* Program the AGP BAR */ + WREG32_SOC15_RLC(GC, GET_INST(GC, i), regMC_VM_AGP_BASE, 0); + WREG32_SOC15_RLC(GC, GET_INST(GC, i), regMC_VM_AGP_BOT, adev->gmc.agp_start >> 24); + WREG32_SOC15_RLC(GC, GET_INST(GC, i), regMC_VM_AGP_TOP, adev->gmc.agp_end >> 24); + + if (!amdgpu_sriov_vf(adev) || adev->asic_type <= CHIP_VEGA10) { + /* Program the system aperture low logical page number. */ + WREG32_SOC15_RLC(GC, GET_INST(GC, i), regMC_VM_SYSTEM_APERTURE_LOW_ADDR, + min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18); + + if (adev->apu_flags & AMD_APU_IS_RAVEN2) + /* + * Raven2 has a HW issue that it is unable to use the + * vram which is out of MC_VM_SYSTEM_APERTURE_HIGH_ADDR. + * So here is the workaround that increase system + * aperture high address (add 1) to get rid of the VM + * fault and hardware hang. + */ + WREG32_SOC15_RLC(GC, GET_INST(GC, i), + regMC_VM_SYSTEM_APERTURE_HIGH_ADDR, + max((adev->gmc.fb_end >> 18) + 0x1, + adev->gmc.agp_end >> 18)); + else + WREG32_SOC15_RLC(GC, GET_INST(GC, i), + regMC_VM_SYSTEM_APERTURE_HIGH_ADDR, + max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18); + + /* Set default page address. */ + value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr); + WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB, + (u32)(value >> 12)); + WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB, + (u32)(value >> 44)); + + /* Program "protection fault". */ + WREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32, + (u32)(adev->dummy_page_addr >> 12)); + WREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32, + (u32)((u64)adev->dummy_page_addr >> 44)); + + tmp = RREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_CNTL2); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL2, + ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1); + WREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_CNTL2, tmp); + } + + /* In the case squeezing vram into GART aperture, we don't use + * FB aperture and AGP aperture. Disable them. + */ + if (adev->gmc.pdb0_bo) { + WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_FB_LOCATION_TOP, 0); + WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_FB_LOCATION_BASE, 0x00FFFFFF); + WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_AGP_TOP, 0); + WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_AGP_BOT, 0xFFFFFF); + WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_SYSTEM_APERTURE_LOW_ADDR, 0x3FFFFFFF); + WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_SYSTEM_APERTURE_HIGH_ADDR, 0); + } } } -static void gfxhub_v1_2_init_tlb_regs(struct amdgpu_device *adev) +static void gfxhub_v1_2_xcc_init_tlb_regs(struct amdgpu_device *adev, + uint32_t xcc_mask) { uint32_t tmp; + int i; - /* Setup TLB control */ - tmp = RREG32_SOC15(GC, 0, regMC_VM_MX_L1_TLB_CNTL); - - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, - ENABLE_ADVANCED_DRIVER_MODEL, 1); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, - SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, - MTYPE, MTYPE_UC);/* XXX for emulation. */ - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1); - - WREG32_SOC15_RLC(GC, 0, regMC_VM_MX_L1_TLB_CNTL, tmp); + for_each_inst(i, xcc_mask) { + /* Setup TLB control */ + tmp = RREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_MX_L1_TLB_CNTL); + + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + ENABLE_L1_TLB, 1); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + SYSTEM_ACCESS_MODE, 3); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + ENABLE_ADVANCED_DRIVER_MODEL, 1); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + MTYPE, MTYPE_UC);/* XXX for emulation. */ + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1); + + WREG32_SOC15_RLC(GC, GET_INST(GC, i), regMC_VM_MX_L1_TLB_CNTL, tmp); + } } -static void gfxhub_v1_2_init_cache_regs(struct amdgpu_device *adev) +static void gfxhub_v1_2_xcc_init_cache_regs(struct amdgpu_device *adev, + uint32_t xcc_mask) { uint32_t tmp; + int i; - /* Setup L2 cache */ - tmp = RREG32_SOC15(GC, 0, regVM_L2_CNTL); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 1); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 1); - /* XXX for emulation, Refer to closed source code.*/ - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, L2_PDE0_CACHE_TAG_GENERATION_MODE, - 0); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0); - WREG32_SOC15_RLC(GC, 0, regVM_L2_CNTL, tmp); - - tmp = RREG32_SOC15(GC, 0, regVM_L2_CNTL2); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_L2_CACHE, 1); - WREG32_SOC15_RLC(GC, 0, regVM_L2_CNTL2, tmp); - - tmp = regVM_L2_CNTL3_DEFAULT; - if (adev->gmc.translate_further) { - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 12); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, - L2_CACHE_BIGK_FRAGMENT_SIZE, 9); - } else { - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 9); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, - L2_CACHE_BIGK_FRAGMENT_SIZE, 6); + for_each_inst(i, xcc_mask) { + /* Setup L2 cache */ + tmp = RREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_CNTL); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 1); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 1); + /* XXX for emulation, Refer to closed source code.*/ + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, L2_PDE0_CACHE_TAG_GENERATION_MODE, + 0); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0); + WREG32_SOC15_RLC(GC, GET_INST(GC, i), regVM_L2_CNTL, tmp); + + tmp = RREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_CNTL2); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_L2_CACHE, 1); + WREG32_SOC15_RLC(GC, GET_INST(GC, i), regVM_L2_CNTL2, tmp); + + tmp = regVM_L2_CNTL3_DEFAULT; + if (adev->gmc.translate_further) { + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 12); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, + L2_CACHE_BIGK_FRAGMENT_SIZE, 9); + } else { + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 9); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, + L2_CACHE_BIGK_FRAGMENT_SIZE, 6); + } + WREG32_SOC15_RLC(GC, GET_INST(GC, i), regVM_L2_CNTL3, tmp); + + tmp = regVM_L2_CNTL4_DEFAULT; + /* For AMD APP APUs setup WC memory */ + if (adev->gmc.xgmi.connected_to_cpu || adev->gmc.is_app_apu) { + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 1); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 1); + } else { + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0); + } + WREG32_SOC15_RLC(GC, GET_INST(GC, i), regVM_L2_CNTL4, tmp); } - WREG32_SOC15_RLC(GC, 0, regVM_L2_CNTL3, tmp); - - tmp = regVM_L2_CNTL4_DEFAULT; - if (adev->gmc.xgmi.connected_to_cpu) { - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 1); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 1); - } else { - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0); - } - WREG32_SOC15_RLC(GC, 0, regVM_L2_CNTL4, tmp); } -static void gfxhub_v1_2_enable_system_domain(struct amdgpu_device *adev) +static void gfxhub_v1_2_xcc_enable_system_domain(struct amdgpu_device *adev, + uint32_t xcc_mask) { uint32_t tmp; + int i; - tmp = RREG32_SOC15(GC, 0, regVM_CONTEXT0_CNTL); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, - adev->gmc.vmid0_page_table_depth); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, PAGE_TABLE_BLOCK_SIZE, - adev->gmc.vmid0_page_table_block_size); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, - RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0); - WREG32_SOC15(GC, 0, regVM_CONTEXT0_CNTL, tmp); + for_each_inst(i, xcc_mask) { + tmp = RREG32_SOC15(GC, GET_INST(GC, i), regVM_CONTEXT0_CNTL); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, + adev->gmc.vmid0_page_table_depth); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, PAGE_TABLE_BLOCK_SIZE, + adev->gmc.vmid0_page_table_block_size); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, + RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0); + WREG32_SOC15(GC, GET_INST(GC, i), regVM_CONTEXT0_CNTL, tmp); + } } -static void gfxhub_v1_2_disable_identity_aperture(struct amdgpu_device *adev) +static void +gfxhub_v1_2_xcc_disable_identity_aperture(struct amdgpu_device *adev, + uint32_t xcc_mask) { - WREG32_SOC15(GC, 0, regVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32, - 0XFFFFFFFF); - WREG32_SOC15(GC, 0, regVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32, - 0x0000000F); - - WREG32_SOC15(GC, 0, regVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32, - 0); - WREG32_SOC15(GC, 0, regVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32, - 0); - - WREG32_SOC15(GC, 0, regVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32, 0); - WREG32_SOC15(GC, 0, regVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32, 0); + int i; + for_each_inst(i, xcc_mask) { + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32, + 0XFFFFFFFF); + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32, + 0x0000000F); + + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32, + 0); + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32, + 0); + + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32, 0); + WREG32_SOC15(GC, GET_INST(GC, i), + regVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32, 0); + } } -static void gfxhub_v1_2_setup_vmid_config(struct amdgpu_device *adev) +static void gfxhub_v1_2_xcc_setup_vmid_config(struct amdgpu_device *adev, + uint32_t xcc_mask) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub; unsigned num_level, block_size; uint32_t tmp; - int i; + int i, j; num_level = adev->vm_manager.num_level; block_size = adev->vm_manager.block_size; @@ -264,124 +326,205 @@ static void gfxhub_v1_2_setup_vmid_config(struct amdgpu_device *adev) else block_size -= 9; - for (i = 0; i <= 14; i++) { - tmp = RREG32_SOC15_OFFSET(GC, 0, regVM_CONTEXT1_CNTL, i); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH, - num_level); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, - 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - PAGE_TABLE_BLOCK_SIZE, - block_size); - /* Send no-retry XNACK on fault to suppress VM fault storm. - * On Aldebaran, XNACK can be enabled in the SQ per-process. - * Retry faults need to be enabled for that to work. - */ - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, - !adev->gmc.noretry || - adev->asic_type == CHIP_ALDEBARAN); - WREG32_SOC15_OFFSET(GC, 0, regVM_CONTEXT1_CNTL, - i * hub->ctx_distance, tmp); - WREG32_SOC15_OFFSET(GC, 0, - regVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32, - i * hub->ctx_addr_distance, 0); - WREG32_SOC15_OFFSET(GC, 0, - regVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32, - i * hub->ctx_addr_distance, 0); - WREG32_SOC15_OFFSET(GC, 0, - regVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32, - i * hub->ctx_addr_distance, - lower_32_bits(adev->vm_manager.max_pfn - 1)); - WREG32_SOC15_OFFSET(GC, 0, - regVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32, - i * hub->ctx_addr_distance, - upper_32_bits(adev->vm_manager.max_pfn - 1)); + for_each_inst(j, xcc_mask) { + hub = &adev->vmhub[AMDGPU_GFXHUB(j)]; + for (i = 0; i <= 14; i++) { + tmp = RREG32_SOC15_OFFSET(GC, GET_INST(GC, j), regVM_CONTEXT1_CNTL, i); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH, + num_level); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, + 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + PAGE_TABLE_BLOCK_SIZE, + block_size); + /* Send no-retry XNACK on fault to suppress VM fault storm. + * On 9.4.2 and 9.4.3, XNACK can be enabled in + * the SQ per-process. + * Retry faults need to be enabled for that to work. + */ + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, + !adev->gmc.noretry || + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2) || + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3)); + WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), regVM_CONTEXT1_CNTL, + i * hub->ctx_distance, tmp); + WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), + regVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32, + i * hub->ctx_addr_distance, 0); + WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), + regVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32, + i * hub->ctx_addr_distance, 0); + WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), + regVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32, + i * hub->ctx_addr_distance, + lower_32_bits(adev->vm_manager.max_pfn - 1)); + WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), + regVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32, + i * hub->ctx_addr_distance, + upper_32_bits(adev->vm_manager.max_pfn - 1)); + } } } -static void gfxhub_v1_2_program_invalidation(struct amdgpu_device *adev) +static void gfxhub_v1_2_xcc_program_invalidation(struct amdgpu_device *adev, + uint32_t xcc_mask) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; - unsigned i; - - for (i = 0 ; i < 18; ++i) { - WREG32_SOC15_OFFSET(GC, 0, regVM_INVALIDATE_ENG0_ADDR_RANGE_LO32, - i * hub->eng_addr_distance, 0xffffffff); - WREG32_SOC15_OFFSET(GC, 0, regVM_INVALIDATE_ENG0_ADDR_RANGE_HI32, - i * hub->eng_addr_distance, 0x1f); + struct amdgpu_vmhub *hub; + unsigned int i, j; + + for_each_inst(j, xcc_mask) { + hub = &adev->vmhub[AMDGPU_GFXHUB(j)]; + + for (i = 0 ; i < 18; ++i) { + WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), regVM_INVALIDATE_ENG0_ADDR_RANGE_LO32, + i * hub->eng_addr_distance, 0xffffffff); + WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), regVM_INVALIDATE_ENG0_ADDR_RANGE_HI32, + i * hub->eng_addr_distance, 0x1f); + } } } -static int gfxhub_v1_2_gart_enable(struct amdgpu_device *adev) +static int gfxhub_v1_2_xcc_gart_enable(struct amdgpu_device *adev, + uint32_t xcc_mask) { - if (amdgpu_sriov_vf(adev) && adev->asic_type != CHIP_ARCTURUS) { - /* - * MC_VM_FB_LOCATION_BASE/TOP is NULL for VF, becuase they are - * VF copy registers so vbios post doesn't program them, for - * SRIOV driver need to program them - */ - WREG32_SOC15_RLC(GC, 0, regMC_VM_FB_LOCATION_BASE, - adev->gmc.vram_start >> 24); - WREG32_SOC15_RLC(GC, 0, regMC_VM_FB_LOCATION_TOP, - adev->gmc.vram_end >> 24); + uint32_t tmp_mask; + int i; + + tmp_mask = xcc_mask; + /* + * MC_VM_FB_LOCATION_BASE/TOP is NULL for VF, because they are + * VF copy registers so vbios post doesn't program them, for + * SRIOV driver need to program them + */ + if (amdgpu_sriov_vf(adev)) { + for_each_inst(i, tmp_mask) { + i = ffs(tmp_mask) - 1; + WREG32_SOC15_RLC(GC, GET_INST(GC, i), regMC_VM_FB_LOCATION_BASE, + adev->gmc.vram_start >> 24); + WREG32_SOC15_RLC(GC, GET_INST(GC, i), regMC_VM_FB_LOCATION_TOP, + adev->gmc.vram_end >> 24); + } } /* GART Enable. */ - gfxhub_v1_2_init_gart_aperture_regs(adev); - gfxhub_v1_2_init_system_aperture_regs(adev); - gfxhub_v1_2_init_tlb_regs(adev); + gfxhub_v1_2_xcc_init_gart_aperture_regs(adev, xcc_mask); + gfxhub_v1_2_xcc_init_system_aperture_regs(adev, xcc_mask); + gfxhub_v1_2_xcc_init_tlb_regs(adev, xcc_mask); if (!amdgpu_sriov_vf(adev)) - gfxhub_v1_2_init_cache_regs(adev); + gfxhub_v1_2_xcc_init_cache_regs(adev, xcc_mask); - gfxhub_v1_2_enable_system_domain(adev); + gfxhub_v1_2_xcc_enable_system_domain(adev, xcc_mask); if (!amdgpu_sriov_vf(adev)) - gfxhub_v1_2_disable_identity_aperture(adev); - gfxhub_v1_2_setup_vmid_config(adev); - gfxhub_v1_2_program_invalidation(adev); + gfxhub_v1_2_xcc_disable_identity_aperture(adev, xcc_mask); + gfxhub_v1_2_xcc_setup_vmid_config(adev, xcc_mask); + gfxhub_v1_2_xcc_program_invalidation(adev, xcc_mask); return 0; } +static int gfxhub_v1_2_gart_enable(struct amdgpu_device *adev) +{ + uint32_t xcc_mask; + + xcc_mask = GENMASK(NUM_XCC(adev->gfx.xcc_mask) - 1, 0); + return gfxhub_v1_2_xcc_gart_enable(adev, xcc_mask); +} + +static void gfxhub_v1_2_xcc_gart_disable(struct amdgpu_device *adev, + uint32_t xcc_mask) +{ + struct amdgpu_vmhub *hub; + u32 tmp; + u32 i, j; + + for_each_inst(j, xcc_mask) { + hub = &adev->vmhub[AMDGPU_GFXHUB(j)]; + /* Disable all tables */ + for (i = 0; i < 16; i++) + WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), regVM_CONTEXT0_CNTL, + i * hub->ctx_distance, 0); + + /* Setup TLB control */ + tmp = RREG32_SOC15(GC, GET_INST(GC, j), regMC_VM_MX_L1_TLB_CNTL); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0); + tmp = REG_SET_FIELD(tmp, + MC_VM_MX_L1_TLB_CNTL, + ENABLE_ADVANCED_DRIVER_MODEL, + 0); + WREG32_SOC15_RLC(GC, GET_INST(GC, j), regMC_VM_MX_L1_TLB_CNTL, tmp); + + /* Setup L2 cache */ + tmp = RREG32_SOC15(GC, GET_INST(GC, j), regVM_L2_CNTL); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 0); + WREG32_SOC15(GC, GET_INST(GC, j), regVM_L2_CNTL, tmp); + WREG32_SOC15(GC, GET_INST(GC, j), regVM_L2_CNTL3, 0); + } +} + static void gfxhub_v1_2_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + uint32_t xcc_mask; + + xcc_mask = GENMASK(NUM_XCC(adev->gfx.xcc_mask) - 1, 0); + gfxhub_v1_2_xcc_gart_disable(adev, xcc_mask); +} + +static void gfxhub_v1_2_xcc_set_fault_enable_default(struct amdgpu_device *adev, + bool value, + uint32_t xcc_mask) +{ u32 tmp; - u32 i; - - /* Disable all tables */ - for (i = 0; i < 16; i++) - WREG32_SOC15_OFFSET(GC, 0, regVM_CONTEXT0_CNTL, - i * hub->ctx_distance, 0); - - /* Setup TLB control */ - tmp = RREG32_SOC15(GC, 0, regMC_VM_MX_L1_TLB_CNTL); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0); - tmp = REG_SET_FIELD(tmp, - MC_VM_MX_L1_TLB_CNTL, - ENABLE_ADVANCED_DRIVER_MODEL, - 0); - WREG32_SOC15_RLC(GC, 0, regMC_VM_MX_L1_TLB_CNTL, tmp); - - /* Setup L2 cache */ - tmp = RREG32_SOC15(GC, 0, regVM_L2_CNTL); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 0); - WREG32_SOC15(GC, 0, regVM_L2_CNTL, tmp); - WREG32_SOC15(GC, 0, regVM_L2_CNTL3, 0); + int i; + + for_each_inst(i, xcc_mask) { + tmp = RREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_CNTL); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, + VM_L2_PROTECTION_FAULT_CNTL, + TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT, + value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + READ_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value); + if (!value) { + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + CRASH_ON_NO_RETRY_FAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + CRASH_ON_RETRY_FAULT, 1); + } + WREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_CNTL, tmp); + } } /** @@ -393,72 +536,100 @@ static void gfxhub_v1_2_gart_disable(struct amdgpu_device *adev) static void gfxhub_v1_2_set_fault_enable_default(struct amdgpu_device *adev, bool value) { - u32 tmp; - tmp = RREG32_SOC15(GC, 0, regVM_L2_PROTECTION_FAULT_CNTL); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, - VM_L2_PROTECTION_FAULT_CNTL, - TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT, - value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - READ_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value); - if (!value) { - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - CRASH_ON_NO_RETRY_FAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - CRASH_ON_RETRY_FAULT, 1); - } - WREG32_SOC15(GC, 0, regVM_L2_PROTECTION_FAULT_CNTL, tmp); + uint32_t xcc_mask; + + xcc_mask = GENMASK(NUM_XCC(adev->gfx.xcc_mask) - 1, 0); + gfxhub_v1_2_xcc_set_fault_enable_default(adev, value, xcc_mask); } -static void gfxhub_v1_2_init(struct amdgpu_device *adev) +static void gfxhub_v1_2_xcc_init(struct amdgpu_device *adev, uint32_t xcc_mask) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub; + int i; + + for_each_inst(i, xcc_mask) { + hub = &adev->vmhub[AMDGPU_GFXHUB(i)]; - hub->ctx0_ptb_addr_lo32 = - SOC15_REG_OFFSET(GC, 0, + hub->ctx0_ptb_addr_lo32 = + SOC15_REG_OFFSET(GC, GET_INST(GC, i), regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32); - hub->ctx0_ptb_addr_hi32 = - SOC15_REG_OFFSET(GC, 0, + hub->ctx0_ptb_addr_hi32 = + SOC15_REG_OFFSET(GC, GET_INST(GC, i), regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32); - hub->vm_inv_eng0_sem = - SOC15_REG_OFFSET(GC, 0, regVM_INVALIDATE_ENG0_SEM); - hub->vm_inv_eng0_req = - SOC15_REG_OFFSET(GC, 0, regVM_INVALIDATE_ENG0_REQ); - hub->vm_inv_eng0_ack = - SOC15_REG_OFFSET(GC, 0, regVM_INVALIDATE_ENG0_ACK); - hub->vm_context0_cntl = - SOC15_REG_OFFSET(GC, 0, regVM_CONTEXT0_CNTL); - hub->vm_l2_pro_fault_status = - SOC15_REG_OFFSET(GC, 0, regVM_L2_PROTECTION_FAULT_STATUS); - hub->vm_l2_pro_fault_cntl = - SOC15_REG_OFFSET(GC, 0, regVM_L2_PROTECTION_FAULT_CNTL); - - hub->ctx_distance = regVM_CONTEXT1_CNTL - regVM_CONTEXT0_CNTL; - hub->ctx_addr_distance = regVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 - - regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32; - hub->eng_distance = regVM_INVALIDATE_ENG1_REQ - regVM_INVALIDATE_ENG0_REQ; - hub->eng_addr_distance = regVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 - - regVM_INVALIDATE_ENG0_ADDR_RANGE_LO32; + hub->vm_inv_eng0_sem = + SOC15_REG_OFFSET(GC, GET_INST(GC, i), regVM_INVALIDATE_ENG0_SEM); + hub->vm_inv_eng0_req = + SOC15_REG_OFFSET(GC, GET_INST(GC, i), regVM_INVALIDATE_ENG0_REQ); + hub->vm_inv_eng0_ack = + SOC15_REG_OFFSET(GC, GET_INST(GC, i), regVM_INVALIDATE_ENG0_ACK); + hub->vm_context0_cntl = + SOC15_REG_OFFSET(GC, GET_INST(GC, i), regVM_CONTEXT0_CNTL); + hub->vm_l2_pro_fault_status = + SOC15_REG_OFFSET(GC, GET_INST(GC, i), + regVM_L2_PROTECTION_FAULT_STATUS); + hub->vm_l2_pro_fault_cntl = + SOC15_REG_OFFSET(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_CNTL); + + hub->ctx_distance = regVM_CONTEXT1_CNTL - + regVM_CONTEXT0_CNTL; + hub->ctx_addr_distance = + regVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 - + regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32; + hub->eng_distance = regVM_INVALIDATE_ENG1_REQ - + regVM_INVALIDATE_ENG0_REQ; + hub->eng_addr_distance = + regVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 - + regVM_INVALIDATE_ENG0_ADDR_RANGE_LO32; + } } +static void gfxhub_v1_2_init(struct amdgpu_device *adev) +{ + uint32_t xcc_mask; + + xcc_mask = GENMASK(NUM_XCC(adev->gfx.xcc_mask) - 1, 0); + gfxhub_v1_2_xcc_init(adev, xcc_mask); +} + +static int gfxhub_v1_2_get_xgmi_info(struct amdgpu_device *adev) +{ + u32 max_num_physical_nodes; + u32 max_physical_node_id; + u32 xgmi_lfb_cntl; + u32 max_region; + u64 seg_size; + + xgmi_lfb_cntl = RREG32_SOC15(GC, GET_INST(GC, 0), regMC_VM_XGMI_LFB_CNTL); + seg_size = REG_GET_FIELD( + RREG32_SOC15(GC, GET_INST(GC, 0), regMC_VM_XGMI_LFB_SIZE), + MC_VM_XGMI_LFB_SIZE, PF_LFB_SIZE) << 24; + max_region = + REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL, PF_MAX_REGION); + + + + max_num_physical_nodes = 8; + max_physical_node_id = 7; + + /* PF_MAX_REGION=0 means xgmi is disabled */ + if (max_region || adev->gmc.xgmi.connected_to_cpu) { + adev->gmc.xgmi.num_physical_nodes = max_region + 1; + + if (adev->gmc.xgmi.num_physical_nodes > max_num_physical_nodes) + return -EINVAL; + + adev->gmc.xgmi.physical_node_id = + REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL, + PF_LFB_REGION); + + if (adev->gmc.xgmi.physical_node_id > max_physical_node_id) + return -EINVAL; + + adev->gmc.xgmi.node_segment_size = seg_size; + } + + return 0; +} const struct amdgpu_gfxhub_funcs gfxhub_v1_2_funcs = { .get_mc_fb_offset = gfxhub_v1_2_get_mc_fb_offset, @@ -467,5 +638,38 @@ const struct amdgpu_gfxhub_funcs gfxhub_v1_2_funcs = { .gart_disable = gfxhub_v1_2_gart_disable, .set_fault_enable_default = gfxhub_v1_2_set_fault_enable_default, .init = gfxhub_v1_2_init, - .get_xgmi_info = gfxhub_v1_1_get_xgmi_info, + .get_xgmi_info = gfxhub_v1_2_get_xgmi_info, +}; + +static int gfxhub_v1_2_xcp_resume(void *handle, uint32_t inst_mask) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + bool value; + + if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS) + value = false; + else + value = true; + + gfxhub_v1_2_xcc_set_fault_enable_default(adev, value, inst_mask); + + if (!amdgpu_sriov_vf(adev)) + return gfxhub_v1_2_xcc_gart_enable(adev, inst_mask); + + return 0; +} + +static int gfxhub_v1_2_xcp_suspend(void *handle, uint32_t inst_mask) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + if (!amdgpu_sriov_vf(adev)) + gfxhub_v1_2_xcc_gart_disable(adev, inst_mask); + + return 0; +} + +struct amdgpu_xcp_ip_funcs gfxhub_v1_2_xcp_funcs = { + .suspend = &gfxhub_v1_2_xcp_suspend, + .resume = &gfxhub_v1_2_xcp_resume }; diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.h b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.h index e2d508f5a7ee..997e9f90c990 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.h +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.h @@ -26,4 +26,6 @@ extern const struct amdgpu_gfxhub_funcs gfxhub_v1_2_funcs; +extern struct amdgpu_xcp_ip_funcs gfxhub_v1_2_xcp_funcs; + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c index 9b3a02527318..f173a61c6c15 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c @@ -120,7 +120,7 @@ static u64 gfxhub_v2_0_get_mc_fb_offset(struct amdgpu_device *adev) static void gfxhub_v2_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, hub->ctx_addr_distance * vmid, @@ -282,7 +282,7 @@ static void gfxhub_v2_0_disable_identity_aperture(struct amdgpu_device *adev) static void gfxhub_v2_0_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; int i; uint32_t tmp; @@ -331,7 +331,7 @@ static void gfxhub_v2_0_setup_vmid_config(struct amdgpu_device *adev) static void gfxhub_v2_0_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; unsigned i; for (i = 0 ; i < 18; ++i) { @@ -360,7 +360,7 @@ static int gfxhub_v2_0_gart_enable(struct amdgpu_device *adev) static void gfxhub_v2_0_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; u32 tmp; u32 i; @@ -433,7 +433,7 @@ static const struct amdgpu_vmhub_funcs gfxhub_v2_0_vmhub_funcs = { static void gfxhub_v2_0_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(GC, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c index 4aacbbec31e2..d8fc3e8088cd 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c @@ -123,7 +123,7 @@ static u64 gfxhub_v2_1_get_mc_fb_offset(struct amdgpu_device *adev) static void gfxhub_v2_1_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, hub->ctx_addr_distance * vmid, @@ -291,7 +291,7 @@ static void gfxhub_v2_1_disable_identity_aperture(struct amdgpu_device *adev) static void gfxhub_v2_1_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; int i; uint32_t tmp; @@ -340,7 +340,7 @@ static void gfxhub_v2_1_setup_vmid_config(struct amdgpu_device *adev) static void gfxhub_v2_1_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; unsigned i; for (i = 0 ; i < 18; ++i) { @@ -381,7 +381,7 @@ static int gfxhub_v2_1_gart_enable(struct amdgpu_device *adev) static void gfxhub_v2_1_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; u32 tmp; u32 i; @@ -462,7 +462,7 @@ static const struct amdgpu_vmhub_funcs gfxhub_v2_1_vmhub_funcs = { static void gfxhub_v2_1_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(GC, 0, @@ -651,7 +651,7 @@ static void gfxhub_v2_1_restore_regs(struct amdgpu_device *adev) static void gfxhub_v2_1_halt(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; int i; uint32_t tmp; int time = 1000; diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c index 13712640fa46..c53147f9c9fc 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c @@ -119,7 +119,7 @@ static u64 gfxhub_v3_0_get_mc_fb_offset(struct amdgpu_device *adev) static void gfxhub_v3_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, hub->ctx_addr_distance * vmid, @@ -290,7 +290,7 @@ static void gfxhub_v3_0_disable_identity_aperture(struct amdgpu_device *adev) static void gfxhub_v3_0_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; int i; uint32_t tmp; @@ -339,7 +339,7 @@ static void gfxhub_v3_0_setup_vmid_config(struct amdgpu_device *adev) static void gfxhub_v3_0_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; unsigned i; for (i = 0 ; i < 18; ++i) { @@ -380,7 +380,7 @@ static int gfxhub_v3_0_gart_enable(struct amdgpu_device *adev) static void gfxhub_v3_0_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; u32 tmp; u32 i; @@ -463,7 +463,7 @@ static const struct amdgpu_vmhub_funcs gfxhub_v3_0_vmhub_funcs = { static void gfxhub_v3_0_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(GC, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.c index 6e0bd628c889..ae777487d72e 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.c @@ -122,7 +122,7 @@ static u64 gfxhub_v3_0_3_get_mc_fb_offset(struct amdgpu_device *adev) static void gfxhub_v3_0_3_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, hub->ctx_addr_distance * vmid, @@ -295,7 +295,7 @@ static void gfxhub_v3_0_3_disable_identity_aperture(struct amdgpu_device *adev) static void gfxhub_v3_0_3_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; int i; uint32_t tmp; @@ -344,7 +344,7 @@ static void gfxhub_v3_0_3_setup_vmid_config(struct amdgpu_device *adev) static void gfxhub_v3_0_3_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; unsigned i; for (i = 0 ; i < 18; ++i) { @@ -373,7 +373,7 @@ static int gfxhub_v3_0_3_gart_enable(struct amdgpu_device *adev) static void gfxhub_v3_0_3_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; u32 tmp; u32 i; @@ -451,7 +451,7 @@ static const struct amdgpu_vmhub_funcs gfxhub_v3_0_3_vmhub_funcs = { static void gfxhub_v3_0_3_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(GC, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c index b213dcf8ca06..0c8a47989576 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c @@ -76,7 +76,7 @@ gmc_v10_0_vm_fault_interrupt_state(struct amdgpu_device *adev, switch (state) { case AMDGPU_IRQ_STATE_DISABLE: /* MM HUB */ - amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB_0, false); + amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB0(0), false); /* GFX HUB */ /* This works because this interrupt is only * enabled at init/resume and disabled in @@ -84,11 +84,11 @@ gmc_v10_0_vm_fault_interrupt_state(struct amdgpu_device *adev, * change over the course of suspend/resume. */ if (!adev->in_s0ix) - amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB_0, false); + amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB(0), false); break; case AMDGPU_IRQ_STATE_ENABLE: /* MM HUB */ - amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB_0, true); + amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB0(0), true); /* GFX HUB */ /* This works because this interrupt is only * enabled at init/resume and disabled in @@ -96,7 +96,7 @@ gmc_v10_0_vm_fault_interrupt_state(struct amdgpu_device *adev, * change over the course of suspend/resume. */ if (!adev->in_s0ix) - amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB_0, true); + amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB(0), true); break; default: break; @@ -139,7 +139,7 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev, /* Try to handle the recoverable page faults by filling page * tables */ - if (amdgpu_vm_handle_fault(adev, entry->pasid, addr, write_fault)) + if (amdgpu_vm_handle_fault(adev, entry->pasid, 0, 0, addr, write_fault)) return 1; } @@ -149,7 +149,7 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev, * be updated to avoid reading an incorrect value due to * the new fast GRBM interface. */ - if ((entry->vmid_src == AMDGPU_GFXHUB_0) && + if ((entry->vmid_src == AMDGPU_GFXHUB(0)) && (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 3, 0))) RREG32(hub->vm_l2_pro_fault_status); @@ -212,8 +212,7 @@ static void gmc_v10_0_set_irq_funcs(struct amdgpu_device *adev) static bool gmc_v10_0_use_invalidate_semaphore(struct amdgpu_device *adev, uint32_t vmhub) { - return ((vmhub == AMDGPU_MMHUB_0 || - vmhub == AMDGPU_MMHUB_1) && + return ((vmhub == AMDGPU_MMHUB0(0)) && (!amdgpu_sriov_vf(adev))); } @@ -249,7 +248,7 @@ static void gmc_v10_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid, unsigned int i; unsigned char hub_ip = 0; - hub_ip = (vmhub == AMDGPU_GFXHUB_0) ? + hub_ip = (vmhub == AMDGPU_GFXHUB(0)) ? GC_HWIP : MMHUB_HWIP; spin_lock(&adev->gmc.invalidate_lock); @@ -284,7 +283,7 @@ static void gmc_v10_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid, * Issue a dummy read to wait for the ACK register to be cleared * to avoid a false ACK due to the new fast GRBM interface. */ - if ((vmhub == AMDGPU_GFXHUB_0) && + if ((vmhub == AMDGPU_GFXHUB(0)) && (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 3, 0))) RREG32_RLC_NO_KIQ(hub->vm_inv_eng0_req + hub->eng_distance * eng, hub_ip); @@ -343,7 +342,7 @@ static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, /* For SRIOV run time, driver shouldn't access the register through MMIO * Directly use kiq to do the vm invalidation instead */ - if (adev->gfx.kiq.ring.sched.ready && !adev->enable_mes && + if (adev->gfx.kiq[0].ring.sched.ready && !adev->enable_mes && (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev)) && down_read_trylock(&adev->reset_domain->sem)) { struct amdgpu_vmhub *hub = &adev->vmhub[vmhub]; @@ -361,19 +360,19 @@ static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, mutex_lock(&adev->mman.gtt_window_lock); - if (vmhub == AMDGPU_MMHUB_0) { - gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_MMHUB_0, 0); + if (vmhub == AMDGPU_MMHUB0(0)) { + gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_MMHUB0(0), 0); mutex_unlock(&adev->mman.gtt_window_lock); return; } - BUG_ON(vmhub != AMDGPU_GFXHUB_0); + BUG_ON(vmhub != AMDGPU_GFXHUB(0)); if (!adev->mman.buffer_funcs_enabled || !adev->ib_pool_ready || amdgpu_in_reset(adev) || ring->sched.ready == false) { - gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_GFXHUB_0, 0); + gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_GFXHUB(0), 0); mutex_unlock(&adev->mman.gtt_window_lock); return; } @@ -383,7 +382,7 @@ static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, * translation. Avoid this by doing the invalidation from the SDMA * itself. */ - r = amdgpu_job_alloc_with_ib(ring->adev, &adev->mman.entity, + r = amdgpu_job_alloc_with_ib(ring->adev, &adev->mman.high_pr, AMDGPU_FENCE_OWNER_UNDEFINED, 16 * 4, AMDGPU_IB_POOL_IMMEDIATE, &job); @@ -415,12 +414,13 @@ error_alloc: * @pasid: pasid to be flush * @flush_type: the flush type * @all_hub: Used with PACKET3_INVALIDATE_TLBS_ALL_HUB() + * @inst: is used to select which instance of KIQ to use for the invalidation * * Flush the TLB for the requested pasid. */ static int gmc_v10_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t pasid, uint32_t flush_type, - bool all_hub) + bool all_hub, uint32_t inst) { int vmid, i; signed long r; @@ -428,11 +428,11 @@ static int gmc_v10_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t queried_pasid; bool ret; u32 usec_timeout = amdgpu_sriov_vf(adev) ? SRIOV_USEC_TIMEOUT : adev->usec_timeout; - struct amdgpu_ring *ring = &adev->gfx.kiq.ring; - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_ring *ring = &adev->gfx.kiq[0].ring; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; if (amdgpu_emu_mode == 0 && ring->sched.ready) { - spin_lock(&adev->gfx.kiq.ring_lock); + spin_lock(&adev->gfx.kiq[0].ring_lock); /* 2 dwords flush + 8 dwords fence */ amdgpu_ring_alloc(ring, kiq->pmf->invalidate_tlbs_size + 8); kiq->pmf->kiq_invalidate_tlbs(ring, @@ -440,12 +440,12 @@ static int gmc_v10_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT); if (r) { amdgpu_ring_undo(ring); - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&adev->gfx.kiq[0].ring_lock); return -ETIME; } amdgpu_ring_commit(ring); - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&adev->gfx.kiq[0].ring_lock); r = amdgpu_fence_wait_polling(ring, seq, usec_timeout); if (r < 1) { dev_err(adev->dev, "wait for kiq fence error: %ld.\n", r); @@ -461,12 +461,12 @@ static int gmc_v10_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, &queried_pasid); if (ret && queried_pasid == pasid) { if (all_hub) { - for (i = 0; i < adev->num_vmhubs; i++) + for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) gmc_v10_0_flush_gpu_tlb(adev, vmid, i, flush_type); } else { gmc_v10_0_flush_gpu_tlb(adev, vmid, - AMDGPU_GFXHUB_0, flush_type); + AMDGPU_GFXHUB(0), flush_type); } if (!adev->enable_mes) break; @@ -534,7 +534,7 @@ static void gmc_v10_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid if (ring->is_mes_queue) return; - if (ring->vm_hub == AMDGPU_GFXHUB_0) + if (ring->vm_hub == AMDGPU_GFXHUB(0)) reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT) + vmid; else reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT_MM) + vmid; @@ -929,7 +929,8 @@ static int gmc_v10_0_sw_init(void *handle) case IP_VERSION(10, 3, 6): case IP_VERSION(10, 3, 3): case IP_VERSION(10, 3, 7): - adev->num_vmhubs = 2; + set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask); + set_bit(AMDGPU_MMHUB0(0), adev->vmhubs_mask); /* * To fulfill 4-level page support, * vm size is 256TB (48bit), maximum size of Navi10/Navi14/Navi12, @@ -1075,9 +1076,9 @@ static int gmc_v10_0_gart_enable(struct amdgpu_device *adev) if (!adev->in_s0ix) adev->gfxhub.funcs->set_fault_enable_default(adev, value); adev->mmhub.funcs->set_fault_enable_default(adev, value); - gmc_v10_0_flush_gpu_tlb(adev, 0, AMDGPU_MMHUB_0, 0); + gmc_v10_0_flush_gpu_tlb(adev, 0, AMDGPU_MMHUB0(0), 0); if (!adev->in_s0ix) - gmc_v10_0_flush_gpu_tlb(adev, 0, AMDGPU_GFXHUB_0, 0); + gmc_v10_0_flush_gpu_tlb(adev, 0, AMDGPU_GFXHUB(0), 0); DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n", (unsigned)(adev->gmc.gart_size >> 20), diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c index d95f9fe8f1c5..c571f0d95994 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c @@ -31,6 +31,8 @@ #include "umc_v8_10.h" #include "athub/athub_3_0_0_sh_mask.h" #include "athub/athub_3_0_0_offset.h" +#include "dcn/dcn_3_2_0_offset.h" +#include "dcn/dcn_3_2_0_sh_mask.h" #include "oss/osssys_6_0_0_offset.h" #include "ivsrcid/vmc/irqsrcs_vmc_1_0.h" #include "navi10_enum.h" @@ -62,7 +64,7 @@ gmc_v11_0_vm_fault_interrupt_state(struct amdgpu_device *adev, switch (state) { case AMDGPU_IRQ_STATE_DISABLE: /* MM HUB */ - amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB_0, false); + amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB0(0), false); /* GFX HUB */ /* This works because this interrupt is only * enabled at init/resume and disabled in @@ -70,11 +72,11 @@ gmc_v11_0_vm_fault_interrupt_state(struct amdgpu_device *adev, * change over the course of suspend/resume. */ if (!adev->in_s0ix) - amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB_0, false); + amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB(0), false); break; case AMDGPU_IRQ_STATE_ENABLE: /* MM HUB */ - amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB_0, true); + amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB0(0), true); /* GFX HUB */ /* This works because this interrupt is only * enabled at init/resume and disabled in @@ -82,7 +84,7 @@ gmc_v11_0_vm_fault_interrupt_state(struct amdgpu_device *adev, * change over the course of suspend/resume. */ if (!adev->in_s0ix) - amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB_0, true); + amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB(0), true); break; default: break; @@ -108,7 +110,7 @@ static int gmc_v11_0_process_interrupt(struct amdgpu_device *adev, * be updated to avoid reading an incorrect value due to * the new fast GRBM interface. */ - if (entry->vmid_src == AMDGPU_GFXHUB_0) + if (entry->vmid_src == AMDGPU_GFXHUB(0)) RREG32(hub->vm_l2_pro_fault_status); status = RREG32(hub->vm_l2_pro_fault_status); @@ -168,7 +170,7 @@ static void gmc_v11_0_set_irq_funcs(struct amdgpu_device *adev) static bool gmc_v11_0_use_invalidate_semaphore(struct amdgpu_device *adev, uint32_t vmhub) { - return ((vmhub == AMDGPU_MMHUB_0) && + return ((vmhub == AMDGPU_MMHUB0(0)) && (!amdgpu_sriov_vf(adev))); } @@ -200,7 +202,7 @@ static void gmc_v11_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid, unsigned int i; unsigned char hub_ip = 0; - hub_ip = (vmhub == AMDGPU_GFXHUB_0) ? + hub_ip = (vmhub == AMDGPU_GFXHUB(0)) ? GC_HWIP : MMHUB_HWIP; spin_lock(&adev->gmc.invalidate_lock); @@ -249,7 +251,7 @@ static void gmc_v11_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid, hub->eng_distance * eng, 0, hub_ip); /* Issue additional private vm invalidation to MMHUB */ - if ((vmhub != AMDGPU_GFXHUB_0) && + if ((vmhub != AMDGPU_GFXHUB(0)) && (hub->vm_l2_bank_select_reserved_cid2) && !amdgpu_sriov_vf(adev)) { inv_req = RREG32_NO_KIQ(hub->vm_l2_bank_select_reserved_cid2); @@ -282,7 +284,7 @@ static void gmc_v11_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid, static void gmc_v11_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, uint32_t vmhub, uint32_t flush_type) { - if ((vmhub == AMDGPU_GFXHUB_0) && !adev->gfx.is_poweron) + if ((vmhub == AMDGPU_GFXHUB(0)) && !adev->gfx.is_poweron) return; /* flush hdp cache */ @@ -291,7 +293,7 @@ static void gmc_v11_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, /* For SRIOV run time, driver shouldn't access the register through MMIO * Directly use kiq to do the vm invalidation instead */ - if ((adev->gfx.kiq.ring.sched.ready || adev->mes.ring.sched.ready) && + if ((adev->gfx.kiq[0].ring.sched.ready || adev->mes.ring.sched.ready) && (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev))) { struct amdgpu_vmhub *hub = &adev->vmhub[vmhub]; const unsigned eng = 17; @@ -317,23 +319,24 @@ static void gmc_v11_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, * @pasid: pasid to be flush * @flush_type: the flush type * @all_hub: flush all hubs + * @inst: is used to select which instance of KIQ to use for the invalidation * * Flush the TLB for the requested pasid. */ static int gmc_v11_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t pasid, uint32_t flush_type, - bool all_hub) + bool all_hub, uint32_t inst) { int vmid, i; signed long r; uint32_t seq; uint16_t queried_pasid; bool ret; - struct amdgpu_ring *ring = &adev->gfx.kiq.ring; - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_ring *ring = &adev->gfx.kiq[0].ring; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; if (amdgpu_emu_mode == 0 && ring->sched.ready) { - spin_lock(&adev->gfx.kiq.ring_lock); + spin_lock(&adev->gfx.kiq[0].ring_lock); /* 2 dwords flush + 8 dwords fence */ amdgpu_ring_alloc(ring, kiq->pmf->invalidate_tlbs_size + 8); kiq->pmf->kiq_invalidate_tlbs(ring, @@ -341,12 +344,12 @@ static int gmc_v11_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT); if (r) { amdgpu_ring_undo(ring); - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&adev->gfx.kiq[0].ring_lock); return -ETIME; } amdgpu_ring_commit(ring); - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&adev->gfx.kiq[0].ring_lock); r = amdgpu_fence_wait_polling(ring, seq, adev->usec_timeout); if (r < 1) { dev_err(adev->dev, "wait for kiq fence error: %ld.\n", r); @@ -362,12 +365,12 @@ static int gmc_v11_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, &queried_pasid); if (ret && queried_pasid == pasid) { if (all_hub) { - for (i = 0; i < adev->num_vmhubs; i++) + for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) gmc_v11_0_flush_gpu_tlb(adev, vmid, i, flush_type); } else { gmc_v11_0_flush_gpu_tlb(adev, vmid, - AMDGPU_GFXHUB_0, flush_type); + AMDGPU_GFXHUB(0), flush_type); } } } @@ -433,7 +436,7 @@ static void gmc_v11_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid if (ring->is_mes_queue) return; - if (ring->vm_hub == AMDGPU_GFXHUB_0) + if (ring->vm_hub == AMDGPU_GFXHUB(0)) reg = SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT) + vmid; else reg = SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT_MM) + vmid; @@ -546,7 +549,24 @@ static void gmc_v11_0_get_vm_pte(struct amdgpu_device *adev, static unsigned gmc_v11_0_get_vbios_fb_size(struct amdgpu_device *adev) { - return 0; + u32 d1vga_control = RREG32_SOC15(DCE, 0, regD1VGA_CONTROL); + unsigned size; + + if (REG_GET_FIELD(d1vga_control, D1VGA_CONTROL, D1VGA_MODE_ENABLE)) { + size = AMDGPU_VBIOS_VGA_ALLOCATION; + } else { + u32 viewport; + u32 pitch; + + viewport = RREG32_SOC15(DCE, 0, regHUBP0_DCSURF_PRI_VIEWPORT_DIMENSION); + pitch = RREG32_SOC15(DCE, 0, regHUBPREQ0_DCSURF_SURFACE_PITCH); + size = (REG_GET_FIELD(viewport, + HUBP0_DCSURF_PRI_VIEWPORT_DIMENSION, PRI_VIEWPORT_HEIGHT) * + REG_GET_FIELD(pitch, HUBPREQ0_DCSURF_SURFACE_PITCH, PITCH) * + 4); + } + + return size; } static const struct amdgpu_gmc_funcs gmc_v11_0_gmc_funcs = { @@ -760,7 +780,8 @@ static int gmc_v11_0_sw_init(void *handle) case IP_VERSION(11, 0, 2): case IP_VERSION(11, 0, 3): case IP_VERSION(11, 0, 4): - adev->num_vmhubs = 2; + set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask); + set_bit(AMDGPU_MMHUB0(0), adev->vmhubs_mask); /* * To fulfill 4-level page support, * vm size is 256TB (48bit), maximum size, @@ -867,7 +888,7 @@ static int gmc_v11_0_sw_fini(void *handle) static void gmc_v11_0_init_golden_registers(struct amdgpu_device *adev) { if (amdgpu_sriov_vf(adev)) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; WREG32(hub->vm_contexts_disable, 0); return; @@ -902,7 +923,7 @@ static int gmc_v11_0_gart_enable(struct amdgpu_device *adev) false : true; adev->mmhub.funcs->set_fault_enable_default(adev, value); - gmc_v11_0_flush_gpu_tlb(adev, 0, AMDGPU_MMHUB_0, 0); + gmc_v11_0_flush_gpu_tlb(adev, 0, AMDGPU_MMHUB0(0), 0); DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n", (unsigned)(adev->gmc.gart_size >> 20), diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c index b7dad4e67813..aa754c95a0b3 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c @@ -808,7 +808,7 @@ static int gmc_v6_0_sw_init(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - adev->num_vmhubs = 1; + set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask); if (adev->flags & AMD_IS_APU) { adev->gmc.vram_type = AMDGPU_VRAM_TYPE_UNKNOWN; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c index 402960b0174e..acd2b407860f 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c @@ -419,12 +419,13 @@ static int gmc_v7_0_mc_init(struct amdgpu_device *adev) * @pasid: pasid to be flush * @flush_type: type of flush * @all_hub: flush all hubs + * @inst: is used to select which instance of KIQ to use for the invalidation * * Flush the TLB for the requested pasid. */ static int gmc_v7_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t pasid, uint32_t flush_type, - bool all_hub) + bool all_hub, uint32_t inst) { int vmid; unsigned int tmp; @@ -977,7 +978,7 @@ static int gmc_v7_0_sw_init(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - adev->num_vmhubs = 1; + set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask); if (adev->flags & AMD_IS_APU) { adev->gmc.vram_type = AMDGPU_VRAM_TYPE_UNKNOWN; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index 504c1b34dab7..85dead2a5702 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -617,12 +617,13 @@ static int gmc_v8_0_mc_init(struct amdgpu_device *adev) * @pasid: pasid to be flush * @flush_type: type of flush * @all_hub: flush all hubs + * @inst: is used to select which instance of KIQ to use for the invalidation * * Flush the TLB for the requested pasid. */ static int gmc_v8_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t pasid, uint32_t flush_type, - bool all_hub) + bool all_hub, uint32_t inst) { int vmid; unsigned int tmp; @@ -1093,7 +1094,7 @@ static int gmc_v8_0_sw_init(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - adev->num_vmhubs = 1; + set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask); if (adev->flags & AMD_IS_APU) { adev->gmc.vram_type = AMDGPU_VRAM_TYPE_UNKNOWN; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c index 2fe21cefd772..67e669e0141c 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c @@ -79,6 +79,7 @@ #define mmHUBP0_DCSURF_PRI_VIEWPORT_DIMENSION_DCN2 0x05ea #define mmHUBP0_DCSURF_PRI_VIEWPORT_DIMENSION_DCN2_BASE_IDX 2 +#define MAX_MEM_RANGES 8 static const char *gfxhub_client_ids[] = { "CB", @@ -481,7 +482,7 @@ static int gmc_v9_0_vm_fault_interrupt_state(struct amdgpu_device *adev, switch (state) { case AMDGPU_IRQ_STATE_DISABLE: - for (j = 0; j < adev->num_vmhubs; j++) { + for_each_set_bit(j, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) { hub = &adev->vmhub[j]; for (i = 0; i < 16; i++) { reg = hub->vm_context0_cntl + i; @@ -491,25 +492,25 @@ static int gmc_v9_0_vm_fault_interrupt_state(struct amdgpu_device *adev, * fini/suspend, so the overall state doesn't * change over the course of suspend/resume. */ - if (adev->in_s0ix && (j == AMDGPU_GFXHUB_0)) + if (adev->in_s0ix && (j == AMDGPU_GFXHUB(0))) continue; - if (j == AMDGPU_GFXHUB_0) - tmp = RREG32_SOC15_IP(GC, reg); - else + if (j >= AMDGPU_MMHUB0(0)) tmp = RREG32_SOC15_IP(MMHUB, reg); + else + tmp = RREG32_SOC15_IP(GC, reg); tmp &= ~bits; - if (j == AMDGPU_GFXHUB_0) - WREG32_SOC15_IP(GC, reg, tmp); - else + if (j >= AMDGPU_MMHUB0(0)) WREG32_SOC15_IP(MMHUB, reg, tmp); + else + WREG32_SOC15_IP(GC, reg, tmp); } } break; case AMDGPU_IRQ_STATE_ENABLE: - for (j = 0; j < adev->num_vmhubs; j++) { + for_each_set_bit(j, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) { hub = &adev->vmhub[j]; for (i = 0; i < 16; i++) { reg = hub->vm_context0_cntl + i; @@ -519,20 +520,20 @@ static int gmc_v9_0_vm_fault_interrupt_state(struct amdgpu_device *adev, * fini/suspend, so the overall state doesn't * change over the course of suspend/resume. */ - if (adev->in_s0ix && (j == AMDGPU_GFXHUB_0)) + if (adev->in_s0ix && (j == AMDGPU_GFXHUB(0))) continue; - if (j == AMDGPU_GFXHUB_0) - tmp = RREG32_SOC15_IP(GC, reg); - else + if (j >= AMDGPU_MMHUB0(0)) tmp = RREG32_SOC15_IP(MMHUB, reg); + else + tmp = RREG32_SOC15_IP(GC, reg); tmp |= bits; - if (j == AMDGPU_GFXHUB_0) - WREG32_SOC15_IP(GC, reg, tmp); - else + if (j >= AMDGPU_MMHUB0(0)) WREG32_SOC15_IP(MMHUB, reg, tmp); + else + WREG32_SOC15_IP(GC, reg, tmp); } } break; @@ -556,11 +557,31 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, const char *hub_name; u64 addr; uint32_t cam_index = 0; - int ret; + int ret, xcc_id = 0; + uint32_t node_id; + + node_id = entry->node_id; addr = (u64)entry->src_data[0] << 12; addr |= ((u64)entry->src_data[1] & 0xf) << 44; + if (entry->client_id == SOC15_IH_CLIENTID_VMC) { + hub_name = "mmhub0"; + hub = &adev->vmhub[AMDGPU_MMHUB0(node_id / 4)]; + } else if (entry->client_id == SOC15_IH_CLIENTID_VMC1) { + hub_name = "mmhub1"; + hub = &adev->vmhub[AMDGPU_MMHUB1(0)]; + } else { + hub_name = "gfxhub0"; + if (adev->gfx.funcs->ih_node_to_logical_xcc) { + xcc_id = adev->gfx.funcs->ih_node_to_logical_xcc(adev, + node_id); + if (xcc_id < 0) + xcc_id = 0; + } + hub = &adev->vmhub[xcc_id]; + } + if (retry_fault) { if (adev->irq.retry_cam_enabled) { /* Delegate it to a different ring if the hardware hasn't @@ -573,7 +594,8 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, cam_index = entry->src_data[2] & 0x3ff; - ret = amdgpu_vm_handle_fault(adev, entry->pasid, addr, write_fault); + ret = amdgpu_vm_handle_fault(adev, entry->pasid, entry->vmid, node_id, + addr, write_fault); WDOORBELL32(adev->irq.retry_cam_doorbell_index, cam_index); if (ret) return 1; @@ -595,7 +617,8 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, /* Try to handle the recoverable page faults by filling page * tables */ - if (amdgpu_vm_handle_fault(adev, entry->pasid, addr, write_fault)) + if (amdgpu_vm_handle_fault(adev, entry->pasid, entry->vmid, node_id, + addr, write_fault)) return 1; } } @@ -603,16 +626,6 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, if (!printk_ratelimit()) return 0; - if (entry->client_id == SOC15_IH_CLIENTID_VMC) { - hub_name = "mmhub0"; - hub = &adev->vmhub[AMDGPU_MMHUB_0]; - } else if (entry->client_id == SOC15_IH_CLIENTID_VMC1) { - hub_name = "mmhub1"; - hub = &adev->vmhub[AMDGPU_MMHUB_1]; - } else { - hub_name = "gfxhub0"; - hub = &adev->vmhub[AMDGPU_GFXHUB_0]; - } memset(&task_info, 0, sizeof(struct amdgpu_task_info)); amdgpu_vm_get_task_info(adev, entry->pasid, &task_info); @@ -628,6 +641,11 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, addr, entry->client_id, soc15_ih_clientid_name[entry->client_id]); + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3)) + dev_err(adev->dev, " cookie node_id %d fault from die %s%d%s\n", + node_id, node_id % 4 == 3 ? "RSV" : "AID", node_id / 4, + node_id % 4 == 1 ? ".XCD0" : node_id % 4 == 2 ? ".XCD1" : ""); + if (amdgpu_sriov_vf(adev)) return 0; @@ -636,7 +654,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, * be updated to avoid reading an incorrect value due to * the new fast GRBM interface. */ - if ((entry->vmid_src == AMDGPU_GFXHUB_0) && + if ((entry->vmid_src == AMDGPU_GFXHUB(0)) && (adev->ip_versions[GC_HWIP][0] < IP_VERSION(9, 4, 2))) RREG32(hub->vm_l2_pro_fault_status); @@ -645,11 +663,10 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, rw = REG_GET_FIELD(status, VM_L2_PROTECTION_FAULT_STATUS, RW); WREG32_P(hub->vm_l2_pro_fault_cntl, 1, ~1); - dev_err(adev->dev, "VM_L2_PROTECTION_FAULT_STATUS:0x%08X\n", status); - if (hub == &adev->vmhub[AMDGPU_GFXHUB_0]) { + if (entry->vmid_src == AMDGPU_GFXHUB(0)) { dev_err(adev->dev, "\t Faulty UTCL2 client ID: %s (0x%x)\n", cid >= ARRAY_SIZE(gfxhub_client_ids) ? "unknown" : gfxhub_client_ids[cid], @@ -759,8 +776,8 @@ static bool gmc_v9_0_use_invalidate_semaphore(struct amdgpu_device *adev, adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3)) return false; - return ((vmhub == AMDGPU_MMHUB_0 || - vmhub == AMDGPU_MMHUB_1) && + return ((vmhub == AMDGPU_MMHUB0(0) || + vmhub == AMDGPU_MMHUB1(0)) && (!amdgpu_sriov_vf(adev)) && (!(!(adev->apu_flags & AMD_APU_IS_RAVEN2) && (adev->apu_flags & AMD_APU_IS_PICASSO)))); @@ -803,7 +820,7 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, u32 j, inv_req, inv_req2, tmp; struct amdgpu_vmhub *hub; - BUG_ON(vmhub >= adev->num_vmhubs); + BUG_ON(vmhub >= AMDGPU_MAX_VMHUBS); hub = &adev->vmhub[vmhub]; if (adev->gmc.xgmi.num_physical_nodes && @@ -816,6 +833,11 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, */ inv_req = gmc_v9_0_get_invalidate_req(vmid, 2); inv_req2 = gmc_v9_0_get_invalidate_req(vmid, flush_type); + } else if (flush_type == 2 && + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3) && + adev->rev_id == 0) { + inv_req = gmc_v9_0_get_invalidate_req(vmid, 0); + inv_req2 = gmc_v9_0_get_invalidate_req(vmid, flush_type); } else { inv_req = gmc_v9_0_get_invalidate_req(vmid, flush_type); inv_req2 = 0; @@ -824,7 +846,7 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, /* This is necessary for a HW workaround under SRIOV as well * as GFXOFF under bare metal */ - if (adev->gfx.kiq.ring.sched.ready && + if (adev->gfx.kiq[0].ring.sched.ready && (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev)) && down_read_trylock(&adev->reset_domain->sem)) { uint32_t req = hub->vm_inv_eng0_req + hub->eng_distance * eng; @@ -849,11 +871,10 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, if (use_semaphore) { for (j = 0; j < adev->usec_timeout; j++) { /* a read return value of 1 means semaphore acquire */ - if (vmhub == AMDGPU_GFXHUB_0) - tmp = RREG32_SOC15_IP_NO_KIQ(GC, hub->vm_inv_eng0_sem + hub->eng_distance * eng); - else + if (vmhub >= AMDGPU_MMHUB0(0)) tmp = RREG32_SOC15_IP_NO_KIQ(MMHUB, hub->vm_inv_eng0_sem + hub->eng_distance * eng); - + else + tmp = RREG32_SOC15_IP_NO_KIQ(GC, hub->vm_inv_eng0_sem + hub->eng_distance * eng); if (tmp & 0x1) break; udelay(1); @@ -864,27 +885,26 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, } do { - if (vmhub == AMDGPU_GFXHUB_0) - WREG32_SOC15_IP_NO_KIQ(GC, hub->vm_inv_eng0_req + hub->eng_distance * eng, inv_req); - else + if (vmhub >= AMDGPU_MMHUB0(0)) WREG32_SOC15_IP_NO_KIQ(MMHUB, hub->vm_inv_eng0_req + hub->eng_distance * eng, inv_req); + else + WREG32_SOC15_IP_NO_KIQ(GC, hub->vm_inv_eng0_req + hub->eng_distance * eng, inv_req); /* * Issue a dummy read to wait for the ACK register to * be cleared to avoid a false ACK due to the new fast * GRBM interface. */ - if ((vmhub == AMDGPU_GFXHUB_0) && + if ((vmhub == AMDGPU_GFXHUB(0)) && (adev->ip_versions[GC_HWIP][0] < IP_VERSION(9, 4, 2))) RREG32_NO_KIQ(hub->vm_inv_eng0_req + hub->eng_distance * eng); for (j = 0; j < adev->usec_timeout; j++) { - if (vmhub == AMDGPU_GFXHUB_0) - tmp = RREG32_SOC15_IP_NO_KIQ(GC, hub->vm_inv_eng0_ack + hub->eng_distance * eng); - else + if (vmhub >= AMDGPU_MMHUB0(0)) tmp = RREG32_SOC15_IP_NO_KIQ(MMHUB, hub->vm_inv_eng0_ack + hub->eng_distance * eng); - + else + tmp = RREG32_SOC15_IP_NO_KIQ(GC, hub->vm_inv_eng0_ack + hub->eng_distance * eng); if (tmp & (1 << vmid)) break; udelay(1); @@ -900,10 +920,10 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, * add semaphore release after invalidation, * write with 0 means semaphore release */ - if (vmhub == AMDGPU_GFXHUB_0) - WREG32_SOC15_IP_NO_KIQ(GC, hub->vm_inv_eng0_sem + hub->eng_distance * eng, 0); - else + if (vmhub >= AMDGPU_MMHUB0(0)) WREG32_SOC15_IP_NO_KIQ(MMHUB, hub->vm_inv_eng0_sem + hub->eng_distance * eng, 0); + else + WREG32_SOC15_IP_NO_KIQ(GC, hub->vm_inv_eng0_sem + hub->eng_distance * eng, 0); } spin_unlock(&adev->gmc.invalidate_lock); @@ -921,12 +941,13 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid, * @pasid: pasid to be flush * @flush_type: the flush type * @all_hub: flush all hubs + * @inst: is used to select which instance of KIQ to use for the invalidation * * Flush the TLB for the requested pasid. */ static int gmc_v9_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t pasid, uint32_t flush_type, - bool all_hub) + bool all_hub, uint32_t inst) { int vmid, i; signed long r; @@ -934,8 +955,8 @@ static int gmc_v9_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t queried_pasid; bool ret; u32 usec_timeout = amdgpu_sriov_vf(adev) ? SRIOV_USEC_TIMEOUT : adev->usec_timeout; - struct amdgpu_ring *ring = &adev->gfx.kiq.ring; - struct amdgpu_kiq *kiq = &adev->gfx.kiq; + struct amdgpu_ring *ring = &adev->gfx.kiq[inst].ring; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[inst]; if (amdgpu_in_reset(adev)) return -EIO; @@ -955,24 +976,31 @@ static int gmc_v9_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, if (vega20_xgmi_wa) ndw += kiq->pmf->invalidate_tlbs_size; - spin_lock(&adev->gfx.kiq.ring_lock); + spin_lock(&adev->gfx.kiq[inst].ring_lock); /* 2 dwords flush + 8 dwords fence */ amdgpu_ring_alloc(ring, ndw); if (vega20_xgmi_wa) kiq->pmf->kiq_invalidate_tlbs(ring, pasid, 2, all_hub); + + if (flush_type == 2 && + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3) && + adev->rev_id == 0) + kiq->pmf->kiq_invalidate_tlbs(ring, + pasid, 0, all_hub); + kiq->pmf->kiq_invalidate_tlbs(ring, pasid, flush_type, all_hub); r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT); if (r) { amdgpu_ring_undo(ring); - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&adev->gfx.kiq[inst].ring_lock); up_read(&adev->reset_domain->sem); return -ETIME; } amdgpu_ring_commit(ring); - spin_unlock(&adev->gfx.kiq.ring_lock); + spin_unlock(&adev->gfx.kiq[inst].ring_lock); r = amdgpu_fence_wait_polling(ring, seq, usec_timeout); if (r < 1) { dev_err(adev->dev, "wait for kiq fence error: %ld.\n", r); @@ -989,12 +1017,12 @@ static int gmc_v9_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev, &queried_pasid); if (ret && queried_pasid == pasid) { if (all_hub) { - for (i = 0; i < adev->num_vmhubs; i++) + for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) gmc_v9_0_flush_gpu_tlb(adev, vmid, i, flush_type); } else { gmc_v9_0_flush_gpu_tlb(adev, vmid, - AMDGPU_GFXHUB_0, flush_type); + AMDGPU_GFXHUB(0), flush_type); } break; } @@ -1060,10 +1088,10 @@ static void gmc_v9_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid, uint32_t reg; /* Do nothing because there's no lut register for mmhub1. */ - if (ring->vm_hub == AMDGPU_MMHUB_1) + if (ring->vm_hub == AMDGPU_MMHUB1(0)) return; - if (ring->vm_hub == AMDGPU_GFXHUB_0) + if (ring->vm_hub == AMDGPU_GFXHUB(0)) reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT) + vmid; else reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT_MM) + vmid; @@ -1159,13 +1187,14 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev, bool is_vram = bo->tbo.resource->mem_type == TTM_PL_VRAM; bool coherent = bo->flags & AMDGPU_GEM_CREATE_COHERENT; bool uncached = bo->flags & AMDGPU_GEM_CREATE_UNCACHED; - unsigned int mtype; + struct amdgpu_vm *vm = mapping->bo_va->base.vm; + unsigned int mtype_local, mtype; bool snoop = false; + bool is_local; switch (adev->ip_versions[GC_HWIP][0]) { case IP_VERSION(9, 4, 1): case IP_VERSION(9, 4, 2): - case IP_VERSION(9, 4, 3): if (is_vram) { if (bo_adev == adev) { if (uncached) @@ -1200,6 +1229,43 @@ static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev, snoop = true; } break; + case IP_VERSION(9, 4, 3): + /* Only local VRAM BOs or system memory on non-NUMA APUs + * can be assumed to be local in their entirety. Choose + * MTYPE_NC as safe fallback for all system memory BOs on + * NUMA systems. Their MTYPE can be overridden per-page in + * gmc_v9_0_override_vm_pte_flags. + */ + mtype_local = MTYPE_RW; + if (amdgpu_mtype_local == 1) { + DRM_INFO_ONCE("Using MTYPE_NC for local memory\n"); + mtype_local = MTYPE_NC; + } else if (amdgpu_mtype_local == 2) { + DRM_INFO_ONCE("Using MTYPE_CC for local memory\n"); + mtype_local = MTYPE_CC; + } else { + DRM_INFO_ONCE("Using MTYPE_RW for local memory\n"); + } + is_local = (!is_vram && (adev->flags & AMD_IS_APU) && + num_possible_nodes() <= 1) || + (is_vram && adev == bo_adev && + KFD_XCP_MEM_ID(adev, bo->xcp_id) == vm->mem_id); + snoop = true; + if (uncached) { + mtype = MTYPE_UC; + } else if (adev->flags & AMD_IS_APU) { + mtype = is_local ? mtype_local : MTYPE_NC; + } else { + /* dGPU */ + if (is_local) + mtype = mtype_local; + else if (is_vram) + mtype = MTYPE_NC; + else + mtype = MTYPE_UC; + } + + break; default: if (uncached || coherent) mtype = MTYPE_UC; @@ -1241,6 +1307,72 @@ static void gmc_v9_0_get_vm_pte(struct amdgpu_device *adev, mapping, flags); } +static void gmc_v9_0_override_vm_pte_flags(struct amdgpu_device *adev, + struct amdgpu_vm *vm, + uint64_t addr, uint64_t *flags) +{ + int local_node, nid; + + /* Only GFX 9.4.3 APUs associate GPUs with NUMA nodes. Local system + * memory can use more efficient MTYPEs. + */ + if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 3)) + return; + + /* Only direct-mapped memory allows us to determine the NUMA node from + * the DMA address. + */ + if (!adev->ram_is_direct_mapped) { + dev_dbg(adev->dev, "RAM is not direct mapped\n"); + return; + } + + /* Only override mappings with MTYPE_NC, which is the safe default for + * cacheable memory. + */ + if ((*flags & AMDGPU_PTE_MTYPE_VG10_MASK) != + AMDGPU_PTE_MTYPE_VG10(MTYPE_NC)) { + dev_dbg(adev->dev, "MTYPE is not NC\n"); + return; + } + + /* FIXME: Only supported on native mode for now. For carve-out, the + * NUMA affinity of the GPU/VM needs to come from the PCI info because + * memory partitions are not associated with different NUMA nodes. + */ + if (adev->gmc.is_app_apu && vm->mem_id >= 0) { + local_node = adev->gmc.mem_partitions[vm->mem_id].numa.node; + } else { + dev_dbg(adev->dev, "Only native mode APU is supported.\n"); + return; + } + + /* Only handle real RAM. Mappings of PCIe resources don't have struct + * page or NUMA nodes. + */ + if (!page_is_ram(addr >> PAGE_SHIFT)) { + dev_dbg(adev->dev, "Page is not RAM.\n"); + return; + } + nid = pfn_to_nid(addr >> PAGE_SHIFT); + dev_dbg(adev->dev, "vm->mem_id=%d, local_node=%d, nid=%d\n", + vm->mem_id, local_node, nid); + if (nid == local_node) { + uint64_t old_flags = *flags; + unsigned int mtype_local = MTYPE_RW; + + if (amdgpu_mtype_local == 1) + mtype_local = MTYPE_NC; + else if (amdgpu_mtype_local == 2) + mtype_local = MTYPE_CC; + + *flags = (*flags & ~AMDGPU_PTE_MTYPE_VG10_MASK) | + AMDGPU_PTE_MTYPE_VG10(mtype_local); + dev_dbg(adev->dev, "flags updated from %llx to %llx\n", + old_flags, *flags); + } +} + static unsigned gmc_v9_0_get_vbios_fb_size(struct amdgpu_device *adev) { u32 d1vga_control = RREG32_SOC15(DCE, 0, mmD1VGA_CONTROL); @@ -1283,6 +1415,27 @@ static unsigned gmc_v9_0_get_vbios_fb_size(struct amdgpu_device *adev) return size; } +static enum amdgpu_memory_partition +gmc_v9_0_get_memory_partition(struct amdgpu_device *adev, u32 *supp_modes) +{ + enum amdgpu_memory_partition mode = UNKNOWN_MEMORY_PARTITION_MODE; + + if (adev->nbio.funcs->get_memory_partition_mode) + mode = adev->nbio.funcs->get_memory_partition_mode(adev, + supp_modes); + + return mode; +} + +static enum amdgpu_memory_partition +gmc_v9_0_query_memory_partition(struct amdgpu_device *adev) +{ + if (amdgpu_sriov_vf(adev)) + return AMDGPU_NPS1_PARTITION_MODE; + + return gmc_v9_0_get_memory_partition(adev, NULL); +} + static const struct amdgpu_gmc_funcs gmc_v9_0_gmc_funcs = { .flush_gpu_tlb = gmc_v9_0_flush_gpu_tlb, .flush_gpu_tlb_pasid = gmc_v9_0_flush_gpu_tlb_pasid, @@ -1291,7 +1444,9 @@ static const struct amdgpu_gmc_funcs gmc_v9_0_gmc_funcs = { .map_mtype = gmc_v9_0_map_mtype, .get_vm_pde = gmc_v9_0_get_vm_pde, .get_vm_pte = gmc_v9_0_get_vm_pte, + .override_vm_pte_flags = gmc_v9_0_override_vm_pte_flags, .get_vbios_fb_size = gmc_v9_0_get_vbios_fb_size, + .query_mem_partition_mode = &gmc_v9_0_query_memory_partition, }; static void gmc_v9_0_set_gmc_funcs(struct amdgpu_device *adev) @@ -1372,6 +1527,9 @@ static void gmc_v9_0_set_mmhub_ras_funcs(struct amdgpu_device *adev) case IP_VERSION(9, 4, 2): adev->mmhub.ras = &mmhub_v1_7_ras; break; + case IP_VERSION(1, 8, 0): + adev->mmhub.ras = &mmhub_v1_8_ras; + break; default: /* mmhub ras is not available */ break; @@ -1419,9 +1577,13 @@ static int gmc_v9_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - /* ARCT and VEGA20 don't have XGMI defined in their IP discovery tables */ - if (adev->asic_type == CHIP_VEGA20 || - adev->asic_type == CHIP_ARCTURUS) + /* + * 9.4.0, 9.4.1 and 9.4.3 don't have XGMI defined + * in their IP discovery tables + */ + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 0) || + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1) || + adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3)) adev->gmc.xgmi.supported = true; if (adev->ip_versions[XGMI_HWIP][0] == IP_VERSION(6, 1, 0)) { @@ -1430,6 +1592,20 @@ static int gmc_v9_0_early_init(void *handle) adev->smuio.funcs->is_host_gpu_xgmi_supported(adev); } + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3)) { + enum amdgpu_pkg_type pkg_type = + adev->smuio.funcs->get_pkg_type(adev); + /* On GFXIP 9.4.3. APU, there is no physical VRAM domain present + * and the APU, can be in used two possible modes: + * - carveout mode + * - native APU mode + * "is_app_apu" can be used to identify the APU in the native + * mode. + */ + adev->gmc.is_app_apu = (pkg_type == AMDGPU_PKG_TYPE_APU && + !pci_resource_len(adev->pdev, 0)); + } + gmc_v9_0_set_gmc_funcs(adev); gmc_v9_0_set_irq_funcs(adev); gmc_v9_0_set_umc_funcs(adev); @@ -1525,8 +1701,13 @@ static int gmc_v9_0_mc_init(struct amdgpu_device *adev) int r; /* size in MB on si */ - adev->gmc.mc_vram_size = - adev->nbio.funcs->get_memsize(adev) * 1024ULL * 1024ULL; + if (!adev->gmc.is_app_apu) { + adev->gmc.mc_vram_size = + adev->nbio.funcs->get_memsize(adev) * 1024ULL * 1024ULL; + } else { + DRM_DEBUG("Set mc_vram_size = 0 for APP APU\n"); + adev->gmc.mc_vram_size = 0; + } adev->gmc.real_vram_size = adev->gmc.mc_vram_size; if (!(adev->flags & AMD_IS_APU) && @@ -1551,7 +1732,8 @@ static int gmc_v9_0_mc_init(struct amdgpu_device *adev) */ /* check whether both host-gpu and gpu-gpu xgmi links exist */ - if (((adev->flags & AMD_IS_APU) && !amdgpu_passthrough(adev)) || + if ((!amdgpu_sriov_vf(adev) && + (adev->flags & AMD_IS_APU) && !amdgpu_passthrough(adev)) || (adev->gmc.xgmi.supported && adev->gmc.xgmi.connected_to_cpu)) { adev->gmc.aper_base = @@ -1618,12 +1800,18 @@ static int gmc_v9_0_gart_init(struct amdgpu_device *adev) adev->gart.gart_pte_flags = AMDGPU_PTE_MTYPE_VG10(MTYPE_UC) | AMDGPU_PTE_EXECUTABLE; - r = amdgpu_gart_table_vram_alloc(adev); - if (r) - return r; + if (!adev->gmc.real_vram_size) { + dev_info(adev->dev, "Put GART in system memory for APU\n"); + r = amdgpu_gart_table_ram_alloc(adev); + if (r) + dev_err(adev->dev, "Failed to allocate GART in system memory\n"); + } else { + r = amdgpu_gart_table_vram_alloc(adev); + if (r) + return r; - if (adev->gmc.xgmi.connected_to_cpu) { - r = amdgpu_gmc_pdb0_alloc(adev); + if (adev->gmc.xgmi.connected_to_cpu) + r = amdgpu_gmc_pdb0_alloc(adev); } return r; @@ -1644,10 +1832,178 @@ static void gmc_v9_0_save_registers(struct amdgpu_device *adev) adev->gmc.sdpif_register = RREG32_SOC15(DCE, 0, mmDCHUBBUB_SDPIF_MMIO_CNTRL_0); } +static bool gmc_v9_0_validate_partition_info(struct amdgpu_device *adev) +{ + enum amdgpu_memory_partition mode; + u32 supp_modes; + bool valid; + + mode = gmc_v9_0_get_memory_partition(adev, &supp_modes); + + /* Mode detected by hardware not present in supported modes */ + if ((mode != UNKNOWN_MEMORY_PARTITION_MODE) && + !(BIT(mode - 1) & supp_modes)) + return false; + + switch (mode) { + case UNKNOWN_MEMORY_PARTITION_MODE: + case AMDGPU_NPS1_PARTITION_MODE: + valid = (adev->gmc.num_mem_partitions == 1); + break; + case AMDGPU_NPS2_PARTITION_MODE: + valid = (adev->gmc.num_mem_partitions == 2); + break; + case AMDGPU_NPS4_PARTITION_MODE: + valid = (adev->gmc.num_mem_partitions == 3 || + adev->gmc.num_mem_partitions == 4); + break; + default: + valid = false; + } + + return valid; +} + +static bool gmc_v9_0_is_node_present(int *node_ids, int num_ids, int nid) +{ + int i; + + /* Check if node with id 'nid' is present in 'node_ids' array */ + for (i = 0; i < num_ids; ++i) + if (node_ids[i] == nid) + return true; + + return false; +} + +static void +gmc_v9_0_init_acpi_mem_ranges(struct amdgpu_device *adev, + struct amdgpu_mem_partition_info *mem_ranges) +{ + int num_ranges = 0, ret, mem_groups; + struct amdgpu_numa_info numa_info; + int node_ids[MAX_MEM_RANGES]; + int num_xcc, xcc_id; + uint32_t xcc_mask; + + num_xcc = NUM_XCC(adev->gfx.xcc_mask); + xcc_mask = (1U << num_xcc) - 1; + mem_groups = hweight32(adev->aid_mask); + + for_each_inst(xcc_id, xcc_mask) { + ret = amdgpu_acpi_get_mem_info(adev, xcc_id, &numa_info); + if (ret) + continue; + + if (numa_info.nid == NUMA_NO_NODE) { + mem_ranges[0].size = numa_info.size; + mem_ranges[0].numa.node = numa_info.nid; + num_ranges = 1; + break; + } + + if (gmc_v9_0_is_node_present(node_ids, num_ranges, + numa_info.nid)) + continue; + + node_ids[num_ranges] = numa_info.nid; + mem_ranges[num_ranges].numa.node = numa_info.nid; + mem_ranges[num_ranges].size = numa_info.size; + ++num_ranges; + } + + adev->gmc.num_mem_partitions = num_ranges; + + /* If there is only partition, don't use entire size */ + if (adev->gmc.num_mem_partitions == 1) { + mem_ranges[0].size = mem_ranges[0].size * (mem_groups - 1); + do_div(mem_ranges[0].size, mem_groups); + } +} + +static void +gmc_v9_0_init_sw_mem_ranges(struct amdgpu_device *adev, + struct amdgpu_mem_partition_info *mem_ranges) +{ + enum amdgpu_memory_partition mode; + u32 start_addr = 0, size; + int i; + + mode = gmc_v9_0_query_memory_partition(adev); + + switch (mode) { + case UNKNOWN_MEMORY_PARTITION_MODE: + case AMDGPU_NPS1_PARTITION_MODE: + adev->gmc.num_mem_partitions = 1; + break; + case AMDGPU_NPS2_PARTITION_MODE: + adev->gmc.num_mem_partitions = 2; + break; + case AMDGPU_NPS4_PARTITION_MODE: + if (adev->flags & AMD_IS_APU) + adev->gmc.num_mem_partitions = 3; + else + adev->gmc.num_mem_partitions = 4; + break; + default: + adev->gmc.num_mem_partitions = 1; + break; + } + + size = adev->gmc.real_vram_size >> AMDGPU_GPU_PAGE_SHIFT; + size /= adev->gmc.num_mem_partitions; + + for (i = 0; i < adev->gmc.num_mem_partitions; ++i) { + mem_ranges[i].range.fpfn = start_addr; + mem_ranges[i].size = ((u64)size << AMDGPU_GPU_PAGE_SHIFT); + mem_ranges[i].range.lpfn = start_addr + size - 1; + start_addr += size; + } + + /* Adjust the last one */ + mem_ranges[adev->gmc.num_mem_partitions - 1].range.lpfn = + (adev->gmc.real_vram_size >> AMDGPU_GPU_PAGE_SHIFT) - 1; + mem_ranges[adev->gmc.num_mem_partitions - 1].size = + adev->gmc.real_vram_size - + ((u64)mem_ranges[adev->gmc.num_mem_partitions - 1].range.fpfn + << AMDGPU_GPU_PAGE_SHIFT); +} + +static int gmc_v9_0_init_mem_ranges(struct amdgpu_device *adev) +{ + bool valid; + + adev->gmc.mem_partitions = kzalloc( + MAX_MEM_RANGES * sizeof(struct amdgpu_mem_partition_info), + GFP_KERNEL); + + if (!adev->gmc.mem_partitions) + return -ENOMEM; + + /* TODO : Get the range from PSP/Discovery for dGPU */ + if (adev->gmc.is_app_apu) + gmc_v9_0_init_acpi_mem_ranges(adev, adev->gmc.mem_partitions); + else + gmc_v9_0_init_sw_mem_ranges(adev, adev->gmc.mem_partitions); + + if (amdgpu_sriov_vf(adev)) + valid = true; + else + valid = gmc_v9_0_validate_partition_info(adev); + if (!valid) { + /* TODO: handle invalid case */ + dev_WARN(adev->dev, + "Mem ranges not matching with hardware config"); + } + + return 0; +} + static int gmc_v9_0_sw_init(void *handle) { int r, vram_width = 0, vram_type = 0, vram_vendor = 0, dma_addr_bits; struct amdgpu_device *adev = (struct amdgpu_device *)handle; + unsigned long inst_mask = adev->aid_mask; adev->gfxhub.funcs->init(adev); @@ -1655,38 +2011,54 @@ static int gmc_v9_0_sw_init(void *handle) spin_lock_init(&adev->gmc.invalidate_lock); - r = amdgpu_atomfirmware_get_vram_info(adev, - &vram_width, &vram_type, &vram_vendor); - if (amdgpu_sriov_vf(adev)) - /* For Vega10 SR-IOV, vram_width can't be read from ATOM as RAVEN, - * and DF related registers is not readable, seems hardcord is the - * only way to set the correct vram_width - */ - adev->gmc.vram_width = 2048; - else if (amdgpu_emu_mode != 1) - adev->gmc.vram_width = vram_width; + if (!(adev->bios) || adev->gmc.is_app_apu) { + if (adev->flags & AMD_IS_APU) { + if (adev->gmc.is_app_apu) { + adev->gmc.vram_type = AMDGPU_VRAM_TYPE_HBM; + adev->gmc.vram_width = 128 * 64; + } else { + adev->gmc.vram_type = AMDGPU_VRAM_TYPE_DDR4; + adev->gmc.vram_width = 64 * 64; + } + } else { + adev->gmc.vram_type = AMDGPU_VRAM_TYPE_HBM; + adev->gmc.vram_width = 128 * 64; + } + } else { + r = amdgpu_atomfirmware_get_vram_info(adev, + &vram_width, &vram_type, &vram_vendor); + if (amdgpu_sriov_vf(adev)) + /* For Vega10 SR-IOV, vram_width can't be read from ATOM as RAVEN, + * and DF related registers is not readable, seems hardcord is the + * only way to set the correct vram_width + */ + adev->gmc.vram_width = 2048; + else if (amdgpu_emu_mode != 1) + adev->gmc.vram_width = vram_width; - if (!adev->gmc.vram_width) { - int chansize, numchan; + if (!adev->gmc.vram_width) { + int chansize, numchan; - /* hbm memory channel size */ - if (adev->flags & AMD_IS_APU) - chansize = 64; - else - chansize = 128; - if (adev->df.funcs && - adev->df.funcs->get_hbm_channel_number) { - numchan = adev->df.funcs->get_hbm_channel_number(adev); - adev->gmc.vram_width = numchan * chansize; + /* hbm memory channel size */ + if (adev->flags & AMD_IS_APU) + chansize = 64; + else + chansize = 128; + if (adev->df.funcs && + adev->df.funcs->get_hbm_channel_number) { + numchan = adev->df.funcs->get_hbm_channel_number(adev); + adev->gmc.vram_width = numchan * chansize; + } } - } - adev->gmc.vram_type = vram_type; - adev->gmc.vram_vendor = vram_vendor; + adev->gmc.vram_type = vram_type; + adev->gmc.vram_vendor = vram_vendor; + } switch (adev->ip_versions[GC_HWIP][0]) { case IP_VERSION(9, 1, 0): case IP_VERSION(9, 2, 2): - adev->num_vmhubs = 2; + set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask); + set_bit(AMDGPU_MMHUB0(0), adev->vmhubs_mask); if (adev->rev_id == 0x0 || adev->rev_id == 0x1) { amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48); @@ -1702,9 +2074,8 @@ static int gmc_v9_0_sw_init(void *handle) case IP_VERSION(9, 4, 0): case IP_VERSION(9, 3, 0): case IP_VERSION(9, 4, 2): - case IP_VERSION(9, 4, 3): - adev->num_vmhubs = 2; - + set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask); + set_bit(AMDGPU_MMHUB0(0), adev->vmhubs_mask); /* * To fulfill 4-level page support, @@ -1720,12 +2091,24 @@ static int gmc_v9_0_sw_init(void *handle) adev->gmc.translate_further = adev->vm_manager.num_level > 1; break; case IP_VERSION(9, 4, 1): - adev->num_vmhubs = 3; + set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask); + set_bit(AMDGPU_MMHUB0(0), adev->vmhubs_mask); + set_bit(AMDGPU_MMHUB1(0), adev->vmhubs_mask); /* Keep the vm size same with Vega20 */ amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48); adev->gmc.translate_further = adev->vm_manager.num_level > 1; break; + case IP_VERSION(9, 4, 3): + bitmap_set(adev->vmhubs_mask, AMDGPU_GFXHUB(0), + NUM_XCC(adev->gfx.xcc_mask)); + + inst_mask <<= AMDGPU_MMHUB0(0); + bitmap_or(adev->vmhubs_mask, adev->vmhubs_mask, &inst_mask, 32); + + amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48); + adev->gmc.translate_further = adev->vm_manager.num_level > 1; + break; default: break; } @@ -1764,7 +2147,7 @@ static int gmc_v9_0_sw_init(void *handle) */ adev->gmc.mc_mask = 0xffffffffffffULL; /* 48 bit MC */ - dma_addr_bits = adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2) ? 48:44; + dma_addr_bits = adev->ip_versions[GC_HWIP][0] >= IP_VERSION(9, 4, 2) ? 48:44; r = dma_set_mask_and_coherent(adev->dev, DMA_BIT_MASK(dma_addr_bits)); if (r) { printk(KERN_WARNING "amdgpu: No suitable DMA available.\n"); @@ -1778,6 +2161,12 @@ static int gmc_v9_0_sw_init(void *handle) amdgpu_gmc_get_vbios_allocations(adev); + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3)) { + r = gmc_v9_0_init_mem_ranges(adev); + if (r) + return r; + } + /* Memory manager */ r = amdgpu_bo_init(adev); if (r) @@ -1810,6 +2199,9 @@ static int gmc_v9_0_sw_init(void *handle) if (r) return r; + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3)) + amdgpu_gmc_sysfs_init(adev); + return 0; } @@ -1817,10 +2209,20 @@ static int gmc_v9_0_sw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3)) + amdgpu_gmc_sysfs_fini(adev); + adev->gmc.num_mem_partitions = 0; + kfree(adev->gmc.mem_partitions); + amdgpu_gmc_ras_fini(adev); amdgpu_gem_force_release(adev); amdgpu_vm_manager_fini(adev); - amdgpu_gart_table_vram_free(adev); + if (!adev->gmc.real_vram_size) { + dev_info(adev->dev, "Put GART in system memory for APU free\n"); + amdgpu_gart_table_ram_free(adev); + } else { + amdgpu_gart_table_vram_free(adev); + } amdgpu_bo_free_kernel(&adev->gmc.pdb0_bo, NULL, &adev->gmc.ptr_pdb0); amdgpu_bo_fini(adev); @@ -1946,8 +2348,8 @@ static int gmc_v9_0_hw_init(void *handle) adev->gfxhub.funcs->set_fault_enable_default(adev, value); adev->mmhub.funcs->set_fault_enable_default(adev, value); } - for (i = 0; i < adev->num_vmhubs; ++i) { - if (adev->in_s0ix && (i == AMDGPU_GFXHUB_0)) + for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) { + if (adev->in_s0ix && (i == AMDGPU_GFXHUB(0))) continue; gmc_v9_0_flush_gpu_tlb(adev, 0, i, 0); } diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c index a3076eb8af6a..77595e9622da 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c @@ -437,7 +437,7 @@ static int jpeg_v1_0_process_interrupt(struct amdgpu_device *adev, switch (entry->src_id) { case 126: - amdgpu_fence_process(&adev->jpeg.inst->ring_dec); + amdgpu_fence_process(adev->jpeg.inst->ring_dec); break; default: DRM_ERROR("Unhandled interrupt: %d %d\n", @@ -460,6 +460,7 @@ int jpeg_v1_0_early_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; adev->jpeg.num_jpeg_inst = 1; + adev->jpeg.num_jpeg_rings = 1; jpeg_v1_0_set_dec_ring_funcs(adev); jpeg_v1_0_set_irq_funcs(adev); @@ -484,15 +485,15 @@ int jpeg_v1_0_sw_init(void *handle) if (r) return r; - ring = &adev->jpeg.inst->ring_dec; - ring->vm_hub = AMDGPU_MMHUB_0; + ring = adev->jpeg.inst->ring_dec; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "jpeg_dec"); r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq, 0, AMDGPU_RING_PRIO_DEFAULT, NULL); if (r) return r; - adev->jpeg.internal.jpeg_pitch = adev->jpeg.inst->external.jpeg_pitch = + adev->jpeg.internal.jpeg_pitch[0] = adev->jpeg.inst->external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, 0, mmUVD_JPEG_PITCH); return 0; @@ -509,7 +510,7 @@ void jpeg_v1_0_sw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - amdgpu_ring_fini(&adev->jpeg.inst[0].ring_dec); + amdgpu_ring_fini(adev->jpeg.inst->ring_dec); } /** @@ -522,7 +523,7 @@ void jpeg_v1_0_sw_fini(void *handle) */ void jpeg_v1_0_start(struct amdgpu_device *adev, int mode) { - struct amdgpu_ring *ring = &adev->jpeg.inst->ring_dec; + struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec; if (mode == 0) { WREG32_SOC15(JPEG, 0, mmUVD_LMI_JRBC_RB_VMID, 0); @@ -579,7 +580,7 @@ static const struct amdgpu_ring_funcs jpeg_v1_0_decode_ring_vm_funcs = { static void jpeg_v1_0_set_dec_ring_funcs(struct amdgpu_device *adev) { - adev->jpeg.inst->ring_dec.funcs = &jpeg_v1_0_decode_ring_vm_funcs; + adev->jpeg.inst->ring_dec->funcs = &jpeg_v1_0_decode_ring_vm_funcs; DRM_INFO("JPEG decode is enabled in VM mode\n"); } diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c index 0eddf7c824a7..c25d4a07350b 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c @@ -49,6 +49,7 @@ static int jpeg_v2_0_early_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; adev->jpeg.num_jpeg_inst = 1; + adev->jpeg.num_jpeg_rings = 1; jpeg_v2_0_set_dec_ring_funcs(adev); jpeg_v2_0_set_irq_funcs(adev); @@ -83,18 +84,18 @@ static int jpeg_v2_0_sw_init(void *handle) if (r) return r; - ring = &adev->jpeg.inst->ring_dec; + ring = adev->jpeg.inst->ring_dec; ring->use_doorbell = true; ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1; - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "jpeg_dec"); r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq, 0, AMDGPU_RING_PRIO_DEFAULT, NULL); if (r) return r; - adev->jpeg.internal.jpeg_pitch = mmUVD_JPEG_PITCH_INTERNAL_OFFSET; - adev->jpeg.inst->external.jpeg_pitch = SOC15_REG_OFFSET(JPEG, 0, mmUVD_JPEG_PITCH); + adev->jpeg.internal.jpeg_pitch[0] = mmUVD_JPEG_PITCH_INTERNAL_OFFSET; + adev->jpeg.inst->external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, 0, mmUVD_JPEG_PITCH); return 0; } @@ -129,7 +130,7 @@ static int jpeg_v2_0_sw_fini(void *handle) static int jpeg_v2_0_hw_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct amdgpu_ring *ring = &adev->jpeg.inst->ring_dec; + struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec; int r; adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell, @@ -312,7 +313,7 @@ static void jpeg_v2_0_enable_clock_gating(struct amdgpu_device *adev) */ static int jpeg_v2_0_start(struct amdgpu_device *adev) { - struct amdgpu_ring *ring = &adev->jpeg.inst->ring_dec; + struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec; int r; if (adev->pm.dpm_enabled) @@ -729,7 +730,7 @@ static int jpeg_v2_0_process_interrupt(struct amdgpu_device *adev, switch (entry->src_id) { case VCN_2_0__SRCID__JPEG_DECODE: - amdgpu_fence_process(&adev->jpeg.inst->ring_dec); + amdgpu_fence_process(adev->jpeg.inst->ring_dec); break; default: DRM_ERROR("Unhandled interrupt: %d %d\n", @@ -791,7 +792,7 @@ static const struct amdgpu_ring_funcs jpeg_v2_0_dec_ring_vm_funcs = { static void jpeg_v2_0_set_dec_ring_funcs(struct amdgpu_device *adev) { - adev->jpeg.inst->ring_dec.funcs = &jpeg_v2_0_dec_ring_vm_funcs; + adev->jpeg.inst->ring_dec->funcs = &jpeg_v2_0_dec_ring_vm_funcs; DRM_INFO("JPEG decode is enabled in VM mode\n"); } diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c index b040f51d9aa9..aadb74de52bc 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c @@ -60,6 +60,7 @@ static int jpeg_v2_5_early_init(void *handle) u32 harvest; int i; + adev->jpeg.num_jpeg_rings = 1; adev->jpeg.num_jpeg_inst = JPEG25_MAX_HW_INSTANCES_ARCTURUS; for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) { harvest = RREG32_SOC15(JPEG, i, mmCC_UVD_HARVESTING); @@ -102,13 +103,13 @@ static int jpeg_v2_5_sw_init(void *handle) /* JPEG DJPEG POISON EVENT */ r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_jpeg[i], - VCN_2_6__SRCID_DJPEG0_POISON, &adev->jpeg.inst[i].irq); + VCN_2_6__SRCID_DJPEG0_POISON, &adev->jpeg.inst[i].ras_poison_irq); if (r) return r; /* JPEG EJPEG POISON EVENT */ r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_jpeg[i], - VCN_2_6__SRCID_EJPEG0_POISON, &adev->jpeg.inst[i].irq); + VCN_2_6__SRCID_EJPEG0_POISON, &adev->jpeg.inst[i].ras_poison_irq); if (r) return r; } @@ -125,12 +126,12 @@ static int jpeg_v2_5_sw_init(void *handle) if (adev->jpeg.harvest_config & (1 << i)) continue; - ring = &adev->jpeg.inst[i].ring_dec; + ring = adev->jpeg.inst[i].ring_dec; ring->use_doorbell = true; if (adev->ip_versions[UVD_HWIP][0] == IP_VERSION(2, 5, 0)) - ring->vm_hub = AMDGPU_MMHUB_1; + ring->vm_hub = AMDGPU_MMHUB1(0); else - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1 + 8 * i; sprintf(ring->name, "jpeg_dec_%d", i); r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst[i].irq, @@ -138,8 +139,8 @@ static int jpeg_v2_5_sw_init(void *handle) if (r) return r; - adev->jpeg.internal.jpeg_pitch = mmUVD_JPEG_PITCH_INTERNAL_OFFSET; - adev->jpeg.inst[i].external.jpeg_pitch = SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_PITCH); + adev->jpeg.internal.jpeg_pitch[0] = mmUVD_JPEG_PITCH_INTERNAL_OFFSET; + adev->jpeg.inst[i].external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_PITCH); } r = amdgpu_jpeg_ras_sw_init(adev); @@ -186,7 +187,7 @@ static int jpeg_v2_5_hw_init(void *handle) if (adev->jpeg.harvest_config & (1 << i)) continue; - ring = &adev->jpeg.inst[i].ring_dec; + ring = adev->jpeg.inst[i].ring_dec; adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell, (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i, i); @@ -221,6 +222,9 @@ static int jpeg_v2_5_hw_fini(void *handle) if (adev->jpeg.cur_state != AMD_PG_STATE_GATE && RREG32_SOC15(JPEG, i, mmUVD_JRBC_STATUS)) jpeg_v2_5_set_powergating_state(adev, AMD_PG_STATE_GATE); + + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG)) + amdgpu_irq_put(adev, &adev->jpeg.inst[i].ras_poison_irq, 0); } return 0; @@ -326,7 +330,7 @@ static int jpeg_v2_5_start(struct amdgpu_device *adev) if (adev->jpeg.harvest_config & (1 << i)) continue; - ring = &adev->jpeg.inst[i].ring_dec; + ring = adev->jpeg.inst[i].ring_dec; /* disable anti hang mechanism */ WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_POWER_STATUS), 0, ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); @@ -569,6 +573,14 @@ static int jpeg_v2_5_set_interrupt_state(struct amdgpu_device *adev, return 0; } +static int jpeg_v2_6_set_ras_interrupt_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned int type, + enum amdgpu_interrupt_state state) +{ + return 0; +} + static int jpeg_v2_5_process_interrupt(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) @@ -591,11 +603,7 @@ static int jpeg_v2_5_process_interrupt(struct amdgpu_device *adev, switch (entry->src_id) { case VCN_2_0__SRCID__JPEG_DECODE: - amdgpu_fence_process(&adev->jpeg.inst[ip_instance].ring_dec); - break; - case VCN_2_6__SRCID_DJPEG0_POISON: - case VCN_2_6__SRCID_EJPEG0_POISON: - amdgpu_jpeg_process_poison_irq(adev, source, entry); + amdgpu_fence_process(adev->jpeg.inst[ip_instance].ring_dec); break; default: DRM_ERROR("Unhandled interrupt: %d %d\n", @@ -712,10 +720,10 @@ static void jpeg_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev) if (adev->jpeg.harvest_config & (1 << i)) continue; if (adev->asic_type == CHIP_ARCTURUS) - adev->jpeg.inst[i].ring_dec.funcs = &jpeg_v2_5_dec_ring_vm_funcs; + adev->jpeg.inst[i].ring_dec->funcs = &jpeg_v2_5_dec_ring_vm_funcs; else /* CHIP_ALDEBARAN */ - adev->jpeg.inst[i].ring_dec.funcs = &jpeg_v2_6_dec_ring_vm_funcs; - adev->jpeg.inst[i].ring_dec.me = i; + adev->jpeg.inst[i].ring_dec->funcs = &jpeg_v2_6_dec_ring_vm_funcs; + adev->jpeg.inst[i].ring_dec->me = i; DRM_INFO("JPEG(%d) JPEG decode is enabled in VM mode\n", i); } } @@ -725,6 +733,11 @@ static const struct amdgpu_irq_src_funcs jpeg_v2_5_irq_funcs = { .process = jpeg_v2_5_process_interrupt, }; +static const struct amdgpu_irq_src_funcs jpeg_v2_6_ras_irq_funcs = { + .set = jpeg_v2_6_set_ras_interrupt_state, + .process = amdgpu_jpeg_process_poison_irq, +}; + static void jpeg_v2_5_set_irq_funcs(struct amdgpu_device *adev) { int i; @@ -735,6 +748,9 @@ static void jpeg_v2_5_set_irq_funcs(struct amdgpu_device *adev) adev->jpeg.inst[i].irq.num_types = 1; adev->jpeg.inst[i].irq.funcs = &jpeg_v2_5_irq_funcs; + + adev->jpeg.inst[i].ras_poison_irq.num_types = 1; + adev->jpeg.inst[i].ras_poison_irq.funcs = &jpeg_v2_6_ras_irq_funcs; } } @@ -800,6 +816,7 @@ const struct amdgpu_ras_block_hw_ops jpeg_v2_6_ras_hw_ops = { static struct amdgpu_jpeg_ras jpeg_v2_6_ras = { .ras_block = { .hw_ops = &jpeg_v2_6_ras_hw_ops, + .ras_late_init = amdgpu_jpeg_ras_late_init, }, }; diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c index 1c2292cc5f2c..79791379fc2b 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c @@ -64,6 +64,7 @@ static int jpeg_v3_0_early_init(void *handle) } adev->jpeg.num_jpeg_inst = 1; + adev->jpeg.num_jpeg_rings = 1; jpeg_v3_0_set_dec_ring_funcs(adev); jpeg_v3_0_set_irq_funcs(adev); @@ -98,18 +99,18 @@ static int jpeg_v3_0_sw_init(void *handle) if (r) return r; - ring = &adev->jpeg.inst->ring_dec; + ring = adev->jpeg.inst->ring_dec; ring->use_doorbell = true; ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1; - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "jpeg_dec"); r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq, 0, AMDGPU_RING_PRIO_DEFAULT, NULL); if (r) return r; - adev->jpeg.internal.jpeg_pitch = mmUVD_JPEG_PITCH_INTERNAL_OFFSET; - adev->jpeg.inst->external.jpeg_pitch = SOC15_REG_OFFSET(JPEG, 0, mmUVD_JPEG_PITCH); + adev->jpeg.internal.jpeg_pitch[0] = mmUVD_JPEG_PITCH_INTERNAL_OFFSET; + adev->jpeg.inst->external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, 0, mmUVD_JPEG_PITCH); return 0; } @@ -144,7 +145,7 @@ static int jpeg_v3_0_sw_fini(void *handle) static int jpeg_v3_0_hw_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct amdgpu_ring *ring = &adev->jpeg.inst->ring_dec; + struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec; int r; adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell, @@ -330,7 +331,7 @@ static int jpeg_v3_0_enable_static_power_gating(struct amdgpu_device *adev) */ static int jpeg_v3_0_start(struct amdgpu_device *adev) { - struct amdgpu_ring *ring = &adev->jpeg.inst->ring_dec; + struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec; int r; if (adev->pm.dpm_enabled) @@ -527,7 +528,7 @@ static int jpeg_v3_0_process_interrupt(struct amdgpu_device *adev, switch (entry->src_id) { case VCN_2_0__SRCID__JPEG_DECODE: - amdgpu_fence_process(&adev->jpeg.inst->ring_dec); + amdgpu_fence_process(adev->jpeg.inst->ring_dec); break; default: DRM_ERROR("Unhandled interrupt: %d %d\n", @@ -589,7 +590,7 @@ static const struct amdgpu_ring_funcs jpeg_v3_0_dec_ring_vm_funcs = { static void jpeg_v3_0_set_dec_ring_funcs(struct amdgpu_device *adev) { - adev->jpeg.inst->ring_dec.funcs = &jpeg_v3_0_dec_ring_vm_funcs; + adev->jpeg.inst->ring_dec->funcs = &jpeg_v3_0_dec_ring_vm_funcs; DRM_INFO("JPEG decode is enabled in VM mode\n"); } diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c index 77e1e64aa1d1..a707d407fbd0 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c @@ -58,6 +58,7 @@ static int jpeg_v4_0_early_init(void *handle) adev->jpeg.num_jpeg_inst = 1; + adev->jpeg.num_jpeg_rings = 1; jpeg_v4_0_set_dec_ring_funcs(adev); jpeg_v4_0_set_irq_funcs(adev); @@ -87,13 +88,13 @@ static int jpeg_v4_0_sw_init(void *handle) /* JPEG DJPEG POISON EVENT */ r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN, - VCN_4_0__SRCID_DJPEG0_POISON, &adev->jpeg.inst->irq); + VCN_4_0__SRCID_DJPEG0_POISON, &adev->jpeg.inst->ras_poison_irq); if (r) return r; /* JPEG EJPEG POISON EVENT */ r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN, - VCN_4_0__SRCID_EJPEG0_POISON, &adev->jpeg.inst->irq); + VCN_4_0__SRCID_EJPEG0_POISON, &adev->jpeg.inst->ras_poison_irq); if (r) return r; @@ -105,10 +106,10 @@ static int jpeg_v4_0_sw_init(void *handle) if (r) return r; - ring = &adev->jpeg.inst->ring_dec; + ring = adev->jpeg.inst->ring_dec; ring->use_doorbell = true; ring->doorbell_index = amdgpu_sriov_vf(adev) ? (((adev->doorbell_index.vcn.vcn_ring0_1) << 1) + 4) : ((adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1); - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "jpeg_dec"); r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq, 0, @@ -116,8 +117,8 @@ static int jpeg_v4_0_sw_init(void *handle) if (r) return r; - adev->jpeg.internal.jpeg_pitch = regUVD_JPEG_PITCH_INTERNAL_OFFSET; - adev->jpeg.inst->external.jpeg_pitch = SOC15_REG_OFFSET(JPEG, 0, regUVD_JPEG_PITCH); + adev->jpeg.internal.jpeg_pitch[0] = regUVD_JPEG_PITCH_INTERNAL_OFFSET; + adev->jpeg.inst->external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, 0, regUVD_JPEG_PITCH); r = amdgpu_jpeg_ras_sw_init(adev); if (r) @@ -156,7 +157,7 @@ static int jpeg_v4_0_sw_fini(void *handle) static int jpeg_v4_0_hw_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct amdgpu_ring *ring = &adev->jpeg.inst->ring_dec; + struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec; int r; if (amdgpu_sriov_vf(adev)) { @@ -202,7 +203,8 @@ static int jpeg_v4_0_hw_fini(void *handle) RREG32_SOC15(JPEG, 0, regUVD_JRBC_STATUS)) jpeg_v4_0_set_powergating_state(adev, AMD_PG_STATE_GATE); } - amdgpu_irq_put(adev, &adev->jpeg.inst->irq, 0); + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG)) + amdgpu_irq_put(adev, &adev->jpeg.inst->ras_poison_irq, 0); return 0; } @@ -363,7 +365,7 @@ static int jpeg_v4_0_enable_static_power_gating(struct amdgpu_device *adev) */ static int jpeg_v4_0_start(struct amdgpu_device *adev) { - struct amdgpu_ring *ring = &adev->jpeg.inst->ring_dec; + struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec; int r; if (adev->pm.dpm_enabled) @@ -441,7 +443,7 @@ static int jpeg_v4_0_start_sriov(struct amdgpu_device *adev) table_size = 0; - ring = &adev->jpeg.inst->ring_dec; + ring = adev->jpeg.inst->ring_dec; MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(JPEG, 0, regUVD_LMI_JRBC_RB_64BIT_BAR_LOW), @@ -670,6 +672,14 @@ static int jpeg_v4_0_set_interrupt_state(struct amdgpu_device *adev, return 0; } +static int jpeg_v4_0_set_ras_interrupt_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned int type, + enum amdgpu_interrupt_state state) +{ + return 0; +} + static int jpeg_v4_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) @@ -678,11 +688,7 @@ static int jpeg_v4_0_process_interrupt(struct amdgpu_device *adev, switch (entry->src_id) { case VCN_4_0__SRCID__JPEG_DECODE: - amdgpu_fence_process(&adev->jpeg.inst->ring_dec); - break; - case VCN_4_0__SRCID_DJPEG0_POISON: - case VCN_4_0__SRCID_EJPEG0_POISON: - amdgpu_jpeg_process_poison_irq(adev, source, entry); + amdgpu_fence_process(adev->jpeg.inst->ring_dec); break; default: DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n", @@ -744,7 +750,7 @@ static const struct amdgpu_ring_funcs jpeg_v4_0_dec_ring_vm_funcs = { static void jpeg_v4_0_set_dec_ring_funcs(struct amdgpu_device *adev) { - adev->jpeg.inst->ring_dec.funcs = &jpeg_v4_0_dec_ring_vm_funcs; + adev->jpeg.inst->ring_dec->funcs = &jpeg_v4_0_dec_ring_vm_funcs; DRM_DEV_INFO(adev->dev, "JPEG decode is enabled in VM mode\n"); } @@ -753,10 +759,18 @@ static const struct amdgpu_irq_src_funcs jpeg_v4_0_irq_funcs = { .process = jpeg_v4_0_process_interrupt, }; +static const struct amdgpu_irq_src_funcs jpeg_v4_0_ras_irq_funcs = { + .set = jpeg_v4_0_set_ras_interrupt_state, + .process = amdgpu_jpeg_process_poison_irq, +}; + static void jpeg_v4_0_set_irq_funcs(struct amdgpu_device *adev) { adev->jpeg.inst->irq.num_types = 1; adev->jpeg.inst->irq.funcs = &jpeg_v4_0_irq_funcs; + + adev->jpeg.inst->ras_poison_irq.num_types = 1; + adev->jpeg.inst->ras_poison_irq.funcs = &jpeg_v4_0_ras_irq_funcs; } const struct amdgpu_ip_block_version jpeg_v4_0_ip_block = { @@ -811,6 +825,7 @@ const struct amdgpu_ras_block_hw_ops jpeg_v4_0_ras_hw_ops = { static struct amdgpu_jpeg_ras jpeg_v4_0_ras = { .ras_block = { .hw_ops = &jpeg_v4_0_ras_hw_ops, + .ras_late_init = amdgpu_jpeg_ras_late_init, }, }; diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c new file mode 100644 index 000000000000..ce2b22f7e4e4 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c @@ -0,0 +1,1074 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#include "amdgpu.h" +#include "amdgpu_jpeg.h" +#include "soc15.h" +#include "soc15d.h" +#include "jpeg_v4_0_3.h" + +#include "vcn/vcn_4_0_3_offset.h" +#include "vcn/vcn_4_0_3_sh_mask.h" +#include "ivsrcid/vcn/irqsrcs_vcn_4_0.h" + +enum jpeg_engin_status { + UVD_PGFSM_STATUS__UVDJ_PWR_ON = 0, + UVD_PGFSM_STATUS__UVDJ_PWR_OFF = 2, +}; + +static void jpeg_v4_0_3_set_dec_ring_funcs(struct amdgpu_device *adev); +static void jpeg_v4_0_3_set_irq_funcs(struct amdgpu_device *adev); +static int jpeg_v4_0_3_set_powergating_state(void *handle, + enum amd_powergating_state state); +static void jpeg_v4_0_3_set_ras_funcs(struct amdgpu_device *adev); + +static int amdgpu_ih_srcid_jpeg[] = { + VCN_4_0__SRCID__JPEG_DECODE, + VCN_4_0__SRCID__JPEG1_DECODE, + VCN_4_0__SRCID__JPEG2_DECODE, + VCN_4_0__SRCID__JPEG3_DECODE, + VCN_4_0__SRCID__JPEG4_DECODE, + VCN_4_0__SRCID__JPEG5_DECODE, + VCN_4_0__SRCID__JPEG6_DECODE, + VCN_4_0__SRCID__JPEG7_DECODE +}; + +/** + * jpeg_v4_0_3_early_init - set function pointers + * + * @handle: amdgpu_device pointer + * + * Set ring and irq function pointers + */ +static int jpeg_v4_0_3_early_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + adev->jpeg.num_jpeg_rings = AMDGPU_MAX_JPEG_RINGS; + + jpeg_v4_0_3_set_dec_ring_funcs(adev); + jpeg_v4_0_3_set_irq_funcs(adev); + jpeg_v4_0_3_set_ras_funcs(adev); + + return 0; +} + +/** + * jpeg_v4_0_3_sw_init - sw init for JPEG block + * + * @handle: amdgpu_device pointer + * + * Load firmware and sw initialization + */ +static int jpeg_v4_0_3_sw_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_ring *ring; + int i, j, r, jpeg_inst; + + for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { + /* JPEG TRAP */ + r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN, + amdgpu_ih_srcid_jpeg[j], &adev->jpeg.inst->irq); + if (r) + return r; + } + + r = amdgpu_jpeg_sw_init(adev); + if (r) + return r; + + r = amdgpu_jpeg_resume(adev); + if (r) + return r; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + jpeg_inst = GET_INST(JPEG, i); + + for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { + ring = &adev->jpeg.inst[i].ring_dec[j]; + ring->use_doorbell = true; + ring->vm_hub = AMDGPU_MMHUB0(adev->jpeg.inst[i].aid_id); + ring->doorbell_index = + (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + + 1 + j + 9 * jpeg_inst; + sprintf(ring->name, "jpeg_dec_%d.%d", adev->jpeg.inst[i].aid_id, j); + r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq, 0, + AMDGPU_RING_PRIO_DEFAULT, NULL); + if (r) + return r; + + adev->jpeg.internal.jpeg_pitch[j] = + regUVD_JRBC0_UVD_JRBC_SCRATCH0_INTERNAL_OFFSET; + adev->jpeg.inst[i].external.jpeg_pitch[j] = + SOC15_REG_OFFSET1( + JPEG, jpeg_inst, + regUVD_JRBC0_UVD_JRBC_SCRATCH0, + (j ? (0x40 * j - 0xc80) : 0)); + } + } + + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG)) { + r = amdgpu_jpeg_ras_sw_init(adev); + if (r) { + dev_err(adev->dev, "Failed to initialize jpeg ras block!\n"); + return r; + } + } + + return 0; +} + +/** + * jpeg_v4_0_3_sw_fini - sw fini for JPEG block + * + * @handle: amdgpu_device pointer + * + * JPEG suspend and free up sw allocation + */ +static int jpeg_v4_0_3_sw_fini(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int r; + + r = amdgpu_jpeg_suspend(adev); + if (r) + return r; + + r = amdgpu_jpeg_sw_fini(adev); + + return r; +} + +/** + * jpeg_v4_0_3_hw_init - start and test JPEG block + * + * @handle: amdgpu_device pointer + * + */ +static int jpeg_v4_0_3_hw_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_ring *ring; + int i, j, r, jpeg_inst; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + jpeg_inst = GET_INST(JPEG, i); + + ring = adev->jpeg.inst[i].ring_dec; + + if (ring->use_doorbell) + adev->nbio.funcs->vcn_doorbell_range( + adev, ring->use_doorbell, + (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + + 9 * jpeg_inst, + adev->jpeg.inst[i].aid_id); + + for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { + ring = &adev->jpeg.inst[i].ring_dec[j]; + if (ring->use_doorbell) + WREG32_SOC15_OFFSET( + VCN, GET_INST(VCN, i), + regVCN_JPEG_DB_CTRL, + (ring->pipe ? (ring->pipe - 0x15) : 0), + ring->doorbell_index + << VCN_JPEG_DB_CTRL__OFFSET__SHIFT | + VCN_JPEG_DB_CTRL__EN_MASK); + r = amdgpu_ring_test_helper(ring); + if (r) + return r; + } + } + DRM_DEV_INFO(adev->dev, "JPEG decode initialized successfully.\n"); + + return 0; +} + +/** + * jpeg_v4_0_3_hw_fini - stop the hardware block + * + * @handle: amdgpu_device pointer + * + * Stop the JPEG block, mark ring as not ready any more + */ +static int jpeg_v4_0_3_hw_fini(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int ret = 0; + + cancel_delayed_work_sync(&adev->jpeg.idle_work); + + if (adev->jpeg.cur_state != AMD_PG_STATE_GATE) + ret = jpeg_v4_0_3_set_powergating_state(adev, AMD_PG_STATE_GATE); + + return ret; +} + +/** + * jpeg_v4_0_3_suspend - suspend JPEG block + * + * @handle: amdgpu_device pointer + * + * HW fini and suspend JPEG block + */ +static int jpeg_v4_0_3_suspend(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int r; + + r = jpeg_v4_0_3_hw_fini(adev); + if (r) + return r; + + r = amdgpu_jpeg_suspend(adev); + + return r; +} + +/** + * jpeg_v4_0_3_resume - resume JPEG block + * + * @handle: amdgpu_device pointer + * + * Resume firmware and hw init JPEG block + */ +static int jpeg_v4_0_3_resume(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int r; + + r = amdgpu_jpeg_resume(adev); + if (r) + return r; + + r = jpeg_v4_0_3_hw_init(adev); + + return r; +} + +static void jpeg_v4_0_3_disable_clock_gating(struct amdgpu_device *adev, int inst_idx) +{ + int i, jpeg_inst; + uint32_t data; + + jpeg_inst = GET_INST(JPEG, inst_idx); + data = RREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_CTRL); + if (adev->cg_flags & AMD_CG_SUPPORT_JPEG_MGCG) { + data |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + data &= (~(JPEG_CGC_CTRL__JPEG0_DEC_MODE_MASK << 1)); + } else { + data &= ~JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + } + + data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT; + data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT; + WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_CTRL, data); + + data = RREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_GATE); + data &= ~(JPEG_CGC_GATE__JMCIF_MASK | JPEG_CGC_GATE__JRBBM_MASK); + for (i = 0; i < adev->jpeg.num_jpeg_rings; ++i) + data &= ~(JPEG_CGC_GATE__JPEG0_DEC_MASK << i); + WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_GATE, data); +} + +static void jpeg_v4_0_3_enable_clock_gating(struct amdgpu_device *adev, int inst_idx) +{ + int i, jpeg_inst; + uint32_t data; + + jpeg_inst = GET_INST(JPEG, inst_idx); + data = RREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_CTRL); + if (adev->cg_flags & AMD_CG_SUPPORT_JPEG_MGCG) { + data |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + data |= (JPEG_CGC_CTRL__JPEG0_DEC_MODE_MASK << 1); + } else { + data &= ~JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + } + + data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT; + data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT; + WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_CTRL, data); + + data = RREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_GATE); + data |= (JPEG_CGC_GATE__JMCIF_MASK | JPEG_CGC_GATE__JRBBM_MASK); + for (i = 0; i < adev->jpeg.num_jpeg_rings; ++i) + data |= (JPEG_CGC_GATE__JPEG0_DEC_MASK << i); + WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_GATE, data); +} + +/** + * jpeg_v4_0_3_start - start JPEG block + * + * @adev: amdgpu_device pointer + * + * Setup and start the JPEG block + */ +static int jpeg_v4_0_3_start(struct amdgpu_device *adev) +{ + struct amdgpu_ring *ring; + int i, j, jpeg_inst; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + jpeg_inst = GET_INST(JPEG, i); + + WREG32_SOC15(JPEG, jpeg_inst, regUVD_PGFSM_CONFIG, + 1 << UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG__SHIFT); + SOC15_WAIT_ON_RREG( + JPEG, jpeg_inst, regUVD_PGFSM_STATUS, + UVD_PGFSM_STATUS__UVDJ_PWR_ON + << UVD_PGFSM_STATUS__UVDJ_PWR_STATUS__SHIFT, + UVD_PGFSM_STATUS__UVDJ_PWR_STATUS_MASK); + + /* disable anti hang mechanism */ + WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, + regUVD_JPEG_POWER_STATUS), + 0, ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); + + /* JPEG disable CGC */ + jpeg_v4_0_3_disable_clock_gating(adev, i); + + /* MJPEG global tiling registers */ + WREG32_SOC15(JPEG, jpeg_inst, regJPEG_DEC_GFX8_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(JPEG, jpeg_inst, regJPEG_DEC_GFX10_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + + /* enable JMI channel */ + WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JMI_CNTL), 0, + ~UVD_JMI_CNTL__SOFT_RESET_MASK); + + for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { + unsigned int reg_offset = (j?(0x40 * j - 0xc80):0); + + ring = &adev->jpeg.inst[i].ring_dec[j]; + + /* enable System Interrupt for JRBC */ + WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, + regJPEG_SYS_INT_EN), + JPEG_SYS_INT_EN__DJRBC0_MASK << j, + ~(JPEG_SYS_INT_EN__DJRBC0_MASK << j)); + + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JMI0_UVD_LMI_JRBC_RB_VMID, + reg_offset, 0); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JRBC0_UVD_JRBC_RB_CNTL, + reg_offset, + (0x00000001L | 0x00000002L)); + WREG32_SOC15_OFFSET( + JPEG, jpeg_inst, + regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_LOW, + reg_offset, lower_32_bits(ring->gpu_addr)); + WREG32_SOC15_OFFSET( + JPEG, jpeg_inst, + regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH, + reg_offset, upper_32_bits(ring->gpu_addr)); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JRBC0_UVD_JRBC_RB_RPTR, + reg_offset, 0); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JRBC0_UVD_JRBC_RB_WPTR, + reg_offset, 0); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JRBC0_UVD_JRBC_RB_CNTL, + reg_offset, 0x00000002L); + WREG32_SOC15_OFFSET(JPEG, jpeg_inst, + regUVD_JRBC0_UVD_JRBC_RB_SIZE, + reg_offset, ring->ring_size / 4); + ring->wptr = RREG32_SOC15_OFFSET( + JPEG, jpeg_inst, regUVD_JRBC0_UVD_JRBC_RB_WPTR, + reg_offset); + } + } + + return 0; +} + +/** + * jpeg_v4_0_3_stop - stop JPEG block + * + * @adev: amdgpu_device pointer + * + * stop the JPEG block + */ +static int jpeg_v4_0_3_stop(struct amdgpu_device *adev) +{ + int i, jpeg_inst; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + jpeg_inst = GET_INST(JPEG, i); + /* reset JMI */ + WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JMI_CNTL), + UVD_JMI_CNTL__SOFT_RESET_MASK, + ~UVD_JMI_CNTL__SOFT_RESET_MASK); + + jpeg_v4_0_3_enable_clock_gating(adev, i); + + /* enable anti hang mechanism */ + WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, + regUVD_JPEG_POWER_STATUS), + UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK, + ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); + + WREG32_SOC15(JPEG, jpeg_inst, regUVD_PGFSM_CONFIG, + 2 << UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG__SHIFT); + SOC15_WAIT_ON_RREG( + JPEG, jpeg_inst, regUVD_PGFSM_STATUS, + UVD_PGFSM_STATUS__UVDJ_PWR_OFF + << UVD_PGFSM_STATUS__UVDJ_PWR_STATUS__SHIFT, + UVD_PGFSM_STATUS__UVDJ_PWR_STATUS_MASK); + } + + return 0; +} + +/** + * jpeg_v4_0_3_dec_ring_get_rptr - get read pointer + * + * @ring: amdgpu_ring pointer + * + * Returns the current hardware read pointer + */ +static uint64_t jpeg_v4_0_3_dec_ring_get_rptr(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + + return RREG32_SOC15_OFFSET( + JPEG, GET_INST(JPEG, ring->me), regUVD_JRBC0_UVD_JRBC_RB_RPTR, + ring->pipe ? (0x40 * ring->pipe - 0xc80) : 0); +} + +/** + * jpeg_v4_0_3_dec_ring_get_wptr - get write pointer + * + * @ring: amdgpu_ring pointer + * + * Returns the current hardware write pointer + */ +static uint64_t jpeg_v4_0_3_dec_ring_get_wptr(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + + if (ring->use_doorbell) + return adev->wb.wb[ring->wptr_offs]; + else + return RREG32_SOC15_OFFSET( + JPEG, GET_INST(JPEG, ring->me), + regUVD_JRBC0_UVD_JRBC_RB_WPTR, + ring->pipe ? (0x40 * ring->pipe - 0xc80) : 0); +} + +/** + * jpeg_v4_0_3_dec_ring_set_wptr - set write pointer + * + * @ring: amdgpu_ring pointer + * + * Commits the write pointer to the hardware + */ +static void jpeg_v4_0_3_dec_ring_set_wptr(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + + if (ring->use_doorbell) { + adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr); + WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr)); + } else { + WREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me), + regUVD_JRBC0_UVD_JRBC_RB_WPTR, + (ring->pipe ? (0x40 * ring->pipe - 0xc80) : + 0), + lower_32_bits(ring->wptr)); + } +} + +/** + * jpeg_v4_0_3_dec_ring_insert_start - insert a start command + * + * @ring: amdgpu_ring pointer + * + * Write a start command to the ring. + */ +static void jpeg_v4_0_3_dec_ring_insert_start(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x62a04); /* PCTL0_MMHUB_DEEPSLEEP_IB */ + + amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x80004000); +} + +/** + * jpeg_v4_0_3_dec_ring_insert_end - insert a end command + * + * @ring: amdgpu_ring pointer + * + * Write a end command to the ring. + */ +static void jpeg_v4_0_3_dec_ring_insert_end(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x62a04); + + amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x00004000); +} + +/** + * jpeg_v4_0_3_dec_ring_emit_fence - emit an fence & trap command + * + * @ring: amdgpu_ring pointer + * @addr: address + * @seq: sequence number + * @flags: fence related flags + * + * Write a fence and a trap command to the ring. + */ +static void jpeg_v4_0_3_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq, + unsigned int flags) +{ + WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT); + + amdgpu_ring_write(ring, PACKETJ(regUVD_JPEG_GPCOM_DATA0_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, seq); + + amdgpu_ring_write(ring, PACKETJ(regUVD_JPEG_GPCOM_DATA1_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, seq); + + amdgpu_ring_write(ring, PACKETJ(regUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, lower_32_bits(addr)); + + amdgpu_ring_write(ring, PACKETJ(regUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, upper_32_bits(addr)); + + amdgpu_ring_write(ring, PACKETJ(regUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x8); + + amdgpu_ring_write(ring, PACKETJ(regUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET, + 0, PACKETJ_CONDITION_CHECK0, PACKETJ_TYPE4)); + amdgpu_ring_write(ring, 0); + + if (ring->adev->jpeg.inst[ring->me].aid_id) { + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_EXTERNAL_MCM_ADDR_INTERNAL_OFFSET, + 0, PACKETJ_CONDITION_CHECK0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x4); + } else { + amdgpu_ring_write(ring, PACKETJ(0, 0, 0, PACKETJ_TYPE6)); + amdgpu_ring_write(ring, 0); + } + + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x3fbc); + + if (ring->adev->jpeg.inst[ring->me].aid_id) { + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_EXTERNAL_MCM_ADDR_INTERNAL_OFFSET, + 0, PACKETJ_CONDITION_CHECK0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x0); + } else { + amdgpu_ring_write(ring, PACKETJ(0, 0, 0, PACKETJ_TYPE6)); + amdgpu_ring_write(ring, 0); + } + + amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x1); + + amdgpu_ring_write(ring, PACKETJ(0, 0, 0, PACKETJ_TYPE7)); + amdgpu_ring_write(ring, 0); +} + +/** + * jpeg_v4_0_3_dec_ring_emit_ib - execute indirect buffer + * + * @ring: amdgpu_ring pointer + * @job: job to retrieve vmid from + * @ib: indirect buffer to execute + * @flags: unused + * + * Write ring commands to execute the indirect buffer. + */ +static void jpeg_v4_0_3_dec_ring_emit_ib(struct amdgpu_ring *ring, + struct amdgpu_job *job, + struct amdgpu_ib *ib, + uint32_t flags) +{ + unsigned int vmid = AMDGPU_JOB_GET_VMID(job); + + amdgpu_ring_write(ring, PACKETJ(regUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, (vmid | (vmid << 4))); + + amdgpu_ring_write(ring, PACKETJ(regUVD_LMI_JPEG_VMID_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, (vmid | (vmid << 4))); + + amdgpu_ring_write(ring, PACKETJ(regUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr)); + + amdgpu_ring_write(ring, PACKETJ(regUVD_LMI_JRBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr)); + + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_IB_SIZE_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, ib->length_dw); + + amdgpu_ring_write(ring, PACKETJ(regUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_LOW_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, lower_32_bits(ring->gpu_addr)); + + amdgpu_ring_write(ring, PACKETJ(regUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_HIGH_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, upper_32_bits(ring->gpu_addr)); + + amdgpu_ring_write(ring, PACKETJ(0, 0, PACKETJ_CONDITION_CHECK0, PACKETJ_TYPE2)); + amdgpu_ring_write(ring, 0); + + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x01400200); + + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x2); + + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_STATUS_INTERNAL_OFFSET, + 0, PACKETJ_CONDITION_CHECK3, PACKETJ_TYPE3)); + amdgpu_ring_write(ring, 0x2); +} + +static void jpeg_v4_0_3_dec_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg, + uint32_t val, uint32_t mask) +{ + uint32_t reg_offset = (reg << 2); + + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, 0x01400200); + + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + amdgpu_ring_write(ring, val); + + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + if (reg_offset >= 0x10000 && reg_offset <= 0x105ff) { + amdgpu_ring_write(ring, 0); + amdgpu_ring_write(ring, + PACKETJ((reg_offset >> 2), 0, 0, PACKETJ_TYPE3)); + } else { + amdgpu_ring_write(ring, reg_offset); + amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, + 0, 0, PACKETJ_TYPE3)); + } + amdgpu_ring_write(ring, mask); +} + +static void jpeg_v4_0_3_dec_ring_emit_vm_flush(struct amdgpu_ring *ring, + unsigned int vmid, uint64_t pd_addr) +{ + struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub]; + uint32_t data0, data1, mask; + + pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr); + + /* wait for register write */ + data0 = hub->ctx0_ptb_addr_lo32 + vmid * hub->ctx_addr_distance; + data1 = lower_32_bits(pd_addr); + mask = 0xffffffff; + jpeg_v4_0_3_dec_ring_emit_reg_wait(ring, data0, data1, mask); +} + +static void jpeg_v4_0_3_dec_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_t val) +{ + uint32_t reg_offset = (reg << 2); + + amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET, + 0, 0, PACKETJ_TYPE0)); + if (reg_offset >= 0x10000 && reg_offset <= 0x105ff) { + amdgpu_ring_write(ring, 0); + amdgpu_ring_write(ring, + PACKETJ((reg_offset >> 2), 0, 0, PACKETJ_TYPE0)); + } else { + amdgpu_ring_write(ring, reg_offset); + amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, + 0, 0, PACKETJ_TYPE0)); + } + amdgpu_ring_write(ring, val); +} + +static void jpeg_v4_0_3_dec_ring_nop(struct amdgpu_ring *ring, uint32_t count) +{ + int i; + + WARN_ON(ring->wptr % 2 || count % 2); + + for (i = 0; i < count / 2; i++) { + amdgpu_ring_write(ring, PACKETJ(0, 0, 0, PACKETJ_TYPE6)); + amdgpu_ring_write(ring, 0); + } +} + +static bool jpeg_v4_0_3_is_idle(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + bool ret = false; + int i, j; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { + unsigned int reg_offset = (j?(0x40 * j - 0xc80):0); + + ret &= ((RREG32_SOC15_OFFSET( + JPEG, GET_INST(JPEG, i), + regUVD_JRBC0_UVD_JRBC_STATUS, + reg_offset) & + UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK) == + UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK); + } + } + + return ret; +} + +static int jpeg_v4_0_3_wait_for_idle(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int ret = 0; + int i, j; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { + unsigned int reg_offset = (j?(0x40 * j - 0xc80):0); + + ret &= SOC15_WAIT_ON_RREG_OFFSET( + JPEG, GET_INST(JPEG, i), + regUVD_JRBC0_UVD_JRBC_STATUS, reg_offset, + UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK, + UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK); + } + } + return ret; +} + +static int jpeg_v4_0_3_set_clockgating_state(void *handle, + enum amd_clockgating_state state) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + bool enable = (state == AMD_CG_STATE_GATE) ? true : false; + int i; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (enable) { + if (!jpeg_v4_0_3_is_idle(handle)) + return -EBUSY; + jpeg_v4_0_3_enable_clock_gating(adev, i); + } else { + jpeg_v4_0_3_disable_clock_gating(adev, i); + } + } + return 0; +} + +static int jpeg_v4_0_3_set_powergating_state(void *handle, + enum amd_powergating_state state) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int ret; + + if (state == adev->jpeg.cur_state) + return 0; + + if (state == AMD_PG_STATE_GATE) + ret = jpeg_v4_0_3_stop(adev); + else + ret = jpeg_v4_0_3_start(adev); + + if (!ret) + adev->jpeg.cur_state = state; + + return ret; +} + +static int jpeg_v4_0_3_set_interrupt_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned int type, + enum amdgpu_interrupt_state state) +{ + return 0; +} + +static int jpeg_v4_0_3_process_interrupt(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + struct amdgpu_iv_entry *entry) +{ + uint32_t i, inst; + + i = node_id_to_phys_map[entry->node_id]; + DRM_DEV_DEBUG(adev->dev, "IH: JPEG TRAP\n"); + + for (inst = 0; inst < adev->jpeg.num_jpeg_inst; ++inst) + if (adev->jpeg.inst[inst].aid_id == i) + break; + + if (inst >= adev->jpeg.num_jpeg_inst) { + dev_WARN_ONCE(adev->dev, 1, + "Interrupt received for unknown JPEG instance %d", + entry->node_id); + return 0; + } + + switch (entry->src_id) { + case VCN_4_0__SRCID__JPEG_DECODE: + amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[0]); + break; + case VCN_4_0__SRCID__JPEG1_DECODE: + amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[1]); + break; + case VCN_4_0__SRCID__JPEG2_DECODE: + amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[2]); + break; + case VCN_4_0__SRCID__JPEG3_DECODE: + amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[3]); + break; + case VCN_4_0__SRCID__JPEG4_DECODE: + amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[4]); + break; + case VCN_4_0__SRCID__JPEG5_DECODE: + amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[5]); + break; + case VCN_4_0__SRCID__JPEG6_DECODE: + amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[6]); + break; + case VCN_4_0__SRCID__JPEG7_DECODE: + amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[7]); + break; + default: + DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n", + entry->src_id, entry->src_data[0]); + break; + } + + return 0; +} + +static const struct amd_ip_funcs jpeg_v4_0_3_ip_funcs = { + .name = "jpeg_v4_0_3", + .early_init = jpeg_v4_0_3_early_init, + .late_init = NULL, + .sw_init = jpeg_v4_0_3_sw_init, + .sw_fini = jpeg_v4_0_3_sw_fini, + .hw_init = jpeg_v4_0_3_hw_init, + .hw_fini = jpeg_v4_0_3_hw_fini, + .suspend = jpeg_v4_0_3_suspend, + .resume = jpeg_v4_0_3_resume, + .is_idle = jpeg_v4_0_3_is_idle, + .wait_for_idle = jpeg_v4_0_3_wait_for_idle, + .check_soft_reset = NULL, + .pre_soft_reset = NULL, + .soft_reset = NULL, + .post_soft_reset = NULL, + .set_clockgating_state = jpeg_v4_0_3_set_clockgating_state, + .set_powergating_state = jpeg_v4_0_3_set_powergating_state, +}; + +static const struct amdgpu_ring_funcs jpeg_v4_0_3_dec_ring_vm_funcs = { + .type = AMDGPU_RING_TYPE_VCN_JPEG, + .align_mask = 0xf, + .get_rptr = jpeg_v4_0_3_dec_ring_get_rptr, + .get_wptr = jpeg_v4_0_3_dec_ring_get_wptr, + .set_wptr = jpeg_v4_0_3_dec_ring_set_wptr, + .emit_frame_size = + SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 + + SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 + + 8 + /* jpeg_v4_0_3_dec_ring_emit_vm_flush */ + 22 + 22 + /* jpeg_v4_0_3_dec_ring_emit_fence x2 vm fence */ + 8 + 16, + .emit_ib_size = 22, /* jpeg_v4_0_3_dec_ring_emit_ib */ + .emit_ib = jpeg_v4_0_3_dec_ring_emit_ib, + .emit_fence = jpeg_v4_0_3_dec_ring_emit_fence, + .emit_vm_flush = jpeg_v4_0_3_dec_ring_emit_vm_flush, + .test_ring = amdgpu_jpeg_dec_ring_test_ring, + .test_ib = amdgpu_jpeg_dec_ring_test_ib, + .insert_nop = jpeg_v4_0_3_dec_ring_nop, + .insert_start = jpeg_v4_0_3_dec_ring_insert_start, + .insert_end = jpeg_v4_0_3_dec_ring_insert_end, + .pad_ib = amdgpu_ring_generic_pad_ib, + .begin_use = amdgpu_jpeg_ring_begin_use, + .end_use = amdgpu_jpeg_ring_end_use, + .emit_wreg = jpeg_v4_0_3_dec_ring_emit_wreg, + .emit_reg_wait = jpeg_v4_0_3_dec_ring_emit_reg_wait, + .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, +}; + +static void jpeg_v4_0_3_set_dec_ring_funcs(struct amdgpu_device *adev) +{ + int i, j, jpeg_inst; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { + adev->jpeg.inst[i].ring_dec[j].funcs = &jpeg_v4_0_3_dec_ring_vm_funcs; + adev->jpeg.inst[i].ring_dec[j].me = i; + adev->jpeg.inst[i].ring_dec[j].pipe = j; + } + jpeg_inst = GET_INST(JPEG, i); + adev->jpeg.inst[i].aid_id = + jpeg_inst / adev->jpeg.num_inst_per_aid; + } + DRM_DEV_INFO(adev->dev, "JPEG decode is enabled in VM mode\n"); +} + +static const struct amdgpu_irq_src_funcs jpeg_v4_0_3_irq_funcs = { + .set = jpeg_v4_0_3_set_interrupt_state, + .process = jpeg_v4_0_3_process_interrupt, +}; + +static void jpeg_v4_0_3_set_irq_funcs(struct amdgpu_device *adev) +{ + int i; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + adev->jpeg.inst->irq.num_types += adev->jpeg.num_jpeg_rings; + } + adev->jpeg.inst->irq.funcs = &jpeg_v4_0_3_irq_funcs; +} + +const struct amdgpu_ip_block_version jpeg_v4_0_3_ip_block = { + .type = AMD_IP_BLOCK_TYPE_JPEG, + .major = 4, + .minor = 0, + .rev = 3, + .funcs = &jpeg_v4_0_3_ip_funcs, +}; + +static const struct amdgpu_ras_err_status_reg_entry jpeg_v4_0_3_ue_reg_list[] = { + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG0S, regVCN_UE_ERR_STATUS_HI_JPEG0S), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG0S"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG0D, regVCN_UE_ERR_STATUS_HI_JPEG0D), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG0D"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG1S, regVCN_UE_ERR_STATUS_HI_JPEG1S), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG1S"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG1D, regVCN_UE_ERR_STATUS_HI_JPEG1D), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG1D"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG2S, regVCN_UE_ERR_STATUS_HI_JPEG2S), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG2S"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG2D, regVCN_UE_ERR_STATUS_HI_JPEG2D), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG2D"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG3S, regVCN_UE_ERR_STATUS_HI_JPEG3S), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG3S"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG3D, regVCN_UE_ERR_STATUS_HI_JPEG3D), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG3D"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG4S, regVCN_UE_ERR_STATUS_HI_JPEG4S), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG4S"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG4D, regVCN_UE_ERR_STATUS_HI_JPEG4D), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG4D"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG5S, regVCN_UE_ERR_STATUS_HI_JPEG5S), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG5S"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG5D, regVCN_UE_ERR_STATUS_HI_JPEG5D), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG5D"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG6S, regVCN_UE_ERR_STATUS_HI_JPEG6S), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG6S"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG6D, regVCN_UE_ERR_STATUS_HI_JPEG6D), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG6D"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG7S, regVCN_UE_ERR_STATUS_HI_JPEG7S), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG7S"}, + {AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG7D, regVCN_UE_ERR_STATUS_HI_JPEG7D), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG7D"}, +}; + +static void jpeg_v4_0_3_inst_query_ras_error_count(struct amdgpu_device *adev, + uint32_t jpeg_inst, + void *ras_err_status) +{ + struct ras_err_data *err_data = (struct ras_err_data *)ras_err_status; + + /* jpeg v4_0_3 only support uncorrectable errors */ + amdgpu_ras_inst_query_ras_error_count(adev, + jpeg_v4_0_3_ue_reg_list, + ARRAY_SIZE(jpeg_v4_0_3_ue_reg_list), + NULL, 0, GET_INST(VCN, jpeg_inst), + AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE, + &err_data->ue_count); +} + +static void jpeg_v4_0_3_query_ras_error_count(struct amdgpu_device *adev, + void *ras_err_status) +{ + uint32_t i; + + if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG)) { + dev_warn(adev->dev, "JPEG RAS is not supported\n"); + return; + } + + for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) + jpeg_v4_0_3_inst_query_ras_error_count(adev, i, ras_err_status); +} + +static void jpeg_v4_0_3_inst_reset_ras_error_count(struct amdgpu_device *adev, + uint32_t jpeg_inst) +{ + amdgpu_ras_inst_reset_ras_error_count(adev, + jpeg_v4_0_3_ue_reg_list, + ARRAY_SIZE(jpeg_v4_0_3_ue_reg_list), + GET_INST(VCN, jpeg_inst)); +} + +static void jpeg_v4_0_3_reset_ras_error_count(struct amdgpu_device *adev) +{ + uint32_t i; + + if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG)) { + dev_warn(adev->dev, "JPEG RAS is not supported\n"); + return; + } + + for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) + jpeg_v4_0_3_inst_reset_ras_error_count(adev, i); +} + +static const struct amdgpu_ras_block_hw_ops jpeg_v4_0_3_ras_hw_ops = { + .query_ras_error_count = jpeg_v4_0_3_query_ras_error_count, + .reset_ras_error_count = jpeg_v4_0_3_reset_ras_error_count, +}; + +static struct amdgpu_jpeg_ras jpeg_v4_0_3_ras = { + .ras_block = { + .hw_ops = &jpeg_v4_0_3_ras_hw_ops, + }, +}; + +static void jpeg_v4_0_3_set_ras_funcs(struct amdgpu_device *adev) +{ + adev->jpeg.ras = &jpeg_v4_0_3_ras; +} diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h new file mode 100644 index 000000000000..22483dc66351 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h @@ -0,0 +1,51 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef __JPEG_V4_0_3_H__ +#define __JPEG_V4_0_3_H__ + +#define regUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET 0x1bfff +#define regUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET 0x404d +#define regUVD_JPEG_GPCOM_DATA0_INTERNAL_OFFSET 0x404e +#define regUVD_JPEG_GPCOM_DATA1_INTERNAL_OFFSET 0x404f +#define regUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_INTERNAL_OFFSET 0x40ab +#define regUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x40ac +#define regUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET 0x40a4 +#define regUVD_LMI_JPEG_VMID_INTERNAL_OFFSET 0x40a6 +#define regUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET 0x40b6 +#define regUVD_LMI_JRBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x40b7 +#define regUVD_JRBC_IB_SIZE_INTERNAL_OFFSET 0x4082 +#define regUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_LOW_INTERNAL_OFFSET 0x42d4 +#define regUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x42d5 +#define regUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET 0x4085 +#define regUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET 0x4084 +#define regUVD_JRBC_STATUS_INTERNAL_OFFSET 0x4089 +#define regUVD_JPEG_PITCH_INTERNAL_OFFSET 0x4043 +#define regUVD_JRBC0_UVD_JRBC_SCRATCH0_INTERNAL_OFFSET 0x4094 +#define regUVD_JRBC_EXTERNAL_MCM_ADDR_INTERNAL_OFFSET 0x1bffe + +#define JRBC_DEC_EXTERNAL_REG_WRITE_ADDR 0x18000 + +extern const struct amdgpu_ip_block_version jpeg_v4_0_3_ip_block; + +#endif /* __JPEG_V4_0_3_H__ */ diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c b/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c index 2e2062636d5f..36a123e6c8ee 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c @@ -149,7 +149,7 @@ static int mes_v10_1_add_hw_queue(struct amdgpu_mes *mes, { struct amdgpu_device *adev = mes->adev; union MESAPI__ADD_QUEUE mes_add_queue_pkt; - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; uint32_t vm_cntx_cntl = hub->vm_cntx_cntl; memset(&mes_add_queue_pkt, 0, sizeof(mes_add_queue_pkt)); @@ -632,6 +632,8 @@ static int mes_v10_1_mqd_init(struct amdgpu_ring *ring) uint64_t hqd_gpu_addr, wb_gpu_addr, eop_base_addr; uint32_t tmp; + memset(mqd, 0, sizeof(*mqd)); + mqd->header = 0xC0310800; mqd->compute_pipelinestat_enable = 0x00000001; mqd->compute_static_thread_mgmt_se0 = 0xffffffff; @@ -728,6 +730,7 @@ static int mes_v10_1_mqd_init(struct amdgpu_ring *ring) /* offset: 184 - this is used for CP_HQD_GFX_CONTROL */ mqd->cp_hqd_suspend_cntl_stack_offset = tmp; + amdgpu_device_flush_hdp(ring->adev, NULL); return 0; } @@ -797,8 +800,8 @@ static void mes_v10_1_queue_init_register(struct amdgpu_ring *ring) static int mes_v10_1_kiq_enable_queue(struct amdgpu_device *adev) { - struct amdgpu_kiq *kiq = &adev->gfx.kiq; - struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; + struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring; int r; if (!kiq->pmf || !kiq->pmf->kiq_map_queues) @@ -812,13 +815,7 @@ static int mes_v10_1_kiq_enable_queue(struct amdgpu_device *adev) kiq->pmf->kiq_map_queues(kiq_ring, &adev->mes.ring); - r = amdgpu_ring_test_ring(kiq_ring); - if (r) { - DRM_ERROR("kfq enable failed\n"); - kiq_ring->sched.ready = false; - } - - return r; + return amdgpu_ring_test_helper(kiq_ring); } static int mes_v10_1_queue_init(struct amdgpu_device *adev) @@ -863,9 +860,9 @@ static int mes_v10_1_kiq_ring_init(struct amdgpu_device *adev) { struct amdgpu_ring *ring; - spin_lock_init(&adev->gfx.kiq.ring_lock); + spin_lock_init(&adev->gfx.kiq[0].ring_lock); - ring = &adev->gfx.kiq.ring; + ring = &adev->gfx.kiq[0].ring; ring->me = 3; ring->pipe = 1; @@ -891,7 +888,7 @@ static int mes_v10_1_mqd_sw_init(struct amdgpu_device *adev, struct amdgpu_ring *ring; if (pipe == AMDGPU_MES_KIQ_PIPE) - ring = &adev->gfx.kiq.ring; + ring = &adev->gfx.kiq[0].ring; else if (pipe == AMDGPU_MES_SCHED_PIPE) ring = &adev->mes.ring; else @@ -901,6 +898,7 @@ static int mes_v10_1_mqd_sw_init(struct amdgpu_device *adev, return 0; r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE, + AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT, &ring->mqd_obj, &ring->mqd_gpu_addr, &ring->mqd_ptr); if (r) { @@ -974,15 +972,15 @@ static int mes_v10_1_sw_fini(void *handle) amdgpu_ucode_release(&adev->mes.fw[pipe]); } - amdgpu_bo_free_kernel(&adev->gfx.kiq.ring.mqd_obj, - &adev->gfx.kiq.ring.mqd_gpu_addr, - &adev->gfx.kiq.ring.mqd_ptr); + amdgpu_bo_free_kernel(&adev->gfx.kiq[0].ring.mqd_obj, + &adev->gfx.kiq[0].ring.mqd_gpu_addr, + &adev->gfx.kiq[0].ring.mqd_ptr); amdgpu_bo_free_kernel(&adev->mes.ring.mqd_obj, &adev->mes.ring.mqd_gpu_addr, &adev->mes.ring.mqd_ptr); - amdgpu_ring_fini(&adev->gfx.kiq.ring); + amdgpu_ring_fini(&adev->gfx.kiq[0].ring); amdgpu_ring_fini(&adev->mes.ring); amdgpu_mes_fini(adev); @@ -1038,7 +1036,7 @@ static int mes_v10_1_kiq_hw_init(struct amdgpu_device *adev) mes_v10_1_enable(adev, true); - mes_v10_1_kiq_setting(&adev->gfx.kiq.ring); + mes_v10_1_kiq_setting(&adev->gfx.kiq[0].ring); r = mes_v10_1_queue_init(adev); if (r) @@ -1090,7 +1088,7 @@ static int mes_v10_1_hw_init(void *handle) * MES uses KIQ ring exclusively so driver cannot access KIQ ring * with MES enabled. */ - adev->gfx.kiq.ring.sched.ready = false; + adev->gfx.kiq[0].ring.sched.ready = false; adev->mes.ring.sched.ready = true; return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c index 45280f047180..1bdaa00c0b46 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c @@ -164,7 +164,7 @@ static int mes_v11_0_add_hw_queue(struct amdgpu_mes *mes, { struct amdgpu_device *adev = mes->adev; union MESAPI__ADD_QUEUE mes_add_queue_pkt; - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)]; uint32_t vm_cntx_cntl = hub->vm_cntx_cntl; memset(&mes_add_queue_pkt, 0, sizeof(mes_add_queue_pkt)); @@ -202,17 +202,14 @@ static int mes_v11_0_add_hw_queue(struct amdgpu_mes *mes, mes_add_queue_pkt.gws_size = input->gws_size; mes_add_queue_pkt.trap_handler_addr = input->tba_addr; mes_add_queue_pkt.tma_addr = input->tma_addr; + mes_add_queue_pkt.trap_en = input->trap_en; + mes_add_queue_pkt.skip_process_ctx_clear = input->skip_process_ctx_clear; mes_add_queue_pkt.is_kfd_process = input->is_kfd_process; /* For KFD, gds_size is re-used for queue size (needed in MES for AQL queues) */ mes_add_queue_pkt.is_aql_queue = input->is_aql_queue; mes_add_queue_pkt.gds_size = input->queue_size; - if (!(((adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) >= 4) && - (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(11, 0, 0)) && - (adev->ip_versions[GC_HWIP][0] <= IP_VERSION(11, 0, 3)))) - mes_add_queue_pkt.trap_en = 1; - /* For KFD, gds_size is re-used for queue size (needed in MES for AQL queues) */ mes_add_queue_pkt.is_aql_queue = input->is_aql_queue; mes_add_queue_pkt.gds_size = input->queue_size; @@ -339,6 +336,19 @@ static int mes_v11_0_misc_op(struct amdgpu_mes *mes, misc_pkt.wait_reg_mem.reg_offset1 = input->wrm_reg.reg0; misc_pkt.wait_reg_mem.reg_offset2 = input->wrm_reg.reg1; break; + case MES_MISC_OP_SET_SHADER_DEBUGGER: + misc_pkt.opcode = MESAPI_MISC__SET_SHADER_DEBUGGER; + misc_pkt.set_shader_debugger.process_context_addr = + input->set_shader_debugger.process_context_addr; + misc_pkt.set_shader_debugger.flags.u32all = + input->set_shader_debugger.flags.u32all; + misc_pkt.set_shader_debugger.spi_gdbg_per_vmid_cntl = + input->set_shader_debugger.spi_gdbg_per_vmid_cntl; + memcpy(misc_pkt.set_shader_debugger.tcp_watch_cntl, + input->set_shader_debugger.tcp_watch_cntl, + sizeof(misc_pkt.set_shader_debugger.tcp_watch_cntl)); + misc_pkt.set_shader_debugger.trap_en = input->set_shader_debugger.trap_en; + break; default: DRM_ERROR("unsupported misc op (%d) \n", input->op); return -EINVAL; @@ -704,6 +714,8 @@ static int mes_v11_0_mqd_init(struct amdgpu_ring *ring) uint64_t hqd_gpu_addr, wb_gpu_addr, eop_base_addr; uint32_t tmp; + memset(mqd, 0, sizeof(*mqd)); + mqd->header = 0xC0310800; mqd->compute_pipelinestat_enable = 0x00000001; mqd->compute_static_thread_mgmt_se0 = 0xffffffff; @@ -797,6 +809,7 @@ static int mes_v11_0_mqd_init(struct amdgpu_ring *ring) mqd->cp_hqd_iq_timer = regCP_HQD_IQ_TIMER_DEFAULT; mqd->cp_hqd_quantum = regCP_HQD_QUANTUM_DEFAULT; + amdgpu_device_flush_hdp(ring->adev, NULL); return 0; } @@ -864,8 +877,8 @@ static void mes_v11_0_queue_init_register(struct amdgpu_ring *ring) static int mes_v11_0_kiq_enable_queue(struct amdgpu_device *adev) { - struct amdgpu_kiq *kiq = &adev->gfx.kiq; - struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; + struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring; int r; if (!kiq->pmf || !kiq->pmf->kiq_map_queues) @@ -879,12 +892,7 @@ static int mes_v11_0_kiq_enable_queue(struct amdgpu_device *adev) kiq->pmf->kiq_map_queues(kiq_ring, &adev->mes.ring); - r = amdgpu_ring_test_ring(kiq_ring); - if (r) { - DRM_ERROR("kfq enable failed\n"); - kiq_ring->sched.ready = false; - } - return r; + return amdgpu_ring_test_helper(kiq_ring); } static int mes_v11_0_queue_init(struct amdgpu_device *adev, @@ -894,7 +902,7 @@ static int mes_v11_0_queue_init(struct amdgpu_device *adev, int r; if (pipe == AMDGPU_MES_KIQ_PIPE) - ring = &adev->gfx.kiq.ring; + ring = &adev->gfx.kiq[0].ring; else if (pipe == AMDGPU_MES_SCHED_PIPE) ring = &adev->mes.ring; else @@ -961,9 +969,9 @@ static int mes_v11_0_kiq_ring_init(struct amdgpu_device *adev) { struct amdgpu_ring *ring; - spin_lock_init(&adev->gfx.kiq.ring_lock); + spin_lock_init(&adev->gfx.kiq[0].ring_lock); - ring = &adev->gfx.kiq.ring; + ring = &adev->gfx.kiq[0].ring; ring->me = 3; ring->pipe = 1; @@ -989,7 +997,7 @@ static int mes_v11_0_mqd_sw_init(struct amdgpu_device *adev, struct amdgpu_ring *ring; if (pipe == AMDGPU_MES_KIQ_PIPE) - ring = &adev->gfx.kiq.ring; + ring = &adev->gfx.kiq[0].ring; else if (pipe == AMDGPU_MES_SCHED_PIPE) ring = &adev->mes.ring; else @@ -999,6 +1007,7 @@ static int mes_v11_0_mqd_sw_init(struct amdgpu_device *adev, return 0; r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE, + AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT, &ring->mqd_obj, &ring->mqd_gpu_addr, &ring->mqd_ptr); if (r) { @@ -1074,15 +1083,15 @@ static int mes_v11_0_sw_fini(void *handle) amdgpu_ucode_release(&adev->mes.fw[pipe]); } - amdgpu_bo_free_kernel(&adev->gfx.kiq.ring.mqd_obj, - &adev->gfx.kiq.ring.mqd_gpu_addr, - &adev->gfx.kiq.ring.mqd_ptr); + amdgpu_bo_free_kernel(&adev->gfx.kiq[0].ring.mqd_obj, + &adev->gfx.kiq[0].ring.mqd_gpu_addr, + &adev->gfx.kiq[0].ring.mqd_ptr); amdgpu_bo_free_kernel(&adev->mes.ring.mqd_obj, &adev->mes.ring.mqd_gpu_addr, &adev->mes.ring.mqd_ptr); - amdgpu_ring_fini(&adev->gfx.kiq.ring); + amdgpu_ring_fini(&adev->gfx.kiq[0].ring); amdgpu_ring_fini(&adev->mes.ring); if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) { @@ -1175,7 +1184,7 @@ static int mes_v11_0_kiq_hw_init(struct amdgpu_device *adev) mes_v11_0_enable(adev, true); - mes_v11_0_kiq_setting(&adev->gfx.kiq.ring); + mes_v11_0_kiq_setting(&adev->gfx.kiq[0].ring); r = mes_v11_0_queue_init(adev, AMDGPU_MES_KIQ_PIPE); if (r) @@ -1196,7 +1205,7 @@ static int mes_v11_0_kiq_hw_fini(struct amdgpu_device *adev) } if (amdgpu_sriov_vf(adev)) { - mes_v11_0_kiq_dequeue(&adev->gfx.kiq.ring); + mes_v11_0_kiq_dequeue(&adev->gfx.kiq[0].ring); mes_v11_0_kiq_clear(adev); } @@ -1244,7 +1253,7 @@ static int mes_v11_0_hw_init(void *handle) * MES uses KIQ ring exclusively so driver cannot access KIQ ring * with MES enabled. */ - adev->gfx.kiq.ring.sched.ready = false; + adev->gfx.kiq[0].ring.sched.ready = false; adev->mes.ring.sched.ready = true; return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c index 15e7cbeae75b..fb91b31056ca 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c @@ -54,7 +54,7 @@ static u64 mmhub_v1_0_get_fb_location(struct amdgpu_device *adev) static void mmhub_v1_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; WREG32_SOC15_OFFSET(MMHUB, 0, mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, hub->ctx_addr_distance * vmid, @@ -229,7 +229,7 @@ static void mmhub_v1_0_disable_identity_aperture(struct amdgpu_device *adev) static void mmhub_v1_0_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; unsigned num_level, block_size; uint32_t tmp; int i; @@ -285,7 +285,7 @@ static void mmhub_v1_0_setup_vmid_config(struct amdgpu_device *adev) static void mmhub_v1_0_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; unsigned i; for (i = 0; i < 18; ++i) { @@ -338,7 +338,7 @@ static int mmhub_v1_0_gart_enable(struct amdgpu_device *adev) static void mmhub_v1_0_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; u32 tmp; u32 i; @@ -415,7 +415,7 @@ static void mmhub_v1_0_set_fault_enable_default(struct amdgpu_device *adev, bool static void mmhub_v1_0_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(MMHUB, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c index 73afbf2facc9..9086f2fdfaf4 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c @@ -54,7 +54,7 @@ static u64 mmhub_v1_7_get_fb_location(struct amdgpu_device *adev) static void mmhub_v1_7_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; WREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, hub->ctx_addr_distance * vmid, lower_32_bits(page_table_base)); @@ -261,7 +261,7 @@ static void mmhub_v1_7_disable_identity_aperture(struct amdgpu_device *adev) static void mmhub_v1_7_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; unsigned num_level, block_size; uint32_t tmp; int i; @@ -319,7 +319,7 @@ static void mmhub_v1_7_setup_vmid_config(struct amdgpu_device *adev) static void mmhub_v1_7_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; unsigned i; for (i = 0; i < 18; ++i) { @@ -348,7 +348,7 @@ static int mmhub_v1_7_gart_enable(struct amdgpu_device *adev) static void mmhub_v1_7_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; u32 tmp; u32 i; @@ -425,7 +425,7 @@ static void mmhub_v1_7_set_fault_enable_default(struct amdgpu_device *adev, bool static void mmhub_v1_7_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(MMHUB, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c index 342d1702104c..5e8b493f8699 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c @@ -29,6 +29,7 @@ #include "soc15_common.h" #include "soc15.h" +#include "amdgpu_ras.h" #define regVM_L2_CNTL3_DEFAULT 0x80100007 #define regVM_L2_CNTL4_DEFAULT 0x000000c1 @@ -53,18 +54,30 @@ static u64 mmhub_v1_8_get_fb_location(struct amdgpu_device *adev) static void mmhub_v1_8_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; - - WREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, - hub->ctx_addr_distance * vmid, lower_32_bits(page_table_base)); + struct amdgpu_vmhub *hub; + u32 inst_mask; + int i; - WREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32, - hub->ctx_addr_distance * vmid, upper_32_bits(page_table_base)); + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) { + hub = &adev->vmhub[AMDGPU_MMHUB0(i)]; + WREG32_SOC15_OFFSET(MMHUB, i, + regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, + hub->ctx_addr_distance * vmid, + lower_32_bits(page_table_base)); + + WREG32_SOC15_OFFSET(MMHUB, i, + regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32, + hub->ctx_addr_distance * vmid, + upper_32_bits(page_table_base)); + } } static void mmhub_v1_8_init_gart_aperture_regs(struct amdgpu_device *adev) { uint64_t pt_base; + u32 inst_mask; + int i; if (adev->gmc.pdb0_bo) pt_base = amdgpu_gmc_pd_addr(adev->gmc.pdb0_bo); @@ -76,187 +89,248 @@ static void mmhub_v1_8_init_gart_aperture_regs(struct amdgpu_device *adev) /* If use GART for FB translation, vmid0 page table covers both * vram and system memory (gart) */ - if (adev->gmc.pdb0_bo) { - WREG32_SOC15(MMHUB, 0, regVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32, - (u32)(adev->gmc.fb_start >> 12)); - WREG32_SOC15(MMHUB, 0, regVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32, - (u32)(adev->gmc.fb_start >> 44)); - - WREG32_SOC15(MMHUB, 0, regVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32, - (u32)(adev->gmc.gart_end >> 12)); - WREG32_SOC15(MMHUB, 0, regVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32, - (u32)(adev->gmc.gart_end >> 44)); - - } else { - WREG32_SOC15(MMHUB, 0, regVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32, - (u32)(adev->gmc.gart_start >> 12)); - WREG32_SOC15(MMHUB, 0, regVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32, - (u32)(adev->gmc.gart_start >> 44)); - - WREG32_SOC15(MMHUB, 0, regVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32, - (u32)(adev->gmc.gart_end >> 12)); - WREG32_SOC15(MMHUB, 0, regVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32, - (u32)(adev->gmc.gart_end >> 44)); + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) { + if (adev->gmc.pdb0_bo) { + WREG32_SOC15(MMHUB, i, + regVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32, + (u32)(adev->gmc.fb_start >> 12)); + WREG32_SOC15(MMHUB, i, + regVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32, + (u32)(adev->gmc.fb_start >> 44)); + + WREG32_SOC15(MMHUB, i, + regVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32, + (u32)(adev->gmc.gart_end >> 12)); + WREG32_SOC15(MMHUB, i, + regVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32, + (u32)(adev->gmc.gart_end >> 44)); + + } else { + WREG32_SOC15(MMHUB, i, + regVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32, + (u32)(adev->gmc.gart_start >> 12)); + WREG32_SOC15(MMHUB, i, + regVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32, + (u32)(adev->gmc.gart_start >> 44)); + + WREG32_SOC15(MMHUB, i, + regVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32, + (u32)(adev->gmc.gart_end >> 12)); + WREG32_SOC15(MMHUB, i, + regVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32, + (u32)(adev->gmc.gart_end >> 44)); + } } } static void mmhub_v1_8_init_system_aperture_regs(struct amdgpu_device *adev) { + uint32_t tmp, inst_mask; uint64_t value; - uint32_t tmp; + int i; - /* Program the AGP BAR */ - WREG32_SOC15(MMHUB, 0, regMC_VM_AGP_BASE, 0); - WREG32_SOC15(MMHUB, 0, regMC_VM_AGP_BOT, adev->gmc.agp_start >> 24); - WREG32_SOC15(MMHUB, 0, regMC_VM_AGP_TOP, adev->gmc.agp_end >> 24); + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) { + /* Program the AGP BAR */ + WREG32_SOC15(MMHUB, i, regMC_VM_AGP_BASE, 0); + WREG32_SOC15(MMHUB, i, regMC_VM_AGP_BOT, + adev->gmc.agp_start >> 24); + WREG32_SOC15(MMHUB, i, regMC_VM_AGP_TOP, + adev->gmc.agp_end >> 24); - /* Program the system aperture low logical page number. */ - WREG32_SOC15(MMHUB, 0, regMC_VM_SYSTEM_APERTURE_LOW_ADDR, - min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18); + if (amdgpu_sriov_vf(adev)) + return; - WREG32_SOC15(MMHUB, 0, regMC_VM_SYSTEM_APERTURE_HIGH_ADDR, - max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18); + /* Program the system aperture low logical page number. */ + WREG32_SOC15(MMHUB, i, regMC_VM_SYSTEM_APERTURE_LOW_ADDR, + min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18); - /* In the case squeezing vram into GART aperture, we don't use - * FB aperture and AGP aperture. Disable them. - */ - if (adev->gmc.pdb0_bo) { - WREG32_SOC15(MMHUB, 0, regMC_VM_AGP_BOT, 0xFFFFFF); - WREG32_SOC15(MMHUB, 0, regMC_VM_AGP_TOP, 0); - WREG32_SOC15(MMHUB, 0, regMC_VM_FB_LOCATION_TOP, 0); - WREG32_SOC15(MMHUB, 0, regMC_VM_FB_LOCATION_BASE, 0x00FFFFFF); - WREG32_SOC15(MMHUB, 0, regMC_VM_SYSTEM_APERTURE_LOW_ADDR, 0x3FFFFFFF); - WREG32_SOC15(MMHUB, 0, regMC_VM_SYSTEM_APERTURE_HIGH_ADDR, 0); - } - if (amdgpu_sriov_vf(adev)) - return; + WREG32_SOC15(MMHUB, i, regMC_VM_SYSTEM_APERTURE_HIGH_ADDR, + max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18); - /* Set default page address. */ - value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr); - WREG32_SOC15(MMHUB, 0, regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB, - (u32)(value >> 12)); - WREG32_SOC15(MMHUB, 0, regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB, - (u32)(value >> 44)); - - /* Program "protection fault". */ - WREG32_SOC15(MMHUB, 0, regVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32, - (u32)(adev->dummy_page_addr >> 12)); - WREG32_SOC15(MMHUB, 0, regVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32, - (u32)((u64)adev->dummy_page_addr >> 44)); - - tmp = RREG32_SOC15(MMHUB, 0, regVM_L2_PROTECTION_FAULT_CNTL2); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL2, - ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1); - WREG32_SOC15(MMHUB, 0, regVM_L2_PROTECTION_FAULT_CNTL2, tmp); + /* In the case squeezing vram into GART aperture, we don't use + * FB aperture and AGP aperture. Disable them. + */ + if (adev->gmc.pdb0_bo) { + WREG32_SOC15(MMHUB, i, regMC_VM_AGP_BOT, 0xFFFFFF); + WREG32_SOC15(MMHUB, i, regMC_VM_AGP_TOP, 0); + WREG32_SOC15(MMHUB, i, regMC_VM_FB_LOCATION_TOP, 0); + WREG32_SOC15(MMHUB, i, regMC_VM_FB_LOCATION_BASE, + 0x00FFFFFF); + WREG32_SOC15(MMHUB, i, + regMC_VM_SYSTEM_APERTURE_LOW_ADDR, + 0x3FFFFFFF); + WREG32_SOC15(MMHUB, i, + regMC_VM_SYSTEM_APERTURE_HIGH_ADDR, 0); + } + + /* Set default page address. */ + value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr); + WREG32_SOC15(MMHUB, i, regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB, + (u32)(value >> 12)); + WREG32_SOC15(MMHUB, i, regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB, + (u32)(value >> 44)); + + /* Program "protection fault". */ + WREG32_SOC15(MMHUB, i, + regVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32, + (u32)(adev->dummy_page_addr >> 12)); + WREG32_SOC15(MMHUB, i, + regVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32, + (u32)((u64)adev->dummy_page_addr >> 44)); + + tmp = RREG32_SOC15(MMHUB, i, regVM_L2_PROTECTION_FAULT_CNTL2); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL2, + ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1); + WREG32_SOC15(MMHUB, i, regVM_L2_PROTECTION_FAULT_CNTL2, tmp); + } } static void mmhub_v1_8_init_tlb_regs(struct amdgpu_device *adev) { - uint32_t tmp; + uint32_t tmp, inst_mask; + int i; /* Setup TLB control */ - tmp = RREG32_SOC15(MMHUB, 0, regMC_VM_MX_L1_TLB_CNTL); - - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, - ENABLE_ADVANCED_DRIVER_MODEL, 1); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, - SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, - MTYPE, MTYPE_UC);/* XXX for emulation. */ - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1); - - WREG32_SOC15(MMHUB, 0, regMC_VM_MX_L1_TLB_CNTL, tmp); + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) { + tmp = RREG32_SOC15(MMHUB, i, regMC_VM_MX_L1_TLB_CNTL); + + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, + 1); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + SYSTEM_ACCESS_MODE, 3); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + ENABLE_ADVANCED_DRIVER_MODEL, 1); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + SYSTEM_APERTURE_UNMAPPED_ACCESS, 0); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + MTYPE, MTYPE_UC);/* XXX for emulation. */ + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1); + + WREG32_SOC15(MMHUB, i, regMC_VM_MX_L1_TLB_CNTL, tmp); + } } static void mmhub_v1_8_init_cache_regs(struct amdgpu_device *adev) { - uint32_t tmp; + uint32_t tmp, inst_mask; + int i; if (amdgpu_sriov_vf(adev)) return; /* Setup L2 cache */ - tmp = RREG32_SOC15(MMHUB, 0, regVM_L2_CNTL); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 1); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 1); - /* XXX for emulation, Refer to closed source code.*/ - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, L2_PDE0_CACHE_TAG_GENERATION_MODE, - 0); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0); - WREG32_SOC15(MMHUB, 0, regVM_L2_CNTL, tmp); - - tmp = RREG32_SOC15(MMHUB, 0, regVM_L2_CNTL2); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_L2_CACHE, 1); - WREG32_SOC15(MMHUB, 0, regVM_L2_CNTL2, tmp); - - tmp = regVM_L2_CNTL3_DEFAULT; - if (adev->gmc.translate_further) { - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 12); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, - L2_CACHE_BIGK_FRAGMENT_SIZE, 9); - } else { - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 9); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, - L2_CACHE_BIGK_FRAGMENT_SIZE, 6); - } - WREG32_SOC15(MMHUB, 0, regVM_L2_CNTL3, tmp); - - tmp = regVM_L2_CNTL4_DEFAULT; - if (adev->gmc.xgmi.connected_to_cpu) { - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, - VMC_TAP_PDE_REQUEST_PHYSICAL, 1); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, - VMC_TAP_PTE_REQUEST_PHYSICAL, 1); - } else { - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, - VMC_TAP_PDE_REQUEST_PHYSICAL, 0); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, - VMC_TAP_PTE_REQUEST_PHYSICAL, 0); + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) { + tmp = RREG32_SOC15(MMHUB, i, regVM_L2_CNTL); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 1); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, + ENABLE_L2_FRAGMENT_PROCESSING, 1); + /* XXX for emulation, Refer to closed source code.*/ + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, + L2_PDE0_CACHE_TAG_GENERATION_MODE, 0); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, PDE_FAULT_CLASSIFICATION, + 0); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, + CONTEXT1_IDENTITY_ACCESS_MODE, 1); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, + IDENTITY_MODE_FRAGMENT_SIZE, 0); + WREG32_SOC15(MMHUB, i, regVM_L2_CNTL, tmp); + + tmp = RREG32_SOC15(MMHUB, i, regVM_L2_CNTL2); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, + 1); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_L2_CACHE, 1); + WREG32_SOC15(MMHUB, i, regVM_L2_CNTL2, tmp); + + tmp = regVM_L2_CNTL3_DEFAULT; + if (adev->gmc.translate_further) { + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 12); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, + L2_CACHE_BIGK_FRAGMENT_SIZE, 9); + } else { + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 9); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, + L2_CACHE_BIGK_FRAGMENT_SIZE, 6); + } + WREG32_SOC15(MMHUB, i, regVM_L2_CNTL3, tmp); + + tmp = regVM_L2_CNTL4_DEFAULT; + /* For AMD APP APUs setup WC memory */ + if (adev->gmc.xgmi.connected_to_cpu || adev->gmc.is_app_apu) { + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, + VMC_TAP_PDE_REQUEST_PHYSICAL, 1); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, + VMC_TAP_PTE_REQUEST_PHYSICAL, 1); + } else { + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, + VMC_TAP_PDE_REQUEST_PHYSICAL, 0); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, + VMC_TAP_PTE_REQUEST_PHYSICAL, 0); + } + WREG32_SOC15(MMHUB, i, regVM_L2_CNTL4, tmp); } - WREG32_SOC15(MMHUB, 0, regVM_L2_CNTL4, tmp); } static void mmhub_v1_8_enable_system_domain(struct amdgpu_device *adev) { - uint32_t tmp; - - tmp = RREG32_SOC15(MMHUB, 0, regVM_CONTEXT0_CNTL); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, - adev->gmc.vmid0_page_table_depth); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, PAGE_TABLE_BLOCK_SIZE, - adev->gmc.vmid0_page_table_block_size); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, - RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0); - WREG32_SOC15(MMHUB, 0, regVM_CONTEXT0_CNTL, tmp); + uint32_t tmp, inst_mask; + int i; + + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) { + tmp = RREG32_SOC15(MMHUB, i, regVM_CONTEXT0_CNTL); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, + adev->gmc.vmid0_page_table_depth); + tmp = REG_SET_FIELD(tmp, + VM_CONTEXT0_CNTL, PAGE_TABLE_BLOCK_SIZE, + adev->gmc.vmid0_page_table_block_size); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, + RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0); + WREG32_SOC15(MMHUB, i, regVM_CONTEXT0_CNTL, tmp); + } } static void mmhub_v1_8_disable_identity_aperture(struct amdgpu_device *adev) { + u32 inst_mask; + int i; + if (amdgpu_sriov_vf(adev)) return; - WREG32_SOC15(MMHUB, 0, regVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32, 0xFFFFFFFF); - WREG32_SOC15(MMHUB, 0, regVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32, 0x0000000F); - - WREG32_SOC15(MMHUB, 0, regVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32, 0); - WREG32_SOC15(MMHUB, 0, regVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32, 0); - - WREG32_SOC15(MMHUB, 0, regVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32, 0); - WREG32_SOC15(MMHUB, 0, regVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32, 0); + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) { + WREG32_SOC15(MMHUB, i, + regVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32, + 0XFFFFFFFF); + WREG32_SOC15(MMHUB, i, + regVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32, + 0x0000000F); + + WREG32_SOC15(MMHUB, i, + regVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32, + 0); + WREG32_SOC15(MMHUB, i, + regVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32, + 0); + + WREG32_SOC15(MMHUB, i, + regVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32, 0); + WREG32_SOC15(MMHUB, i, + regVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32, 0); + } } static void mmhub_v1_8_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; - unsigned num_level, block_size; - uint32_t tmp; - int i; + struct amdgpu_vmhub *hub; + unsigned int num_level, block_size; + uint32_t tmp, inst_mask; + int i, j; num_level = adev->vm_manager.num_level; block_size = adev->vm_manager.block_size; @@ -265,60 +339,75 @@ static void mmhub_v1_8_setup_vmid_config(struct amdgpu_device *adev) else block_size -= 9; - for (i = 0; i <= 14; i++) { - tmp = RREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT1_CNTL, i); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH, - num_level); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, - 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - PAGE_TABLE_BLOCK_SIZE, - block_size); - /* On Aldebaran, XNACK can be enabled in the SQ per-process. - * Retry faults need to be enabled for that to work. - */ - tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, - RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, - 1); - WREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT1_CNTL, - i * hub->ctx_distance, tmp); - WREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32, - i * hub->ctx_addr_distance, 0); - WREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32, - i * hub->ctx_addr_distance, 0); - WREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32, - i * hub->ctx_addr_distance, - lower_32_bits(adev->vm_manager.max_pfn - 1)); - WREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32, - i * hub->ctx_addr_distance, - upper_32_bits(adev->vm_manager.max_pfn - 1)); + inst_mask = adev->aid_mask; + for_each_inst(j, inst_mask) { + hub = &adev->vmhub[AMDGPU_MMHUB0(j)]; + for (i = 0; i <= 14; i++) { + tmp = RREG32_SOC15_OFFSET(MMHUB, j, regVM_CONTEXT1_CNTL, + i); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + ENABLE_CONTEXT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + PAGE_TABLE_DEPTH, num_level); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + PAGE_TABLE_BLOCK_SIZE, + block_size); + /* On 9.4.3, XNACK can be enabled in the SQ + * per-process. Retry faults need to be enabled for + * that to work. + */ + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, + RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 1); + WREG32_SOC15_OFFSET(MMHUB, j, regVM_CONTEXT1_CNTL, + i * hub->ctx_distance, tmp); + WREG32_SOC15_OFFSET(MMHUB, j, + regVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32, + i * hub->ctx_addr_distance, 0); + WREG32_SOC15_OFFSET(MMHUB, j, + regVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32, + i * hub->ctx_addr_distance, 0); + WREG32_SOC15_OFFSET(MMHUB, j, + regVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32, + i * hub->ctx_addr_distance, + lower_32_bits(adev->vm_manager.max_pfn - 1)); + WREG32_SOC15_OFFSET(MMHUB, j, + regVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32, + i * hub->ctx_addr_distance, + upper_32_bits(adev->vm_manager.max_pfn - 1)); + } } } static void mmhub_v1_8_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; - unsigned i; - - for (i = 0; i < 18; ++i) { - WREG32_SOC15_OFFSET(MMHUB, 0, regVM_INVALIDATE_ENG0_ADDR_RANGE_LO32, - i * hub->eng_addr_distance, 0xffffffff); - WREG32_SOC15_OFFSET(MMHUB, 0, regVM_INVALIDATE_ENG0_ADDR_RANGE_HI32, - i * hub->eng_addr_distance, 0x1f); + struct amdgpu_vmhub *hub; + u32 i, j, inst_mask; + + inst_mask = adev->aid_mask; + for_each_inst(j, inst_mask) { + hub = &adev->vmhub[AMDGPU_MMHUB0(j)]; + for (i = 0; i < 18; ++i) { + WREG32_SOC15_OFFSET(MMHUB, j, + regVM_INVALIDATE_ENG0_ADDR_RANGE_LO32, + i * hub->eng_addr_distance, 0xffffffff); + WREG32_SOC15_OFFSET(MMHUB, j, + regVM_INVALIDATE_ENG0_ADDR_RANGE_HI32, + i * hub->eng_addr_distance, 0x1f); + } } } @@ -352,28 +441,34 @@ static int mmhub_v1_8_gart_enable(struct amdgpu_device *adev) static void mmhub_v1_8_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub; u32 tmp; - u32 i; + u32 i, j, inst_mask; /* Disable all tables */ - for (i = 0; i < 16; i++) - WREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT0_CNTL, - i * hub->ctx_distance, 0); - - /* Setup TLB control */ - tmp = RREG32_SOC15(MMHUB, 0, regMC_VM_MX_L1_TLB_CNTL); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0); - tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, - ENABLE_ADVANCED_DRIVER_MODEL, 0); - WREG32_SOC15(MMHUB, 0, regMC_VM_MX_L1_TLB_CNTL, tmp); - - if (!amdgpu_sriov_vf(adev)) { - /* Setup L2 cache */ - tmp = RREG32_SOC15(MMHUB, 0, regVM_L2_CNTL); - tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 0); - WREG32_SOC15(MMHUB, 0, regVM_L2_CNTL, tmp); - WREG32_SOC15(MMHUB, 0, regVM_L2_CNTL3, 0); + inst_mask = adev->aid_mask; + for_each_inst(j, inst_mask) { + hub = &adev->vmhub[AMDGPU_MMHUB0(j)]; + for (i = 0; i < 16; i++) + WREG32_SOC15_OFFSET(MMHUB, j, regVM_CONTEXT0_CNTL, + i * hub->ctx_distance, 0); + + /* Setup TLB control */ + tmp = RREG32_SOC15(MMHUB, j, regMC_VM_MX_L1_TLB_CNTL); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, + 0); + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, + ENABLE_ADVANCED_DRIVER_MODEL, 0); + WREG32_SOC15(MMHUB, j, regMC_VM_MX_L1_TLB_CNTL, tmp); + + if (!amdgpu_sriov_vf(adev)) { + /* Setup L2 cache */ + tmp = RREG32_SOC15(MMHUB, j, regVM_L2_CNTL); + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, + 0); + WREG32_SOC15(MMHUB, j, regVM_L2_CNTL, tmp); + WREG32_SOC15(MMHUB, j, regVM_L2_CNTL3, 0); + } } } @@ -385,73 +480,83 @@ static void mmhub_v1_8_gart_disable(struct amdgpu_device *adev) */ static void mmhub_v1_8_set_fault_enable_default(struct amdgpu_device *adev, bool value) { - u32 tmp; + u32 tmp, inst_mask; + int i; if (amdgpu_sriov_vf(adev)) return; - tmp = RREG32_SOC15(MMHUB, 0, regVM_L2_PROTECTION_FAULT_CNTL); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT, - value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - READ_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value); - tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value); - if (!value) { + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) { + tmp = RREG32_SOC15(MMHUB, i, regVM_L2_PROTECTION_FAULT_CNTL); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT, + value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value); tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - CRASH_ON_NO_RETRY_FAULT, 1); + DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value); tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, - CRASH_ON_RETRY_FAULT, 1); + VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + READ_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value); + if (!value) { + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + CRASH_ON_NO_RETRY_FAULT, 1); + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL, + CRASH_ON_RETRY_FAULT, 1); + } + + WREG32_SOC15(MMHUB, i, regVM_L2_PROTECTION_FAULT_CNTL, tmp); } - - WREG32_SOC15(MMHUB, 0, regVM_L2_PROTECTION_FAULT_CNTL, tmp); } static void mmhub_v1_8_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; - - hub->ctx0_ptb_addr_lo32 = - SOC15_REG_OFFSET(MMHUB, 0, - regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32); - hub->ctx0_ptb_addr_hi32 = - SOC15_REG_OFFSET(MMHUB, 0, - regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32); - hub->vm_inv_eng0_req = - SOC15_REG_OFFSET(MMHUB, 0, regVM_INVALIDATE_ENG0_REQ); - hub->vm_inv_eng0_ack = - SOC15_REG_OFFSET(MMHUB, 0, regVM_INVALIDATE_ENG0_ACK); - hub->vm_context0_cntl = - SOC15_REG_OFFSET(MMHUB, 0, regVM_CONTEXT0_CNTL); - hub->vm_l2_pro_fault_status = - SOC15_REG_OFFSET(MMHUB, 0, regVM_L2_PROTECTION_FAULT_STATUS); - hub->vm_l2_pro_fault_cntl = - SOC15_REG_OFFSET(MMHUB, 0, regVM_L2_PROTECTION_FAULT_CNTL); - - hub->ctx_distance = regVM_CONTEXT1_CNTL - regVM_CONTEXT0_CNTL; - hub->ctx_addr_distance = regVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 - - regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32; - hub->eng_distance = regVM_INVALIDATE_ENG1_REQ - regVM_INVALIDATE_ENG0_REQ; - hub->eng_addr_distance = regVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 - - regVM_INVALIDATE_ENG0_ADDR_RANGE_LO32; + struct amdgpu_vmhub *hub; + u32 inst_mask; + int i; + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) { + hub = &adev->vmhub[AMDGPU_MMHUB0(i)]; + + hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(MMHUB, i, + regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32); + hub->ctx0_ptb_addr_hi32 = SOC15_REG_OFFSET(MMHUB, i, + regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32); + hub->vm_inv_eng0_req = + SOC15_REG_OFFSET(MMHUB, i, regVM_INVALIDATE_ENG0_REQ); + hub->vm_inv_eng0_ack = + SOC15_REG_OFFSET(MMHUB, i, regVM_INVALIDATE_ENG0_ACK); + hub->vm_context0_cntl = + SOC15_REG_OFFSET(MMHUB, i, regVM_CONTEXT0_CNTL); + hub->vm_l2_pro_fault_status = SOC15_REG_OFFSET(MMHUB, i, + regVM_L2_PROTECTION_FAULT_STATUS); + hub->vm_l2_pro_fault_cntl = SOC15_REG_OFFSET(MMHUB, i, + regVM_L2_PROTECTION_FAULT_CNTL); + + hub->ctx_distance = regVM_CONTEXT1_CNTL - regVM_CONTEXT0_CNTL; + hub->ctx_addr_distance = + regVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 - + regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32; + hub->eng_distance = regVM_INVALIDATE_ENG1_REQ - + regVM_INVALIDATE_ENG0_REQ; + hub->eng_addr_distance = regVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 - + regVM_INVALIDATE_ENG0_ADDR_RANGE_LO32; + } } static int mmhub_v1_8_set_clockgating(struct amdgpu_device *adev, @@ -475,3 +580,277 @@ const struct amdgpu_mmhub_funcs mmhub_v1_8_funcs = { .set_clockgating = mmhub_v1_8_set_clockgating, .get_clockgating = mmhub_v1_8_get_clockgating, }; + +static const struct amdgpu_ras_err_status_reg_entry mmhub_v1_8_ce_reg_list[] = { + {AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA0_CE_ERR_STATUS_LO, regMMEA0_CE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA0"}, + {AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA1_CE_ERR_STATUS_LO, regMMEA1_CE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA1"}, + {AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA2_CE_ERR_STATUS_LO, regMMEA2_CE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA2"}, + {AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA3_CE_ERR_STATUS_LO, regMMEA3_CE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA3"}, + {AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA4_CE_ERR_STATUS_LO, regMMEA4_CE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA4"}, + {AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMM_CANE_CE_ERR_STATUS_LO, regMM_CANE_CE_ERR_STATUS_HI), + 1, 0, "MM_CANE"}, +}; + +static const struct amdgpu_ras_err_status_reg_entry mmhub_v1_8_ue_reg_list[] = { + {AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA0_UE_ERR_STATUS_LO, regMMEA0_UE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA0"}, + {AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA1_UE_ERR_STATUS_LO, regMMEA1_UE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA1"}, + {AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA2_UE_ERR_STATUS_LO, regMMEA2_UE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA2"}, + {AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA3_UE_ERR_STATUS_LO, regMMEA3_UE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA3"}, + {AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA4_UE_ERR_STATUS_LO, regMMEA4_UE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA4"}, + {AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMM_CANE_UE_ERR_STATUS_LO, regMM_CANE_UE_ERR_STATUS_HI), + 1, 0, "MM_CANE"}, +}; + +static const struct amdgpu_ras_memory_id_entry mmhub_v1_8_ras_memory_list[] = { + {AMDGPU_MMHUB_WGMI_PAGEMEM, "MMEA_WGMI_PAGEMEM"}, + {AMDGPU_MMHUB_RGMI_PAGEMEM, "MMEA_RGMI_PAGEMEM"}, + {AMDGPU_MMHUB_WDRAM_PAGEMEM, "MMEA_WDRAM_PAGEMEM"}, + {AMDGPU_MMHUB_RDRAM_PAGEMEM, "MMEA_RDRAM_PAGEMEM"}, + {AMDGPU_MMHUB_WIO_CMDMEM, "MMEA_WIO_CMDMEM"}, + {AMDGPU_MMHUB_RIO_CMDMEM, "MMEA_RIO_CMDMEM"}, + {AMDGPU_MMHUB_WGMI_CMDMEM, "MMEA_WGMI_CMDMEM"}, + {AMDGPU_MMHUB_RGMI_CMDMEM, "MMEA_RGMI_CMDMEM"}, + {AMDGPU_MMHUB_WDRAM_CMDMEM, "MMEA_WDRAM_CMDMEM"}, + {AMDGPU_MMHUB_RDRAM_CMDMEM, "MMEA_RDRAM_CMDMEM"}, + {AMDGPU_MMHUB_MAM_DMEM0, "MMEA_MAM_DMEM0"}, + {AMDGPU_MMHUB_MAM_DMEM1, "MMEA_MAM_DMEM1"}, + {AMDGPU_MMHUB_MAM_DMEM2, "MMEA_MAM_DMEM2"}, + {AMDGPU_MMHUB_MAM_DMEM3, "MMEA_MAM_DMEM3"}, + {AMDGPU_MMHUB_WRET_TAGMEM, "MMEA_WRET_TAGMEM"}, + {AMDGPU_MMHUB_RRET_TAGMEM, "MMEA_RRET_TAGMEM"}, + {AMDGPU_MMHUB_WIO_DATAMEM, "MMEA_WIO_DATAMEM"}, + {AMDGPU_MMHUB_WGMI_DATAMEM, "MMEA_WGMI_DATAMEM"}, + {AMDGPU_MMHUB_WDRAM_DATAMEM, "MMEA_WDRAM_DATAMEM"}, +}; + +static void mmhub_v1_8_inst_query_ras_error_count(struct amdgpu_device *adev, + uint32_t mmhub_inst, + void *ras_err_status) +{ + struct ras_err_data *err_data = (struct ras_err_data *)ras_err_status; + + amdgpu_ras_inst_query_ras_error_count(adev, + mmhub_v1_8_ce_reg_list, + ARRAY_SIZE(mmhub_v1_8_ce_reg_list), + mmhub_v1_8_ras_memory_list, + ARRAY_SIZE(mmhub_v1_8_ras_memory_list), + mmhub_inst, + AMDGPU_RAS_ERROR__SINGLE_CORRECTABLE, + &err_data->ce_count); + amdgpu_ras_inst_query_ras_error_count(adev, + mmhub_v1_8_ue_reg_list, + ARRAY_SIZE(mmhub_v1_8_ue_reg_list), + mmhub_v1_8_ras_memory_list, + ARRAY_SIZE(mmhub_v1_8_ras_memory_list), + mmhub_inst, + AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE, + &err_data->ue_count); +} + +static void mmhub_v1_8_query_ras_error_count(struct amdgpu_device *adev, + void *ras_err_status) +{ + uint32_t inst_mask; + uint32_t i; + + if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__MMHUB)) { + dev_warn(adev->dev, "MMHUB RAS is not supported\n"); + return; + } + + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) + mmhub_v1_8_inst_query_ras_error_count(adev, i, ras_err_status); +} + +static void mmhub_v1_8_inst_reset_ras_error_count(struct amdgpu_device *adev, + uint32_t mmhub_inst) +{ + amdgpu_ras_inst_reset_ras_error_count(adev, + mmhub_v1_8_ce_reg_list, + ARRAY_SIZE(mmhub_v1_8_ce_reg_list), + mmhub_inst); + amdgpu_ras_inst_reset_ras_error_count(adev, + mmhub_v1_8_ue_reg_list, + ARRAY_SIZE(mmhub_v1_8_ue_reg_list), + mmhub_inst); +} + +static void mmhub_v1_8_reset_ras_error_count(struct amdgpu_device *adev) +{ + uint32_t inst_mask; + uint32_t i; + + if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__MMHUB)) { + dev_warn(adev->dev, "MMHUB RAS is not supported\n"); + return; + } + + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) + mmhub_v1_8_inst_reset_ras_error_count(adev, i); +} + +static const u32 mmhub_v1_8_mmea_err_status_reg[] __maybe_unused = { + regMMEA0_ERR_STATUS, + regMMEA1_ERR_STATUS, + regMMEA2_ERR_STATUS, + regMMEA3_ERR_STATUS, + regMMEA4_ERR_STATUS, +}; + +static void mmhub_v1_8_inst_query_ras_err_status(struct amdgpu_device *adev, + uint32_t mmhub_inst) +{ + uint32_t reg_value; + uint32_t mmea_err_status_addr_dist; + uint32_t i; + + /* query mmea ras err status */ + mmea_err_status_addr_dist = regMMEA1_ERR_STATUS - regMMEA0_ERR_STATUS; + for (i = 0; i < ARRAY_SIZE(mmhub_v1_8_mmea_err_status_reg); i++) { + reg_value = RREG32_SOC15_OFFSET(MMHUB, mmhub_inst, + regMMEA0_ERR_STATUS, + i * mmea_err_status_addr_dist); + if (REG_GET_FIELD(reg_value, MMEA0_ERR_STATUS, SDP_RDRSP_STATUS) || + REG_GET_FIELD(reg_value, MMEA0_ERR_STATUS, SDP_WRRSP_STATUS) || + REG_GET_FIELD(reg_value, MMEA0_ERR_STATUS, SDP_RDRSP_DATAPARITY_ERROR)) { + dev_warn(adev->dev, + "Detected MMEA%d err in MMHUB%d, status: 0x%x\n", + i, mmhub_inst, reg_value); + } + } + + /* query mm_cane ras err status */ + reg_value = RREG32_SOC15(MMHUB, mmhub_inst, regMM_CANE_ERR_STATUS); + if (REG_GET_FIELD(reg_value, MM_CANE_ERR_STATUS, SDPM_RDRSP_STATUS) || + REG_GET_FIELD(reg_value, MM_CANE_ERR_STATUS, SDPM_WRRSP_STATUS) || + REG_GET_FIELD(reg_value, MM_CANE_ERR_STATUS, SDPM_RDRSP_DATAPARITY_ERROR)) { + dev_warn(adev->dev, + "Detected MM CANE err in MMHUB%d, status: 0x%x\n", + mmhub_inst, reg_value); + } +} + +static void mmhub_v1_8_query_ras_error_status(struct amdgpu_device *adev) +{ + uint32_t inst_mask; + uint32_t i; + + if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__MMHUB)) { + dev_warn(adev->dev, "MMHUB RAS is not supported\n"); + return; + } + + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) + mmhub_v1_8_inst_query_ras_err_status(adev, i); +} + +static void mmhub_v1_8_inst_reset_ras_err_status(struct amdgpu_device *adev, + uint32_t mmhub_inst) +{ + uint32_t mmea_cgtt_clk_cntl_addr_dist; + uint32_t mmea_err_status_addr_dist; + uint32_t reg_value; + uint32_t i; + + /* reset mmea ras err status */ + mmea_cgtt_clk_cntl_addr_dist = regMMEA1_CGTT_CLK_CTRL - regMMEA0_CGTT_CLK_CTRL; + mmea_err_status_addr_dist = regMMEA1_ERR_STATUS - regMMEA0_ERR_STATUS; + for (i = 0; i < ARRAY_SIZE(mmhub_v1_8_mmea_err_status_reg); i++) { + /* force clk branch on for response path + * set MMEA0_CGTT_CLK_CTRL.SOFT_OVERRIDE_RETURN = 1 + */ + reg_value = RREG32_SOC15_OFFSET(MMHUB, mmhub_inst, + regMMEA0_CGTT_CLK_CTRL, + i * mmea_cgtt_clk_cntl_addr_dist); + reg_value = REG_SET_FIELD(reg_value, MMEA0_CGTT_CLK_CTRL, + SOFT_OVERRIDE_RETURN, 1); + WREG32_SOC15_OFFSET(MMHUB, mmhub_inst, + regMMEA0_CGTT_CLK_CTRL, + i * mmea_cgtt_clk_cntl_addr_dist, + reg_value); + + /* set MMEA0_ERR_STATUS.CLEAR_ERROR_STATUS = 1 */ + reg_value = RREG32_SOC15_OFFSET(MMHUB, mmhub_inst, + regMMEA0_ERR_STATUS, + i * mmea_err_status_addr_dist); + reg_value = REG_SET_FIELD(reg_value, MMEA0_ERR_STATUS, + CLEAR_ERROR_STATUS, 1); + WREG32_SOC15_OFFSET(MMHUB, mmhub_inst, + regMMEA0_ERR_STATUS, + i * mmea_err_status_addr_dist, + reg_value); + + /* set MMEA0_CGTT_CLK_CTRL.SOFT_OVERRIDE_RETURN = 0 */ + reg_value = RREG32_SOC15_OFFSET(MMHUB, mmhub_inst, + regMMEA0_CGTT_CLK_CTRL, + i * mmea_cgtt_clk_cntl_addr_dist); + reg_value = REG_SET_FIELD(reg_value, MMEA0_CGTT_CLK_CTRL, + SOFT_OVERRIDE_RETURN, 0); + WREG32_SOC15_OFFSET(MMHUB, mmhub_inst, + regMMEA0_CGTT_CLK_CTRL, + i * mmea_cgtt_clk_cntl_addr_dist, + reg_value); + } + + /* reset mm_cane ras err status + * force clk branch on for response path + * set MM_CANE_ICG_CTRL.SOFT_OVERRIDE_ATRET = 1 + */ + reg_value = RREG32_SOC15(MMHUB, mmhub_inst, regMM_CANE_ICG_CTRL); + reg_value = REG_SET_FIELD(reg_value, MM_CANE_ICG_CTRL, + SOFT_OVERRIDE_ATRET, 1); + WREG32_SOC15(MMHUB, mmhub_inst, regMM_CANE_ICG_CTRL, reg_value); + + /* set MM_CANE_ERR_STATUS.CLEAR_ERROR_STATUS = 1 */ + reg_value = RREG32_SOC15(MMHUB, mmhub_inst, regMM_CANE_ERR_STATUS); + reg_value = REG_SET_FIELD(reg_value, MM_CANE_ERR_STATUS, + CLEAR_ERROR_STATUS, 1); + WREG32_SOC15(MMHUB, mmhub_inst, regMM_CANE_ERR_STATUS, reg_value); + + /* set MM_CANE_ICG_CTRL.SOFT_OVERRIDE_ATRET = 0 */ + reg_value = RREG32_SOC15(MMHUB, mmhub_inst, regMM_CANE_ICG_CTRL); + reg_value = REG_SET_FIELD(reg_value, MM_CANE_ICG_CTRL, + SOFT_OVERRIDE_ATRET, 0); + WREG32_SOC15(MMHUB, mmhub_inst, regMM_CANE_ICG_CTRL, reg_value); +} + +static void mmhub_v1_8_reset_ras_error_status(struct amdgpu_device *adev) +{ + uint32_t inst_mask; + uint32_t i; + + if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__MMHUB)) { + dev_warn(adev->dev, "MMHUB RAS is not supported\n"); + return; + } + + inst_mask = adev->aid_mask; + for_each_inst(i, inst_mask) + mmhub_v1_8_inst_reset_ras_err_status(adev, i); +} + +static const struct amdgpu_ras_block_hw_ops mmhub_v1_8_ras_hw_ops = { + .query_ras_error_count = mmhub_v1_8_query_ras_error_count, + .reset_ras_error_count = mmhub_v1_8_reset_ras_error_count, + .query_ras_error_status = mmhub_v1_8_query_ras_error_status, + .reset_ras_error_status = mmhub_v1_8_reset_ras_error_status, +}; + +struct amdgpu_mmhub_ras mmhub_v1_8_ras = { + .ras_block = { + .hw_ops = &mmhub_v1_8_ras_hw_ops, + }, +}; diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.h b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.h index 0bb36200e4e5..126f0075ac50 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.h +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.h @@ -24,5 +24,6 @@ #define __MMHUB_V1_8_H__ extern const struct amdgpu_mmhub_funcs mmhub_v1_8_funcs; +extern struct amdgpu_mmhub_ras mmhub_v1_8_ras; #endif diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c index 278e32db878d..8f76c6ecf50a 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c @@ -187,7 +187,7 @@ mmhub_v2_0_print_l2_protection_fault_status(struct amdgpu_device *adev, static void mmhub_v2_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; WREG32_SOC15_OFFSET_RLC(MMHUB, 0, mmMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, hub->ctx_addr_distance * vmid, @@ -362,7 +362,7 @@ static void mmhub_v2_0_disable_identity_aperture(struct amdgpu_device *adev) static void mmhub_v2_0_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; int i; uint32_t tmp; @@ -412,7 +412,7 @@ static void mmhub_v2_0_setup_vmid_config(struct amdgpu_device *adev) static void mmhub_v2_0_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; unsigned i; for (i = 0; i < 18; ++i) { @@ -441,7 +441,7 @@ static int mmhub_v2_0_gart_enable(struct amdgpu_device *adev) static void mmhub_v2_0_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; u32 tmp; u32 i; @@ -520,7 +520,7 @@ static const struct amdgpu_vmhub_funcs mmhub_v2_0_vmhub_funcs = { static void mmhub_v2_0_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(MMHUB, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c index fcf2813e70db..8bd0fc8d9d25 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c @@ -121,7 +121,7 @@ static void mmhub_v2_3_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, hub->ctx_addr_distance * vmid, lower_32_bits(page_table_base)); @@ -280,7 +280,7 @@ static void mmhub_v2_3_disable_identity_aperture(struct amdgpu_device *adev) static void mmhub_v2_3_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; int i; uint32_t tmp; @@ -330,7 +330,7 @@ static void mmhub_v2_3_setup_vmid_config(struct amdgpu_device *adev) static void mmhub_v2_3_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; unsigned i; for (i = 0; i < 18; ++i) { @@ -373,7 +373,7 @@ static int mmhub_v2_3_gart_enable(struct amdgpu_device *adev) static void mmhub_v2_3_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; u32 tmp; u32 i; @@ -446,7 +446,7 @@ static const struct amdgpu_vmhub_funcs mmhub_v2_3_vmhub_funcs = { static void mmhub_v2_3_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(MMHUB, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c index 17a792616979..441379e91cfa 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c @@ -136,7 +136,7 @@ mmhub_v3_0_print_l2_protection_fault_status(struct amdgpu_device *adev, static void mmhub_v3_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, hub->ctx_addr_distance * vmid, @@ -319,7 +319,7 @@ static void mmhub_v3_0_disable_identity_aperture(struct amdgpu_device *adev) static void mmhub_v3_0_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; int i; uint32_t tmp; @@ -369,7 +369,7 @@ static void mmhub_v3_0_setup_vmid_config(struct amdgpu_device *adev) static void mmhub_v3_0_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; unsigned i; for (i = 0; i < 18; ++i) { @@ -398,7 +398,7 @@ static int mmhub_v3_0_gart_enable(struct amdgpu_device *adev) static void mmhub_v3_0_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; u32 tmp; u32 i; @@ -477,7 +477,7 @@ static const struct amdgpu_vmhub_funcs mmhub_v3_0_vmhub_funcs = { static void mmhub_v3_0_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(MMHUB, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c index 26509b6b8c24..12c7f4b46ea9 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c @@ -138,7 +138,7 @@ static void mmhub_v3_0_1_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, hub->ctx_addr_distance * vmid, @@ -306,7 +306,7 @@ static void mmhub_v3_0_1_disable_identity_aperture(struct amdgpu_device *adev) static void mmhub_v3_0_1_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; int i; uint32_t tmp; @@ -356,7 +356,7 @@ static void mmhub_v3_0_1_setup_vmid_config(struct amdgpu_device *adev) static void mmhub_v3_0_1_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; unsigned i; for (i = 0; i < 18; ++i) { @@ -385,7 +385,7 @@ static int mmhub_v3_0_1_gart_enable(struct amdgpu_device *adev) static void mmhub_v3_0_1_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; u32 tmp; u32 i; @@ -459,7 +459,7 @@ static const struct amdgpu_vmhub_funcs mmhub_v3_0_1_vmhub_funcs = { static void mmhub_v3_0_1_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(MMHUB, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c index 26abbc6a47ab..5dadc85abf7e 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c @@ -129,7 +129,7 @@ mmhub_v3_0_2_print_l2_protection_fault_status(struct amdgpu_device *adev, static void mmhub_v3_0_2_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid, uint64_t page_table_base) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, hub->ctx_addr_distance * vmid, @@ -311,7 +311,7 @@ static void mmhub_v3_0_2_disable_identity_aperture(struct amdgpu_device *adev) static void mmhub_v3_0_2_setup_vmid_config(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; int i; uint32_t tmp; @@ -361,7 +361,7 @@ static void mmhub_v3_0_2_setup_vmid_config(struct amdgpu_device *adev) static void mmhub_v3_0_2_program_invalidation(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; unsigned i; for (i = 0; i < 18; ++i) { @@ -390,7 +390,7 @@ static int mmhub_v3_0_2_gart_enable(struct amdgpu_device *adev) static void mmhub_v3_0_2_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; u32 tmp; u32 i; @@ -469,7 +469,7 @@ static const struct amdgpu_vmhub_funcs mmhub_v3_0_2_vmhub_funcs = { static void mmhub_v3_0_2_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(MMHUB, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c index 72083e96222f..e790f890aec6 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c @@ -57,7 +57,7 @@ static u64 mmhub_v9_4_get_fb_location(struct amdgpu_device *adev) static void mmhub_v9_4_setup_hubid_vm_pt_regs(struct amdgpu_device *adev, int hubid, uint32_t vmid, uint64_t value) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; WREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, @@ -294,7 +294,7 @@ static void mmhub_v9_4_disable_identity_aperture(struct amdgpu_device *adev, static void mmhub_v9_4_setup_vmid_config(struct amdgpu_device *adev, int hubid) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; unsigned int num_level, block_size; uint32_t tmp; int i; @@ -363,7 +363,7 @@ static void mmhub_v9_4_setup_vmid_config(struct amdgpu_device *adev, int hubid) static void mmhub_v9_4_program_invalidation(struct amdgpu_device *adev, int hubid) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; unsigned i; for (i = 0; i < 18; ++i) { @@ -404,7 +404,7 @@ static int mmhub_v9_4_gart_enable(struct amdgpu_device *adev) static void mmhub_v9_4_gart_disable(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0]; + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)]; u32 tmp; u32 i, j; @@ -507,8 +507,8 @@ static void mmhub_v9_4_set_fault_enable_default(struct amdgpu_device *adev, bool static void mmhub_v9_4_init(struct amdgpu_device *adev) { - struct amdgpu_vmhub *hub[MMHUB_NUM_INSTANCES] = - {&adev->vmhub[AMDGPU_MMHUB_0], &adev->vmhub[AMDGPU_MMHUB_1]}; + struct amdgpu_vmhub *hub[MMHUB_NUM_INSTANCES] = { + &adev->vmhub[AMDGPU_MMHUB0(0)], &adev->vmhub[AMDGPU_MMHUB1(0)]}; int i; for (i = 0; i < MMHUB_NUM_INSTANCES; i++) { diff --git a/drivers/gpu/drm/amd/amdgpu/mmsch_v3_0.h b/drivers/gpu/drm/amd/amdgpu/mmsch_v3_0.h index 3e4e858a6965..a773ef61b78c 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmsch_v3_0.h +++ b/drivers/gpu/drm/amd/amdgpu/mmsch_v3_0.h @@ -30,6 +30,8 @@ #define MMSCH_VERSION_MINOR 0 #define MMSCH_VERSION (MMSCH_VERSION_MAJOR << 16 | MMSCH_VERSION_MINOR) +#define MMSCH_V3_0_VCN_INSTANCES 0x2 + enum mmsch_v3_0_command_type { MMSCH_COMMAND__DIRECT_REG_WRITE = 0, MMSCH_COMMAND__DIRECT_REG_POLLING = 2, @@ -47,7 +49,7 @@ struct mmsch_v3_0_table_info { struct mmsch_v3_0_init_header { uint32_t version; uint32_t total_size; - struct mmsch_v3_0_table_info inst[AMDGPU_MAX_VCN_INSTANCES]; + struct mmsch_v3_0_table_info inst[MMSCH_V3_0_VCN_INSTANCES]; }; struct mmsch_v3_0_cmd_direct_reg_header { diff --git a/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h b/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h index 83653a50a1a2..796d4f8791e5 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h +++ b/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h @@ -43,6 +43,8 @@ #define MMSCH_VF_MAILBOX_RESP__OK 0x1 #define MMSCH_VF_MAILBOX_RESP__INCOMPLETE 0x2 +#define MMSCH_V4_0_VCN_INSTANCES 0x2 + enum mmsch_v4_0_command_type { MMSCH_COMMAND__DIRECT_REG_WRITE = 0, MMSCH_COMMAND__DIRECT_REG_POLLING = 2, @@ -60,7 +62,7 @@ struct mmsch_v4_0_table_info { struct mmsch_v4_0_init_header { uint32_t version; uint32_t total_size; - struct mmsch_v4_0_table_info inst[AMDGPU_MAX_VCN_INSTANCES]; + struct mmsch_v4_0_table_info inst[MMSCH_V4_0_VCN_INSTANCES]; struct mmsch_v4_0_table_info jpegdec; }; diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c index 24d12075ca3a..cd1a02d30420 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c @@ -30,6 +30,8 @@ #include "ivsrcid/nbio/irqsrcs_nbif_7_4.h" #include <uapi/linux/kfd_ioctl.h> +#define NPS_MODE_MASK 0x000000FFL + static void nbio_v7_9_remap_hdp_registers(struct amdgpu_device *adev) { WREG32_SOC15(NBIO, 0, regBIF_BX0_REMAP_HDP_MEM_FLUSH_CNTL, @@ -66,6 +68,13 @@ static void nbio_v7_9_sdma_doorbell_range(struct amdgpu_device *adev, int instan bool use_doorbell, int doorbell_index, int doorbell_size) { u32 doorbell_range = 0, doorbell_ctrl = 0; + int aid_id, dev_inst; + + dev_inst = GET_INST(SDMA0, instance); + aid_id = adev->sdma.instance[instance].aid_id; + + if (use_doorbell == false) + return; doorbell_range = REG_SET_FIELD(doorbell_range, DOORBELL0_CTRL_ENTRY_0, @@ -80,9 +89,10 @@ static void nbio_v7_9_sdma_doorbell_range(struct amdgpu_device *adev, int instan REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, S2A_DOORBELL_PORT1_RANGE_SIZE, doorbell_size); - switch (instance) { + switch (dev_inst % adev->sdma.num_inst_per_aid) { case 0: - WREG32_SOC15(NBIO, 0, regDOORBELL0_CTRL_ENTRY_1, doorbell_range); + WREG32_SOC15_OFFSET(NBIO, 0, regDOORBELL0_CTRL_ENTRY_1, + 4 * aid_id, doorbell_range); doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, @@ -94,10 +104,12 @@ static void nbio_v7_9_sdma_doorbell_range(struct amdgpu_device *adev, int instan S2A_DOORBELL_ENTRY_1_CTRL, S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE, 0x1); - WREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_1_CTRL, doorbell_ctrl); + WREG32_SOC15_EXT(NBIO, aid_id, regS2A_DOORBELL_ENTRY_1_CTRL, + aid_id, doorbell_ctrl); break; case 1: - WREG32_SOC15(NBIO, 0, regDOORBELL0_CTRL_ENTRY_2, doorbell_range); + WREG32_SOC15_OFFSET(NBIO, 0, regDOORBELL0_CTRL_ENTRY_2, + 4 * aid_id, doorbell_range); doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, @@ -109,10 +121,12 @@ static void nbio_v7_9_sdma_doorbell_range(struct amdgpu_device *adev, int instan S2A_DOORBELL_ENTRY_1_CTRL, S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE, 0x2); - WREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_2_CTRL, doorbell_ctrl); + WREG32_SOC15_EXT(NBIO, aid_id, regS2A_DOORBELL_ENTRY_2_CTRL, + aid_id, doorbell_ctrl); break; case 2: - WREG32_SOC15(NBIO, 0, regDOORBELL0_CTRL_ENTRY_3, doorbell_range); + WREG32_SOC15_OFFSET(NBIO, 0, regDOORBELL0_CTRL_ENTRY_3, + 4 * aid_id, doorbell_range); doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, @@ -124,10 +138,12 @@ static void nbio_v7_9_sdma_doorbell_range(struct amdgpu_device *adev, int instan S2A_DOORBELL_ENTRY_1_CTRL, S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE, 0x8); - WREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_5_CTRL, doorbell_ctrl); + WREG32_SOC15_EXT(NBIO, aid_id, regS2A_DOORBELL_ENTRY_5_CTRL, + aid_id, doorbell_ctrl); break; case 3: - WREG32_SOC15(NBIO, 0, regDOORBELL0_CTRL_ENTRY_4, doorbell_range); + WREG32_SOC15_OFFSET(NBIO, 0, regDOORBELL0_CTRL_ENTRY_4, + 4 * aid_id, doorbell_range); doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, @@ -139,11 +155,12 @@ static void nbio_v7_9_sdma_doorbell_range(struct amdgpu_device *adev, int instan S2A_DOORBELL_ENTRY_1_CTRL, S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE, 0x9); - WREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_6_CTRL, doorbell_ctrl); + WREG32_SOC15_EXT(NBIO, aid_id, regS2A_DOORBELL_ENTRY_6_CTRL, + aid_id, doorbell_ctrl); break; default: break; - }; + } return; } @@ -152,6 +169,7 @@ static void nbio_v7_9_vcn_doorbell_range(struct amdgpu_device *adev, bool use_do int doorbell_index, int instance) { u32 doorbell_range = 0, doorbell_ctrl = 0; + u32 aid_id = instance; if (use_doorbell) { doorbell_range = REG_SET_FIELD(doorbell_range, @@ -161,7 +179,12 @@ static void nbio_v7_9_vcn_doorbell_range(struct amdgpu_device *adev, bool use_do doorbell_range = REG_SET_FIELD(doorbell_range, DOORBELL0_CTRL_ENTRY_0, BIF_DOORBELL0_RANGE_SIZE_ENTRY, - 0x8); + 0x9); + if (aid_id) + doorbell_range = REG_SET_FIELD(doorbell_range, + DOORBELL0_CTRL_ENTRY_0, + DOORBELL0_FENCE_ENABLE_ENTRY, + 0x4); doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, @@ -174,10 +197,15 @@ static void nbio_v7_9_vcn_doorbell_range(struct amdgpu_device *adev, bool use_do S2A_DOORBELL_PORT1_RANGE_OFFSET, 0x4); doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, - S2A_DOORBELL_PORT1_RANGE_SIZE, 0x8); + S2A_DOORBELL_PORT1_RANGE_SIZE, 0x9); doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE, 0x4); + + WREG32_SOC15_OFFSET(NBIO, 0, regDOORBELL0_CTRL_ENTRY_17, + aid_id, doorbell_range); + WREG32_SOC15_EXT(NBIO, aid_id, regS2A_DOORBELL_ENTRY_4_CTRL, + aid_id, doorbell_ctrl); } else { doorbell_range = REG_SET_FIELD(doorbell_range, DOORBELL0_CTRL_ENTRY_0, @@ -185,10 +213,12 @@ static void nbio_v7_9_vcn_doorbell_range(struct amdgpu_device *adev, bool use_do doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, S2A_DOORBELL_PORT1_RANGE_SIZE, 0); - } - WREG32_SOC15(NBIO, 0, regDOORBELL0_CTRL_ENTRY_17, doorbell_range); - WREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_4_CTRL, doorbell_ctrl); + WREG32_SOC15_OFFSET(NBIO, 0, regDOORBELL0_CTRL_ENTRY_17, + aid_id, doorbell_range); + WREG32_SOC15_EXT(NBIO, aid_id, regS2A_DOORBELL_ENTRY_4_CTRL, + aid_id, doorbell_ctrl); + } } static void nbio_v7_9_enable_doorbell_aperture(struct amdgpu_device *adev, @@ -235,7 +265,7 @@ static void nbio_v7_9_ih_doorbell_range(struct amdgpu_device *adev, ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range, DOORBELL0_CTRL_ENTRY_0, BIF_DOORBELL0_RANGE_SIZE_ENTRY, - 0x4); + 0x8); ih_doorbell_ctrl = REG_SET_FIELD(ih_doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, @@ -248,7 +278,7 @@ static void nbio_v7_9_ih_doorbell_range(struct amdgpu_device *adev, S2A_DOORBELL_PORT1_RANGE_OFFSET, 0); ih_doorbell_ctrl = REG_SET_FIELD(ih_doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, - S2A_DOORBELL_PORT1_RANGE_SIZE, 0x4); + S2A_DOORBELL_PORT1_RANGE_SIZE, 0x8); ih_doorbell_ctrl = REG_SET_FIELD(ih_doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL, S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE, 0); @@ -319,6 +349,11 @@ static u32 nbio_v7_9_get_pcie_data_offset(struct amdgpu_device *adev) return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX0_PCIE_DATA2); } +static u32 nbio_v7_9_get_pcie_index_hi_offset(struct amdgpu_device *adev) +{ + return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX0_PCIE_INDEX2_HI); +} + const struct nbio_hdp_flush_reg nbio_v7_9_hdp_flush_reg = { .ref_and_mask_cp0 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP0_MASK, .ref_and_mask_cp1 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP1_MASK, @@ -347,11 +382,57 @@ static void nbio_v7_9_enable_doorbell_interrupt(struct amdgpu_device *adev, DOORBELL_INTERRUPT_DISABLE, enable ? 0 : 1); } +static int nbio_v7_9_get_compute_partition_mode(struct amdgpu_device *adev) +{ + u32 tmp, px; + + tmp = RREG32_SOC15(NBIO, 0, regBIF_BX_PF0_PARTITION_COMPUTE_STATUS); + px = REG_GET_FIELD(tmp, BIF_BX_PF0_PARTITION_COMPUTE_STATUS, + PARTITION_MODE); + + return px; +} + +static u32 nbio_v7_9_get_memory_partition_mode(struct amdgpu_device *adev, + u32 *supp_modes) +{ + u32 tmp; + + tmp = RREG32_SOC15(NBIO, 0, regBIF_BX_PF0_PARTITION_MEM_STATUS); + tmp = REG_GET_FIELD(tmp, BIF_BX_PF0_PARTITION_MEM_STATUS, NPS_MODE); + + if (supp_modes) { + *supp_modes = + RREG32_SOC15(NBIO, 0, regBIF_BX_PF0_PARTITION_MEM_CAP); + } + + return ffs(tmp); +} + +static void nbio_v7_9_init_registers(struct amdgpu_device *adev) +{ + u32 inst_mask; + int i; + + WREG32_SOC15(NBIO, 0, regXCC_DOORBELL_FENCE, + 0xff & ~(adev->gfx.xcc_mask)); + + WREG32_SOC15(NBIO, 0, regBIFC_GFX_INT_MONITOR_MASK, 0x7ff); + + inst_mask = adev->aid_mask & ~1U; + for_each_inst(i, inst_mask) { + WREG32_SOC15_EXT(NBIO, i, regXCC_DOORBELL_FENCE, i, + XCC_DOORBELL_FENCE__SHUB_SLV_MODE_MASK); + + } +} + const struct amdgpu_nbio_funcs nbio_v7_9_funcs = { .get_hdp_flush_req_offset = nbio_v7_9_get_hdp_flush_req_offset, .get_hdp_flush_done_offset = nbio_v7_9_get_hdp_flush_done_offset, .get_pcie_index_offset = nbio_v7_9_get_pcie_index_offset, .get_pcie_data_offset = nbio_v7_9_get_pcie_data_offset, + .get_pcie_index_hi_offset = nbio_v7_9_get_pcie_index_hi_offset, .get_rev_id = nbio_v7_9_get_rev_id, .mc_access_enable = nbio_v7_9_mc_access_enable, .get_memsize = nbio_v7_9_get_memsize, @@ -366,4 +447,7 @@ const struct amdgpu_nbio_funcs nbio_v7_9_funcs = { .get_clockgating_state = nbio_v7_9_get_clockgating_state, .ih_control = nbio_v7_9_ih_control, .remap_hdp_registers = nbio_v7_9_remap_hdp_registers, + .get_compute_partition_mode = nbio_v7_9_get_compute_partition_mode, + .get_memory_partition_mode = nbio_v7_9_get_memory_partition_mode, + .init_registers = nbio_v7_9_init_registers, }; diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c index 0fb6013441f0..51523b27a186 100644 --- a/drivers/gpu/drm/amd/amdgpu/nv.c +++ b/drivers/gpu/drm/amd/amdgpu/nv.c @@ -341,11 +341,6 @@ void nv_grbm_select(struct amdgpu_device *adev, WREG32_SOC15(GC, 0, mmGRBM_GFX_CNTL, grbm_gfx_cntl); } -static void nv_vga_set_state(struct amdgpu_device *adev, bool state) -{ - /* todo */ -} - static bool nv_read_disabled_bios(struct amdgpu_device *adev) { /* todo */ @@ -381,12 +376,12 @@ static uint32_t nv_read_indexed_register(struct amdgpu_device *adev, u32 se_num, mutex_lock(&adev->grbm_idx_mutex); if (se_num != 0xffffffff || sh_num != 0xffffffff) - amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0); val = RREG32(reg_offset); if (se_num != 0xffffffff || sh_num != 0xffffffff) - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); return val; } @@ -632,9 +627,9 @@ static int nv_update_umd_stable_pstate(struct amdgpu_device *adev, bool enter) { if (enter) - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); else - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); if (adev->gfx.funcs->update_perfmon_mgcg) adev->gfx.funcs->update_perfmon_mgcg(adev, !enter); @@ -654,7 +649,6 @@ static const struct amdgpu_asic_funcs nv_asic_funcs = .read_register = &nv_read_register, .reset = &nv_asic_reset, .reset_method = &nv_asic_reset_method, - .set_vga_state = &nv_vga_set_state, .get_xclk = &nv_get_xclk, .set_uvd_clocks = &nv_set_uvd_clocks, .set_vce_clocks = &nv_set_vce_clocks, diff --git a/drivers/gpu/drm/amd/amdgpu/nvd.h b/drivers/gpu/drm/amd/amdgpu/nvd.h index fd6b58243b03..631dafb92299 100644 --- a/drivers/gpu/drm/amd/amdgpu/nvd.h +++ b/drivers/gpu/drm/amd/amdgpu/nvd.h @@ -462,6 +462,9 @@ # define PACKET3_QUERY_STATUS_ENG_SEL(x) ((x) << 25) #define PACKET3_RUN_LIST 0xA5 #define PACKET3_MAP_PROCESS_VM 0xA6 - +/* GFX11 */ +#define PACKET3_SET_Q_PREEMPTION_MODE 0xF0 +# define PACKET3_SET_Q_PREEMPTION_MODE_IB_VMID(x) ((x) << 0) +# define PACKET3_SET_Q_PREEMPTION_MODE_INIT_SHADOW_MEM (1 << 0) #endif diff --git a/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h b/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h index 22c775f39119..18917df785ec 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h +++ b/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h @@ -102,6 +102,7 @@ enum psp_gfx_cmd_id GFX_CMD_ID_LOAD_TOC = 0x00000020, /* Load TOC and obtain TMR size */ GFX_CMD_ID_AUTOLOAD_RLC = 0x00000021, /* Indicates all graphics fw loaded, start RLC autoload */ GFX_CMD_ID_BOOT_CFG = 0x00000022, /* Boot Config */ + GFX_CMD_ID_SRIOV_SPATIAL_PART = 0x00000027, /* Configure spatial partitioning mode */ }; /* PSP boot config sub-commands */ @@ -338,6 +339,13 @@ struct psp_gfx_cmd_boot_cfg uint32_t boot_config_valid; /* dynamic boot configuration valid bits bitmask */ }; +struct psp_gfx_cmd_sriov_spatial_part { + uint32_t mode; + uint32_t override_ips; + uint32_t override_xcds_avail; + uint32_t override_this_aid; +}; + /* All GFX ring buffer commands. */ union psp_gfx_commands { @@ -351,6 +359,7 @@ union psp_gfx_commands struct psp_gfx_cmd_setup_tmr cmd_setup_vmr; struct psp_gfx_cmd_load_toc cmd_load_toc; struct psp_gfx_cmd_boot_cfg boot_cfg; + struct psp_gfx_cmd_sriov_spatial_part cmd_spatial_part; }; struct psp_gfx_uresp_reserved diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c index e1b7fca09666..5f10883da6a2 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c @@ -57,7 +57,13 @@ static int psp_v10_0_init_microcode(struct psp_context *psp) if (err) return err; - return psp_init_ta_microcode(psp, ucode_prefix); + err = psp_init_ta_microcode(psp, ucode_prefix); + if ((adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 1, 0)) && + (adev->pdev->revision == 0xa1) && + (psp->securedisplay_context.context.bin_desc.fw_version >= 0x27000008)) { + adev->psp.securedisplay_context.context.bin_desc.size_bytes = 0; + } + return err; } static int psp_v10_0_ring_create(struct psp_context *psp, diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c index caee76ab7110..f9cb0d2c89d1 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c @@ -624,10 +624,11 @@ static int psp_v13_0_exec_spi_cmd(struct psp_context *psp, int cmd) WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_73, 1); if (cmd == C2PMSG_CMD_SPI_UPDATE_FLASH_IMAGE) - return 0; - - ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_115), - MBOX_READY_FLAG, MBOX_READY_MASK, false); + ret = psp_wait_for_spirom_update(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_115), + MBOX_READY_FLAG, MBOX_READY_MASK, PSP_SPIROM_UPDATE_TIMEOUT); + else + ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_115), + MBOX_READY_FLAG, MBOX_READY_MASK, false); if (ret) { dev_err(adev->dev, "SPI cmd %x timed out, ret = %d", cmd, ret); return ret; diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.h b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.h index b2414a729ca1..de5677ce4330 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.h +++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.h @@ -25,6 +25,8 @@ #include "amdgpu_psp.h" +#define PSP_SPIROM_UPDATE_TIMEOUT 60000 /* 60s */ + void psp_v13_0_set_psp_funcs(struct psp_context *psp); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c index fd2a7b66ac56..51afc92994a8 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c @@ -466,8 +466,6 @@ static int sdma_v2_4_gfx_resume(struct amdgpu_device *adev) #endif /* enable DMA IBs */ WREG32(mmSDMA0_GFX_IB_CNTL + sdma_offsets[i], ib_cntl); - - ring->sched.ready = true; } sdma_v2_4_enable(adev, true); diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c index e572389089d2..344202870aeb 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c @@ -734,8 +734,6 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev) #endif /* enable DMA IBs */ WREG32(mmSDMA0_GFX_IB_CNTL + sdma_offsets[i], ib_cntl); - - ring->sched.ready = true; } /* unhalt the MEs */ diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c index 9295ac7edd56..cd37f45e01a1 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c @@ -1114,8 +1114,6 @@ static void sdma_v4_0_gfx_resume(struct amdgpu_device *adev, unsigned int i) #endif /* enable DMA IBs */ WREG32_SDMA(i, mmSDMA0_GFX_IB_CNTL, ib_cntl); - - ring->sched.ready = true; } /** @@ -1202,8 +1200,6 @@ static void sdma_v4_0_page_resume(struct amdgpu_device *adev, unsigned int i) #endif /* enable DMA IBs */ WREG32_SDMA(i, mmSDMA0_PAGE_IB_CNTL, ib_cntl); - - ring->sched.ready = true; } static void @@ -1825,12 +1821,12 @@ static int sdma_v4_0_sw_init(void *handle) /* * On Arcturus, SDMA instance 5~7 has a different vmhub - * type(AMDGPU_MMHUB_1). + * type(AMDGPU_MMHUB1). */ if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 2) && i >= 5) - ring->vm_hub = AMDGPU_MMHUB_1; + ring->vm_hub = AMDGPU_MMHUB1(0); else - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "sdma%d", i); r = amdgpu_ring_init(adev, ring, 1024, &adev->sdma.trap_irq, @@ -1847,13 +1843,23 @@ static int sdma_v4_0_sw_init(void *handle) /* paging queue use same doorbell index/routing as gfx queue * with 0x400 (4096 dwords) offset on second doorbell page */ - ring->doorbell_index = adev->doorbell_index.sdma_engine[i] << 1; - ring->doorbell_index += 0x400; + if (adev->ip_versions[SDMA0_HWIP][0] >= IP_VERSION(4, 0, 0) && + adev->ip_versions[SDMA0_HWIP][0] < IP_VERSION(4, 2, 0)) { + ring->doorbell_index = + adev->doorbell_index.sdma_engine[i] << 1; + ring->doorbell_index += 0x400; + } else { + /* From vega20, the sdma_doorbell_range in 1st + * doorbell page is reserved for page queue. + */ + ring->doorbell_index = + (adev->doorbell_index.sdma_engine[i] + 1) << 1; + } if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 2) && i >= 5) - ring->vm_hub = AMDGPU_MMHUB_1; + ring->vm_hub = AMDGPU_MMHUB1(0); else - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "page%d", i); r = amdgpu_ring_init(adev, ring, 1024, @@ -2306,7 +2312,7 @@ const struct amd_ip_funcs sdma_v4_0_ip_funcs = { static const struct amdgpu_ring_funcs sdma_v4_0_ring_funcs = { .type = AMDGPU_RING_TYPE_SDMA, - .align_mask = 0xf, + .align_mask = 0xff, .nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP), .support_64bit_ptrs = true, .secure_submission_supported = true, @@ -2338,7 +2344,7 @@ static const struct amdgpu_ring_funcs sdma_v4_0_ring_funcs = { static const struct amdgpu_ring_funcs sdma_v4_0_page_ring_funcs = { .type = AMDGPU_RING_TYPE_SDMA, - .align_mask = 0xf, + .align_mask = 0xff, .nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP), .support_64bit_ptrs = true, .secure_submission_supported = true, diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4.c index 6f9895cdddb1..0ddb6955a6d3 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4.c @@ -141,6 +141,10 @@ static const struct soc15_ras_field_entry sdma_v4_4_ras_fields[] = { SOC15_REG_FIELD(SDMA0_EDC_COUNTER2, SDMA_UTCL1_RDBST_FIFO_SED), 0, 0, }, + { "SDMA_UTCL1_WR_FIFO_SED", SOC15_REG_ENTRY(SDMA0, 0, regSDMA0_EDC_COUNTER2), + SOC15_REG_FIELD(SDMA0_EDC_COUNTER2, SDMA_UTCL1_WR_FIFO_SED), + 0, 0, + }, { "SDMA_DATA_LUT_FIFO_SED", SOC15_REG_ENTRY(SDMA0, 0, regSDMA0_EDC_COUNTER2), SOC15_REG_FIELD(SDMA0_EDC_COUNTER2, SDMA_DATA_LUT_FIFO_SED), 0, 0, diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c index 64dcaa2670dd..ea5e12390d18 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c @@ -27,6 +27,7 @@ #include <linux/pci.h> #include "amdgpu.h" +#include "amdgpu_xcp.h" #include "amdgpu_ucode.h" #include "amdgpu_trace.h" @@ -53,11 +54,14 @@ static void sdma_v4_4_2_set_ring_funcs(struct amdgpu_device *adev); static void sdma_v4_4_2_set_buffer_funcs(struct amdgpu_device *adev); static void sdma_v4_4_2_set_vm_pte_funcs(struct amdgpu_device *adev); static void sdma_v4_4_2_set_irq_funcs(struct amdgpu_device *adev); +static void sdma_v4_4_2_set_ras_funcs(struct amdgpu_device *adev); static u32 sdma_v4_4_2_get_reg_offset(struct amdgpu_device *adev, u32 instance, u32 offset) { - return (adev->reg_offset[SDMA0_HWIP][instance][0] + offset); + u32 dev_inst = GET_INST(SDMA0, instance); + + return (adev->reg_offset[SDMA0_HWIP][dev_inst][0] + offset); } static unsigned sdma_v4_4_2_seq_to_irq_id(int seq_num) @@ -92,13 +96,25 @@ static int sdma_v4_4_2_irq_id_to_seq(unsigned client_id) } } -static void sdma_v4_4_2_init_golden_registers(struct amdgpu_device *adev) +static void sdma_v4_4_2_inst_init_golden_registers(struct amdgpu_device *adev, + uint32_t inst_mask) { - switch (adev->ip_versions[SDMA0_HWIP][0]) { - case IP_VERSION(4, 4, 2): - break; - default: - break; + u32 val; + int i; + + for (i = 0; i < adev->sdma.num_instances; i++) { + val = RREG32_SDMA(i, regSDMA_GB_ADDR_CONFIG); + val = REG_SET_FIELD(val, SDMA_GB_ADDR_CONFIG, NUM_BANKS, 4); + val = REG_SET_FIELD(val, SDMA_GB_ADDR_CONFIG, + PIPE_INTERLEAVE_SIZE, 0); + WREG32_SDMA(i, regSDMA_GB_ADDR_CONFIG, val); + + val = RREG32_SDMA(i, regSDMA_GB_ADDR_CONFIG_READ); + val = REG_SET_FIELD(val, SDMA_GB_ADDR_CONFIG_READ, NUM_BANKS, + 4); + val = REG_SET_FIELD(val, SDMA_GB_ADDR_CONFIG_READ, + PIPE_INTERLEAVE_SIZE, 0); + WREG32_SDMA(i, regSDMA_GB_ADDR_CONFIG_READ, val); } } @@ -399,19 +415,21 @@ static void sdma_v4_4_2_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 /** - * sdma_v4_4_2_gfx_stop - stop the gfx async dma engines + * sdma_v4_4_2_inst_gfx_stop - stop the gfx async dma engines * * @adev: amdgpu_device pointer + * @inst_mask: mask of dma engine instances to be disabled * * Stop the gfx async dma ring buffers. */ -static void sdma_v4_4_2_gfx_stop(struct amdgpu_device *adev) +static void sdma_v4_4_2_inst_gfx_stop(struct amdgpu_device *adev, + uint32_t inst_mask) { struct amdgpu_ring *sdma[AMDGPU_MAX_SDMA_INSTANCES]; u32 rb_cntl, ib_cntl; int i, unset = 0; - for (i = 0; i < adev->sdma.num_instances; i++) { + for_each_inst(i, inst_mask) { sdma[i] = &adev->sdma.instance[i].ring; if ((adev->mman.buffer_funcs_ring == sdma[i]) && unset != 1) { @@ -429,32 +447,36 @@ static void sdma_v4_4_2_gfx_stop(struct amdgpu_device *adev) } /** - * sdma_v4_4_2_rlc_stop - stop the compute async dma engines + * sdma_v4_4_2_inst_rlc_stop - stop the compute async dma engines * * @adev: amdgpu_device pointer + * @inst_mask: mask of dma engine instances to be disabled * * Stop the compute async dma queues. */ -static void sdma_v4_4_2_rlc_stop(struct amdgpu_device *adev) +static void sdma_v4_4_2_inst_rlc_stop(struct amdgpu_device *adev, + uint32_t inst_mask) { /* XXX todo */ } /** - * sdma_v4_4_2_page_stop - stop the page async dma engines + * sdma_v4_4_2_inst_page_stop - stop the page async dma engines * * @adev: amdgpu_device pointer + * @inst_mask: mask of dma engine instances to be disabled * * Stop the page async dma ring buffers. */ -static void sdma_v4_4_2_page_stop(struct amdgpu_device *adev) +static void sdma_v4_4_2_inst_page_stop(struct amdgpu_device *adev, + uint32_t inst_mask) { struct amdgpu_ring *sdma[AMDGPU_MAX_SDMA_INSTANCES]; u32 rb_cntl, ib_cntl; int i; bool unset = false; - for (i = 0; i < adev->sdma.num_instances; i++) { + for_each_inst(i, inst_mask) { sdma[i] = &adev->sdma.instance[i].page; if ((adev->mman.buffer_funcs_ring == sdma[i]) && @@ -475,14 +497,16 @@ static void sdma_v4_4_2_page_stop(struct amdgpu_device *adev) } /** - * sdma_v4_4_2_ctx_switch_enable - stop the async dma engines context switch + * sdma_v4_4_2_inst_ctx_switch_enable - stop the async dma engines context switch * * @adev: amdgpu_device pointer * @enable: enable/disable the DMA MEs context switch. + * @inst_mask: mask of dma engine instances to be enabled * * Halt or unhalt the async dma engines context switch. */ -static void sdma_v4_4_2_ctx_switch_enable(struct amdgpu_device *adev, bool enable) +static void sdma_v4_4_2_inst_ctx_switch_enable(struct amdgpu_device *adev, + bool enable, uint32_t inst_mask) { u32 f32_cntl, phase_quantum = 0; int i; @@ -511,7 +535,7 @@ static void sdma_v4_4_2_ctx_switch_enable(struct amdgpu_device *adev, bool enabl unit << SDMA_PHASE0_QUANTUM__UNIT__SHIFT; } - for (i = 0; i < adev->sdma.num_instances; i++) { + for_each_inst(i, inst_mask) { f32_cntl = RREG32_SDMA(i, regSDMA_CNTL); f32_cntl = REG_SET_FIELD(f32_cntl, SDMA_CNTL, AUTO_CTXSW_ENABLE, enable ? 1 : 0); @@ -525,30 +549,36 @@ static void sdma_v4_4_2_ctx_switch_enable(struct amdgpu_device *adev, bool enabl /* Extend page fault timeout to avoid interrupt storm */ WREG32_SDMA(i, regSDMA_UTCL1_TIMEOUT, 0x00800080); } - } /** - * sdma_v4_4_2_enable - stop the async dma engines + * sdma_v4_4_2_inst_enable - stop the async dma engines * * @adev: amdgpu_device pointer * @enable: enable/disable the DMA MEs. + * @inst_mask: mask of dma engine instances to be enabled * * Halt or unhalt the async dma engines. */ -static void sdma_v4_4_2_enable(struct amdgpu_device *adev, bool enable) +static void sdma_v4_4_2_inst_enable(struct amdgpu_device *adev, bool enable, + uint32_t inst_mask) { u32 f32_cntl; int i; if (!enable) { - sdma_v4_4_2_gfx_stop(adev); - sdma_v4_4_2_rlc_stop(adev); + sdma_v4_4_2_inst_gfx_stop(adev, inst_mask); + sdma_v4_4_2_inst_rlc_stop(adev, inst_mask); if (adev->sdma.has_page_queue) - sdma_v4_4_2_page_stop(adev); + sdma_v4_4_2_inst_page_stop(adev, inst_mask); + + /* SDMA FW needs to respond to FREEZE requests during reset. + * Keep it running during reset */ + if (!amdgpu_sriov_vf(adev) && amdgpu_in_reset(adev)) + return; } - for (i = 0; i < adev->sdma.num_instances; i++) { + for_each_inst(i, inst_mask) { f32_cntl = RREG32_SDMA(i, regSDMA_F32_CNTL); f32_cntl = REG_SET_FIELD(f32_cntl, SDMA_F32_CNTL, HALT, enable ? 0 : 1); WREG32_SDMA(i, regSDMA_F32_CNTL, f32_cntl); @@ -659,8 +689,6 @@ static void sdma_v4_4_2_gfx_resume(struct amdgpu_device *adev, unsigned int i) #endif /* enable DMA IBs */ WREG32_SDMA(i, regSDMA_GFX_IB_CNTL, ib_cntl); - - ring->sched.ready = true; } /** @@ -750,8 +778,6 @@ static void sdma_v4_4_2_page_resume(struct amdgpu_device *adev, unsigned int i) #endif /* enable DMA IBs */ WREG32_SDMA(i, regSDMA_PAGE_IB_CNTL, ib_cntl); - - ring->sched.ready = true; } static void sdma_v4_4_2_init_pg(struct amdgpu_device *adev) @@ -760,14 +786,16 @@ static void sdma_v4_4_2_init_pg(struct amdgpu_device *adev) } /** - * sdma_v4_4_2_rlc_resume - setup and start the async dma engines + * sdma_v4_4_2_inst_rlc_resume - setup and start the async dma engines * * @adev: amdgpu_device pointer + * @inst_mask: mask of dma engine instances to be enabled * * Set up the compute DMA queues and enable them. * Returns 0 for success, error for failure. */ -static int sdma_v4_4_2_rlc_resume(struct amdgpu_device *adev) +static int sdma_v4_4_2_inst_rlc_resume(struct amdgpu_device *adev, + uint32_t inst_mask) { sdma_v4_4_2_init_pg(adev); @@ -775,14 +803,16 @@ static int sdma_v4_4_2_rlc_resume(struct amdgpu_device *adev) } /** - * sdma_v4_4_2_load_microcode - load the sDMA ME ucode + * sdma_v4_4_2_inst_load_microcode - load the sDMA ME ucode * * @adev: amdgpu_device pointer + * @inst_mask: mask of dma engine instances to be enabled * * Loads the sDMA0/1 ucode. * Returns 0 for success, -EINVAL if the ucode is not available. */ -static int sdma_v4_4_2_load_microcode(struct amdgpu_device *adev) +static int sdma_v4_4_2_inst_load_microcode(struct amdgpu_device *adev, + uint32_t inst_mask) { const struct sdma_firmware_header_v1_0 *hdr; const __le32 *fw_data; @@ -790,9 +820,9 @@ static int sdma_v4_4_2_load_microcode(struct amdgpu_device *adev) int i, j; /* halt the MEs */ - sdma_v4_4_2_enable(adev, false); + sdma_v4_4_2_inst_enable(adev, false, inst_mask); - for (i = 0; i < adev->sdma.num_instances; i++) { + for_each_inst(i, inst_mask) { if (!adev->sdma.instance[i].fw) return -EINVAL; @@ -818,38 +848,42 @@ static int sdma_v4_4_2_load_microcode(struct amdgpu_device *adev) } /** - * sdma_v4_4_2_start - setup and start the async dma engines + * sdma_v4_4_2_inst_start - setup and start the async dma engines * * @adev: amdgpu_device pointer + * @inst_mask: mask of dma engine instances to be enabled * * Set up the DMA engines and enable them. * Returns 0 for success, error for failure. */ -static int sdma_v4_4_2_start(struct amdgpu_device *adev) +static int sdma_v4_4_2_inst_start(struct amdgpu_device *adev, + uint32_t inst_mask) { struct amdgpu_ring *ring; + uint32_t tmp_mask; int i, r = 0; if (amdgpu_sriov_vf(adev)) { - sdma_v4_4_2_ctx_switch_enable(adev, false); - sdma_v4_4_2_enable(adev, false); + sdma_v4_4_2_inst_ctx_switch_enable(adev, false, inst_mask); + sdma_v4_4_2_inst_enable(adev, false, inst_mask); } else { /* bypass sdma microcode loading on Gopher */ if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP && - !(adev->pdev->device == 0x49) && !(adev->pdev->device == 0x50)) { - r = sdma_v4_4_2_load_microcode(adev); + adev->sdma.instance[0].fw) { + r = sdma_v4_4_2_inst_load_microcode(adev, inst_mask); if (r) return r; } /* unhalt the MEs */ - sdma_v4_4_2_enable(adev, true); + sdma_v4_4_2_inst_enable(adev, true, inst_mask); /* enable sdma ring preemption */ - sdma_v4_4_2_ctx_switch_enable(adev, true); + sdma_v4_4_2_inst_ctx_switch_enable(adev, true, inst_mask); } /* start the gfx rings and rlc compute queues */ - for (i = 0; i < adev->sdma.num_instances; i++) { + tmp_mask = inst_mask; + for_each_inst(i, tmp_mask) { uint32_t temp; WREG32_SDMA(i, regSDMA_SEM_WAIT_FAIL_TIMER_CNTL, 0); @@ -860,6 +894,8 @@ static int sdma_v4_4_2_start(struct amdgpu_device *adev) /* set utc l1 enable flag always to 1 */ temp = RREG32_SDMA(i, regSDMA_CNTL); temp = REG_SET_FIELD(temp, SDMA_CNTL, UTC_L1_ENABLE, 1); + /* enable context empty interrupt during initialization */ + temp = REG_SET_FIELD(temp, SDMA_CNTL, CTXEMPTY_INT_ENABLE, 1); WREG32_SDMA(i, regSDMA_CNTL, temp); if (!amdgpu_sriov_vf(adev)) { @@ -876,15 +912,16 @@ static int sdma_v4_4_2_start(struct amdgpu_device *adev) } if (amdgpu_sriov_vf(adev)) { - sdma_v4_4_2_ctx_switch_enable(adev, true); - sdma_v4_4_2_enable(adev, true); + sdma_v4_4_2_inst_ctx_switch_enable(adev, true, inst_mask); + sdma_v4_4_2_inst_enable(adev, true, inst_mask); } else { - r = sdma_v4_4_2_rlc_resume(adev); + r = sdma_v4_4_2_inst_rlc_resume(adev, inst_mask); if (r) return r; } - for (i = 0; i < adev->sdma.num_instances; i++) { + tmp_mask = inst_mask; + for_each_inst(i, tmp_mask) { ring = &adev->sdma.instance[i].ring; r = amdgpu_ring_test_helper(ring); @@ -1221,6 +1258,7 @@ static int sdma_v4_4_2_early_init(void *handle) sdma_v4_4_2_set_buffer_funcs(adev); sdma_v4_4_2_set_vm_pte_funcs(adev); sdma_v4_4_2_set_irq_funcs(adev); + sdma_v4_4_2_set_ras_funcs(adev); return 0; } @@ -1253,9 +1291,10 @@ static int sdma_v4_4_2_sw_init(void *handle) struct amdgpu_ring *ring; int r, i; struct amdgpu_device *adev = (struct amdgpu_device *)handle; + u32 aid_id; /* SDMA trap event */ - for (i = 0; i < adev->sdma.num_instances; i++) { + for (i = 0; i < adev->sdma.num_inst_per_aid; i++) { r = amdgpu_irq_add_id(adev, sdma_v4_4_2_seq_to_irq_id(i), SDMA0_4_0__SRCID__SDMA_TRAP, &adev->sdma.trap_irq); @@ -1264,7 +1303,7 @@ static int sdma_v4_4_2_sw_init(void *handle) } /* SDMA SRAM ECC event */ - for (i = 0; i < adev->sdma.num_instances; i++) { + for (i = 0; i < adev->sdma.num_inst_per_aid; i++) { r = amdgpu_irq_add_id(adev, sdma_v4_4_2_seq_to_irq_id(i), SDMA0_4_0__SRCID__SDMA_SRAM_ECC, &adev->sdma.ecc_irq); @@ -1273,7 +1312,7 @@ static int sdma_v4_4_2_sw_init(void *handle) } /* SDMA VM_HOLE/DOORBELL_INV/POLL_TIMEOUT/SRBM_WRITE_PROTECTION event*/ - for (i = 0; i < adev->sdma.num_instances; i++) { + for (i = 0; i < adev->sdma.num_inst_per_aid; i++) { r = amdgpu_irq_add_id(adev, sdma_v4_4_2_seq_to_irq_id(i), SDMA0_4_0__SRCID__SDMA_VM_HOLE, &adev->sdma.vm_hole_irq); @@ -1303,15 +1342,17 @@ static int sdma_v4_4_2_sw_init(void *handle) ring = &adev->sdma.instance[i].ring; ring->ring_obj = NULL; ring->use_doorbell = true; + aid_id = adev->sdma.instance[i].aid_id; DRM_DEBUG("SDMA %d use_doorbell being set to: [%s]\n", i, ring->use_doorbell?"true":"false"); /* doorbell size is 2 dwords, get DWORD offset */ ring->doorbell_index = adev->doorbell_index.sdma_engine[i] << 1; - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(aid_id); - sprintf(ring->name, "sdma%d", i); + sprintf(ring->name, "sdma%d.%d", aid_id, + i % adev->sdma.num_inst_per_aid); r = amdgpu_ring_init(adev, ring, 1024, &adev->sdma.trap_irq, AMDGPU_SDMA_IRQ_INSTANCE0 + i, AMDGPU_RING_PRIO_DEFAULT, NULL); @@ -1323,14 +1364,15 @@ static int sdma_v4_4_2_sw_init(void *handle) ring->ring_obj = NULL; ring->use_doorbell = true; - /* paging queue use same doorbell index/routing as gfx queue - * with 0x400 (4096 dwords) offset on second doorbell page + /* doorbell index of page queue is assigned right after + * gfx queue on the same instance */ - ring->doorbell_index = adev->doorbell_index.sdma_engine[i] << 1; - ring->doorbell_index += 0x400; - ring->vm_hub = AMDGPU_MMHUB_0; + ring->doorbell_index = + (adev->doorbell_index.sdma_engine[i] + 1) << 1; + ring->vm_hub = AMDGPU_MMHUB0(aid_id); - sprintf(ring->name, "page%d", i); + sprintf(ring->name, "page%d.%d", aid_id, + i % adev->sdma.num_inst_per_aid); r = amdgpu_ring_init(adev, ring, 1024, &adev->sdma.trap_irq, AMDGPU_SDMA_IRQ_INSTANCE0 + i, @@ -1340,6 +1382,11 @@ static int sdma_v4_4_2_sw_init(void *handle) } } + if (amdgpu_sdma_ras_sw_init(adev)) { + dev_err(adev->dev, "fail to initialize sdma ras block\n"); + return -EINVAL; + } + return r; } @@ -1366,14 +1413,13 @@ static int sdma_v4_4_2_hw_init(void *handle) { int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; + uint32_t inst_mask; - if (adev->flags & AMD_IS_APU) - amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, false); - + inst_mask = GENMASK(adev->sdma.num_instances - 1, 0); if (!amdgpu_sriov_vf(adev)) - sdma_v4_4_2_init_golden_registers(adev); + sdma_v4_4_2_inst_init_golden_registers(adev, inst_mask); - r = sdma_v4_4_2_start(adev); + r = sdma_v4_4_2_inst_start(adev, inst_mask); return r; } @@ -1381,26 +1427,36 @@ static int sdma_v4_4_2_hw_init(void *handle) static int sdma_v4_4_2_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + uint32_t inst_mask; int i; if (amdgpu_sriov_vf(adev)) return 0; - for (i = 0; i < adev->sdma.num_instances; i++) { - amdgpu_irq_put(adev, &adev->sdma.ecc_irq, - AMDGPU_SDMA_IRQ_INSTANCE0 + i); + inst_mask = GENMASK(adev->sdma.num_instances - 1, 0); + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA)) { + for (i = 0; i < adev->sdma.num_instances; i++) { + amdgpu_irq_put(adev, &adev->sdma.ecc_irq, + AMDGPU_SDMA_IRQ_INSTANCE0 + i); + } } - sdma_v4_4_2_ctx_switch_enable(adev, false); - sdma_v4_4_2_enable(adev, false); + sdma_v4_4_2_inst_ctx_switch_enable(adev, false, inst_mask); + sdma_v4_4_2_inst_enable(adev, false, inst_mask); return 0; } +static int sdma_v4_4_2_set_clockgating_state(void *handle, + enum amd_clockgating_state state); + static int sdma_v4_4_2_suspend(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + if (amdgpu_in_reset(adev)) + sdma_v4_4_2_set_clockgating_state(adev, AMD_CG_STATE_UNGATE); + return sdma_v4_4_2_hw_fini(adev); } @@ -1471,13 +1527,31 @@ static int sdma_v4_4_2_process_trap_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) { - uint32_t instance; + uint32_t instance, i; DRM_DEBUG("IH: SDMA trap\n"); instance = sdma_v4_4_2_irq_id_to_seq(entry->client_id); + + /* Client id gives the SDMA instance in AID. To know the exact SDMA + * instance, interrupt entry gives the node id which corresponds to the AID instance. + * Match node id with the AID id associated with the SDMA instance. */ + for (i = instance; i < adev->sdma.num_instances; + i += adev->sdma.num_inst_per_aid) { + if (adev->sdma.instance[i].aid_id == + node_id_to_phys_map[entry->node_id]) + break; + } + + if (i >= adev->sdma.num_instances) { + dev_WARN_ONCE( + adev->dev, 1, + "Couldn't find the right sdma instance in trap handler"); + return 0; + } + switch (entry->ring_id) { case 0: - amdgpu_fence_process(&adev->sdma.instance[instance].ring); + amdgpu_fence_process(&adev->sdma.instance[i].ring); break; default: break; @@ -1496,7 +1570,7 @@ static int sdma_v4_4_2_process_ras_data_cb(struct amdgpu_device *adev, * be disabled and the driver should only look for the aggregated * interrupt via sync flood */ - if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX)) + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA)) goto out; instance = sdma_v4_4_2_irq_id_to_seq(entry->client_id); @@ -1535,15 +1609,22 @@ static int sdma_v4_4_2_set_ecc_irq_state(struct amdgpu_device *adev, unsigned type, enum amdgpu_interrupt_state state) { - u32 sdma_edc_config; + u32 sdma_cntl; - sdma_edc_config = RREG32_SDMA(type, regCC_SDMA_EDC_CONFIG); - /* - * FIXME: This was inherited from Aldebaran, but no this field - * definition in the regspec of both Aldebaran and SDMA 4.4.2 - */ - sdma_edc_config |= (state == AMDGPU_IRQ_STATE_ENABLE) ? (1 << 2) : 0; - WREG32_SDMA(type, regCC_SDMA_EDC_CONFIG, sdma_edc_config); + sdma_cntl = RREG32_SDMA(type, regSDMA_CNTL); + switch (state) { + case AMDGPU_IRQ_STATE_DISABLE: + sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA_CNTL, + DRAM_ECC_INT_ENABLE, 0); + WREG32_SDMA(type, regSDMA_CNTL, sdma_cntl); + break; + /* sdma ecc interrupt is enabled by default + * driver doesn't need to do anything to + * enable the interrupt */ + case AMDGPU_IRQ_STATE_ENABLE: + default: + break; + } return 0; } @@ -1615,19 +1696,49 @@ static int sdma_v4_4_2_process_srbm_write_irq(struct amdgpu_device *adev, return 0; } -static void sdma_v4_4_2_update_medium_grain_clock_gating( - struct amdgpu_device *adev, - bool enable) +static void sdma_v4_4_2_inst_update_medium_grain_light_sleep( + struct amdgpu_device *adev, bool enable, uint32_t inst_mask) { uint32_t data, def; int i; - if (enable && (adev->cg_flags & AMD_CG_SUPPORT_SDMA_MGCG)) { - for (i = 0; i < adev->sdma.num_instances; i++) { + /* leave as default if it is not driver controlled */ + if (!(adev->cg_flags & AMD_CG_SUPPORT_SDMA_LS)) + return; + + if (enable) { + for_each_inst(i, inst_mask) { + /* 1-not override: enable sdma mem light sleep */ + def = data = RREG32_SDMA(i, regSDMA_POWER_CNTL); + data |= SDMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK; + if (def != data) + WREG32_SDMA(i, regSDMA_POWER_CNTL, data); + } + } else { + for_each_inst(i, inst_mask) { + /* 0-override:disable sdma mem light sleep */ + def = data = RREG32_SDMA(i, regSDMA_POWER_CNTL); + data &= ~SDMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK; + if (def != data) + WREG32_SDMA(i, regSDMA_POWER_CNTL, data); + } + } +} + +static void sdma_v4_4_2_inst_update_medium_grain_clock_gating( + struct amdgpu_device *adev, bool enable, uint32_t inst_mask) +{ + uint32_t data, def; + int i; + + /* leave as default if it is not driver controlled */ + if (!(adev->cg_flags & AMD_CG_SUPPORT_SDMA_MGCG)) + return; + + if (enable) { + for_each_inst(i, inst_mask) { def = data = RREG32_SDMA(i, regSDMA_CLK_CTRL); - data &= ~(SDMA_CLK_CTRL__SOFT_OVERRIDE7_MASK | - SDMA_CLK_CTRL__SOFT_OVERRIDE6_MASK | - SDMA_CLK_CTRL__SOFT_OVERRIDE5_MASK | + data &= ~(SDMA_CLK_CTRL__SOFT_OVERRIDE5_MASK | SDMA_CLK_CTRL__SOFT_OVERRIDE4_MASK | SDMA_CLK_CTRL__SOFT_OVERRIDE3_MASK | SDMA_CLK_CTRL__SOFT_OVERRIDE2_MASK | @@ -1637,11 +1748,9 @@ static void sdma_v4_4_2_update_medium_grain_clock_gating( WREG32_SDMA(i, regSDMA_CLK_CTRL, data); } } else { - for (i = 0; i < adev->sdma.num_instances; i++) { + for_each_inst(i, inst_mask) { def = data = RREG32_SDMA(i, regSDMA_CLK_CTRL); - data |= (SDMA_CLK_CTRL__SOFT_OVERRIDE7_MASK | - SDMA_CLK_CTRL__SOFT_OVERRIDE6_MASK | - SDMA_CLK_CTRL__SOFT_OVERRIDE5_MASK | + data |= (SDMA_CLK_CTRL__SOFT_OVERRIDE5_MASK | SDMA_CLK_CTRL__SOFT_OVERRIDE4_MASK | SDMA_CLK_CTRL__SOFT_OVERRIDE3_MASK | SDMA_CLK_CTRL__SOFT_OVERRIDE2_MASK | @@ -1653,45 +1762,21 @@ static void sdma_v4_4_2_update_medium_grain_clock_gating( } } - -static void sdma_v4_4_2_update_medium_grain_light_sleep( - struct amdgpu_device *adev, - bool enable) -{ - uint32_t data, def; - int i; - - if (enable && (adev->cg_flags & AMD_CG_SUPPORT_SDMA_LS)) { - for (i = 0; i < adev->sdma.num_instances; i++) { - /* 1-not override: enable sdma mem light sleep */ - def = data = RREG32_SDMA(0, regSDMA_POWER_CNTL); - data |= SDMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK; - if (def != data) - WREG32_SDMA(0, regSDMA_POWER_CNTL, data); - } - } else { - for (i = 0; i < adev->sdma.num_instances; i++) { - /* 0-override:disable sdma mem light sleep */ - def = data = RREG32_SDMA(0, regSDMA_POWER_CNTL); - data &= ~SDMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK; - if (def != data) - WREG32_SDMA(0, regSDMA_POWER_CNTL, data); - } - } -} - static int sdma_v4_4_2_set_clockgating_state(void *handle, enum amd_clockgating_state state) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + uint32_t inst_mask; if (amdgpu_sriov_vf(adev)) return 0; - sdma_v4_4_2_update_medium_grain_clock_gating(adev, - state == AMD_CG_STATE_GATE); - sdma_v4_4_2_update_medium_grain_light_sleep(adev, - state == AMD_CG_STATE_GATE); + inst_mask = GENMASK(adev->sdma.num_instances - 1, 0); + + sdma_v4_4_2_inst_update_medium_grain_clock_gating( + adev, state == AMD_CG_STATE_GATE, inst_mask); + sdma_v4_4_2_inst_update_medium_grain_light_sleep( + adev, state == AMD_CG_STATE_GATE, inst_mask); return 0; } @@ -1710,12 +1795,12 @@ static void sdma_v4_4_2_get_clockgating_state(void *handle, u64 *flags) *flags = 0; /* AMD_CG_SUPPORT_SDMA_MGCG */ - data = RREG32(SOC15_REG_OFFSET(SDMA0, 0, regSDMA_CLK_CTRL)); - if (!(data & SDMA_CLK_CTRL__SOFT_OVERRIDE7_MASK)) + data = RREG32(SOC15_REG_OFFSET(SDMA0, GET_INST(SDMA0, 0), regSDMA_CLK_CTRL)); + if (!(data & SDMA_CLK_CTRL__SOFT_OVERRIDE5_MASK)) *flags |= AMD_CG_SUPPORT_SDMA_MGCG; /* AMD_CG_SUPPORT_SDMA_LS */ - data = RREG32(SOC15_REG_OFFSET(SDMA0, 0, regSDMA_POWER_CNTL)); + data = RREG32(SOC15_REG_OFFSET(SDMA0, GET_INST(SDMA0, 0), regSDMA_POWER_CNTL)); if (data & SDMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK) *flags |= AMD_CG_SUPPORT_SDMA_LS; } @@ -1740,7 +1825,7 @@ const struct amd_ip_funcs sdma_v4_4_2_ip_funcs = { static const struct amdgpu_ring_funcs sdma_v4_4_2_ring_funcs = { .type = AMDGPU_RING_TYPE_SDMA, - .align_mask = 0xf, + .align_mask = 0xff, .nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP), .support_64bit_ptrs = true, .get_rptr = sdma_v4_4_2_ring_get_rptr, @@ -1771,7 +1856,7 @@ static const struct amdgpu_ring_funcs sdma_v4_4_2_ring_funcs = { static const struct amdgpu_ring_funcs sdma_v4_4_2_page_ring_funcs = { .type = AMDGPU_RING_TYPE_SDMA, - .align_mask = 0xf, + .align_mask = 0xff, .nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP), .support_64bit_ptrs = true, .get_rptr = sdma_v4_4_2_ring_get_rptr, @@ -1802,7 +1887,7 @@ static const struct amdgpu_ring_funcs sdma_v4_4_2_page_ring_funcs = { static void sdma_v4_4_2_set_ring_funcs(struct amdgpu_device *adev) { - int i; + int i, dev_inst; for (i = 0; i < adev->sdma.num_instances; i++) { adev->sdma.instance[i].ring.funcs = &sdma_v4_4_2_ring_funcs; @@ -1812,6 +1897,11 @@ static void sdma_v4_4_2_set_ring_funcs(struct amdgpu_device *adev) &sdma_v4_4_2_page_ring_funcs; adev->sdma.instance[i].page.me = i; } + + dev_inst = GET_INST(SDMA0, i); + /* AID to which SDMA belongs depends on physical instance */ + adev->sdma.instance[i].aid_id = + dev_inst / adev->sdma.num_inst_per_aid; } } @@ -1965,3 +2055,146 @@ const struct amdgpu_ip_block_version sdma_v4_4_2_ip_block = { .rev = 0, .funcs = &sdma_v4_4_2_ip_funcs, }; + +static int sdma_v4_4_2_xcp_resume(void *handle, uint32_t inst_mask) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int r; + + if (!amdgpu_sriov_vf(adev)) + sdma_v4_4_2_inst_init_golden_registers(adev, inst_mask); + + r = sdma_v4_4_2_inst_start(adev, inst_mask); + + return r; +} + +static int sdma_v4_4_2_xcp_suspend(void *handle, uint32_t inst_mask) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + uint32_t tmp_mask = inst_mask; + int i; + + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA)) { + for_each_inst(i, tmp_mask) { + amdgpu_irq_put(adev, &adev->sdma.ecc_irq, + AMDGPU_SDMA_IRQ_INSTANCE0 + i); + } + } + + sdma_v4_4_2_inst_ctx_switch_enable(adev, false, inst_mask); + sdma_v4_4_2_inst_enable(adev, false, inst_mask); + + return 0; +} + +struct amdgpu_xcp_ip_funcs sdma_v4_4_2_xcp_funcs = { + .suspend = &sdma_v4_4_2_xcp_suspend, + .resume = &sdma_v4_4_2_xcp_resume +}; + +static const struct amdgpu_ras_err_status_reg_entry sdma_v4_2_2_ue_reg_list[] = { + {AMDGPU_RAS_REG_ENTRY(SDMA0, 0, regSDMA_UE_ERR_STATUS_LO, regSDMA_UE_ERR_STATUS_HI), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SDMA"}, +}; + +static const struct amdgpu_ras_memory_id_entry sdma_v4_4_2_ras_memory_list[] = { + {AMDGPU_SDMA_MBANK_DATA_BUF0, "SDMA_MBANK_DATA_BUF0"}, + {AMDGPU_SDMA_MBANK_DATA_BUF1, "SDMA_MBANK_DATA_BUF1"}, + {AMDGPU_SDMA_MBANK_DATA_BUF2, "SDMA_MBANK_DATA_BUF2"}, + {AMDGPU_SDMA_MBANK_DATA_BUF3, "SDMA_MBANK_DATA_BUF3"}, + {AMDGPU_SDMA_MBANK_DATA_BUF4, "SDMA_MBANK_DATA_BUF4"}, + {AMDGPU_SDMA_MBANK_DATA_BUF5, "SDMA_MBANK_DATA_BUF5"}, + {AMDGPU_SDMA_MBANK_DATA_BUF6, "SDMA_MBANK_DATA_BUF6"}, + {AMDGPU_SDMA_MBANK_DATA_BUF7, "SDMA_MBANK_DATA_BUF7"}, + {AMDGPU_SDMA_MBANK_DATA_BUF8, "SDMA_MBANK_DATA_BUF8"}, + {AMDGPU_SDMA_MBANK_DATA_BUF9, "SDMA_MBANK_DATA_BUF9"}, + {AMDGPU_SDMA_MBANK_DATA_BUF10, "SDMA_MBANK_DATA_BUF10"}, + {AMDGPU_SDMA_MBANK_DATA_BUF11, "SDMA_MBANK_DATA_BUF11"}, + {AMDGPU_SDMA_MBANK_DATA_BUF12, "SDMA_MBANK_DATA_BUF12"}, + {AMDGPU_SDMA_MBANK_DATA_BUF13, "SDMA_MBANK_DATA_BUF13"}, + {AMDGPU_SDMA_MBANK_DATA_BUF14, "SDMA_MBANK_DATA_BUF14"}, + {AMDGPU_SDMA_MBANK_DATA_BUF15, "SDMA_MBANK_DATA_BUF15"}, + {AMDGPU_SDMA_UCODE_BUF, "SDMA_UCODE_BUF"}, + {AMDGPU_SDMA_RB_CMD_BUF, "SDMA_RB_CMD_BUF"}, + {AMDGPU_SDMA_IB_CMD_BUF, "SDMA_IB_CMD_BUF"}, + {AMDGPU_SDMA_UTCL1_RD_FIFO, "SDMA_UTCL1_RD_FIFO"}, + {AMDGPU_SDMA_UTCL1_RDBST_FIFO, "SDMA_UTCL1_RDBST_FIFO"}, + {AMDGPU_SDMA_UTCL1_WR_FIFO, "SDMA_UTCL1_WR_FIFO"}, + {AMDGPU_SDMA_DATA_LUT_FIFO, "SDMA_DATA_LUT_FIFO"}, + {AMDGPU_SDMA_SPLIT_DAT_BUF, "SDMA_SPLIT_DAT_BUF"}, +}; + +static void sdma_v4_4_2_inst_query_ras_error_count(struct amdgpu_device *adev, + uint32_t sdma_inst, + void *ras_err_status) +{ + struct ras_err_data *err_data = (struct ras_err_data *)ras_err_status; + uint32_t sdma_dev_inst = GET_INST(SDMA0, sdma_inst); + + /* sdma v4_4_2 doesn't support query ce counts */ + amdgpu_ras_inst_query_ras_error_count(adev, + sdma_v4_2_2_ue_reg_list, + ARRAY_SIZE(sdma_v4_2_2_ue_reg_list), + sdma_v4_4_2_ras_memory_list, + ARRAY_SIZE(sdma_v4_4_2_ras_memory_list), + sdma_dev_inst, + AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE, + &err_data->ue_count); +} + +static void sdma_v4_4_2_query_ras_error_count(struct amdgpu_device *adev, + void *ras_err_status) +{ + uint32_t inst_mask; + int i = 0; + + inst_mask = GENMASK(adev->sdma.num_instances - 1, 0); + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA)) { + for_each_inst(i, inst_mask) + sdma_v4_4_2_inst_query_ras_error_count(adev, i, ras_err_status); + } else { + dev_warn(adev->dev, "SDMA RAS is not supported\n"); + } +} + +static void sdma_v4_4_2_inst_reset_ras_error_count(struct amdgpu_device *adev, + uint32_t sdma_inst) +{ + uint32_t sdma_dev_inst = GET_INST(SDMA0, sdma_inst); + + amdgpu_ras_inst_reset_ras_error_count(adev, + sdma_v4_2_2_ue_reg_list, + ARRAY_SIZE(sdma_v4_2_2_ue_reg_list), + sdma_dev_inst); +} + +static void sdma_v4_4_2_reset_ras_error_count(struct amdgpu_device *adev) +{ + uint32_t inst_mask; + int i = 0; + + inst_mask = GENMASK(adev->sdma.num_instances - 1, 0); + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA)) { + for_each_inst(i, inst_mask) + sdma_v4_4_2_inst_reset_ras_error_count(adev, i); + } else { + dev_warn(adev->dev, "SDMA RAS is not supported\n"); + } +} + +static const struct amdgpu_ras_block_hw_ops sdma_v4_4_2_ras_hw_ops = { + .query_ras_error_count = sdma_v4_4_2_query_ras_error_count, + .reset_ras_error_count = sdma_v4_4_2_reset_ras_error_count, +}; + +static struct amdgpu_sdma_ras sdma_v4_4_2_ras = { + .ras_block = { + .hw_ops = &sdma_v4_4_2_ras_hw_ops, + }, +}; + +static void sdma_v4_4_2_set_ras_funcs(struct amdgpu_device *adev) +{ + adev->sdma.ras = &sdma_v4_4_2_ras; +} diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.h b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.h index 4814e8a074d6..d516145529bb 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.h +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.h @@ -27,4 +27,6 @@ extern const struct amd_ip_funcs sdma_v4_4_2_ip_funcs; extern const struct amdgpu_ip_block_version sdma_v4_4_2_ip_block; +extern struct amdgpu_xcp_ip_funcs sdma_v4_4_2_xcp_funcs; + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c index 92e1299be021..5c4d4df9cf94 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c @@ -819,8 +819,6 @@ static int sdma_v5_0_gfx_resume(struct amdgpu_device *adev) /* enable DMA IBs */ WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL), ib_cntl); - ring->sched.ready = true; - if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */ sdma_v5_0_ctx_switch_enable(adev, true); sdma_v5_0_enable(adev, true); @@ -1389,7 +1387,7 @@ static int sdma_v5_0_sw_init(void *handle) (adev->doorbell_index.sdma_engine[0] << 1) //get DWORD offset : (adev->doorbell_index.sdma_engine[1] << 1); // get DWORD offset - ring->vm_hub = AMDGPU_GFXHUB_0; + ring->vm_hub = AMDGPU_GFXHUB(0); sprintf(ring->name, "sdma%d", i); r = amdgpu_ring_init(adev, ring, 1024, &adev->sdma.trap_irq, (i == 0) ? AMDGPU_SDMA_IRQ_INSTANCE0 : diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c index ca7e8757d78e..a7b230e5a26d 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c @@ -617,18 +617,14 @@ static int sdma_v5_2_gfx_resume(struct amdgpu_device *adev) /* enable DMA IBs */ WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL), ib_cntl); - ring->sched.ready = true; - if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */ sdma_v5_2_ctx_switch_enable(adev, true); sdma_v5_2_enable(adev, true); } - r = amdgpu_ring_test_ring(ring); - if (r) { - ring->sched.ready = false; + r = amdgpu_ring_test_helper(ring); + if (r) return r; - } if (adev->mman.buffer_funcs_ring == ring) amdgpu_ttm_set_buffer_funcs_status(adev, true); @@ -1253,7 +1249,7 @@ static int sdma_v5_2_sw_init(void *handle) ring->doorbell_index = (adev->doorbell_index.sdma_engine[i] << 1); //get DWORD offset - ring->vm_hub = AMDGPU_GFXHUB_0; + ring->vm_hub = AMDGPU_GFXHUB(0); sprintf(ring->name, "sdma%d", i); r = amdgpu_ring_init(adev, ring, 1024, &adev->sdma.trap_irq, AMDGPU_SDMA_IRQ_INSTANCE0 + i, diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c index 3d9a80511a45..3b03dda854fd 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c @@ -238,6 +238,8 @@ static void sdma_v6_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count) * * @ring: amdgpu ring pointer * @ib: IB object to schedule + * @flags: unused + * @job: job to retrieve vmid from * * Schedule an IB in the DMA ring. */ @@ -585,16 +587,12 @@ static int sdma_v6_0_gfx_resume(struct amdgpu_device *adev) /* enable DMA IBs */ WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL), ib_cntl); - ring->sched.ready = true; - if (amdgpu_sriov_vf(adev)) sdma_v6_0_enable(adev, true); r = amdgpu_ring_test_helper(ring); - if (r) { - ring->sched.ready = false; + if (r) return r; - } if (adev->mman.buffer_funcs_ring == ring) amdgpu_ttm_set_buffer_funcs_status(adev, true); @@ -942,6 +940,7 @@ static int sdma_v6_0_ring_test_ring(struct amdgpu_ring *ring) * sdma_v6_0_ring_test_ib - test an IB on the DMA engine * * @ring: amdgpu_ring structure holding ring information + * @timeout: timeout value in jiffies, or MAX_SCHEDULE_TIMEOUT * * Test a simple IB in the DMA ring. * Returns 0 on success, error on failure. @@ -1122,6 +1121,7 @@ static void sdma_v6_0_vm_set_pte_pde(struct amdgpu_ib *ib, /** * sdma_v6_0_ring_pad_ib - pad the IB * @ib: indirect buffer to fill with padding + * @ring: amdgpu ring pointer * * Pad the IB with NOPs to a boundary multiple of 8. */ @@ -1171,6 +1171,8 @@ static void sdma_v6_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring) * sdma_v6_0_ring_emit_vm_flush - vm flush using sDMA * * @ring: amdgpu_ring pointer + * @vmid: vmid number to use + * @pd_addr: address * * Update the page table base and flush the VM TLB * using sDMA. @@ -1298,7 +1300,7 @@ static int sdma_v6_0_sw_init(void *handle) ring->doorbell_index = (adev->doorbell_index.sdma_engine[i] << 1); // get DWORD offset - ring->vm_hub = AMDGPU_GFXHUB_0; + ring->vm_hub = AMDGPU_GFXHUB(0); sprintf(ring->name, "sdma%d", i); r = amdgpu_ring_init(adev, ring, 1024, &adev->sdma.trap_irq, diff --git a/drivers/gpu/drm/amd/amdgpu/si.c b/drivers/gpu/drm/amd/amdgpu/si.c index 7f99e130acd0..f64b87b11b1b 100644 --- a/drivers/gpu/drm/amd/amdgpu/si.c +++ b/drivers/gpu/drm/amd/amdgpu/si.c @@ -1181,12 +1181,12 @@ static uint32_t si_get_register_value(struct amdgpu_device *adev, mutex_lock(&adev->grbm_idx_mutex); if (se_num != 0xffffffff || sh_num != 0xffffffff) - amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0); val = RREG32(reg_offset); if (se_num != 0xffffffff || sh_num != 0xffffffff) - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); return val; } else { diff --git a/drivers/gpu/drm/amd/amdgpu/si_dma.c b/drivers/gpu/drm/amd/amdgpu/si_dma.c index abca8b529721..42c4547f32ec 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_dma.c +++ b/drivers/gpu/drm/amd/amdgpu/si_dma.c @@ -174,8 +174,6 @@ static int si_dma_start(struct amdgpu_device *adev) WREG32(DMA_RB_WPTR + sdma_offsets[i], ring->wptr << 2); WREG32(DMA_RB_CNTL + sdma_offsets[i], rb_cntl | DMA_RB_ENABLE); - ring->sched.ready = true; - r = amdgpu_ring_test_helper(ring); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_3.c b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_3.c new file mode 100644 index 000000000000..4368a5891eeb --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_3.c @@ -0,0 +1,103 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#include "amdgpu.h" +#include "smuio_v13_0_3.h" +#include "soc15_common.h" +#include "smuio/smuio_13_0_3_offset.h" +#include "smuio/smuio_13_0_3_sh_mask.h" + +#define PKG_TYPE_MASK 0x00000003L + +/** + * smuio_v13_0_3_get_die_id - query die id from FCH. + * + * @adev: amdgpu device pointer + * + * Returns die id + */ +static u32 smuio_v13_0_3_get_die_id(struct amdgpu_device *adev) +{ + u32 data, die_id; + + data = RREG32_SOC15(SMUIO, 0, regSMUIO_MCM_CONFIG); + die_id = REG_GET_FIELD(data, SMUIO_MCM_CONFIG, DIE_ID); + + return die_id; +} + +/** + * smuio_v13_0_3_get_socket_id - query socket id from FCH + * + * @adev: amdgpu device pointer + * + * Returns socket id + */ +static u32 smuio_v13_0_3_get_socket_id(struct amdgpu_device *adev) +{ + u32 data, socket_id; + + data = RREG32_SOC15(SMUIO, 0, regSMUIO_MCM_CONFIG); + socket_id = REG_GET_FIELD(data, SMUIO_MCM_CONFIG, SOCKET_ID); + + return socket_id; +} + +/** + * smuio_v13_0_3_get_pkg_type - query package type set by MP1/bootcode + * + * @adev: amdgpu device pointer + * + * Returns package type + */ + +static enum amdgpu_pkg_type smuio_v13_0_3_get_pkg_type(struct amdgpu_device *adev) +{ + enum amdgpu_pkg_type pkg_type; + u32 data; + + data = RREG32_SOC15(SMUIO, 0, regSMUIO_MCM_CONFIG); + data = REG_GET_FIELD(data, SMUIO_MCM_CONFIG, PKG_TYPE); + /* pkg_type[4:0] + * + * bit 1 == 1 APU form factor + * + * b0100 - b1111 - Reserved + */ + switch (data & PKG_TYPE_MASK) { + case 0x2: + pkg_type = AMDGPU_PKG_TYPE_APU; + break; + default: + pkg_type = AMDGPU_PKG_TYPE_UNKNOWN; + break; + } + + return pkg_type; +} + + +const struct amdgpu_smuio_funcs smuio_v13_0_3_funcs = { + .get_die_id = smuio_v13_0_3_get_die_id, + .get_socket_id = smuio_v13_0_3_get_socket_id, + .get_pkg_type = smuio_v13_0_3_get_pkg_type, +}; diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_3.h b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_3.h new file mode 100644 index 000000000000..795f66c5a58b --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_3.h @@ -0,0 +1,30 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#ifndef __SMUIO_V13_0_3_H__ +#define __SMUIO_V13_0_3_H__ + +#include "soc15_common.h" + +extern const struct amdgpu_smuio_funcs smuio_v13_0_3_funcs; + +#endif /* __SMUIO_V13_0_3_H__ */ diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c index 6d15d5cd9e07..afcaeadda4c7 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.c +++ b/drivers/gpu/drm/amd/amdgpu/soc15.c @@ -153,6 +153,24 @@ static const struct amdgpu_video_codecs rn_video_codecs_decode = .codec_array = rn_video_codecs_decode_array, }; +static const struct amdgpu_video_codec_info vcn_4_0_3_video_codecs_decode_array[] = { + {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)}, + {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)}, + {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)}, + {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)}, + {codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)}, +}; + +static const struct amdgpu_video_codecs vcn_4_0_3_video_codecs_decode = { + .codec_count = ARRAY_SIZE(vcn_4_0_3_video_codecs_decode_array), + .codec_array = vcn_4_0_3_video_codecs_decode_array, +}; + +static const struct amdgpu_video_codecs vcn_4_0_3_video_codecs_encode = { + .codec_count = 0, + .codec_array = NULL, +}; + static int soc15_query_video_codecs(struct amdgpu_device *adev, bool encode, const struct amdgpu_video_codecs **codecs) { @@ -185,6 +203,12 @@ static int soc15_query_video_codecs(struct amdgpu_device *adev, bool encode, else *codecs = &rn_video_codecs_decode; return 0; + case IP_VERSION(4, 0, 3): + if (encode) + *codecs = &vcn_4_0_3_video_codecs_encode; + else + *codecs = &vcn_4_0_3_video_codecs_decode; + return 0; default: return -EINVAL; } @@ -301,17 +325,18 @@ static u32 soc15_get_xclk(struct amdgpu_device *adev) u32 reference_clock = adev->clock.spll.reference_freq; if (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(12, 0, 0) || - adev->ip_versions[MP1_HWIP][0] == IP_VERSION(12, 0, 1) || - adev->ip_versions[MP1_HWIP][0] == IP_VERSION(10, 0, 0) || - adev->ip_versions[MP1_HWIP][0] == IP_VERSION(10, 0, 1)) + adev->ip_versions[MP1_HWIP][0] == IP_VERSION(12, 0, 1)) return 10000; + if (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(10, 0, 0) || + adev->ip_versions[MP1_HWIP][0] == IP_VERSION(10, 0, 1)) + return reference_clock / 4; return reference_clock; } void soc15_grbm_select(struct amdgpu_device *adev, - u32 me, u32 pipe, u32 queue, u32 vmid) + u32 me, u32 pipe, u32 queue, u32 vmid, int xcc_id) { u32 grbm_gfx_cntl = 0; grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, PIPEID, pipe); @@ -319,12 +344,7 @@ void soc15_grbm_select(struct amdgpu_device *adev, grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, VMID, vmid); grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, QUEUEID, queue); - WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_CNTL, grbm_gfx_cntl); -} - -static void soc15_vga_set_state(struct amdgpu_device *adev, bool state) -{ - /* todo */ + WREG32_SOC15_RLC_SHADOW(GC, xcc_id, mmGRBM_GFX_CNTL, grbm_gfx_cntl); } static bool soc15_read_disabled_bios(struct amdgpu_device *adev) @@ -363,12 +383,12 @@ static uint32_t soc15_read_indexed_register(struct amdgpu_device *adev, u32 se_n mutex_lock(&adev->grbm_idx_mutex); if (se_num != 0xffffffff || sh_num != 0xffffffff) - amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0); val = RREG32(reg_offset); if (se_num != 0xffffffff || sh_num != 0xffffffff) - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); return val; } @@ -532,6 +552,15 @@ soc15_asic_reset_method(struct amdgpu_device *adev) if (connected_to_cpu) return AMD_RESET_METHOD_MODE2; break; + case IP_VERSION(13, 0, 6): + /* Use gpu_recovery param to target a reset method. + * Enable triggering of GPU reset only if specified + * by module parameter. + */ + if (amdgpu_gpu_recovery == 4 || amdgpu_gpu_recovery == 5) + return AMD_RESET_METHOD_MODE2; + else + return AMD_RESET_METHOD_NONE; default: break; } @@ -816,7 +845,6 @@ static const struct amdgpu_asic_funcs soc15_asic_funcs = .read_register = &soc15_read_register, .reset = &soc15_asic_reset, .reset_method = &soc15_asic_reset_method, - .set_vga_state = &soc15_vga_set_state, .get_xclk = &soc15_get_xclk, .set_uvd_clocks = &soc15_set_uvd_clocks, .set_vce_clocks = &soc15_set_vce_clocks, @@ -838,7 +866,6 @@ static const struct amdgpu_asic_funcs vega20_asic_funcs = .read_register = &soc15_read_register, .reset = &soc15_asic_reset, .reset_method = &soc15_asic_reset_method, - .set_vga_state = &soc15_vga_set_state, .get_xclk = &soc15_get_xclk, .set_uvd_clocks = &soc15_set_uvd_clocks, .set_vce_clocks = &soc15_set_vce_clocks, @@ -853,6 +880,28 @@ static const struct amdgpu_asic_funcs vega20_asic_funcs = .query_video_codecs = &soc15_query_video_codecs, }; +static const struct amdgpu_asic_funcs aqua_vanjaram_asic_funcs = +{ + .read_disabled_bios = &soc15_read_disabled_bios, + .read_bios_from_rom = &amdgpu_soc15_read_bios_from_rom, + .read_register = &soc15_read_register, + .reset = &soc15_asic_reset, + .reset_method = &soc15_asic_reset_method, + .get_xclk = &soc15_get_xclk, + .set_uvd_clocks = &soc15_set_uvd_clocks, + .set_vce_clocks = &soc15_set_vce_clocks, + .get_config_memsize = &soc15_get_config_memsize, + .need_full_reset = &soc15_need_full_reset, + .init_doorbell_index = &aqua_vanjaram_doorbell_index_init, + .get_pcie_usage = &vega20_get_pcie_usage, + .need_reset_on_init = &soc15_need_reset_on_init, + .get_pcie_replay_count = &soc15_get_pcie_replay_count, + .supports_baco = &soc15_supports_baco, + .pre_asic_init = &soc15_pre_asic_init, + .query_video_codecs = &soc15_query_video_codecs, + .encode_ext_smn_addressing = &aqua_vanjaram_encode_ext_smn_addressing, +}; + static int soc15_common_early_init(void *handle) { #define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE) @@ -866,6 +915,8 @@ static int soc15_common_early_init(void *handle) adev->smc_wreg = NULL; adev->pcie_rreg = &amdgpu_device_indirect_rreg; adev->pcie_wreg = &amdgpu_device_indirect_wreg; + adev->pcie_rreg_ext = &amdgpu_device_indirect_rreg_ext; + adev->pcie_wreg_ext = &amdgpu_device_indirect_wreg_ext; adev->pcie_rreg64 = &amdgpu_device_indirect_rreg64; adev->pcie_wreg64 = &amdgpu_device_indirect_wreg64; adev->uvd_ctx_rreg = &soc15_uvd_ctx_rreg; @@ -1094,9 +1145,18 @@ static int soc15_common_early_init(void *handle) adev->external_rev_id = adev->rev_id + 0x3c; break; case IP_VERSION(9, 4, 3): - adev->asic_funcs = &vega20_asic_funcs; - adev->cg_flags = 0; - adev->pg_flags = 0; + adev->asic_funcs = &aqua_vanjaram_asic_funcs; + adev->cg_flags = + AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_CGCG | + AMD_CG_SUPPORT_GFX_CGLS | AMD_CG_SUPPORT_SDMA_MGCG | + AMD_CG_SUPPORT_GFX_FGCG | AMD_CG_SUPPORT_REPEATER_FGCG | + AMD_CG_SUPPORT_VCN_MGCG | AMD_CG_SUPPORT_JPEG_MGCG | + AMD_CG_SUPPORT_IH_CG; + adev->pg_flags = + AMD_PG_SUPPORT_VCN | + AMD_PG_SUPPORT_VCN_DPG | + AMD_PG_SUPPORT_JPEG; + adev->external_rev_id = adev->rev_id + 0x46; break; default: /* FIXME: not supported yet */ diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.h b/drivers/gpu/drm/amd/amdgpu/soc15.h index efc2a253e8db..eac54042c6c0 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.h +++ b/drivers/gpu/drm/amd/amdgpu/soc15.h @@ -100,7 +100,7 @@ struct soc15_ras_field_entry { #define SOC15_RAS_REG_FIELD_VAL(val, entry, field) SOC15_REG_FIELD_VAL((val), (entry).field##_count_mask, (entry).field##_count_shift) void soc15_grbm_select(struct amdgpu_device *adev, - u32 me, u32 pipe, u32 queue, u32 vmid); + u32 me, u32 pipe, u32 queue, u32 vmid, int xcc_id); void soc15_set_virt_ops(struct amdgpu_device *adev); void soc15_program_register_sequence(struct amdgpu_device *adev, @@ -111,7 +111,11 @@ int vega10_reg_base_init(struct amdgpu_device *adev); int vega20_reg_base_init(struct amdgpu_device *adev); int arct_reg_base_init(struct amdgpu_device *adev); int aldebaran_reg_base_init(struct amdgpu_device *adev); +void aqua_vanjaram_ip_map_init(struct amdgpu_device *adev); +u64 aqua_vanjaram_encode_ext_smn_addressing(int ext_id); +int aqua_vanjaram_init_soc_config(struct amdgpu_device *adev); void vega10_doorbell_index_init(struct amdgpu_device *adev); void vega20_doorbell_index_init(struct amdgpu_device *adev); +void aqua_vanjaram_doorbell_index_init(struct amdgpu_device *adev); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/soc15_common.h b/drivers/gpu/drm/amd/amdgpu/soc15_common.h index 9fefd403e14f..96948a59f8dd 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15_common.h +++ b/drivers/gpu/drm/amd/amdgpu/soc15_common.h @@ -24,8 +24,18 @@ #ifndef __SOC15_COMMON_H__ #define __SOC15_COMMON_H__ +/* GET_INST returns the physical instance corresponding to a logical instance */ +#define GET_INST(ip, inst) \ + (adev->ip_map.logical_to_dev_inst ? \ + adev->ip_map.logical_to_dev_inst(adev, ip##_HWIP, inst) : inst) +#define GET_MASK(ip, mask) \ + (adev->ip_map.logical_to_dev_mask ? \ + adev->ip_map.logical_to_dev_mask(adev, ip##_HWIP, mask) : mask) + /* Register Access Macros */ #define SOC15_REG_OFFSET(ip, inst, reg) (adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) +#define SOC15_REG_OFFSET1(ip, inst, reg, offset) \ + (adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + (reg)+(offset)) #define __WREG32_SOC15_RLC__(reg, value, flag, hwip) \ ((amdgpu_sriov_vf(adev) && adev->gfx.rlc.funcs && adev->gfx.rlc.rlcg_reg_access_supported) ? \ @@ -66,7 +76,8 @@ AMDGPU_REGS_NO_KIQ, ip##_HWIP) #define RREG32_SOC15_OFFSET(ip, inst, reg, offset) \ - __RREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, 0, ip##_HWIP) + __RREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + (reg)) + \ + (offset), 0, ip##_HWIP) #define WREG32_SOC15(ip, inst, reg, value) \ __WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg), \ @@ -86,31 +97,15 @@ __WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, \ value, 0, ip##_HWIP) -#define SOC15_WAIT_ON_RREG(ip, inst, reg, expected_value, mask) \ -({ int ret = 0; \ - do { \ - uint32_t old_ = 0; \ - uint32_t tmp_ = RREG32(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg); \ - uint32_t loop = adev->usec_timeout; \ - ret = 0; \ - while ((tmp_ & (mask)) != (expected_value)) { \ - if (old_ != tmp_) { \ - loop = adev->usec_timeout; \ - old_ = tmp_; \ - } else \ - udelay(1); \ - tmp_ = RREG32(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg); \ - loop--; \ - if (!loop) { \ - DRM_WARN("Register(%d) [%s] failed to reach value 0x%08x != 0x%08x\n", \ - inst, #reg, (unsigned)expected_value, (unsigned)(tmp_ & (mask))); \ - ret = -ETIMEDOUT; \ - break; \ - } \ - } \ - } while (0); \ - ret; \ -}) +#define SOC15_WAIT_ON_RREG(ip, inst, reg, expected_value, mask) \ + amdgpu_device_wait_on_rreg(adev, inst, \ + (adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + (reg)), \ + #reg, expected_value, mask) + +#define SOC15_WAIT_ON_RREG_OFFSET(ip, inst, reg, offset, expected_value, mask) \ + amdgpu_device_wait_on_rreg(adev, inst, \ + (adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + (reg) + (offset)), \ + #reg, expected_value, mask) #define WREG32_RLC(reg, value) \ __WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_RLC, GC_HWIP) @@ -157,10 +152,10 @@ do { \ uint32_t target_reg = adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg;\ if (amdgpu_sriov_fullaccess(adev)) { \ - uint32_t r2 = adev->reg_offset[GC_HWIP][0][prefix##SCRATCH_REG1_BASE_IDX] + prefix##SCRATCH_REG2; \ - uint32_t r3 = adev->reg_offset[GC_HWIP][0][prefix##SCRATCH_REG1_BASE_IDX] + prefix##SCRATCH_REG3; \ - uint32_t grbm_cntl = adev->reg_offset[GC_HWIP][0][prefix##GRBM_GFX_CNTL_BASE_IDX] + prefix##GRBM_GFX_CNTL; \ - uint32_t grbm_idx = adev->reg_offset[GC_HWIP][0][prefix##GRBM_GFX_INDEX_BASE_IDX] + prefix##GRBM_GFX_INDEX; \ + uint32_t r2 = adev->reg_offset[GC_HWIP][inst][prefix##SCRATCH_REG1_BASE_IDX] + prefix##SCRATCH_REG2; \ + uint32_t r3 = adev->reg_offset[GC_HWIP][inst][prefix##SCRATCH_REG1_BASE_IDX] + prefix##SCRATCH_REG3; \ + uint32_t grbm_cntl = adev->reg_offset[GC_HWIP][inst][prefix##GRBM_GFX_CNTL_BASE_IDX] + prefix##GRBM_GFX_CNTL; \ + uint32_t grbm_idx = adev->reg_offset[GC_HWIP][inst][prefix##GRBM_GFX_INDEX_BASE_IDX] + prefix##GRBM_GFX_INDEX; \ if (target_reg == grbm_cntl) \ WREG32(r2, value); \ else if (target_reg == grbm_idx) \ @@ -176,13 +171,13 @@ #define WREG32_SOC15_RLC(ip, inst, reg, value) \ do { \ - uint32_t target_reg = adev->reg_offset[ip##_HWIP][0][reg##_BASE_IDX] + reg;\ + uint32_t target_reg = adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg;\ __WREG32_SOC15_RLC__(target_reg, value, AMDGPU_REGS_RLC, ip##_HWIP); \ } while (0) #define WREG32_SOC15_RLC_EX(prefix, ip, inst, reg, value) \ do { \ - uint32_t target_reg = adev->reg_offset[GC_HWIP][0][reg##_BASE_IDX] + reg;\ + uint32_t target_reg = adev->reg_offset[GC_HWIP][inst][reg##_BASE_IDX] + reg;\ WREG32_RLC_EX(prefix, target_reg, value); \ } while (0) @@ -199,4 +194,14 @@ #define RREG32_SOC15_OFFSET_RLC(ip, inst, reg, offset) \ __RREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, AMDGPU_REGS_RLC, ip##_HWIP) +/* inst equals to ext for some IPs */ +#define RREG32_SOC15_EXT(ip, inst, reg, ext) \ + RREG32_PCIE_EXT((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) * 4 \ + + adev->asic_funcs->encode_ext_smn_addressing(ext)) \ + +#define WREG32_SOC15_EXT(ip, inst, reg, ext, value) \ + WREG32_PCIE_EXT((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) * 4 \ + + adev->asic_funcs->encode_ext_smn_addressing(ext), \ + value) \ + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.c b/drivers/gpu/drm/amd/amdgpu/soc21.c index d77162536514..e5e5d68a4d70 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc21.c +++ b/drivers/gpu/drm/amd/amdgpu/soc21.c @@ -248,11 +248,6 @@ void soc21_grbm_select(struct amdgpu_device *adev, WREG32_SOC15(GC, 0, regGRBM_GFX_CNTL, grbm_gfx_cntl); } -static void soc21_vga_set_state(struct amdgpu_device *adev, bool state) -{ - /* todo */ -} - static bool soc21_read_disabled_bios(struct amdgpu_device *adev) { /* todo */ @@ -288,12 +283,12 @@ static uint32_t soc21_read_indexed_register(struct amdgpu_device *adev, u32 se_n mutex_lock(&adev->grbm_idx_mutex); if (se_num != 0xffffffff || sh_num != 0xffffffff) - amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0); val = RREG32(reg_offset); if (se_num != 0xffffffff || sh_num != 0xffffffff) - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); return val; } @@ -542,9 +537,9 @@ static int soc21_update_umd_stable_pstate(struct amdgpu_device *adev, bool enter) { if (enter) - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); else - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); if (adev->gfx.funcs->update_perfmon_mgcg) adev->gfx.funcs->update_perfmon_mgcg(adev, !enter); @@ -559,7 +554,6 @@ static const struct amdgpu_asic_funcs soc21_asic_funcs = .read_register = &soc21_read_register, .reset = &soc21_asic_reset, .reset_method = &soc21_asic_reset_method, - .set_vga_state = &soc21_vga_set_state, .get_xclk = &soc21_get_xclk, .set_uvd_clocks = &soc21_set_uvd_clocks, .set_vce_clocks = &soc21_set_vce_clocks, diff --git a/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h b/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h index 30d0482ac466..879bb7af297c 100644 --- a/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h +++ b/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h @@ -129,6 +129,8 @@ struct ta_ras_trigger_error_input { struct ta_ras_init_flags { uint8_t poison_mode_en; uint8_t dgpu_mode; + uint16_t xcc_mask; + uint8_t channel_dis_num; }; struct ta_ras_output_flags { diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_10.c b/drivers/gpu/drm/amd/amdgpu/umc_v8_10.c index d51ae0bc36f7..46bfdee79bfd 100644 --- a/drivers/gpu/drm/amd/amdgpu/umc_v8_10.c +++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_10.c @@ -444,6 +444,11 @@ static void umc_v8_10_ecc_info_query_ras_error_address(struct amdgpu_device *ade umc_v8_10_ecc_info_query_error_address, ras_error_status); } +static void umc_v8_10_set_eeprom_table_version(struct amdgpu_ras_eeprom_table_header *hdr) +{ + hdr->version = RAS_TABLE_VER_V2_1; +} + const struct amdgpu_ras_block_hw_ops umc_v8_10_ras_hw_ops = { .query_ras_error_count = umc_v8_10_query_ras_error_count, .query_ras_error_address = umc_v8_10_query_ras_error_address, @@ -457,4 +462,5 @@ struct amdgpu_umc_ras umc_v8_10_ras = { .query_ras_poison_mode = umc_v8_10_query_ras_poison_mode, .ecc_info_query_ras_error_count = umc_v8_10_ecc_info_query_ras_error_count, .ecc_info_query_ras_error_address = umc_v8_10_ecc_info_query_ras_error_address, + .set_eeprom_table_version = umc_v8_10_set_eeprom_table_version, }; diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_10.h b/drivers/gpu/drm/amd/amdgpu/umc_v8_10.h index c6dfd433fec7..dc12e0af5451 100644 --- a/drivers/gpu/drm/amd/amdgpu/umc_v8_10.h +++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_10.h @@ -33,7 +33,8 @@ /* Total channel instances for all available umc nodes */ #define UMC_V8_10_TOTAL_CHANNEL_NUM(adev) \ - (UMC_V8_10_CHANNEL_INSTANCE_NUM * UMC_V8_10_UMC_INSTANCE_NUM * (adev)->gmc.num_umc) + (UMC_V8_10_CHANNEL_INSTANCE_NUM * UMC_V8_10_UMC_INSTANCE_NUM * \ + (adev)->gmc.num_umc - hweight32((adev)->gmc.m_half_use) * 2) /* UMC regiser per channel offset */ #define UMC_V8_10_PER_CHANNEL_OFFSET 0x400 diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c index e32b656b3dab..abaa4463e906 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c @@ -444,7 +444,7 @@ static int uvd_v7_0_sw_init(void *handle) continue; if (!amdgpu_sriov_vf(adev)) { ring = &adev->uvd.inst[j].ring; - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "uvd_%d", ring->me); r = amdgpu_ring_init(adev, ring, 512, &adev->uvd.inst[j].irq, 0, @@ -455,7 +455,7 @@ static int uvd_v7_0_sw_init(void *handle) for (i = 0; i < adev->uvd.num_enc_rings; ++i) { ring = &adev->uvd.inst[j].ring_enc[i]; - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "uvd_enc_%d.%d", ring->me, i); if (amdgpu_sriov_vf(adev)) { ring->use_doorbell = true; diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c index 57b85bb6a1e4..e0b70cd3b697 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c @@ -466,7 +466,7 @@ static int vce_v4_0_sw_init(void *handle) enum amdgpu_ring_priority_level hw_prio = amdgpu_vce_get_ring_prio(i); ring = &adev->vce.ring[i]; - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "vce%d", i); if (amdgpu_sriov_vf(adev)) { /* DOORBELL only works under SRIOV */ diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c index 761c28fa6ec1..16feb491adf5 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c @@ -120,7 +120,7 @@ static int vcn_v1_0_sw_init(void *handle) return r; ring = &adev->vcn.inst->ring_dec; - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "vcn_dec"); r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst->irq, 0, AMDGPU_RING_PRIO_DEFAULT, NULL); @@ -142,7 +142,7 @@ static int vcn_v1_0_sw_init(void *handle) enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(i); ring = &adev->vcn.inst->ring_enc[i]; - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "vcn_enc%d", i); r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst->irq, 0, hw_prio, NULL); @@ -211,7 +211,7 @@ static int vcn_v1_0_hw_init(void *handle) goto done; } - ring = &adev->jpeg.inst->ring_dec; + ring = adev->jpeg.inst->ring_dec; r = amdgpu_ring_test_helper(ring); if (r) goto done; @@ -1304,7 +1304,7 @@ static int vcn_v1_0_pause_dpg_mode(struct amdgpu_device *adev, UVD_DPG_PAUSE__JPEG_PAUSE_DPG_ACK_MASK); /* Restore */ - ring = &adev->jpeg.inst->ring_dec; + ring = adev->jpeg.inst->ring_dec; WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_VMID, 0); WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK | @@ -1802,7 +1802,7 @@ static void vcn_v1_0_idle_work_handler(struct work_struct *work) else new_state.fw_based = VCN_DPG_STATE__UNPAUSE; - if (amdgpu_fence_count_emitted(&adev->jpeg.inst->ring_dec)) + if (amdgpu_fence_count_emitted(adev->jpeg.inst->ring_dec)) new_state.jpeg = VCN_DPG_STATE__PAUSE; else new_state.jpeg = VCN_DPG_STATE__UNPAUSE; @@ -1810,7 +1810,7 @@ static void vcn_v1_0_idle_work_handler(struct work_struct *work) adev->vcn.pause_dpg_mode(adev, 0, &new_state); } - fences += amdgpu_fence_count_emitted(&adev->jpeg.inst->ring_dec); + fences += amdgpu_fence_count_emitted(adev->jpeg.inst->ring_dec); fences += amdgpu_fence_count_emitted(&adev->vcn.inst->ring_dec); if (fences == 0) { @@ -1832,7 +1832,7 @@ static void vcn_v1_0_ring_begin_use(struct amdgpu_ring *ring) mutex_lock(&adev->vcn.vcn1_jpeg1_workaround); - if (amdgpu_fence_wait_empty(&ring->adev->jpeg.inst->ring_dec)) + if (amdgpu_fence_wait_empty(ring->adev->jpeg.inst->ring_dec)) DRM_ERROR("VCN dec: jpeg dec ring may not be empty\n"); vcn_v1_0_set_pg_for_begin_use(ring, set_clocks); @@ -1864,7 +1864,7 @@ void vcn_v1_0_set_pg_for_begin_use(struct amdgpu_ring *ring, bool set_clocks) else new_state.fw_based = VCN_DPG_STATE__UNPAUSE; - if (amdgpu_fence_count_emitted(&adev->jpeg.inst->ring_dec)) + if (amdgpu_fence_count_emitted(adev->jpeg.inst->ring_dec)) new_state.jpeg = VCN_DPG_STATE__PAUSE; else new_state.jpeg = VCN_DPG_STATE__UNPAUSE; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c index 7c2b3aa48083..c975aed2f6c7 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c @@ -129,7 +129,7 @@ static int vcn_v2_0_sw_init(void *handle) ring->use_doorbell = true; ring->doorbell_index = adev->doorbell_index.vcn.vcn_ring0_1 << 1; - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "vcn_dec"); r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst->irq, 0, @@ -160,7 +160,7 @@ static int vcn_v2_0_sw_init(void *handle) ring = &adev->vcn.inst->ring_enc[i]; ring->use_doorbell = true; - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); if (!amdgpu_sriov_vf(adev)) ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + i; else diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c index ab0b45d0ead1..bb1875f926f1 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c @@ -143,7 +143,7 @@ static int vcn_v2_5_sw_init(void *handle) /* VCN POISON TRAP */ r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[j], - VCN_2_6__SRCID_UVD_POISON, &adev->vcn.inst[j].irq); + VCN_2_6__SRCID_UVD_POISON, &adev->vcn.inst[j].ras_poison_irq); if (r) return r; } @@ -188,9 +188,9 @@ static int vcn_v2_5_sw_init(void *handle) (amdgpu_sriov_vf(adev) ? 2*j : 8*j); if (adev->ip_versions[UVD_HWIP][0] == IP_VERSION(2, 5, 0)) - ring->vm_hub = AMDGPU_MMHUB_1; + ring->vm_hub = AMDGPU_MMHUB1(0); else - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "vcn_dec_%d", j); r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[j].irq, @@ -208,9 +208,9 @@ static int vcn_v2_5_sw_init(void *handle) (amdgpu_sriov_vf(adev) ? (1 + i + 2*j) : (2 + i + 8*j)); if (adev->ip_versions[UVD_HWIP][0] == IP_VERSION(2, 5, 0)) - ring->vm_hub = AMDGPU_MMHUB_1; + ring->vm_hub = AMDGPU_MMHUB1(0); else - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "vcn_enc_%d.%d", j, i); r = amdgpu_ring_init(adev, ring, 512, @@ -354,6 +354,9 @@ static int vcn_v2_5_hw_fini(void *handle) (adev->vcn.cur_state != AMD_PG_STATE_GATE && RREG32_SOC15(VCN, i, mmUVD_STATUS))) vcn_v2_5_set_powergating_state(adev, AMD_PG_STATE_GATE); + + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) + amdgpu_irq_put(adev, &adev->vcn.inst[i].ras_poison_irq, 0); } return 0; @@ -1807,6 +1810,14 @@ static int vcn_v2_5_set_interrupt_state(struct amdgpu_device *adev, return 0; } +static int vcn_v2_6_set_ras_interrupt_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned int type, + enum amdgpu_interrupt_state state) +{ + return 0; +} + static int vcn_v2_5_process_interrupt(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) @@ -1837,9 +1848,6 @@ static int vcn_v2_5_process_interrupt(struct amdgpu_device *adev, case VCN_2_0__SRCID__UVD_ENC_LOW_LATENCY: amdgpu_fence_process(&adev->vcn.inst[ip_instance].ring_enc[1]); break; - case VCN_2_6__SRCID_UVD_POISON: - amdgpu_vcn_process_poison_irq(adev, source, entry); - break; default: DRM_ERROR("Unhandled interrupt: %d %d\n", entry->src_id, entry->src_data[0]); @@ -1854,6 +1862,11 @@ static const struct amdgpu_irq_src_funcs vcn_v2_5_irq_funcs = { .process = vcn_v2_5_process_interrupt, }; +static const struct amdgpu_irq_src_funcs vcn_v2_6_ras_irq_funcs = { + .set = vcn_v2_6_set_ras_interrupt_state, + .process = amdgpu_vcn_process_poison_irq, +}; + static void vcn_v2_5_set_irq_funcs(struct amdgpu_device *adev) { int i; @@ -1863,6 +1876,9 @@ static void vcn_v2_5_set_irq_funcs(struct amdgpu_device *adev) continue; adev->vcn.inst[i].irq.num_types = adev->vcn.num_enc_rings + 1; adev->vcn.inst[i].irq.funcs = &vcn_v2_5_irq_funcs; + + adev->vcn.inst[i].ras_poison_irq.num_types = adev->vcn.num_enc_rings + 1; + adev->vcn.inst[i].ras_poison_irq.funcs = &vcn_v2_6_ras_irq_funcs; } } @@ -1965,6 +1981,7 @@ const struct amdgpu_ras_block_hw_ops vcn_v2_6_ras_hw_ops = { static struct amdgpu_vcn_ras vcn_v2_6_ras = { .ras_block = { .hw_ops = &vcn_v2_6_ras_hw_ops, + .ras_late_init = amdgpu_vcn_ras_late_init, }, }; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c index 3eab186261aa..c8f63b3c6f69 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c @@ -189,7 +189,7 @@ static int vcn_v3_0_sw_init(void *handle) } else { ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i; } - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "vcn_dec_%d", i); r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0, AMDGPU_RING_PRIO_DEFAULT, @@ -213,7 +213,7 @@ static int vcn_v3_0_sw_init(void *handle) } else { ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + j + 8 * i; } - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "vcn_enc_%d.%d", i, j); r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0, hw_prio, &adev->vcn.inst[i].sched_score); @@ -1313,7 +1313,7 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev) header.version = MMSCH_VERSION; header.total_size = sizeof(struct mmsch_v3_0_init_header) >> 2; - for (i = 0; i < AMDGPU_MAX_VCN_INSTANCES; i++) { + for (i = 0; i < MMSCH_V3_0_VCN_INSTANCES; i++) { header.inst[i].init_status = 0; header.inst[i].table_offset = 0; header.inst[i].table_size = 0; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c index bf0674039598..b48bb5212488 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c @@ -129,7 +129,11 @@ static int vcn_v4_0_sw_init(void *handle) if (adev->vcn.harvest_config & (1 << i)) continue; - atomic_set(&adev->vcn.inst[i].sched_score, 0); + /* Init instance 0 sched_score to 1, so it's scheduled after other instances */ + if (i == 0) + atomic_set(&adev->vcn.inst[i].sched_score, 1); + else + atomic_set(&adev->vcn.inst[i].sched_score, 0); /* VCN UNIFIED TRAP */ r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i], @@ -139,7 +143,7 @@ static int vcn_v4_0_sw_init(void *handle) /* VCN POISON TRAP */ r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i], - VCN_4_0__SRCID_UVD_POISON, &adev->vcn.inst[i].irq); + VCN_4_0__SRCID_UVD_POISON, &adev->vcn.inst[i].ras_poison_irq); if (r) return r; @@ -149,7 +153,7 @@ static int vcn_v4_0_sw_init(void *handle) ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + i * (adev->vcn.num_enc_rings + 1) + 1; else ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + 8 * i; - ring->vm_hub = AMDGPU_MMHUB_0; + ring->vm_hub = AMDGPU_MMHUB0(0); sprintf(ring->name, "vcn_unified_%d", i); r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0, @@ -305,8 +309,8 @@ static int vcn_v4_0_hw_fini(void *handle) vcn_v4_0_set_powergating_state(adev, AMD_PG_STATE_GATE); } } - - amdgpu_irq_put(adev, &adev->vcn.inst[i].irq, 0); + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) + amdgpu_irq_put(adev, &adev->vcn.inst[i].ras_poison_irq, 0); } return 0; @@ -1239,7 +1243,7 @@ static int vcn_v4_0_start_sriov(struct amdgpu_device *adev) header.version = MMSCH_VERSION; header.total_size = sizeof(struct mmsch_v4_0_init_header) >> 2; - for (i = 0; i < AMDGPU_MAX_VCN_INSTANCES; i++) { + for (i = 0; i < MMSCH_V4_0_VCN_INSTANCES; i++) { header.inst[i].init_status = 0; header.inst[i].table_offset = 0; header.inst[i].table_size = 0; @@ -1976,6 +1980,24 @@ static int vcn_v4_0_set_interrupt_state(struct amdgpu_device *adev, struct amdgp } /** + * vcn_v4_0_set_ras_interrupt_state - set VCN block RAS interrupt state + * + * @adev: amdgpu_device pointer + * @source: interrupt sources + * @type: interrupt types + * @state: interrupt states + * + * Set VCN block RAS interrupt state + */ +static int vcn_v4_0_set_ras_interrupt_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned int type, + enum amdgpu_interrupt_state state) +{ + return 0; +} + +/** * vcn_v4_0_process_interrupt - process VCN block interrupt * * @adev: amdgpu_device pointer @@ -2007,9 +2029,6 @@ static int vcn_v4_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_ case VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE: amdgpu_fence_process(&adev->vcn.inst[ip_instance].ring_enc[0]); break; - case VCN_4_0__SRCID_UVD_POISON: - amdgpu_vcn_process_poison_irq(adev, source, entry); - break; default: DRM_ERROR("Unhandled interrupt: %d %d\n", entry->src_id, entry->src_data[0]); @@ -2024,6 +2043,11 @@ static const struct amdgpu_irq_src_funcs vcn_v4_0_irq_funcs = { .process = vcn_v4_0_process_interrupt, }; +static const struct amdgpu_irq_src_funcs vcn_v4_0_ras_irq_funcs = { + .set = vcn_v4_0_set_ras_interrupt_state, + .process = amdgpu_vcn_process_poison_irq, +}; + /** * vcn_v4_0_set_irq_funcs - set VCN block interrupt irq functions * @@ -2041,6 +2065,9 @@ static void vcn_v4_0_set_irq_funcs(struct amdgpu_device *adev) adev->vcn.inst[i].irq.num_types = adev->vcn.num_enc_rings + 1; adev->vcn.inst[i].irq.funcs = &vcn_v4_0_irq_funcs; + + adev->vcn.inst[i].ras_poison_irq.num_types = adev->vcn.num_enc_rings + 1; + adev->vcn.inst[i].ras_poison_irq.funcs = &vcn_v4_0_ras_irq_funcs; } } @@ -2114,6 +2141,7 @@ const struct amdgpu_ras_block_hw_ops vcn_v4_0_ras_hw_ops = { static struct amdgpu_vcn_ras vcn_v4_0_ras = { .ras_block = { .hw_ops = &vcn_v4_0_ras_hw_ops, + .ras_late_init = amdgpu_vcn_ras_late_init, }, }; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c new file mode 100644 index 000000000000..5d67b8b8a3d6 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c @@ -0,0 +1,1541 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#include <linux/firmware.h> +#include <drm/drm_drv.h> + +#include "amdgpu.h" +#include "amdgpu_vcn.h" +#include "amdgpu_pm.h" +#include "soc15.h" +#include "soc15d.h" +#include "soc15_hw_ip.h" +#include "vcn_v2_0.h" + +#include "vcn/vcn_4_0_3_offset.h" +#include "vcn/vcn_4_0_3_sh_mask.h" +#include "ivsrcid/vcn/irqsrcs_vcn_4_0.h" + +#define mmUVD_DPG_LMA_CTL regUVD_DPG_LMA_CTL +#define mmUVD_DPG_LMA_CTL_BASE_IDX regUVD_DPG_LMA_CTL_BASE_IDX +#define mmUVD_DPG_LMA_DATA regUVD_DPG_LMA_DATA +#define mmUVD_DPG_LMA_DATA_BASE_IDX regUVD_DPG_LMA_DATA_BASE_IDX + +#define VCN_VID_SOC_ADDRESS_2_0 0x1fb00 +#define VCN1_VID_SOC_ADDRESS_3_0 0x48300 + +static void vcn_v4_0_3_set_unified_ring_funcs(struct amdgpu_device *adev); +static void vcn_v4_0_3_set_irq_funcs(struct amdgpu_device *adev); +static int vcn_v4_0_3_set_powergating_state(void *handle, + enum amd_powergating_state state); +static int vcn_v4_0_3_pause_dpg_mode(struct amdgpu_device *adev, + int inst_idx, struct dpg_pause_state *new_state); +static void vcn_v4_0_3_unified_ring_set_wptr(struct amdgpu_ring *ring); +static void vcn_v4_0_3_set_ras_funcs(struct amdgpu_device *adev); +static void vcn_v4_0_3_enable_ras(struct amdgpu_device *adev, + int inst_idx, bool indirect); +/** + * vcn_v4_0_3_early_init - set function pointers + * + * @handle: amdgpu_device pointer + * + * Set ring and irq function pointers + */ +static int vcn_v4_0_3_early_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + /* re-use enc ring as unified ring */ + adev->vcn.num_enc_rings = 1; + + vcn_v4_0_3_set_unified_ring_funcs(adev); + vcn_v4_0_3_set_irq_funcs(adev); + vcn_v4_0_3_set_ras_funcs(adev); + + return amdgpu_vcn_early_init(adev); +} + +/** + * vcn_v4_0_3_sw_init - sw init for VCN block + * + * @handle: amdgpu_device pointer + * + * Load firmware and sw initialization + */ +static int vcn_v4_0_3_sw_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_ring *ring; + int i, r, vcn_inst; + + r = amdgpu_vcn_sw_init(adev); + if (r) + return r; + + amdgpu_vcn_setup_ucode(adev); + + r = amdgpu_vcn_resume(adev); + if (r) + return r; + + /* VCN DEC TRAP */ + r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN, + VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst->irq); + if (r) + return r; + + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + volatile struct amdgpu_vcn4_fw_shared *fw_shared; + + vcn_inst = GET_INST(VCN, i); + + ring = &adev->vcn.inst[i].ring_enc[0]; + ring->use_doorbell = true; + ring->doorbell_index = + (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + + 9 * vcn_inst; + ring->vm_hub = AMDGPU_MMHUB0(adev->vcn.inst[i].aid_id); + sprintf(ring->name, "vcn_unified_%d", adev->vcn.inst[i].aid_id); + r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst->irq, 0, + AMDGPU_RING_PRIO_DEFAULT, + &adev->vcn.inst[i].sched_score); + if (r) + return r; + + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared->present_flag_0 = cpu_to_le32(AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE); + fw_shared->sq.is_enabled = true; + + if (amdgpu_vcnfw_log) + amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]); + } + + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + adev->vcn.pause_dpg_mode = vcn_v4_0_3_pause_dpg_mode; + + if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) { + r = amdgpu_vcn_ras_sw_init(adev); + if (r) { + dev_err(adev->dev, "Failed to initialize vcn ras block!\n"); + return r; + } + } + + return 0; +} + +/** + * vcn_v4_0_3_sw_fini - sw fini for VCN block + * + * @handle: amdgpu_device pointer + * + * VCN suspend and free up sw allocation + */ +static int vcn_v4_0_3_sw_fini(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int i, r, idx; + + if (drm_dev_enter(&adev->ddev, &idx)) { + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + volatile struct amdgpu_vcn4_fw_shared *fw_shared; + + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared->present_flag_0 = 0; + fw_shared->sq.is_enabled = cpu_to_le32(false); + } + drm_dev_exit(idx); + } + + r = amdgpu_vcn_suspend(adev); + if (r) + return r; + + r = amdgpu_vcn_sw_fini(adev); + + return r; +} + +/** + * vcn_v4_0_3_hw_init - start and test VCN block + * + * @handle: amdgpu_device pointer + * + * Initialize the hardware, boot up the VCPU and do some testing + */ +static int vcn_v4_0_3_hw_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_ring *ring; + int i, r, vcn_inst; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + vcn_inst = GET_INST(VCN, i); + ring = &adev->vcn.inst[i].ring_enc[0]; + + if (ring->use_doorbell) { + adev->nbio.funcs->vcn_doorbell_range( + adev, ring->use_doorbell, + (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + + 9 * vcn_inst, + adev->vcn.inst[i].aid_id); + + WREG32_SOC15( + VCN, GET_INST(VCN, ring->me), + regVCN_RB1_DB_CTRL, + ring->doorbell_index + << VCN_RB1_DB_CTRL__OFFSET__SHIFT | + VCN_RB1_DB_CTRL__EN_MASK); + + /* Read DB_CTRL to flush the write DB_CTRL command. */ + RREG32_SOC15( + VCN, GET_INST(VCN, ring->me), + regVCN_RB1_DB_CTRL); + } + + r = amdgpu_ring_test_helper(ring); + if (r) + goto done; + } + +done: + if (!r) + DRM_DEV_INFO(adev->dev, "VCN decode initialized successfully(under %s).\n", + (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)?"DPG Mode":"SPG Mode"); + + return r; +} + +/** + * vcn_v4_0_3_hw_fini - stop the hardware block + * + * @handle: amdgpu_device pointer + * + * Stop the VCN block, mark ring as not ready any more + */ +static int vcn_v4_0_3_hw_fini(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + cancel_delayed_work_sync(&adev->vcn.idle_work); + + if (adev->vcn.cur_state != AMD_PG_STATE_GATE) + vcn_v4_0_3_set_powergating_state(adev, AMD_PG_STATE_GATE); + + return 0; +} + +/** + * vcn_v4_0_3_suspend - suspend VCN block + * + * @handle: amdgpu_device pointer + * + * HW fini and suspend VCN block + */ +static int vcn_v4_0_3_suspend(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int r; + + r = vcn_v4_0_3_hw_fini(adev); + if (r) + return r; + + r = amdgpu_vcn_suspend(adev); + + return r; +} + +/** + * vcn_v4_0_3_resume - resume VCN block + * + * @handle: amdgpu_device pointer + * + * Resume firmware and hw init VCN block + */ +static int vcn_v4_0_3_resume(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int r; + + r = amdgpu_vcn_resume(adev); + if (r) + return r; + + r = vcn_v4_0_3_hw_init(adev); + + return r; +} + +/** + * vcn_v4_0_3_mc_resume - memory controller programming + * + * @adev: amdgpu_device pointer + * @inst_idx: instance number + * + * Let the VCN memory controller know it's offsets + */ +static void vcn_v4_0_3_mc_resume(struct amdgpu_device *adev, int inst_idx) +{ + uint32_t offset, size, vcn_inst; + const struct common_firmware_header *hdr; + + hdr = (const struct common_firmware_header *)adev->vcn.fw->data; + size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8); + + vcn_inst = GET_INST(VCN, inst_idx); + /* cache window 0: fw */ + if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { + WREG32_SOC15( + VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW, + (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx] + .tmr_mc_addr_lo)); + WREG32_SOC15( + VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH, + (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx] + .tmr_mc_addr_hi)); + WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET0, 0); + offset = 0; + } else { + WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW, + lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr)); + WREG32_SOC15(VCN, vcn_inst, + regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH, + upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr)); + offset = size; + WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET0, + AMDGPU_UVD_FIRMWARE_OFFSET >> 3); + } + WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_SIZE0, size); + + /* cache window 1: stack */ + WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW, + lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset)); + WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH, + upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset)); + WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET1, 0); + WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_SIZE1, + AMDGPU_VCN_STACK_SIZE); + + /* cache window 2: context */ + WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW, + lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + + AMDGPU_VCN_STACK_SIZE)); + WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH, + upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + + AMDGPU_VCN_STACK_SIZE)); + WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET2, 0); + WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_SIZE2, + AMDGPU_VCN_CONTEXT_SIZE); + + /* non-cache window */ + WREG32_SOC15( + VCN, vcn_inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW, + lower_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr)); + WREG32_SOC15( + VCN, vcn_inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH, + upper_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr)); + WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_NONCACHE_OFFSET0, 0); + WREG32_SOC15( + VCN, vcn_inst, regUVD_VCPU_NONCACHE_SIZE0, + AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared))); +} + +/** + * vcn_v4_0_3_mc_resume_dpg_mode - memory controller programming for dpg mode + * + * @adev: amdgpu_device pointer + * @inst_idx: instance number index + * @indirect: indirectly write sram + * + * Let the VCN memory controller know it's offsets with dpg mode + */ +static void vcn_v4_0_3_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect) +{ + uint32_t offset, size; + const struct common_firmware_header *hdr; + + hdr = (const struct common_firmware_header *)adev->vcn.fw->data; + size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8); + + /* cache window 0: fw */ + if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { + if (!indirect) { + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), + (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + + inst_idx].tmr_mc_addr_lo), 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), + (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + + inst_idx].tmr_mc_addr_hi), 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect); + } else { + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), 0, 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), 0, 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect); + } + offset = 0; + } else { + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), + lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), + upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect); + offset = size; + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_CACHE_OFFSET0), + AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0, indirect); + } + + if (!indirect) + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_CACHE_SIZE0), size, 0, indirect); + else + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_CACHE_SIZE0), 0, 0, indirect); + + /* cache window 1: stack */ + if (!indirect) { + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), + lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), + upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect); + } else { + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), 0, 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), 0, 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect); + } + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE, 0, indirect); + + /* cache window 2: context */ + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW), + lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + + AMDGPU_VCN_STACK_SIZE), 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH), + upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + + AMDGPU_VCN_STACK_SIZE), 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_CACHE_OFFSET2), 0, 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE, 0, indirect); + + /* non-cache window */ + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW), + lower_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH), + upper_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_NONCACHE_SIZE0), + AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)), 0, indirect); + + /* VCN global tiling registers */ + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_GFX8_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_GFX10_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect); +} + +/** + * vcn_v4_0_3_disable_clock_gating - disable VCN clock gating + * + * @adev: amdgpu_device pointer + * @inst_idx: instance number + * + * Disable clock gating for VCN block + */ +static void vcn_v4_0_3_disable_clock_gating(struct amdgpu_device *adev, int inst_idx) +{ + uint32_t data; + int vcn_inst; + + if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG) + return; + + vcn_inst = GET_INST(VCN, inst_idx); + + /* VCN disable CGC */ + data = RREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL); + data &= ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK; + data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT; + data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT; + WREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL, data); + + data = RREG32_SOC15(VCN, vcn_inst, regUVD_CGC_GATE); + data &= ~(UVD_CGC_GATE__SYS_MASK + | UVD_CGC_GATE__MPEG2_MASK + | UVD_CGC_GATE__REGS_MASK + | UVD_CGC_GATE__RBC_MASK + | UVD_CGC_GATE__LMI_MC_MASK + | UVD_CGC_GATE__LMI_UMC_MASK + | UVD_CGC_GATE__MPC_MASK + | UVD_CGC_GATE__LBSI_MASK + | UVD_CGC_GATE__LRBBM_MASK + | UVD_CGC_GATE__WCB_MASK + | UVD_CGC_GATE__VCPU_MASK + | UVD_CGC_GATE__MMSCH_MASK); + + WREG32_SOC15(VCN, vcn_inst, regUVD_CGC_GATE, data); + SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_CGC_GATE, 0, 0xFFFFFFFF); + + data = RREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL); + data &= ~(UVD_CGC_CTRL__SYS_MODE_MASK + | UVD_CGC_CTRL__MPEG2_MODE_MASK + | UVD_CGC_CTRL__REGS_MODE_MASK + | UVD_CGC_CTRL__RBC_MODE_MASK + | UVD_CGC_CTRL__LMI_MC_MODE_MASK + | UVD_CGC_CTRL__LMI_UMC_MODE_MASK + | UVD_CGC_CTRL__MPC_MODE_MASK + | UVD_CGC_CTRL__LBSI_MODE_MASK + | UVD_CGC_CTRL__LRBBM_MODE_MASK + | UVD_CGC_CTRL__WCB_MODE_MASK + | UVD_CGC_CTRL__VCPU_MODE_MASK + | UVD_CGC_CTRL__MMSCH_MODE_MASK); + WREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL, data); + + data = RREG32_SOC15(VCN, vcn_inst, regUVD_SUVD_CGC_GATE); + data |= (UVD_SUVD_CGC_GATE__SRE_MASK + | UVD_SUVD_CGC_GATE__SIT_MASK + | UVD_SUVD_CGC_GATE__SMP_MASK + | UVD_SUVD_CGC_GATE__SCM_MASK + | UVD_SUVD_CGC_GATE__SDB_MASK + | UVD_SUVD_CGC_GATE__SRE_H264_MASK + | UVD_SUVD_CGC_GATE__SRE_HEVC_MASK + | UVD_SUVD_CGC_GATE__SIT_H264_MASK + | UVD_SUVD_CGC_GATE__SIT_HEVC_MASK + | UVD_SUVD_CGC_GATE__SCM_H264_MASK + | UVD_SUVD_CGC_GATE__SCM_HEVC_MASK + | UVD_SUVD_CGC_GATE__SDB_H264_MASK + | UVD_SUVD_CGC_GATE__SDB_HEVC_MASK + | UVD_SUVD_CGC_GATE__ENT_MASK + | UVD_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK + | UVD_SUVD_CGC_GATE__SITE_MASK + | UVD_SUVD_CGC_GATE__SRE_VP9_MASK + | UVD_SUVD_CGC_GATE__SCM_VP9_MASK + | UVD_SUVD_CGC_GATE__SIT_VP9_DEC_MASK + | UVD_SUVD_CGC_GATE__SDB_VP9_MASK + | UVD_SUVD_CGC_GATE__IME_HEVC_MASK); + WREG32_SOC15(VCN, vcn_inst, regUVD_SUVD_CGC_GATE, data); + + data = RREG32_SOC15(VCN, vcn_inst, regUVD_SUVD_CGC_CTRL); + data &= ~(UVD_SUVD_CGC_CTRL__SRE_MODE_MASK + | UVD_SUVD_CGC_CTRL__SIT_MODE_MASK + | UVD_SUVD_CGC_CTRL__SMP_MODE_MASK + | UVD_SUVD_CGC_CTRL__SCM_MODE_MASK + | UVD_SUVD_CGC_CTRL__SDB_MODE_MASK + | UVD_SUVD_CGC_CTRL__ENT_MODE_MASK + | UVD_SUVD_CGC_CTRL__IME_MODE_MASK + | UVD_SUVD_CGC_CTRL__SITE_MODE_MASK); + WREG32_SOC15(VCN, vcn_inst, regUVD_SUVD_CGC_CTRL, data); +} + +/** + * vcn_v4_0_3_disable_clock_gating_dpg_mode - disable VCN clock gating dpg mode + * + * @adev: amdgpu_device pointer + * @sram_sel: sram select + * @inst_idx: instance number index + * @indirect: indirectly write sram + * + * Disable clock gating for VCN block with dpg mode + */ +static void vcn_v4_0_3_disable_clock_gating_dpg_mode(struct amdgpu_device *adev, uint8_t sram_sel, + int inst_idx, uint8_t indirect) +{ + uint32_t reg_data = 0; + + if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG) + return; + + /* enable sw clock gating control */ + reg_data = 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + reg_data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT; + reg_data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT; + reg_data &= ~(UVD_CGC_CTRL__SYS_MODE_MASK | + UVD_CGC_CTRL__MPEG2_MODE_MASK | + UVD_CGC_CTRL__REGS_MODE_MASK | + UVD_CGC_CTRL__RBC_MODE_MASK | + UVD_CGC_CTRL__LMI_MC_MODE_MASK | + UVD_CGC_CTRL__LMI_UMC_MODE_MASK | + UVD_CGC_CTRL__IDCT_MODE_MASK | + UVD_CGC_CTRL__MPRD_MODE_MASK | + UVD_CGC_CTRL__MPC_MODE_MASK | + UVD_CGC_CTRL__LBSI_MODE_MASK | + UVD_CGC_CTRL__LRBBM_MODE_MASK | + UVD_CGC_CTRL__WCB_MODE_MASK | + UVD_CGC_CTRL__VCPU_MODE_MASK); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_CGC_CTRL), reg_data, sram_sel, indirect); + + /* turn off clock gating */ + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_CGC_GATE), 0, sram_sel, indirect); + + /* turn on SUVD clock gating */ + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_SUVD_CGC_GATE), 1, sram_sel, indirect); + + /* turn on sw mode in UVD_SUVD_CGC_CTRL */ + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_SUVD_CGC_CTRL), 0, sram_sel, indirect); +} + +/** + * vcn_v4_0_3_enable_clock_gating - enable VCN clock gating + * + * @adev: amdgpu_device pointer + * @inst_idx: instance number + * + * Enable clock gating for VCN block + */ +static void vcn_v4_0_3_enable_clock_gating(struct amdgpu_device *adev, int inst_idx) +{ + uint32_t data; + int vcn_inst; + + if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG) + return; + + vcn_inst = GET_INST(VCN, inst_idx); + + /* enable VCN CGC */ + data = RREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL); + data |= 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT; + data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT; + WREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL, data); + + data = RREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL); + data |= (UVD_CGC_CTRL__SYS_MODE_MASK + | UVD_CGC_CTRL__MPEG2_MODE_MASK + | UVD_CGC_CTRL__REGS_MODE_MASK + | UVD_CGC_CTRL__RBC_MODE_MASK + | UVD_CGC_CTRL__LMI_MC_MODE_MASK + | UVD_CGC_CTRL__LMI_UMC_MODE_MASK + | UVD_CGC_CTRL__MPC_MODE_MASK + | UVD_CGC_CTRL__LBSI_MODE_MASK + | UVD_CGC_CTRL__LRBBM_MODE_MASK + | UVD_CGC_CTRL__WCB_MODE_MASK + | UVD_CGC_CTRL__VCPU_MODE_MASK); + WREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL, data); + + data = RREG32_SOC15(VCN, vcn_inst, regUVD_SUVD_CGC_CTRL); + data |= (UVD_SUVD_CGC_CTRL__SRE_MODE_MASK + | UVD_SUVD_CGC_CTRL__SIT_MODE_MASK + | UVD_SUVD_CGC_CTRL__SMP_MODE_MASK + | UVD_SUVD_CGC_CTRL__SCM_MODE_MASK + | UVD_SUVD_CGC_CTRL__SDB_MODE_MASK + | UVD_SUVD_CGC_CTRL__ENT_MODE_MASK + | UVD_SUVD_CGC_CTRL__IME_MODE_MASK + | UVD_SUVD_CGC_CTRL__SITE_MODE_MASK); + WREG32_SOC15(VCN, vcn_inst, regUVD_SUVD_CGC_CTRL, data); +} + +/** + * vcn_v4_0_3_start_dpg_mode - VCN start with dpg mode + * + * @adev: amdgpu_device pointer + * @inst_idx: instance number index + * @indirect: indirectly write sram + * + * Start VCN block with dpg mode + */ +static int vcn_v4_0_3_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect) +{ + volatile struct amdgpu_vcn4_fw_shared *fw_shared = + adev->vcn.inst[inst_idx].fw_shared.cpu_addr; + struct amdgpu_ring *ring; + int vcn_inst; + uint32_t tmp; + + vcn_inst = GET_INST(VCN, inst_idx); + /* disable register anti-hang mechanism */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_POWER_STATUS), 1, + ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK); + /* enable dynamic power gating mode */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_POWER_STATUS); + tmp |= UVD_POWER_STATUS__UVD_PG_MODE_MASK; + tmp |= UVD_POWER_STATUS__UVD_PG_EN_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_POWER_STATUS, tmp); + + if (indirect) { + DRM_DEV_DEBUG(adev->dev, "VCN %d start: on AID %d", + inst_idx, adev->vcn.inst[inst_idx].aid_id); + adev->vcn.inst[inst_idx].dpg_sram_curr_addr = + (uint32_t *)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr; + /* Use dummy register 0xDEADBEEF passing AID selection to PSP FW */ + WREG32_SOC15_DPG_MODE(inst_idx, 0xDEADBEEF, + adev->vcn.inst[inst_idx].aid_id, 0, true); + } + + /* enable clock gating */ + vcn_v4_0_3_disable_clock_gating_dpg_mode(adev, 0, inst_idx, indirect); + + /* enable VCPU clock */ + tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT); + tmp |= UVD_VCPU_CNTL__CLK_EN_MASK; + tmp |= UVD_VCPU_CNTL__BLK_RST_MASK; + + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_CNTL), tmp, 0, indirect); + + /* disable master interrupt */ + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_MASTINT_EN), 0, 0, indirect); + + /* setup regUVD_LMI_CTRL */ + tmp = (UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | + UVD_LMI_CTRL__REQ_MODE_MASK | + UVD_LMI_CTRL__CRC_RESET_MASK | + UVD_LMI_CTRL__MASK_MC_URGENT_MASK | + UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK | + (8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) | + 0x00100000L); + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_CTRL), tmp, 0, indirect); + + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_MPC_CNTL), + 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT, 0, indirect); + + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_MPC_SET_MUXA0), + ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)), 0, indirect); + + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_MPC_SET_MUXB0), + ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)), 0, indirect); + + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_MPC_SET_MUX), + ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) | + (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)), 0, indirect); + + vcn_v4_0_3_mc_resume_dpg_mode(adev, inst_idx, indirect); + + tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT); + tmp |= UVD_VCPU_CNTL__CLK_EN_MASK; + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_VCPU_CNTL), tmp, 0, indirect); + + /* enable LMI MC and UMC channels */ + tmp = 0x1f << UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM__SHIFT; + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_LMI_CTRL2), tmp, 0, indirect); + + vcn_v4_0_3_enable_ras(adev, inst_idx, indirect); + + /* enable master interrupt */ + WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( + VCN, 0, regUVD_MASTINT_EN), + UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect); + + if (indirect) + psp_update_vcn_sram(adev, 0, adev->vcn.inst[inst_idx].dpg_sram_gpu_addr, + (uint32_t)((uintptr_t)adev->vcn.inst[inst_idx].dpg_sram_curr_addr - + (uintptr_t)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr)); + + ring = &adev->vcn.inst[inst_idx].ring_enc[0]; + + /* program the RB_BASE for ring buffer */ + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO, + lower_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_HI, + upper_32_bits(ring->gpu_addr)); + + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_SIZE, + ring->ring_size / sizeof(uint32_t)); + + /* resetting ring, fw should not check RB ring */ + tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE); + tmp &= ~(VCN_RB_ENABLE__RB_EN_MASK); + WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp); + fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET; + + /* Initialize the ring buffer's read and write pointers */ + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR, 0); + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, 0); + ring->wptr = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR); + + tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE); + tmp |= VCN_RB_ENABLE__RB_EN_MASK; + WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp); + fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF); + + /*resetting done, fw can check RB ring */ + fw_shared->sq.queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET); + + return 0; +} + +/** + * vcn_v4_0_3_start - VCN start + * + * @adev: amdgpu_device pointer + * + * Start VCN block + */ +static int vcn_v4_0_3_start(struct amdgpu_device *adev) +{ + volatile struct amdgpu_vcn4_fw_shared *fw_shared; + struct amdgpu_ring *ring; + int i, j, k, r, vcn_inst; + uint32_t tmp; + + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_uvd(adev, true); + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { + r = vcn_v4_0_3_start_dpg_mode(adev, i, adev->vcn.indirect_sram); + continue; + } + + vcn_inst = GET_INST(VCN, i); + /* set VCN status busy */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS) | + UVD_STATUS__UVD_BUSY; + WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, tmp); + + /*SW clock gating */ + vcn_v4_0_3_disable_clock_gating(adev, i); + + /* enable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__CLK_EN_MASK, + ~UVD_VCPU_CNTL__CLK_EN_MASK); + + /* disable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), 0, + ~UVD_MASTINT_EN__VCPU_EN_MASK); + + /* enable LMI MC and UMC channels */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_LMI_CTRL2), 0, + ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); + + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); + tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); + + /* setup regUVD_LMI_CTRL */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL); + WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL, + tmp | UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | + UVD_LMI_CTRL__MASK_MC_URGENT_MASK | + UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); + + /* setup regUVD_MPC_CNTL */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_MPC_CNTL); + tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK; + tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT; + WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_CNTL, tmp); + + /* setup UVD_MPC_SET_MUXA0 */ + WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_SET_MUXA0, + ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT))); + + /* setup UVD_MPC_SET_MUXB0 */ + WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_SET_MUXB0, + ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT))); + + /* setup UVD_MPC_SET_MUX */ + WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_SET_MUX, + ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) | + (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT))); + + vcn_v4_0_3_mc_resume(adev, i); + + /* VCN global tiling registers */ + WREG32_SOC15(VCN, vcn_inst, regUVD_GFX8_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(VCN, vcn_inst, regUVD_GFX10_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + + /* unblock VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), 0, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + + /* release VCPU reset to boot */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + for (j = 0; j < 10; ++j) { + uint32_t status; + + for (k = 0; k < 100; ++k) { + status = RREG32_SOC15(VCN, vcn_inst, + regUVD_STATUS); + if (status & 2) + break; + mdelay(10); + } + r = 0; + if (status & 2) + break; + + DRM_DEV_ERROR(adev->dev, + "VCN decode not responding, trying to reset the VCPU!!!\n"); + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, + regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + mdelay(10); + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, + regUVD_VCPU_CNTL), + 0, ~UVD_VCPU_CNTL__BLK_RST_MASK); + + mdelay(10); + r = -1; + } + + if (r) { + DRM_DEV_ERROR(adev->dev, "VCN decode not responding, giving up!!!\n"); + return r; + } + + /* enable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), + UVD_MASTINT_EN__VCPU_EN_MASK, + ~UVD_MASTINT_EN__VCPU_EN_MASK); + + /* clear the busy bit of VCN_STATUS */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_STATUS), 0, + ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); + + ring = &adev->vcn.inst[i].ring_enc[0]; + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + + /* program the RB_BASE for ring buffer */ + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO, + lower_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_HI, + upper_32_bits(ring->gpu_addr)); + + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_SIZE, + ring->ring_size / sizeof(uint32_t)); + + /* resetting ring, fw should not check RB ring */ + tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE); + tmp &= ~(VCN_RB_ENABLE__RB_EN_MASK); + WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp); + + /* Initialize the ring buffer's read and write pointers */ + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR, 0); + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, 0); + + tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE); + tmp |= VCN_RB_ENABLE__RB_EN_MASK; + WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp); + + ring->wptr = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR); + fw_shared->sq.queue_mode &= + cpu_to_le32(~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF)); + + } + return 0; +} + +/** + * vcn_v4_0_3_stop_dpg_mode - VCN stop with dpg mode + * + * @adev: amdgpu_device pointer + * @inst_idx: instance number index + * + * Stop VCN block with dpg mode + */ +static int vcn_v4_0_3_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx) +{ + uint32_t tmp; + int vcn_inst; + + vcn_inst = GET_INST(VCN, inst_idx); + + /* Wait for power status to be 1 */ + SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_POWER_STATUS, 1, + UVD_POWER_STATUS__UVD_POWER_STATUS_MASK); + + /* wait for read ptr to be equal to write ptr */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR); + SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_RB_RPTR, tmp, 0xFFFFFFFF); + + SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_POWER_STATUS, 1, + UVD_POWER_STATUS__UVD_POWER_STATUS_MASK); + + /* disable dynamic power gating mode */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_POWER_STATUS), 0, + ~UVD_POWER_STATUS__UVD_PG_MODE_MASK); + return 0; +} + +/** + * vcn_v4_0_3_stop - VCN stop + * + * @adev: amdgpu_device pointer + * + * Stop VCN block + */ +static int vcn_v4_0_3_stop(struct amdgpu_device *adev) +{ + volatile struct amdgpu_vcn4_fw_shared *fw_shared; + int i, r = 0, vcn_inst; + uint32_t tmp; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + vcn_inst = GET_INST(VCN, i); + + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF; + + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { + vcn_v4_0_3_stop_dpg_mode(adev, i); + continue; + } + + /* wait for vcn idle */ + r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_STATUS, + UVD_STATUS__IDLE, 0x7); + if (r) + goto Done; + + tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | + UVD_LMI_STATUS__READ_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, + tmp); + if (r) + goto Done; + + /* stall UMC channel */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2); + tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2, tmp); + tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK | + UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, + tmp); + if (r) + goto Done; + + /* Unblock VCPU Register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), + UVD_RB_ARB_CTRL__VCPU_DIS_MASK, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + + /* release VCPU reset to boot */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + /* disable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0, + ~(UVD_VCPU_CNTL__CLK_EN_MASK)); + + /* reset LMI UMC/LMI/VCPU */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); + + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); + + /* clear VCN status */ + WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, 0); + + /* apply HW clock gating */ + vcn_v4_0_3_enable_clock_gating(adev, i); + } +Done: + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_uvd(adev, false); + + return 0; +} + +/** + * vcn_v4_0_3_pause_dpg_mode - VCN pause with dpg mode + * + * @adev: amdgpu_device pointer + * @inst_idx: instance number index + * @new_state: pause state + * + * Pause dpg mode for VCN block + */ +static int vcn_v4_0_3_pause_dpg_mode(struct amdgpu_device *adev, int inst_idx, + struct dpg_pause_state *new_state) +{ + + return 0; +} + +/** + * vcn_v4_0_3_unified_ring_get_rptr - get unified read pointer + * + * @ring: amdgpu_ring pointer + * + * Returns the current hardware unified read pointer + */ +static uint64_t vcn_v4_0_3_unified_ring_get_rptr(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + + if (ring != &adev->vcn.inst[ring->me].ring_enc[0]) + DRM_ERROR("wrong ring id is identified in %s", __func__); + + return RREG32_SOC15(VCN, GET_INST(VCN, ring->me), regUVD_RB_RPTR); +} + +/** + * vcn_v4_0_3_unified_ring_get_wptr - get unified write pointer + * + * @ring: amdgpu_ring pointer + * + * Returns the current hardware unified write pointer + */ +static uint64_t vcn_v4_0_3_unified_ring_get_wptr(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + + if (ring != &adev->vcn.inst[ring->me].ring_enc[0]) + DRM_ERROR("wrong ring id is identified in %s", __func__); + + if (ring->use_doorbell) + return *ring->wptr_cpu_addr; + else + return RREG32_SOC15(VCN, GET_INST(VCN, ring->me), + regUVD_RB_WPTR); +} + +/** + * vcn_v4_0_3_unified_ring_set_wptr - set enc write pointer + * + * @ring: amdgpu_ring pointer + * + * Commits the enc write pointer to the hardware + */ +static void vcn_v4_0_3_unified_ring_set_wptr(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + + if (ring != &adev->vcn.inst[ring->me].ring_enc[0]) + DRM_ERROR("wrong ring id is identified in %s", __func__); + + if (ring->use_doorbell) { + *ring->wptr_cpu_addr = lower_32_bits(ring->wptr); + WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr)); + } else { + WREG32_SOC15(VCN, GET_INST(VCN, ring->me), regUVD_RB_WPTR, + lower_32_bits(ring->wptr)); + } +} + +static const struct amdgpu_ring_funcs vcn_v4_0_3_unified_ring_vm_funcs = { + .type = AMDGPU_RING_TYPE_VCN_ENC, + .align_mask = 0x3f, + .nop = VCN_ENC_CMD_NO_OP, + .get_rptr = vcn_v4_0_3_unified_ring_get_rptr, + .get_wptr = vcn_v4_0_3_unified_ring_get_wptr, + .set_wptr = vcn_v4_0_3_unified_ring_set_wptr, + .emit_frame_size = + SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + + SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 + + 4 + /* vcn_v2_0_enc_ring_emit_vm_flush */ + 5 + 5 + /* vcn_v2_0_enc_ring_emit_fence x2 vm fence */ + 1, /* vcn_v2_0_enc_ring_insert_end */ + .emit_ib_size = 5, /* vcn_v2_0_enc_ring_emit_ib */ + .emit_ib = vcn_v2_0_enc_ring_emit_ib, + .emit_fence = vcn_v2_0_enc_ring_emit_fence, + .emit_vm_flush = vcn_v2_0_enc_ring_emit_vm_flush, + .test_ring = amdgpu_vcn_enc_ring_test_ring, + .test_ib = amdgpu_vcn_unified_ring_test_ib, + .insert_nop = amdgpu_ring_insert_nop, + .insert_end = vcn_v2_0_enc_ring_insert_end, + .pad_ib = amdgpu_ring_generic_pad_ib, + .begin_use = amdgpu_vcn_ring_begin_use, + .end_use = amdgpu_vcn_ring_end_use, + .emit_wreg = vcn_v2_0_enc_ring_emit_wreg, + .emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait, + .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, +}; + +/** + * vcn_v4_0_3_set_unified_ring_funcs - set unified ring functions + * + * @adev: amdgpu_device pointer + * + * Set unified ring functions + */ +static void vcn_v4_0_3_set_unified_ring_funcs(struct amdgpu_device *adev) +{ + int i, vcn_inst; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + adev->vcn.inst[i].ring_enc[0].funcs = &vcn_v4_0_3_unified_ring_vm_funcs; + adev->vcn.inst[i].ring_enc[0].me = i; + vcn_inst = GET_INST(VCN, i); + adev->vcn.inst[i].aid_id = + vcn_inst / adev->vcn.num_inst_per_aid; + } + DRM_DEV_INFO(adev->dev, "VCN decode is enabled in VM mode\n"); +} + +/** + * vcn_v4_0_3_is_idle - check VCN block is idle + * + * @handle: amdgpu_device pointer + * + * Check whether VCN block is idle + */ +static bool vcn_v4_0_3_is_idle(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int i, ret = 1; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + ret &= (RREG32_SOC15(VCN, GET_INST(VCN, i), regUVD_STATUS) == + UVD_STATUS__IDLE); + } + + return ret; +} + +/** + * vcn_v4_0_3_wait_for_idle - wait for VCN block idle + * + * @handle: amdgpu_device pointer + * + * Wait for VCN block idle + */ +static int vcn_v4_0_3_wait_for_idle(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int i, ret = 0; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + ret = SOC15_WAIT_ON_RREG(VCN, GET_INST(VCN, i), regUVD_STATUS, + UVD_STATUS__IDLE, UVD_STATUS__IDLE); + if (ret) + return ret; + } + + return ret; +} + +/* vcn_v4_0_3_set_clockgating_state - set VCN block clockgating state + * + * @handle: amdgpu_device pointer + * @state: clock gating state + * + * Set VCN block clockgating state + */ +static int vcn_v4_0_3_set_clockgating_state(void *handle, + enum amd_clockgating_state state) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + bool enable = (state == AMD_CG_STATE_GATE) ? true : false; + int i; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + if (enable) { + if (RREG32_SOC15(VCN, GET_INST(VCN, i), + regUVD_STATUS) != UVD_STATUS__IDLE) + return -EBUSY; + vcn_v4_0_3_enable_clock_gating(adev, i); + } else { + vcn_v4_0_3_disable_clock_gating(adev, i); + } + } + return 0; +} + +/** + * vcn_v4_0_3_set_powergating_state - set VCN block powergating state + * + * @handle: amdgpu_device pointer + * @state: power gating state + * + * Set VCN block powergating state + */ +static int vcn_v4_0_3_set_powergating_state(void *handle, + enum amd_powergating_state state) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int ret; + + if (state == adev->vcn.cur_state) + return 0; + + if (state == AMD_PG_STATE_GATE) + ret = vcn_v4_0_3_stop(adev); + else + ret = vcn_v4_0_3_start(adev); + + if (!ret) + adev->vcn.cur_state = state; + + return ret; +} + +/** + * vcn_v4_0_3_set_interrupt_state - set VCN block interrupt state + * + * @adev: amdgpu_device pointer + * @source: interrupt sources + * @type: interrupt types + * @state: interrupt states + * + * Set VCN block interrupt state + */ +static int vcn_v4_0_3_set_interrupt_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned int type, + enum amdgpu_interrupt_state state) +{ + return 0; +} + +/** + * vcn_v4_0_3_process_interrupt - process VCN block interrupt + * + * @adev: amdgpu_device pointer + * @source: interrupt sources + * @entry: interrupt entry from clients and sources + * + * Process VCN block interrupt + */ +static int vcn_v4_0_3_process_interrupt(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + struct amdgpu_iv_entry *entry) +{ + uint32_t i, inst; + + i = node_id_to_phys_map[entry->node_id]; + + DRM_DEV_DEBUG(adev->dev, "IH: VCN TRAP\n"); + + for (inst = 0; inst < adev->vcn.num_vcn_inst; ++inst) + if (adev->vcn.inst[inst].aid_id == i) + break; + + if (inst >= adev->vcn.num_vcn_inst) { + dev_WARN_ONCE(adev->dev, 1, + "Interrupt received for unknown VCN instance %d", + entry->node_id); + return 0; + } + + switch (entry->src_id) { + case VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE: + amdgpu_fence_process(&adev->vcn.inst[inst].ring_enc[0]); + break; + default: + DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n", + entry->src_id, entry->src_data[0]); + break; + } + + return 0; +} + +static const struct amdgpu_irq_src_funcs vcn_v4_0_3_irq_funcs = { + .set = vcn_v4_0_3_set_interrupt_state, + .process = vcn_v4_0_3_process_interrupt, +}; + +/** + * vcn_v4_0_3_set_irq_funcs - set VCN block interrupt irq functions + * + * @adev: amdgpu_device pointer + * + * Set VCN block interrupt irq functions + */ +static void vcn_v4_0_3_set_irq_funcs(struct amdgpu_device *adev) +{ + int i; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + adev->vcn.inst->irq.num_types++; + } + adev->vcn.inst->irq.funcs = &vcn_v4_0_3_irq_funcs; +} + +static const struct amd_ip_funcs vcn_v4_0_3_ip_funcs = { + .name = "vcn_v4_0_3", + .early_init = vcn_v4_0_3_early_init, + .late_init = NULL, + .sw_init = vcn_v4_0_3_sw_init, + .sw_fini = vcn_v4_0_3_sw_fini, + .hw_init = vcn_v4_0_3_hw_init, + .hw_fini = vcn_v4_0_3_hw_fini, + .suspend = vcn_v4_0_3_suspend, + .resume = vcn_v4_0_3_resume, + .is_idle = vcn_v4_0_3_is_idle, + .wait_for_idle = vcn_v4_0_3_wait_for_idle, + .check_soft_reset = NULL, + .pre_soft_reset = NULL, + .soft_reset = NULL, + .post_soft_reset = NULL, + .set_clockgating_state = vcn_v4_0_3_set_clockgating_state, + .set_powergating_state = vcn_v4_0_3_set_powergating_state, +}; + +const struct amdgpu_ip_block_version vcn_v4_0_3_ip_block = { + .type = AMD_IP_BLOCK_TYPE_VCN, + .major = 4, + .minor = 0, + .rev = 3, + .funcs = &vcn_v4_0_3_ip_funcs, +}; + +static const struct amdgpu_ras_err_status_reg_entry vcn_v4_0_3_ue_reg_list[] = { + {AMDGPU_RAS_REG_ENTRY(VCN, 0, regVCN_UE_ERR_STATUS_LO_VIDD, regVCN_UE_ERR_STATUS_HI_VIDD), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "VIDD"}, + {AMDGPU_RAS_REG_ENTRY(VCN, 0, regVCN_UE_ERR_STATUS_LO_VIDV, regVCN_UE_ERR_STATUS_HI_VIDV), + 1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "VIDV"}, +}; + +static void vcn_v4_0_3_inst_query_ras_error_count(struct amdgpu_device *adev, + uint32_t vcn_inst, + void *ras_err_status) +{ + struct ras_err_data *err_data = (struct ras_err_data *)ras_err_status; + + /* vcn v4_0_3 only support query uncorrectable errors */ + amdgpu_ras_inst_query_ras_error_count(adev, + vcn_v4_0_3_ue_reg_list, + ARRAY_SIZE(vcn_v4_0_3_ue_reg_list), + NULL, 0, GET_INST(VCN, vcn_inst), + AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE, + &err_data->ue_count); +} + +static void vcn_v4_0_3_query_ras_error_count(struct amdgpu_device *adev, + void *ras_err_status) +{ + uint32_t i; + + if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) { + dev_warn(adev->dev, "VCN RAS is not supported\n"); + return; + } + + for (i = 0; i < adev->vcn.num_vcn_inst; i++) + vcn_v4_0_3_inst_query_ras_error_count(adev, i, ras_err_status); +} + +static void vcn_v4_0_3_inst_reset_ras_error_count(struct amdgpu_device *adev, + uint32_t vcn_inst) +{ + amdgpu_ras_inst_reset_ras_error_count(adev, + vcn_v4_0_3_ue_reg_list, + ARRAY_SIZE(vcn_v4_0_3_ue_reg_list), + GET_INST(VCN, vcn_inst)); +} + +static void vcn_v4_0_3_reset_ras_error_count(struct amdgpu_device *adev) +{ + uint32_t i; + + if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) { + dev_warn(adev->dev, "VCN RAS is not supported\n"); + return; + } + + for (i = 0; i < adev->vcn.num_vcn_inst; i++) + vcn_v4_0_3_inst_reset_ras_error_count(adev, i); +} + +static const struct amdgpu_ras_block_hw_ops vcn_v4_0_3_ras_hw_ops = { + .query_ras_error_count = vcn_v4_0_3_query_ras_error_count, + .reset_ras_error_count = vcn_v4_0_3_reset_ras_error_count, +}; + +static struct amdgpu_vcn_ras vcn_v4_0_3_ras = { + .ras_block = { + .hw_ops = &vcn_v4_0_3_ras_hw_ops, + }, +}; + +static void vcn_v4_0_3_set_ras_funcs(struct amdgpu_device *adev) +{ + adev->vcn.ras = &vcn_v4_0_3_ras; +} + +static void vcn_v4_0_3_enable_ras(struct amdgpu_device *adev, + int inst_idx, bool indirect) +{ + uint32_t tmp; + + if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) + return; + + tmp = VCN_RAS_CNTL__VCPU_VCODEC_REARM_MASK | + VCN_RAS_CNTL__VCPU_VCODEC_IH_EN_MASK | + VCN_RAS_CNTL__VCPU_VCODEC_PMI_EN_MASK | + VCN_RAS_CNTL__VCPU_VCODEC_STALL_EN_MASK; + WREG32_SOC15_DPG_MODE(inst_idx, + SOC15_DPG_MODE_OFFSET(VCN, 0, regVCN_RAS_CNTL), + tmp, 0, indirect); + + tmp = UVD_SYS_INT_EN__RASCNTL_VCPU_VCODEC_EN_MASK; + WREG32_SOC15_DPG_MODE(inst_idx, + SOC15_DPG_MODE_OFFSET(VCN, 0, regUVD_SYS_INT_EN), + tmp, 0, indirect); +} diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.h b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.h new file mode 100644 index 000000000000..0b046114373a --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.h @@ -0,0 +1,29 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef __VCN_V4_0_3_H__ +#define __VCN_V4_0_3_H__ + +extern const struct amdgpu_ip_block_version vcn_v4_0_3_ip_block; + +#endif /* __VCN_V4_0_3_H__ */ diff --git a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c index 536128447b71..4d719df376a7 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c @@ -334,7 +334,8 @@ static int vega20_ih_irq_init(struct amdgpu_device *adev) vega20_setup_retry_doorbell(adev->irq.retry_cam_doorbell_index)); /* Enable IH Retry CAM */ - if (adev->ip_versions[OSSSYS_HWIP][0] == IP_VERSION(4, 4, 0)) + if (adev->ip_versions[OSSSYS_HWIP][0] == IP_VERSION(4, 4, 0) || + adev->ip_versions[OSSSYS_HWIP][0] == IP_VERSION(4, 4, 2)) WREG32_FIELD15(OSSSYS, 0, IH_RETRY_INT_CAM_CNTL_ALDEBARAN, ENABLE, 1); else @@ -526,6 +527,7 @@ static int vega20_ih_early_init(void *handle) static int vega20_ih_sw_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + bool use_bus_addr = true; int r; r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_IH, 0, @@ -533,14 +535,18 @@ static int vega20_ih_sw_init(void *handle) if (r) return r; - r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 256 * 1024, true); + if ((adev->flags & AMD_IS_APU) && + (adev->ip_versions[OSSSYS_HWIP][0] == IP_VERSION(4, 4, 2))) + use_bus_addr = false; + + r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 256 * 1024, use_bus_addr); if (r) return r; adev->irq.ih.use_doorbell = true; adev->irq.ih.doorbell_index = adev->doorbell_index.ih << 1; - r = amdgpu_ih_ring_init(adev, &adev->irq.ih1, PAGE_SIZE, true); + r = amdgpu_ih_ring_init(adev, &adev->irq.ih1, PAGE_SIZE, use_bus_addr); if (r) return r; @@ -559,7 +565,7 @@ static int vega20_ih_sw_init(void *handle) /* initialize ih control registers offset */ vega20_ih_init_register_offset(adev); - r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, PAGE_SIZE, true); + r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, PAGE_SIZE, use_bus_addr); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c index 531f173ade2d..6a8494f98d3e 100644 --- a/drivers/gpu/drm/amd/amdgpu/vi.c +++ b/drivers/gpu/drm/amd/amdgpu/vi.c @@ -542,8 +542,15 @@ static u32 vi_get_xclk(struct amdgpu_device *adev) u32 reference_clock = adev->clock.spll.reference_freq; u32 tmp; - if (adev->flags & AMD_IS_APU) - return reference_clock; + if (adev->flags & AMD_IS_APU) { + switch (adev->asic_type) { + case CHIP_STONEY: + /* vbios says 48Mhz, but the actual freq is 100Mhz */ + return 10000; + default: + return reference_clock; + } + } tmp = RREG32_SMC(ixCG_CLKPIN_CNTL_2); if (REG_GET_FIELD(tmp, CG_CLKPIN_CNTL_2, MUX_TCLK_TO_XCLK)) @@ -580,11 +587,6 @@ void vi_srbm_select(struct amdgpu_device *adev, WREG32(mmSRBM_GFX_CNTL, srbm_gfx_cntl); } -static void vi_vga_set_state(struct amdgpu_device *adev, bool state) -{ - /* todo */ -} - static bool vi_read_disabled_bios(struct amdgpu_device *adev) { u32 bus_cntl; @@ -762,12 +764,12 @@ static uint32_t vi_get_register_value(struct amdgpu_device *adev, mutex_lock(&adev->grbm_idx_mutex); if (se_num != 0xffffffff || sh_num != 0xffffffff) - amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0); val = RREG32(reg_offset); if (se_num != 0xffffffff || sh_num != 0xffffffff) - amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0); mutex_unlock(&adev->grbm_idx_mutex); return val; } else { @@ -1435,7 +1437,6 @@ static const struct amdgpu_asic_funcs vi_asic_funcs = .read_register = &vi_read_register, .reset = &vi_asic_reset, .reset_method = &vi_asic_reset_method, - .set_vga_state = &vi_vga_set_state, .get_xclk = &vi_get_xclk, .set_uvd_clocks = &vi_set_uvd_clocks, .set_vce_clocks = &vi_set_vce_clocks, diff --git a/drivers/gpu/drm/amd/amdkfd/Makefile b/drivers/gpu/drm/amd/amdkfd/Makefile index e758c2a24cd0..2ec8f27c5366 100644 --- a/drivers/gpu/drm/amd/amdkfd/Makefile +++ b/drivers/gpu/drm/amd/amdkfd/Makefile @@ -53,9 +53,11 @@ AMDKFD_FILES := $(AMDKFD_PATH)/kfd_module.o \ $(AMDKFD_PATH)/kfd_events.o \ $(AMDKFD_PATH)/cik_event_interrupt.o \ $(AMDKFD_PATH)/kfd_int_process_v9.o \ + $(AMDKFD_PATH)/kfd_int_process_v10.o \ $(AMDKFD_PATH)/kfd_int_process_v11.o \ $(AMDKFD_PATH)/kfd_smi_events.o \ - $(AMDKFD_PATH)/kfd_crat.o + $(AMDKFD_PATH)/kfd_crat.o \ + $(AMDKFD_PATH)/kfd_debug.o ifneq ($(CONFIG_AMD_IOMMU_V2),) AMDKFD_FILES += $(AMDKFD_PATH)/kfd_iommu.o diff --git a/drivers/gpu/drm/amd/amdkfd/cik_event_interrupt.c b/drivers/gpu/drm/amd/amdkfd/cik_event_interrupt.c index 5c8023cba196..795382b55e0a 100644 --- a/drivers/gpu/drm/amd/amdkfd/cik_event_interrupt.c +++ b/drivers/gpu/drm/amd/amdkfd/cik_event_interrupt.c @@ -26,7 +26,7 @@ #include "amdgpu_amdkfd.h" #include "kfd_smi_events.h" -static bool cik_event_interrupt_isr(struct kfd_dev *dev, +static bool cik_event_interrupt_isr(struct kfd_node *dev, const uint32_t *ih_ring_entry, uint32_t *patched_ihre, bool *patched_flag) @@ -85,7 +85,7 @@ static bool cik_event_interrupt_isr(struct kfd_dev *dev, !amdgpu_no_queue_eviction_on_vm_fault); } -static void cik_event_interrupt_wq(struct kfd_dev *dev, +static void cik_event_interrupt_wq(struct kfd_node *dev, const uint32_t *ih_ring_entry) { const struct cik_ih_ring_entry *ihre = @@ -118,9 +118,9 @@ static void cik_event_interrupt_wq(struct kfd_dev *dev, return; if (info.vmid == vmid) - kfd_signal_vm_fault_event(dev, pasid, &info); + kfd_signal_vm_fault_event(dev, pasid, &info, NULL); else - kfd_signal_vm_fault_event(dev, pasid, NULL); + kfd_signal_vm_fault_event(dev, pasid, NULL, NULL); } } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c index 1b54a9aaae70..6a27b000a246 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c @@ -44,6 +44,7 @@ #include "amdgpu_amdkfd.h" #include "kfd_smi_events.h" #include "amdgpu_dma_buf.h" +#include "kfd_debug.h" static long kfd_ioctl(struct file *, unsigned int, unsigned long); static int kfd_open(struct inode *, struct file *); @@ -142,15 +143,13 @@ static int kfd_open(struct inode *inode, struct file *filep) return -EPERM; } - process = kfd_create_process(filep); + process = kfd_create_process(current); if (IS_ERR(process)) return PTR_ERR(process); - if (kfd_is_locked()) { - dev_dbg(kfd_device, "kfd is locked!\n" - "process %d unreferenced", process->pasid); + if (kfd_process_init_cwsr_apu(process, filep)) { kfd_unref_process(process); - return -EAGAIN; + return -EFAULT; } /* filep now owns the reference returned by kfd_create_process */ @@ -186,7 +185,12 @@ static int kfd_ioctl_get_version(struct file *filep, struct kfd_process *p, static int set_queue_properties_from_user(struct queue_properties *q_properties, struct kfd_ioctl_create_queue_args *args) { - if (args->queue_percentage > KFD_MAX_QUEUE_PERCENTAGE) { + /* + * Repurpose queue percentage to accommodate new features: + * bit 0-7: queue percentage + * bit 8-15: pm4_target_xcc + */ + if ((args->queue_percentage & 0xFF) > KFD_MAX_QUEUE_PERCENTAGE) { pr_err("Queue percentage must be between 0 to KFD_MAX_QUEUE_PERCENTAGE\n"); return -EINVAL; } @@ -236,7 +240,9 @@ static int set_queue_properties_from_user(struct queue_properties *q_properties, q_properties->is_interop = false; q_properties->is_gws = false; - q_properties->queue_percent = args->queue_percentage; + q_properties->queue_percent = args->queue_percentage & 0xFF; + /* bit 8-15 are repurposed to be PM4 target XCC */ + q_properties->pm4_target_xcc = (args->queue_percentage >> 8) & 0xFF; q_properties->priority = args->queue_priority; q_properties->queue_address = args->ring_base_address; q_properties->queue_size = args->ring_size; @@ -293,7 +299,7 @@ static int kfd_ioctl_create_queue(struct file *filep, struct kfd_process *p, void *data) { struct kfd_ioctl_create_queue_args *args = data; - struct kfd_dev *dev; + struct kfd_node *dev; int err = 0; unsigned int queue_id; struct kfd_process_device *pdd; @@ -328,7 +334,7 @@ static int kfd_ioctl_create_queue(struct file *filep, struct kfd_process *p, } if (!pdd->doorbell_index && - kfd_alloc_process_doorbells(dev, &pdd->doorbell_index) < 0) { + kfd_alloc_process_doorbells(dev->kfd, &pdd->doorbell_index) < 0) { err = -ENOMEM; goto err_alloc_doorbells; } @@ -336,7 +342,7 @@ static int kfd_ioctl_create_queue(struct file *filep, struct kfd_process *p, /* Starting with GFX11, wptr BOs must be mapped to GART for MES to determine work * on unmapped queues for usermode queue oversubscription (no aggregated doorbell) */ - if (dev->shared_resources.enable_mes && + if (dev->kfd->shared_resources.enable_mes && ((dev->adev->mes.sched_version & AMDGPU_MES_API_VERSION_MASK) >> AMDGPU_MES_API_VERSION_SHIFT) >= 2) { struct amdgpu_bo_va_mapping *wptr_mapping; @@ -404,6 +410,7 @@ static int kfd_ioctl_create_queue(struct file *filep, struct kfd_process *p, pr_debug("Write ptr address == 0x%016llX\n", args->write_pointer_address); + kfd_dbg_ev_raise(KFD_EC_MASK(EC_QUEUE_NEW), p, dev, queue_id, false, NULL, 0); return 0; err_create_queue: @@ -442,7 +449,12 @@ static int kfd_ioctl_update_queue(struct file *filp, struct kfd_process *p, struct kfd_ioctl_update_queue_args *args = data; struct queue_properties properties; - if (args->queue_percentage > KFD_MAX_QUEUE_PERCENTAGE) { + /* + * Repurpose queue percentage to accommodate new features: + * bit 0-7: queue percentage + * bit 8-15: pm4_target_xcc + */ + if ((args->queue_percentage & 0xFF) > KFD_MAX_QUEUE_PERCENTAGE) { pr_err("Queue percentage must be between 0 to KFD_MAX_QUEUE_PERCENTAGE\n"); return -EINVAL; } @@ -466,7 +478,9 @@ static int kfd_ioctl_update_queue(struct file *filp, struct kfd_process *p, properties.queue_address = args->ring_base_address; properties.queue_size = args->ring_size; - properties.queue_percent = args->queue_percentage; + properties.queue_percent = args->queue_percentage & 0xFF; + /* bit 8-15 are repurposed to be PM4 target XCC */ + properties.pm4_target_xcc = (args->queue_percentage >> 8) & 0xFF; properties.priority = args->queue_priority; pr_debug("Updating queue id %d for pasid 0x%x\n", @@ -524,8 +538,6 @@ static int kfd_ioctl_set_cu_mask(struct file *filp, struct kfd_process *p, goto out; } - minfo.update_flag = UPDATE_FLAG_CU_MASK; - mutex_lock(&p->mutex); retval = pqm_update_mqd(&p->pqm, args->queue_id, &minfo); @@ -887,7 +899,7 @@ static int kfd_ioctl_set_scratch_backing_va(struct file *filep, { struct kfd_ioctl_set_scratch_backing_va_args *args = data; struct kfd_process_device *pdd; - struct kfd_dev *dev; + struct kfd_node *dev; long err; mutex_lock(&p->mutex); @@ -1006,19 +1018,26 @@ err_drm_file: return ret; } -bool kfd_dev_is_large_bar(struct kfd_dev *dev) +bool kfd_dev_is_large_bar(struct kfd_node *dev) { if (debug_largebar) { pr_debug("Simulate large-bar allocation on non large-bar machine\n"); return true; } - if (dev->use_iommu_v2) + if (dev->kfd->use_iommu_v2) return false; if (dev->local_mem_info.local_mem_size_private == 0 && - dev->local_mem_info.local_mem_size_public > 0) + dev->local_mem_info.local_mem_size_public > 0) + return true; + + if (dev->local_mem_info.local_mem_size_public == 0 && + dev->kfd->adev->gmc.is_app_apu) { + pr_debug("APP APU, Consider like a large bar system\n"); return true; + } + return false; } @@ -1030,7 +1049,8 @@ static int kfd_ioctl_get_available_memory(struct file *filep, if (!pdd) return -EINVAL; - args->available = amdgpu_amdkfd_get_available_memory(pdd->dev->adev); + args->available = amdgpu_amdkfd_get_available_memory(pdd->dev->adev, + pdd->dev->node_id); kfd_unlock_pdd(pdd); return 0; } @@ -1041,7 +1061,7 @@ static int kfd_ioctl_alloc_memory_of_gpu(struct file *filep, struct kfd_ioctl_alloc_memory_of_gpu_args *args = data; struct kfd_process_device *pdd; void *mem; - struct kfd_dev *dev; + struct kfd_node *dev; int idr_handle; long err; uint64_t offset = args->mmap_offset; @@ -1105,7 +1125,7 @@ static int kfd_ioctl_alloc_memory_of_gpu(struct file *filep, } if (flags & KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL) { - if (args->size != kfd_doorbell_process_slice(dev)) { + if (args->size != kfd_doorbell_process_slice(dev->kfd)) { err = -EINVAL; goto err_unlock; } @@ -1231,7 +1251,7 @@ static int kfd_ioctl_map_memory_to_gpu(struct file *filep, struct kfd_ioctl_map_memory_to_gpu_args *args = data; struct kfd_process_device *pdd, *peer_pdd; void *mem; - struct kfd_dev *dev; + struct kfd_node *dev; long err = 0; int i; uint32_t *devices_arr = NULL; @@ -1405,7 +1425,7 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep, args->n_success = i+1; } - flush_tlb = kfd_flush_tlb_after_unmap(pdd->dev); + flush_tlb = kfd_flush_tlb_after_unmap(pdd->dev->kfd); if (flush_tlb) { err = amdgpu_amdkfd_gpuvm_sync_memory(pdd->dev->adev, (struct kgd_mem *) mem, true); @@ -1445,7 +1465,7 @@ static int kfd_ioctl_alloc_queue_gws(struct file *filep, int retval; struct kfd_ioctl_alloc_queue_gws_args *args = data; struct queue *q; - struct kfd_dev *dev; + struct kfd_node *dev; mutex_lock(&p->mutex); q = pqm_get_user_queue(&p->pqm, args->queue_id); @@ -1467,6 +1487,11 @@ static int kfd_ioctl_alloc_queue_gws(struct file *filep, goto out_unlock; } + if (!kfd_dbg_has_gws_support(dev) && p->debug_trap_enabled) { + retval = -EBUSY; + goto out_unlock; + } + retval = pqm_set_gws(&p->pqm, args->queue_id, args->num_gws ? dev->gws : NULL); mutex_unlock(&p->mutex); @@ -1482,10 +1507,11 @@ static int kfd_ioctl_get_dmabuf_info(struct file *filep, struct kfd_process *p, void *data) { struct kfd_ioctl_get_dmabuf_info_args *args = data; - struct kfd_dev *dev = NULL; + struct kfd_node *dev = NULL; struct amdgpu_device *dmabuf_adev; void *metadata_buffer = NULL; uint32_t flags; + int8_t xcp_id; unsigned int i; int r; @@ -1506,17 +1532,14 @@ static int kfd_ioctl_get_dmabuf_info(struct file *filep, r = amdgpu_amdkfd_get_dmabuf_info(dev->adev, args->dmabuf_fd, &dmabuf_adev, &args->size, metadata_buffer, args->metadata_size, - &args->metadata_size, &flags); + &args->metadata_size, &flags, &xcp_id); if (r) goto exit; - /* Reverse-lookup gpu_id from kgd pointer */ - dev = kfd_device_by_adev(dmabuf_adev); - if (!dev) { - r = -EINVAL; - goto exit; - } - args->gpu_id = dev->id; + if (xcp_id >= 0) + args->gpu_id = dmabuf_adev->kfd.dev->nodes[xcp_id]->id; + else + args->gpu_id = dmabuf_adev->kfd.dev->nodes[0]->id; args->flags = flags; /* Copy metadata buffer to user mode */ @@ -1596,7 +1619,7 @@ static int kfd_ioctl_export_dmabuf(struct file *filep, struct kfd_ioctl_export_dmabuf_args *args = data; struct kfd_process_device *pdd; struct dma_buf *dmabuf; - struct kfd_dev *dev; + struct kfd_node *dev; void *mem; int ret = 0; @@ -2178,7 +2201,7 @@ static int criu_restore_devices(struct kfd_process *p, } for (i = 0; i < args->num_devices; i++) { - struct kfd_dev *dev; + struct kfd_node *dev; struct kfd_process_device *pdd; struct file *drm_file; @@ -2240,7 +2263,7 @@ static int criu_restore_devices(struct kfd_process *p, } if (!pdd->doorbell_index && - kfd_alloc_process_doorbells(pdd->dev, &pdd->doorbell_index) < 0) { + kfd_alloc_process_doorbells(pdd->dev->kfd, &pdd->doorbell_index) < 0) { ret = -ENOMEM; goto exit; } @@ -2268,7 +2291,8 @@ static int criu_restore_memory_of_gpu(struct kfd_process_device *pdd, u64 offset; if (bo_bucket->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL) { - if (bo_bucket->size != kfd_doorbell_process_slice(pdd->dev)) + if (bo_bucket->size != + kfd_doorbell_process_slice(pdd->dev->kfd)) return -EINVAL; offset = kfd_get_process_doorbells(pdd); @@ -2350,7 +2374,7 @@ static int criu_restore_bo(struct kfd_process *p, /* now map these BOs to GPU/s */ for (j = 0; j < p->n_pdds; j++) { - struct kfd_dev *peer; + struct kfd_node *peer; struct kfd_process_device *peer_pdd; if (!bo_priv->mapped_gpuids[j]) @@ -2715,6 +2739,356 @@ static int kfd_ioctl_criu(struct file *filep, struct kfd_process *p, void *data) return ret; } +static int runtime_enable(struct kfd_process *p, uint64_t r_debug, + bool enable_ttmp_setup) +{ + int i = 0, ret = 0; + + if (p->is_runtime_retry) + goto retry; + + if (p->runtime_info.runtime_state != DEBUG_RUNTIME_STATE_DISABLED) + return -EBUSY; + + for (i = 0; i < p->n_pdds; i++) { + struct kfd_process_device *pdd = p->pdds[i]; + + if (pdd->qpd.queue_count) + return -EEXIST; + } + + p->runtime_info.runtime_state = DEBUG_RUNTIME_STATE_ENABLED; + p->runtime_info.r_debug = r_debug; + p->runtime_info.ttmp_setup = enable_ttmp_setup; + + if (p->runtime_info.ttmp_setup) { + for (i = 0; i < p->n_pdds; i++) { + struct kfd_process_device *pdd = p->pdds[i]; + + if (!kfd_dbg_is_rlc_restore_supported(pdd->dev)) { + amdgpu_gfx_off_ctrl(pdd->dev->adev, false); + pdd->dev->kfd2kgd->enable_debug_trap( + pdd->dev->adev, + true, + pdd->dev->vm_info.last_vmid_kfd); + } else if (kfd_dbg_is_per_vmid_supported(pdd->dev)) { + pdd->spi_dbg_override = pdd->dev->kfd2kgd->enable_debug_trap( + pdd->dev->adev, + false, + 0); + } + } + } + +retry: + if (p->debug_trap_enabled) { + if (!p->is_runtime_retry) { + kfd_dbg_trap_activate(p); + kfd_dbg_ev_raise(KFD_EC_MASK(EC_PROCESS_RUNTIME), + p, NULL, 0, false, NULL, 0); + } + + mutex_unlock(&p->mutex); + ret = down_interruptible(&p->runtime_enable_sema); + mutex_lock(&p->mutex); + + p->is_runtime_retry = !!ret; + } + + return ret; +} + +static int runtime_disable(struct kfd_process *p) +{ + int i = 0, ret; + bool was_enabled = p->runtime_info.runtime_state == DEBUG_RUNTIME_STATE_ENABLED; + + p->runtime_info.runtime_state = DEBUG_RUNTIME_STATE_DISABLED; + p->runtime_info.r_debug = 0; + + if (p->debug_trap_enabled) { + if (was_enabled) + kfd_dbg_trap_deactivate(p, false, 0); + + if (!p->is_runtime_retry) + kfd_dbg_ev_raise(KFD_EC_MASK(EC_PROCESS_RUNTIME), + p, NULL, 0, false, NULL, 0); + + mutex_unlock(&p->mutex); + ret = down_interruptible(&p->runtime_enable_sema); + mutex_lock(&p->mutex); + + p->is_runtime_retry = !!ret; + if (ret) + return ret; + } + + if (was_enabled && p->runtime_info.ttmp_setup) { + for (i = 0; i < p->n_pdds; i++) { + struct kfd_process_device *pdd = p->pdds[i]; + + if (!kfd_dbg_is_rlc_restore_supported(pdd->dev)) + amdgpu_gfx_off_ctrl(pdd->dev->adev, true); + } + } + + p->runtime_info.ttmp_setup = false; + + /* disable ttmp setup */ + for (i = 0; i < p->n_pdds; i++) { + struct kfd_process_device *pdd = p->pdds[i]; + + if (kfd_dbg_is_per_vmid_supported(pdd->dev)) { + pdd->spi_dbg_override = + pdd->dev->kfd2kgd->disable_debug_trap( + pdd->dev->adev, + false, + pdd->dev->vm_info.last_vmid_kfd); + + if (!pdd->dev->kfd->shared_resources.enable_mes) + debug_refresh_runlist(pdd->dev->dqm); + else + kfd_dbg_set_mes_debug_mode(pdd); + } + } + + return 0; +} + +static int kfd_ioctl_runtime_enable(struct file *filep, struct kfd_process *p, void *data) +{ + struct kfd_ioctl_runtime_enable_args *args = data; + int r; + + mutex_lock(&p->mutex); + + if (args->mode_mask & KFD_RUNTIME_ENABLE_MODE_ENABLE_MASK) + r = runtime_enable(p, args->r_debug, + !!(args->mode_mask & KFD_RUNTIME_ENABLE_MODE_TTMP_SAVE_MASK)); + else + r = runtime_disable(p); + + mutex_unlock(&p->mutex); + + return r; +} + +static int kfd_ioctl_set_debug_trap(struct file *filep, struct kfd_process *p, void *data) +{ + struct kfd_ioctl_dbg_trap_args *args = data; + struct task_struct *thread = NULL; + struct mm_struct *mm = NULL; + struct pid *pid = NULL; + struct kfd_process *target = NULL; + struct kfd_process_device *pdd = NULL; + int r = 0; + + if (sched_policy == KFD_SCHED_POLICY_NO_HWS) { + pr_err("Debugging does not support sched_policy %i", sched_policy); + return -EINVAL; + } + + pid = find_get_pid(args->pid); + if (!pid) { + pr_debug("Cannot find pid info for %i\n", args->pid); + r = -ESRCH; + goto out; + } + + thread = get_pid_task(pid, PIDTYPE_PID); + if (!thread) { + r = -ESRCH; + goto out; + } + + mm = get_task_mm(thread); + if (!mm) { + r = -ESRCH; + goto out; + } + + if (args->op == KFD_IOC_DBG_TRAP_ENABLE) { + bool create_process; + + rcu_read_lock(); + create_process = thread && thread != current && ptrace_parent(thread) == current; + rcu_read_unlock(); + + target = create_process ? kfd_create_process(thread) : + kfd_lookup_process_by_pid(pid); + } else { + target = kfd_lookup_process_by_pid(pid); + } + + if (IS_ERR_OR_NULL(target)) { + pr_debug("Cannot find process PID %i to debug\n", args->pid); + r = target ? PTR_ERR(target) : -ESRCH; + goto out; + } + + /* Check if target is still PTRACED. */ + rcu_read_lock(); + if (target != p && args->op != KFD_IOC_DBG_TRAP_DISABLE + && ptrace_parent(target->lead_thread) != current) { + pr_err("PID %i is not PTRACED and cannot be debugged\n", args->pid); + r = -EPERM; + } + rcu_read_unlock(); + + if (r) + goto out; + + mutex_lock(&target->mutex); + + if (args->op != KFD_IOC_DBG_TRAP_ENABLE && !target->debug_trap_enabled) { + pr_err("PID %i not debug enabled for op %i\n", args->pid, args->op); + r = -EINVAL; + goto unlock_out; + } + + if (target->runtime_info.runtime_state != DEBUG_RUNTIME_STATE_ENABLED && + (args->op == KFD_IOC_DBG_TRAP_SET_WAVE_LAUNCH_OVERRIDE || + args->op == KFD_IOC_DBG_TRAP_SET_WAVE_LAUNCH_MODE || + args->op == KFD_IOC_DBG_TRAP_SUSPEND_QUEUES || + args->op == KFD_IOC_DBG_TRAP_RESUME_QUEUES || + args->op == KFD_IOC_DBG_TRAP_SET_NODE_ADDRESS_WATCH || + args->op == KFD_IOC_DBG_TRAP_CLEAR_NODE_ADDRESS_WATCH || + args->op == KFD_IOC_DBG_TRAP_SET_FLAGS)) { + r = -EPERM; + goto unlock_out; + } + + if (args->op == KFD_IOC_DBG_TRAP_SET_NODE_ADDRESS_WATCH || + args->op == KFD_IOC_DBG_TRAP_CLEAR_NODE_ADDRESS_WATCH) { + int user_gpu_id = kfd_process_get_user_gpu_id(target, + args->op == KFD_IOC_DBG_TRAP_SET_NODE_ADDRESS_WATCH ? + args->set_node_address_watch.gpu_id : + args->clear_node_address_watch.gpu_id); + + pdd = kfd_process_device_data_by_id(target, user_gpu_id); + if (user_gpu_id == -EINVAL || !pdd) { + r = -ENODEV; + goto unlock_out; + } + } + + switch (args->op) { + case KFD_IOC_DBG_TRAP_ENABLE: + if (target != p) + target->debugger_process = p; + + r = kfd_dbg_trap_enable(target, + args->enable.dbg_fd, + (void __user *)args->enable.rinfo_ptr, + &args->enable.rinfo_size); + if (!r) + target->exception_enable_mask = args->enable.exception_mask; + + break; + case KFD_IOC_DBG_TRAP_DISABLE: + r = kfd_dbg_trap_disable(target); + break; + case KFD_IOC_DBG_TRAP_SEND_RUNTIME_EVENT: + r = kfd_dbg_send_exception_to_runtime(target, + args->send_runtime_event.gpu_id, + args->send_runtime_event.queue_id, + args->send_runtime_event.exception_mask); + break; + case KFD_IOC_DBG_TRAP_SET_EXCEPTIONS_ENABLED: + kfd_dbg_set_enabled_debug_exception_mask(target, + args->set_exceptions_enabled.exception_mask); + break; + case KFD_IOC_DBG_TRAP_SET_WAVE_LAUNCH_OVERRIDE: + r = kfd_dbg_trap_set_wave_launch_override(target, + args->launch_override.override_mode, + args->launch_override.enable_mask, + args->launch_override.support_request_mask, + &args->launch_override.enable_mask, + &args->launch_override.support_request_mask); + break; + case KFD_IOC_DBG_TRAP_SET_WAVE_LAUNCH_MODE: + r = kfd_dbg_trap_set_wave_launch_mode(target, + args->launch_mode.launch_mode); + break; + case KFD_IOC_DBG_TRAP_SUSPEND_QUEUES: + r = suspend_queues(target, + args->suspend_queues.num_queues, + args->suspend_queues.grace_period, + args->suspend_queues.exception_mask, + (uint32_t *)args->suspend_queues.queue_array_ptr); + + break; + case KFD_IOC_DBG_TRAP_RESUME_QUEUES: + r = resume_queues(target, args->resume_queues.num_queues, + (uint32_t *)args->resume_queues.queue_array_ptr); + break; + case KFD_IOC_DBG_TRAP_SET_NODE_ADDRESS_WATCH: + r = kfd_dbg_trap_set_dev_address_watch(pdd, + args->set_node_address_watch.address, + args->set_node_address_watch.mask, + &args->set_node_address_watch.id, + args->set_node_address_watch.mode); + break; + case KFD_IOC_DBG_TRAP_CLEAR_NODE_ADDRESS_WATCH: + r = kfd_dbg_trap_clear_dev_address_watch(pdd, + args->clear_node_address_watch.id); + break; + case KFD_IOC_DBG_TRAP_SET_FLAGS: + r = kfd_dbg_trap_set_flags(target, &args->set_flags.flags); + break; + case KFD_IOC_DBG_TRAP_QUERY_DEBUG_EVENT: + r = kfd_dbg_ev_query_debug_event(target, + &args->query_debug_event.queue_id, + &args->query_debug_event.gpu_id, + args->query_debug_event.exception_mask, + &args->query_debug_event.exception_mask); + break; + case KFD_IOC_DBG_TRAP_QUERY_EXCEPTION_INFO: + r = kfd_dbg_trap_query_exception_info(target, + args->query_exception_info.source_id, + args->query_exception_info.exception_code, + args->query_exception_info.clear_exception, + (void __user *)args->query_exception_info.info_ptr, + &args->query_exception_info.info_size); + break; + case KFD_IOC_DBG_TRAP_GET_QUEUE_SNAPSHOT: + r = pqm_get_queue_snapshot(&target->pqm, + args->queue_snapshot.exception_mask, + (void __user *)args->queue_snapshot.snapshot_buf_ptr, + &args->queue_snapshot.num_queues, + &args->queue_snapshot.entry_size); + break; + case KFD_IOC_DBG_TRAP_GET_DEVICE_SNAPSHOT: + r = kfd_dbg_trap_device_snapshot(target, + args->device_snapshot.exception_mask, + (void __user *)args->device_snapshot.snapshot_buf_ptr, + &args->device_snapshot.num_devices, + &args->device_snapshot.entry_size); + break; + default: + pr_err("Invalid option: %i\n", args->op); + r = -EINVAL; + } + +unlock_out: + mutex_unlock(&target->mutex); + +out: + if (thread) + put_task_struct(thread); + + if (mm) + mmput(mm); + + if (pid) + put_pid(pid); + + if (target) + kfd_unref_process(target); + + return r; +} + #define AMDKFD_IOCTL_DEF(ioctl, _func, _flags) \ [_IOC_NR(ioctl)] = {.cmd = ioctl, .func = _func, .flags = _flags, \ .cmd_drv = 0, .name = #ioctl} @@ -2827,6 +3201,12 @@ static const struct amdkfd_ioctl_desc amdkfd_ioctls[] = { AMDKFD_IOCTL_DEF(AMDKFD_IOC_EXPORT_DMABUF, kfd_ioctl_export_dmabuf, 0), + + AMDKFD_IOCTL_DEF(AMDKFD_IOC_RUNTIME_ENABLE, + kfd_ioctl_runtime_enable, 0), + + AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_TRAP, + kfd_ioctl_set_debug_trap, 0), }; #define AMDKFD_CORE_IOCTL_COUNT ARRAY_SIZE(amdkfd_ioctls) @@ -2947,7 +3327,7 @@ err_i1: return retcode; } -static int kfd_mmio_mmap(struct kfd_dev *dev, struct kfd_process *process, +static int kfd_mmio_mmap(struct kfd_node *dev, struct kfd_process *process, struct vm_area_struct *vma) { phys_addr_t address; @@ -2981,7 +3361,7 @@ static int kfd_mmio_mmap(struct kfd_dev *dev, struct kfd_process *process, static int kfd_mmap(struct file *filp, struct vm_area_struct *vma) { struct kfd_process *process; - struct kfd_dev *dev = NULL; + struct kfd_node *dev = NULL; unsigned long mmap_offset; unsigned int gpu_id; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c index 475e47027354..49f40d9f16e8 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c @@ -1166,7 +1166,7 @@ static int kfd_parse_subtype_iolink(struct crat_subtype_iolink *iolink, if (props->iolink_type == CRAT_IOLINK_TYPE_PCIEXPRESS) props->weight = 20; else if (props->iolink_type == CRAT_IOLINK_TYPE_XGMI) - props->weight = 15 * iolink->num_hops_xgmi; + props->weight = iolink->weight_xgmi; else props->weight = node_distance(id_from, id_to); @@ -1405,7 +1405,7 @@ static int kfd_fill_gpu_cache_info_from_gfx_config(struct kfd_dev *kdev, return i; } -int kfd_get_gpu_cache_info(struct kfd_dev *kdev, struct kfd_gpu_cache_info **pcache_info) +int kfd_get_gpu_cache_info(struct kfd_node *kdev, struct kfd_gpu_cache_info **pcache_info) { int num_of_cache_types = 0; @@ -1524,7 +1524,7 @@ int kfd_get_gpu_cache_info(struct kfd_dev *kdev, struct kfd_gpu_cache_info **pca case IP_VERSION(11, 0, 3): case IP_VERSION(11, 0, 4): num_of_cache_types = - kfd_fill_gpu_cache_info_from_gfx_config(kdev, *pcache_info); + kfd_fill_gpu_cache_info_from_gfx_config(kdev->kfd, *pcache_info); break; default: *pcache_info = dummy_cache_info; @@ -1858,7 +1858,7 @@ static int kfd_create_vcrat_image_cpu(void *pcrat_image, size_t *size) } static int kfd_fill_gpu_memory_affinity(int *avail_size, - struct kfd_dev *kdev, uint8_t type, uint64_t size, + struct kfd_node *kdev, uint8_t type, uint64_t size, struct crat_subtype_memory *sub_type_hdr, uint32_t proximity_domain, const struct kfd_local_mem_info *local_mem_info) @@ -1887,7 +1887,7 @@ static int kfd_fill_gpu_memory_affinity(int *avail_size, } #ifdef CONFIG_ACPI_NUMA -static void kfd_find_numa_node_in_srat(struct kfd_dev *kdev) +static void kfd_find_numa_node_in_srat(struct kfd_node *kdev) { struct acpi_table_header *table_header = NULL; struct acpi_subtable_header *sub_header = NULL; @@ -1972,6 +1972,9 @@ static void kfd_find_numa_node_in_srat(struct kfd_dev *kdev) } #endif +#define KFD_CRAT_INTRA_SOCKET_WEIGHT 13 +#define KFD_CRAT_XGMI_WEIGHT 15 + /* kfd_fill_gpu_direct_io_link - Fill in direct io link from GPU * to its NUMA node * @avail_size: Available size in the memory @@ -1982,7 +1985,7 @@ static void kfd_find_numa_node_in_srat(struct kfd_dev *kdev) * Return 0 if successful else return -ve value */ static int kfd_fill_gpu_direct_io_link_to_cpu(int *avail_size, - struct kfd_dev *kdev, + struct kfd_node *kdev, struct crat_subtype_iolink *sub_type_hdr, uint32_t proximity_domain) { @@ -2002,7 +2005,16 @@ static int kfd_fill_gpu_direct_io_link_to_cpu(int *avail_size, /* Fill in IOLINK subtype. * TODO: Fill-in other fields of iolink subtype */ - if (kdev->adev->gmc.xgmi.connected_to_cpu) { + if (kdev->adev->gmc.xgmi.connected_to_cpu || + (KFD_GC_VERSION(kdev) == IP_VERSION(9, 4, 3) && + kdev->adev->smuio.funcs->get_pkg_type(kdev->adev) == + AMDGPU_PKG_TYPE_APU)) { + bool ext_cpu = KFD_GC_VERSION(kdev) != IP_VERSION(9, 4, 3); + int mem_bw = 819200, weight = ext_cpu ? KFD_CRAT_XGMI_WEIGHT : + KFD_CRAT_INTRA_SOCKET_WEIGHT; + uint32_t bandwidth = ext_cpu ? amdgpu_amdkfd_get_xgmi_bandwidth_mbytes( + kdev->adev, NULL, true) : mem_bw; + /* * with host gpu xgmi link, host can access gpu memory whether * or not pcie bar type is large, so always create bidirectional @@ -2010,14 +2022,9 @@ static int kfd_fill_gpu_direct_io_link_to_cpu(int *avail_size, */ sub_type_hdr->flags |= CRAT_IOLINK_FLAGS_BI_DIRECTIONAL; sub_type_hdr->io_interface_type = CRAT_IOLINK_TYPE_XGMI; - sub_type_hdr->num_hops_xgmi = 1; - if (KFD_GC_VERSION(kdev) == IP_VERSION(9, 4, 2)) { - sub_type_hdr->minimum_bandwidth_mbs = - amdgpu_amdkfd_get_xgmi_bandwidth_mbytes( - kdev->adev, NULL, true); - sub_type_hdr->maximum_bandwidth_mbs = - sub_type_hdr->minimum_bandwidth_mbs; - } + sub_type_hdr->weight_xgmi = weight; + sub_type_hdr->minimum_bandwidth_mbs = bandwidth; + sub_type_hdr->maximum_bandwidth_mbs = bandwidth; } else { sub_type_hdr->io_interface_type = CRAT_IOLINK_TYPE_PCIEXPRESS; sub_type_hdr->minimum_bandwidth_mbs = @@ -2029,7 +2036,8 @@ static int kfd_fill_gpu_direct_io_link_to_cpu(int *avail_size, sub_type_hdr->proximity_domain_from = proximity_domain; #ifdef CONFIG_ACPI_NUMA - if (kdev->adev->pdev->dev.numa_node == NUMA_NO_NODE) + if (kdev->adev->pdev->dev.numa_node == NUMA_NO_NODE && + num_possible_nodes() > 1) kfd_find_numa_node_in_srat(kdev); #endif #ifdef CONFIG_NUMA @@ -2044,12 +2052,14 @@ static int kfd_fill_gpu_direct_io_link_to_cpu(int *avail_size, } static int kfd_fill_gpu_xgmi_link_to_gpu(int *avail_size, - struct kfd_dev *kdev, - struct kfd_dev *peer_kdev, + struct kfd_node *kdev, + struct kfd_node *peer_kdev, struct crat_subtype_iolink *sub_type_hdr, uint32_t proximity_domain_from, uint32_t proximity_domain_to) { + bool use_ta_info = kdev->kfd->num_nodes == 1; + *avail_size -= sizeof(struct crat_subtype_iolink); if (*avail_size < 0) return -ENOMEM; @@ -2064,12 +2074,25 @@ static int kfd_fill_gpu_xgmi_link_to_gpu(int *avail_size, sub_type_hdr->io_interface_type = CRAT_IOLINK_TYPE_XGMI; sub_type_hdr->proximity_domain_from = proximity_domain_from; sub_type_hdr->proximity_domain_to = proximity_domain_to; - sub_type_hdr->num_hops_xgmi = - amdgpu_amdkfd_get_xgmi_hops_count(kdev->adev, peer_kdev->adev); - sub_type_hdr->maximum_bandwidth_mbs = - amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(kdev->adev, peer_kdev->adev, false); - sub_type_hdr->minimum_bandwidth_mbs = sub_type_hdr->maximum_bandwidth_mbs ? - amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(kdev->adev, NULL, true) : 0; + + if (use_ta_info) { + sub_type_hdr->weight_xgmi = KFD_CRAT_XGMI_WEIGHT * + amdgpu_amdkfd_get_xgmi_hops_count(kdev->adev, peer_kdev->adev); + sub_type_hdr->maximum_bandwidth_mbs = + amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(kdev->adev, + peer_kdev->adev, false); + sub_type_hdr->minimum_bandwidth_mbs = sub_type_hdr->maximum_bandwidth_mbs ? + amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(kdev->adev, NULL, true) : 0; + } else { + bool is_single_hop = kdev->kfd == peer_kdev->kfd; + int weight = is_single_hop ? KFD_CRAT_INTRA_SOCKET_WEIGHT : + (2 * KFD_CRAT_INTRA_SOCKET_WEIGHT) + KFD_CRAT_XGMI_WEIGHT; + int mem_bw = 819200; + + sub_type_hdr->weight_xgmi = weight; + sub_type_hdr->maximum_bandwidth_mbs = is_single_hop ? mem_bw : 0; + sub_type_hdr->minimum_bandwidth_mbs = is_single_hop ? mem_bw : 0; + } return 0; } @@ -2081,7 +2104,7 @@ static int kfd_fill_gpu_xgmi_link_to_gpu(int *avail_size, * [OUT] actual size of data filled in crat_image */ static int kfd_create_vcrat_image_gpu(void *pcrat_image, - size_t *size, struct kfd_dev *kdev, + size_t *size, struct kfd_node *kdev, uint32_t proximity_domain) { struct crat_header *crat_table = (struct crat_header *)pcrat_image; @@ -2153,7 +2176,7 @@ static int kfd_create_vcrat_image_gpu(void *pcrat_image, /* Check if this node supports IOMMU. During parsing this flag will * translate to HSA_CAP_ATS_PRESENT */ - if (!kfd_iommu_check_device(kdev)) + if (!kfd_iommu_check_device(kdev->kfd)) cu->hsa_capability |= CRAT_CU_FLAGS_IOMMU_PRESENT; crat_table->length += sub_type_hdr->length; @@ -2216,12 +2239,12 @@ static int kfd_create_vcrat_image_gpu(void *pcrat_image, * (from other GPU to this GPU) will be added * in kfd_parse_subtype_iolink. */ - if (kdev->hive_id) { + if (kdev->kfd->hive_id) { for (nid = 0; nid < proximity_domain; ++nid) { peer_dev = kfd_topology_device_by_proximity_domain_no_lock(nid); if (!peer_dev->gpu) continue; - if (peer_dev->gpu->hive_id != kdev->hive_id) + if (peer_dev->gpu->kfd->hive_id != kdev->kfd->hive_id) continue; sub_type_hdr = (typeof(sub_type_hdr))( (char *)sub_type_hdr + @@ -2255,12 +2278,12 @@ static int kfd_create_vcrat_image_gpu(void *pcrat_image, * (COMPUTE_UNIT_CPU | COMPUTE_UNIT_GPU) - Create VCRAT for APU * -- this option is not currently implemented. * The assumption is that all AMD APUs will have CRAT - * @kdev: Valid kfd_device required if flags contain COMPUTE_UNIT_GPU + * @kdev: Valid kfd_node required if flags contain COMPUTE_UNIT_GPU * * Return 0 if successful else return -ve value */ int kfd_create_crat_image_virtual(void **crat_image, size_t *size, - int flags, struct kfd_dev *kdev, + int flags, struct kfd_node *kdev, uint32_t proximity_domain) { void *pcrat_image = NULL; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.h b/drivers/gpu/drm/amd/amdkfd/kfd_crat.h index 8d1e8ba58dee..fc719389b5d6 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.h @@ -275,7 +275,7 @@ struct crat_subtype_iolink { uint32_t maximum_bandwidth_mbs; uint32_t recommended_transfer_size; uint8_t reserved2[CRAT_IOLINK_RESERVED_LENGTH - 1]; - uint8_t num_hops_xgmi; + uint8_t weight_xgmi; }; /* @@ -293,7 +293,7 @@ struct crat_subtype_generic { #pragma pack() -struct kfd_dev; +struct kfd_node; /* Static table to describe GPU Cache information */ struct kfd_gpu_cache_info { @@ -305,14 +305,14 @@ struct kfd_gpu_cache_info { */ uint32_t num_cu_shared; }; -int kfd_get_gpu_cache_info(struct kfd_dev *kdev, struct kfd_gpu_cache_info **pcache_info); +int kfd_get_gpu_cache_info(struct kfd_node *kdev, struct kfd_gpu_cache_info **pcache_info); int kfd_create_crat_image_acpi(void **crat_image, size_t *size); void kfd_destroy_crat_image(void *crat_image); int kfd_parse_crat_table(void *crat_image, struct list_head *device_list, uint32_t proximity_domain); int kfd_create_crat_image_virtual(void **crat_image, size_t *size, - int flags, struct kfd_dev *kdev, + int flags, struct kfd_node *kdev, uint32_t proximity_domain); #endif /* KFD_CRAT_H_INCLUDED */ diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_debug.c b/drivers/gpu/drm/amd/amdkfd/kfd_debug.c new file mode 100644 index 000000000000..fff3ccc04fa9 --- /dev/null +++ b/drivers/gpu/drm/amd/amdkfd/kfd_debug.c @@ -0,0 +1,1118 @@ +/* + * Copyright 2023 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ + +#include "kfd_debug.h" +#include "kfd_device_queue_manager.h" +#include "kfd_topology.h" +#include <linux/file.h> +#include <uapi/linux/kfd_ioctl.h> + +#define MAX_WATCH_ADDRESSES 4 + +int kfd_dbg_ev_query_debug_event(struct kfd_process *process, + unsigned int *queue_id, + unsigned int *gpu_id, + uint64_t exception_clear_mask, + uint64_t *event_status) +{ + struct process_queue_manager *pqm; + struct process_queue_node *pqn; + int i; + + if (!(process && process->debug_trap_enabled)) + return -ENODATA; + + mutex_lock(&process->event_mutex); + *event_status = 0; + *queue_id = 0; + *gpu_id = 0; + + /* find and report queue events */ + pqm = &process->pqm; + list_for_each_entry(pqn, &pqm->queues, process_queue_list) { + uint64_t tmp = process->exception_enable_mask; + + if (!pqn->q) + continue; + + tmp &= pqn->q->properties.exception_status; + + if (!tmp) + continue; + + *event_status = pqn->q->properties.exception_status; + *queue_id = pqn->q->properties.queue_id; + *gpu_id = pqn->q->device->id; + pqn->q->properties.exception_status &= ~exception_clear_mask; + goto out; + } + + /* find and report device events */ + for (i = 0; i < process->n_pdds; i++) { + struct kfd_process_device *pdd = process->pdds[i]; + uint64_t tmp = process->exception_enable_mask + & pdd->exception_status; + + if (!tmp) + continue; + + *event_status = pdd->exception_status; + *gpu_id = pdd->dev->id; + pdd->exception_status &= ~exception_clear_mask; + goto out; + } + + /* report process events */ + if (process->exception_enable_mask & process->exception_status) { + *event_status = process->exception_status; + process->exception_status &= ~exception_clear_mask; + } + +out: + mutex_unlock(&process->event_mutex); + return *event_status ? 0 : -EAGAIN; +} + +void debug_event_write_work_handler(struct work_struct *work) +{ + struct kfd_process *process; + + static const char write_data = '.'; + loff_t pos = 0; + + process = container_of(work, + struct kfd_process, + debug_event_workarea); + + kernel_write(process->dbg_ev_file, &write_data, 1, &pos); +} + +/* update process/device/queue exception status, write to descriptor + * only if exception_status is enabled. + */ +bool kfd_dbg_ev_raise(uint64_t event_mask, + struct kfd_process *process, struct kfd_node *dev, + unsigned int source_id, bool use_worker, + void *exception_data, size_t exception_data_size) +{ + struct process_queue_manager *pqm; + struct process_queue_node *pqn; + int i; + static const char write_data = '.'; + loff_t pos = 0; + bool is_subscribed = true; + + if (!(process && process->debug_trap_enabled)) + return false; + + mutex_lock(&process->event_mutex); + + if (event_mask & KFD_EC_MASK_DEVICE) { + for (i = 0; i < process->n_pdds; i++) { + struct kfd_process_device *pdd = process->pdds[i]; + + if (pdd->dev != dev) + continue; + + pdd->exception_status |= event_mask & KFD_EC_MASK_DEVICE; + + if (event_mask & KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION)) { + if (!pdd->vm_fault_exc_data) { + pdd->vm_fault_exc_data = kmemdup( + exception_data, + exception_data_size, + GFP_KERNEL); + if (!pdd->vm_fault_exc_data) + pr_debug("Failed to allocate exception data memory"); + } else { + pr_debug("Debugger exception data not saved\n"); + print_hex_dump_bytes("exception data: ", + DUMP_PREFIX_OFFSET, + exception_data, + exception_data_size); + } + } + break; + } + } else if (event_mask & KFD_EC_MASK_PROCESS) { + process->exception_status |= event_mask & KFD_EC_MASK_PROCESS; + } else { + pqm = &process->pqm; + list_for_each_entry(pqn, &pqm->queues, + process_queue_list) { + int target_id; + + if (!pqn->q) + continue; + + target_id = event_mask & KFD_EC_MASK(EC_QUEUE_NEW) ? + pqn->q->properties.queue_id : + pqn->q->doorbell_id; + + if (pqn->q->device != dev || target_id != source_id) + continue; + + pqn->q->properties.exception_status |= event_mask; + break; + } + } + + if (process->exception_enable_mask & event_mask) { + if (use_worker) + schedule_work(&process->debug_event_workarea); + else + kernel_write(process->dbg_ev_file, + &write_data, + 1, + &pos); + } else { + is_subscribed = false; + } + + mutex_unlock(&process->event_mutex); + + return is_subscribed; +} + +/* set pending event queue entry from ring entry */ +bool kfd_set_dbg_ev_from_interrupt(struct kfd_node *dev, + unsigned int pasid, + uint32_t doorbell_id, + uint64_t trap_mask, + void *exception_data, + size_t exception_data_size) +{ + struct kfd_process *p; + bool signaled_to_debugger_or_runtime = false; + + p = kfd_lookup_process_by_pasid(pasid); + + if (!p) + return false; + + if (!kfd_dbg_ev_raise(trap_mask, p, dev, doorbell_id, true, + exception_data, exception_data_size)) { + struct process_queue_manager *pqm; + struct process_queue_node *pqn; + + if (!!(trap_mask & KFD_EC_MASK_QUEUE) && + p->runtime_info.runtime_state == DEBUG_RUNTIME_STATE_ENABLED) { + mutex_lock(&p->mutex); + + pqm = &p->pqm; + list_for_each_entry(pqn, &pqm->queues, + process_queue_list) { + + if (!(pqn->q && pqn->q->device == dev && + pqn->q->doorbell_id == doorbell_id)) + continue; + + kfd_send_exception_to_runtime(p, pqn->q->properties.queue_id, + trap_mask); + + signaled_to_debugger_or_runtime = true; + + break; + } + + mutex_unlock(&p->mutex); + } else if (trap_mask & KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION)) { + kfd_dqm_evict_pasid(dev->dqm, p->pasid); + kfd_signal_vm_fault_event(dev, p->pasid, NULL, + exception_data); + + signaled_to_debugger_or_runtime = true; + } + } else { + signaled_to_debugger_or_runtime = true; + } + + kfd_unref_process(p); + + return signaled_to_debugger_or_runtime; +} + +int kfd_dbg_send_exception_to_runtime(struct kfd_process *p, + unsigned int dev_id, + unsigned int queue_id, + uint64_t error_reason) +{ + if (error_reason & KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION)) { + struct kfd_process_device *pdd = NULL; + struct kfd_hsa_memory_exception_data *data; + int i; + + for (i = 0; i < p->n_pdds; i++) { + if (p->pdds[i]->dev->id == dev_id) { + pdd = p->pdds[i]; + break; + } + } + + if (!pdd) + return -ENODEV; + + data = (struct kfd_hsa_memory_exception_data *) + pdd->vm_fault_exc_data; + + kfd_dqm_evict_pasid(pdd->dev->dqm, p->pasid); + kfd_signal_vm_fault_event(pdd->dev, p->pasid, NULL, data); + error_reason &= ~KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION); + } + + if (error_reason & (KFD_EC_MASK(EC_PROCESS_RUNTIME))) { + /* + * block should only happen after the debugger receives runtime + * enable notice. + */ + up(&p->runtime_enable_sema); + error_reason &= ~KFD_EC_MASK(EC_PROCESS_RUNTIME); + } + + if (error_reason) + return kfd_send_exception_to_runtime(p, queue_id, error_reason); + + return 0; +} + +static int kfd_dbg_set_queue_workaround(struct queue *q, bool enable) +{ + struct mqd_update_info minfo = {0}; + int err; + + if (!q) + return 0; + + if (KFD_GC_VERSION(q->device) < IP_VERSION(11, 0, 0) || + KFD_GC_VERSION(q->device) >= IP_VERSION(12, 0, 0)) + return 0; + + if (enable && q->properties.is_user_cu_masked) + return -EBUSY; + + minfo.update_flag = enable ? UPDATE_FLAG_DBG_WA_ENABLE : UPDATE_FLAG_DBG_WA_DISABLE; + + q->properties.is_dbg_wa = enable; + err = q->device->dqm->ops.update_queue(q->device->dqm, q, &minfo); + if (err) + q->properties.is_dbg_wa = false; + + return err; +} + +static int kfd_dbg_set_workaround(struct kfd_process *target, bool enable) +{ + struct process_queue_manager *pqm = &target->pqm; + struct process_queue_node *pqn; + int r = 0; + + list_for_each_entry(pqn, &pqm->queues, process_queue_list) { + r = kfd_dbg_set_queue_workaround(pqn->q, enable); + if (enable && r) + goto unwind; + } + + return 0; + +unwind: + list_for_each_entry(pqn, &pqm->queues, process_queue_list) + kfd_dbg_set_queue_workaround(pqn->q, false); + + if (enable) + target->runtime_info.runtime_state = r == -EBUSY ? + DEBUG_RUNTIME_STATE_ENABLED_BUSY : + DEBUG_RUNTIME_STATE_ENABLED_ERROR; + + return r; +} + +int kfd_dbg_set_mes_debug_mode(struct kfd_process_device *pdd) +{ + uint32_t spi_dbg_cntl = pdd->spi_dbg_override | pdd->spi_dbg_launch_mode; + uint32_t flags = pdd->process->dbg_flags; + bool sq_trap_en = !!spi_dbg_cntl; + + if (!kfd_dbg_is_per_vmid_supported(pdd->dev)) + return 0; + + return amdgpu_mes_set_shader_debugger(pdd->dev->adev, pdd->proc_ctx_gpu_addr, spi_dbg_cntl, + pdd->watch_points, flags, sq_trap_en); +} + +#define KFD_DEBUGGER_INVALID_WATCH_POINT_ID -1 +static int kfd_dbg_get_dev_watch_id(struct kfd_process_device *pdd, int *watch_id) +{ + int i; + + *watch_id = KFD_DEBUGGER_INVALID_WATCH_POINT_ID; + + spin_lock(&pdd->dev->kfd->watch_points_lock); + + for (i = 0; i < MAX_WATCH_ADDRESSES; i++) { + /* device watchpoint in use so skip */ + if ((pdd->dev->kfd->alloc_watch_ids >> i) & 0x1) + continue; + + pdd->alloc_watch_ids |= 0x1 << i; + pdd->dev->kfd->alloc_watch_ids |= 0x1 << i; + *watch_id = i; + spin_unlock(&pdd->dev->kfd->watch_points_lock); + return 0; + } + + spin_unlock(&pdd->dev->kfd->watch_points_lock); + + return -ENOMEM; +} + +static void kfd_dbg_clear_dev_watch_id(struct kfd_process_device *pdd, int watch_id) +{ + spin_lock(&pdd->dev->kfd->watch_points_lock); + + /* process owns device watch point so safe to clear */ + if ((pdd->alloc_watch_ids >> watch_id) & 0x1) { + pdd->alloc_watch_ids &= ~(0x1 << watch_id); + pdd->dev->kfd->alloc_watch_ids &= ~(0x1 << watch_id); + } + + spin_unlock(&pdd->dev->kfd->watch_points_lock); +} + +static bool kfd_dbg_owns_dev_watch_id(struct kfd_process_device *pdd, int watch_id) +{ + bool owns_watch_id = false; + + spin_lock(&pdd->dev->kfd->watch_points_lock); + owns_watch_id = watch_id < MAX_WATCH_ADDRESSES && + ((pdd->alloc_watch_ids >> watch_id) & 0x1); + + spin_unlock(&pdd->dev->kfd->watch_points_lock); + + return owns_watch_id; +} + +int kfd_dbg_trap_clear_dev_address_watch(struct kfd_process_device *pdd, + uint32_t watch_id) +{ + int r; + + if (!kfd_dbg_owns_dev_watch_id(pdd, watch_id)) + return -EINVAL; + + if (!pdd->dev->kfd->shared_resources.enable_mes) { + r = debug_lock_and_unmap(pdd->dev->dqm); + if (r) + return r; + } + + amdgpu_gfx_off_ctrl(pdd->dev->adev, false); + pdd->watch_points[watch_id] = pdd->dev->kfd2kgd->clear_address_watch( + pdd->dev->adev, + watch_id); + amdgpu_gfx_off_ctrl(pdd->dev->adev, true); + + if (!pdd->dev->kfd->shared_resources.enable_mes) + r = debug_map_and_unlock(pdd->dev->dqm); + else + r = kfd_dbg_set_mes_debug_mode(pdd); + + kfd_dbg_clear_dev_watch_id(pdd, watch_id); + + return r; +} + +int kfd_dbg_trap_set_dev_address_watch(struct kfd_process_device *pdd, + uint64_t watch_address, + uint32_t watch_address_mask, + uint32_t *watch_id, + uint32_t watch_mode) +{ + int r = kfd_dbg_get_dev_watch_id(pdd, watch_id); + + if (r) + return r; + + if (!pdd->dev->kfd->shared_resources.enable_mes) { + r = debug_lock_and_unmap(pdd->dev->dqm); + if (r) { + kfd_dbg_clear_dev_watch_id(pdd, *watch_id); + return r; + } + } + + amdgpu_gfx_off_ctrl(pdd->dev->adev, false); + pdd->watch_points[*watch_id] = pdd->dev->kfd2kgd->set_address_watch( + pdd->dev->adev, + watch_address, + watch_address_mask, + *watch_id, + watch_mode, + pdd->dev->vm_info.last_vmid_kfd); + amdgpu_gfx_off_ctrl(pdd->dev->adev, true); + + if (!pdd->dev->kfd->shared_resources.enable_mes) + r = debug_map_and_unlock(pdd->dev->dqm); + else + r = kfd_dbg_set_mes_debug_mode(pdd); + + /* HWS is broken so no point in HW rollback but release the watchpoint anyways */ + if (r) + kfd_dbg_clear_dev_watch_id(pdd, *watch_id); + + return 0; +} + +static void kfd_dbg_clear_process_address_watch(struct kfd_process *target) +{ + int i, j; + + for (i = 0; i < target->n_pdds; i++) + for (j = 0; j < MAX_WATCH_ADDRESSES; j++) + kfd_dbg_trap_clear_dev_address_watch(target->pdds[i], j); +} + +int kfd_dbg_trap_set_flags(struct kfd_process *target, uint32_t *flags) +{ + uint32_t prev_flags = target->dbg_flags; + int i, r = 0, rewind_count = 0; + + for (i = 0; i < target->n_pdds; i++) { + if (!kfd_dbg_is_per_vmid_supported(target->pdds[i]->dev) && + (*flags & KFD_DBG_TRAP_FLAG_SINGLE_MEM_OP)) { + *flags = prev_flags; + return -EACCES; + } + } + + target->dbg_flags = *flags & KFD_DBG_TRAP_FLAG_SINGLE_MEM_OP; + *flags = prev_flags; + for (i = 0; i < target->n_pdds; i++) { + struct kfd_process_device *pdd = target->pdds[i]; + + if (!kfd_dbg_is_per_vmid_supported(pdd->dev)) + continue; + + if (!pdd->dev->kfd->shared_resources.enable_mes) + r = debug_refresh_runlist(pdd->dev->dqm); + else + r = kfd_dbg_set_mes_debug_mode(pdd); + + if (r) { + target->dbg_flags = prev_flags; + break; + } + + rewind_count++; + } + + /* Rewind flags */ + if (r) { + target->dbg_flags = prev_flags; + + for (i = 0; i < rewind_count; i++) { + struct kfd_process_device *pdd = target->pdds[i]; + + if (!kfd_dbg_is_per_vmid_supported(pdd->dev)) + continue; + + if (!pdd->dev->kfd->shared_resources.enable_mes) + debug_refresh_runlist(pdd->dev->dqm); + else + kfd_dbg_set_mes_debug_mode(pdd); + } + } + + return r; +} + +/* kfd_dbg_trap_deactivate: + * target: target process + * unwind: If this is unwinding a failed kfd_dbg_trap_enable() + * unwind_count: + * If unwind == true, how far down the pdd list we need + * to unwind + * else: ignored + */ +void kfd_dbg_trap_deactivate(struct kfd_process *target, bool unwind, int unwind_count) +{ + int i; + + if (!unwind) { + uint32_t flags = 0; + int resume_count = resume_queues(target, 0, NULL); + + if (resume_count) + pr_debug("Resumed %d queues\n", resume_count); + + cancel_work_sync(&target->debug_event_workarea); + kfd_dbg_clear_process_address_watch(target); + kfd_dbg_trap_set_wave_launch_mode(target, 0); + + kfd_dbg_trap_set_flags(target, &flags); + } + + for (i = 0; i < target->n_pdds; i++) { + struct kfd_process_device *pdd = target->pdds[i]; + + /* If this is an unwind, and we have unwound the required + * enable calls on the pdd list, we need to stop now + * otherwise we may mess up another debugger session. + */ + if (unwind && i == unwind_count) + break; + + kfd_process_set_trap_debug_flag(&pdd->qpd, false); + + /* GFX off is already disabled by debug activate if not RLC restore supported. */ + if (kfd_dbg_is_rlc_restore_supported(pdd->dev)) + amdgpu_gfx_off_ctrl(pdd->dev->adev, false); + pdd->spi_dbg_override = + pdd->dev->kfd2kgd->disable_debug_trap( + pdd->dev->adev, + target->runtime_info.ttmp_setup, + pdd->dev->vm_info.last_vmid_kfd); + amdgpu_gfx_off_ctrl(pdd->dev->adev, true); + + if (!kfd_dbg_is_per_vmid_supported(pdd->dev) && + release_debug_trap_vmid(pdd->dev->dqm, &pdd->qpd)) + pr_err("Failed to release debug vmid on [%i]\n", pdd->dev->id); + + if (!pdd->dev->kfd->shared_resources.enable_mes) + debug_refresh_runlist(pdd->dev->dqm); + else + kfd_dbg_set_mes_debug_mode(pdd); + } + + kfd_dbg_set_workaround(target, false); +} + +static void kfd_dbg_clean_exception_status(struct kfd_process *target) +{ + struct process_queue_manager *pqm; + struct process_queue_node *pqn; + int i; + + for (i = 0; i < target->n_pdds; i++) { + struct kfd_process_device *pdd = target->pdds[i]; + + kfd_process_drain_interrupts(pdd); + + pdd->exception_status = 0; + } + + pqm = &target->pqm; + list_for_each_entry(pqn, &pqm->queues, process_queue_list) { + if (!pqn->q) + continue; + + pqn->q->properties.exception_status = 0; + } + + target->exception_status = 0; +} + +int kfd_dbg_trap_disable(struct kfd_process *target) +{ + if (!target->debug_trap_enabled) + return 0; + + /* + * Defer deactivation to runtime if runtime not enabled otherwise reset + * attached running target runtime state to enable for re-attach. + */ + if (target->runtime_info.runtime_state == DEBUG_RUNTIME_STATE_ENABLED) + kfd_dbg_trap_deactivate(target, false, 0); + else if (target->runtime_info.runtime_state != DEBUG_RUNTIME_STATE_DISABLED) + target->runtime_info.runtime_state = DEBUG_RUNTIME_STATE_ENABLED; + + fput(target->dbg_ev_file); + target->dbg_ev_file = NULL; + + if (target->debugger_process) { + atomic_dec(&target->debugger_process->debugged_process_count); + target->debugger_process = NULL; + } + + target->debug_trap_enabled = false; + kfd_dbg_clean_exception_status(target); + kfd_unref_process(target); + + return 0; +} + +int kfd_dbg_trap_activate(struct kfd_process *target) +{ + int i, r = 0; + + r = kfd_dbg_set_workaround(target, true); + if (r) + return r; + + for (i = 0; i < target->n_pdds; i++) { + struct kfd_process_device *pdd = target->pdds[i]; + + if (!kfd_dbg_is_per_vmid_supported(pdd->dev)) { + r = reserve_debug_trap_vmid(pdd->dev->dqm, &pdd->qpd); + + if (r) { + target->runtime_info.runtime_state = (r == -EBUSY) ? + DEBUG_RUNTIME_STATE_ENABLED_BUSY : + DEBUG_RUNTIME_STATE_ENABLED_ERROR; + + goto unwind_err; + } + } + + /* Disable GFX OFF to prevent garbage read/writes to debug registers. + * If RLC restore of debug registers is not supported and runtime enable + * hasn't done so already on ttmp setup request, restore the trap config registers. + * + * If RLC restore of debug registers is not supported, keep gfx off disabled for + * the debug session. + */ + amdgpu_gfx_off_ctrl(pdd->dev->adev, false); + if (!(kfd_dbg_is_rlc_restore_supported(pdd->dev) || + target->runtime_info.ttmp_setup)) + pdd->dev->kfd2kgd->enable_debug_trap(pdd->dev->adev, true, + pdd->dev->vm_info.last_vmid_kfd); + + pdd->spi_dbg_override = pdd->dev->kfd2kgd->enable_debug_trap( + pdd->dev->adev, + false, + pdd->dev->vm_info.last_vmid_kfd); + + if (kfd_dbg_is_rlc_restore_supported(pdd->dev)) + amdgpu_gfx_off_ctrl(pdd->dev->adev, true); + + /* + * Setting the debug flag in the trap handler requires that the TMA has been + * allocated, which occurs during CWSR initialization. + * In the event that CWSR has not been initialized at this point, setting the + * flag will be called again during CWSR initialization if the target process + * is still debug enabled. + */ + kfd_process_set_trap_debug_flag(&pdd->qpd, true); + + if (!pdd->dev->kfd->shared_resources.enable_mes) + r = debug_refresh_runlist(pdd->dev->dqm); + else + r = kfd_dbg_set_mes_debug_mode(pdd); + + if (r) { + target->runtime_info.runtime_state = + DEBUG_RUNTIME_STATE_ENABLED_ERROR; + goto unwind_err; + } + } + + return 0; + +unwind_err: + /* Enabling debug failed, we need to disable on + * all GPUs so the enable is all or nothing. + */ + kfd_dbg_trap_deactivate(target, true, i); + return r; +} + +int kfd_dbg_trap_enable(struct kfd_process *target, uint32_t fd, + void __user *runtime_info, uint32_t *runtime_size) +{ + struct file *f; + uint32_t copy_size; + int i, r = 0; + + if (target->debug_trap_enabled) + return -EALREADY; + + /* Enable pre-checks */ + for (i = 0; i < target->n_pdds; i++) { + struct kfd_process_device *pdd = target->pdds[i]; + + if (!KFD_IS_SOC15(pdd->dev)) + return -ENODEV; + + if (!kfd_dbg_has_gws_support(pdd->dev) && pdd->qpd.num_gws) + return -EBUSY; + } + + copy_size = min((size_t)(*runtime_size), sizeof(target->runtime_info)); + + f = fget(fd); + if (!f) { + pr_err("Failed to get file for (%i)\n", fd); + return -EBADF; + } + + target->dbg_ev_file = f; + + /* defer activation to runtime if not runtime enabled */ + if (target->runtime_info.runtime_state == DEBUG_RUNTIME_STATE_ENABLED) + kfd_dbg_trap_activate(target); + + /* We already hold the process reference but hold another one for the + * debug session. + */ + kref_get(&target->ref); + target->debug_trap_enabled = true; + + if (target->debugger_process) + atomic_inc(&target->debugger_process->debugged_process_count); + + if (copy_to_user(runtime_info, (void *)&target->runtime_info, copy_size)) { + kfd_dbg_trap_deactivate(target, false, 0); + r = -EFAULT; + } + + *runtime_size = sizeof(target->runtime_info); + + return r; +} + +static int kfd_dbg_validate_trap_override_request(struct kfd_process *p, + uint32_t trap_override, + uint32_t trap_mask_request, + uint32_t *trap_mask_supported) +{ + int i = 0; + + *trap_mask_supported = 0xffffffff; + + for (i = 0; i < p->n_pdds; i++) { + struct kfd_process_device *pdd = p->pdds[i]; + int err = pdd->dev->kfd2kgd->validate_trap_override_request( + pdd->dev->adev, + trap_override, + trap_mask_supported); + + if (err) + return err; + } + + if (trap_mask_request & ~*trap_mask_supported) + return -EACCES; + + return 0; +} + +int kfd_dbg_trap_set_wave_launch_override(struct kfd_process *target, + uint32_t trap_override, + uint32_t trap_mask_bits, + uint32_t trap_mask_request, + uint32_t *trap_mask_prev, + uint32_t *trap_mask_supported) +{ + int r = 0, i; + + r = kfd_dbg_validate_trap_override_request(target, + trap_override, + trap_mask_request, + trap_mask_supported); + + if (r) + return r; + + for (i = 0; i < target->n_pdds; i++) { + struct kfd_process_device *pdd = target->pdds[i]; + + amdgpu_gfx_off_ctrl(pdd->dev->adev, false); + pdd->spi_dbg_override = pdd->dev->kfd2kgd->set_wave_launch_trap_override( + pdd->dev->adev, + pdd->dev->vm_info.last_vmid_kfd, + trap_override, + trap_mask_bits, + trap_mask_request, + trap_mask_prev, + pdd->spi_dbg_override); + amdgpu_gfx_off_ctrl(pdd->dev->adev, true); + + if (!pdd->dev->kfd->shared_resources.enable_mes) + r = debug_refresh_runlist(pdd->dev->dqm); + else + r = kfd_dbg_set_mes_debug_mode(pdd); + + if (r) + break; + } + + return r; +} + +int kfd_dbg_trap_set_wave_launch_mode(struct kfd_process *target, + uint8_t wave_launch_mode) +{ + int r = 0, i; + + if (wave_launch_mode != KFD_DBG_TRAP_WAVE_LAUNCH_MODE_NORMAL && + wave_launch_mode != KFD_DBG_TRAP_WAVE_LAUNCH_MODE_HALT && + wave_launch_mode != KFD_DBG_TRAP_WAVE_LAUNCH_MODE_DEBUG) + return -EINVAL; + + for (i = 0; i < target->n_pdds; i++) { + struct kfd_process_device *pdd = target->pdds[i]; + + amdgpu_gfx_off_ctrl(pdd->dev->adev, false); + pdd->spi_dbg_launch_mode = pdd->dev->kfd2kgd->set_wave_launch_mode( + pdd->dev->adev, + wave_launch_mode, + pdd->dev->vm_info.last_vmid_kfd); + amdgpu_gfx_off_ctrl(pdd->dev->adev, true); + + if (!pdd->dev->kfd->shared_resources.enable_mes) + r = debug_refresh_runlist(pdd->dev->dqm); + else + r = kfd_dbg_set_mes_debug_mode(pdd); + + if (r) + break; + } + + return r; +} + +int kfd_dbg_trap_query_exception_info(struct kfd_process *target, + uint32_t source_id, + uint32_t exception_code, + bool clear_exception, + void __user *info, + uint32_t *info_size) +{ + bool found = false; + int r = 0; + uint32_t copy_size, actual_info_size = 0; + uint64_t *exception_status_ptr = NULL; + + if (!target) + return -EINVAL; + + if (!info || !info_size) + return -EINVAL; + + mutex_lock(&target->event_mutex); + + if (KFD_DBG_EC_TYPE_IS_QUEUE(exception_code)) { + /* Per queue exceptions */ + struct queue *queue = NULL; + int i; + + for (i = 0; i < target->n_pdds; i++) { + struct kfd_process_device *pdd = target->pdds[i]; + struct qcm_process_device *qpd = &pdd->qpd; + + list_for_each_entry(queue, &qpd->queues_list, list) { + if (!found && queue->properties.queue_id == source_id) { + found = true; + break; + } + } + if (found) + break; + } + + if (!found) { + r = -EINVAL; + goto out; + } + + if (!(queue->properties.exception_status & KFD_EC_MASK(exception_code))) { + r = -ENODATA; + goto out; + } + exception_status_ptr = &queue->properties.exception_status; + } else if (KFD_DBG_EC_TYPE_IS_DEVICE(exception_code)) { + /* Per device exceptions */ + struct kfd_process_device *pdd = NULL; + int i; + + for (i = 0; i < target->n_pdds; i++) { + pdd = target->pdds[i]; + if (pdd->dev->id == source_id) { + found = true; + break; + } + } + + if (!found) { + r = -EINVAL; + goto out; + } + + if (!(pdd->exception_status & KFD_EC_MASK(exception_code))) { + r = -ENODATA; + goto out; + } + + if (exception_code == EC_DEVICE_MEMORY_VIOLATION) { + copy_size = min((size_t)(*info_size), pdd->vm_fault_exc_data_size); + + if (copy_to_user(info, pdd->vm_fault_exc_data, copy_size)) { + r = -EFAULT; + goto out; + } + actual_info_size = pdd->vm_fault_exc_data_size; + if (clear_exception) { + kfree(pdd->vm_fault_exc_data); + pdd->vm_fault_exc_data = NULL; + pdd->vm_fault_exc_data_size = 0; + } + } + exception_status_ptr = &pdd->exception_status; + } else if (KFD_DBG_EC_TYPE_IS_PROCESS(exception_code)) { + /* Per process exceptions */ + if (!(target->exception_status & KFD_EC_MASK(exception_code))) { + r = -ENODATA; + goto out; + } + + if (exception_code == EC_PROCESS_RUNTIME) { + copy_size = min((size_t)(*info_size), sizeof(target->runtime_info)); + + if (copy_to_user(info, (void *)&target->runtime_info, copy_size)) { + r = -EFAULT; + goto out; + } + + actual_info_size = sizeof(target->runtime_info); + } + + exception_status_ptr = &target->exception_status; + } else { + pr_debug("Bad exception type [%i]\n", exception_code); + r = -EINVAL; + goto out; + } + + *info_size = actual_info_size; + if (clear_exception) + *exception_status_ptr &= ~KFD_EC_MASK(exception_code); +out: + mutex_unlock(&target->event_mutex); + return r; +} + +int kfd_dbg_trap_device_snapshot(struct kfd_process *target, + uint64_t exception_clear_mask, + void __user *user_info, + uint32_t *number_of_device_infos, + uint32_t *entry_size) +{ + struct kfd_dbg_device_info_entry device_info; + uint32_t tmp_entry_size = *entry_size, tmp_num_devices; + int i, r = 0; + + if (!(target && user_info && number_of_device_infos && entry_size)) + return -EINVAL; + + tmp_num_devices = min_t(size_t, *number_of_device_infos, target->n_pdds); + *number_of_device_infos = target->n_pdds; + *entry_size = min_t(size_t, *entry_size, sizeof(device_info)); + + if (!tmp_num_devices) + return 0; + + memset(&device_info, 0, sizeof(device_info)); + + mutex_lock(&target->event_mutex); + + /* Run over all pdd of the process */ + for (i = 0; i < tmp_num_devices; i++) { + struct kfd_process_device *pdd = target->pdds[i]; + struct kfd_topology_device *topo_dev = kfd_topology_device_by_id(pdd->dev->id); + + device_info.gpu_id = pdd->dev->id; + device_info.exception_status = pdd->exception_status; + device_info.lds_base = pdd->lds_base; + device_info.lds_limit = pdd->lds_limit; + device_info.scratch_base = pdd->scratch_base; + device_info.scratch_limit = pdd->scratch_limit; + device_info.gpuvm_base = pdd->gpuvm_base; + device_info.gpuvm_limit = pdd->gpuvm_limit; + device_info.location_id = topo_dev->node_props.location_id; + device_info.vendor_id = topo_dev->node_props.vendor_id; + device_info.device_id = topo_dev->node_props.device_id; + device_info.revision_id = pdd->dev->adev->pdev->revision; + device_info.subsystem_vendor_id = pdd->dev->adev->pdev->subsystem_vendor; + device_info.subsystem_device_id = pdd->dev->adev->pdev->subsystem_device; + device_info.fw_version = pdd->dev->kfd->mec_fw_version; + device_info.gfx_target_version = + topo_dev->node_props.gfx_target_version; + device_info.simd_count = topo_dev->node_props.simd_count; + device_info.max_waves_per_simd = + topo_dev->node_props.max_waves_per_simd; + device_info.array_count = topo_dev->node_props.array_count; + device_info.simd_arrays_per_engine = + topo_dev->node_props.simd_arrays_per_engine; + device_info.num_xcc = NUM_XCC(pdd->dev->xcc_mask); + device_info.capability = topo_dev->node_props.capability; + device_info.debug_prop = topo_dev->node_props.debug_prop; + + if (exception_clear_mask) + pdd->exception_status &= ~exception_clear_mask; + + if (copy_to_user(user_info, &device_info, *entry_size)) { + r = -EFAULT; + break; + } + + user_info += tmp_entry_size; + } + + mutex_unlock(&target->event_mutex); + + return r; +} + +void kfd_dbg_set_enabled_debug_exception_mask(struct kfd_process *target, + uint64_t exception_set_mask) +{ + uint64_t found_mask = 0; + struct process_queue_manager *pqm; + struct process_queue_node *pqn; + static const char write_data = '.'; + loff_t pos = 0; + int i; + + mutex_lock(&target->event_mutex); + + found_mask |= target->exception_status; + + pqm = &target->pqm; + list_for_each_entry(pqn, &pqm->queues, process_queue_list) { + if (!pqn->q) + continue; + + found_mask |= pqn->q->properties.exception_status; + } + + for (i = 0; i < target->n_pdds; i++) { + struct kfd_process_device *pdd = target->pdds[i]; + + found_mask |= pdd->exception_status; + } + + if (exception_set_mask & found_mask) + kernel_write(target->dbg_ev_file, &write_data, 1, &pos); + + target->exception_enable_mask = exception_set_mask; + + mutex_unlock(&target->event_mutex); +} diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_debug.h b/drivers/gpu/drm/amd/amdkfd/kfd_debug.h new file mode 100644 index 000000000000..a289e59ceb79 --- /dev/null +++ b/drivers/gpu/drm/amd/amdkfd/kfd_debug.h @@ -0,0 +1,123 @@ +/* + * Copyright 2023 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef KFD_DEBUG_EVENTS_H_INCLUDED +#define KFD_DEBUG_EVENTS_H_INCLUDED + +#include "kfd_priv.h" + +void kfd_dbg_trap_deactivate(struct kfd_process *target, bool unwind, int unwind_count); +int kfd_dbg_trap_activate(struct kfd_process *target); +int kfd_dbg_ev_query_debug_event(struct kfd_process *process, + unsigned int *queue_id, + unsigned int *gpu_id, + uint64_t exception_clear_mask, + uint64_t *event_status); +bool kfd_set_dbg_ev_from_interrupt(struct kfd_node *dev, + unsigned int pasid, + uint32_t doorbell_id, + uint64_t trap_mask, + void *exception_data, + size_t exception_data_size); +bool kfd_dbg_ev_raise(uint64_t event_mask, + struct kfd_process *process, struct kfd_node *dev, + unsigned int source_id, bool use_worker, + void *exception_data, + size_t exception_data_size); +int kfd_dbg_trap_disable(struct kfd_process *target); +int kfd_dbg_trap_enable(struct kfd_process *target, uint32_t fd, + void __user *runtime_info, + uint32_t *runtime_info_size); +int kfd_dbg_trap_set_wave_launch_override(struct kfd_process *target, + uint32_t trap_override, + uint32_t trap_mask_bits, + uint32_t trap_mask_request, + uint32_t *trap_mask_prev, + uint32_t *trap_mask_supported); +int kfd_dbg_trap_set_wave_launch_mode(struct kfd_process *target, + uint8_t wave_launch_mode); +int kfd_dbg_trap_clear_dev_address_watch(struct kfd_process_device *pdd, + uint32_t watch_id); +int kfd_dbg_trap_set_dev_address_watch(struct kfd_process_device *pdd, + uint64_t watch_address, + uint32_t watch_address_mask, + uint32_t *watch_id, + uint32_t watch_mode); +int kfd_dbg_trap_set_flags(struct kfd_process *target, uint32_t *flags); +int kfd_dbg_trap_query_exception_info(struct kfd_process *target, + uint32_t source_id, + uint32_t exception_code, + bool clear_exception, + void __user *info, + uint32_t *info_size); +int kfd_dbg_send_exception_to_runtime(struct kfd_process *p, + unsigned int dev_id, + unsigned int queue_id, + uint64_t error_reason); + +static inline bool kfd_dbg_is_per_vmid_supported(struct kfd_node *dev) +{ + return KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 2) || + KFD_GC_VERSION(dev) >= IP_VERSION(11, 0, 0); +} + +void debug_event_write_work_handler(struct work_struct *work); +int kfd_dbg_trap_device_snapshot(struct kfd_process *target, + uint64_t exception_clear_mask, + void __user *user_info, + uint32_t *number_of_device_infos, + uint32_t *entry_size); + +void kfd_dbg_set_enabled_debug_exception_mask(struct kfd_process *target, + uint64_t exception_set_mask); +/* + * If GFX off is enabled, chips that do not support RLC restore for the debug + * registers will disable GFX off temporarily for the entire debug session. + * See disable_on_trap_action_entry and enable_on_trap_action_exit for details. + */ +static inline bool kfd_dbg_is_rlc_restore_supported(struct kfd_node *dev) +{ + return !(KFD_GC_VERSION(dev) == IP_VERSION(10, 1, 10) || + KFD_GC_VERSION(dev) == IP_VERSION(10, 1, 1)); +} + +static inline bool kfd_dbg_has_gws_support(struct kfd_node *dev) +{ + if ((KFD_GC_VERSION(dev) == IP_VERSION(9, 0, 1) + && dev->kfd->mec2_fw_version < 0x81b6) || + (KFD_GC_VERSION(dev) >= IP_VERSION(9, 1, 0) + && KFD_GC_VERSION(dev) <= IP_VERSION(9, 2, 2) + && dev->kfd->mec2_fw_version < 0x1b6) || + (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 0) + && dev->kfd->mec2_fw_version < 0x1b6) || + (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 1) + && dev->kfd->mec2_fw_version < 0x30) || + (KFD_GC_VERSION(dev) >= IP_VERSION(11, 0, 0) && + KFD_GC_VERSION(dev) < IP_VERSION(12, 0, 0))) + return false; + + /* Assume debugging and cooperative launch supported otherwise. */ + return true; +} + +int kfd_dbg_set_mes_debug_mode(struct kfd_process_device *pdd); +#endif diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c b/drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c index ad5a40a685ac..4a5a0a4e00f2 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c @@ -43,7 +43,7 @@ static int kfd_debugfs_hang_hws_read(struct seq_file *m, void *data) static ssize_t kfd_debugfs_hang_hws_write(struct file *file, const char __user *user_buf, size_t size, loff_t *ppos) { - struct kfd_dev *dev; + struct kfd_node *dev; char tmp[16]; uint32_t gpu_id; int ret = -EINVAL; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c index 00f528eb9812..9d4abfd8b55e 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c @@ -32,8 +32,10 @@ #include "kfd_iommu.h" #include "amdgpu_amdkfd.h" #include "kfd_smi_events.h" +#include "kfd_svm.h" #include "kfd_migrate.h" #include "amdgpu.h" +#include "amdgpu_xcp.h" #define MQD_SIZE_ALIGNED 768 @@ -42,7 +44,7 @@ * once locked, kfd driver will stop any further GPU execution. * create process (open) will return -EAGAIN. */ -static atomic_t kfd_locked = ATOMIC_INIT(0); +static int kfd_locked; #ifdef CONFIG_DRM_AMDGPU_CIK extern const struct kfd2kgd_calls gfx_v7_kfd2kgd; @@ -51,6 +53,7 @@ extern const struct kfd2kgd_calls gfx_v8_kfd2kgd; extern const struct kfd2kgd_calls gfx_v9_kfd2kgd; extern const struct kfd2kgd_calls arcturus_kfd2kgd; extern const struct kfd2kgd_calls aldebaran_kfd2kgd; +extern const struct kfd2kgd_calls gc_9_4_3_kfd2kgd; extern const struct kfd2kgd_calls gfx_v10_kfd2kgd; extern const struct kfd2kgd_calls gfx_v10_3_kfd2kgd; extern const struct kfd2kgd_calls gfx_v11_kfd2kgd; @@ -60,7 +63,7 @@ static int kfd_gtt_sa_init(struct kfd_dev *kfd, unsigned int buf_size, static void kfd_gtt_sa_fini(struct kfd_dev *kfd); static int kfd_resume_iommu(struct kfd_dev *kfd); -static int kfd_resume(struct kfd_dev *kfd); +static int kfd_resume(struct kfd_node *kfd); static void kfd_device_info_set_sdma_info(struct kfd_dev *kfd) { @@ -81,6 +84,7 @@ static void kfd_device_info_set_sdma_info(struct kfd_dev *kfd) case IP_VERSION(4, 2, 0):/* VEGA20 */ case IP_VERSION(4, 2, 2):/* ARCTURUS */ case IP_VERSION(4, 4, 0):/* ALDEBARAN */ + case IP_VERSION(4, 4, 2): case IP_VERSION(5, 0, 0):/* NAVI10 */ case IP_VERSION(5, 0, 1):/* CYAN_SKILLFISH */ case IP_VERSION(5, 0, 2):/* NAVI14 */ @@ -102,20 +106,19 @@ static void kfd_device_info_set_sdma_info(struct kfd_dev *kfd) kfd->device_info.num_sdma_queues_per_engine = 8; } + bitmap_zero(kfd->device_info.reserved_sdma_queues_bitmap, KFD_MAX_SDMA_QUEUES); + switch (sdma_version) { case IP_VERSION(6, 0, 0): + case IP_VERSION(6, 0, 1): case IP_VERSION(6, 0, 2): case IP_VERSION(6, 0, 3): /* Reserve 1 for paging and 1 for gfx */ kfd->device_info.num_reserved_sdma_queues_per_engine = 2; /* BIT(0)=engine-0 queue-0; BIT(1)=engine-1 queue-0; BIT(2)=engine-0 queue-1; ... */ - kfd->device_info.reserved_sdma_queues_bitmap = 0xFULL; - break; - case IP_VERSION(6, 0, 1): - /* Reserve 1 for paging and 1 for gfx */ - kfd->device_info.num_reserved_sdma_queues_per_engine = 2; - /* BIT(0)=engine-0 queue-0; BIT(1)=engine-0 queue-1; ... */ - kfd->device_info.reserved_sdma_queues_bitmap = 0x3ULL; + bitmap_set(kfd->device_info.reserved_sdma_queues_bitmap, 0, + kfd->adev->sdma.num_instances * + kfd->device_info.num_reserved_sdma_queues_per_engine); break; default: break; @@ -135,6 +138,9 @@ static void kfd_device_info_set_event_interrupt_class(struct kfd_dev *kfd) case IP_VERSION(9, 4, 0): /* VEGA20 */ case IP_VERSION(9, 4, 1): /* ARCTURUS */ case IP_VERSION(9, 4, 2): /* ALDEBARAN */ + case IP_VERSION(9, 4, 3): /* GC 9.4.3 */ + kfd->device_info.event_interrupt_class = &event_interrupt_class_v9; + break; case IP_VERSION(10, 3, 1): /* VANGOGH */ case IP_VERSION(10, 3, 3): /* YELLOW_CARP */ case IP_VERSION(10, 3, 6): /* GC 10.3.6 */ @@ -148,7 +154,7 @@ static void kfd_device_info_set_event_interrupt_class(struct kfd_dev *kfd) case IP_VERSION(10, 3, 2): /* NAVY_FLOUNDER */ case IP_VERSION(10, 3, 4): /* DIMGREY_CAVEFISH */ case IP_VERSION(10, 3, 5): /* BEIGE_GOBY */ - kfd->device_info.event_interrupt_class = &event_interrupt_class_v9; + kfd->device_info.event_interrupt_class = &event_interrupt_class_v10; break; case IP_VERSION(11, 0, 0): case IP_VERSION(11, 0, 1): @@ -327,8 +333,10 @@ struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf) f2g = &aldebaran_kfd2kgd; break; case IP_VERSION(9, 4, 3): - gfx_target_version = 90400; - f2g = &aldebaran_kfd2kgd; + gfx_target_version = adev->rev_id >= 1 ? 90402 + : adev->flags & AMD_IS_APU ? 90400 + : 90401; + f2g = &gc_9_4_3_kfd2kgd; break; /* Navi10 */ case IP_VERSION(10, 1, 10): @@ -406,8 +414,15 @@ struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf) f2g = &gfx_v11_kfd2kgd; break; case IP_VERSION(11, 0, 3): - /* Note: Compiler version is 11.0.1 while HW version is 11.0.3 */ - gfx_target_version = 110001; + if ((adev->pdev->device == 0x7460 && + adev->pdev->revision == 0x00) || + (adev->pdev->device == 0x7461 && + adev->pdev->revision == 0x00)) + /* Note: Compiler version is 11.0.5 while HW version is 11.0.3 */ + gfx_target_version = 110005; + else + /* Note: Compiler version is 11.0.1 while HW version is 11.0.3 */ + gfx_target_version = 110001; f2g = &gfx_v11_kfd2kgd; break; default: @@ -440,8 +455,6 @@ struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf) memset(&kfd->doorbell_available_index, 0, sizeof(kfd->doorbell_available_index)); - atomic_set(&kfd->sram_ecc_flag, 0); - ida_init(&kfd->doorbell_ida); return kfd; @@ -488,41 +501,112 @@ static void kfd_cwsr_init(struct kfd_dev *kfd) } } -static int kfd_gws_init(struct kfd_dev *kfd) +static int kfd_gws_init(struct kfd_node *node) { int ret = 0; + struct kfd_dev *kfd = node->kfd; - if (kfd->dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) + if (node->dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) return 0; - if (hws_gws_support || (KFD_IS_SOC15(kfd) && - ((KFD_GC_VERSION(kfd) == IP_VERSION(9, 0, 1) + if (hws_gws_support || (KFD_IS_SOC15(node) && + ((KFD_GC_VERSION(node) == IP_VERSION(9, 0, 1) && kfd->mec2_fw_version >= 0x81b3) || - (KFD_GC_VERSION(kfd) <= IP_VERSION(9, 4, 0) + (KFD_GC_VERSION(node) <= IP_VERSION(9, 4, 0) && kfd->mec2_fw_version >= 0x1b3) || - (KFD_GC_VERSION(kfd) == IP_VERSION(9, 4, 1) + (KFD_GC_VERSION(node) == IP_VERSION(9, 4, 1) && kfd->mec2_fw_version >= 0x30) || - (KFD_GC_VERSION(kfd) == IP_VERSION(9, 4, 2) + (KFD_GC_VERSION(node) == IP_VERSION(9, 4, 2) && kfd->mec2_fw_version >= 0x28) || - (KFD_GC_VERSION(kfd) >= IP_VERSION(10, 3, 0) - && KFD_GC_VERSION(kfd) < IP_VERSION(11, 0, 0) + (KFD_GC_VERSION(node) >= IP_VERSION(10, 3, 0) + && KFD_GC_VERSION(node) < IP_VERSION(11, 0, 0) && kfd->mec2_fw_version >= 0x6b)))) - ret = amdgpu_amdkfd_alloc_gws(kfd->adev, - kfd->adev->gds.gws_size, &kfd->gws); + ret = amdgpu_amdkfd_alloc_gws(node->adev, + node->adev->gds.gws_size, &node->gws); return ret; } -static void kfd_smi_init(struct kfd_dev *dev) +static void kfd_smi_init(struct kfd_node *dev) { INIT_LIST_HEAD(&dev->smi_clients); spin_lock_init(&dev->smi_lock); } +static int kfd_init_node(struct kfd_node *node) +{ + int err = -1; + + if (kfd_interrupt_init(node)) { + dev_err(kfd_device, "Error initializing interrupts\n"); + goto kfd_interrupt_error; + } + + node->dqm = device_queue_manager_init(node); + if (!node->dqm) { + dev_err(kfd_device, "Error initializing queue manager\n"); + goto device_queue_manager_error; + } + + if (kfd_gws_init(node)) { + dev_err(kfd_device, "Could not allocate %d gws\n", + node->adev->gds.gws_size); + goto gws_error; + } + + if (kfd_resume(node)) + goto kfd_resume_error; + + if (kfd_topology_add_device(node)) { + dev_err(kfd_device, "Error adding device to topology\n"); + goto kfd_topology_add_device_error; + } + + kfd_smi_init(node); + + return 0; + +kfd_topology_add_device_error: +kfd_resume_error: +gws_error: + device_queue_manager_uninit(node->dqm); +device_queue_manager_error: + kfd_interrupt_exit(node); +kfd_interrupt_error: + if (node->gws) + amdgpu_amdkfd_free_gws(node->adev, node->gws); + + /* Cleanup the node memory here */ + kfree(node); + return err; +} + +static void kfd_cleanup_nodes(struct kfd_dev *kfd, unsigned int num_nodes) +{ + struct kfd_node *knode; + unsigned int i; + + for (i = 0; i < num_nodes; i++) { + knode = kfd->nodes[i]; + device_queue_manager_uninit(knode->dqm); + kfd_interrupt_exit(knode); + kfd_topology_remove_device(knode); + if (knode->gws) + amdgpu_amdkfd_free_gws(knode->adev, knode->gws); + kfree(knode); + kfd->nodes[i] = NULL; + } +} + bool kgd2kfd_device_init(struct kfd_dev *kfd, const struct kgd2kfd_shared_resources *gpu_resources) { - unsigned int size, map_process_packet_size; + unsigned int size, map_process_packet_size, i; + struct kfd_node *node; + uint32_t first_vmid_kfd, last_vmid_kfd, vmid_num_kfd; + unsigned int max_proc_per_quantum; + int partition_mode; + int xcp_idx; kfd->mec_fw_version = amdgpu_amdkfd_get_fw_version(kfd->adev, KGD_ENGINE_MEC1); @@ -532,10 +616,14 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd, KGD_ENGINE_SDMA1); kfd->shared_resources = *gpu_resources; - kfd->vm_info.first_vmid_kfd = ffs(gpu_resources->compute_vmid_bitmap)-1; - kfd->vm_info.last_vmid_kfd = fls(gpu_resources->compute_vmid_bitmap)-1; - kfd->vm_info.vmid_num_kfd = kfd->vm_info.last_vmid_kfd - - kfd->vm_info.first_vmid_kfd + 1; + kfd->num_nodes = amdgpu_xcp_get_num_xcp(kfd->adev->xcp_mgr); + + if (kfd->num_nodes == 0) { + dev_err(kfd_device, + "KFD num nodes cannot be 0, num_xcc_in_node: %d\n", + kfd->adev->gfx.num_xcc_per_xcp); + goto out; + } /* Allow BIF to recode atomics to PCIe 3.0 AtomicOps. * 32 and 64-bit requests are possible and must be @@ -554,11 +642,34 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd, return false; } + first_vmid_kfd = ffs(gpu_resources->compute_vmid_bitmap)-1; + last_vmid_kfd = fls(gpu_resources->compute_vmid_bitmap)-1; + vmid_num_kfd = last_vmid_kfd - first_vmid_kfd + 1; + + /* For GFX9.4.3, we need special handling for VMIDs depending on + * partition mode. + * In CPX mode, the VMID range needs to be shared between XCDs. + * Additionally, there are 13 VMIDs (3-15) available for KFD. To + * divide them equally, we change starting VMID to 4 and not use + * VMID 3. + * If the VMID range changes for GFX9.4.3, then this code MUST be + * revisited. + */ + if (kfd->adev->xcp_mgr) { + partition_mode = amdgpu_xcp_query_partition_mode(kfd->adev->xcp_mgr, + AMDGPU_XCP_FL_LOCKED); + if (partition_mode == AMDGPU_CPX_PARTITION_MODE && + kfd->num_nodes != 1) { + vmid_num_kfd /= 2; + first_vmid_kfd = last_vmid_kfd + 1 - vmid_num_kfd*2; + } + } + /* Verify module parameters regarding mapped process number*/ if (hws_max_conc_proc >= 0) - kfd->max_proc_per_quantum = min((u32)hws_max_conc_proc, kfd->vm_info.vmid_num_kfd); + max_proc_per_quantum = min((u32)hws_max_conc_proc, vmid_num_kfd); else - kfd->max_proc_per_quantum = kfd->vm_info.vmid_num_kfd; + max_proc_per_quantum = vmid_num_kfd; /* calculate max size of mqds needed for queues */ size = max_num_of_queues_per_device * @@ -606,27 +717,15 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd, if (amdgpu_use_xgmi_p2p) kfd->hive_id = kfd->adev->gmc.xgmi.hive_id; - kfd->noretry = kfd->adev->gmc.noretry; - - if (kfd_interrupt_init(kfd)) { - dev_err(kfd_device, "Error initializing interrupts\n"); - goto kfd_interrupt_error; - } - - kfd->dqm = device_queue_manager_init(kfd); - if (!kfd->dqm) { - dev_err(kfd_device, "Error initializing queue manager\n"); - goto device_queue_manager_error; - } - - /* If supported on this device, allocate global GWS that is shared - * by all KFD processes + /* + * For GFX9.4.3, the KFD abstracts all partitions within a socket as + * xGMI connected in the topology so assign a unique hive id per + * device based on the pci device location if device is in PCIe mode. */ - if (kfd_gws_init(kfd)) { - dev_err(kfd_device, "Could not allocate %d gws\n", - kfd->adev->gds.gws_size); - goto gws_error; - } + if (!kfd->hive_id && (KFD_GC_VERSION(kfd) == IP_VERSION(9, 4, 3)) && kfd->num_nodes > 1) + kfd->hive_id = pci_dev_id(kfd->adev->pdev); + + kfd->noretry = kfd->adev->gmc.noretry; /* If CRAT is broken, won't set iommu enabled */ kfd_double_confirm_iommu_support(kfd); @@ -639,48 +738,100 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd, kfd_cwsr_init(kfd); - svm_migrate_init(kfd->adev); + dev_info(kfd_device, "Total number of KFD nodes to be created: %d\n", + kfd->num_nodes); + + /* Allocate the KFD nodes */ + for (i = 0, xcp_idx = 0; i < kfd->num_nodes; i++) { + node = kzalloc(sizeof(struct kfd_node), GFP_KERNEL); + if (!node) + goto node_alloc_error; + + node->node_id = i; + node->adev = kfd->adev; + node->kfd = kfd; + node->kfd2kgd = kfd->kfd2kgd; + node->vm_info.vmid_num_kfd = vmid_num_kfd; + node->xcp = amdgpu_get_next_xcp(kfd->adev->xcp_mgr, &xcp_idx); + /* TODO : Check if error handling is needed */ + if (node->xcp) { + amdgpu_xcp_get_inst_details(node->xcp, AMDGPU_XCP_GFX, + &node->xcc_mask); + ++xcp_idx; + } else { + node->xcc_mask = + (1U << NUM_XCC(kfd->adev->gfx.xcc_mask)) - 1; + } - if (kfd_resume_iommu(kfd)) - goto device_iommu_error; + if (node->xcp) { + dev_info(kfd_device, "KFD node %d partition %d size %lldM\n", + node->node_id, node->xcp->mem_id, + KFD_XCP_MEMORY_SIZE(node->adev, node->node_id) >> 20); + } - if (kfd_resume(kfd)) - goto kfd_resume_error; + if (KFD_GC_VERSION(kfd) == IP_VERSION(9, 4, 3) && + partition_mode == AMDGPU_CPX_PARTITION_MODE && + kfd->num_nodes != 1) { + /* For GFX9.4.3 and CPX mode, first XCD gets VMID range + * 4-9 and second XCD gets VMID range 10-15. + */ - amdgpu_amdkfd_get_local_mem_info(kfd->adev, &kfd->local_mem_info); + node->vm_info.first_vmid_kfd = (i%2 == 0) ? + first_vmid_kfd : + first_vmid_kfd+vmid_num_kfd; + node->vm_info.last_vmid_kfd = (i%2 == 0) ? + last_vmid_kfd-vmid_num_kfd : + last_vmid_kfd; + node->compute_vmid_bitmap = + ((0x1 << (node->vm_info.last_vmid_kfd + 1)) - 1) - + ((0x1 << (node->vm_info.first_vmid_kfd)) - 1); + } else { + node->vm_info.first_vmid_kfd = first_vmid_kfd; + node->vm_info.last_vmid_kfd = last_vmid_kfd; + node->compute_vmid_bitmap = + gpu_resources->compute_vmid_bitmap; + } + node->max_proc_per_quantum = max_proc_per_quantum; + atomic_set(&node->sram_ecc_flag, 0); - if (kfd_topology_add_device(kfd)) { - dev_err(kfd_device, "Error adding device to topology\n"); - goto kfd_topology_add_device_error; + amdgpu_amdkfd_get_local_mem_info(kfd->adev, + &node->local_mem_info, node->xcp); + + /* Initialize the KFD node */ + if (kfd_init_node(node)) { + dev_err(kfd_device, "Error initializing KFD node\n"); + goto node_init_error; + } + kfd->nodes[i] = node; } - kfd_smi_init(kfd); + svm_range_set_max_pages(kfd->adev); + + if (kfd_resume_iommu(kfd)) + goto kfd_resume_iommu_error; + + spin_lock_init(&kfd->watch_points_lock); kfd->init_complete = true; dev_info(kfd_device, "added device %x:%x\n", kfd->adev->pdev->vendor, kfd->adev->pdev->device); pr_debug("Starting kfd with the following scheduling policy %d\n", - kfd->dqm->sched_policy); + node->dqm->sched_policy); goto out; -kfd_topology_add_device_error: -kfd_resume_error: +kfd_resume_iommu_error: +node_init_error: +node_alloc_error: + kfd_cleanup_nodes(kfd, i); device_iommu_error: -gws_error: - device_queue_manager_uninit(kfd->dqm); -device_queue_manager_error: - kfd_interrupt_exit(kfd); -kfd_interrupt_error: kfd_doorbell_fini(kfd); kfd_doorbell_error: kfd_gtt_sa_fini(kfd); kfd_gtt_sa_init_error: amdgpu_amdkfd_free_gtt_mem(kfd->adev, kfd->gtt_mem); alloc_gtt_mem_failure: - if (kfd->gws) - amdgpu_amdkfd_free_gws(kfd->adev, kfd->gws); dev_err(kfd_device, "device %x:%x NOT added due to errors\n", kfd->adev->pdev->vendor, kfd->adev->pdev->device); @@ -691,15 +842,13 @@ out: void kgd2kfd_device_exit(struct kfd_dev *kfd) { if (kfd->init_complete) { - device_queue_manager_uninit(kfd->dqm); - kfd_interrupt_exit(kfd); - kfd_topology_remove_device(kfd); + /* Cleanup KFD nodes */ + kfd_cleanup_nodes(kfd, kfd->num_nodes); + /* Cleanup common/shared resources */ kfd_doorbell_fini(kfd); ida_destroy(&kfd->doorbell_ida); kfd_gtt_sa_fini(kfd); amdgpu_amdkfd_free_gtt_mem(kfd->adev, kfd->gtt_mem); - if (kfd->gws) - amdgpu_amdkfd_free_gws(kfd->adev, kfd->gws); } kfree(kfd); @@ -707,16 +856,23 @@ void kgd2kfd_device_exit(struct kfd_dev *kfd) int kgd2kfd_pre_reset(struct kfd_dev *kfd) { + struct kfd_node *node; + int i; + if (!kfd->init_complete) return 0; - kfd_smi_event_update_gpu_reset(kfd, false); - - kfd->dqm->ops.pre_reset(kfd->dqm); + for (i = 0; i < kfd->num_nodes; i++) { + node = kfd->nodes[i]; + kfd_smi_event_update_gpu_reset(node, false); + node->dqm->ops.pre_reset(node->dqm); + } kgd2kfd_suspend(kfd, false); - kfd_signal_reset_event(kfd); + for (i = 0; i < kfd->num_nodes; i++) + kfd_signal_reset_event(kfd->nodes[i]); + return 0; } @@ -729,57 +885,83 @@ int kgd2kfd_pre_reset(struct kfd_dev *kfd) int kgd2kfd_post_reset(struct kfd_dev *kfd) { int ret; + struct kfd_node *node; + int i; if (!kfd->init_complete) return 0; - ret = kfd_resume(kfd); - if (ret) - return ret; - atomic_dec(&kfd_locked); + for (i = 0; i < kfd->num_nodes; i++) { + ret = kfd_resume(kfd->nodes[i]); + if (ret) + return ret; + } - atomic_set(&kfd->sram_ecc_flag, 0); + mutex_lock(&kfd_processes_mutex); + --kfd_locked; + mutex_unlock(&kfd_processes_mutex); - kfd_smi_event_update_gpu_reset(kfd, true); + for (i = 0; i < kfd->num_nodes; i++) { + node = kfd->nodes[i]; + atomic_set(&node->sram_ecc_flag, 0); + kfd_smi_event_update_gpu_reset(node, true); + } return 0; } bool kfd_is_locked(void) { - return (atomic_read(&kfd_locked) > 0); + lockdep_assert_held(&kfd_processes_mutex); + return (kfd_locked > 0); } void kgd2kfd_suspend(struct kfd_dev *kfd, bool run_pm) { + struct kfd_node *node; + int i; + int count; + if (!kfd->init_complete) return; /* for runtime suspend, skip locking kfd */ if (!run_pm) { + mutex_lock(&kfd_processes_mutex); + count = ++kfd_locked; + mutex_unlock(&kfd_processes_mutex); + /* For first KFD device suspend all the KFD processes */ - if (atomic_inc_return(&kfd_locked) == 1) + if (count == 1) kfd_suspend_all_processes(); } - kfd->dqm->ops.stop(kfd->dqm); + for (i = 0; i < kfd->num_nodes; i++) { + node = kfd->nodes[i]; + node->dqm->ops.stop(node->dqm); + } kfd_iommu_suspend(kfd); } int kgd2kfd_resume(struct kfd_dev *kfd, bool run_pm) { - int ret, count; + int ret, count, i; if (!kfd->init_complete) return 0; - ret = kfd_resume(kfd); - if (ret) - return ret; + for (i = 0; i < kfd->num_nodes; i++) { + ret = kfd_resume(kfd->nodes[i]); + if (ret) + return ret; + } /* for runtime resume, skip unlocking kfd */ if (!run_pm) { - count = atomic_dec_return(&kfd_locked); + mutex_lock(&kfd_processes_mutex); + count = --kfd_locked; + mutex_unlock(&kfd_processes_mutex); + WARN_ONCE(count < 0, "KFD suspend / resume ref. error"); if (count == 0) ret = kfd_resume_all_processes(); @@ -808,15 +990,15 @@ static int kfd_resume_iommu(struct kfd_dev *kfd) return err; } -static int kfd_resume(struct kfd_dev *kfd) +static int kfd_resume(struct kfd_node *node) { int err = 0; - err = kfd->dqm->ops.start(kfd->dqm); + err = node->dqm->ops.start(node->dqm); if (err) dev_err(kfd_device, "Error starting queue manager for device %x:%x\n", - kfd->adev->pdev->vendor, kfd->adev->pdev->device); + node->adev->pdev->vendor, node->adev->pdev->device); return err; } @@ -839,9 +1021,10 @@ static inline void kfd_queue_work(struct workqueue_struct *wq, /* This is called directly from KGD at ISR. */ void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry) { - uint32_t patched_ihre[KFD_MAX_RING_ENTRY_SIZE]; + uint32_t patched_ihre[KFD_MAX_RING_ENTRY_SIZE], i; bool is_patched = false; unsigned long flags; + struct kfd_node *node; if (!kfd->init_complete) return; @@ -851,16 +1034,22 @@ void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry) return; } - spin_lock_irqsave(&kfd->interrupt_lock, flags); - - if (kfd->interrupts_active - && interrupt_is_wanted(kfd, ih_ring_entry, - patched_ihre, &is_patched) - && enqueue_ih_ring_entry(kfd, - is_patched ? patched_ihre : ih_ring_entry)) - kfd_queue_work(kfd->ih_wq, &kfd->interrupt_work); + for (i = 0; i < kfd->num_nodes; i++) { + node = kfd->nodes[i]; + spin_lock_irqsave(&node->interrupt_lock, flags); + + if (node->interrupts_active + && interrupt_is_wanted(node, ih_ring_entry, + patched_ihre, &is_patched) + && enqueue_ih_ring_entry(node, + is_patched ? patched_ihre : ih_ring_entry)) { + kfd_queue_work(node->ih_wq, &node->interrupt_work); + spin_unlock_irqrestore(&node->interrupt_lock, flags); + return; + } + spin_unlock_irqrestore(&node->interrupt_lock, flags); + } - spin_unlock_irqrestore(&kfd->interrupt_lock, flags); } int kgd2kfd_quiesce_mm(struct mm_struct *mm, uint32_t trigger) @@ -998,10 +1187,11 @@ static inline uint32_t *kfd_gtt_sa_calc_cpu_addr(void *start_addr, return (uint32_t *) ((uint64_t) start_addr + bit_num * chunk_size); } -int kfd_gtt_sa_allocate(struct kfd_dev *kfd, unsigned int size, +int kfd_gtt_sa_allocate(struct kfd_node *node, unsigned int size, struct kfd_mem_obj **mem_obj) { unsigned int found, start_search, cur_size; + struct kfd_dev *kfd = node->kfd; if (size == 0) return -EINVAL; @@ -1101,8 +1291,10 @@ kfd_gtt_no_free_chunk: return -ENOMEM; } -int kfd_gtt_sa_free(struct kfd_dev *kfd, struct kfd_mem_obj *mem_obj) +int kfd_gtt_sa_free(struct kfd_node *node, struct kfd_mem_obj *mem_obj) { + struct kfd_dev *kfd = node->kfd; + /* Act like kfree when trying to free a NULL object */ if (!mem_obj) return 0; @@ -1124,29 +1316,40 @@ int kfd_gtt_sa_free(struct kfd_dev *kfd, struct kfd_mem_obj *mem_obj) void kgd2kfd_set_sram_ecc_flag(struct kfd_dev *kfd) { + /* + * TODO: Currently update SRAM ECC flag for first node. + * This needs to be updated later when we can + * identify SRAM ECC error on other nodes also. + */ if (kfd) - atomic_inc(&kfd->sram_ecc_flag); + atomic_inc(&kfd->nodes[0]->sram_ecc_flag); } -void kfd_inc_compute_active(struct kfd_dev *kfd) +void kfd_inc_compute_active(struct kfd_node *node) { - if (atomic_inc_return(&kfd->compute_profile) == 1) - amdgpu_amdkfd_set_compute_idle(kfd->adev, false); + if (atomic_inc_return(&node->kfd->compute_profile) == 1) + amdgpu_amdkfd_set_compute_idle(node->adev, false); } -void kfd_dec_compute_active(struct kfd_dev *kfd) +void kfd_dec_compute_active(struct kfd_node *node) { - int count = atomic_dec_return(&kfd->compute_profile); + int count = atomic_dec_return(&node->kfd->compute_profile); if (count == 0) - amdgpu_amdkfd_set_compute_idle(kfd->adev, true); + amdgpu_amdkfd_set_compute_idle(node->adev, true); WARN_ONCE(count < 0, "Compute profile ref. count error"); } void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint64_t throttle_bitmask) { + /* + * TODO: For now, raise the throttling event only on first node. + * This will need to change after we are able to determine + * which node raised the throttling event. + */ if (kfd && kfd->init_complete) - kfd_smi_event_update_thermal_throttling(kfd, throttle_bitmask); + kfd_smi_event_update_thermal_throttling(kfd->nodes[0], + throttle_bitmask); } /* kfd_get_num_sdma_engines returns the number of PCIe optimized SDMA and @@ -1154,19 +1357,41 @@ void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint64_t throttle_bitmask) * When the device has more than two engines, we reserve two for PCIe to enable * full-duplex and the rest are used as XGMI. */ -unsigned int kfd_get_num_sdma_engines(struct kfd_dev *kdev) +unsigned int kfd_get_num_sdma_engines(struct kfd_node *node) { /* If XGMI is not supported, all SDMA engines are PCIe */ - if (!kdev->adev->gmc.xgmi.supported) - return kdev->adev->sdma.num_instances; + if (!node->adev->gmc.xgmi.supported) + return node->adev->sdma.num_instances/(int)node->kfd->num_nodes; - return min(kdev->adev->sdma.num_instances, 2); + return min(node->adev->sdma.num_instances/(int)node->kfd->num_nodes, 2); } -unsigned int kfd_get_num_xgmi_sdma_engines(struct kfd_dev *kdev) +unsigned int kfd_get_num_xgmi_sdma_engines(struct kfd_node *node) { /* After reserved for PCIe, the rest of engines are XGMI */ - return kdev->adev->sdma.num_instances - kfd_get_num_sdma_engines(kdev); + return node->adev->sdma.num_instances/(int)node->kfd->num_nodes - + kfd_get_num_sdma_engines(node); +} + +int kgd2kfd_check_and_lock_kfd(void) +{ + mutex_lock(&kfd_processes_mutex); + if (!hash_empty(kfd_processes_table) || kfd_is_locked()) { + mutex_unlock(&kfd_processes_mutex); + return -EBUSY; + } + + ++kfd_locked; + mutex_unlock(&kfd_processes_mutex); + + return 0; +} + +void kgd2kfd_unlock_kfd(void) +{ + mutex_lock(&kfd_processes_mutex); + --kfd_locked; + mutex_unlock(&kfd_processes_mutex); } #if defined(CONFIG_DEBUG_FS) @@ -1174,7 +1399,7 @@ unsigned int kfd_get_num_xgmi_sdma_engines(struct kfd_dev *kdev) /* This function will send a package to HIQ to hang the HWS * which will trigger a GPU reset and bring the HWS back to normal state */ -int kfd_debugfs_hang_hws(struct kfd_dev *dev) +int kfd_debugfs_hang_hws(struct kfd_node *dev) { if (dev->dqm->sched_policy != KFD_SCHED_POLICY_HWS) { pr_err("HWS is not enabled"); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c index 7a95698d83f7..f515cb8f30ca 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c @@ -36,6 +36,7 @@ #include "kfd_kernel_queue.h" #include "amdgpu_amdkfd.h" #include "mes_api_def.h" +#include "kfd_debug.h" /* Size of the per-pipe EOP queue */ #define CIK_HPD_EOP_BYTES_LOG2 11 @@ -46,10 +47,13 @@ static int set_pasid_vmid_mapping(struct device_queue_manager *dqm, static int execute_queues_cpsch(struct device_queue_manager *dqm, enum kfd_unmap_queues_filter filter, - uint32_t filter_param); + uint32_t filter_param, + uint32_t grace_period); static int unmap_queues_cpsch(struct device_queue_manager *dqm, enum kfd_unmap_queues_filter filter, - uint32_t filter_param, bool reset); + uint32_t filter_param, + uint32_t grace_period, + bool reset); static int map_queues_cpsch(struct device_queue_manager *dqm); @@ -74,31 +78,31 @@ enum KFD_MQD_TYPE get_mqd_type_from_queue_type(enum kfd_queue_type type) static bool is_pipe_enabled(struct device_queue_manager *dqm, int mec, int pipe) { int i; - int pipe_offset = (mec * dqm->dev->shared_resources.num_pipe_per_mec - + pipe) * dqm->dev->shared_resources.num_queue_per_pipe; + int pipe_offset = (mec * dqm->dev->kfd->shared_resources.num_pipe_per_mec + + pipe) * dqm->dev->kfd->shared_resources.num_queue_per_pipe; /* queue is available for KFD usage if bit is 1 */ - for (i = 0; i < dqm->dev->shared_resources.num_queue_per_pipe; ++i) + for (i = 0; i < dqm->dev->kfd->shared_resources.num_queue_per_pipe; ++i) if (test_bit(pipe_offset + i, - dqm->dev->shared_resources.cp_queue_bitmap)) + dqm->dev->kfd->shared_resources.cp_queue_bitmap)) return true; return false; } unsigned int get_cp_queues_num(struct device_queue_manager *dqm) { - return bitmap_weight(dqm->dev->shared_resources.cp_queue_bitmap, + return bitmap_weight(dqm->dev->kfd->shared_resources.cp_queue_bitmap, KGD_MAX_QUEUES); } unsigned int get_queues_per_pipe(struct device_queue_manager *dqm) { - return dqm->dev->shared_resources.num_queue_per_pipe; + return dqm->dev->kfd->shared_resources.num_queue_per_pipe; } unsigned int get_pipes_per_mec(struct device_queue_manager *dqm) { - return dqm->dev->shared_resources.num_pipe_per_mec; + return dqm->dev->kfd->shared_resources.num_pipe_per_mec; } static unsigned int get_num_all_sdma_engines(struct device_queue_manager *dqm) @@ -110,29 +114,40 @@ static unsigned int get_num_all_sdma_engines(struct device_queue_manager *dqm) unsigned int get_num_sdma_queues(struct device_queue_manager *dqm) { return kfd_get_num_sdma_engines(dqm->dev) * - dqm->dev->device_info.num_sdma_queues_per_engine; + dqm->dev->kfd->device_info.num_sdma_queues_per_engine; } unsigned int get_num_xgmi_sdma_queues(struct device_queue_manager *dqm) { return kfd_get_num_xgmi_sdma_engines(dqm->dev) * - dqm->dev->device_info.num_sdma_queues_per_engine; + dqm->dev->kfd->device_info.num_sdma_queues_per_engine; } -static inline uint64_t get_reserved_sdma_queues_bitmap(struct device_queue_manager *dqm) +static void init_sdma_bitmaps(struct device_queue_manager *dqm) { - return dqm->dev->device_info.reserved_sdma_queues_bitmap; + bitmap_zero(dqm->sdma_bitmap, KFD_MAX_SDMA_QUEUES); + bitmap_set(dqm->sdma_bitmap, 0, get_num_sdma_queues(dqm)); + + bitmap_zero(dqm->xgmi_sdma_bitmap, KFD_MAX_SDMA_QUEUES); + bitmap_set(dqm->xgmi_sdma_bitmap, 0, get_num_xgmi_sdma_queues(dqm)); + + /* Mask out the reserved queues */ + bitmap_andnot(dqm->sdma_bitmap, dqm->sdma_bitmap, + dqm->dev->kfd->device_info.reserved_sdma_queues_bitmap, + KFD_MAX_SDMA_QUEUES); } void program_sh_mem_settings(struct device_queue_manager *dqm, struct qcm_process_device *qpd) { - return dqm->dev->kfd2kgd->program_sh_mem_settings( - dqm->dev->adev, qpd->vmid, - qpd->sh_mem_config, - qpd->sh_mem_ape1_base, - qpd->sh_mem_ape1_limit, - qpd->sh_mem_bases); + uint32_t xcc_mask = dqm->dev->xcc_mask; + int xcc_id; + + for_each_inst(xcc_id, xcc_mask) + dqm->dev->kfd2kgd->program_sh_mem_settings( + dqm->dev->adev, qpd->vmid, qpd->sh_mem_config, + qpd->sh_mem_ape1_base, qpd->sh_mem_ape1_limit, + qpd->sh_mem_bases, xcc_id); } static void kfd_hws_hang(struct device_queue_manager *dqm) @@ -211,6 +226,9 @@ static int add_queue_mes(struct device_queue_manager *dqm, struct queue *q, queue_input.paging = false; queue_input.tba_addr = qpd->tba_addr; queue_input.tma_addr = qpd->tma_addr; + queue_input.trap_en = KFD_GC_VERSION(q->device) < IP_VERSION(11, 0, 0) || + KFD_GC_VERSION(q->device) > IP_VERSION(11, 0, 3); + queue_input.skip_process_ctx_clear = qpd->pqm->process->debug_trap_enabled; queue_type = convert_to_mes_queue_type(q->properties.type); if (queue_type < 0) { @@ -330,7 +348,7 @@ static int allocate_doorbell(struct qcm_process_device *qpd, struct queue *q, uint32_t const *restore_id) { - struct kfd_dev *dev = qpd->dqm->dev; + struct kfd_node *dev = qpd->dqm->dev; if (!KFD_IS_SOC15(dev)) { /* On pre-SOC15 chips we need to use the queue ID to @@ -349,8 +367,17 @@ static int allocate_doorbell(struct qcm_process_device *qpd, * for a SDMA engine is 512. */ - uint32_t *idx_offset = dev->shared_resources.sdma_doorbell_idx; - uint32_t valid_id = idx_offset[q->properties.sdma_engine_id] + uint32_t *idx_offset = dev->kfd->shared_resources.sdma_doorbell_idx; + + /* + * q->properties.sdma_engine_id corresponds to the virtual + * sdma engine number. However, for doorbell allocation, + * we need the physical sdma engine id in order to get the + * correct doorbell offset. + */ + uint32_t valid_id = idx_offset[qpd->dqm->dev->node_id * + get_num_all_sdma_engines(qpd->dqm) + + q->properties.sdma_engine_id] + (q->properties.sdma_queue_id & 1) * KFD_QUEUE_DOORBELL_MIRROR_OFFSET + (q->properties.sdma_queue_id >> 1); @@ -382,7 +409,7 @@ static int allocate_doorbell(struct qcm_process_device *qpd, } q->properties.doorbell_off = - kfd_get_doorbell_dw_offset_in_bar(dev, qpd_to_pdd(qpd), + kfd_get_doorbell_dw_offset_in_bar(dev->kfd, qpd_to_pdd(qpd), q->doorbell_id); return 0; } @@ -391,7 +418,7 @@ static void deallocate_doorbell(struct qcm_process_device *qpd, struct queue *q) { unsigned int old; - struct kfd_dev *dev = qpd->dqm->dev; + struct kfd_node *dev = qpd->dqm->dev; if (!KFD_IS_SOC15(dev) || q->properties.type == KFD_QUEUE_TYPE_SDMA || @@ -405,10 +432,14 @@ static void deallocate_doorbell(struct qcm_process_device *qpd, static void program_trap_handler_settings(struct device_queue_manager *dqm, struct qcm_process_device *qpd) { + uint32_t xcc_mask = dqm->dev->xcc_mask; + int xcc_id; + if (dqm->dev->kfd2kgd->program_trap_handler_settings) - dqm->dev->kfd2kgd->program_trap_handler_settings( - dqm->dev->adev, qpd->vmid, - qpd->tba_addr, qpd->tma_addr); + for_each_inst(xcc_id, xcc_mask) + dqm->dev->kfd2kgd->program_trap_handler_settings( + dqm->dev->adev, qpd->vmid, qpd->tba_addr, + qpd->tma_addr, xcc_id); } static int allocate_vmid(struct device_queue_manager *dqm, @@ -441,7 +472,7 @@ static int allocate_vmid(struct device_queue_manager *dqm, program_sh_mem_settings(dqm, qpd); - if (KFD_IS_SOC15(dqm->dev) && dqm->dev->cwsr_enabled) + if (KFD_IS_SOC15(dqm->dev) && dqm->dev->kfd->cwsr_enabled) program_trap_handler_settings(dqm, qpd); /* qpd->page_table_base is set earlier when register_process() @@ -460,7 +491,7 @@ static int allocate_vmid(struct device_queue_manager *dqm, return 0; } -static int flush_texture_cache_nocpsch(struct kfd_dev *kdev, +static int flush_texture_cache_nocpsch(struct kfd_node *kdev, struct qcm_process_device *qpd) { const struct packet_manager_funcs *pmf = qpd->dqm->packet_mgr.pmf; @@ -661,7 +692,7 @@ static inline void deallocate_hqd(struct device_queue_manager *dqm, #define SQ_IND_CMD_CMD_KILL 0x00000003 #define SQ_IND_CMD_MODE_BROADCAST 0x00000001 -static int dbgdev_wave_reset_wavefronts(struct kfd_dev *dev, struct kfd_process *p) +static int dbgdev_wave_reset_wavefronts(struct kfd_node *dev, struct kfd_process *p) { int status = 0; unsigned int vmid; @@ -671,6 +702,8 @@ static int dbgdev_wave_reset_wavefronts(struct kfd_dev *dev, struct kfd_process struct kfd_process_device *pdd; int first_vmid_to_scan = dev->vm_info.first_vmid_kfd; int last_vmid_to_scan = dev->vm_info.last_vmid_kfd; + uint32_t xcc_mask = dev->xcc_mask; + int xcc_id; reg_sq_cmd.u32All = 0; reg_gfx_index.u32All = 0; @@ -715,9 +748,10 @@ static int dbgdev_wave_reset_wavefronts(struct kfd_dev *dev, struct kfd_process reg_sq_cmd.bits.cmd = SQ_IND_CMD_CMD_KILL; reg_sq_cmd.bits.vm_id = vmid; - dev->kfd2kgd->wave_control_execute(dev->adev, - reg_gfx_index.u32All, - reg_sq_cmd.u32All); + for_each_inst(xcc_id, xcc_mask) + dev->kfd2kgd->wave_control_execute( + dev->adev, reg_gfx_index.u32All, + reg_sq_cmd.u32All, xcc_id); return 0; } @@ -837,9 +871,9 @@ static int update_queue(struct device_queue_manager *dqm, struct queue *q, /* Make sure the queue is unmapped before updating the MQD */ if (dqm->sched_policy != KFD_SCHED_POLICY_NO_HWS) { - if (!dqm->dev->shared_resources.enable_mes) + if (!dqm->dev->kfd->shared_resources.enable_mes) retval = unmap_queues_cpsch(dqm, - KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, false); + KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD, false); else if (prev_active) retval = remove_queue_mes(dqm, q, &pdd->qpd); @@ -858,7 +892,7 @@ static int update_queue(struct device_queue_manager *dqm, struct queue *q, } retval = mqd_mgr->destroy_mqd(mqd_mgr, q->mqd, - (dqm->dev->cwsr_enabled ? + (dqm->dev->kfd->cwsr_enabled ? KFD_PREEMPT_TYPE_WAVEFRONT_SAVE : KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN), KFD_UNMAP_LATENCY_MS, q->pipe, q->queue); @@ -895,7 +929,7 @@ static int update_queue(struct device_queue_manager *dqm, struct queue *q, } if (dqm->sched_policy != KFD_SCHED_POLICY_NO_HWS) { - if (!dqm->dev->shared_resources.enable_mes) + if (!dqm->dev->kfd->shared_resources.enable_mes) retval = map_queues_cpsch(dqm); else if (q->properties.is_active) retval = add_queue_mes(dqm, q, &pdd->qpd); @@ -917,6 +951,92 @@ out_unlock: return retval; } +/* suspend_single_queue does not lock the dqm like the + * evict_process_queues_cpsch or evict_process_queues_nocpsch. You should + * lock the dqm before calling, and unlock after calling. + * + * The reason we don't lock the dqm is because this function may be + * called on multiple queues in a loop, so rather than locking/unlocking + * multiple times, we will just keep the dqm locked for all of the calls. + */ +static int suspend_single_queue(struct device_queue_manager *dqm, + struct kfd_process_device *pdd, + struct queue *q) +{ + bool is_new; + + if (q->properties.is_suspended) + return 0; + + pr_debug("Suspending PASID %u queue [%i]\n", + pdd->process->pasid, + q->properties.queue_id); + + is_new = q->properties.exception_status & KFD_EC_MASK(EC_QUEUE_NEW); + + if (is_new || q->properties.is_being_destroyed) { + pr_debug("Suspend: skip %s queue id %i\n", + is_new ? "new" : "destroyed", + q->properties.queue_id); + return -EBUSY; + } + + q->properties.is_suspended = true; + if (q->properties.is_active) { + if (dqm->dev->kfd->shared_resources.enable_mes) { + int r = remove_queue_mes(dqm, q, &pdd->qpd); + + if (r) + return r; + } + + decrement_queue_count(dqm, &pdd->qpd, q); + q->properties.is_active = false; + } + + return 0; +} + +/* resume_single_queue does not lock the dqm like the functions + * restore_process_queues_cpsch or restore_process_queues_nocpsch. You should + * lock the dqm before calling, and unlock after calling. + * + * The reason we don't lock the dqm is because this function may be + * called on multiple queues in a loop, so rather than locking/unlocking + * multiple times, we will just keep the dqm locked for all of the calls. + */ +static int resume_single_queue(struct device_queue_manager *dqm, + struct qcm_process_device *qpd, + struct queue *q) +{ + struct kfd_process_device *pdd; + + if (!q->properties.is_suspended) + return 0; + + pdd = qpd_to_pdd(qpd); + + pr_debug("Restoring from suspend PASID %u queue [%i]\n", + pdd->process->pasid, + q->properties.queue_id); + + q->properties.is_suspended = false; + + if (QUEUE_IS_ACTIVE(q->properties)) { + if (dqm->dev->kfd->shared_resources.enable_mes) { + int r = add_queue_mes(dqm, q, &pdd->qpd); + + if (r) + return r; + } + + q->properties.is_active = true; + increment_queue_count(dqm, qpd, q); + } + + return 0; +} + static int evict_process_queues_nocpsch(struct device_queue_manager *dqm, struct qcm_process_device *qpd) { @@ -951,7 +1071,7 @@ static int evict_process_queues_nocpsch(struct device_queue_manager *dqm, continue; retval = mqd_mgr->destroy_mqd(mqd_mgr, q->mqd, - (dqm->dev->cwsr_enabled ? + (dqm->dev->kfd->cwsr_enabled ? KFD_PREEMPT_TYPE_WAVEFRONT_SAVE : KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN), KFD_UNMAP_LATENCY_MS, q->pipe, q->queue); @@ -979,6 +1099,14 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm, goto out; pdd = qpd_to_pdd(qpd); + + /* The debugger creates processes that temporarily have not acquired + * all VMs for all devices and has no VMs itself. + * Skip queue eviction on process eviction. + */ + if (!pdd->drm_priv) + goto out; + pr_debug_ratelimited("Evicting PASID 0x%x queues\n", pdd->process->pasid); @@ -993,7 +1121,7 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm, q->properties.is_active = false; decrement_queue_count(dqm, qpd, q); - if (dqm->dev->shared_resources.enable_mes) { + if (dqm->dev->kfd->shared_resources.enable_mes) { retval = remove_queue_mes(dqm, q, qpd); if (retval) { pr_err("Failed to evict queue %d\n", @@ -1003,11 +1131,12 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm, } } pdd->last_evict_timestamp = get_jiffies_64(); - if (!dqm->dev->shared_resources.enable_mes) + if (!dqm->dev->kfd->shared_resources.enable_mes) retval = execute_queues_cpsch(dqm, qpd->is_debug ? KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES : - KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0); + KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, + USE_DEFAULT_GRACE_PERIOD); out: dqm_unlock(dqm); @@ -1100,13 +1229,10 @@ static int restore_process_queues_cpsch(struct device_queue_manager *dqm, { struct queue *q; struct kfd_process_device *pdd; - uint64_t pd_base; uint64_t eviction_duration; int retval = 0; pdd = qpd_to_pdd(qpd); - /* Retrieve PD base */ - pd_base = amdgpu_amdkfd_gpuvm_get_process_page_dir(pdd->drm_priv); dqm_lock(dqm); if (WARN_ON_ONCE(!qpd->evicted)) /* already restored, do nothing */ @@ -1116,12 +1242,19 @@ static int restore_process_queues_cpsch(struct device_queue_manager *dqm, goto out; } + /* The debugger creates processes that temporarily have not acquired + * all VMs for all devices and has no VMs itself. + * Skip queue restore on process restore. + */ + if (!pdd->drm_priv) + goto vm_not_acquired; + pr_debug_ratelimited("Restoring PASID 0x%x queues\n", pdd->process->pasid); /* Update PD Base in QPD */ - qpd->page_table_base = pd_base; - pr_debug("Updated PD address to 0x%llx\n", pd_base); + qpd->page_table_base = amdgpu_amdkfd_gpuvm_get_process_page_dir(pdd->drm_priv); + pr_debug("Updated PD address to 0x%llx\n", qpd->page_table_base); /* activate all active queues on the qpd */ list_for_each_entry(q, &qpd->queues_list, list) { @@ -1132,7 +1265,7 @@ static int restore_process_queues_cpsch(struct device_queue_manager *dqm, q->properties.is_active = true; increment_queue_count(dqm, &pdd->qpd, q); - if (dqm->dev->shared_resources.enable_mes) { + if (dqm->dev->kfd->shared_resources.enable_mes) { retval = add_queue_mes(dqm, q, qpd); if (retval) { pr_err("Failed to restore queue %d\n", @@ -1141,12 +1274,13 @@ static int restore_process_queues_cpsch(struct device_queue_manager *dqm, } } } - if (!dqm->dev->shared_resources.enable_mes) + if (!dqm->dev->kfd->shared_resources.enable_mes) retval = execute_queues_cpsch(dqm, - KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0); - qpd->evicted = 0; + KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD); eviction_duration = get_jiffies_64() - pdd->last_evict_timestamp; atomic64_add(eviction_duration, &pdd->evict_duration_counter); +vm_not_acquired: + qpd->evicted = 0; out: dqm_unlock(dqm); return retval; @@ -1229,35 +1363,32 @@ static int set_pasid_vmid_mapping(struct device_queue_manager *dqm, u32 pasid, unsigned int vmid) { - return dqm->dev->kfd2kgd->set_pasid_vmid_mapping( - dqm->dev->adev, pasid, vmid); -} + uint32_t xcc_mask = dqm->dev->xcc_mask; + int xcc_id, ret; -static void init_interrupts(struct device_queue_manager *dqm) -{ - unsigned int i; + for_each_inst(xcc_id, xcc_mask) { + ret = dqm->dev->kfd2kgd->set_pasid_vmid_mapping( + dqm->dev->adev, pasid, vmid, xcc_id); + if (ret) + break; + } - for (i = 0 ; i < get_pipes_per_mec(dqm) ; i++) - if (is_pipe_enabled(dqm, 0, i)) - dqm->dev->kfd2kgd->init_interrupts(dqm->dev->adev, i); + return ret; } -static void init_sdma_bitmaps(struct device_queue_manager *dqm) +static void init_interrupts(struct device_queue_manager *dqm) { - unsigned int num_sdma_queues = - min_t(unsigned int, sizeof(dqm->sdma_bitmap)*8, - get_num_sdma_queues(dqm)); - unsigned int num_xgmi_sdma_queues = - min_t(unsigned int, sizeof(dqm->xgmi_sdma_bitmap)*8, - get_num_xgmi_sdma_queues(dqm)); - - if (num_sdma_queues) - dqm->sdma_bitmap = GENMASK_ULL(num_sdma_queues-1, 0); - if (num_xgmi_sdma_queues) - dqm->xgmi_sdma_bitmap = GENMASK_ULL(num_xgmi_sdma_queues-1, 0); + uint32_t xcc_mask = dqm->dev->xcc_mask; + unsigned int i, xcc_id; - dqm->sdma_bitmap &= ~get_reserved_sdma_queues_bitmap(dqm); - pr_info("sdma_bitmap: %llx\n", dqm->sdma_bitmap); + for_each_inst(xcc_id, xcc_mask) { + for (i = 0 ; i < get_pipes_per_mec(dqm) ; i++) { + if (is_pipe_enabled(dqm, 0, i)) { + dqm->dev->kfd2kgd->init_interrupts( + dqm->dev->adev, i, xcc_id); + } + } + } } static int initialize_nocpsch(struct device_queue_manager *dqm) @@ -1282,7 +1413,7 @@ static int initialize_nocpsch(struct device_queue_manager *dqm) for (queue = 0; queue < get_queues_per_pipe(dqm); queue++) if (test_bit(pipe_offset + queue, - dqm->dev->shared_resources.cp_queue_bitmap)) + dqm->dev->kfd->shared_resources.cp_queue_bitmap)) dqm->allocated_queues[pipe] |= 1 << queue; } @@ -1322,9 +1453,16 @@ static int start_nocpsch(struct device_queue_manager *dqm) static int stop_nocpsch(struct device_queue_manager *dqm) { + dqm_lock(dqm); + if (!dqm->sched_running) { + dqm_unlock(dqm); + return 0; + } + if (dqm->dev->adev->asic_type == CHIP_HAWAII) pm_uninit(&dqm->packet_mgr, false); dqm->sched_running = false; + dqm_unlock(dqm); return 0; } @@ -1342,46 +1480,48 @@ static int allocate_sdma_queue(struct device_queue_manager *dqm, int bit; if (q->properties.type == KFD_QUEUE_TYPE_SDMA) { - if (dqm->sdma_bitmap == 0) { + if (bitmap_empty(dqm->sdma_bitmap, KFD_MAX_SDMA_QUEUES)) { pr_err("No more SDMA queue to allocate\n"); return -ENOMEM; } if (restore_sdma_id) { /* Re-use existing sdma_id */ - if (!(dqm->sdma_bitmap & (1ULL << *restore_sdma_id))) { + if (!test_bit(*restore_sdma_id, dqm->sdma_bitmap)) { pr_err("SDMA queue already in use\n"); return -EBUSY; } - dqm->sdma_bitmap &= ~(1ULL << *restore_sdma_id); + clear_bit(*restore_sdma_id, dqm->sdma_bitmap); q->sdma_id = *restore_sdma_id; } else { /* Find first available sdma_id */ - bit = __ffs64(dqm->sdma_bitmap); - dqm->sdma_bitmap &= ~(1ULL << bit); + bit = find_first_bit(dqm->sdma_bitmap, + get_num_sdma_queues(dqm)); + clear_bit(bit, dqm->sdma_bitmap); q->sdma_id = bit; } - q->properties.sdma_engine_id = q->sdma_id % - kfd_get_num_sdma_engines(dqm->dev); + q->properties.sdma_engine_id = + q->sdma_id % kfd_get_num_sdma_engines(dqm->dev); q->properties.sdma_queue_id = q->sdma_id / kfd_get_num_sdma_engines(dqm->dev); } else if (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) { - if (dqm->xgmi_sdma_bitmap == 0) { + if (bitmap_empty(dqm->xgmi_sdma_bitmap, KFD_MAX_SDMA_QUEUES)) { pr_err("No more XGMI SDMA queue to allocate\n"); return -ENOMEM; } if (restore_sdma_id) { /* Re-use existing sdma_id */ - if (!(dqm->xgmi_sdma_bitmap & (1ULL << *restore_sdma_id))) { + if (!test_bit(*restore_sdma_id, dqm->xgmi_sdma_bitmap)) { pr_err("SDMA queue already in use\n"); return -EBUSY; } - dqm->xgmi_sdma_bitmap &= ~(1ULL << *restore_sdma_id); + clear_bit(*restore_sdma_id, dqm->xgmi_sdma_bitmap); q->sdma_id = *restore_sdma_id; } else { - bit = __ffs64(dqm->xgmi_sdma_bitmap); - dqm->xgmi_sdma_bitmap &= ~(1ULL << bit); + bit = find_first_bit(dqm->xgmi_sdma_bitmap, + get_num_xgmi_sdma_queues(dqm)); + clear_bit(bit, dqm->xgmi_sdma_bitmap); q->sdma_id = bit; } /* sdma_engine_id is sdma id including @@ -1409,11 +1549,11 @@ static void deallocate_sdma_queue(struct device_queue_manager *dqm, if (q->properties.type == KFD_QUEUE_TYPE_SDMA) { if (q->sdma_id >= get_num_sdma_queues(dqm)) return; - dqm->sdma_bitmap |= (1ULL << q->sdma_id); + set_bit(q->sdma_id, dqm->sdma_bitmap); } else if (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) { if (q->sdma_id >= get_num_xgmi_sdma_queues(dqm)) return; - dqm->xgmi_sdma_bitmap |= (1ULL << q->sdma_id); + set_bit(q->sdma_id, dqm->xgmi_sdma_bitmap); } } @@ -1426,14 +1566,14 @@ static int set_sched_resources(struct device_queue_manager *dqm) int i, mec; struct scheduling_resources res; - res.vmid_mask = dqm->dev->shared_resources.compute_vmid_bitmap; + res.vmid_mask = dqm->dev->compute_vmid_bitmap; res.queue_mask = 0; for (i = 0; i < KGD_MAX_QUEUES; ++i) { - mec = (i / dqm->dev->shared_resources.num_queue_per_pipe) - / dqm->dev->shared_resources.num_pipe_per_mec; + mec = (i / dqm->dev->kfd->shared_resources.num_queue_per_pipe) + / dqm->dev->kfd->shared_resources.num_pipe_per_mec; - if (!test_bit(i, dqm->dev->shared_resources.cp_queue_bitmap)) + if (!test_bit(i, dqm->dev->kfd->shared_resources.cp_queue_bitmap)) continue; /* only acquire queues from the first MEC */ @@ -1475,9 +1615,13 @@ static int initialize_cpsch(struct device_queue_manager *dqm) dqm->gws_queue_count = 0; dqm->active_runlist = false; INIT_WORK(&dqm->hw_exception_work, kfd_process_hw_exception); + dqm->trap_debug_vmid = 0; init_sdma_bitmaps(dqm); + if (dqm->dev->kfd2kgd->get_iq_wait_times) + dqm->dev->kfd2kgd->get_iq_wait_times(dqm->dev->adev, + &dqm->wait_times); return 0; } @@ -1489,7 +1633,7 @@ static int start_cpsch(struct device_queue_manager *dqm) dqm_lock(dqm); - if (!dqm->dev->shared_resources.enable_mes) { + if (!dqm->dev->kfd->shared_resources.enable_mes) { retval = pm_init(&dqm->packet_mgr, dqm); if (retval) goto fail_packet_manager_init; @@ -1516,14 +1660,15 @@ static int start_cpsch(struct device_queue_manager *dqm) dqm->is_hws_hang = false; dqm->is_resetting = false; dqm->sched_running = true; - if (!dqm->dev->shared_resources.enable_mes) - execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0); + + if (!dqm->dev->kfd->shared_resources.enable_mes) + execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD); dqm_unlock(dqm); return 0; fail_allocate_vidmem: fail_set_sched_resources: - if (!dqm->dev->shared_resources.enable_mes) + if (!dqm->dev->kfd->shared_resources.enable_mes) pm_uninit(&dqm->packet_mgr, false); fail_packet_manager_init: dqm_unlock(dqm); @@ -1541,8 +1686,8 @@ static int stop_cpsch(struct device_queue_manager *dqm) } if (!dqm->is_hws_hang) { - if (!dqm->dev->shared_resources.enable_mes) - unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0, false); + if (!dqm->dev->kfd->shared_resources.enable_mes) + unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD, false); else remove_all_queues_mes(dqm); } @@ -1550,11 +1695,11 @@ static int stop_cpsch(struct device_queue_manager *dqm) hanging = dqm->is_hws_hang || dqm->is_resetting; dqm->sched_running = false; - if (!dqm->dev->shared_resources.enable_mes) + if (!dqm->dev->kfd->shared_resources.enable_mes) pm_release_ib(&dqm->packet_mgr); kfd_gtt_sa_free(dqm->dev, dqm->fence_mem); - if (!dqm->dev->shared_resources.enable_mes) + if (!dqm->dev->kfd->shared_resources.enable_mes) pm_uninit(&dqm->packet_mgr, hanging); dqm_unlock(dqm); @@ -1584,7 +1729,8 @@ static int create_kernel_queue_cpsch(struct device_queue_manager *dqm, list_add(&kq->list, &qpd->priv_queue_list); increment_queue_count(dqm, qpd, kq->queue); qpd->is_debug = true; - execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0); + execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, + USE_DEFAULT_GRACE_PERIOD); dqm_unlock(dqm); return 0; @@ -1598,7 +1744,8 @@ static void destroy_kernel_queue_cpsch(struct device_queue_manager *dqm, list_del(&kq->list); decrement_queue_count(dqm, qpd, kq->queue); qpd->is_debug = false; - execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0); + execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0, + USE_DEFAULT_GRACE_PERIOD); /* * Unconditionally decrement this counter, regardless of the queue's * type. @@ -1658,6 +1805,9 @@ static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q, * updates the is_evicted flag but is a no-op otherwise. */ q->properties.is_evicted = !!qpd->evicted; + q->properties.is_dbg_wa = qpd->pqm->process->debug_trap_enabled && + KFD_GC_VERSION(q->device) >= IP_VERSION(11, 0, 0) && + KFD_GC_VERSION(q->device) <= IP_VERSION(11, 0, 3); if (qd) mqd_mgr->restore_mqd(mqd_mgr, &q->mqd, q->mqd_mem_obj, &q->gart_mqd_addr, @@ -1673,9 +1823,9 @@ static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q, if (q->properties.is_active) { increment_queue_count(dqm, qpd, q); - if (!dqm->dev->shared_resources.enable_mes) + if (!dqm->dev->kfd->shared_resources.enable_mes) retval = execute_queues_cpsch(dqm, - KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0); + KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD); else retval = add_queue_mes(dqm, q, qpd); if (retval) @@ -1764,7 +1914,9 @@ static int map_queues_cpsch(struct device_queue_manager *dqm) /* dqm->lock mutex has to be locked before calling this function */ static int unmap_queues_cpsch(struct device_queue_manager *dqm, enum kfd_unmap_queues_filter filter, - uint32_t filter_param, bool reset) + uint32_t filter_param, + uint32_t grace_period, + bool reset) { int retval = 0; struct mqd_manager *mqd_mgr; @@ -1776,6 +1928,12 @@ static int unmap_queues_cpsch(struct device_queue_manager *dqm, if (!dqm->active_runlist) return retval; + if (grace_period != USE_DEFAULT_GRACE_PERIOD) { + retval = pm_update_grace_period(&dqm->packet_mgr, grace_period); + if (retval) + return retval; + } + retval = pm_send_unmap_queue(&dqm->packet_mgr, filter, filter_param, reset); if (retval) return retval; @@ -1808,6 +1966,13 @@ static int unmap_queues_cpsch(struct device_queue_manager *dqm, return -ETIME; } + /* We need to reset the grace period value for this device */ + if (grace_period != USE_DEFAULT_GRACE_PERIOD) { + if (pm_update_grace_period(&dqm->packet_mgr, + USE_DEFAULT_GRACE_PERIOD)) + pr_err("Failed to reset grace period\n"); + } + pm_release_ib(&dqm->packet_mgr); dqm->active_runlist = false; @@ -1823,7 +1988,7 @@ static int reset_queues_cpsch(struct device_queue_manager *dqm, dqm_lock(dqm); retval = unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_BY_PASID, - pasid, true); + pasid, USE_DEFAULT_GRACE_PERIOD, true); dqm_unlock(dqm); return retval; @@ -1832,19 +1997,45 @@ static int reset_queues_cpsch(struct device_queue_manager *dqm, /* dqm->lock mutex has to be locked before calling this function */ static int execute_queues_cpsch(struct device_queue_manager *dqm, enum kfd_unmap_queues_filter filter, - uint32_t filter_param) + uint32_t filter_param, + uint32_t grace_period) { int retval; if (dqm->is_hws_hang) return -EIO; - retval = unmap_queues_cpsch(dqm, filter, filter_param, false); + retval = unmap_queues_cpsch(dqm, filter, filter_param, grace_period, false); if (retval) return retval; return map_queues_cpsch(dqm); } +static int wait_on_destroy_queue(struct device_queue_manager *dqm, + struct queue *q) +{ + struct kfd_process_device *pdd = kfd_get_process_device_data(q->device, + q->process); + int ret = 0; + + if (pdd->qpd.is_debug) + return ret; + + q->properties.is_being_destroyed = true; + + if (pdd->process->debug_trap_enabled && q->properties.is_suspended) { + dqm_unlock(dqm); + mutex_unlock(&q->process->mutex); + ret = wait_event_interruptible(dqm->destroy_wait, + !q->properties.is_suspended); + + mutex_lock(&q->process->mutex); + dqm_lock(dqm); + } + + return ret; +} + static int destroy_queue_cpsch(struct device_queue_manager *dqm, struct qcm_process_device *qpd, struct queue *q) @@ -1864,11 +2055,16 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm, q->properties.queue_id); } - retval = 0; - /* remove queue from list to prevent rescheduling after preemption */ dqm_lock(dqm); + retval = wait_on_destroy_queue(dqm, q); + + if (retval) { + dqm_unlock(dqm); + return retval; + } + if (qpd->is_debug) { /* * error, currently we do not allow to destroy a queue @@ -1893,10 +2089,11 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm, list_del(&q->list); qpd->queue_count--; if (q->properties.is_active) { - if (!dqm->dev->shared_resources.enable_mes) { - decrement_queue_count(dqm, qpd, q); + decrement_queue_count(dqm, qpd, q); + if (!dqm->dev->kfd->shared_resources.enable_mes) { retval = execute_queues_cpsch(dqm, - KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0); + KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, + USE_DEFAULT_GRACE_PERIOD); if (retval == -ETIME) qpd->reset_wavefronts = true; } else { @@ -1914,7 +2111,14 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm, dqm_unlock(dqm); - /* Do free_mqd after dqm_unlock(dqm) to avoid circular locking */ + /* + * Do free_mqd and raise delete event after dqm_unlock(dqm) to avoid + * circular locking + */ + kfd_dbg_ev_raise(KFD_EC_MASK(EC_DEVICE_QUEUE_DELETE), + qpd->pqm->process, q->device, + -1, false, NULL, 0); + mqd_mgr->free_mqd(mqd_mgr, q->mqd, q->mqd_mem_obj); return retval; @@ -2056,7 +2260,7 @@ static int get_wave_state(struct device_queue_manager *dqm, mqd_mgr = dqm->mqd_mgrs[KFD_MQD_TYPE_CP]; if (q->properties.type != KFD_QUEUE_TYPE_COMPUTE || - q->properties.is_active || !q->device->cwsr_enabled || + q->properties.is_active || !q->device->kfd->cwsr_enabled || !mqd_mgr->get_wave_state) { dqm_unlock(dqm); return -EINVAL; @@ -2069,8 +2273,8 @@ static int get_wave_state(struct device_queue_manager *dqm, * and the queue should be protected against destruction by the process * lock. */ - return mqd_mgr->get_wave_state(mqd_mgr, q->mqd, ctl_stack, - ctl_stack_used_size, save_area_used_size); + return mqd_mgr->get_wave_state(mqd_mgr, q->mqd, &q->properties, + ctl_stack, ctl_stack_used_size, save_area_used_size); } static void get_queue_checkpoint_info(struct device_queue_manager *dqm, @@ -2105,7 +2309,7 @@ static int checkpoint_mqd(struct device_queue_manager *dqm, dqm_lock(dqm); - if (q->properties.is_active || !q->device->cwsr_enabled) { + if (q->properties.is_active || !q->device->kfd->cwsr_enabled) { r = -EINVAL; goto dqm_unlock; } @@ -2158,7 +2362,7 @@ static int process_termination_cpsch(struct device_queue_manager *dqm, if (q->properties.is_active) { decrement_queue_count(dqm, qpd, q); - if (dqm->dev->shared_resources.enable_mes) { + if (dqm->dev->kfd->shared_resources.enable_mes) { retval = remove_queue_mes(dqm, q, qpd); if (retval) pr_err("Failed to remove queue %d\n", @@ -2180,8 +2384,8 @@ static int process_termination_cpsch(struct device_queue_manager *dqm, } } - if (!dqm->dev->shared_resources.enable_mes) - retval = execute_queues_cpsch(dqm, filter, 0); + if (!dqm->dev->kfd->shared_resources.enable_mes) + retval = execute_queues_cpsch(dqm, filter, 0, USE_DEFAULT_GRACE_PERIOD); if ((!dqm->is_hws_hang) && (retval || qpd->reset_wavefronts)) { pr_warn("Resetting wave fronts (cpsch) on dev %p\n", dqm->dev); @@ -2242,12 +2446,13 @@ out_free: static int allocate_hiq_sdma_mqd(struct device_queue_manager *dqm) { int retval; - struct kfd_dev *dev = dqm->dev; + struct kfd_node *dev = dqm->dev; struct kfd_mem_obj *mem_obj = &dqm->hiq_sdma_mqd; uint32_t size = dqm->mqd_mgrs[KFD_MQD_TYPE_SDMA]->mqd_size * get_num_all_sdma_engines(dqm) * - dev->device_info.num_sdma_queues_per_engine + - dqm->mqd_mgrs[KFD_MQD_TYPE_HIQ]->mqd_size; + dev->kfd->device_info.num_sdma_queues_per_engine + + (dqm->mqd_mgrs[KFD_MQD_TYPE_HIQ]->mqd_size * + NUM_XCC(dqm->dev->xcc_mask)); retval = amdgpu_amdkfd_alloc_gtt_mem(dev->adev, size, &(mem_obj->gtt_mem), &(mem_obj->gpu_addr), @@ -2256,7 +2461,7 @@ static int allocate_hiq_sdma_mqd(struct device_queue_manager *dqm) return retval; } -struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev) +struct device_queue_manager *device_queue_manager_init(struct kfd_node *dev) { struct device_queue_manager *dqm; @@ -2373,20 +2578,22 @@ struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev) if (init_mqd_managers(dqm)) goto out_free; - if (!dev->shared_resources.enable_mes && allocate_hiq_sdma_mqd(dqm)) { + if (!dev->kfd->shared_resources.enable_mes && allocate_hiq_sdma_mqd(dqm)) { pr_err("Failed to allocate hiq sdma mqd trunk buffer\n"); goto out_free; } - if (!dqm->ops.initialize(dqm)) + if (!dqm->ops.initialize(dqm)) { + init_waitqueue_head(&dqm->destroy_wait); return dqm; + } out_free: kfree(dqm); return NULL; } -static void deallocate_hiq_sdma_mqd(struct kfd_dev *dev, +static void deallocate_hiq_sdma_mqd(struct kfd_node *dev, struct kfd_mem_obj *mqd) { WARN(!mqd, "No hiq sdma mqd trunk to free"); @@ -2396,8 +2603,9 @@ static void deallocate_hiq_sdma_mqd(struct kfd_dev *dev, void device_queue_manager_uninit(struct device_queue_manager *dqm) { + dqm->ops.stop(dqm); dqm->ops.uninitialize(dqm); - if (!dqm->dev->shared_resources.enable_mes) + if (!dqm->dev->kfd->shared_resources.enable_mes) deallocate_hiq_sdma_mqd(dqm->dev, &dqm->hiq_sdma_mqd); kfree(dqm); } @@ -2426,6 +2634,498 @@ static void kfd_process_hw_exception(struct work_struct *work) amdgpu_amdkfd_gpu_reset(dqm->dev->adev); } +int reserve_debug_trap_vmid(struct device_queue_manager *dqm, + struct qcm_process_device *qpd) +{ + int r; + int updated_vmid_mask; + + if (dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) { + pr_err("Unsupported on sched_policy: %i\n", dqm->sched_policy); + return -EINVAL; + } + + dqm_lock(dqm); + + if (dqm->trap_debug_vmid != 0) { + pr_err("Trap debug id already reserved\n"); + r = -EBUSY; + goto out_unlock; + } + + r = unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0, + USE_DEFAULT_GRACE_PERIOD, false); + if (r) + goto out_unlock; + + updated_vmid_mask = dqm->dev->kfd->shared_resources.compute_vmid_bitmap; + updated_vmid_mask &= ~(1 << dqm->dev->vm_info.last_vmid_kfd); + + dqm->dev->kfd->shared_resources.compute_vmid_bitmap = updated_vmid_mask; + dqm->trap_debug_vmid = dqm->dev->vm_info.last_vmid_kfd; + r = set_sched_resources(dqm); + if (r) + goto out_unlock; + + r = map_queues_cpsch(dqm); + if (r) + goto out_unlock; + + pr_debug("Reserved VMID for trap debug: %i\n", dqm->trap_debug_vmid); + +out_unlock: + dqm_unlock(dqm); + return r; +} + +/* + * Releases vmid for the trap debugger + */ +int release_debug_trap_vmid(struct device_queue_manager *dqm, + struct qcm_process_device *qpd) +{ + int r; + int updated_vmid_mask; + uint32_t trap_debug_vmid; + + if (dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) { + pr_err("Unsupported on sched_policy: %i\n", dqm->sched_policy); + return -EINVAL; + } + + dqm_lock(dqm); + trap_debug_vmid = dqm->trap_debug_vmid; + if (dqm->trap_debug_vmid == 0) { + pr_err("Trap debug id is not reserved\n"); + r = -EINVAL; + goto out_unlock; + } + + r = unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0, + USE_DEFAULT_GRACE_PERIOD, false); + if (r) + goto out_unlock; + + updated_vmid_mask = dqm->dev->kfd->shared_resources.compute_vmid_bitmap; + updated_vmid_mask |= (1 << dqm->dev->vm_info.last_vmid_kfd); + + dqm->dev->kfd->shared_resources.compute_vmid_bitmap = updated_vmid_mask; + dqm->trap_debug_vmid = 0; + r = set_sched_resources(dqm); + if (r) + goto out_unlock; + + r = map_queues_cpsch(dqm); + if (r) + goto out_unlock; + + pr_debug("Released VMID for trap debug: %i\n", trap_debug_vmid); + +out_unlock: + dqm_unlock(dqm); + return r; +} + +#define QUEUE_NOT_FOUND -1 +/* invalidate queue operation in array */ +static void q_array_invalidate(uint32_t num_queues, uint32_t *queue_ids) +{ + int i; + + for (i = 0; i < num_queues; i++) + queue_ids[i] |= KFD_DBG_QUEUE_INVALID_MASK; +} + +/* find queue index in array */ +static int q_array_get_index(unsigned int queue_id, + uint32_t num_queues, + uint32_t *queue_ids) +{ + int i; + + for (i = 0; i < num_queues; i++) + if (queue_id == (queue_ids[i] & ~KFD_DBG_QUEUE_INVALID_MASK)) + return i; + + return QUEUE_NOT_FOUND; +} + +struct copy_context_work_handler_workarea { + struct work_struct copy_context_work; + struct kfd_process *p; +}; + +static void copy_context_work_handler (struct work_struct *work) +{ + struct copy_context_work_handler_workarea *workarea; + struct mqd_manager *mqd_mgr; + struct queue *q; + struct mm_struct *mm; + struct kfd_process *p; + uint32_t tmp_ctl_stack_used_size, tmp_save_area_used_size; + int i; + + workarea = container_of(work, + struct copy_context_work_handler_workarea, + copy_context_work); + + p = workarea->p; + mm = get_task_mm(p->lead_thread); + + if (!mm) + return; + + kthread_use_mm(mm); + for (i = 0; i < p->n_pdds; i++) { + struct kfd_process_device *pdd = p->pdds[i]; + struct device_queue_manager *dqm = pdd->dev->dqm; + struct qcm_process_device *qpd = &pdd->qpd; + + list_for_each_entry(q, &qpd->queues_list, list) { + mqd_mgr = dqm->mqd_mgrs[KFD_MQD_TYPE_CP]; + + /* We ignore the return value from get_wave_state + * because + * i) right now, it always returns 0, and + * ii) if we hit an error, we would continue to the + * next queue anyway. + */ + mqd_mgr->get_wave_state(mqd_mgr, + q->mqd, + &q->properties, + (void __user *) q->properties.ctx_save_restore_area_address, + &tmp_ctl_stack_used_size, + &tmp_save_area_used_size); + } + } + kthread_unuse_mm(mm); + mmput(mm); +} + +static uint32_t *get_queue_ids(uint32_t num_queues, uint32_t *usr_queue_id_array) +{ + size_t array_size = num_queues * sizeof(uint32_t); + uint32_t *queue_ids = NULL; + + if (!usr_queue_id_array) + return NULL; + + queue_ids = kzalloc(array_size, GFP_KERNEL); + if (!queue_ids) + return ERR_PTR(-ENOMEM); + + if (copy_from_user(queue_ids, usr_queue_id_array, array_size)) + return ERR_PTR(-EFAULT); + + return queue_ids; +} + +int resume_queues(struct kfd_process *p, + uint32_t num_queues, + uint32_t *usr_queue_id_array) +{ + uint32_t *queue_ids = NULL; + int total_resumed = 0; + int i; + + if (usr_queue_id_array) { + queue_ids = get_queue_ids(num_queues, usr_queue_id_array); + + if (IS_ERR(queue_ids)) + return PTR_ERR(queue_ids); + + /* mask all queues as invalid. unmask per successful request */ + q_array_invalidate(num_queues, queue_ids); + } + + for (i = 0; i < p->n_pdds; i++) { + struct kfd_process_device *pdd = p->pdds[i]; + struct device_queue_manager *dqm = pdd->dev->dqm; + struct qcm_process_device *qpd = &pdd->qpd; + struct queue *q; + int r, per_device_resumed = 0; + + dqm_lock(dqm); + + /* unmask queues that resume or already resumed as valid */ + list_for_each_entry(q, &qpd->queues_list, list) { + int q_idx = QUEUE_NOT_FOUND; + + if (queue_ids) + q_idx = q_array_get_index( + q->properties.queue_id, + num_queues, + queue_ids); + + if (!queue_ids || q_idx != QUEUE_NOT_FOUND) { + int err = resume_single_queue(dqm, &pdd->qpd, q); + + if (queue_ids) { + if (!err) { + queue_ids[q_idx] &= + ~KFD_DBG_QUEUE_INVALID_MASK; + } else { + queue_ids[q_idx] |= + KFD_DBG_QUEUE_ERROR_MASK; + break; + } + } + + if (dqm->dev->kfd->shared_resources.enable_mes) { + wake_up_all(&dqm->destroy_wait); + if (!err) + total_resumed++; + } else { + per_device_resumed++; + } + } + } + + if (!per_device_resumed) { + dqm_unlock(dqm); + continue; + } + + r = execute_queues_cpsch(dqm, + KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, + 0, + USE_DEFAULT_GRACE_PERIOD); + if (r) { + pr_err("Failed to resume process queues\n"); + if (queue_ids) { + list_for_each_entry(q, &qpd->queues_list, list) { + int q_idx = q_array_get_index( + q->properties.queue_id, + num_queues, + queue_ids); + + /* mask queue as error on resume fail */ + if (q_idx != QUEUE_NOT_FOUND) + queue_ids[q_idx] |= + KFD_DBG_QUEUE_ERROR_MASK; + } + } + } else { + wake_up_all(&dqm->destroy_wait); + total_resumed += per_device_resumed; + } + + dqm_unlock(dqm); + } + + if (queue_ids) { + if (copy_to_user((void __user *)usr_queue_id_array, queue_ids, + num_queues * sizeof(uint32_t))) + pr_err("copy_to_user failed on queue resume\n"); + + kfree(queue_ids); + } + + return total_resumed; +} + +int suspend_queues(struct kfd_process *p, + uint32_t num_queues, + uint32_t grace_period, + uint64_t exception_clear_mask, + uint32_t *usr_queue_id_array) +{ + uint32_t *queue_ids = get_queue_ids(num_queues, usr_queue_id_array); + int total_suspended = 0; + int i; + + if (IS_ERR(queue_ids)) + return PTR_ERR(queue_ids); + + /* mask all queues as invalid. umask on successful request */ + q_array_invalidate(num_queues, queue_ids); + + for (i = 0; i < p->n_pdds; i++) { + struct kfd_process_device *pdd = p->pdds[i]; + struct device_queue_manager *dqm = pdd->dev->dqm; + struct qcm_process_device *qpd = &pdd->qpd; + struct queue *q; + int r, per_device_suspended = 0; + + mutex_lock(&p->event_mutex); + dqm_lock(dqm); + + /* unmask queues that suspend or already suspended */ + list_for_each_entry(q, &qpd->queues_list, list) { + int q_idx = q_array_get_index(q->properties.queue_id, + num_queues, + queue_ids); + + if (q_idx != QUEUE_NOT_FOUND) { + int err = suspend_single_queue(dqm, pdd, q); + bool is_mes = dqm->dev->kfd->shared_resources.enable_mes; + + if (!err) { + queue_ids[q_idx] &= ~KFD_DBG_QUEUE_INVALID_MASK; + if (exception_clear_mask && is_mes) + q->properties.exception_status &= + ~exception_clear_mask; + + if (is_mes) + total_suspended++; + else + per_device_suspended++; + } else if (err != -EBUSY) { + r = err; + queue_ids[q_idx] |= KFD_DBG_QUEUE_ERROR_MASK; + break; + } + } + } + + if (!per_device_suspended) { + dqm_unlock(dqm); + mutex_unlock(&p->event_mutex); + if (total_suspended) + amdgpu_amdkfd_debug_mem_fence(dqm->dev->adev); + continue; + } + + r = execute_queues_cpsch(dqm, + KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, + grace_period); + + if (r) + pr_err("Failed to suspend process queues.\n"); + else + total_suspended += per_device_suspended; + + list_for_each_entry(q, &qpd->queues_list, list) { + int q_idx = q_array_get_index(q->properties.queue_id, + num_queues, queue_ids); + + if (q_idx == QUEUE_NOT_FOUND) + continue; + + /* mask queue as error on suspend fail */ + if (r) + queue_ids[q_idx] |= KFD_DBG_QUEUE_ERROR_MASK; + else if (exception_clear_mask) + q->properties.exception_status &= + ~exception_clear_mask; + } + + dqm_unlock(dqm); + mutex_unlock(&p->event_mutex); + amdgpu_device_flush_hdp(dqm->dev->adev, NULL); + } + + if (total_suspended) { + struct copy_context_work_handler_workarea copy_context_worker; + + INIT_WORK_ONSTACK( + ©_context_worker.copy_context_work, + copy_context_work_handler); + + copy_context_worker.p = p; + + schedule_work(©_context_worker.copy_context_work); + + + flush_work(©_context_worker.copy_context_work); + destroy_work_on_stack(©_context_worker.copy_context_work); + } + + if (copy_to_user((void __user *)usr_queue_id_array, queue_ids, + num_queues * sizeof(uint32_t))) + pr_err("copy_to_user failed on queue suspend\n"); + + kfree(queue_ids); + + return total_suspended; +} + +static uint32_t set_queue_type_for_user(struct queue_properties *q_props) +{ + switch (q_props->type) { + case KFD_QUEUE_TYPE_COMPUTE: + return q_props->format == KFD_QUEUE_FORMAT_PM4 + ? KFD_IOC_QUEUE_TYPE_COMPUTE + : KFD_IOC_QUEUE_TYPE_COMPUTE_AQL; + case KFD_QUEUE_TYPE_SDMA: + return KFD_IOC_QUEUE_TYPE_SDMA; + case KFD_QUEUE_TYPE_SDMA_XGMI: + return KFD_IOC_QUEUE_TYPE_SDMA_XGMI; + default: + WARN_ONCE(true, "queue type not recognized!"); + return 0xffffffff; + }; +} + +void set_queue_snapshot_entry(struct queue *q, + uint64_t exception_clear_mask, + struct kfd_queue_snapshot_entry *qss_entry) +{ + qss_entry->ring_base_address = q->properties.queue_address; + qss_entry->write_pointer_address = (uint64_t)q->properties.write_ptr; + qss_entry->read_pointer_address = (uint64_t)q->properties.read_ptr; + qss_entry->ctx_save_restore_address = + q->properties.ctx_save_restore_area_address; + qss_entry->ctx_save_restore_area_size = + q->properties.ctx_save_restore_area_size; + qss_entry->exception_status = q->properties.exception_status; + qss_entry->queue_id = q->properties.queue_id; + qss_entry->gpu_id = q->device->id; + qss_entry->ring_size = (uint32_t)q->properties.queue_size; + qss_entry->queue_type = set_queue_type_for_user(&q->properties); + q->properties.exception_status &= ~exception_clear_mask; +} + +int debug_lock_and_unmap(struct device_queue_manager *dqm) +{ + int r; + + if (dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) { + pr_err("Unsupported on sched_policy: %i\n", dqm->sched_policy); + return -EINVAL; + } + + if (!kfd_dbg_is_per_vmid_supported(dqm->dev)) + return 0; + + dqm_lock(dqm); + + r = unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0, 0, false); + if (r) + dqm_unlock(dqm); + + return r; +} + +int debug_map_and_unlock(struct device_queue_manager *dqm) +{ + int r; + + if (dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) { + pr_err("Unsupported on sched_policy: %i\n", dqm->sched_policy); + return -EINVAL; + } + + if (!kfd_dbg_is_per_vmid_supported(dqm->dev)) + return 0; + + r = map_queues_cpsch(dqm); + + dqm_unlock(dqm); + + return r; +} + +int debug_refresh_runlist(struct device_queue_manager *dqm) +{ + int r = debug_lock_and_unmap(dqm); + + if (r) + return r; + + return debug_map_and_unlock(dqm); +} + #if defined(CONFIG_DEBUG_FS) static void seq_reg_dump(struct seq_file *m, @@ -2452,52 +3152,66 @@ static void seq_reg_dump(struct seq_file *m, int dqm_debugfs_hqds(struct seq_file *m, void *data) { struct device_queue_manager *dqm = data; + uint32_t xcc_mask = dqm->dev->xcc_mask; uint32_t (*dump)[2], n_regs; int pipe, queue; - int r = 0; + int r = 0, xcc_id; + uint32_t sdma_engine_start; if (!dqm->sched_running) { seq_puts(m, " Device is stopped\n"); return 0; } - r = dqm->dev->kfd2kgd->hqd_dump(dqm->dev->adev, - KFD_CIK_HIQ_PIPE, KFD_CIK_HIQ_QUEUE, - &dump, &n_regs); - if (!r) { - seq_printf(m, " HIQ on MEC %d Pipe %d Queue %d\n", - KFD_CIK_HIQ_PIPE/get_pipes_per_mec(dqm)+1, - KFD_CIK_HIQ_PIPE%get_pipes_per_mec(dqm), - KFD_CIK_HIQ_QUEUE); - seq_reg_dump(m, dump, n_regs); + for_each_inst(xcc_id, xcc_mask) { + r = dqm->dev->kfd2kgd->hqd_dump(dqm->dev->adev, + KFD_CIK_HIQ_PIPE, + KFD_CIK_HIQ_QUEUE, &dump, + &n_regs, xcc_id); + if (!r) { + seq_printf( + m, + " Inst %d, HIQ on MEC %d Pipe %d Queue %d\n", + xcc_id, + KFD_CIK_HIQ_PIPE / get_pipes_per_mec(dqm) + 1, + KFD_CIK_HIQ_PIPE % get_pipes_per_mec(dqm), + KFD_CIK_HIQ_QUEUE); + seq_reg_dump(m, dump, n_regs); - kfree(dump); - } + kfree(dump); + } - for (pipe = 0; pipe < get_pipes_per_mec(dqm); pipe++) { - int pipe_offset = pipe * get_queues_per_pipe(dqm); + for (pipe = 0; pipe < get_pipes_per_mec(dqm); pipe++) { + int pipe_offset = pipe * get_queues_per_pipe(dqm); - for (queue = 0; queue < get_queues_per_pipe(dqm); queue++) { - if (!test_bit(pipe_offset + queue, - dqm->dev->shared_resources.cp_queue_bitmap)) - continue; + for (queue = 0; queue < get_queues_per_pipe(dqm); queue++) { + if (!test_bit(pipe_offset + queue, + dqm->dev->kfd->shared_resources.cp_queue_bitmap)) + continue; - r = dqm->dev->kfd2kgd->hqd_dump( - dqm->dev->adev, pipe, queue, &dump, &n_regs); - if (r) - break; + r = dqm->dev->kfd2kgd->hqd_dump(dqm->dev->adev, + pipe, queue, + &dump, &n_regs, + xcc_id); + if (r) + break; - seq_printf(m, " CP Pipe %d, Queue %d\n", - pipe, queue); - seq_reg_dump(m, dump, n_regs); + seq_printf(m, + " Inst %d, CP Pipe %d, Queue %d\n", + xcc_id, pipe, queue); + seq_reg_dump(m, dump, n_regs); - kfree(dump); + kfree(dump); + } } } - for (pipe = 0; pipe < get_num_all_sdma_engines(dqm); pipe++) { + sdma_engine_start = dqm->dev->node_id * get_num_all_sdma_engines(dqm); + for (pipe = sdma_engine_start; + pipe < (sdma_engine_start + get_num_all_sdma_engines(dqm)); + pipe++) { for (queue = 0; - queue < dqm->dev->device_info.num_sdma_queues_per_engine; + queue < dqm->dev->kfd->device_info.num_sdma_queues_per_engine; queue++) { r = dqm->dev->kfd2kgd->hqd_sdma_dump( dqm->dev->adev, pipe, queue, &dump, &n_regs); @@ -2526,7 +3240,8 @@ int dqm_debugfs_hang_hws(struct device_queue_manager *dqm) return r; } dqm->active_runlist = true; - r = execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0); + r = execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, + 0, USE_DEFAULT_GRACE_PERIOD); dqm_unlock(dqm); return r; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h index a537b9ef3e16..7dd4b177219d 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h @@ -37,6 +37,7 @@ #define KFD_MES_PROCESS_QUANTUM 100000 #define KFD_MES_GANG_QUANTUM 10000 +#define USE_DEFAULT_GRACE_PERIOD 0xffffffff struct device_process_node { struct qcm_process_device *qpd; @@ -207,7 +208,7 @@ struct device_queue_manager_asic_ops { struct queue *q, struct qcm_process_device *qpd); struct mqd_manager * (*mqd_manager_init)(enum KFD_MQD_TYPE type, - struct kfd_dev *dev); + struct kfd_node *dev); }; /** @@ -228,7 +229,7 @@ struct device_queue_manager { struct mqd_manager *mqd_mgrs[KFD_MQD_TYPE_MAX]; struct packet_manager packet_mgr; - struct kfd_dev *dev; + struct kfd_node *dev; struct mutex lock_hidden; /* use dqm_lock/unlock(dqm) */ struct list_head queues; unsigned int saved_flags; @@ -239,8 +240,8 @@ struct device_queue_manager { unsigned int total_queue_count; unsigned int next_pipe_to_allocate; unsigned int *allocated_queues; - uint64_t sdma_bitmap; - uint64_t xgmi_sdma_bitmap; + DECLARE_BITMAP(sdma_bitmap, KFD_MAX_SDMA_QUEUES); + DECLARE_BITMAP(xgmi_sdma_bitmap, KFD_MAX_SDMA_QUEUES); /* the pasid mapping for each kfd vmid */ uint16_t vmid_pasid[VMID_NUM]; uint64_t pipelines_addr; @@ -249,6 +250,7 @@ struct device_queue_manager { struct kfd_mem_obj *fence_mem; bool active_runlist; int sched_policy; + uint32_t trap_debug_vmid; /* hw exception */ bool is_hws_hang; @@ -256,6 +258,13 @@ struct device_queue_manager { struct work_struct hw_exception_work; struct kfd_mem_obj hiq_sdma_mqd; bool sched_running; + + /* used for GFX 9.4.3 only */ + uint32_t current_logical_xcc_start; + + uint32_t wait_times; + + wait_queue_head_t destroy_wait; }; void device_queue_manager_init_cik( @@ -279,6 +288,24 @@ unsigned int get_queues_per_pipe(struct device_queue_manager *dqm); unsigned int get_pipes_per_mec(struct device_queue_manager *dqm); unsigned int get_num_sdma_queues(struct device_queue_manager *dqm); unsigned int get_num_xgmi_sdma_queues(struct device_queue_manager *dqm); +int reserve_debug_trap_vmid(struct device_queue_manager *dqm, + struct qcm_process_device *qpd); +int release_debug_trap_vmid(struct device_queue_manager *dqm, + struct qcm_process_device *qpd); +int suspend_queues(struct kfd_process *p, + uint32_t num_queues, + uint32_t grace_period, + uint64_t exception_clear_mask, + uint32_t *usr_queue_id_array); +int resume_queues(struct kfd_process *p, + uint32_t num_queues, + uint32_t *usr_queue_id_array); +void set_queue_snapshot_entry(struct queue *q, + uint64_t exception_clear_mask, + struct kfd_queue_snapshot_entry *qss_entry); +int debug_lock_and_unmap(struct device_queue_manager *dqm); +int debug_map_and_unlock(struct device_queue_manager *dqm); +int debug_refresh_runlist(struct device_queue_manager *dqm); static inline unsigned int get_sh_mem_bases_32(struct kfd_process_device *pdd) { diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c index 8b2dd2670ab7..8af643388768 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c @@ -24,9 +24,7 @@ #include "kfd_device_queue_manager.h" #include "vega10_enum.h" -#include "gc/gc_9_0_offset.h" -#include "gc/gc_9_0_sh_mask.h" -#include "sdma0/sdma0_4_0_sh_mask.h" +#include "gc/gc_9_4_3_sh_mask.h" static int update_qpd_v9(struct device_queue_manager *dqm, struct qcm_process_device *qpd); @@ -62,9 +60,13 @@ static int update_qpd_v9(struct device_queue_manager *dqm, qpd->sh_mem_config = SH_MEM_ALIGNMENT_MODE_UNALIGNED << SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT; - if (dqm->dev->noretry && !dqm->dev->use_iommu_v2) + if (dqm->dev->kfd->noretry && !dqm->dev->kfd->use_iommu_v2) qpd->sh_mem_config |= 1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT; + if (KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 3)) + qpd->sh_mem_config |= + (1 << SH_MEM_CONFIG__F8_MODE__SHIFT); + qpd->sh_mem_ape1_limit = 0; qpd->sh_mem_ape1_base = 0; } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c b/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c index 38c9e1ca6691..6421b620388d 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c @@ -138,7 +138,7 @@ void kfd_doorbell_fini(struct kfd_dev *kfd) iounmap(kfd->doorbell_kernel_ptr); } -int kfd_doorbell_mmap(struct kfd_dev *dev, struct kfd_process *process, +int kfd_doorbell_mmap(struct kfd_node *dev, struct kfd_process *process, struct vm_area_struct *vma) { phys_addr_t address; @@ -148,7 +148,7 @@ int kfd_doorbell_mmap(struct kfd_dev *dev, struct kfd_process *process, * For simplicitly we only allow mapping of the entire doorbell * allocation of a single device & process. */ - if (vma->vm_end - vma->vm_start != kfd_doorbell_process_slice(dev)) + if (vma->vm_end - vma->vm_start != kfd_doorbell_process_slice(dev->kfd)) return -EINVAL; pdd = kfd_get_process_device_data(dev, process); @@ -170,13 +170,13 @@ int kfd_doorbell_mmap(struct kfd_dev *dev, struct kfd_process *process, " vm_flags == 0x%04lX\n" " size == 0x%04lX\n", (unsigned long long) vma->vm_start, address, vma->vm_flags, - kfd_doorbell_process_slice(dev)); + kfd_doorbell_process_slice(dev->kfd)); return io_remap_pfn_range(vma, vma->vm_start, address >> PAGE_SHIFT, - kfd_doorbell_process_slice(dev), + kfd_doorbell_process_slice(dev->kfd), vma->vm_page_prot); } @@ -278,14 +278,14 @@ uint64_t kfd_get_number_elems(struct kfd_dev *kfd) phys_addr_t kfd_get_process_doorbells(struct kfd_process_device *pdd) { if (!pdd->doorbell_index) { - int r = kfd_alloc_process_doorbells(pdd->dev, + int r = kfd_alloc_process_doorbells(pdd->dev->kfd, &pdd->doorbell_index); if (r < 0) return 0; } - return pdd->dev->doorbell_base + - pdd->doorbell_index * kfd_doorbell_process_slice(pdd->dev); + return pdd->dev->kfd->doorbell_base + + pdd->doorbell_index * kfd_doorbell_process_slice(pdd->dev->kfd); } int kfd_alloc_process_doorbells(struct kfd_dev *kfd, unsigned int *doorbell_index) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_events.c b/drivers/gpu/drm/amd/amdkfd/kfd_events.c index c894cf8f7c50..8081a9408006 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_events.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_events.c @@ -41,6 +41,7 @@ struct kfd_event_waiter { wait_queue_entry_t wait; struct kfd_event *event; /* Event to wait for */ bool activated; /* Becomes true when event is signaled */ + bool event_age_enabled; /* set to true when last_event_age is non-zero */ }; /* @@ -348,7 +349,7 @@ static int kfd_event_page_set(struct kfd_process *p, void *kernel_address, int kfd_kmap_event_page(struct kfd_process *p, uint64_t event_page_offset) { - struct kfd_dev *kfd; + struct kfd_node *kfd; struct kfd_process_device *pdd; void *mem, *kern_addr; uint64_t size; @@ -431,6 +432,7 @@ int kfd_event_create(struct file *devkfd, struct kfd_process *p, if (!ret) { *event_id = ev->event_id; *event_trigger_data = ev->event_id; + ev->event_age = 1; } else { kfree(ev); } @@ -629,6 +631,11 @@ static void set_event(struct kfd_event *ev) * updating the wait queues in kfd_wait_on_events. */ ev->signaled = !ev->auto_reset || !waitqueue_active(&ev->wq); + if (!(++ev->event_age)) { + /* Never wrap back to reserved/default event age 0/1 */ + ev->event_age = 2; + WARN_ONCE(1, "event_age wrap back!"); + } list_for_each_entry(waiter, &ev->wq.head, wait.entry) WRITE_ONCE(waiter->activated, true); @@ -791,9 +798,9 @@ static struct kfd_event_waiter *alloc_event_waiters(uint32_t num_events) static int init_event_waiter(struct kfd_process *p, struct kfd_event_waiter *waiter, - uint32_t event_id) + struct kfd_event_data *event_data) { - struct kfd_event *ev = lookup_event_by_id(p, event_id); + struct kfd_event *ev = lookup_event_by_id(p, event_data->event_id); if (!ev) return -EINVAL; @@ -802,6 +809,15 @@ static int init_event_waiter(struct kfd_process *p, waiter->event = ev; waiter->activated = ev->signaled; ev->signaled = ev->signaled && !ev->auto_reset; + + /* last_event_age = 0 reserved for backward compatible */ + if (waiter->event->type == KFD_EVENT_TYPE_SIGNAL && + event_data->signal_event_data.last_event_age) { + waiter->event_age_enabled = true; + if (ev->event_age != event_data->signal_event_data.last_event_age) + waiter->activated = true; + } + if (!waiter->activated) add_wait_queue(&ev->wq, &waiter->wait); spin_unlock(&ev->lock); @@ -849,22 +865,29 @@ static int copy_signaled_event_data(uint32_t num_events, struct kfd_event_waiter *event_waiters, struct kfd_event_data __user *data) { - struct kfd_hsa_memory_exception_data *src; - struct kfd_hsa_memory_exception_data __user *dst; + void *src; + void __user *dst; struct kfd_event_waiter *waiter; struct kfd_event *event; - uint32_t i; + uint32_t i, size = 0; for (i = 0; i < num_events; i++) { waiter = &event_waiters[i]; event = waiter->event; if (!event) return -EINVAL; /* event was destroyed */ - if (waiter->activated && event->type == KFD_EVENT_TYPE_MEMORY) { - dst = &data[i].memory_exception_data; - src = &event->memory_exception_data; - if (copy_to_user(dst, src, - sizeof(struct kfd_hsa_memory_exception_data))) + if (waiter->activated) { + if (event->type == KFD_EVENT_TYPE_MEMORY) { + dst = &data[i].memory_exception_data; + src = &event->memory_exception_data; + size = sizeof(struct kfd_hsa_memory_exception_data); + } else if (event->type == KFD_EVENT_TYPE_SIGNAL && + waiter->event_age_enabled) { + dst = &data[i].signal_event_data.last_event_age; + src = &event->event_age; + size = sizeof(u64); + } + if (size && copy_to_user(dst, src, size)) return -EFAULT; } } @@ -942,8 +965,7 @@ int kfd_wait_on_events(struct kfd_process *p, goto out_unlock; } - ret = init_event_waiter(p, &event_waiters[i], - event_data.event_id); + ret = init_event_waiter(p, &event_waiters[i], &event_data); if (ret) goto out_unlock; } @@ -1125,7 +1147,7 @@ static void lookup_events_by_type_and_signal(struct kfd_process *p, } #ifdef KFD_SUPPORT_IOMMU_V2 -void kfd_signal_iommu_event(struct kfd_dev *dev, u32 pasid, +void kfd_signal_iommu_event(struct kfd_node *dev, u32 pasid, unsigned long address, bool is_write_requested, bool is_execute_requested) { @@ -1221,8 +1243,9 @@ void kfd_signal_hw_exception_event(u32 pasid) kfd_unref_process(p); } -void kfd_signal_vm_fault_event(struct kfd_dev *dev, u32 pasid, - struct kfd_vm_fault_info *info) +void kfd_signal_vm_fault_event(struct kfd_node *dev, u32 pasid, + struct kfd_vm_fault_info *info, + struct kfd_hsa_memory_exception_data *data) { struct kfd_event *ev; uint32_t id; @@ -1239,19 +1262,24 @@ void kfd_signal_vm_fault_event(struct kfd_dev *dev, u32 pasid, return; } - memset(&memory_exception_data, 0, sizeof(memory_exception_data)); - memory_exception_data.gpu_id = user_gpu_id; - memory_exception_data.failure.imprecise = true; - /* Set failure reason */ - if (info) { - memory_exception_data.va = (info->page_addr) << PAGE_SHIFT; - memory_exception_data.failure.NotPresent = - info->prot_valid ? 1 : 0; - memory_exception_data.failure.NoExecute = - info->prot_exec ? 1 : 0; - memory_exception_data.failure.ReadOnly = - info->prot_write ? 1 : 0; - memory_exception_data.failure.imprecise = 0; + /* SoC15 chips and onwards will pass in data from now on. */ + if (!data) { + memset(&memory_exception_data, 0, sizeof(memory_exception_data)); + memory_exception_data.gpu_id = user_gpu_id; + memory_exception_data.failure.imprecise = true; + + /* Set failure reason */ + if (info) { + memory_exception_data.va = (info->page_addr) << + PAGE_SHIFT; + memory_exception_data.failure.NotPresent = + info->prot_valid ? 1 : 0; + memory_exception_data.failure.NoExecute = + info->prot_exec ? 1 : 0; + memory_exception_data.failure.ReadOnly = + info->prot_write ? 1 : 0; + memory_exception_data.failure.imprecise = 0; + } } rcu_read_lock(); @@ -1260,7 +1288,8 @@ void kfd_signal_vm_fault_event(struct kfd_dev *dev, u32 pasid, idr_for_each_entry_continue(&p->event_idr, ev, id) if (ev->type == KFD_EVENT_TYPE_MEMORY) { spin_lock(&ev->lock); - ev->memory_exception_data = memory_exception_data; + ev->memory_exception_data = data ? *data : + memory_exception_data; set_event(ev); spin_unlock(&ev->lock); } @@ -1269,7 +1298,7 @@ void kfd_signal_vm_fault_event(struct kfd_dev *dev, u32 pasid, kfd_unref_process(p); } -void kfd_signal_reset_event(struct kfd_dev *dev) +void kfd_signal_reset_event(struct kfd_node *dev) { struct kfd_hsa_hw_exception_data hw_exception_data; struct kfd_hsa_memory_exception_data memory_exception_data; @@ -1325,7 +1354,7 @@ void kfd_signal_reset_event(struct kfd_dev *dev) srcu_read_unlock(&kfd_processes_srcu, idx); } -void kfd_signal_poison_consumed_event(struct kfd_dev *dev, u32 pasid) +void kfd_signal_poison_consumed_event(struct kfd_node *dev, u32 pasid) { struct kfd_process *p = kfd_lookup_process_by_pasid(pasid); struct kfd_hsa_memory_exception_data memory_exception_data; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_events.h b/drivers/gpu/drm/amd/amdkfd/kfd_events.h index 1c62c8dd6460..52ccfd397c2b 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_events.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_events.h @@ -53,6 +53,7 @@ struct signal_page; struct kfd_event { u32 event_id; + u64 event_age; bool signaled; bool auto_reset; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c index 8aebe408c544..da2ca00d79e5 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c @@ -322,21 +322,21 @@ static void kfd_init_apertures_vi(struct kfd_process_device *pdd, uint8_t id) pdd->lds_base = MAKE_LDS_APP_BASE_VI(); pdd->lds_limit = MAKE_LDS_APP_LIMIT(pdd->lds_base); - if (!pdd->dev->use_iommu_v2) { + if (!pdd->dev->kfd->use_iommu_v2) { /* dGPUs: SVM aperture starting at 0 * with small reserved space for kernel. * Set them to CANONICAL addresses. */ pdd->gpuvm_base = SVM_USER_BASE; pdd->gpuvm_limit = - pdd->dev->shared_resources.gpuvm_size - 1; + pdd->dev->kfd->shared_resources.gpuvm_size - 1; } else { /* set them to non CANONICAL addresses, and no SVM is * allocated. */ pdd->gpuvm_base = MAKE_GPUVM_APP_BASE_VI(id + 1); pdd->gpuvm_limit = MAKE_GPUVM_APP_LIMIT(pdd->gpuvm_base, - pdd->dev->shared_resources.gpuvm_size); + pdd->dev->kfd->shared_resources.gpuvm_size); } pdd->scratch_base = MAKE_SCRATCH_APP_BASE_VI(); @@ -356,7 +356,7 @@ static void kfd_init_apertures_v9(struct kfd_process_device *pdd, uint8_t id) */ pdd->gpuvm_base = SVM_USER_BASE; pdd->gpuvm_limit = - pdd->dev->shared_resources.gpuvm_size - 1; + pdd->dev->kfd->shared_resources.gpuvm_size - 1; pdd->scratch_base = MAKE_SCRATCH_APP_BASE_V9(); pdd->scratch_limit = MAKE_SCRATCH_APP_LIMIT(pdd->scratch_base); @@ -365,7 +365,7 @@ static void kfd_init_apertures_v9(struct kfd_process_device *pdd, uint8_t id) int kfd_init_apertures(struct kfd_process *process) { uint8_t id = 0; - struct kfd_dev *dev; + struct kfd_node *dev; struct kfd_process_device *pdd; /*Iterating over all devices*/ @@ -417,7 +417,7 @@ int kfd_init_apertures(struct kfd_process *process) } } - if (!dev->use_iommu_v2) { + if (!dev->kfd->use_iommu_v2) { /* dGPUs: the reserved space for kernel * before SVM */ diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c new file mode 100644 index 000000000000..c7991e07b6be --- /dev/null +++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c @@ -0,0 +1,405 @@ +/* + * Copyright 2023 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ + +#include "kfd_events.h" +#include "kfd_debug.h" +#include "soc15_int.h" +#include "kfd_device_queue_manager.h" + +/* + * GFX10 SQ Interrupts + * + * There are 3 encoding types of interrupts sourced from SQ sent as a 44-bit + * packet to the Interrupt Handler: + * Auto - Generated by the SQG (various cmd overflows, timestamps etc) + * Wave - Generated by S_SENDMSG through a shader program + * Error - HW generated errors (Illegal instructions, Memviols, EDC etc) + * + * The 44-bit packet is mapped as {context_id1[7:0],context_id0[31:0]} plus + * 4-bits for VMID (SOC15_VMID_FROM_IH_ENTRY) as such: + * + * - context_id1[7:6] + * Encoding type (0 = Auto, 1 = Wave, 2 = Error) + * + * - context_id0[24] + * PRIV bit indicates that Wave S_SEND or error occurred within trap + * + * - context_id0[22:0] + * 23-bit data with the following layout per encoding type: + * Auto - only context_id0[8:0] is used, which reports various interrupts + * generated by SQG. The rest is 0. + * Wave - user data sent from m0 via S_SENDMSG + * Error - Error type (context_id0[22:19]), Error Details (rest of bits) + * + * The other context_id bits show coordinates (SE/SH/CU/SIMD/WGP) for wave + * S_SENDMSG and Errors. These are 0 for Auto. + */ + +enum SQ_INTERRUPT_WORD_ENCODING { + SQ_INTERRUPT_WORD_ENCODING_AUTO = 0x0, + SQ_INTERRUPT_WORD_ENCODING_INST, + SQ_INTERRUPT_WORD_ENCODING_ERROR, +}; + +enum SQ_INTERRUPT_ERROR_TYPE { + SQ_INTERRUPT_ERROR_TYPE_EDC_FUE = 0x0, + SQ_INTERRUPT_ERROR_TYPE_ILLEGAL_INST, + SQ_INTERRUPT_ERROR_TYPE_MEMVIOL, + SQ_INTERRUPT_ERROR_TYPE_EDC_FED, +}; + +/* SQ_INTERRUPT_WORD_AUTO_CTXID */ +#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE__SHIFT 0 +#define SQ_INTERRUPT_WORD_AUTO_CTXID0__WLT__SHIFT 1 +#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_BUF0_FULL__SHIFT 2 +#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_BUF1_FULL__SHIFT 3 +#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_UTC_ERROR__SHIFT 7 +#define SQ_INTERRUPT_WORD_AUTO_CTXID1__SE_ID__SHIFT 4 +#define SQ_INTERRUPT_WORD_AUTO_CTXID1__ENCODING__SHIFT 6 + +#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_MASK 0x00000001 +#define SQ_INTERRUPT_WORD_AUTO_CTXID0__WLT_MASK 0x00000002 +#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_BUF0_FULL_MASK 0x00000004 +#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_BUF1_FULL_MASK 0x00000008 +#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_UTC_ERROR_MASK 0x00000080 +#define SQ_INTERRUPT_WORD_AUTO_CTXID1__SE_ID_MASK 0x030 +#define SQ_INTERRUPT_WORD_AUTO_CTXID1__ENCODING_MASK 0x0c0 + +/* SQ_INTERRUPT_WORD_WAVE_CTXID */ +#define SQ_INTERRUPT_WORD_WAVE_CTXID0__DATA__SHIFT 0 +#define SQ_INTERRUPT_WORD_WAVE_CTXID0__SA_ID__SHIFT 23 +#define SQ_INTERRUPT_WORD_WAVE_CTXID0__PRIV__SHIFT 24 +#define SQ_INTERRUPT_WORD_WAVE_CTXID0__WAVE_ID__SHIFT 25 +#define SQ_INTERRUPT_WORD_WAVE_CTXID0__SIMD_ID__SHIFT 30 +#define SQ_INTERRUPT_WORD_WAVE_CTXID1__WGP_ID__SHIFT 0 +#define SQ_INTERRUPT_WORD_WAVE_CTXID1__SE_ID__SHIFT 4 +#define SQ_INTERRUPT_WORD_WAVE_CTXID1__ENCODING__SHIFT 6 + +#define SQ_INTERRUPT_WORD_WAVE_CTXID0__DATA_MASK 0x000007fffff +#define SQ_INTERRUPT_WORD_WAVE_CTXID0__SA_ID_MASK 0x0000800000 +#define SQ_INTERRUPT_WORD_WAVE_CTXID0__PRIV_MASK 0x00001000000 +#define SQ_INTERRUPT_WORD_WAVE_CTXID0__WAVE_ID_MASK 0x0003e000000 +#define SQ_INTERRUPT_WORD_WAVE_CTXID0__SIMD_ID_MASK 0x000c0000000 +#define SQ_INTERRUPT_WORD_WAVE_CTXID1__WGP_ID_MASK 0x00f +#define SQ_INTERRUPT_WORD_WAVE_CTXID1__SE_ID_MASK 0x030 +#define SQ_INTERRUPT_WORD_WAVE_CTXID1__ENCODING_MASK 0x0c0 + +#define KFD_CTXID0__ERR_TYPE_MASK 0x780000 +#define KFD_CTXID0__ERR_TYPE__SHIFT 19 + +/* GFX10 SQ interrupt ENC type bit (context_id1[7:6]) for wave s_sendmsg */ +#define KFD_CONTEXT_ID1_ENC_TYPE_WAVE_MASK 0x40 +/* GFX10 SQ interrupt PRIV bit (context_id0[24]) for s_sendmsg inside trap */ +#define KFD_CONTEXT_ID0_PRIV_MASK 0x1000000 +/* + * The debugger will send user data(m0) with PRIV=1 to indicate it requires + * notification from the KFD with the following queue id (DOORBELL_ID) and + * trap code (TRAP_CODE). + */ +#define KFD_CONTEXT_ID0_DEBUG_DOORBELL_MASK 0x0003ff +#define KFD_CONTEXT_ID0_DEBUG_TRAP_CODE_SHIFT 10 +#define KFD_CONTEXT_ID0_DEBUG_TRAP_CODE_MASK 0x07fc00 +#define KFD_DEBUG_DOORBELL_ID(ctxid0) ((ctxid0) & \ + KFD_CONTEXT_ID0_DEBUG_DOORBELL_MASK) +#define KFD_DEBUG_TRAP_CODE(ctxid0) (((ctxid0) & \ + KFD_CONTEXT_ID0_DEBUG_TRAP_CODE_MASK) \ + >> KFD_CONTEXT_ID0_DEBUG_TRAP_CODE_SHIFT) +#define KFD_DEBUG_CP_BAD_OP_ECODE_MASK 0x3fffc00 +#define KFD_DEBUG_CP_BAD_OP_ECODE_SHIFT 10 +#define KFD_DEBUG_CP_BAD_OP_ECODE(ctxid0) (((ctxid0) & \ + KFD_DEBUG_CP_BAD_OP_ECODE_MASK) \ + >> KFD_DEBUG_CP_BAD_OP_ECODE_SHIFT) + +static void event_interrupt_poison_consumption(struct kfd_node *dev, + uint16_t pasid, uint16_t client_id) +{ + int old_poison, ret = -EINVAL; + struct kfd_process *p = kfd_lookup_process_by_pasid(pasid); + + if (!p) + return; + + /* all queues of a process will be unmapped in one time */ + old_poison = atomic_cmpxchg(&p->poison, 0, 1); + kfd_unref_process(p); + if (old_poison) + return; + + switch (client_id) { + case SOC15_IH_CLIENTID_SE0SH: + case SOC15_IH_CLIENTID_SE1SH: + case SOC15_IH_CLIENTID_SE2SH: + case SOC15_IH_CLIENTID_SE3SH: + case SOC15_IH_CLIENTID_UTCL2: + ret = kfd_dqm_evict_pasid(dev->dqm, pasid); + break; + case SOC15_IH_CLIENTID_SDMA0: + case SOC15_IH_CLIENTID_SDMA1: + case SOC15_IH_CLIENTID_SDMA2: + case SOC15_IH_CLIENTID_SDMA3: + case SOC15_IH_CLIENTID_SDMA4: + break; + default: + break; + } + + kfd_signal_poison_consumed_event(dev, pasid); + + /* resetting queue passes, do page retirement without gpu reset + * resetting queue fails, fallback to gpu reset solution + */ + if (!ret) { + dev_warn(dev->adev->dev, + "RAS poison consumption, unmap queue flow succeeded: client id %d\n", + client_id); + amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, false); + } else { + dev_warn(dev->adev->dev, + "RAS poison consumption, fall back to gpu reset flow: client id %d\n", + client_id); + amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, true); + } +} + +static bool event_interrupt_isr_v10(struct kfd_node *dev, + const uint32_t *ih_ring_entry, + uint32_t *patched_ihre, + bool *patched_flag) +{ + uint16_t source_id, client_id, pasid, vmid; + const uint32_t *data = ih_ring_entry; + + source_id = SOC15_SOURCE_ID_FROM_IH_ENTRY(ih_ring_entry); + client_id = SOC15_CLIENT_ID_FROM_IH_ENTRY(ih_ring_entry); + + /* Only handle interrupts from KFD VMIDs */ + vmid = SOC15_VMID_FROM_IH_ENTRY(ih_ring_entry); + if (!KFD_IRQ_IS_FENCE(client_id, source_id) && + (vmid < dev->vm_info.first_vmid_kfd || + vmid > dev->vm_info.last_vmid_kfd)) + return false; + + pasid = SOC15_PASID_FROM_IH_ENTRY(ih_ring_entry); + + /* Only handle clients we care about */ + if (client_id != SOC15_IH_CLIENTID_GRBM_CP && + client_id != SOC15_IH_CLIENTID_SDMA0 && + client_id != SOC15_IH_CLIENTID_SDMA1 && + client_id != SOC15_IH_CLIENTID_SDMA2 && + client_id != SOC15_IH_CLIENTID_SDMA3 && + client_id != SOC15_IH_CLIENTID_SDMA4 && + client_id != SOC15_IH_CLIENTID_SDMA5 && + client_id != SOC15_IH_CLIENTID_SDMA6 && + client_id != SOC15_IH_CLIENTID_SDMA7 && + client_id != SOC15_IH_CLIENTID_VMC && + client_id != SOC15_IH_CLIENTID_VMC1 && + client_id != SOC15_IH_CLIENTID_UTCL2 && + client_id != SOC15_IH_CLIENTID_SE0SH && + client_id != SOC15_IH_CLIENTID_SE1SH && + client_id != SOC15_IH_CLIENTID_SE2SH && + client_id != SOC15_IH_CLIENTID_SE3SH) + return false; + + pr_debug("client id 0x%x, source id %d, vmid %d, pasid 0x%x. raw data:\n", + client_id, source_id, vmid, pasid); + pr_debug("%8X, %8X, %8X, %8X, %8X, %8X, %8X, %8X.\n", + data[0], data[1], data[2], data[3], + data[4], data[5], data[6], data[7]); + + /* If there is no valid PASID, it's likely a bug */ + if (WARN_ONCE(pasid == 0, "Bug: No PASID in KFD interrupt")) + return 0; + + /* Interrupt types we care about: various signals and faults. + * They will be forwarded to a work queue (see below). + */ + return source_id == SOC15_INTSRC_CP_END_OF_PIPE || + source_id == SOC15_INTSRC_SDMA_TRAP || + source_id == SOC15_INTSRC_SQ_INTERRUPT_MSG || + source_id == SOC15_INTSRC_CP_BAD_OPCODE || + client_id == SOC15_IH_CLIENTID_VMC || + client_id == SOC15_IH_CLIENTID_VMC1 || + client_id == SOC15_IH_CLIENTID_UTCL2 || + KFD_IRQ_IS_FENCE(client_id, source_id); +} + +static void event_interrupt_wq_v10(struct kfd_node *dev, + const uint32_t *ih_ring_entry) +{ + uint16_t source_id, client_id, pasid, vmid; + uint32_t context_id0, context_id1; + uint32_t encoding, sq_intr_err_type; + + source_id = SOC15_SOURCE_ID_FROM_IH_ENTRY(ih_ring_entry); + client_id = SOC15_CLIENT_ID_FROM_IH_ENTRY(ih_ring_entry); + pasid = SOC15_PASID_FROM_IH_ENTRY(ih_ring_entry); + vmid = SOC15_VMID_FROM_IH_ENTRY(ih_ring_entry); + context_id0 = SOC15_CONTEXT_ID0_FROM_IH_ENTRY(ih_ring_entry); + context_id1 = SOC15_CONTEXT_ID1_FROM_IH_ENTRY(ih_ring_entry); + + if (client_id == SOC15_IH_CLIENTID_GRBM_CP || + client_id == SOC15_IH_CLIENTID_SE0SH || + client_id == SOC15_IH_CLIENTID_SE1SH || + client_id == SOC15_IH_CLIENTID_SE2SH || + client_id == SOC15_IH_CLIENTID_SE3SH) { + if (source_id == SOC15_INTSRC_CP_END_OF_PIPE) + kfd_signal_event_interrupt(pasid, context_id0, 32); + else if (source_id == SOC15_INTSRC_SQ_INTERRUPT_MSG) { + encoding = REG_GET_FIELD(context_id1, + SQ_INTERRUPT_WORD_WAVE_CTXID1, ENCODING); + switch (encoding) { + case SQ_INTERRUPT_WORD_ENCODING_AUTO: + pr_debug( + "sq_intr: auto, se %d, ttrace %d, wlt %d, ttrac_buf0_full %d, ttrac_buf1_full %d, ttrace_utc_err %d\n", + REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_AUTO_CTXID1, + SE_ID), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + THREAD_TRACE), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + WLT), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + THREAD_TRACE_BUF0_FULL), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + THREAD_TRACE_BUF1_FULL), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, + THREAD_TRACE_UTC_ERROR)); + break; + case SQ_INTERRUPT_WORD_ENCODING_INST: + pr_debug("sq_intr: inst, se %d, data 0x%x, sa %d, priv %d, wave_id %d, simd_id %d, wgp_id %d\n", + REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1, + SE_ID), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, + DATA), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, + SA_ID), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, + PRIV), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, + WAVE_ID), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, + SIMD_ID), + REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1, + WGP_ID)); + if (context_id0 & SQ_INTERRUPT_WORD_WAVE_CTXID0__PRIV_MASK) { + if (kfd_set_dbg_ev_from_interrupt(dev, pasid, + KFD_DEBUG_DOORBELL_ID(context_id0), + KFD_DEBUG_TRAP_CODE(context_id0), + NULL, 0)) + return; + } + break; + case SQ_INTERRUPT_WORD_ENCODING_ERROR: + sq_intr_err_type = REG_GET_FIELD(context_id0, KFD_CTXID0, + ERR_TYPE); + pr_warn("sq_intr: error, se %d, data 0x%x, sa %d, priv %d, wave_id %d, simd_id %d, wgp_id %d, err_type %d\n", + REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1, + SE_ID), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, + DATA), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, + SA_ID), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, + PRIV), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, + WAVE_ID), + REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, + SIMD_ID), + REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1, + WGP_ID), + sq_intr_err_type); + if (sq_intr_err_type != SQ_INTERRUPT_ERROR_TYPE_ILLEGAL_INST && + sq_intr_err_type != SQ_INTERRUPT_ERROR_TYPE_MEMVIOL) { + event_interrupt_poison_consumption(dev, pasid, source_id); + return; + } + break; + default: + break; + } + kfd_signal_event_interrupt(pasid, context_id0 & 0x7fffff, 23); + } else if (source_id == SOC15_INTSRC_CP_BAD_OPCODE) { + kfd_set_dbg_ev_from_interrupt(dev, pasid, + KFD_DEBUG_DOORBELL_ID(context_id0), + KFD_EC_MASK(KFD_DEBUG_CP_BAD_OP_ECODE(context_id0)), + NULL, + 0); + } + } else if (client_id == SOC15_IH_CLIENTID_SDMA0 || + client_id == SOC15_IH_CLIENTID_SDMA1 || + client_id == SOC15_IH_CLIENTID_SDMA2 || + client_id == SOC15_IH_CLIENTID_SDMA3 || + (client_id == SOC15_IH_CLIENTID_SDMA3_Sienna_Cichlid && + KFD_GC_VERSION(dev) == IP_VERSION(10, 3, 0)) || + client_id == SOC15_IH_CLIENTID_SDMA4 || + client_id == SOC15_IH_CLIENTID_SDMA5 || + client_id == SOC15_IH_CLIENTID_SDMA6 || + client_id == SOC15_IH_CLIENTID_SDMA7) { + if (source_id == SOC15_INTSRC_SDMA_TRAP) { + kfd_signal_event_interrupt(pasid, context_id0 & 0xfffffff, 28); + } else if (source_id == SOC15_INTSRC_SDMA_ECC) { + event_interrupt_poison_consumption(dev, pasid, source_id); + return; + } + } else if (client_id == SOC15_IH_CLIENTID_VMC || + client_id == SOC15_IH_CLIENTID_VMC1 || + client_id == SOC15_IH_CLIENTID_UTCL2) { + struct kfd_vm_fault_info info = {0}; + uint16_t ring_id = SOC15_RING_ID_FROM_IH_ENTRY(ih_ring_entry); + struct kfd_hsa_memory_exception_data exception_data; + + if (client_id == SOC15_IH_CLIENTID_UTCL2 && + amdgpu_amdkfd_ras_query_utcl2_poison_status(dev->adev)) { + event_interrupt_poison_consumption(dev, pasid, client_id); + return; + } + + info.vmid = vmid; + info.mc_id = client_id; + info.page_addr = ih_ring_entry[4] | + (uint64_t)(ih_ring_entry[5] & 0xf) << 32; + info.prot_valid = ring_id & 0x08; + info.prot_read = ring_id & 0x10; + info.prot_write = ring_id & 0x20; + + memset(&exception_data, 0, sizeof(exception_data)); + exception_data.gpu_id = dev->id; + exception_data.va = (info.page_addr) << PAGE_SHIFT; + exception_data.failure.NotPresent = info.prot_valid ? 1 : 0; + exception_data.failure.NoExecute = info.prot_exec ? 1 : 0; + exception_data.failure.ReadOnly = info.prot_write ? 1 : 0; + exception_data.failure.imprecise = 0; + + kfd_set_dbg_ev_from_interrupt(dev, + pasid, + -1, + KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION), + &exception_data, + sizeof(exception_data)); + } else if (KFD_IRQ_IS_FENCE(client_id, source_id)) { + kfd_process_close_interrupt_drain(pasid); + } +} + +const struct kfd_event_interrupt_class event_interrupt_class_v10 = { + .interrupt_isr = event_interrupt_isr_v10, + .interrupt_wq = event_interrupt_wq_v10, +}; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c index 0d53f6067422..f933bd231fb9 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c @@ -26,6 +26,7 @@ #include "kfd_device_queue_manager.h" #include "ivsrcid/vmc/irqsrcs_vmc_1_0.h" #include "kfd_smi_events.h" +#include "kfd_debug.h" /* * GFX11 SQ Interrupts @@ -187,7 +188,7 @@ static void print_sq_intr_info_error(uint32_t context_id0, uint32_t context_id1) REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID1, WGP_ID)); } -static void event_interrupt_poison_consumption_v11(struct kfd_dev *dev, +static void event_interrupt_poison_consumption_v11(struct kfd_node *dev, uint16_t pasid, uint16_t source_id) { int ret = -EINVAL; @@ -225,7 +226,7 @@ static void event_interrupt_poison_consumption_v11(struct kfd_dev *dev, amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, true); } -static bool event_interrupt_isr_v11(struct kfd_dev *dev, +static bool event_interrupt_isr_v11(struct kfd_node *dev, const uint32_t *ih_ring_entry, uint32_t *patched_ihre, bool *patched_flag) @@ -238,7 +239,7 @@ static bool event_interrupt_isr_v11(struct kfd_dev *dev, client_id = SOC15_CLIENT_ID_FROM_IH_ENTRY(ih_ring_entry); /* Only handle interrupts from KFD VMIDs */ vmid = SOC15_VMID_FROM_IH_ENTRY(ih_ring_entry); - if (/*!KFD_IRQ_IS_FENCE(client_id, source_id) &&*/ + if (!KFD_IRQ_IS_FENCE(client_id, source_id) && (vmid < dev->vm_info.first_vmid_kfd || vmid > dev->vm_info.last_vmid_kfd)) return false; @@ -267,19 +268,19 @@ static bool event_interrupt_isr_v11(struct kfd_dev *dev, source_id == SOC15_INTSRC_SQ_INTERRUPT_MSG || source_id == SOC15_INTSRC_CP_BAD_OPCODE || source_id == SOC21_INTSRC_SDMA_TRAP || - /* KFD_IRQ_IS_FENCE(client_id, source_id) || */ + KFD_IRQ_IS_FENCE(client_id, source_id) || (((client_id == SOC21_IH_CLIENTID_VMC) || ((client_id == SOC21_IH_CLIENTID_GFX) && (source_id == UTCL2_1_0__SRCID__FAULT))) && !amdgpu_no_queue_eviction_on_vm_fault); } -static void event_interrupt_wq_v11(struct kfd_dev *dev, +static void event_interrupt_wq_v11(struct kfd_node *dev, const uint32_t *ih_ring_entry) { uint16_t source_id, client_id, ring_id, pasid, vmid; uint32_t context_id0, context_id1; - uint8_t sq_int_enc, sq_int_errtype, sq_int_priv; + uint8_t sq_int_enc, sq_int_priv, sq_int_errtype; struct kfd_vm_fault_info info = {0}; struct kfd_hsa_memory_exception_data exception_data; @@ -312,9 +313,9 @@ static void event_interrupt_wq_v11(struct kfd_dev *dev, exception_data.failure.ReadOnly = info.prot_write ? 1 : 0; exception_data.failure.imprecise = 0; - /*kfd_set_dbg_ev_from_interrupt(dev, pasid, -1, + kfd_set_dbg_ev_from_interrupt(dev, pasid, -1, KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION), - &exception_data, sizeof(exception_data));*/ + &exception_data, sizeof(exception_data)); kfd_smi_event_update_vmfault(dev, pasid); /* GRBM, SDMA, SE, PMM */ @@ -324,11 +325,11 @@ static void event_interrupt_wq_v11(struct kfd_dev *dev, /* CP */ if (source_id == SOC15_INTSRC_CP_END_OF_PIPE) kfd_signal_event_interrupt(pasid, context_id0, 32); - /*else if (source_id == SOC15_INTSRC_CP_BAD_OPCODE) + else if (source_id == SOC15_INTSRC_CP_BAD_OPCODE) kfd_set_dbg_ev_from_interrupt(dev, pasid, KFD_CTXID0_DOORBELL_ID(context_id0), KFD_EC_MASK(KFD_CTXID0_CP_BAD_OP_ECODE(context_id0)), - NULL, 0);*/ + NULL, 0); /* SDMA */ else if (source_id == SOC21_INTSRC_SDMA_TRAP) @@ -350,11 +351,11 @@ static void event_interrupt_wq_v11(struct kfd_dev *dev, print_sq_intr_info_inst(context_id0, context_id1); sq_int_priv = REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, PRIV); - /*if (sq_int_priv && (kfd_set_dbg_ev_from_interrupt(dev, pasid, + if (sq_int_priv && (kfd_set_dbg_ev_from_interrupt(dev, pasid, KFD_CTXID0_DOORBELL_ID(context_id0), KFD_CTXID0_TRAP_CODE(context_id0), NULL, 0))) - return;*/ + return; break; case SQ_INTERRUPT_WORD_ENCODING_ERROR: print_sq_intr_info_error(context_id0, context_id1); @@ -373,8 +374,8 @@ static void event_interrupt_wq_v11(struct kfd_dev *dev, kfd_signal_event_interrupt(pasid, context_id0 & 0xffffff, 24); } - /*} else if (KFD_IRQ_IS_FENCE(client_id, source_id)) { - kfd_process_close_interrupt_drain(pasid);*/ + } else if (KFD_IRQ_IS_FENCE(client_id, source_id)) { + kfd_process_close_interrupt_drain(pasid); } } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c index 0b75a37b689b..d5c9f30552e3 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c @@ -23,10 +23,40 @@ #include "kfd_priv.h" #include "kfd_events.h" +#include "kfd_debug.h" #include "soc15_int.h" #include "kfd_device_queue_manager.h" #include "kfd_smi_events.h" +/* + * GFX9 SQ Interrupts + * + * There are 3 encoding types of interrupts sourced from SQ sent as a 44-bit + * packet to the Interrupt Handler: + * Auto - Generated by the SQG (various cmd overflows, timestamps etc) + * Wave - Generated by S_SENDMSG through a shader program + * Error - HW generated errors (Illegal instructions, Memviols, EDC etc) + * + * The 44-bit packet is mapped as {context_id1[7:0],context_id0[31:0]} plus + * 4-bits for VMID (SOC15_VMID_FROM_IH_ENTRY) as such: + * + * - context_id0[27:26] + * Encoding type (0 = Auto, 1 = Wave, 2 = Error) + * + * - context_id0[13] + * PRIV bit indicates that Wave S_SEND or error occurred within trap + * + * - {context_id1[7:0],context_id0[31:28],context_id0[11:0]} + * 24-bit data with the following layout per encoding type: + * Auto - only context_id0[8:0] is used, which reports various interrupts + * generated by SQG. The rest is 0. + * Wave - user data sent from m0 via S_SENDMSG + * Error - Error type (context_id1[7:4]), Error Details (rest of bits) + * + * The other context_id bits show coordinates (SE/SH/CU/SIMD/WAVE) for wave + * S_SENDMSG and Errors. These are 0 for Auto. + */ + enum SQ_INTERRUPT_WORD_ENCODING { SQ_INTERRUPT_WORD_ENCODING_AUTO = 0x0, SQ_INTERRUPT_WORD_ENCODING_INST, @@ -84,13 +114,33 @@ enum SQ_INTERRUPT_ERROR_TYPE { #define SQ_INTERRUPT_WORD_WAVE_CTXID__SE_ID_MASK 0x03000000 #define SQ_INTERRUPT_WORD_WAVE_CTXID__ENCODING_MASK 0x0c000000 +/* GFX9 SQ interrupt 24-bit data from context_id<0,1> */ #define KFD_CONTEXT_ID_GET_SQ_INT_DATA(ctx0, ctx1) \ ((ctx0 & 0xfff) | ((ctx0 >> 16) & 0xf000) | ((ctx1 << 16) & 0xff0000)) #define KFD_SQ_INT_DATA__ERR_TYPE_MASK 0xF00000 #define KFD_SQ_INT_DATA__ERR_TYPE__SHIFT 20 -static void event_interrupt_poison_consumption_v9(struct kfd_dev *dev, +/* + * The debugger will send user data(m0) with PRIV=1 to indicate it requires + * notification from the KFD with the following queue id (DOORBELL_ID) and + * trap code (TRAP_CODE). + */ +#define KFD_INT_DATA_DEBUG_DOORBELL_MASK 0x0003ff +#define KFD_INT_DATA_DEBUG_TRAP_CODE_SHIFT 10 +#define KFD_INT_DATA_DEBUG_TRAP_CODE_MASK 0x07fc00 +#define KFD_DEBUG_DOORBELL_ID(sq_int_data) ((sq_int_data) & \ + KFD_INT_DATA_DEBUG_DOORBELL_MASK) +#define KFD_DEBUG_TRAP_CODE(sq_int_data) (((sq_int_data) & \ + KFD_INT_DATA_DEBUG_TRAP_CODE_MASK) \ + >> KFD_INT_DATA_DEBUG_TRAP_CODE_SHIFT) +#define KFD_DEBUG_CP_BAD_OP_ECODE_MASK 0x3fffc00 +#define KFD_DEBUG_CP_BAD_OP_ECODE_SHIFT 10 +#define KFD_DEBUG_CP_BAD_OP_ECODE(ctxid0) (((ctxid0) & \ + KFD_DEBUG_CP_BAD_OP_ECODE_MASK) \ + >> KFD_DEBUG_CP_BAD_OP_ECODE_SHIFT) + +static void event_interrupt_poison_consumption_v9(struct kfd_node *dev, uint16_t pasid, uint16_t client_id) { int old_poison, ret = -EINVAL; @@ -160,7 +210,7 @@ static bool context_id_expected(struct kfd_dev *dev) } } -static bool event_interrupt_isr_v9(struct kfd_dev *dev, +static bool event_interrupt_isr_v9(struct kfd_node *dev, const uint32_t *ih_ring_entry, uint32_t *patched_ihre, bool *patched_flag) @@ -168,14 +218,16 @@ static bool event_interrupt_isr_v9(struct kfd_dev *dev, uint16_t source_id, client_id, pasid, vmid; const uint32_t *data = ih_ring_entry; + source_id = SOC15_SOURCE_ID_FROM_IH_ENTRY(ih_ring_entry); + client_id = SOC15_CLIENT_ID_FROM_IH_ENTRY(ih_ring_entry); + /* Only handle interrupts from KFD VMIDs */ vmid = SOC15_VMID_FROM_IH_ENTRY(ih_ring_entry); - if (vmid < dev->vm_info.first_vmid_kfd || - vmid > dev->vm_info.last_vmid_kfd) + if (!KFD_IRQ_IS_FENCE(client_id, source_id) && + (vmid < dev->vm_info.first_vmid_kfd || + vmid > dev->vm_info.last_vmid_kfd)) return false; - source_id = SOC15_SOURCE_ID_FROM_IH_ENTRY(ih_ring_entry); - client_id = SOC15_CLIENT_ID_FROM_IH_ENTRY(ih_ring_entry); pasid = SOC15_PASID_FROM_IH_ENTRY(ih_ring_entry); /* Only handle clients we care about */ @@ -194,7 +246,8 @@ static bool event_interrupt_isr_v9(struct kfd_dev *dev, client_id != SOC15_IH_CLIENTID_SE0SH && client_id != SOC15_IH_CLIENTID_SE1SH && client_id != SOC15_IH_CLIENTID_SE2SH && - client_id != SOC15_IH_CLIENTID_SE3SH) + client_id != SOC15_IH_CLIENTID_SE3SH && + !KFD_IRQ_IS_FENCE(client_id, source_id)) return false; /* This is a known issue for gfx9. Under non HWS, pasid is not set @@ -206,7 +259,7 @@ static bool event_interrupt_isr_v9(struct kfd_dev *dev, *patched_flag = true; memcpy(patched_ihre, ih_ring_entry, - dev->device_info.ih_ring_entry_size); + dev->kfd->device_info.ih_ring_entry_size); pasid = dev->dqm->vmid_pasid[vmid]; @@ -235,7 +288,7 @@ static bool event_interrupt_isr_v9(struct kfd_dev *dev, uint32_t context_id = SOC15_CONTEXT_ID0_FROM_IH_ENTRY(ih_ring_entry); - if (context_id == 0 && context_id_expected(dev)) + if (context_id == 0 && context_id_expected(dev->kfd)) return false; } @@ -247,13 +300,14 @@ static bool event_interrupt_isr_v9(struct kfd_dev *dev, source_id == SOC15_INTSRC_SDMA_ECC || source_id == SOC15_INTSRC_SQ_INTERRUPT_MSG || source_id == SOC15_INTSRC_CP_BAD_OPCODE || + KFD_IRQ_IS_FENCE(client_id, source_id) || ((client_id == SOC15_IH_CLIENTID_VMC || client_id == SOC15_IH_CLIENTID_VMC1 || client_id == SOC15_IH_CLIENTID_UTCL2) && !amdgpu_no_queue_eviction_on_vm_fault); } -static void event_interrupt_wq_v9(struct kfd_dev *dev, +static void event_interrupt_wq_v9(struct kfd_node *dev, const uint32_t *ih_ring_entry) { uint16_t source_id, client_id, pasid, vmid; @@ -302,6 +356,13 @@ static void event_interrupt_wq_v9(struct kfd_dev *dev, REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SIMD_ID), REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, CU_ID), sq_int_data); + if (context_id0 & SQ_INTERRUPT_WORD_WAVE_CTXID__PRIV_MASK) { + if (kfd_set_dbg_ev_from_interrupt(dev, pasid, + KFD_DEBUG_DOORBELL_ID(sq_int_data), + KFD_DEBUG_TRAP_CODE(sq_int_data), + NULL, 0)) + return; + } break; case SQ_INTERRUPT_WORD_ENCODING_ERROR: sq_intr_err = REG_GET_FIELD(sq_int_data, KFD_SQ_INT_DATA, ERR_TYPE); @@ -324,8 +385,12 @@ static void event_interrupt_wq_v9(struct kfd_dev *dev, break; } kfd_signal_event_interrupt(pasid, context_id0 & 0xffffff, 24); - } else if (source_id == SOC15_INTSRC_CP_BAD_OPCODE) - kfd_signal_hw_exception_event(pasid); + } else if (source_id == SOC15_INTSRC_CP_BAD_OPCODE) { + kfd_set_dbg_ev_from_interrupt(dev, pasid, + KFD_DEBUG_DOORBELL_ID(context_id0), + KFD_EC_MASK(KFD_DEBUG_CP_BAD_OP_ECODE(context_id0)), + NULL, 0); + } } else if (client_id == SOC15_IH_CLIENTID_SDMA0 || client_id == SOC15_IH_CLIENTID_SDMA1 || client_id == SOC15_IH_CLIENTID_SDMA2 || @@ -345,6 +410,7 @@ static void event_interrupt_wq_v9(struct kfd_dev *dev, client_id == SOC15_IH_CLIENTID_UTCL2) { struct kfd_vm_fault_info info = {0}; uint16_t ring_id = SOC15_RING_ID_FROM_IH_ENTRY(ih_ring_entry); + struct kfd_hsa_memory_exception_data exception_data; if (client_id == SOC15_IH_CLIENTID_UTCL2 && amdgpu_amdkfd_ras_query_utcl2_poison_status(dev->adev)) { @@ -360,9 +426,23 @@ static void event_interrupt_wq_v9(struct kfd_dev *dev, info.prot_read = ring_id & 0x10; info.prot_write = ring_id & 0x20; + memset(&exception_data, 0, sizeof(exception_data)); + exception_data.gpu_id = dev->id; + exception_data.va = (info.page_addr) << PAGE_SHIFT; + exception_data.failure.NotPresent = info.prot_valid ? 1 : 0; + exception_data.failure.NoExecute = info.prot_exec ? 1 : 0; + exception_data.failure.ReadOnly = info.prot_write ? 1 : 0; + exception_data.failure.imprecise = 0; + + kfd_set_dbg_ev_from_interrupt(dev, + pasid, + -1, + KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION), + &exception_data, + sizeof(exception_data)); kfd_smi_event_update_vmfault(dev, pasid); - kfd_dqm_evict_pasid(dev->dqm, pasid); - kfd_signal_vm_fault_event(dev, pasid, &info); + } else if (KFD_IRQ_IS_FENCE(client_id, source_id)) { + kfd_process_close_interrupt_drain(pasid); } } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c index 34772fe74296..dd3c43c1ad70 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c @@ -50,29 +50,29 @@ static void interrupt_wq(struct work_struct *); -int kfd_interrupt_init(struct kfd_dev *kfd) +int kfd_interrupt_init(struct kfd_node *node) { int r; - r = kfifo_alloc(&kfd->ih_fifo, - KFD_IH_NUM_ENTRIES * kfd->device_info.ih_ring_entry_size, + r = kfifo_alloc(&node->ih_fifo, + KFD_IH_NUM_ENTRIES * node->kfd->device_info.ih_ring_entry_size, GFP_KERNEL); if (r) { - dev_err(kfd->adev->dev, "Failed to allocate IH fifo\n"); + dev_err(node->adev->dev, "Failed to allocate IH fifo\n"); return r; } - kfd->ih_wq = alloc_workqueue("KFD IH", WQ_HIGHPRI, 1); - if (unlikely(!kfd->ih_wq)) { - kfifo_free(&kfd->ih_fifo); - dev_err(kfd->adev->dev, "Failed to allocate KFD IH workqueue\n"); + node->ih_wq = alloc_workqueue("KFD IH", WQ_HIGHPRI, 1); + if (unlikely(!node->ih_wq)) { + kfifo_free(&node->ih_fifo); + dev_err(node->adev->dev, "Failed to allocate KFD IH workqueue\n"); return -ENOMEM; } - spin_lock_init(&kfd->interrupt_lock); + spin_lock_init(&node->interrupt_lock); - INIT_WORK(&kfd->interrupt_work, interrupt_wq); + INIT_WORK(&node->interrupt_work, interrupt_wq); - kfd->interrupts_active = true; + node->interrupts_active = true; /* * After this function returns, the interrupt will be enabled. This @@ -84,7 +84,7 @@ int kfd_interrupt_init(struct kfd_dev *kfd) return 0; } -void kfd_interrupt_exit(struct kfd_dev *kfd) +void kfd_interrupt_exit(struct kfd_node *node) { /* * Stop the interrupt handler from writing to the ring and scheduling @@ -93,31 +93,31 @@ void kfd_interrupt_exit(struct kfd_dev *kfd) */ unsigned long flags; - spin_lock_irqsave(&kfd->interrupt_lock, flags); - kfd->interrupts_active = false; - spin_unlock_irqrestore(&kfd->interrupt_lock, flags); + spin_lock_irqsave(&node->interrupt_lock, flags); + node->interrupts_active = false; + spin_unlock_irqrestore(&node->interrupt_lock, flags); /* * flush_work ensures that there are no outstanding * work-queue items that will access interrupt_ring. New work items * can't be created because we stopped interrupt handling above. */ - flush_workqueue(kfd->ih_wq); + flush_workqueue(node->ih_wq); - kfifo_free(&kfd->ih_fifo); + kfifo_free(&node->ih_fifo); } /* * Assumption: single reader/writer. This function is not re-entrant */ -bool enqueue_ih_ring_entry(struct kfd_dev *kfd, const void *ih_ring_entry) +bool enqueue_ih_ring_entry(struct kfd_node *node, const void *ih_ring_entry) { int count; - count = kfifo_in(&kfd->ih_fifo, ih_ring_entry, - kfd->device_info.ih_ring_entry_size); - if (count != kfd->device_info.ih_ring_entry_size) { - dev_dbg_ratelimited(kfd->adev->dev, + count = kfifo_in(&node->ih_fifo, ih_ring_entry, + node->kfd->device_info.ih_ring_entry_size); + if (count != node->kfd->device_info.ih_ring_entry_size) { + dev_dbg_ratelimited(node->adev->dev, "Interrupt ring overflow, dropping interrupt %d\n", count); return false; @@ -129,32 +129,32 @@ bool enqueue_ih_ring_entry(struct kfd_dev *kfd, const void *ih_ring_entry) /* * Assumption: single reader/writer. This function is not re-entrant */ -static bool dequeue_ih_ring_entry(struct kfd_dev *kfd, void *ih_ring_entry) +static bool dequeue_ih_ring_entry(struct kfd_node *node, void *ih_ring_entry) { int count; - count = kfifo_out(&kfd->ih_fifo, ih_ring_entry, - kfd->device_info.ih_ring_entry_size); + count = kfifo_out(&node->ih_fifo, ih_ring_entry, + node->kfd->device_info.ih_ring_entry_size); - WARN_ON(count && count != kfd->device_info.ih_ring_entry_size); + WARN_ON(count && count != node->kfd->device_info.ih_ring_entry_size); - return count == kfd->device_info.ih_ring_entry_size; + return count == node->kfd->device_info.ih_ring_entry_size; } static void interrupt_wq(struct work_struct *work) { - struct kfd_dev *dev = container_of(work, struct kfd_dev, + struct kfd_node *dev = container_of(work, struct kfd_node, interrupt_work); uint32_t ih_ring_entry[KFD_MAX_RING_ENTRY_SIZE]; unsigned long start_jiffies = jiffies; - if (dev->device_info.ih_ring_entry_size > sizeof(ih_ring_entry)) { + if (dev->kfd->device_info.ih_ring_entry_size > sizeof(ih_ring_entry)) { dev_err_once(dev->adev->dev, "Ring entry too small\n"); return; } while (dequeue_ih_ring_entry(dev, ih_ring_entry)) { - dev->device_info.event_interrupt_class->interrupt_wq(dev, + dev->kfd->device_info.event_interrupt_class->interrupt_wq(dev, ih_ring_entry); if (time_is_before_jiffies(start_jiffies + HZ)) { /* If we spent more than a second processing signals, @@ -166,14 +166,14 @@ static void interrupt_wq(struct work_struct *work) } } -bool interrupt_is_wanted(struct kfd_dev *dev, +bool interrupt_is_wanted(struct kfd_node *dev, const uint32_t *ih_ring_entry, uint32_t *patched_ihre, bool *flag) { /* integer and bitwise OR so there is no boolean short-circuiting */ unsigned int wanted = 0; - wanted |= dev->device_info.event_interrupt_class->interrupt_isr(dev, + wanted |= dev->kfd->device_info.event_interrupt_class->interrupt_isr(dev, ih_ring_entry, patched_ihre, flag); return wanted != 0; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_iommu.c b/drivers/gpu/drm/amd/amdkfd/kfd_iommu.c index ec1bf611624e..808ee010520a 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_iommu.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_iommu.c @@ -109,11 +109,11 @@ int kfd_iommu_device_init(struct kfd_dev *kfd) */ int kfd_iommu_bind_process_to_device(struct kfd_process_device *pdd) { - struct kfd_dev *dev = pdd->dev; + struct kfd_node *dev = pdd->dev; struct kfd_process *p = pdd->process; int err; - if (!dev->use_iommu_v2 || pdd->bound == PDD_BOUND) + if (!dev->kfd->use_iommu_v2 || pdd->bound == PDD_BOUND) return 0; if (unlikely(pdd->bound == PDD_BOUND_SUSPENDED)) { @@ -121,6 +121,12 @@ int kfd_iommu_bind_process_to_device(struct kfd_process_device *pdd) return -EINVAL; } + if (!kfd_is_first_node(dev)) { + dev_warn_once(kfd_device, + "IOMMU supported only on first node\n"); + return 0; + } + err = amd_iommu_bind_pasid(dev->adev->pdev, p->pasid, p->lead_thread); if (!err) pdd->bound = PDD_BOUND; @@ -138,7 +144,8 @@ void kfd_iommu_unbind_process(struct kfd_process *p) int i; for (i = 0; i < p->n_pdds; i++) - if (p->pdds[i]->bound == PDD_BOUND) + if ((p->pdds[i]->bound == PDD_BOUND) && + (kfd_is_first_node((p->pdds[i]->dev)))) amd_iommu_unbind_pasid(p->pdds[i]->dev->adev->pdev, p->pasid); } @@ -146,7 +153,7 @@ void kfd_iommu_unbind_process(struct kfd_process *p) /* Callback for process shutdown invoked by the IOMMU driver */ static void iommu_pasid_shutdown_callback(struct pci_dev *pdev, u32 pasid) { - struct kfd_dev *dev = kfd_device_by_pci_dev(pdev); + struct kfd_node *dev = kfd_device_by_pci_dev(pdev); struct kfd_process *p; struct kfd_process_device *pdd; @@ -182,7 +189,7 @@ static void iommu_pasid_shutdown_callback(struct pci_dev *pdev, u32 pasid) static int iommu_invalid_ppr_cb(struct pci_dev *pdev, u32 pasid, unsigned long address, u16 flags) { - struct kfd_dev *dev; + struct kfd_node *dev; dev_warn_ratelimited(kfd_device, "Invalid PPR device %x:%x.%x pasid 0x%x address 0x%lX flags 0x%X", @@ -205,7 +212,7 @@ static int iommu_invalid_ppr_cb(struct pci_dev *pdev, u32 pasid, * Bind processes do the device that have been temporarily unbound * (PDD_BOUND_SUSPENDED) in kfd_unbind_processes_from_device. */ -static int kfd_bind_processes_to_device(struct kfd_dev *kfd) +static int kfd_bind_processes_to_device(struct kfd_node *knode) { struct kfd_process_device *pdd; struct kfd_process *p; @@ -216,14 +223,14 @@ static int kfd_bind_processes_to_device(struct kfd_dev *kfd) hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) { mutex_lock(&p->mutex); - pdd = kfd_get_process_device_data(kfd, p); + pdd = kfd_get_process_device_data(knode, p); if (WARN_ON(!pdd) || pdd->bound != PDD_BOUND_SUSPENDED) { mutex_unlock(&p->mutex); continue; } - err = amd_iommu_bind_pasid(kfd->adev->pdev, p->pasid, + err = amd_iommu_bind_pasid(knode->adev->pdev, p->pasid, p->lead_thread); if (err < 0) { pr_err("Unexpected pasid 0x%x binding failure\n", @@ -246,7 +253,7 @@ static int kfd_bind_processes_to_device(struct kfd_dev *kfd) * processes will be restored to PDD_BOUND state in * kfd_bind_processes_to_device. */ -static void kfd_unbind_processes_from_device(struct kfd_dev *kfd) +static void kfd_unbind_processes_from_device(struct kfd_node *knode) { struct kfd_process_device *pdd; struct kfd_process *p; @@ -256,7 +263,7 @@ static void kfd_unbind_processes_from_device(struct kfd_dev *kfd) hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) { mutex_lock(&p->mutex); - pdd = kfd_get_process_device_data(kfd, p); + pdd = kfd_get_process_device_data(knode, p); if (WARN_ON(!pdd)) { mutex_unlock(&p->mutex); @@ -281,7 +288,7 @@ void kfd_iommu_suspend(struct kfd_dev *kfd) if (!kfd->use_iommu_v2) return; - kfd_unbind_processes_from_device(kfd); + kfd_unbind_processes_from_device(kfd->nodes[0]); amd_iommu_set_invalidate_ctx_cb(kfd->adev->pdev, NULL); amd_iommu_set_invalid_ppr_cb(kfd->adev->pdev, NULL); @@ -312,7 +319,7 @@ int kfd_iommu_resume(struct kfd_dev *kfd) amd_iommu_set_invalid_ppr_cb(kfd->adev->pdev, iommu_invalid_ppr_cb); - err = kfd_bind_processes_to_device(kfd); + err = kfd_bind_processes_to_device(kfd->nodes[0]); if (err) { amd_iommu_set_invalidate_ctx_cb(kfd->adev->pdev, NULL); amd_iommu_set_invalid_ppr_cb(kfd->adev->pdev, NULL); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c index bcf7bc3302c9..1bea629c49ca 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c @@ -38,7 +38,7 @@ /* Initialize a kernel queue, including allocations of GART memory * needed for the queue. */ -static bool kq_initialize(struct kernel_queue *kq, struct kfd_dev *dev, +static bool kq_initialize(struct kernel_queue *kq, struct kfd_node *dev, enum kfd_queue_type type, unsigned int queue_size) { struct queue_properties prop; @@ -75,7 +75,7 @@ static bool kq_initialize(struct kernel_queue *kq, struct kfd_dev *dev, if (!kq->mqd_mgr) return false; - prop.doorbell_ptr = kfd_get_kernel_doorbell(dev, &prop.doorbell_off); + prop.doorbell_ptr = kfd_get_kernel_doorbell(dev->kfd, &prop.doorbell_off); if (!prop.doorbell_ptr) { pr_err("Failed to initialize doorbell"); @@ -112,7 +112,7 @@ static bool kq_initialize(struct kernel_queue *kq, struct kfd_dev *dev, kq->rptr_kernel = kq->rptr_mem->cpu_ptr; kq->rptr_gpu_addr = kq->rptr_mem->gpu_addr; - retval = kfd_gtt_sa_allocate(dev, dev->device_info.doorbell_size, + retval = kfd_gtt_sa_allocate(dev, dev->kfd->device_info.doorbell_size, &kq->wptr_mem); if (retval != 0) @@ -189,7 +189,7 @@ err_rptr_allocate_vidmem: err_eop_allocate_vidmem: kfd_gtt_sa_free(dev, kq->pq); err_pq_allocate_vidmem: - kfd_release_kernel_doorbell(dev, prop.doorbell_ptr); + kfd_release_kernel_doorbell(dev->kfd, prop.doorbell_ptr); err_get_kernel_doorbell: return false; @@ -220,7 +220,7 @@ static void kq_uninitialize(struct kernel_queue *kq, bool hanging) kfd_gtt_sa_free(kq->dev, kq->eop_mem); kfd_gtt_sa_free(kq->dev, kq->pq); - kfd_release_kernel_doorbell(kq->dev, + kfd_release_kernel_doorbell(kq->dev->kfd, kq->queue->properties.doorbell_ptr); uninit_queue(kq->queue); } @@ -298,7 +298,7 @@ void kq_submit_packet(struct kernel_queue *kq) } pr_debug("\n"); #endif - if (kq->dev->device_info.doorbell_size == 8) { + if (kq->dev->kfd->device_info.doorbell_size == 8) { *kq->wptr64_kernel = kq->pending_wptr64; write_kernel_doorbell64(kq->queue->properties.doorbell_ptr, kq->pending_wptr64); @@ -311,7 +311,7 @@ void kq_submit_packet(struct kernel_queue *kq) void kq_rollback_packet(struct kernel_queue *kq) { - if (kq->dev->device_info.doorbell_size == 8) { + if (kq->dev->kfd->device_info.doorbell_size == 8) { kq->pending_wptr64 = *kq->wptr64_kernel; kq->pending_wptr = *kq->wptr_kernel % (kq->queue->properties.queue_size / 4); @@ -320,7 +320,7 @@ void kq_rollback_packet(struct kernel_queue *kq) } } -struct kernel_queue *kernel_queue_init(struct kfd_dev *dev, +struct kernel_queue *kernel_queue_init(struct kfd_node *dev, enum kfd_queue_type type) { struct kernel_queue *kq; @@ -345,7 +345,7 @@ void kernel_queue_uninit(struct kernel_queue *kq, bool hanging) } /* FIXME: Can this test be removed? */ -static __attribute__((unused)) void test_kq(struct kfd_dev *dev) +static __attribute__((unused)) void test_kq(struct kfd_node *dev) { struct kernel_queue *kq; uint32_t *buffer, i; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.h b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.h index 383202fd1ea2..9a6244430845 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.h @@ -53,7 +53,7 @@ void kq_rollback_packet(struct kernel_queue *kq); struct kernel_queue { /* data */ - struct kfd_dev *dev; + struct kfd_node *dev; struct mqd_manager *mqd_mgr; struct queue *queue; uint64_t pending_wptr64; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c index 54933903bcb8..709ac885ca6d 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c @@ -64,7 +64,7 @@ svm_migrate_gart_map(struct amdgpu_ring *ring, uint64_t npages, num_dw = ALIGN(adev->mman.buffer_funcs->copy_num_dw, 8); num_bytes = npages * 8; - r = amdgpu_job_alloc_with_ib(adev, &adev->mman.entity, + r = amdgpu_job_alloc_with_ib(adev, &adev->mman.high_pr, AMDGPU_FENCE_OWNER_UNDEFINED, num_dw * 4 + num_bytes, AMDGPU_IB_POOL_DELAYED, @@ -206,7 +206,7 @@ svm_migrate_copy_done(struct amdgpu_device *adev, struct dma_fence *mfence) unsigned long svm_migrate_addr_to_pfn(struct amdgpu_device *adev, unsigned long addr) { - return (addr + adev->kfd.dev->pgmap.range.start) >> PAGE_SHIFT; + return (addr + adev->kfd.pgmap.range.start) >> PAGE_SHIFT; } static void @@ -236,7 +236,7 @@ svm_migrate_addr(struct amdgpu_device *adev, struct page *page) unsigned long addr; addr = page_to_pfn(page) << PAGE_SHIFT; - return (addr - adev->kfd.dev->pgmap.range.start); + return (addr - adev->kfd.pgmap.range.start); } static struct page * @@ -287,11 +287,12 @@ static unsigned long svm_migrate_unsuccessful_pages(struct migrate_vma *migrate) } static int -svm_migrate_copy_to_vram(struct amdgpu_device *adev, struct svm_range *prange, +svm_migrate_copy_to_vram(struct kfd_node *node, struct svm_range *prange, struct migrate_vma *migrate, struct dma_fence **mfence, dma_addr_t *scratch, uint64_t ttm_res_offset) { - uint64_t npages = migrate->npages; + uint64_t npages = migrate->cpages; + struct amdgpu_device *adev = node->adev; struct device *dev = adev->dev; struct amdgpu_res_cursor cursor; dma_addr_t *src; @@ -321,7 +322,7 @@ svm_migrate_copy_to_vram(struct amdgpu_device *adev, struct svm_range *prange, DMA_TO_DEVICE); r = dma_mapping_error(dev, src[i]); if (r) { - dev_err(adev->dev, "%s: fail %d dma_map_page\n", + dev_err(dev, "%s: fail %d dma_map_page\n", __func__, r); goto out_free_vram_pages; } @@ -390,12 +391,13 @@ out_free_vram_pages: } static long -svm_migrate_vma_to_vram(struct amdgpu_device *adev, struct svm_range *prange, +svm_migrate_vma_to_vram(struct kfd_node *node, struct svm_range *prange, struct vm_area_struct *vma, uint64_t start, uint64_t end, uint32_t trigger, uint64_t ttm_res_offset) { struct kfd_process *p = container_of(prange->svms, struct kfd_process, svms); uint64_t npages = (end - start) >> PAGE_SHIFT; + struct amdgpu_device *adev = node->adev; struct kfd_process_device *pdd; struct dma_fence *mfence = NULL; struct migrate_vma migrate = { 0 }; @@ -421,9 +423,9 @@ svm_migrate_vma_to_vram(struct amdgpu_device *adev, struct svm_range *prange, migrate.dst = migrate.src + npages; scratch = (dma_addr_t *)(migrate.dst + npages); - kfd_smi_event_migration_start(adev->kfd.dev, p->lead_thread->pid, + kfd_smi_event_migration_start(node, p->lead_thread->pid, start >> PAGE_SHIFT, end >> PAGE_SHIFT, - 0, adev->kfd.dev->id, prange->prefetch_loc, + 0, node->id, prange->prefetch_loc, prange->preferred_loc, trigger); r = migrate_vma_setup(&migrate); @@ -445,7 +447,7 @@ svm_migrate_vma_to_vram(struct amdgpu_device *adev, struct svm_range *prange, else pr_debug("0x%lx pages migrated\n", cpages); - r = svm_migrate_copy_to_vram(adev, prange, &migrate, &mfence, scratch, ttm_res_offset); + r = svm_migrate_copy_to_vram(node, prange, &migrate, &mfence, scratch, ttm_res_offset); migrate_vma_pages(&migrate); pr_debug("successful/cpages/npages 0x%lx/0x%lx/0x%lx\n", @@ -454,9 +456,9 @@ svm_migrate_vma_to_vram(struct amdgpu_device *adev, struct svm_range *prange, svm_migrate_copy_done(adev, mfence); migrate_vma_finalize(&migrate); - kfd_smi_event_migration_end(adev->kfd.dev, p->lead_thread->pid, + kfd_smi_event_migration_end(node, p->lead_thread->pid, start >> PAGE_SHIFT, end >> PAGE_SHIFT, - 0, adev->kfd.dev->id, trigger); + 0, node->id, trigger); svm_range_dma_unmap(adev->dev, scratch, 0, npages); svm_range_free_dma_mappings(prange); @@ -465,7 +467,7 @@ out_free: kvfree(buf); out: if (!r && cpages) { - pdd = svm_range_get_pdd_by_adev(prange, adev); + pdd = svm_range_get_pdd_by_node(prange, node); if (pdd) WRITE_ONCE(pdd->page_in, pdd->page_in + cpages); @@ -492,8 +494,8 @@ svm_migrate_ram_to_vram(struct svm_range *prange, uint32_t best_loc, { unsigned long addr, start, end; struct vm_area_struct *vma; - struct amdgpu_device *adev; uint64_t ttm_res_offset; + struct kfd_node *node; unsigned long cpages = 0; long r = 0; @@ -503,9 +505,9 @@ svm_migrate_ram_to_vram(struct svm_range *prange, uint32_t best_loc, return 0; } - adev = svm_range_get_adev_by_id(prange, best_loc); - if (!adev) { - pr_debug("failed to get device by id 0x%x\n", best_loc); + node = svm_range_get_node_by_id(prange, best_loc); + if (!node) { + pr_debug("failed to get kfd node by id 0x%x\n", best_loc); return -ENODEV; } @@ -515,9 +517,9 @@ svm_migrate_ram_to_vram(struct svm_range *prange, uint32_t best_loc, start = prange->start << PAGE_SHIFT; end = (prange->last + 1) << PAGE_SHIFT; - r = svm_range_vram_node_new(adev, prange, true); + r = svm_range_vram_node_new(node, prange, true); if (r) { - dev_dbg(adev->dev, "fail %ld to alloc vram\n", r); + dev_dbg(node->adev->dev, "fail %ld to alloc vram\n", r); return r; } ttm_res_offset = prange->offset << PAGE_SHIFT; @@ -530,7 +532,7 @@ svm_migrate_ram_to_vram(struct svm_range *prange, uint32_t best_loc, break; next = min(vma->vm_end, end); - r = svm_migrate_vma_to_vram(adev, prange, vma, addr, next, trigger, ttm_res_offset); + r = svm_migrate_vma_to_vram(node, prange, vma, addr, next, trigger, ttm_res_offset); if (r < 0) { pr_debug("failed %ld to migrate\n", r); break; @@ -649,11 +651,13 @@ out_oom: /** * svm_migrate_vma_to_ram - migrate range inside one vma from device to system * - * @adev: amdgpu device to migrate from * @prange: svm range structure * @vma: vm_area_struct that range [start, end] belongs to * @start: range start virtual address in pages * @end: range end virtual address in pages + * @node: kfd node device to migrate from + * @trigger: reason of migration + * @fault_page: is from vmf->page, svm_migrate_to_ram(), this is CPU page fault callback * * Context: Process context, caller hold mmap read lock, prange->migrate_mutex * @@ -663,7 +667,7 @@ out_oom: * positive values - partial migration, number of pages not migrated */ static long -svm_migrate_vma_to_ram(struct amdgpu_device *adev, struct svm_range *prange, +svm_migrate_vma_to_ram(struct kfd_node *node, struct svm_range *prange, struct vm_area_struct *vma, uint64_t start, uint64_t end, uint32_t trigger, struct page *fault_page) { @@ -671,6 +675,7 @@ svm_migrate_vma_to_ram(struct amdgpu_device *adev, struct svm_range *prange, uint64_t npages = (end - start) >> PAGE_SHIFT; unsigned long upages = npages; unsigned long cpages = 0; + struct amdgpu_device *adev = node->adev; struct kfd_process_device *pdd; struct dma_fence *mfence = NULL; struct migrate_vma migrate = { 0 }; @@ -699,9 +704,9 @@ svm_migrate_vma_to_ram(struct amdgpu_device *adev, struct svm_range *prange, migrate.fault_page = fault_page; scratch = (dma_addr_t *)(migrate.dst + npages); - kfd_smi_event_migration_start(adev->kfd.dev, p->lead_thread->pid, + kfd_smi_event_migration_start(node, p->lead_thread->pid, start >> PAGE_SHIFT, end >> PAGE_SHIFT, - adev->kfd.dev->id, 0, prange->prefetch_loc, + node->id, 0, prange->prefetch_loc, prange->preferred_loc, trigger); r = migrate_vma_setup(&migrate); @@ -735,9 +740,9 @@ svm_migrate_vma_to_ram(struct amdgpu_device *adev, struct svm_range *prange, svm_migrate_copy_done(adev, mfence); migrate_vma_finalize(&migrate); - kfd_smi_event_migration_end(adev->kfd.dev, p->lead_thread->pid, + kfd_smi_event_migration_end(node, p->lead_thread->pid, start >> PAGE_SHIFT, end >> PAGE_SHIFT, - adev->kfd.dev->id, 0, trigger); + node->id, 0, trigger); svm_range_dma_unmap(adev->dev, scratch, 0, npages); @@ -745,7 +750,7 @@ out_free: kvfree(buf); out: if (!r && cpages) { - pdd = svm_range_get_pdd_by_adev(prange, adev); + pdd = svm_range_get_pdd_by_node(prange, node); if (pdd) WRITE_ONCE(pdd->page_out, pdd->page_out + cpages); } @@ -757,6 +762,7 @@ out: * @prange: range structure * @mm: process mm, use current->mm if NULL * @trigger: reason of migration + * @fault_page: is from vmf->page, svm_migrate_to_ram(), this is CPU page fault callback * * Context: Process context, caller hold mmap read lock, prange->migrate_mutex * @@ -766,7 +772,7 @@ out: int svm_migrate_vram_to_ram(struct svm_range *prange, struct mm_struct *mm, uint32_t trigger, struct page *fault_page) { - struct amdgpu_device *adev; + struct kfd_node *node; struct vm_area_struct *vma; unsigned long addr; unsigned long start; @@ -780,13 +786,11 @@ int svm_migrate_vram_to_ram(struct svm_range *prange, struct mm_struct *mm, return 0; } - adev = svm_range_get_adev_by_id(prange, prange->actual_loc); - if (!adev) { - pr_debug("failed to get device by id 0x%x\n", - prange->actual_loc); + node = svm_range_get_node_by_id(prange, prange->actual_loc); + if (!node) { + pr_debug("failed to get kfd node by id 0x%x\n", prange->actual_loc); return -ENODEV; } - pr_debug("svms 0x%p prange 0x%p [0x%lx 0x%lx] from gpu 0x%x to ram\n", prange->svms, prange, prange->start, prange->last, prange->actual_loc); @@ -805,7 +809,7 @@ int svm_migrate_vram_to_ram(struct svm_range *prange, struct mm_struct *mm, } next = min(vma->vm_end, end); - r = svm_migrate_vma_to_ram(adev, prange, vma, addr, next, trigger, + r = svm_migrate_vma_to_ram(node, prange, vma, addr, next, trigger, fault_page); if (r < 0) { pr_debug("failed %ld to migrate prange %p\n", r, prange); @@ -987,18 +991,21 @@ static const struct dev_pagemap_ops svm_migrate_pgmap_ops = { /* Each VRAM page uses sizeof(struct page) on system memory */ #define SVM_HMM_PAGE_STRUCT_SIZE(size) ((size)/PAGE_SIZE * sizeof(struct page)) -int svm_migrate_init(struct amdgpu_device *adev) +int kgd2kfd_init_zone_device(struct amdgpu_device *adev) { - struct kfd_dev *kfddev = adev->kfd.dev; + struct amdgpu_kfd_dev *kfddev = &adev->kfd; struct dev_pagemap *pgmap; struct resource *res = NULL; unsigned long size; void *r; - /* Page migration works on Vega10 or newer */ - if (!KFD_IS_SOC15(kfddev)) + /* Page migration works on gfx9 or newer */ + if (adev->ip_versions[GC_HWIP][0] < IP_VERSION(9, 0, 1)) return -EINVAL; + if (adev->gmc.is_app_apu) + return 0; + pgmap = &kfddev->pgmap; memset(pgmap, 0, sizeof(*pgmap)); @@ -1041,8 +1048,6 @@ int svm_migrate_init(struct amdgpu_device *adev) amdgpu_amdkfd_reserve_system_mem(SVM_HMM_PAGE_STRUCT_SIZE(size)); - svm_range_set_max_pages(adev); - pr_info("HMM registered %ldMB device memory\n", size >> 20); return 0; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.h b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.h index a5d7e6d22264..487f26368164 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.h @@ -47,15 +47,6 @@ int svm_migrate_vram_to_ram(struct svm_range *prange, struct mm_struct *mm, unsigned long svm_migrate_addr_to_pfn(struct amdgpu_device *adev, unsigned long addr); -int svm_migrate_init(struct amdgpu_device *adev); - -#else - -static inline int svm_migrate_init(struct amdgpu_device *adev) -{ - return 0; -} - #endif /* IS_ENABLED(CONFIG_HSA_AMD_SVM) */ #endif /* KFD_MIGRATE_H_ */ diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c index 623ccd227b7d..863cf060af48 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c @@ -46,7 +46,7 @@ int pipe_priority_map[] = { KFD_PIPE_PRIORITY_CS_HIGH }; -struct kfd_mem_obj *allocate_hiq_mqd(struct kfd_dev *dev, struct queue_properties *q) +struct kfd_mem_obj *allocate_hiq_mqd(struct kfd_node *dev, struct queue_properties *q) { struct kfd_mem_obj *mqd_mem_obj = NULL; @@ -61,7 +61,7 @@ struct kfd_mem_obj *allocate_hiq_mqd(struct kfd_dev *dev, struct queue_propertie return mqd_mem_obj; } -struct kfd_mem_obj *allocate_sdma_mqd(struct kfd_dev *dev, +struct kfd_mem_obj *allocate_sdma_mqd(struct kfd_node *dev, struct queue_properties *q) { struct kfd_mem_obj *mqd_mem_obj = NULL; @@ -72,11 +72,12 @@ struct kfd_mem_obj *allocate_sdma_mqd(struct kfd_dev *dev, return NULL; offset = (q->sdma_engine_id * - dev->device_info.num_sdma_queues_per_engine + + dev->kfd->device_info.num_sdma_queues_per_engine + q->sdma_queue_id) * dev->dqm->mqd_mgrs[KFD_MQD_TYPE_SDMA]->mqd_size; - offset += dev->dqm->mqd_mgrs[KFD_MQD_TYPE_HIQ]->mqd_size; + offset += dev->dqm->mqd_mgrs[KFD_MQD_TYPE_HIQ]->mqd_size * + NUM_XCC(dev->xcc_mask); mqd_mem_obj->gtt_mem = (void *)((uint64_t)dev->dqm->hiq_sdma_mqd.gtt_mem + offset); @@ -189,7 +190,7 @@ int kfd_hiq_load_mqd_kiq(struct mqd_manager *mm, void *mqd, struct queue_properties *p, struct mm_struct *mms) { return mm->dev->kfd2kgd->hiq_mqd_load(mm->dev->adev, mqd, pipe_id, - queue_id, p->doorbell_off); + queue_id, p->doorbell_off, 0); } int kfd_destroy_mqd_cp(struct mqd_manager *mm, void *mqd, @@ -197,7 +198,7 @@ int kfd_destroy_mqd_cp(struct mqd_manager *mm, void *mqd, uint32_t pipe_id, uint32_t queue_id) { return mm->dev->kfd2kgd->hqd_destroy(mm->dev->adev, mqd, type, timeout, - pipe_id, queue_id); + pipe_id, queue_id, 0); } void kfd_free_mqd_cp(struct mqd_manager *mm, void *mqd, @@ -216,7 +217,7 @@ bool kfd_is_occupied_cp(struct mqd_manager *mm, void *mqd, uint32_t queue_id) { return mm->dev->kfd2kgd->hqd_is_occupied(mm->dev->adev, queue_address, - pipe_id, queue_id); + pipe_id, queue_id, 0); } int kfd_load_mqd_sdma(struct mqd_manager *mm, void *mqd, @@ -246,3 +247,28 @@ bool kfd_is_occupied_sdma(struct mqd_manager *mm, void *mqd, { return mm->dev->kfd2kgd->hqd_sdma_is_occupied(mm->dev->adev, mqd); } + +uint64_t kfd_hiq_mqd_stride(struct kfd_node *dev) +{ + return dev->dqm->mqd_mgrs[KFD_MQD_TYPE_HIQ]->mqd_size; +} + +void kfd_get_hiq_xcc_mqd(struct kfd_node *dev, struct kfd_mem_obj *mqd_mem_obj, + uint32_t virtual_xcc_id) +{ + uint64_t offset; + + offset = kfd_hiq_mqd_stride(dev) * virtual_xcc_id; + + mqd_mem_obj->gtt_mem = (virtual_xcc_id == 0) ? + dev->dqm->hiq_sdma_mqd.gtt_mem : NULL; + mqd_mem_obj->gpu_addr = dev->dqm->hiq_sdma_mqd.gpu_addr + offset; + mqd_mem_obj->cpu_ptr = (uint32_t *)((uintptr_t) + dev->dqm->hiq_sdma_mqd.cpu_ptr + offset); +} + +uint64_t kfd_mqd_stride(struct mqd_manager *mm, + struct queue_properties *q) +{ + return mm->mqd_size; +} diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h index 57f900ccaa10..23158db7da03 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h @@ -68,7 +68,7 @@ */ extern int pipe_priority_map[]; struct mqd_manager { - struct kfd_mem_obj* (*allocate_mqd)(struct kfd_dev *kfd, + struct kfd_mem_obj* (*allocate_mqd)(struct kfd_node *kfd, struct queue_properties *q); void (*init_mqd)(struct mqd_manager *mm, void **mqd, @@ -97,6 +97,7 @@ struct mqd_manager { uint32_t queue_id); int (*get_wave_state)(struct mqd_manager *mm, void *mqd, + struct queue_properties *q, void __user *ctl_stack, u32 *ctl_stack_used_size, u32 *save_area_used_size); @@ -119,16 +120,18 @@ struct mqd_manager { int (*debugfs_show_mqd)(struct seq_file *m, void *data); #endif uint32_t (*read_doorbell_id)(void *mqd); + uint64_t (*mqd_stride)(struct mqd_manager *mm, + struct queue_properties *p); struct mutex mqd_mutex; - struct kfd_dev *dev; + struct kfd_node *dev; uint32_t mqd_size; }; -struct kfd_mem_obj *allocate_hiq_mqd(struct kfd_dev *dev, +struct kfd_mem_obj *allocate_hiq_mqd(struct kfd_node *dev, struct queue_properties *q); -struct kfd_mem_obj *allocate_sdma_mqd(struct kfd_dev *dev, +struct kfd_mem_obj *allocate_sdma_mqd(struct kfd_node *dev, struct queue_properties *q); void free_mqd_hiq_sdma(struct mqd_manager *mm, void *mqd, struct kfd_mem_obj *mqd_mem_obj); @@ -164,4 +167,10 @@ bool kfd_is_occupied_sdma(struct mqd_manager *mm, void *mqd, uint64_t queue_address, uint32_t pipe_id, uint32_t queue_id); +void kfd_get_hiq_xcc_mqd(struct kfd_node *dev, + struct kfd_mem_obj *mqd_mem_obj, uint32_t virtual_xcc_id); + +uint64_t kfd_hiq_mqd_stride(struct kfd_node *dev); +uint64_t kfd_mqd_stride(struct mqd_manager *mm, + struct queue_properties *q); #endif /* KFD_MQD_MANAGER_H_ */ diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c index 4889865c725c..65c9f01a1f86 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c @@ -48,8 +48,7 @@ static void update_cu_mask(struct mqd_manager *mm, void *mqd, struct cik_mqd *m; uint32_t se_mask[4] = {0}; /* 4 is the max # of SEs */ - if (!minfo || (minfo->update_flag != UPDATE_FLAG_CU_MASK) || - !minfo->cu_mask.ptr) + if (!minfo || !minfo->cu_mask.ptr) return; mqd_symmetrically_map_cu_mask(mm, @@ -74,7 +73,7 @@ static void set_priority(struct cik_mqd *m, struct queue_properties *q) m->cp_hqd_queue_priority = q->priority; } -static struct kfd_mem_obj *allocate_mqd(struct kfd_dev *kfd, +static struct kfd_mem_obj *allocate_mqd(struct kfd_node *kfd, struct queue_properties *q) { struct kfd_mem_obj *mqd_mem_obj; @@ -167,7 +166,7 @@ static int load_mqd(struct mqd_manager *mm, void *mqd, uint32_t pipe_id, return mm->dev->kfd2kgd->hqd_load(mm->dev->adev, mqd, pipe_id, queue_id, (uint32_t __user *)p->write_ptr, - wptr_shift, wptr_mask, mms); + wptr_shift, wptr_mask, mms, 0); } static void __update_mqd(struct mqd_manager *mm, void *mqd, @@ -390,7 +389,7 @@ static int debugfs_show_mqd_sdma(struct seq_file *m, void *data) struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type, - struct kfd_dev *dev) + struct kfd_node *dev) { struct mqd_manager *mqd; @@ -428,6 +427,7 @@ struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type, mqd->destroy_mqd = kfd_destroy_mqd_cp; mqd->is_occupied = kfd_is_occupied_cp; mqd->mqd_size = sizeof(struct cik_mqd); + mqd->mqd_stride = kfd_mqd_stride; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd; #endif @@ -442,6 +442,7 @@ struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type, mqd->destroy_mqd = kfd_destroy_mqd_cp; mqd->is_occupied = kfd_is_occupied_cp; mqd->mqd_size = sizeof(struct cik_mqd); + mqd->mqd_stride = kfd_mqd_stride; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd; #endif @@ -457,6 +458,7 @@ struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type, mqd->checkpoint_mqd = checkpoint_mqd_sdma; mqd->restore_mqd = restore_mqd_sdma; mqd->mqd_size = sizeof(struct cik_sdma_rlc_registers); + mqd->mqd_stride = kfd_mqd_stride; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd_sdma; #endif @@ -470,7 +472,7 @@ struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type, } struct mqd_manager *mqd_manager_init_cik_hawaii(enum KFD_MQD_TYPE type, - struct kfd_dev *dev) + struct kfd_node *dev) { struct mqd_manager *mqd; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v10.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v10.c index d3e2b6a599a4..94c0fc2e57b7 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v10.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v10.c @@ -48,8 +48,7 @@ static void update_cu_mask(struct mqd_manager *mm, void *mqd, struct v10_compute_mqd *m; uint32_t se_mask[4] = {0}; /* 4 is the max # of SEs */ - if (!minfo || (minfo->update_flag != UPDATE_FLAG_CU_MASK) || - !minfo->cu_mask.ptr) + if (!minfo || !minfo->cu_mask.ptr) return; mqd_symmetrically_map_cu_mask(mm, @@ -74,7 +73,7 @@ static void set_priority(struct v10_compute_mqd *m, struct queue_properties *q) m->cp_hqd_queue_priority = q->priority; } -static struct kfd_mem_obj *allocate_mqd(struct kfd_dev *kfd, +static struct kfd_mem_obj *allocate_mqd(struct kfd_node *kfd, struct queue_properties *q) { struct kfd_mem_obj *mqd_mem_obj; @@ -117,12 +116,17 @@ static void init_mqd(struct mqd_manager *mm, void **mqd, 1 << CP_HQD_QUANTUM__QUANTUM_SCALE__SHIFT | 1 << CP_HQD_QUANTUM__QUANTUM_DURATION__SHIFT; + /* Set cp_hqd_hq_scheduler0 bit 14 to 1 to have the CP set up the + * DISPATCH_PTR. This is required for the kfd debugger + */ + m->cp_hqd_hq_scheduler0 = 1 << 14; + if (q->format == KFD_QUEUE_FORMAT_AQL) { m->cp_hqd_aql_control = 1 << CP_HQD_AQL_CONTROL__CONTROL0__SHIFT; } - if (mm->dev->cwsr_enabled) { + if (mm->dev->kfd->cwsr_enabled) { m->cp_hqd_persistent_state |= (1 << CP_HQD_PERSISTENT_STATE__QSWITCH_MODE__SHIFT); m->cp_hqd_ctx_save_base_addr_lo = @@ -151,7 +155,7 @@ static int load_mqd(struct mqd_manager *mm, void *mqd, r = mm->dev->kfd2kgd->hqd_load(mm->dev->adev, mqd, pipe_id, queue_id, (uint32_t __user *)p->write_ptr, - wptr_shift, 0, mms); + wptr_shift, 0, mms, 0); return r; } @@ -210,7 +214,7 @@ static void update_mqd(struct mqd_manager *mm, void *mqd, m->cp_hqd_pq_doorbell_control |= 1 << CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_BIF_DROP__SHIFT; } - if (mm->dev->cwsr_enabled) + if (mm->dev->kfd->cwsr_enabled) m->cp_hqd_ctx_save_control = 0; update_cu_mask(mm, mqd, minfo); @@ -227,11 +231,13 @@ static uint32_t read_doorbell_id(void *mqd) } static int get_wave_state(struct mqd_manager *mm, void *mqd, + struct queue_properties *q, void __user *ctl_stack, u32 *ctl_stack_used_size, u32 *save_area_used_size) { struct v10_compute_mqd *m; + struct kfd_context_save_area_header header; m = get_mqd(mqd); @@ -250,6 +256,15 @@ static int get_wave_state(struct mqd_manager *mm, void *mqd, * accessible to user mode */ + header.wave_state.control_stack_size = *ctl_stack_used_size; + header.wave_state.wave_state_size = *save_area_used_size; + + header.wave_state.wave_state_offset = m->cp_hqd_wg_state_offset; + header.wave_state.control_stack_offset = m->cp_hqd_cntl_stack_offset; + + if (copy_to_user(ctl_stack, &header, sizeof(header.wave_state))) + return -EFAULT; + return 0; } @@ -405,7 +420,7 @@ static int debugfs_show_mqd_sdma(struct seq_file *m, void *data) #endif struct mqd_manager *mqd_manager_init_v10(enum KFD_MQD_TYPE type, - struct kfd_dev *dev) + struct kfd_node *dev) { struct mqd_manager *mqd; @@ -432,6 +447,7 @@ struct mqd_manager *mqd_manager_init_v10(enum KFD_MQD_TYPE type, mqd->get_wave_state = get_wave_state; mqd->checkpoint_mqd = checkpoint_mqd; mqd->restore_mqd = restore_mqd; + mqd->mqd_stride = kfd_mqd_stride; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd; #endif @@ -447,6 +463,7 @@ struct mqd_manager *mqd_manager_init_v10(enum KFD_MQD_TYPE type, mqd->destroy_mqd = kfd_destroy_mqd_cp; mqd->is_occupied = kfd_is_occupied_cp; mqd->mqd_size = sizeof(struct v10_compute_mqd); + mqd->mqd_stride = kfd_mqd_stride; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd; #endif @@ -478,6 +495,7 @@ struct mqd_manager *mqd_manager_init_v10(enum KFD_MQD_TYPE type, mqd->checkpoint_mqd = checkpoint_mqd_sdma; mqd->restore_mqd = restore_mqd_sdma; mqd->mqd_size = sizeof(struct v10_sdma_mqd); + mqd->mqd_stride = kfd_mqd_stride; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd_sdma; #endif diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c index 5aa75f72caa1..31fec5e70d13 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c @@ -46,15 +46,33 @@ static void update_cu_mask(struct mqd_manager *mm, void *mqd, { struct v11_compute_mqd *m; uint32_t se_mask[KFD_MAX_NUM_SE] = {0}; + bool has_wa_flag = minfo && (minfo->update_flag & (UPDATE_FLAG_DBG_WA_ENABLE | + UPDATE_FLAG_DBG_WA_DISABLE)); - if (!minfo || (minfo->update_flag != UPDATE_FLAG_CU_MASK) || - !minfo->cu_mask.ptr) + if (!minfo || !(has_wa_flag || minfo->cu_mask.ptr)) return; + m = get_mqd(mqd); + + if (has_wa_flag) { + uint32_t wa_mask = minfo->update_flag == UPDATE_FLAG_DBG_WA_ENABLE ? + 0xffff : 0xffffffff; + + m->compute_static_thread_mgmt_se0 = wa_mask; + m->compute_static_thread_mgmt_se1 = wa_mask; + m->compute_static_thread_mgmt_se2 = wa_mask; + m->compute_static_thread_mgmt_se3 = wa_mask; + m->compute_static_thread_mgmt_se4 = wa_mask; + m->compute_static_thread_mgmt_se5 = wa_mask; + m->compute_static_thread_mgmt_se6 = wa_mask; + m->compute_static_thread_mgmt_se7 = wa_mask; + + return; + } + mqd_symmetrically_map_cu_mask(mm, minfo->cu_mask.ptr, minfo->cu_mask.count, se_mask); - m = get_mqd(mqd); m->compute_static_thread_mgmt_se0 = se_mask[0]; m->compute_static_thread_mgmt_se1 = se_mask[1]; m->compute_static_thread_mgmt_se2 = se_mask[2]; @@ -81,7 +99,7 @@ static void set_priority(struct v11_compute_mqd *m, struct queue_properties *q) m->cp_hqd_queue_priority = q->priority; } -static struct kfd_mem_obj *allocate_mqd(struct kfd_dev *kfd, +static struct kfd_mem_obj *allocate_mqd(struct kfd_node *node, struct queue_properties *q) { struct kfd_mem_obj *mqd_mem_obj; @@ -91,12 +109,12 @@ static struct kfd_mem_obj *allocate_mqd(struct kfd_dev *kfd, * MES write to areas beyond MQD size. So allocate * 1 PAGE_SIZE memory for MQD is MES is enabled. */ - if (kfd->shared_resources.enable_mes) + if (node->kfd->shared_resources.enable_mes) size = PAGE_SIZE; else size = sizeof(struct v11_compute_mqd); - if (kfd_gtt_sa_allocate(kfd, size, &mqd_mem_obj)) + if (kfd_gtt_sa_allocate(node, size, &mqd_mem_obj)) return NULL; return mqd_mem_obj; @@ -109,11 +127,12 @@ static void init_mqd(struct mqd_manager *mm, void **mqd, uint64_t addr; struct v11_compute_mqd *m; int size; + uint32_t wa_mask = q->is_dbg_wa ? 0xffff : 0xffffffff; m = (struct v11_compute_mqd *) mqd_mem_obj->cpu_ptr; addr = mqd_mem_obj->gpu_addr; - if (mm->dev->shared_resources.enable_mes) + if (mm->dev->kfd->shared_resources.enable_mes) size = PAGE_SIZE; else size = sizeof(struct v11_compute_mqd); @@ -122,14 +141,15 @@ static void init_mqd(struct mqd_manager *mm, void **mqd, m->header = 0xC0310800; m->compute_pipelinestat_enable = 1; - m->compute_static_thread_mgmt_se0 = 0xFFFFFFFF; - m->compute_static_thread_mgmt_se1 = 0xFFFFFFFF; - m->compute_static_thread_mgmt_se2 = 0xFFFFFFFF; - m->compute_static_thread_mgmt_se3 = 0xFFFFFFFF; - m->compute_static_thread_mgmt_se4 = 0xFFFFFFFF; - m->compute_static_thread_mgmt_se5 = 0xFFFFFFFF; - m->compute_static_thread_mgmt_se6 = 0xFFFFFFFF; - m->compute_static_thread_mgmt_se7 = 0xFFFFFFFF; + + m->compute_static_thread_mgmt_se0 = wa_mask; + m->compute_static_thread_mgmt_se1 = wa_mask; + m->compute_static_thread_mgmt_se2 = wa_mask; + m->compute_static_thread_mgmt_se3 = wa_mask; + m->compute_static_thread_mgmt_se4 = wa_mask; + m->compute_static_thread_mgmt_se5 = wa_mask; + m->compute_static_thread_mgmt_se6 = wa_mask; + m->compute_static_thread_mgmt_se7 = wa_mask; m->cp_hqd_persistent_state = CP_HQD_PERSISTENT_STATE__PRELOAD_REQ_MASK | 0x55 << CP_HQD_PERSISTENT_STATE__PRELOAD_SIZE__SHIFT; @@ -143,6 +163,11 @@ static void init_mqd(struct mqd_manager *mm, void **mqd, 1 << CP_HQD_QUANTUM__QUANTUM_SCALE__SHIFT | 1 << CP_HQD_QUANTUM__QUANTUM_DURATION__SHIFT; + /* Set cp_hqd_hq_scheduler0 bit 14 to 1 to have the CP set up the + * DISPATCH_PTR. This is required for the kfd debugger + */ + m->cp_hqd_hq_status0 = 1 << 14; + /* * GFX11 RS64 CPFW version >= 509 supports PCIe atomics support * acknowledgment. @@ -155,7 +180,7 @@ static void init_mqd(struct mqd_manager *mm, void **mqd, 1 << CP_HQD_AQL_CONTROL__CONTROL0__SHIFT; } - if (mm->dev->cwsr_enabled) { + if (mm->dev->kfd->cwsr_enabled) { m->cp_hqd_persistent_state |= (1 << CP_HQD_PERSISTENT_STATE__QSWITCH_MODE__SHIFT); m->cp_hqd_ctx_save_base_addr_lo = @@ -184,7 +209,7 @@ static int load_mqd(struct mqd_manager *mm, void *mqd, r = mm->dev->kfd2kgd->hqd_load(mm->dev->adev, mqd, pipe_id, queue_id, (uint32_t __user *)p->write_ptr, - wptr_shift, 0, mms); + wptr_shift, 0, mms, 0); return r; } @@ -243,7 +268,7 @@ static void update_mqd(struct mqd_manager *mm, void *mqd, m->cp_hqd_pq_doorbell_control |= 1 << CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_BIF_DROP__SHIFT; } - if (mm->dev->cwsr_enabled) + if (mm->dev->kfd->cwsr_enabled) m->cp_hqd_ctx_save_control = 0; update_cu_mask(mm, mqd, minfo); @@ -260,12 +285,13 @@ static uint32_t read_doorbell_id(void *mqd) } static int get_wave_state(struct mqd_manager *mm, void *mqd, + struct queue_properties *q, void __user *ctl_stack, u32 *ctl_stack_used_size, u32 *save_area_used_size) { struct v11_compute_mqd *m; - /*struct mqd_user_context_save_area_header header;*/ + struct kfd_context_save_area_header header; m = get_mqd(mqd); @@ -283,16 +309,15 @@ static int get_wave_state(struct mqd_manager *mm, void *mqd, * it's part of the context save area that is already * accessible to user mode */ -/* - header.control_stack_size = *ctl_stack_used_size; - header.wave_state_size = *save_area_used_size; + header.wave_state.control_stack_size = *ctl_stack_used_size; + header.wave_state.wave_state_size = *save_area_used_size; - header.wave_state_offset = m->cp_hqd_wg_state_offset; - header.control_stack_offset = m->cp_hqd_cntl_stack_offset; + header.wave_state.wave_state_offset = m->cp_hqd_wg_state_offset; + header.wave_state.control_stack_offset = m->cp_hqd_cntl_stack_offset; - if (copy_to_user(ctl_stack, &header, sizeof(header))) + if (copy_to_user(ctl_stack, &header, sizeof(header.wave_state))) return -EFAULT; -*/ + return 0; } @@ -319,7 +344,7 @@ static void init_mqd_sdma(struct mqd_manager *mm, void **mqd, m = (struct v11_sdma_mqd *) mqd_mem_obj->cpu_ptr; - if (mm->dev->shared_resources.enable_mes) + if (mm->dev->kfd->shared_resources.enable_mes) size = PAGE_SIZE; else size = sizeof(struct v11_sdma_mqd); @@ -387,7 +412,7 @@ static int debugfs_show_mqd_sdma(struct seq_file *m, void *data) #endif struct mqd_manager *mqd_manager_init_v11(enum KFD_MQD_TYPE type, - struct kfd_dev *dev) + struct kfd_node *dev) { struct mqd_manager *mqd; @@ -463,7 +488,7 @@ struct mqd_manager *mqd_manager_init_v11(enum KFD_MQD_TYPE type, * To allocate SDMA MQDs by generic functions * when MES is enabled. */ - if (dev->shared_resources.enable_mes) { + if (dev->kfd->shared_resources.enable_mes) { mqd->allocate_mqd = allocate_mqd; mqd->free_mqd = kfd_free_mqd_cp; } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c index fdbfd725841f..601bb9f68048 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c @@ -32,6 +32,22 @@ #include "gc/gc_9_0_sh_mask.h" #include "sdma0/sdma0_4_0_sh_mask.h" #include "amdgpu_amdkfd.h" +#include "kfd_device_queue_manager.h" + +static void update_mqd(struct mqd_manager *mm, void *mqd, + struct queue_properties *q, + struct mqd_update_info *minfo); + +static uint64_t mqd_stride_v9(struct mqd_manager *mm, + struct queue_properties *q) +{ + if (mm->dev->kfd->cwsr_enabled && + q->type == KFD_QUEUE_TYPE_COMPUTE) + return ALIGN(q->ctl_stack_size, PAGE_SIZE) + + ALIGN(sizeof(struct v9_mqd), PAGE_SIZE); + + return mm->mqd_size; +} static inline struct v9_mqd *get_mqd(void *mqd) { @@ -49,8 +65,7 @@ static void update_cu_mask(struct mqd_manager *mm, void *mqd, struct v9_mqd *m; uint32_t se_mask[KFD_MAX_NUM_SE] = {0}; - if (!minfo || (minfo->update_flag != UPDATE_FLAG_CU_MASK) || - !minfo->cu_mask.ptr) + if (!minfo || !minfo->cu_mask.ptr) return; mqd_symmetrically_map_cu_mask(mm, @@ -83,7 +98,7 @@ static void set_priority(struct v9_mqd *m, struct queue_properties *q) m->cp_hqd_queue_priority = q->priority; } -static struct kfd_mem_obj *allocate_mqd(struct kfd_dev *kfd, +static struct kfd_mem_obj *allocate_mqd(struct kfd_node *node, struct queue_properties *q) { int retval; @@ -105,28 +120,30 @@ static struct kfd_mem_obj *allocate_mqd(struct kfd_dev *kfd, * pass a special bo flag AMDGPU_GEM_CREATE_CP_MQD_GFX9 to instruct * amdgpu memory functions to do so. */ - if (kfd->cwsr_enabled && (q->type == KFD_QUEUE_TYPE_COMPUTE)) { + if (node->kfd->cwsr_enabled && (q->type == KFD_QUEUE_TYPE_COMPUTE)) { mqd_mem_obj = kzalloc(sizeof(struct kfd_mem_obj), GFP_KERNEL); if (!mqd_mem_obj) return NULL; - retval = amdgpu_amdkfd_alloc_gtt_mem(kfd->adev, - ALIGN(q->ctl_stack_size, PAGE_SIZE) + - ALIGN(sizeof(struct v9_mqd), PAGE_SIZE), + retval = amdgpu_amdkfd_alloc_gtt_mem(node->adev, + (ALIGN(q->ctl_stack_size, PAGE_SIZE) + + ALIGN(sizeof(struct v9_mqd), PAGE_SIZE)) * + NUM_XCC(node->xcc_mask), &(mqd_mem_obj->gtt_mem), &(mqd_mem_obj->gpu_addr), (void *)&(mqd_mem_obj->cpu_ptr), true); + + if (retval) { + kfree(mqd_mem_obj); + return NULL; + } } else { - retval = kfd_gtt_sa_allocate(kfd, sizeof(struct v9_mqd), + retval = kfd_gtt_sa_allocate(node, sizeof(struct v9_mqd), &mqd_mem_obj); - } - - if (retval) { - kfree(mqd_mem_obj); - return NULL; + if (retval) + return NULL; } return mqd_mem_obj; - } static void init_mqd(struct mqd_manager *mm, void **mqd, @@ -135,7 +152,6 @@ static void init_mqd(struct mqd_manager *mm, void **mqd, { uint64_t addr; struct v9_mqd *m; - struct amdgpu_device *adev = (struct amdgpu_device *)mm->dev->adev; m = (struct v9_mqd *) mqd_mem_obj->cpu_ptr; addr = mqd_mem_obj->gpu_addr; @@ -165,31 +181,21 @@ static void init_mqd(struct mqd_manager *mm, void **mqd, 1 << CP_HQD_QUANTUM__QUANTUM_SCALE__SHIFT | 1 << CP_HQD_QUANTUM__QUANTUM_DURATION__SHIFT; - if (q->format == KFD_QUEUE_FORMAT_AQL) { + /* Set cp_hqd_hq_scheduler0 bit 14 to 1 to have the CP set up the + * DISPATCH_PTR. This is required for the kfd debugger + */ + m->cp_hqd_hq_status0 = 1 << 14; + + if (q->format == KFD_QUEUE_FORMAT_AQL) m->cp_hqd_aql_control = 1 << CP_HQD_AQL_CONTROL__CONTROL0__SHIFT; - if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3)) { - /* On GC 9.4.3, DW 41 is re-purposed as - * compute_tg_chunk_size. - * TODO: review this setting when active CUs in the - * partition play a role - */ - m->compute_static_thread_mgmt_se6 = 1; - } - } else { - /* PM4 queue */ - if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 3)) { - m->compute_static_thread_mgmt_se6 = 0; - /* TODO: program pm4_target_xcc */ - } - } if (q->tba_addr) { m->compute_pgm_rsrc2 |= (1 << COMPUTE_PGM_RSRC2__TRAP_PRESENT__SHIFT); } - if (mm->dev->cwsr_enabled && q->ctx_save_restore_area_address) { + if (mm->dev->kfd->cwsr_enabled && q->ctx_save_restore_area_address) { m->cp_hqd_persistent_state |= (1 << CP_HQD_PERSISTENT_STATE__QSWITCH_MODE__SHIFT); m->cp_hqd_ctx_save_base_addr_lo = @@ -205,7 +211,7 @@ static void init_mqd(struct mqd_manager *mm, void **mqd, *mqd = m; if (gart_addr) *gart_addr = addr; - mm->update_mqd(mm, m, q, NULL); + update_mqd(mm, m, q, NULL); } static int load_mqd(struct mqd_manager *mm, void *mqd, @@ -217,14 +223,13 @@ static int load_mqd(struct mqd_manager *mm, void *mqd, return mm->dev->kfd2kgd->hqd_load(mm->dev->adev, mqd, pipe_id, queue_id, (uint32_t __user *)p->write_ptr, - wptr_shift, 0, mms); + wptr_shift, 0, mms, 0); } static void update_mqd(struct mqd_manager *mm, void *mqd, struct queue_properties *q, struct mqd_update_info *minfo) { - struct amdgpu_device *adev = (struct amdgpu_device *)mm->dev->adev; struct v9_mqd *m; m = get_mqd(mqd); @@ -257,9 +262,14 @@ static void update_mqd(struct mqd_manager *mm, void *mqd, * Limit is 0xFF EOP entries (= 0x7F8 dwords). CP will not submit * more than (EOP entry count - 1) so a queue size of 0x800 dwords * is safe, giving a maximum field value of 0xA. + * + * Also, do calculation only if EOP is used (size > 0), otherwise + * the order_base_2 calculation provides incorrect result. + * */ - m->cp_hqd_eop_control = min(0xA, - order_base_2(q->eop_ring_buffer_size / 4) - 1); + m->cp_hqd_eop_control = q->eop_ring_buffer_size ? + min(0xA, order_base_2(q->eop_ring_buffer_size / 4) - 1) : 0; + m->cp_hqd_eop_base_addr_lo = lower_32_bits(q->eop_ring_buffer_address >> 8); m->cp_hqd_eop_base_addr_hi = @@ -270,17 +280,14 @@ static void update_mqd(struct mqd_manager *mm, void *mqd, m->cp_hqd_vmid = q->vmid; if (q->format == KFD_QUEUE_FORMAT_AQL) { - m->cp_hqd_pq_control |= + m->cp_hqd_pq_control |= CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK | 2 << CP_HQD_PQ_CONTROL__SLOT_BASED_WPTR__SHIFT | 1 << CP_HQD_PQ_CONTROL__QUEUE_FULL_EN__SHIFT | 1 << CP_HQD_PQ_CONTROL__WPP_CLAMP_EN__SHIFT; - if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 3)) - m->cp_hqd_pq_control |= - CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK; m->cp_hqd_pq_doorbell_control |= 1 << CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_BIF_DROP__SHIFT; } - if (mm->dev->cwsr_enabled && q->ctx_save_restore_area_address) + if (mm->dev->kfd->cwsr_enabled && q->ctx_save_restore_area_address) m->cp_hqd_ctx_save_control = 0; update_cu_mask(mm, mqd, minfo); @@ -298,11 +305,13 @@ static uint32_t read_doorbell_id(void *mqd) } static int get_wave_state(struct mqd_manager *mm, void *mqd, + struct queue_properties *q, void __user *ctl_stack, u32 *ctl_stack_used_size, u32 *save_area_used_size) { struct v9_mqd *m; + struct kfd_context_save_area_header header; /* Control stack is located one page after MQD. */ void *mqd_ctl_stack = (void *)((uintptr_t)mqd + PAGE_SIZE); @@ -314,7 +323,18 @@ static int get_wave_state(struct mqd_manager *mm, void *mqd, *save_area_used_size = m->cp_hqd_wg_state_offset - m->cp_hqd_cntl_stack_size; - if (copy_to_user(ctl_stack, mqd_ctl_stack, m->cp_hqd_cntl_stack_size)) + header.wave_state.control_stack_size = *ctl_stack_used_size; + header.wave_state.wave_state_size = *save_area_used_size; + + header.wave_state.wave_state_offset = m->cp_hqd_wg_state_offset; + header.wave_state.control_stack_offset = m->cp_hqd_cntl_stack_offset; + + if (copy_to_user(ctl_stack, &header, sizeof(header.wave_state))) + return -EFAULT; + + if (copy_to_user(ctl_stack + m->cp_hqd_cntl_stack_offset, + mqd_ctl_stack + m->cp_hqd_cntl_stack_offset, + *ctl_stack_used_size)) return -EFAULT; return 0; @@ -467,6 +487,288 @@ static void restore_mqd_sdma(struct mqd_manager *mm, void **mqd, qp->is_active = 0; } +static void init_mqd_hiq_v9_4_3(struct mqd_manager *mm, void **mqd, + struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr, + struct queue_properties *q) +{ + struct v9_mqd *m; + int xcc = 0; + struct kfd_mem_obj xcc_mqd_mem_obj; + uint64_t xcc_gart_addr = 0; + + memset(&xcc_mqd_mem_obj, 0x0, sizeof(struct kfd_mem_obj)); + + for (xcc = 0; xcc < NUM_XCC(mm->dev->xcc_mask); xcc++) { + kfd_get_hiq_xcc_mqd(mm->dev, &xcc_mqd_mem_obj, xcc); + + init_mqd(mm, (void **)&m, &xcc_mqd_mem_obj, &xcc_gart_addr, q); + + m->cp_hqd_pq_control |= CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK | + 1 << CP_HQD_PQ_CONTROL__PRIV_STATE__SHIFT | + 1 << CP_HQD_PQ_CONTROL__KMD_QUEUE__SHIFT; + m->cp_mqd_stride_size = kfd_hiq_mqd_stride(mm->dev); + if (xcc == 0) { + /* Set no_update_rptr = 0 in Master XCC */ + m->cp_hqd_pq_control &= ~CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK; + + /* Set the MQD pointer and gart address to XCC0 MQD */ + *mqd = m; + *gart_addr = xcc_gart_addr; + } + } +} + +static int hiq_load_mqd_kiq_v9_4_3(struct mqd_manager *mm, void *mqd, + uint32_t pipe_id, uint32_t queue_id, + struct queue_properties *p, struct mm_struct *mms) +{ + uint32_t xcc_mask = mm->dev->xcc_mask; + int xcc_id, err, inst = 0; + void *xcc_mqd; + uint64_t hiq_mqd_size = kfd_hiq_mqd_stride(mm->dev); + + for_each_inst(xcc_id, xcc_mask) { + xcc_mqd = mqd + hiq_mqd_size * inst; + err = mm->dev->kfd2kgd->hiq_mqd_load(mm->dev->adev, xcc_mqd, + pipe_id, queue_id, + p->doorbell_off, xcc_id); + if (err) { + pr_debug("Failed to load HIQ MQD for XCC: %d\n", inst); + break; + } + ++inst; + } + + return err; +} + +static int destroy_hiq_mqd_v9_4_3(struct mqd_manager *mm, void *mqd, + enum kfd_preempt_type type, unsigned int timeout, + uint32_t pipe_id, uint32_t queue_id) +{ + uint32_t xcc_mask = mm->dev->xcc_mask; + int xcc_id, err, inst = 0; + void *xcc_mqd; + uint64_t hiq_mqd_size = kfd_hiq_mqd_stride(mm->dev); + + for_each_inst(xcc_id, xcc_mask) { + xcc_mqd = mqd + hiq_mqd_size * inst; + err = mm->dev->kfd2kgd->hqd_destroy(mm->dev->adev, xcc_mqd, + type, timeout, pipe_id, + queue_id, xcc_id); + if (err) { + pr_debug("Destroy MQD failed for xcc: %d\n", inst); + break; + } + ++inst; + } + + return err; +} + +static void get_xcc_mqd(struct kfd_mem_obj *mqd_mem_obj, + struct kfd_mem_obj *xcc_mqd_mem_obj, + uint64_t offset) +{ + xcc_mqd_mem_obj->gtt_mem = (offset == 0) ? + mqd_mem_obj->gtt_mem : NULL; + xcc_mqd_mem_obj->gpu_addr = mqd_mem_obj->gpu_addr + offset; + xcc_mqd_mem_obj->cpu_ptr = (uint32_t *)((uintptr_t)mqd_mem_obj->cpu_ptr + + offset); +} + +static void init_mqd_v9_4_3(struct mqd_manager *mm, void **mqd, + struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr, + struct queue_properties *q) +{ + struct v9_mqd *m; + int xcc = 0; + struct kfd_mem_obj xcc_mqd_mem_obj; + uint64_t xcc_gart_addr = 0; + uint64_t xcc_ctx_save_restore_area_address; + uint64_t offset = mm->mqd_stride(mm, q); + uint32_t local_xcc_start = mm->dev->dqm->current_logical_xcc_start++; + + memset(&xcc_mqd_mem_obj, 0x0, sizeof(struct kfd_mem_obj)); + for (xcc = 0; xcc < NUM_XCC(mm->dev->xcc_mask); xcc++) { + get_xcc_mqd(mqd_mem_obj, &xcc_mqd_mem_obj, offset*xcc); + + init_mqd(mm, (void **)&m, &xcc_mqd_mem_obj, &xcc_gart_addr, q); + + m->cp_mqd_stride_size = offset; + + /* + * Update the CWSR address for each XCC if CWSR is enabled + * and CWSR area is allocated in thunk + */ + if (mm->dev->kfd->cwsr_enabled && + q->ctx_save_restore_area_address) { + xcc_ctx_save_restore_area_address = + q->ctx_save_restore_area_address + + (xcc * q->ctx_save_restore_area_size); + + m->cp_hqd_ctx_save_base_addr_lo = + lower_32_bits(xcc_ctx_save_restore_area_address); + m->cp_hqd_ctx_save_base_addr_hi = + upper_32_bits(xcc_ctx_save_restore_area_address); + } + + if (q->format == KFD_QUEUE_FORMAT_AQL) { + m->compute_tg_chunk_size = 1; + m->compute_current_logic_xcc_id = + (local_xcc_start + xcc) % + NUM_XCC(mm->dev->xcc_mask); + + switch (xcc) { + case 0: + /* Master XCC */ + m->cp_hqd_pq_control &= + ~CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK; + break; + default: + break; + } + } else { + /* PM4 Queue */ + m->compute_current_logic_xcc_id = 0; + m->compute_tg_chunk_size = 0; + m->pm4_target_xcc_in_xcp = q->pm4_target_xcc; + } + + if (xcc == 0) { + /* Set the MQD pointer and gart address to XCC0 MQD */ + *mqd = m; + *gart_addr = xcc_gart_addr; + } + } +} + +static void update_mqd_v9_4_3(struct mqd_manager *mm, void *mqd, + struct queue_properties *q, struct mqd_update_info *minfo) +{ + struct v9_mqd *m; + int xcc = 0; + uint64_t size = mm->mqd_stride(mm, q); + + for (xcc = 0; xcc < NUM_XCC(mm->dev->xcc_mask); xcc++) { + m = get_mqd(mqd + size * xcc); + update_mqd(mm, m, q, minfo); + + if (q->format == KFD_QUEUE_FORMAT_AQL) { + switch (xcc) { + case 0: + /* Master XCC */ + m->cp_hqd_pq_control &= + ~CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK; + break; + default: + break; + } + m->compute_tg_chunk_size = 1; + } else { + /* PM4 Queue */ + m->compute_current_logic_xcc_id = 0; + m->compute_tg_chunk_size = 0; + m->pm4_target_xcc_in_xcp = q->pm4_target_xcc; + } + } +} + +static int destroy_mqd_v9_4_3(struct mqd_manager *mm, void *mqd, + enum kfd_preempt_type type, unsigned int timeout, + uint32_t pipe_id, uint32_t queue_id) +{ + uint32_t xcc_mask = mm->dev->xcc_mask; + int xcc_id, err, inst = 0; + void *xcc_mqd; + struct v9_mqd *m; + uint64_t mqd_offset; + + m = get_mqd(mqd); + mqd_offset = m->cp_mqd_stride_size; + + for_each_inst(xcc_id, xcc_mask) { + xcc_mqd = mqd + mqd_offset * inst; + err = mm->dev->kfd2kgd->hqd_destroy(mm->dev->adev, xcc_mqd, + type, timeout, pipe_id, + queue_id, xcc_id); + if (err) { + pr_debug("Destroy MQD failed for xcc: %d\n", inst); + break; + } + ++inst; + } + + return err; +} + +static int load_mqd_v9_4_3(struct mqd_manager *mm, void *mqd, + uint32_t pipe_id, uint32_t queue_id, + struct queue_properties *p, struct mm_struct *mms) +{ + /* AQL write pointer counts in 64B packets, PM4/CP counts in dwords. */ + uint32_t wptr_shift = (p->format == KFD_QUEUE_FORMAT_AQL ? 4 : 0); + uint32_t xcc_mask = mm->dev->xcc_mask; + int xcc_id, err, inst = 0; + void *xcc_mqd; + uint64_t mqd_stride_size = mm->mqd_stride(mm, p); + + for_each_inst(xcc_id, xcc_mask) { + xcc_mqd = mqd + mqd_stride_size * inst; + err = mm->dev->kfd2kgd->hqd_load( + mm->dev->adev, xcc_mqd, pipe_id, queue_id, + (uint32_t __user *)p->write_ptr, wptr_shift, 0, mms, + xcc_id); + if (err) { + pr_debug("Load MQD failed for xcc: %d\n", inst); + break; + } + ++inst; + } + + return err; +} + +static int get_wave_state_v9_4_3(struct mqd_manager *mm, void *mqd, + struct queue_properties *q, + void __user *ctl_stack, + u32 *ctl_stack_used_size, + u32 *save_area_used_size) +{ + int xcc, err = 0; + void *xcc_mqd; + void __user *xcc_ctl_stack; + uint64_t mqd_stride_size = mm->mqd_stride(mm, q); + u32 tmp_ctl_stack_used_size = 0, tmp_save_area_used_size = 0; + + for (xcc = 0; xcc < NUM_XCC(mm->dev->xcc_mask); xcc++) { + xcc_mqd = mqd + mqd_stride_size * xcc; + xcc_ctl_stack = (void __user *)((uintptr_t)ctl_stack + + q->ctx_save_restore_area_size * xcc); + + err = get_wave_state(mm, xcc_mqd, q, xcc_ctl_stack, + &tmp_ctl_stack_used_size, + &tmp_save_area_used_size); + if (err) + break; + + /* + * Set the ctl_stack_used_size and save_area_used_size to + * ctl_stack_used_size and save_area_used_size of XCC 0 when + * passing the info the user-space. + * For multi XCC, user-space would have to look at the header + * info of each Control stack area to determine the control + * stack size and save area used. + */ + if (xcc == 0) { + *ctl_stack_used_size = tmp_ctl_stack_used_size; + *save_area_used_size = tmp_save_area_used_size; + } + } + + return err; +} + #if defined(CONFIG_DEBUG_FS) static int debugfs_show_mqd(struct seq_file *m, void *data) @@ -486,7 +788,7 @@ static int debugfs_show_mqd_sdma(struct seq_file *m, void *data) #endif struct mqd_manager *mqd_manager_init_v9(enum KFD_MQD_TYPE type, - struct kfd_dev *dev) + struct kfd_node *dev) { struct mqd_manager *mqd; @@ -502,34 +804,50 @@ struct mqd_manager *mqd_manager_init_v9(enum KFD_MQD_TYPE type, switch (type) { case KFD_MQD_TYPE_CP: mqd->allocate_mqd = allocate_mqd; - mqd->init_mqd = init_mqd; mqd->free_mqd = kfd_free_mqd_cp; - mqd->load_mqd = load_mqd; - mqd->update_mqd = update_mqd; - mqd->destroy_mqd = kfd_destroy_mqd_cp; mqd->is_occupied = kfd_is_occupied_cp; - mqd->get_wave_state = get_wave_state; mqd->get_checkpoint_info = get_checkpoint_info; mqd->checkpoint_mqd = checkpoint_mqd; mqd->restore_mqd = restore_mqd; mqd->mqd_size = sizeof(struct v9_mqd); + mqd->mqd_stride = mqd_stride_v9; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd; #endif + if (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 3)) { + mqd->init_mqd = init_mqd_v9_4_3; + mqd->load_mqd = load_mqd_v9_4_3; + mqd->update_mqd = update_mqd_v9_4_3; + mqd->destroy_mqd = destroy_mqd_v9_4_3; + mqd->get_wave_state = get_wave_state_v9_4_3; + } else { + mqd->init_mqd = init_mqd; + mqd->load_mqd = load_mqd; + mqd->update_mqd = update_mqd; + mqd->destroy_mqd = kfd_destroy_mqd_cp; + mqd->get_wave_state = get_wave_state; + } break; case KFD_MQD_TYPE_HIQ: mqd->allocate_mqd = allocate_hiq_mqd; - mqd->init_mqd = init_mqd_hiq; mqd->free_mqd = free_mqd_hiq_sdma; - mqd->load_mqd = kfd_hiq_load_mqd_kiq; mqd->update_mqd = update_mqd; - mqd->destroy_mqd = kfd_destroy_mqd_cp; mqd->is_occupied = kfd_is_occupied_cp; mqd->mqd_size = sizeof(struct v9_mqd); + mqd->mqd_stride = kfd_mqd_stride; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd; #endif mqd->read_doorbell_id = read_doorbell_id; + if (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 3)) { + mqd->init_mqd = init_mqd_hiq_v9_4_3; + mqd->load_mqd = hiq_load_mqd_kiq_v9_4_3; + mqd->destroy_mqd = destroy_hiq_mqd_v9_4_3; + } else { + mqd->init_mqd = init_mqd_hiq; + mqd->load_mqd = kfd_hiq_load_mqd_kiq; + mqd->destroy_mqd = kfd_destroy_mqd_cp; + } break; case KFD_MQD_TYPE_DIQ: mqd->allocate_mqd = allocate_mqd; @@ -555,6 +873,7 @@ struct mqd_manager *mqd_manager_init_v9(enum KFD_MQD_TYPE type, mqd->checkpoint_mqd = checkpoint_mqd_sdma; mqd->restore_mqd = restore_mqd_sdma; mqd->mqd_size = sizeof(struct v9_sdma_mqd); + mqd->mqd_stride = kfd_mqd_stride; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd_sdma; #endif diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c index 530ba6f5b57e..d1e962da51dd 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c @@ -51,8 +51,7 @@ static void update_cu_mask(struct mqd_manager *mm, void *mqd, struct vi_mqd *m; uint32_t se_mask[4] = {0}; /* 4 is the max # of SEs */ - if (!minfo || (minfo->update_flag != UPDATE_FLAG_CU_MASK) || - !minfo->cu_mask.ptr) + if (!minfo || !minfo->cu_mask.ptr) return; mqd_symmetrically_map_cu_mask(mm, @@ -77,7 +76,7 @@ static void set_priority(struct vi_mqd *m, struct queue_properties *q) m->cp_hqd_queue_priority = q->priority; } -static struct kfd_mem_obj *allocate_mqd(struct kfd_dev *kfd, +static struct kfd_mem_obj *allocate_mqd(struct kfd_node *kfd, struct queue_properties *q) { struct kfd_mem_obj *mqd_mem_obj; @@ -136,7 +135,7 @@ static void init_mqd(struct mqd_manager *mm, void **mqd, (1 << COMPUTE_PGM_RSRC2__TRAP_PRESENT__SHIFT); } - if (mm->dev->cwsr_enabled && q->ctx_save_restore_area_address) { + if (mm->dev->kfd->cwsr_enabled && q->ctx_save_restore_area_address) { m->cp_hqd_persistent_state |= (1 << CP_HQD_PERSISTENT_STATE__QSWITCH_MODE__SHIFT); m->cp_hqd_ctx_save_base_addr_lo = @@ -165,7 +164,7 @@ static int load_mqd(struct mqd_manager *mm, void *mqd, return mm->dev->kfd2kgd->hqd_load(mm->dev->adev, mqd, pipe_id, queue_id, (uint32_t __user *)p->write_ptr, - wptr_shift, wptr_mask, mms); + wptr_shift, wptr_mask, mms, 0); } static void __update_mqd(struct mqd_manager *mm, void *mqd, @@ -227,7 +226,7 @@ static void __update_mqd(struct mqd_manager *mm, void *mqd, 2 << CP_HQD_PQ_CONTROL__SLOT_BASED_WPTR__SHIFT; } - if (mm->dev->cwsr_enabled && q->ctx_save_restore_area_address) + if (mm->dev->kfd->cwsr_enabled && q->ctx_save_restore_area_address) m->cp_hqd_ctx_save_control = atc_bit << CP_HQD_CTX_SAVE_CONTROL__ATC__SHIFT | mtype << CP_HQD_CTX_SAVE_CONTROL__MTYPE__SHIFT; @@ -261,6 +260,7 @@ static void update_mqd_tonga(struct mqd_manager *mm, void *mqd, } static int get_wave_state(struct mqd_manager *mm, void *mqd, + struct queue_properties *q, void __user *ctl_stack, u32 *ctl_stack_used_size, u32 *save_area_used_size) @@ -446,7 +446,7 @@ static int debugfs_show_mqd_sdma(struct seq_file *m, void *data) #endif struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type, - struct kfd_dev *dev) + struct kfd_node *dev) { struct mqd_manager *mqd; @@ -486,6 +486,7 @@ struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type, mqd->destroy_mqd = kfd_destroy_mqd_cp; mqd->is_occupied = kfd_is_occupied_cp; mqd->mqd_size = sizeof(struct vi_mqd); + mqd->mqd_stride = kfd_mqd_stride; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd; #endif @@ -500,6 +501,7 @@ struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type, mqd->destroy_mqd = kfd_destroy_mqd_cp; mqd->is_occupied = kfd_is_occupied_cp; mqd->mqd_size = sizeof(struct vi_mqd); + mqd->mqd_stride = kfd_mqd_stride; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd; #endif @@ -515,6 +517,7 @@ struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type, mqd->checkpoint_mqd = checkpoint_mqd_sdma; mqd->restore_mqd = restore_mqd_sdma; mqd->mqd_size = sizeof(struct vi_sdma_mqd); + mqd->mqd_stride = kfd_mqd_stride; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd_sdma; #endif @@ -528,7 +531,7 @@ struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type, } struct mqd_manager *mqd_manager_init_vi_tonga(enum KFD_MQD_TYPE type, - struct kfd_dev *dev) + struct kfd_node *dev) { struct mqd_manager *mqd; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c index f612325241aa..401096c103b2 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c @@ -45,7 +45,7 @@ static void pm_calc_rlib_size(struct packet_manager *pm, unsigned int process_count, queue_count, compute_queue_count, gws_queue_count; unsigned int map_queue_size; unsigned int max_proc_per_quantum = 1; - struct kfd_dev *dev = pm->dqm->dev; + struct kfd_node *dev = pm->dqm->dev; process_count = pm->dqm->processes_count; queue_count = pm->dqm->active_queue_count; @@ -370,6 +370,38 @@ out: return retval; } +int pm_update_grace_period(struct packet_manager *pm, uint32_t grace_period) +{ + int retval = 0; + uint32_t *buffer, size; + + size = pm->pmf->set_grace_period_size; + + mutex_lock(&pm->lock); + + if (size) { + kq_acquire_packet_buffer(pm->priv_queue, + size / sizeof(uint32_t), + (unsigned int **)&buffer); + + if (!buffer) { + pr_err("Failed to allocate buffer on kernel queue\n"); + retval = -ENOMEM; + goto out; + } + + retval = pm->pmf->set_grace_period(pm, buffer, grace_period); + if (!retval) + kq_submit_packet(pm->priv_queue); + else + kq_rollback_packet(pm->priv_queue); + } + +out: + mutex_unlock(&pm->lock); + return retval; +} + int pm_send_unmap_queue(struct packet_manager *pm, enum kfd_unmap_queues_filter filter, uint32_t filter_param, bool reset) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c index 18250845a989..29a2d0499b67 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c @@ -34,6 +34,9 @@ static int pm_map_process_v9(struct packet_manager *pm, { struct pm4_mes_map_process *packet; uint64_t vm_page_table_base_addr = qpd->page_table_base; + struct kfd_node *kfd = pm->dqm->dev; + struct kfd_process_device *pdd = + container_of(qpd, struct kfd_process_device, qpd); packet = (struct pm4_mes_map_process *)buffer; memset(buffer, 0, sizeof(struct pm4_mes_map_process)); @@ -49,6 +52,12 @@ static int pm_map_process_v9(struct packet_manager *pm, packet->bitfields14.sdma_enable = 1; packet->bitfields14.num_queues = (qpd->is_debug) ? 0 : qpd->queue_count; + if (kfd->dqm->trap_debug_vmid && pdd->process->debug_trap_enabled && + pdd->process->runtime_info.runtime_state == DEBUG_RUNTIME_STATE_ENABLED) { + packet->bitfields2.debug_vmid = kfd->dqm->trap_debug_vmid; + packet->bitfields2.new_debug = 1; + } + packet->sh_mem_config = qpd->sh_mem_config; packet->sh_mem_bases = qpd->sh_mem_bases; if (qpd->tba_addr) { @@ -79,6 +88,10 @@ static int pm_map_process_aldebaran(struct packet_manager *pm, { struct pm4_mes_map_process_aldebaran *packet; uint64_t vm_page_table_base_addr = qpd->page_table_base; + struct kfd_dev *kfd = pm->dqm->dev->kfd; + struct kfd_process_device *pdd = + container_of(qpd, struct kfd_process_device, qpd); + int i; packet = (struct pm4_mes_map_process_aldebaran *)buffer; memset(buffer, 0, sizeof(struct pm4_mes_map_process_aldebaran)); @@ -93,6 +106,16 @@ static int pm_map_process_aldebaran(struct packet_manager *pm, packet->bitfields14.num_oac = qpd->num_oac; packet->bitfields14.sdma_enable = 1; packet->bitfields14.num_queues = (qpd->is_debug) ? 0 : qpd->queue_count; + packet->spi_gdbg_per_vmid_cntl = pdd->spi_dbg_override | + pdd->spi_dbg_launch_mode; + + if (pdd->process->debug_trap_enabled) { + for (i = 0; i < kfd->device_info.num_of_watch_points; i++) + packet->tcp_watch_cntl[i] = pdd->watch_points[i]; + + packet->bitfields2.single_memops = + !!(pdd->process->dbg_flags & KFD_DBG_TRAP_FLAG_SINGLE_MEM_OP); + } packet->sh_mem_config = qpd->sh_mem_config; packet->sh_mem_bases = qpd->sh_mem_bases; @@ -119,7 +142,7 @@ static int pm_runlist_v9(struct packet_manager *pm, uint32_t *buffer, struct pm4_mes_runlist *packet; int concurrent_proc_cnt = 0; - struct kfd_dev *kfd = pm->dqm->dev; + struct kfd_node *kfd = pm->dqm->dev; /* Determine the number of processes to map together to HW: * it can not exceed the number of VMIDs available to the @@ -220,13 +243,24 @@ static int pm_map_queues_v9(struct packet_manager *pm, uint32_t *buffer, case KFD_QUEUE_TYPE_SDMA: case KFD_QUEUE_TYPE_SDMA_XGMI: use_static = false; /* no static queues under SDMA */ - if (q->properties.sdma_engine_id < 2 && !pm_use_ext_eng(q->device)) + if (q->properties.sdma_engine_id < 2 && + !pm_use_ext_eng(q->device->kfd)) packet->bitfields2.engine_sel = q->properties.sdma_engine_id + engine_sel__mes_map_queues__sdma0_vi; else { - packet->bitfields2.extended_engine_sel = - extended_engine_sel__mes_map_queues__sdma0_to_7_sel; - packet->bitfields2.engine_sel = q->properties.sdma_engine_id; + /* + * For GFX9.4.3, SDMA engine id can be greater than 8. + * For such cases, set extended_engine_sel to 2 and + * ensure engine_sel lies between 0-7. + */ + if (q->properties.sdma_engine_id >= 8) + packet->bitfields2.extended_engine_sel = + extended_engine_sel__mes_map_queues__sdma8_to_15_sel; + else + packet->bitfields2.extended_engine_sel = + extended_engine_sel__mes_map_queues__sdma0_to_7_sel; + + packet->bitfields2.engine_sel = q->properties.sdma_engine_id % 8; } break; default: @@ -251,6 +285,41 @@ static int pm_map_queues_v9(struct packet_manager *pm, uint32_t *buffer, return 0; } +static int pm_set_grace_period_v9(struct packet_manager *pm, + uint32_t *buffer, + uint32_t grace_period) +{ + struct pm4_mec_write_data_mmio *packet; + uint32_t reg_offset = 0; + uint32_t reg_data = 0; + + pm->dqm->dev->kfd2kgd->build_grace_period_packet_info( + pm->dqm->dev->adev, + pm->dqm->wait_times, + grace_period, + ®_offset, + ®_data); + + if (grace_period == USE_DEFAULT_GRACE_PERIOD) + reg_data = pm->dqm->wait_times; + + packet = (struct pm4_mec_write_data_mmio *)buffer; + memset(buffer, 0, sizeof(struct pm4_mec_write_data_mmio)); + + packet->header.u32All = pm_build_pm4_header(IT_WRITE_DATA, + sizeof(struct pm4_mec_write_data_mmio)); + + packet->bitfields2.dst_sel = dst_sel___write_data__mem_mapped_register; + packet->bitfields2.addr_incr = + addr_incr___write_data__do_not_increment_address; + + packet->bitfields3.dst_mmreg_addr = reg_offset; + + packet->data = reg_data; + + return 0; +} + static int pm_unmap_queues_v9(struct packet_manager *pm, uint32_t *buffer, enum kfd_unmap_queues_filter filter, uint32_t filter_param, bool reset) @@ -263,7 +332,8 @@ static int pm_unmap_queues_v9(struct packet_manager *pm, uint32_t *buffer, packet->header.u32All = pm_build_pm4_header(IT_UNMAP_QUEUES, sizeof(struct pm4_mes_unmap_queues)); - packet->bitfields2.extended_engine_sel = pm_use_ext_eng(pm->dqm->dev) ? + packet->bitfields2.extended_engine_sel = + pm_use_ext_eng(pm->dqm->dev->kfd) ? extended_engine_sel__mes_unmap_queues__sdma0_to_7_sel : extended_engine_sel__mes_unmap_queues__legacy_engine_sel; @@ -333,6 +403,7 @@ const struct packet_manager_funcs kfd_v9_pm_funcs = { .set_resources = pm_set_resources_v9, .map_queues = pm_map_queues_v9, .unmap_queues = pm_unmap_queues_v9, + .set_grace_period = pm_set_grace_period_v9, .query_status = pm_query_status_v9, .release_mem = NULL, .map_process_size = sizeof(struct pm4_mes_map_process), @@ -340,6 +411,7 @@ const struct packet_manager_funcs kfd_v9_pm_funcs = { .set_resources_size = sizeof(struct pm4_mes_set_resources), .map_queues_size = sizeof(struct pm4_mes_map_queues), .unmap_queues_size = sizeof(struct pm4_mes_unmap_queues), + .set_grace_period_size = sizeof(struct pm4_mec_write_data_mmio), .query_status_size = sizeof(struct pm4_mes_query_status), .release_mem_size = 0, }; @@ -350,6 +422,7 @@ const struct packet_manager_funcs kfd_aldebaran_pm_funcs = { .set_resources = pm_set_resources_v9, .map_queues = pm_map_queues_v9, .unmap_queues = pm_unmap_queues_v9, + .set_grace_period = pm_set_grace_period_v9, .query_status = pm_query_status_v9, .release_mem = NULL, .map_process_size = sizeof(struct pm4_mes_map_process_aldebaran), @@ -357,6 +430,7 @@ const struct packet_manager_funcs kfd_aldebaran_pm_funcs = { .set_resources_size = sizeof(struct pm4_mes_set_resources), .map_queues_size = sizeof(struct pm4_mes_map_queues), .unmap_queues_size = sizeof(struct pm4_mes_unmap_queues), + .set_grace_period_size = sizeof(struct pm4_mec_write_data_mmio), .query_status_size = sizeof(struct pm4_mes_query_status), .release_mem_size = 0, }; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_vi.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_vi.c index 4f951eaa6ee8..c1199d06d131 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_vi.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_vi.c @@ -77,7 +77,7 @@ static int pm_runlist_vi(struct packet_manager *pm, uint32_t *buffer, { struct pm4_mes_runlist *packet; int concurrent_proc_cnt = 0; - struct kfd_dev *kfd = pm->dqm->dev; + struct kfd_node *kfd = pm->dqm->dev; if (WARN_ON(!ib)) return -EFAULT; @@ -303,6 +303,7 @@ const struct packet_manager_funcs kfd_vi_pm_funcs = { .set_resources = pm_set_resources_vi, .map_queues = pm_map_queues_vi, .unmap_queues = pm_unmap_queues_vi, + .set_grace_period = NULL, .query_status = pm_query_status_vi, .release_mem = pm_release_mem_vi, .map_process_size = sizeof(struct pm4_mes_map_process), @@ -310,6 +311,7 @@ const struct packet_manager_funcs kfd_vi_pm_funcs = { .set_resources_size = sizeof(struct pm4_mes_set_resources), .map_queues_size = sizeof(struct pm4_mes_map_queues), .unmap_queues_size = sizeof(struct pm4_mes_unmap_queues), + .set_grace_period_size = 0, .query_status_size = sizeof(struct pm4_mes_query_status), .release_mem_size = sizeof(struct pm4_mec_release_mem) }; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h index a666710ed403..8b6b2bd5c148 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h @@ -146,7 +146,10 @@ struct pm4_mes_map_process { union { struct { uint32_t pasid:16; - uint32_t reserved1:8; + uint32_t reserved1:2; + uint32_t debug_vmid:4; + uint32_t new_debug:1; + uint32_t reserved2:1; uint32_t diq_enable:1; uint32_t process_quantum:7; } bitfields2; @@ -263,7 +266,8 @@ enum mes_map_queues_engine_sel_enum { enum mes_map_queues_extended_engine_sel_enum { extended_engine_sel__mes_map_queues__legacy_engine_sel = 0, - extended_engine_sel__mes_map_queues__sdma0_to_7_sel = 1 + extended_engine_sel__mes_map_queues__sdma0_to_7_sel = 1, + extended_engine_sel__mes_map_queues__sdma8_to_15_sel = 2 }; struct pm4_mes_map_queues { @@ -583,6 +587,71 @@ struct pm4_mec_release_mem { #endif +#ifndef PM4_MEC_WRITE_DATA_DEFINED +#define PM4_MEC_WRITE_DATA_DEFINED + +enum WRITE_DATA_dst_sel_enum { + dst_sel___write_data__mem_mapped_register = 0, + dst_sel___write_data__tc_l2 = 2, + dst_sel___write_data__gds = 3, + dst_sel___write_data__memory = 5, + dst_sel___write_data__memory_mapped_adc_persistent_state = 6, +}; + +enum WRITE_DATA_addr_incr_enum { + addr_incr___write_data__increment_address = 0, + addr_incr___write_data__do_not_increment_address = 1 +}; + +enum WRITE_DATA_wr_confirm_enum { + wr_confirm___write_data__do_not_wait_for_write_confirmation = 0, + wr_confirm___write_data__wait_for_write_confirmation = 1 +}; + +enum WRITE_DATA_cache_policy_enum { + cache_policy___write_data__lru = 0, + cache_policy___write_data__stream = 1 +}; + + +struct pm4_mec_write_data_mmio { + union { + union PM4_MES_TYPE_3_HEADER header; /*header */ + unsigned int ordinal1; + }; + + union { + struct { + unsigned int reserved1:8; + unsigned int dst_sel:4; + unsigned int reserved2:4; + unsigned int addr_incr:1; + unsigned int reserved3:2; + unsigned int resume_vf:1; + unsigned int wr_confirm:1; + unsigned int reserved4:4; + unsigned int cache_policy:2; + unsigned int reserved5:5; + } bitfields2; + unsigned int ordinal2; + }; + + union { + struct { + unsigned int dst_mmreg_addr:18; + unsigned int reserved6:14; + } bitfields3; + unsigned int ordinal3; + }; + + uint32_t reserved7; + + uint32_t data; + +}; + +#endif + enum { CACHE_FLUSH_AND_INV_TS_EVENT = 0x00000014 }; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h index 94a438956868..7364a5d77c6e 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h @@ -113,6 +113,8 @@ #define KFD_UNMAP_LATENCY_MS (4000) +#define KFD_MAX_SDMA_QUEUES 128 + /* * 512 = 0x200 * The doorbell index distance between SDMA RLC (2*i) and (2*i+1) in the @@ -199,6 +201,8 @@ extern int amdgpu_no_queue_eviction_on_vm_fault; /* Enable eviction debug messages */ extern bool debug_evictions; +extern struct mutex kfd_processes_mutex; + enum cache_policy { cache_policy_coherent, cache_policy_noncoherent @@ -210,11 +214,13 @@ enum cache_policy { ((KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 2)) || \ (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 3))) +struct kfd_node; + struct kfd_event_interrupt_class { - bool (*interrupt_isr)(struct kfd_dev *dev, + bool (*interrupt_isr)(struct kfd_node *dev, const uint32_t *ih_ring_entry, uint32_t *patched_ihre, bool *patched_flag); - void (*interrupt_wq)(struct kfd_dev *dev, + void (*interrupt_wq)(struct kfd_node *dev, const uint32_t *ih_ring_entry); }; @@ -233,11 +239,11 @@ struct kfd_device_info { uint32_t no_atomic_fw_version; unsigned int num_sdma_queues_per_engine; unsigned int num_reserved_sdma_queues_per_engine; - uint64_t reserved_sdma_queues_bitmap; + DECLARE_BITMAP(reserved_sdma_queues_bitmap, KFD_MAX_SDMA_QUEUES); }; -unsigned int kfd_get_num_sdma_engines(struct kfd_dev *kdev); -unsigned int kfd_get_num_xgmi_sdma_engines(struct kfd_dev *kdev); +unsigned int kfd_get_num_sdma_engines(struct kfd_node *kdev); +unsigned int kfd_get_num_xgmi_sdma_engines(struct kfd_node *kdev); struct kfd_mem_obj { uint32_t range_start; @@ -253,13 +259,70 @@ struct kfd_vmid_info { uint32_t vmid_num_kfd; }; +#define MAX_KFD_NODES 8 + +struct kfd_dev; + +struct kfd_node { + unsigned int node_id; + struct amdgpu_device *adev; /* Duplicated here along with keeping + * a copy in kfd_dev to save a hop + */ + const struct kfd2kgd_calls *kfd2kgd; /* Duplicated here along with + * keeping a copy in kfd_dev to + * save a hop + */ + struct kfd_vmid_info vm_info; + unsigned int id; /* topology stub index */ + uint32_t xcc_mask; /* Instance mask of XCCs present */ + struct amdgpu_xcp *xcp; + + /* Interrupts */ + struct kfifo ih_fifo; + struct workqueue_struct *ih_wq; + struct work_struct interrupt_work; + spinlock_t interrupt_lock; + + /* + * Interrupts of interest to KFD are copied + * from the HW ring into a SW ring. + */ + bool interrupts_active; + uint32_t interrupt_bitmap; /* Only used for GFX 9.4.3 */ + + /* QCM Device instance */ + struct device_queue_manager *dqm; + + /* Global GWS resource shared between processes */ + void *gws; + bool gws_debug_workaround; + + /* Clients watching SMI events */ + struct list_head smi_clients; + spinlock_t smi_lock; + uint32_t reset_seq_num; + + /* SRAM ECC flag */ + atomic_t sram_ecc_flag; + + /*spm process id */ + unsigned int spm_pasid; + + /* Maximum process number mapped to HW scheduler */ + unsigned int max_proc_per_quantum; + + unsigned int compute_vmid_bitmap; + + struct kfd_local_mem_info local_mem_info; + + struct kfd_dev *kfd; +}; + struct kfd_dev { struct amdgpu_device *adev; struct kfd_device_info device_info; - unsigned int id; /* topology stub index */ - phys_addr_t doorbell_base; /* Start of actual doorbells used by * KFD. It is aligned for mapping * into user mode @@ -274,8 +337,6 @@ struct kfd_dev { */ struct kgd2kfd_shared_resources shared_resources; - struct kfd_vmid_info vm_info; - struct kfd_local_mem_info local_mem_info; const struct kfd2kgd_calls *kfd2kgd; struct mutex doorbell_mutex; @@ -290,30 +351,13 @@ struct kfd_dev { unsigned int gtt_sa_chunk_size; unsigned int gtt_sa_num_of_chunks; - /* Interrupts */ - struct kfifo ih_fifo; - struct workqueue_struct *ih_wq; - struct work_struct interrupt_work; - spinlock_t interrupt_lock; - - /* QCM Device instance */ - struct device_queue_manager *dqm; - bool init_complete; - /* - * Interrupts of interest to KFD are copied - * from the HW ring into a SW ring. - */ - bool interrupts_active; /* Firmware versions */ uint16_t mec_fw_version; uint16_t mec2_fw_version; uint16_t sdma_fw_version; - /* Maximum process number mapped to HW scheduler */ - unsigned int max_proc_per_quantum; - /* CWSR */ bool cwsr_enabled; const void *cwsr_isa; @@ -327,28 +371,20 @@ struct kfd_dev { /* Use IOMMU v2 flag */ bool use_iommu_v2; - /* SRAM ECC flag */ - atomic_t sram_ecc_flag; - /* Compute Profile ref. count */ atomic_t compute_profile; - /* Global GWS resource shared between processes */ - void *gws; - - /* Clients watching SMI events */ - struct list_head smi_clients; - spinlock_t smi_lock; - - uint32_t reset_seq_num; - struct ida doorbell_ida; unsigned int max_doorbell_slices; int noretry; - /* HMM page migration MEMORY_DEVICE_PRIVATE mapping */ - struct dev_pagemap pgmap; + struct kfd_node *nodes[MAX_KFD_NODES]; + unsigned int num_nodes; + + /* Track per device allocated watch points */ + uint32_t alloc_watch_ids; + spinlock_t watch_points_lock; }; enum kfd_mempool { @@ -478,8 +514,13 @@ struct queue_properties { uint32_t doorbell_off; bool is_interop; bool is_evicted; + bool is_suspended; + bool is_being_destroyed; bool is_active; bool is_gws; + uint32_t pm4_target_xcc; + bool is_dbg_wa; + bool is_user_cu_masked; /* Not relevant for user mode queues in cp scheduling */ unsigned int vmid; /* Relevant only for sdma queues*/ @@ -494,15 +535,18 @@ struct queue_properties { uint32_t ctl_stack_size; uint64_t tba_addr; uint64_t tma_addr; + uint64_t exception_status; }; #define QUEUE_IS_ACTIVE(q) ((q).queue_size > 0 && \ (q).queue_address != 0 && \ (q).queue_percent > 0 && \ - !(q).is_evicted) + !(q).is_evicted && \ + !(q).is_suspended) enum mqd_update_flag { - UPDATE_FLAG_CU_MASK = 0, + UPDATE_FLAG_DBG_WA_ENABLE = 1, + UPDATE_FLAG_DBG_WA_DISABLE = 2, }; struct mqd_update_info { @@ -563,7 +607,7 @@ struct queue { unsigned int doorbell_id; struct kfd_process *process; - struct kfd_dev *device; + struct kfd_node *device; void *gws; /* procfs */ @@ -697,7 +741,7 @@ enum kfd_pdd_bound { /* Data that is per-process-per device. */ struct kfd_process_device { /* The device that owns this data. */ - struct kfd_dev *dev; + struct kfd_node *dev; /* The process that owns this kfd_process_device. */ struct kfd_process *process; @@ -783,6 +827,18 @@ struct kfd_process_device { uint64_t faults; uint64_t page_in; uint64_t page_out; + + /* Exception code status*/ + uint64_t exception_status; + void *vm_fault_exc_data; + size_t vm_fault_exc_data_size; + + /* Tracks debug per-vmid request settings */ + uint32_t spi_dbg_override; + uint32_t spi_dbg_launch_mode; + uint32_t watch_points[4]; + uint32_t alloc_watch_ids; + /* * If this process has been checkpointed before, then the user * application will use the original gpu_id on the @@ -887,19 +943,57 @@ struct kfd_process { */ unsigned long last_restore_timestamp; + /* Indicates device process is debug attached with reserved vmid. */ + bool debug_trap_enabled; + + /* per-process-per device debug event fd file */ + struct file *dbg_ev_file; + + /* If the process is a kfd debugger, we need to know so we can clean + * up at exit time. If a process enables debugging on itself, it does + * its own clean-up, so we don't set the flag here. We track this by + * counting the number of processes this process is debugging. + */ + atomic_t debugged_process_count; + + /* If the process is a debugged, this is the debugger process */ + struct kfd_process *debugger_process; + /* Kobj for our procfs */ struct kobject *kobj; struct kobject *kobj_queues; struct attribute attr_pasid; + /* Keep track cwsr init */ + bool has_cwsr; + + /* Exception code enable mask and status */ + uint64_t exception_enable_mask; + uint64_t exception_status; + + /* Used to drain stale interrupts */ + wait_queue_head_t wait_irq_drain; + bool irq_drain_is_open; + /* shared virtual memory registered by this process */ struct svm_range_list svms; bool xnack_enabled; + /* Work area for debugger event writer worker. */ + struct work_struct debug_event_workarea; + + /* Tracks debug per-vmid request for debug flags */ + bool dbg_flags; + atomic_t poison; /* Queues are in paused stated because we are in the process of doing a CRIU checkpoint */ bool queues_paused; + + /* Tracks runtime enable status */ + struct semaphore runtime_enable_sema; + bool is_runtime_retry; + struct kfd_runtime_info runtime_info; }; #define KFD_PROCESS_TABLE_SIZE 5 /* bits: 32 entries */ @@ -925,20 +1019,19 @@ struct amdkfd_ioctl_desc { unsigned int cmd_drv; const char *name; }; -bool kfd_dev_is_large_bar(struct kfd_dev *dev); +bool kfd_dev_is_large_bar(struct kfd_node *dev); int kfd_process_create_wq(void); void kfd_process_destroy_wq(void); void kfd_cleanup_processes(void); -struct kfd_process *kfd_create_process(struct file *filep); +struct kfd_process *kfd_create_process(struct task_struct *thread); struct kfd_process *kfd_get_process(const struct task_struct *task); struct kfd_process *kfd_lookup_process_by_pasid(u32 pasid); struct kfd_process *kfd_lookup_process_by_mm(const struct mm_struct *mm); int kfd_process_gpuidx_from_gpuid(struct kfd_process *p, uint32_t gpu_id); -int kfd_process_gpuid_from_adev(struct kfd_process *p, - struct amdgpu_device *adev, uint32_t *gpuid, - uint32_t *gpuidx); +int kfd_process_gpuid_from_node(struct kfd_process *p, struct kfd_node *node, + uint32_t *gpuid, uint32_t *gpuidx); static inline int kfd_process_gpuid_from_gpuidx(struct kfd_process *p, uint32_t gpuidx, uint32_t *gpuid) { return gpuidx < p->n_pdds ? p->pdds[gpuidx]->dev->id : -EINVAL; @@ -961,16 +1054,16 @@ int kfd_process_get_user_gpu_id(struct kfd_process *p, uint32_t actual_gpu_id); int kfd_process_device_init_vm(struct kfd_process_device *pdd, struct file *drm_file); -struct kfd_process_device *kfd_bind_process_to_device(struct kfd_dev *dev, +struct kfd_process_device *kfd_bind_process_to_device(struct kfd_node *dev, struct kfd_process *p); -struct kfd_process_device *kfd_get_process_device_data(struct kfd_dev *dev, +struct kfd_process_device *kfd_get_process_device_data(struct kfd_node *dev, struct kfd_process *p); -struct kfd_process_device *kfd_create_process_device_data(struct kfd_dev *dev, +struct kfd_process_device *kfd_create_process_device_data(struct kfd_node *dev, struct kfd_process *p); bool kfd_process_xnack_mode(struct kfd_process *p, bool supported); -int kfd_reserved_mem_mmap(struct kfd_dev *dev, struct kfd_process *process, +int kfd_reserved_mem_mmap(struct kfd_node *dev, struct kfd_process *process, struct vm_area_struct *vma); /* KFD process API for creating and translating handles */ @@ -994,7 +1087,7 @@ void kfd_pasid_free(u32 pasid); size_t kfd_doorbell_process_slice(struct kfd_dev *kfd); int kfd_doorbell_init(struct kfd_dev *kfd); void kfd_doorbell_fini(struct kfd_dev *kfd); -int kfd_doorbell_mmap(struct kfd_dev *dev, struct kfd_process *process, +int kfd_doorbell_mmap(struct kfd_node *dev, struct kfd_process *process, struct vm_area_struct *vma); void __iomem *kfd_get_kernel_doorbell(struct kfd_dev *kfd, unsigned int *doorbell_off); @@ -1012,10 +1105,10 @@ void kfd_free_process_doorbells(struct kfd_dev *kfd, unsigned int doorbell_index); /* GTT Sub-Allocator */ -int kfd_gtt_sa_allocate(struct kfd_dev *kfd, unsigned int size, +int kfd_gtt_sa_allocate(struct kfd_node *node, unsigned int size, struct kfd_mem_obj **mem_obj); -int kfd_gtt_sa_free(struct kfd_dev *kfd, struct kfd_mem_obj *mem_obj); +int kfd_gtt_sa_free(struct kfd_node *node, struct kfd_mem_obj *mem_obj); extern struct device *kfd_device; @@ -1028,27 +1121,53 @@ void kfd_procfs_del_queue(struct queue *q); /* Topology */ int kfd_topology_init(void); void kfd_topology_shutdown(void); -int kfd_topology_add_device(struct kfd_dev *gpu); -int kfd_topology_remove_device(struct kfd_dev *gpu); +int kfd_topology_add_device(struct kfd_node *gpu); +int kfd_topology_remove_device(struct kfd_node *gpu); struct kfd_topology_device *kfd_topology_device_by_proximity_domain( uint32_t proximity_domain); struct kfd_topology_device *kfd_topology_device_by_proximity_domain_no_lock( uint32_t proximity_domain); struct kfd_topology_device *kfd_topology_device_by_id(uint32_t gpu_id); -struct kfd_dev *kfd_device_by_id(uint32_t gpu_id); -struct kfd_dev *kfd_device_by_pci_dev(const struct pci_dev *pdev); -struct kfd_dev *kfd_device_by_adev(const struct amdgpu_device *adev); -int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_dev **kdev); +struct kfd_node *kfd_device_by_id(uint32_t gpu_id); +struct kfd_node *kfd_device_by_pci_dev(const struct pci_dev *pdev); +static inline bool kfd_irq_is_from_node(struct kfd_node *node, uint32_t node_id, + uint32_t vmid) +{ + return (node->interrupt_bitmap & (1 << node_id)) != 0 && + (node->compute_vmid_bitmap & (1 << vmid)) != 0; +} +static inline struct kfd_node *kfd_node_by_irq_ids(struct amdgpu_device *adev, + uint32_t node_id, uint32_t vmid) { + struct kfd_dev *dev = adev->kfd.dev; + uint32_t i; + + if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 3)) + return dev->nodes[0]; + + for (i = 0; i < dev->num_nodes; i++) + if (kfd_irq_is_from_node(dev->nodes[i], node_id, vmid)) + return dev->nodes[i]; + + return NULL; +} +int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_node **kdev); int kfd_numa_node_to_apic_id(int numa_node_id); void kfd_double_confirm_iommu_support(struct kfd_dev *gpu); /* Interrupts */ -int kfd_interrupt_init(struct kfd_dev *dev); -void kfd_interrupt_exit(struct kfd_dev *dev); -bool enqueue_ih_ring_entry(struct kfd_dev *kfd, const void *ih_ring_entry); -bool interrupt_is_wanted(struct kfd_dev *dev, +#define KFD_IRQ_FENCE_CLIENTID 0xff +#define KFD_IRQ_FENCE_SOURCEID 0xff +#define KFD_IRQ_IS_FENCE(client, source) \ + ((client) == KFD_IRQ_FENCE_CLIENTID && \ + (source) == KFD_IRQ_FENCE_SOURCEID) +int kfd_interrupt_init(struct kfd_node *dev); +void kfd_interrupt_exit(struct kfd_node *dev); +bool enqueue_ih_ring_entry(struct kfd_node *kfd, const void *ih_ring_entry); +bool interrupt_is_wanted(struct kfd_node *dev, const uint32_t *ih_ring_entry, uint32_t *patched_ihre, bool *flag); +int kfd_process_drain_interrupts(struct kfd_process_device *pdd); +void kfd_process_close_interrupt_drain(unsigned int pasid); /* amdkfd Apertures */ int kfd_init_apertures(struct kfd_process *process); @@ -1056,6 +1175,11 @@ int kfd_init_apertures(struct kfd_process *process); void kfd_process_set_trap_handler(struct qcm_process_device *qpd, uint64_t tba_addr, uint64_t tma_addr); +void kfd_process_set_trap_debug_flag(struct qcm_process_device *qpd, + bool enabled); + +/* CWSR initialization */ +int kfd_process_init_cwsr_apu(struct kfd_process *process, struct file *filep); /* CRIU */ /* @@ -1174,22 +1298,22 @@ void print_queue_properties(struct queue_properties *q); void print_queue(struct queue *q); struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type, - struct kfd_dev *dev); + struct kfd_node *dev); struct mqd_manager *mqd_manager_init_cik_hawaii(enum KFD_MQD_TYPE type, - struct kfd_dev *dev); + struct kfd_node *dev); struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type, - struct kfd_dev *dev); + struct kfd_node *dev); struct mqd_manager *mqd_manager_init_vi_tonga(enum KFD_MQD_TYPE type, - struct kfd_dev *dev); + struct kfd_node *dev); struct mqd_manager *mqd_manager_init_v9(enum KFD_MQD_TYPE type, - struct kfd_dev *dev); + struct kfd_node *dev); struct mqd_manager *mqd_manager_init_v10(enum KFD_MQD_TYPE type, - struct kfd_dev *dev); + struct kfd_node *dev); struct mqd_manager *mqd_manager_init_v11(enum KFD_MQD_TYPE type, - struct kfd_dev *dev); -struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev); + struct kfd_node *dev); +struct device_queue_manager *device_queue_manager_init(struct kfd_node *dev); void device_queue_manager_uninit(struct device_queue_manager *dqm); -struct kernel_queue *kernel_queue_init(struct kfd_dev *dev, +struct kernel_queue *kernel_queue_init(struct kfd_node *dev, enum kfd_queue_type type); void kernel_queue_uninit(struct kernel_queue *kq, bool hanging); int kfd_dqm_evict_pasid(struct device_queue_manager *dqm, u32 pasid); @@ -1206,7 +1330,7 @@ void kfd_process_dequeue_from_all_devices(struct kfd_process *p); int pqm_init(struct process_queue_manager *pqm, struct kfd_process *p); void pqm_uninit(struct process_queue_manager *pqm); int pqm_create_queue(struct process_queue_manager *pqm, - struct kfd_dev *dev, + struct kfd_node *dev, struct file *f, struct queue_properties *properties, unsigned int *qid, @@ -1231,6 +1355,11 @@ int pqm_get_wave_state(struct process_queue_manager *pqm, void __user *ctl_stack, u32 *ctl_stack_used_size, u32 *save_area_used_size); +int pqm_get_queue_snapshot(struct process_queue_manager *pqm, + uint64_t exception_clear_mask, + void __user *buf, + int *num_qss_entries, + uint32_t *entry_size); int amdkfd_fence_wait_timeout(uint64_t *fence_addr, uint64_t fence_value, @@ -1270,6 +1399,8 @@ struct packet_manager_funcs { int (*unmap_queues)(struct packet_manager *pm, uint32_t *buffer, enum kfd_unmap_queues_filter mode, uint32_t filter_param, bool reset); + int (*set_grace_period)(struct packet_manager *pm, uint32_t *buffer, + uint32_t grace_period); int (*query_status)(struct packet_manager *pm, uint32_t *buffer, uint64_t fence_address, uint64_t fence_value); int (*release_mem)(uint64_t gpu_addr, uint32_t *buffer); @@ -1280,6 +1411,7 @@ struct packet_manager_funcs { int set_resources_size; int map_queues_size; int unmap_queues_size; + int set_grace_period_size; int query_status_size; int release_mem_size; }; @@ -1302,6 +1434,8 @@ int pm_send_unmap_queue(struct packet_manager *pm, void pm_release_ib(struct packet_manager *pm); +int pm_update_grace_period(struct packet_manager *pm, uint32_t grace_period); + /* Following PM funcs can be shared among VI and AI */ unsigned int pm_build_pm4_header(unsigned int opcode, size_t packet_size); @@ -1310,6 +1444,7 @@ uint64_t kfd_get_number_elems(struct kfd_dev *kfd); /* Events */ extern const struct kfd_event_interrupt_class event_interrupt_class_cik; extern const struct kfd_event_interrupt_class event_interrupt_class_v9; +extern const struct kfd_event_interrupt_class event_interrupt_class_v10; extern const struct kfd_event_interrupt_class event_interrupt_class_v11; extern const struct kfd_device_global_init_class device_global_init_class_cik; @@ -1323,7 +1458,7 @@ int kfd_wait_on_events(struct kfd_process *p, uint32_t *wait_result); void kfd_signal_event_interrupt(u32 pasid, uint32_t partial_id, uint32_t valid_id_bits); -void kfd_signal_iommu_event(struct kfd_dev *dev, +void kfd_signal_iommu_event(struct kfd_node *dev, u32 pasid, unsigned long address, bool is_write_requested, bool is_execute_requested); void kfd_signal_hw_exception_event(u32 pasid); @@ -1339,32 +1474,36 @@ int kfd_event_create(struct file *devkfd, struct kfd_process *p, int kfd_get_num_events(struct kfd_process *p); int kfd_event_destroy(struct kfd_process *p, uint32_t event_id); -void kfd_signal_vm_fault_event(struct kfd_dev *dev, u32 pasid, - struct kfd_vm_fault_info *info); +void kfd_signal_vm_fault_event(struct kfd_node *dev, u32 pasid, + struct kfd_vm_fault_info *info, + struct kfd_hsa_memory_exception_data *data); -void kfd_signal_reset_event(struct kfd_dev *dev); +void kfd_signal_reset_event(struct kfd_node *dev); -void kfd_signal_poison_consumed_event(struct kfd_dev *dev, u32 pasid); +void kfd_signal_poison_consumed_event(struct kfd_node *dev, u32 pasid); void kfd_flush_tlb(struct kfd_process_device *pdd, enum TLB_FLUSH_TYPE type); static inline bool kfd_flush_tlb_after_unmap(struct kfd_dev *dev) { - return KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 2) || - (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 1) && - dev->adev->sdma.instance[0].fw_version >= 18) || + return KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 3) || + KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 2) || + (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 1) && dev->sdma_fw_version >= 18) || KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 0); } +int kfd_send_exception_to_runtime(struct kfd_process *p, + unsigned int queue_id, + uint64_t error_reason); bool kfd_is_locked(void); /* Compute profile */ -void kfd_inc_compute_active(struct kfd_dev *dev); -void kfd_dec_compute_active(struct kfd_dev *dev); +void kfd_inc_compute_active(struct kfd_node *dev); +void kfd_dec_compute_active(struct kfd_node *dev); /* Cgroup Support */ /* Check with device cgroup if @kfd device is accessible */ -static inline int kfd_devcgroup_check_permission(struct kfd_dev *kfd) +static inline int kfd_devcgroup_check_permission(struct kfd_node *kfd) { #if defined(CONFIG_CGROUP_DEVICE) || defined(CONFIG_CGROUP_BPF) struct drm_device *ddev = adev_to_drm(kfd->adev); @@ -1377,6 +1516,11 @@ static inline int kfd_devcgroup_check_permission(struct kfd_dev *kfd) #endif } +static inline bool kfd_is_first_node(struct kfd_node *node) +{ + return (node == node->kfd->nodes[0]); +} + /* Debugfs */ #if defined(CONFIG_DEBUG_FS) @@ -1389,7 +1533,7 @@ int dqm_debugfs_hqds(struct seq_file *m, void *data); int kfd_debugfs_rls_by_device(struct seq_file *m, void *data); int pm_debugfs_runlist(struct seq_file *m, void *data); -int kfd_debugfs_hang_hws(struct kfd_dev *dev); +int kfd_debugfs_hang_hws(struct kfd_node *dev); int pm_debugfs_hang_hws(struct packet_manager *pm); int dqm_debugfs_hang_hws(struct device_queue_manager *dqm); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c index 07a9eaf9b7d8..3d3611705d41 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c @@ -44,13 +44,14 @@ struct mm_struct; #include "kfd_iommu.h" #include "kfd_svm.h" #include "kfd_smi_events.h" +#include "kfd_debug.h" /* * List of struct kfd_process (field kfd_process). * Unique/indexed by mm_struct* */ DEFINE_HASHTABLE(kfd_processes_table, KFD_PROCESS_TABLE_SIZE); -static DEFINE_MUTEX(kfd_processes_mutex); +DEFINE_MUTEX(kfd_processes_mutex); DEFINE_SRCU(kfd_processes_srcu); @@ -69,7 +70,6 @@ static struct kfd_process *find_process(const struct task_struct *thread, bool ref); static void kfd_process_ref_release(struct kref *ref); static struct kfd_process *create_process(const struct task_struct *thread); -static int kfd_process_init_cwsr_apu(struct kfd_process *p, struct file *filep); static void evict_process_worker(struct work_struct *work); static void restore_process_worker(struct work_struct *work); @@ -269,7 +269,7 @@ static int kfd_get_cu_occupancy(struct attribute *attr, char *buffer) int cu_cnt; int wave_cnt; int max_waves_per_cu; - struct kfd_dev *dev = NULL; + struct kfd_node *dev = NULL; struct kfd_process *proc = NULL; struct kfd_process_device *pdd = NULL; @@ -290,7 +290,7 @@ static int kfd_get_cu_occupancy(struct attribute *attr, char *buffer) wave_cnt = 0; max_waves_per_cu = 0; dev->kfd2kgd->get_cu_occupancy(dev->adev, proc->pasid, &wave_cnt, - &max_waves_per_cu); + &max_waves_per_cu, 0); /* Translate wave count to number of compute units */ cu_cnt = (wave_cnt + (max_waves_per_cu - 1)) / max_waves_per_cu; @@ -691,7 +691,7 @@ void kfd_process_destroy_wq(void) static void kfd_process_free_gpuvm(struct kgd_mem *mem, struct kfd_process_device *pdd, void **kptr) { - struct kfd_dev *dev = pdd->dev; + struct kfd_node *dev = pdd->dev; if (kptr && *kptr) { amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(mem); @@ -713,7 +713,7 @@ static int kfd_process_alloc_gpuvm(struct kfd_process_device *pdd, uint64_t gpu_va, uint32_t size, uint32_t flags, struct kgd_mem **mem, void **kptr) { - struct kfd_dev *kdev = pdd->dev; + struct kfd_node *kdev = pdd->dev; int err; err = amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(kdev->adev, gpu_va, size, @@ -798,18 +798,19 @@ static void kfd_process_device_destroy_ib_mem(struct kfd_process_device *pdd) kfd_process_free_gpuvm(qpd->ib_mem, pdd, &qpd->ib_kaddr); } -struct kfd_process *kfd_create_process(struct file *filep) +struct kfd_process *kfd_create_process(struct task_struct *thread) { struct kfd_process *process; - struct task_struct *thread = current; int ret; - if (!thread->mm) + if (!(thread->mm && mmget_not_zero(thread->mm))) return ERR_PTR(-EINVAL); /* Only the pthreads threading model is supported. */ - if (thread->group_leader->mm != thread->mm) + if (thread->group_leader->mm != thread->mm) { + mmput(thread->mm); return ERR_PTR(-EINVAL); + } /* * take kfd processes mutex before starting of process creation @@ -818,6 +819,12 @@ struct kfd_process *kfd_create_process(struct file *filep) */ mutex_lock(&kfd_processes_mutex); + if (kfd_is_locked()) { + mutex_unlock(&kfd_processes_mutex); + pr_debug("KFD is locked! Cannot create process"); + return ERR_PTR(-EINVAL); + } + /* A prior open of /dev/kfd could have already created the process. */ process = find_process(thread, false); if (process) { @@ -827,10 +834,6 @@ struct kfd_process *kfd_create_process(struct file *filep) if (IS_ERR(process)) goto out; - ret = kfd_process_init_cwsr_apu(process, filep); - if (ret) - goto out_destroy; - if (!procfs.kobj) goto out; @@ -859,21 +862,16 @@ struct kfd_process *kfd_create_process(struct file *filep) kfd_procfs_add_sysfs_stats(process); kfd_procfs_add_sysfs_files(process); kfd_procfs_add_sysfs_counters(process); + + init_waitqueue_head(&process->wait_irq_drain); } out: if (!IS_ERR(process)) kref_get(&process->ref); mutex_unlock(&kfd_processes_mutex); + mmput(thread->mm); return process; - -out_destroy: - hash_del_rcu(&process->kfd_processes); - mutex_unlock(&kfd_processes_mutex); - synchronize_srcu(&kfd_processes_srcu); - /* kfd_process_free_notifier will trigger the cleanup */ - mmu_notifier_put(&process->mmu_notifier); - return ERR_PTR(ret); } struct kfd_process *kfd_get_process(const struct task_struct *thread) @@ -982,7 +980,7 @@ static void kfd_process_device_free_bos(struct kfd_process_device *pdd) static void kfd_process_kunmap_signal_bo(struct kfd_process *p) { struct kfd_process_device *pdd; - struct kfd_dev *kdev; + struct kfd_node *kdev; void *mem; kdev = kfd_device_by_id(GET_GPU_ID(p->signal_handle)); @@ -1040,9 +1038,9 @@ static void kfd_process_destroy_pdds(struct kfd_process *p) bitmap_free(pdd->qpd.doorbell_bitmap); idr_destroy(&pdd->alloc_idr); - kfd_free_process_doorbells(pdd->dev, pdd->doorbell_index); + kfd_free_process_doorbells(pdd->dev->kfd, pdd->doorbell_index); - if (pdd->dev->shared_resources.enable_mes) + if (pdd->dev->kfd->shared_resources.enable_mes) amdgpu_amdkfd_free_gtt_mem(pdd->dev->adev, pdd->proc_ctx_bo); /* @@ -1169,11 +1167,40 @@ static void kfd_process_free_notifier(struct mmu_notifier *mn) static void kfd_process_notifier_release_internal(struct kfd_process *p) { + int i; + cancel_delayed_work_sync(&p->eviction_work); cancel_delayed_work_sync(&p->restore_work); + for (i = 0; i < p->n_pdds; i++) { + struct kfd_process_device *pdd = p->pdds[i]; + + /* re-enable GFX OFF since runtime enable with ttmp setup disabled it. */ + if (!kfd_dbg_is_rlc_restore_supported(pdd->dev) && p->runtime_info.ttmp_setup) + amdgpu_gfx_off_ctrl(pdd->dev->adev, true); + } + /* Indicate to other users that MM is no longer valid */ p->mm = NULL; + kfd_dbg_trap_disable(p); + + if (atomic_read(&p->debugged_process_count) > 0) { + struct kfd_process *target; + unsigned int temp; + int idx = srcu_read_lock(&kfd_processes_srcu); + + hash_for_each_rcu(kfd_processes_table, temp, target, kfd_processes) { + if (target->debugger_process && target->debugger_process == p) { + mutex_lock_nested(&target->mutex, 1); + kfd_dbg_trap_disable(target); + mutex_unlock(&target->mutex); + if (atomic_read(&p->debugged_process_count) == 0) + break; + } + } + + srcu_read_unlock(&kfd_processes_srcu, idx); + } mmu_notifier_put(&p->mmu_notifier); } @@ -1253,16 +1280,19 @@ void kfd_cleanup_processes(void) mmu_notifier_synchronize(); } -static int kfd_process_init_cwsr_apu(struct kfd_process *p, struct file *filep) +int kfd_process_init_cwsr_apu(struct kfd_process *p, struct file *filep) { unsigned long offset; int i; + if (p->has_cwsr) + return 0; + for (i = 0; i < p->n_pdds; i++) { - struct kfd_dev *dev = p->pdds[i]->dev; + struct kfd_node *dev = p->pdds[i]->dev; struct qcm_process_device *qpd = &p->pdds[i]->qpd; - if (!dev->cwsr_enabled || qpd->cwsr_kaddr || qpd->cwsr_base) + if (!dev->kfd->cwsr_enabled || qpd->cwsr_kaddr || qpd->cwsr_base) continue; offset = KFD_MMAP_TYPE_RESERVED_MEM | KFD_MMAP_GPU_ID(dev->id); @@ -1279,19 +1309,23 @@ static int kfd_process_init_cwsr_apu(struct kfd_process *p, struct file *filep) return err; } - memcpy(qpd->cwsr_kaddr, dev->cwsr_isa, dev->cwsr_isa_size); + memcpy(qpd->cwsr_kaddr, dev->kfd->cwsr_isa, dev->kfd->cwsr_isa_size); + + kfd_process_set_trap_debug_flag(qpd, p->debug_trap_enabled); qpd->tma_addr = qpd->tba_addr + KFD_CWSR_TMA_OFFSET; pr_debug("set tba :0x%llx, tma:0x%llx, cwsr_kaddr:%p for pqm.\n", qpd->tba_addr, qpd->tma_addr, qpd->cwsr_kaddr); } + p->has_cwsr = true; + return 0; } static int kfd_process_device_init_cwsr_dgpu(struct kfd_process_device *pdd) { - struct kfd_dev *dev = pdd->dev; + struct kfd_node *dev = pdd->dev; struct qcm_process_device *qpd = &pdd->qpd; uint32_t flags = KFD_IOC_ALLOC_MEM_FLAGS_GTT | KFD_IOC_ALLOC_MEM_FLAGS_NO_SUBSTITUTE @@ -1300,7 +1334,7 @@ static int kfd_process_device_init_cwsr_dgpu(struct kfd_process_device *pdd) void *kaddr; int ret; - if (!dev->cwsr_enabled || qpd->cwsr_kaddr || !qpd->cwsr_base) + if (!dev->kfd->cwsr_enabled || qpd->cwsr_kaddr || !qpd->cwsr_base) return 0; /* cwsr_base is only set for dGPU */ @@ -1313,7 +1347,10 @@ static int kfd_process_device_init_cwsr_dgpu(struct kfd_process_device *pdd) qpd->cwsr_kaddr = kaddr; qpd->tba_addr = qpd->cwsr_base; - memcpy(qpd->cwsr_kaddr, dev->cwsr_isa, dev->cwsr_isa_size); + memcpy(qpd->cwsr_kaddr, dev->kfd->cwsr_isa, dev->kfd->cwsr_isa_size); + + kfd_process_set_trap_debug_flag(&pdd->qpd, + pdd->process->debug_trap_enabled); qpd->tma_addr = qpd->tba_addr + KFD_CWSR_TMA_OFFSET; pr_debug("set tba :0x%llx, tma:0x%llx, cwsr_kaddr:%p for pqm.\n", @@ -1324,10 +1361,10 @@ static int kfd_process_device_init_cwsr_dgpu(struct kfd_process_device *pdd) static void kfd_process_device_destroy_cwsr_dgpu(struct kfd_process_device *pdd) { - struct kfd_dev *dev = pdd->dev; + struct kfd_node *dev = pdd->dev; struct qcm_process_device *qpd = &pdd->qpd; - if (!dev->cwsr_enabled || !qpd->cwsr_kaddr || !qpd->cwsr_base) + if (!dev->kfd->cwsr_enabled || !qpd->cwsr_kaddr || !qpd->cwsr_base) return; kfd_process_free_gpuvm(qpd->cwsr_mem, pdd, &qpd->cwsr_kaddr); @@ -1371,7 +1408,7 @@ bool kfd_process_xnack_mode(struct kfd_process *p, bool supported) * support retry. */ for (i = 0; i < p->n_pdds; i++) { - struct kfd_dev *dev = p->pdds[i]->dev; + struct kfd_node *dev = p->pdds[i]->dev; /* Only consider GFXv9 and higher GPUs. Older GPUs don't * support the SVM APIs and don't need to be considered @@ -1394,13 +1431,23 @@ bool kfd_process_xnack_mode(struct kfd_process *p, bool supported) if (KFD_GC_VERSION(dev) >= IP_VERSION(10, 1, 1)) return false; - if (dev->noretry) + if (dev->kfd->noretry) return false; } return true; } +void kfd_process_set_trap_debug_flag(struct qcm_process_device *qpd, + bool enabled) +{ + if (qpd->cwsr_kaddr) { + uint64_t *tma = + (uint64_t *)(qpd->cwsr_kaddr + KFD_CWSR_TMA_OFFSET); + tma[2] = enabled; + } +} + /* * On return the kfd_process is fully operational and will be freed when the * mm is released @@ -1428,6 +1475,11 @@ static struct kfd_process *create_process(const struct task_struct *thread) if (err) goto err_event_init; process->is_32bit_user_mode = in_compat_syscall(); + process->debug_trap_enabled = false; + process->debugger_process = NULL; + process->exception_enable_mask = 0; + atomic_set(&process->debugged_process_count, 0); + sema_init(&process->runtime_enable_sema, 0); process->pasid = kfd_pasid_alloc(); if (process->pasid == 0) { @@ -1475,6 +1527,8 @@ static struct kfd_process *create_process(const struct task_struct *thread) kfd_unref_process(process); get_task_struct(process->lead_thread); + INIT_WORK(&process->debug_event_workarea, debug_event_write_work_handler); + return process; err_register_notifier: @@ -1528,7 +1582,7 @@ static int init_doorbell_bitmap(struct qcm_process_device *qpd, return 0; } -struct kfd_process_device *kfd_get_process_device_data(struct kfd_dev *dev, +struct kfd_process_device *kfd_get_process_device_data(struct kfd_node *dev, struct kfd_process *p) { int i; @@ -1540,7 +1594,7 @@ struct kfd_process_device *kfd_get_process_device_data(struct kfd_dev *dev, return NULL; } -struct kfd_process_device *kfd_create_process_device_data(struct kfd_dev *dev, +struct kfd_process_device *kfd_create_process_device_data(struct kfd_node *dev, struct kfd_process *p) { struct kfd_process_device *pdd = NULL; @@ -1552,7 +1606,7 @@ struct kfd_process_device *kfd_create_process_device_data(struct kfd_dev *dev, if (!pdd) return NULL; - if (init_doorbell_bitmap(&pdd->qpd, dev)) { + if (init_doorbell_bitmap(&pdd->qpd, dev->kfd)) { pr_err("Failed to init doorbell for process\n"); goto err_free_pdd; } @@ -1573,7 +1627,7 @@ struct kfd_process_device *kfd_create_process_device_data(struct kfd_dev *dev, pdd->user_gpu_id = dev->id; atomic64_set(&pdd->evict_duration_counter, 0); - if (dev->shared_resources.enable_mes) { + if (dev->kfd->shared_resources.enable_mes) { retval = amdgpu_amdkfd_alloc_gtt_mem(dev->adev, AMDGPU_MES_PROC_CTX_SIZE, &pdd->proc_ctx_bo, @@ -1588,6 +1642,11 @@ struct kfd_process_device *kfd_create_process_device_data(struct kfd_dev *dev, } p->pdds[p->n_pdds++] = pdd; + if (kfd_dbg_is_per_vmid_supported(pdd->dev)) + pdd->spi_dbg_override = pdd->dev->kfd2kgd->disable_debug_trap( + pdd->dev->adev, + false, + 0); /* Init idr used for memory handle translation */ idr_init(&pdd->alloc_idr); @@ -1619,7 +1678,7 @@ int kfd_process_device_init_vm(struct kfd_process_device *pdd, struct amdgpu_fpriv *drv_priv; struct amdgpu_vm *avm; struct kfd_process *p; - struct kfd_dev *dev; + struct kfd_node *dev; int ret; if (!drm_file) @@ -1679,7 +1738,7 @@ err_reserve_ib_mem: * * Assumes that the process lock is held. */ -struct kfd_process_device *kfd_bind_process_to_device(struct kfd_dev *dev, +struct kfd_process_device *kfd_bind_process_to_device(struct kfd_node *dev, struct kfd_process *p) { struct kfd_process_device *pdd; @@ -1885,13 +1944,13 @@ int kfd_process_gpuidx_from_gpuid(struct kfd_process *p, uint32_t gpu_id) } int -kfd_process_gpuid_from_adev(struct kfd_process *p, struct amdgpu_device *adev, - uint32_t *gpuid, uint32_t *gpuidx) +kfd_process_gpuid_from_node(struct kfd_process *p, struct kfd_node *node, + uint32_t *gpuid, uint32_t *gpuidx) { int i; for (i = 0; i < p->n_pdds; i++) - if (p->pdds[i] && p->pdds[i]->dev->adev == adev) { + if (p->pdds[i] && p->pdds[i]->dev == node) { *gpuid = p->pdds[i]->user_gpu_id; *gpuidx = i; return 0; @@ -1961,8 +2020,10 @@ static void restore_process_worker(struct work_struct *work) */ p->last_restore_timestamp = get_jiffies_64(); - ret = amdgpu_amdkfd_gpuvm_restore_process_bos(p->kgd_process_info, - &p->ef); + /* VMs may not have been acquired yet during debugging. */ + if (p->kgd_process_info) + ret = amdgpu_amdkfd_gpuvm_restore_process_bos(p->kgd_process_info, + &p->ef); if (ret) { pr_debug("Failed to restore BOs of pasid 0x%x, retry after %d ms\n", p->pasid, PROCESS_BACK_OFF_TIME_MS); @@ -1988,7 +2049,7 @@ void kfd_suspend_all_processes(void) WARN(debug_evictions, "Evicting all processes"); hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) { cancel_delayed_work_sync(&p->eviction_work); - cancel_delayed_work_sync(&p->restore_work); + flush_delayed_work(&p->restore_work); if (kfd_process_evict_queues(p, KFD_QUEUE_EVICTION_TRIGGER_SUSPEND)) pr_err("Failed to suspend process 0x%x\n", p->pasid); @@ -2016,7 +2077,7 @@ int kfd_resume_all_processes(void) return ret; } -int kfd_reserved_mem_mmap(struct kfd_dev *dev, struct kfd_process *process, +int kfd_reserved_mem_mmap(struct kfd_node *dev, struct kfd_process *process, struct vm_area_struct *vma) { struct kfd_process_device *pdd; @@ -2051,7 +2112,9 @@ void kfd_flush_tlb(struct kfd_process_device *pdd, enum TLB_FLUSH_TYPE type) { struct amdgpu_vm *vm = drm_priv_to_vm(pdd->drm_priv); uint64_t tlb_seq = amdgpu_vm_tlb_seq(vm); - struct kfd_dev *dev = pdd->dev; + struct kfd_node *dev = pdd->dev; + uint32_t xcc_mask = dev->xcc_mask; + int xcc = 0; /* * It can be that we race and lose here, but that is extremely unlikely @@ -2069,11 +2132,126 @@ void kfd_flush_tlb(struct kfd_process_device *pdd, enum TLB_FLUSH_TYPE type) amdgpu_amdkfd_flush_gpu_tlb_vmid(dev->adev, pdd->qpd.vmid); } else { - amdgpu_amdkfd_flush_gpu_tlb_pasid(dev->adev, - pdd->process->pasid, type); + for_each_inst(xcc, xcc_mask) + amdgpu_amdkfd_flush_gpu_tlb_pasid( + dev->adev, pdd->process->pasid, type, xcc); } } +/* assumes caller holds process lock. */ +int kfd_process_drain_interrupts(struct kfd_process_device *pdd) +{ + uint32_t irq_drain_fence[8]; + int r = 0; + + if (!KFD_IS_SOC15(pdd->dev)) + return 0; + + pdd->process->irq_drain_is_open = true; + + memset(irq_drain_fence, 0, sizeof(irq_drain_fence)); + irq_drain_fence[0] = (KFD_IRQ_FENCE_SOURCEID << 8) | + KFD_IRQ_FENCE_CLIENTID; + irq_drain_fence[3] = pdd->process->pasid; + + /* ensure stale irqs scheduled KFD interrupts and send drain fence. */ + if (amdgpu_amdkfd_send_close_event_drain_irq(pdd->dev->adev, + irq_drain_fence)) { + pdd->process->irq_drain_is_open = false; + return 0; + } + + r = wait_event_interruptible(pdd->process->wait_irq_drain, + !READ_ONCE(pdd->process->irq_drain_is_open)); + if (r) + pdd->process->irq_drain_is_open = false; + + return r; +} + +void kfd_process_close_interrupt_drain(unsigned int pasid) +{ + struct kfd_process *p; + + p = kfd_lookup_process_by_pasid(pasid); + + if (!p) + return; + + WRITE_ONCE(p->irq_drain_is_open, false); + wake_up_all(&p->wait_irq_drain); + kfd_unref_process(p); +} + +struct send_exception_work_handler_workarea { + struct work_struct work; + struct kfd_process *p; + unsigned int queue_id; + uint64_t error_reason; +}; + +static void send_exception_work_handler(struct work_struct *work) +{ + struct send_exception_work_handler_workarea *workarea; + struct kfd_process *p; + struct queue *q; + struct mm_struct *mm; + struct kfd_context_save_area_header __user *csa_header; + uint64_t __user *err_payload_ptr; + uint64_t cur_err; + uint32_t ev_id; + + workarea = container_of(work, + struct send_exception_work_handler_workarea, + work); + p = workarea->p; + + mm = get_task_mm(p->lead_thread); + + if (!mm) + return; + + kthread_use_mm(mm); + + q = pqm_get_user_queue(&p->pqm, workarea->queue_id); + + if (!q) + goto out; + + csa_header = (void __user *)q->properties.ctx_save_restore_area_address; + + get_user(err_payload_ptr, (uint64_t __user **)&csa_header->err_payload_addr); + get_user(cur_err, err_payload_ptr); + cur_err |= workarea->error_reason; + put_user(cur_err, err_payload_ptr); + get_user(ev_id, &csa_header->err_event_id); + + kfd_set_event(p, ev_id); + +out: + kthread_unuse_mm(mm); + mmput(mm); +} + +int kfd_send_exception_to_runtime(struct kfd_process *p, + unsigned int queue_id, + uint64_t error_reason) +{ + struct send_exception_work_handler_workarea worker; + + INIT_WORK_ONSTACK(&worker.work, send_exception_work_handler); + + worker.p = p; + worker.queue_id = queue_id; + worker.error_reason = error_reason; + + schedule_work(&worker.work); + flush_work(&worker.work); + destroy_work_on_stack(&worker.work); + + return 0; +} + struct kfd_process_device *kfd_process_device_data_by_id(struct kfd_process *p, uint32_t gpu_id) { int i; @@ -2133,4 +2311,3 @@ int kfd_debugfs_mqds_by_process(struct seq_file *m, void *data) } #endif - diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c index 4236539d9f93..9ad1a2186a24 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c @@ -81,7 +81,7 @@ static int find_available_queue_slot(struct process_queue_manager *pqm, void kfd_process_dequeue_from_device(struct kfd_process_device *pdd) { - struct kfd_dev *dev = pdd->dev; + struct kfd_node *dev = pdd->dev; if (pdd->already_dequeued) return; @@ -93,7 +93,7 @@ void kfd_process_dequeue_from_device(struct kfd_process_device *pdd) int pqm_set_gws(struct process_queue_manager *pqm, unsigned int qid, void *gws) { - struct kfd_dev *dev = NULL; + struct kfd_node *dev = NULL; struct process_queue_node *pqn; struct kfd_process_device *pdd; struct kgd_mem *mem = NULL; @@ -178,7 +178,7 @@ void pqm_uninit(struct process_queue_manager *pqm) } static int init_user_queue(struct process_queue_manager *pqm, - struct kfd_dev *dev, struct queue **q, + struct kfd_node *dev, struct queue **q, struct queue_properties *q_properties, struct file *f, struct amdgpu_bo *wptr_bo, unsigned int qid) @@ -187,6 +187,7 @@ static int init_user_queue(struct process_queue_manager *pqm, /* Doorbell initialized in user space*/ q_properties->doorbell_ptr = NULL; + q_properties->exception_status = KFD_EC_MASK(EC_QUEUE_NEW); /* let DQM handle it*/ q_properties->vmid = 0; @@ -199,7 +200,7 @@ static int init_user_queue(struct process_queue_manager *pqm, (*q)->device = dev; (*q)->process = pqm->process; - if (dev->shared_resources.enable_mes) { + if (dev->kfd->shared_resources.enable_mes) { retval = amdgpu_amdkfd_alloc_gtt_mem(dev->adev, AMDGPU_MES_GANG_CTX_SIZE, &(*q)->gang_ctx_bo, @@ -224,7 +225,7 @@ cleanup: } int pqm_create_queue(struct process_queue_manager *pqm, - struct kfd_dev *dev, + struct kfd_node *dev, struct file *f, struct queue_properties *properties, unsigned int *qid, @@ -242,6 +243,13 @@ int pqm_create_queue(struct process_queue_manager *pqm, enum kfd_queue_type type = properties->type; unsigned int max_queues = 127; /* HWS limit */ + /* + * On GFX 9.4.3, increase the number of queues that + * can be created to 255. No HWS limit on GFX 9.4.3. + */ + if (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 3)) + max_queues = 255; + q = NULL; kq = NULL; @@ -258,7 +266,7 @@ int pqm_create_queue(struct process_queue_manager *pqm, * Hence we also check the type as well */ if ((pdd->qpd.is_debug) || (type == KFD_QUEUE_TYPE_DIQ)) - max_queues = dev->device_info.max_no_of_hqd/2; + max_queues = dev->kfd->device_info.max_no_of_hqd/2; if (pdd->qpd.queue_count >= max_queues) return -ENOSPC; @@ -330,6 +338,10 @@ int pqm_create_queue(struct process_queue_manager *pqm, kq->queue->properties.queue_id = *qid; pqn->kq = kq; pqn->q = NULL; + retval = kfd_process_drain_interrupts(pdd); + if (retval) + break; + retval = dev->dqm->ops.create_kernel_queue(dev->dqm, kq, &pdd->qpd); break; @@ -354,7 +366,7 @@ int pqm_create_queue(struct process_queue_manager *pqm, */ *p_doorbell_offset_in_process = (q->properties.doorbell_off * sizeof(uint32_t)) & - (kfd_doorbell_process_slice(dev) - 1); + (kfd_doorbell_process_slice(dev->kfd) - 1); pr_debug("PQM After DQM create queue\n"); @@ -387,7 +399,7 @@ int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid) struct process_queue_node *pqn; struct kfd_process_device *pdd; struct device_queue_manager *dqm; - struct kfd_dev *dev; + struct kfd_node *dev; int retval; dqm = NULL; @@ -439,7 +451,7 @@ int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid) pdd->qpd.num_gws = 0; } - if (dev->shared_resources.enable_mes) { + if (dev->kfd->shared_resources.enable_mes) { amdgpu_amdkfd_free_gtt_mem(dev->adev, pqn->q->gang_ctx_bo); if (pqn->q->wptr_bo) @@ -477,6 +489,7 @@ int pqm_update_queue_properties(struct process_queue_manager *pqm, pqn->q->properties.queue_size = p->queue_size; pqn->q->properties.queue_percent = p->queue_percent; pqn->q->properties.priority = p->priority; + pqn->q->properties.pm4_target_xcc = p->pm4_target_xcc; retval = pqn->q->device->dqm->ops.update_queue(pqn->q->device->dqm, pqn->q, NULL); @@ -498,8 +511,12 @@ int pqm_update_mqd(struct process_queue_manager *pqm, return -EFAULT; } + /* CUs are masked for debugger requirements so deny user mask */ + if (pqn->q->properties.is_dbg_wa && minfo && minfo->cu_mask.ptr) + return -EBUSY; + /* ASICs that have WGPs must enforce pairwise enabled mask checks. */ - if (minfo && minfo->update_flag == UPDATE_FLAG_CU_MASK && minfo->cu_mask.ptr && + if (minfo && minfo->cu_mask.ptr && KFD_GC_VERSION(pqn->q->device) >= IP_VERSION(10, 0, 0)) { int i; @@ -518,6 +535,9 @@ int pqm_update_mqd(struct process_queue_manager *pqm, if (retval != 0) return retval; + if (minfo && minfo->cu_mask.ptr) + pqn->q->properties.is_user_cu_masked = true; + return 0; } @@ -565,6 +585,46 @@ int pqm_get_wave_state(struct process_queue_manager *pqm, save_area_used_size); } +int pqm_get_queue_snapshot(struct process_queue_manager *pqm, + uint64_t exception_clear_mask, + void __user *buf, + int *num_qss_entries, + uint32_t *entry_size) +{ + struct process_queue_node *pqn; + struct kfd_queue_snapshot_entry src; + uint32_t tmp_entry_size = *entry_size, tmp_qss_entries = *num_qss_entries; + int r = 0; + + *num_qss_entries = 0; + if (!(*entry_size)) + return -EINVAL; + + *entry_size = min_t(size_t, *entry_size, sizeof(struct kfd_queue_snapshot_entry)); + mutex_lock(&pqm->process->event_mutex); + + memset(&src, 0, sizeof(src)); + + list_for_each_entry(pqn, &pqm->queues, process_queue_list) { + if (!pqn->q) + continue; + + if (*num_qss_entries < tmp_qss_entries) { + set_queue_snapshot_entry(pqn->q, exception_clear_mask, &src); + + if (copy_to_user(buf, &src, *entry_size)) { + r = -EFAULT; + break; + } + buf += tmp_entry_size; + } + *num_qss_entries += 1; + } + + mutex_unlock(&pqm->process->event_mutex); + return r; +} + static int get_queue_data_sizes(struct kfd_process_device *pdd, struct queue *q, uint32_t *mqd_size, @@ -859,7 +919,7 @@ int kfd_criu_restore_queue(struct kfd_process *p, } if (!pdd->doorbell_index && - kfd_alloc_process_doorbells(pdd->dev, &pdd->doorbell_index) < 0) { + kfd_alloc_process_doorbells(pdd->dev->kfd, &pdd->doorbell_index) < 0) { ret = -ENOMEM; goto exit; } @@ -927,7 +987,9 @@ int pqm_debugfs_mqds(struct seq_file *m, void *data) struct queue *q; enum KFD_MQD_TYPE mqd_type; struct mqd_manager *mqd_mgr; - int r = 0; + int r = 0, xcc, num_xccs = 1; + void *mqd; + uint64_t size = 0; list_for_each_entry(pqn, &pqm->queues, process_queue_list) { if (pqn->q) { @@ -943,6 +1005,7 @@ int pqm_debugfs_mqds(struct seq_file *m, void *data) seq_printf(m, " Compute queue on device %x\n", q->device->id); mqd_type = KFD_MQD_TYPE_CP; + num_xccs = NUM_XCC(q->device->xcc_mask); break; default: seq_printf(m, @@ -951,6 +1014,8 @@ int pqm_debugfs_mqds(struct seq_file *m, void *data) continue; } mqd_mgr = q->device->dqm->mqd_mgrs[mqd_type]; + size = mqd_mgr->mqd_stride(mqd_mgr, + &q->properties); } else if (pqn->kq) { q = pqn->kq->queue; mqd_mgr = pqn->kq->mqd_mgr; @@ -972,9 +1037,12 @@ int pqm_debugfs_mqds(struct seq_file *m, void *data) continue; } - r = mqd_mgr->debugfs_show_mqd(m, q->mqd); - if (r != 0) - break; + for (xcc = 0; xcc < num_xccs; xcc++) { + mqd = q->mqd + size * xcc; + r = mqd_mgr->debugfs_show_mqd(m, mqd); + if (r != 0) + break; + } } return r; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c index 0472b56de245..d9953c2b2661 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c @@ -36,7 +36,7 @@ struct kfd_smi_client { wait_queue_head_t wait_queue; /* events enabled */ uint64_t events; - struct kfd_dev *dev; + struct kfd_node *dev; spinlock_t lock; struct rcu_head rcu; pid_t pid; @@ -149,7 +149,7 @@ static void kfd_smi_ev_client_free(struct rcu_head *p) static int kfd_smi_ev_release(struct inode *inode, struct file *filep) { struct kfd_smi_client *client = filep->private_data; - struct kfd_dev *dev = client->dev; + struct kfd_node *dev = client->dev; spin_lock(&dev->smi_lock); list_del_rcu(&client->list); @@ -171,7 +171,7 @@ static bool kfd_smi_ev_enabled(pid_t pid, struct kfd_smi_client *client, return events & KFD_SMI_EVENT_MASK_FROM_INDEX(event); } -static void add_event_to_kfifo(pid_t pid, struct kfd_dev *dev, +static void add_event_to_kfifo(pid_t pid, struct kfd_node *dev, unsigned int smi_event, char *event_msg, int len) { struct kfd_smi_client *client; @@ -196,7 +196,7 @@ static void add_event_to_kfifo(pid_t pid, struct kfd_dev *dev, } __printf(4, 5) -static void kfd_smi_event_add(pid_t pid, struct kfd_dev *dev, +static void kfd_smi_event_add(pid_t pid, struct kfd_node *dev, unsigned int event, char *fmt, ...) { char fifo_in[KFD_SMI_EVENT_MSG_SIZE]; @@ -215,7 +215,7 @@ static void kfd_smi_event_add(pid_t pid, struct kfd_dev *dev, add_event_to_kfifo(pid, dev, event, fifo_in, len); } -void kfd_smi_event_update_gpu_reset(struct kfd_dev *dev, bool post_reset) +void kfd_smi_event_update_gpu_reset(struct kfd_node *dev, bool post_reset) { unsigned int event; @@ -228,7 +228,7 @@ void kfd_smi_event_update_gpu_reset(struct kfd_dev *dev, bool post_reset) kfd_smi_event_add(0, dev, event, "%x\n", dev->reset_seq_num); } -void kfd_smi_event_update_thermal_throttling(struct kfd_dev *dev, +void kfd_smi_event_update_thermal_throttling(struct kfd_node *dev, uint64_t throttle_bitmask) { kfd_smi_event_add(0, dev, KFD_SMI_EVENT_THERMAL_THROTTLE, "%llx:%llx\n", @@ -236,7 +236,7 @@ void kfd_smi_event_update_thermal_throttling(struct kfd_dev *dev, amdgpu_dpm_get_thermal_throttling_counter(dev->adev)); } -void kfd_smi_event_update_vmfault(struct kfd_dev *dev, uint16_t pasid) +void kfd_smi_event_update_vmfault(struct kfd_node *dev, uint16_t pasid) { struct amdgpu_task_info task_info; @@ -250,58 +250,58 @@ void kfd_smi_event_update_vmfault(struct kfd_dev *dev, uint16_t pasid) task_info.pid, task_info.task_name); } -void kfd_smi_event_page_fault_start(struct kfd_dev *dev, pid_t pid, +void kfd_smi_event_page_fault_start(struct kfd_node *node, pid_t pid, unsigned long address, bool write_fault, ktime_t ts) { - kfd_smi_event_add(pid, dev, KFD_SMI_EVENT_PAGE_FAULT_START, + kfd_smi_event_add(pid, node, KFD_SMI_EVENT_PAGE_FAULT_START, "%lld -%d @%lx(%x) %c\n", ktime_to_ns(ts), pid, - address, dev->id, write_fault ? 'W' : 'R'); + address, node->id, write_fault ? 'W' : 'R'); } -void kfd_smi_event_page_fault_end(struct kfd_dev *dev, pid_t pid, +void kfd_smi_event_page_fault_end(struct kfd_node *node, pid_t pid, unsigned long address, bool migration) { - kfd_smi_event_add(pid, dev, KFD_SMI_EVENT_PAGE_FAULT_END, + kfd_smi_event_add(pid, node, KFD_SMI_EVENT_PAGE_FAULT_END, "%lld -%d @%lx(%x) %c\n", ktime_get_boottime_ns(), - pid, address, dev->id, migration ? 'M' : 'U'); + pid, address, node->id, migration ? 'M' : 'U'); } -void kfd_smi_event_migration_start(struct kfd_dev *dev, pid_t pid, +void kfd_smi_event_migration_start(struct kfd_node *node, pid_t pid, unsigned long start, unsigned long end, uint32_t from, uint32_t to, uint32_t prefetch_loc, uint32_t preferred_loc, uint32_t trigger) { - kfd_smi_event_add(pid, dev, KFD_SMI_EVENT_MIGRATE_START, + kfd_smi_event_add(pid, node, KFD_SMI_EVENT_MIGRATE_START, "%lld -%d @%lx(%lx) %x->%x %x:%x %d\n", ktime_get_boottime_ns(), pid, start, end - start, from, to, prefetch_loc, preferred_loc, trigger); } -void kfd_smi_event_migration_end(struct kfd_dev *dev, pid_t pid, +void kfd_smi_event_migration_end(struct kfd_node *node, pid_t pid, unsigned long start, unsigned long end, uint32_t from, uint32_t to, uint32_t trigger) { - kfd_smi_event_add(pid, dev, KFD_SMI_EVENT_MIGRATE_END, + kfd_smi_event_add(pid, node, KFD_SMI_EVENT_MIGRATE_END, "%lld -%d @%lx(%lx) %x->%x %d\n", ktime_get_boottime_ns(), pid, start, end - start, from, to, trigger); } -void kfd_smi_event_queue_eviction(struct kfd_dev *dev, pid_t pid, +void kfd_smi_event_queue_eviction(struct kfd_node *node, pid_t pid, uint32_t trigger) { - kfd_smi_event_add(pid, dev, KFD_SMI_EVENT_QUEUE_EVICTION, + kfd_smi_event_add(pid, node, KFD_SMI_EVENT_QUEUE_EVICTION, "%lld -%d %x %d\n", ktime_get_boottime_ns(), pid, - dev->id, trigger); + node->id, trigger); } -void kfd_smi_event_queue_restore(struct kfd_dev *dev, pid_t pid) +void kfd_smi_event_queue_restore(struct kfd_node *node, pid_t pid) { - kfd_smi_event_add(pid, dev, KFD_SMI_EVENT_QUEUE_RESTORE, + kfd_smi_event_add(pid, node, KFD_SMI_EVENT_QUEUE_RESTORE, "%lld -%d %x\n", ktime_get_boottime_ns(), pid, - dev->id); + node->id); } void kfd_smi_event_queue_restore_rescheduled(struct mm_struct *mm) @@ -324,16 +324,16 @@ void kfd_smi_event_queue_restore_rescheduled(struct mm_struct *mm) kfd_unref_process(p); } -void kfd_smi_event_unmap_from_gpu(struct kfd_dev *dev, pid_t pid, +void kfd_smi_event_unmap_from_gpu(struct kfd_node *node, pid_t pid, unsigned long address, unsigned long last, uint32_t trigger) { - kfd_smi_event_add(pid, dev, KFD_SMI_EVENT_UNMAP_FROM_GPU, + kfd_smi_event_add(pid, node, KFD_SMI_EVENT_UNMAP_FROM_GPU, "%lld -%d @%lx(%lx) %x %d\n", ktime_get_boottime_ns(), - pid, address, last - address + 1, dev->id, trigger); + pid, address, last - address + 1, node->id, trigger); } -int kfd_smi_event_open(struct kfd_dev *dev, uint32_t *fd) +int kfd_smi_event_open(struct kfd_node *dev, uint32_t *fd) { struct kfd_smi_client *client; int ret; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h index 76fe4e0ec2d2..fa95c2dfd587 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h @@ -24,29 +24,29 @@ #ifndef KFD_SMI_EVENTS_H_INCLUDED #define KFD_SMI_EVENTS_H_INCLUDED -int kfd_smi_event_open(struct kfd_dev *dev, uint32_t *fd); -void kfd_smi_event_update_vmfault(struct kfd_dev *dev, uint16_t pasid); -void kfd_smi_event_update_thermal_throttling(struct kfd_dev *dev, +int kfd_smi_event_open(struct kfd_node *dev, uint32_t *fd); +void kfd_smi_event_update_vmfault(struct kfd_node *dev, uint16_t pasid); +void kfd_smi_event_update_thermal_throttling(struct kfd_node *dev, uint64_t throttle_bitmask); -void kfd_smi_event_update_gpu_reset(struct kfd_dev *dev, bool post_reset); -void kfd_smi_event_page_fault_start(struct kfd_dev *dev, pid_t pid, +void kfd_smi_event_update_gpu_reset(struct kfd_node *dev, bool post_reset); +void kfd_smi_event_page_fault_start(struct kfd_node *node, pid_t pid, unsigned long address, bool write_fault, ktime_t ts); -void kfd_smi_event_page_fault_end(struct kfd_dev *dev, pid_t pid, +void kfd_smi_event_page_fault_end(struct kfd_node *node, pid_t pid, unsigned long address, bool migration); -void kfd_smi_event_migration_start(struct kfd_dev *dev, pid_t pid, +void kfd_smi_event_migration_start(struct kfd_node *node, pid_t pid, unsigned long start, unsigned long end, uint32_t from, uint32_t to, uint32_t prefetch_loc, uint32_t preferred_loc, uint32_t trigger); -void kfd_smi_event_migration_end(struct kfd_dev *dev, pid_t pid, +void kfd_smi_event_migration_end(struct kfd_node *node, pid_t pid, unsigned long start, unsigned long end, uint32_t from, uint32_t to, uint32_t trigger); -void kfd_smi_event_queue_eviction(struct kfd_dev *dev, pid_t pid, +void kfd_smi_event_queue_eviction(struct kfd_node *node, pid_t pid, uint32_t trigger); -void kfd_smi_event_queue_restore(struct kfd_dev *dev, pid_t pid); +void kfd_smi_event_queue_restore(struct kfd_node *node, pid_t pid); void kfd_smi_event_queue_restore_rescheduled(struct mm_struct *mm); -void kfd_smi_event_unmap_from_gpu(struct kfd_dev *dev, pid_t pid, +void kfd_smi_event_unmap_from_gpu(struct kfd_node *node, pid_t pid, unsigned long address, unsigned long last, uint32_t trigger); #endif diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c index 96a138a39515..5ff1a5a89d96 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c @@ -170,12 +170,11 @@ svm_range_dma_map_dev(struct amdgpu_device *adev, struct svm_range *prange, page = hmm_pfn_to_page(hmm_pfns[i]); if (is_zone_device_page(page)) { - struct amdgpu_device *bo_adev = - amdgpu_ttm_adev(prange->svm_bo->bo->tbo.bdev); + struct amdgpu_device *bo_adev = prange->svm_bo->node->adev; addr[i] = (hmm_pfns[i] << PAGE_SHIFT) + bo_adev->vm_manager.vram_base_offset - - bo_adev->kfd.dev->pgmap.range.start; + bo_adev->kfd.pgmap.range.start; addr[i] |= SVM_RANGE_VRAM_DOMAIN; pr_debug_ratelimited("vram address: 0x%llx\n", addr[i]); continue; @@ -281,7 +280,7 @@ static void svm_range_free(struct svm_range *prange, bool update_mem_usage) if (update_mem_usage && !p->xnack_enabled) { pr_debug("unreserve prange 0x%p size: 0x%llx\n", prange, size); amdgpu_amdkfd_unreserve_mem_limit(NULL, size, - KFD_IOC_ALLOC_MEM_FLAGS_USERPTR); + KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0); } mutex_destroy(&prange->lock); mutex_destroy(&prange->migrate_mutex); @@ -314,7 +313,7 @@ svm_range *svm_range_new(struct svm_range_list *svms, uint64_t start, p = container_of(svms, struct kfd_process, svms); if (!p->xnack_enabled && update_mem_usage && amdgpu_amdkfd_reserve_mem_limit(NULL, size << PAGE_SHIFT, - KFD_IOC_ALLOC_MEM_FLAGS_USERPTR)) { + KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0)) { pr_info("SVM mapping failed, exceeds resident system memory limit\n"); kfree(prange); return NULL; @@ -424,10 +423,8 @@ static void svm_range_bo_unref(struct svm_range_bo *svm_bo) } static bool -svm_range_validate_svm_bo(struct amdgpu_device *adev, struct svm_range *prange) +svm_range_validate_svm_bo(struct kfd_node *node, struct svm_range *prange) { - struct amdgpu_device *bo_adev; - mutex_lock(&prange->lock); if (!prange->svm_bo) { mutex_unlock(&prange->lock); @@ -440,12 +437,11 @@ svm_range_validate_svm_bo(struct amdgpu_device *adev, struct svm_range *prange) } if (svm_bo_ref_unless_zero(prange->svm_bo)) { /* - * Migrate from GPU to GPU, remove range from source bo_adev - * svm_bo range list, and return false to allocate svm_bo from - * destination adev. + * Migrate from GPU to GPU, remove range from source svm_bo->node + * range list, and return false to allocate svm_bo from destination + * node. */ - bo_adev = amdgpu_ttm_adev(prange->svm_bo->bo->tbo.bdev); - if (bo_adev != adev) { + if (prange->svm_bo->node != node) { mutex_unlock(&prange->lock); spin_lock(&prange->svm_bo->list_lock); @@ -513,7 +509,7 @@ static struct svm_range_bo *svm_range_bo_new(void) } int -svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange, +svm_range_vram_node_new(struct kfd_node *node, struct svm_range *prange, bool clear) { struct amdgpu_bo_param bp; @@ -528,7 +524,7 @@ svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange, pr_debug("pasid: %x svms 0x%p [0x%lx 0x%lx]\n", p->pasid, prange->svms, prange->start, prange->last); - if (svm_range_validate_svm_bo(adev, prange)) + if (svm_range_validate_svm_bo(node, prange)) return 0; svm_bo = svm_range_bo_new(); @@ -542,6 +538,7 @@ svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange, kfree(svm_bo); return -ESRCH; } + svm_bo->node = node; svm_bo->eviction_fence = amdgpu_amdkfd_fence_create(dma_fence_context_alloc(1), mm, @@ -558,13 +555,20 @@ svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange, bp.flags |= AMDGPU_GEM_CREATE_DISCARDABLE; bp.type = ttm_bo_type_device; bp.resv = NULL; + if (node->xcp) + bp.xcp_id_plus1 = node->xcp->id + 1; - r = amdgpu_bo_create_user(adev, &bp, &ubo); + r = amdgpu_bo_create_user(node->adev, &bp, &ubo); if (r) { pr_debug("failed %d to create bo\n", r); goto create_bo_failed; } bo = &ubo->bo; + + pr_debug("alloc bo at offset 0x%lx size 0x%lx on partition %d\n", + bo->tbo.resource->start << PAGE_SHIFT, bp.size, + bp.xcp_id_plus1 - 1); + r = amdgpu_bo_reserve(bo, true); if (r) { pr_debug("failed %d to reserve bo\n", r); @@ -617,45 +621,30 @@ void svm_range_vram_node_free(struct svm_range *prange) prange->ttm_res = NULL; } -struct amdgpu_device * -svm_range_get_adev_by_id(struct svm_range *prange, uint32_t gpu_id) +struct kfd_node * +svm_range_get_node_by_id(struct svm_range *prange, uint32_t gpu_id) { - struct kfd_process_device *pdd; struct kfd_process *p; - int32_t gpu_idx; + struct kfd_process_device *pdd; p = container_of(prange->svms, struct kfd_process, svms); - - gpu_idx = kfd_process_gpuidx_from_gpuid(p, gpu_id); - if (gpu_idx < 0) { - pr_debug("failed to get device by id 0x%x\n", gpu_id); - return NULL; - } - pdd = kfd_process_device_from_gpuidx(p, gpu_idx); + pdd = kfd_process_device_data_by_id(p, gpu_id); if (!pdd) { - pr_debug("failed to get device by idx 0x%x\n", gpu_idx); + pr_debug("failed to get kfd process device by id 0x%x\n", gpu_id); return NULL; } - return pdd->dev->adev; + return pdd->dev; } struct kfd_process_device * -svm_range_get_pdd_by_adev(struct svm_range *prange, struct amdgpu_device *adev) +svm_range_get_pdd_by_node(struct svm_range *prange, struct kfd_node *node) { struct kfd_process *p; - int32_t gpu_idx, gpuid; - int r; p = container_of(prange->svms, struct kfd_process, svms); - r = kfd_process_gpuid_from_adev(p, adev, &gpuid, &gpu_idx); - if (r) { - pr_debug("failed to get device id by adev %p\n", adev); - return NULL; - } - - return kfd_process_device_from_gpuidx(p, gpu_idx); + return kfd_get_process_device_data(node, p); } static int svm_range_bo_validate(void *param, struct amdgpu_bo *bo) @@ -735,7 +724,9 @@ svm_range_apply_attrs(struct kfd_process *p, struct svm_range *prange, case KFD_IOCTL_SVM_ATTR_ACCESS: case KFD_IOCTL_SVM_ATTR_ACCESS_IN_PLACE: case KFD_IOCTL_SVM_ATTR_NO_ACCESS: - *update_mapping = true; + if (!p->xnack_enabled) + *update_mapping = true; + gpuidx = kfd_process_gpuidx_from_gpuid(p, attrs[i].value); if (attrs[i].type == KFD_IOCTL_SVM_ATTR_NO_ACCESS) { @@ -818,7 +809,7 @@ svm_range_is_same_attrs(struct kfd_process *p, struct svm_range *prange, } } - return true; + return !prange->is_error_flag; } /** @@ -1146,31 +1137,39 @@ svm_range_split_by_granularity(struct kfd_process *p, struct mm_struct *mm, } return 0; } +static bool +svm_nodes_in_same_hive(struct kfd_node *node_a, struct kfd_node *node_b) +{ + return (node_a->adev == node_b->adev || + amdgpu_xgmi_same_hive(node_a->adev, node_b->adev)); +} static uint64_t -svm_range_get_pte_flags(struct amdgpu_device *adev, struct svm_range *prange, - int domain) +svm_range_get_pte_flags(struct kfd_node *node, + struct svm_range *prange, int domain) { - struct amdgpu_device *bo_adev; + struct kfd_node *bo_node; uint32_t flags = prange->flags; uint32_t mapping_flags = 0; uint64_t pte_flags; bool snoop = (domain != SVM_RANGE_VRAM_DOMAIN); bool coherent = flags & KFD_IOCTL_SVM_FLAG_COHERENT; + bool uncached = false; /*flags & KFD_IOCTL_SVM_FLAG_UNCACHED;*/ + unsigned int mtype_local; if (domain == SVM_RANGE_VRAM_DOMAIN) - bo_adev = amdgpu_ttm_adev(prange->svm_bo->bo->tbo.bdev); + bo_node = prange->svm_bo->node; - switch (KFD_GC_VERSION(adev->kfd.dev)) { + switch (node->adev->ip_versions[GC_HWIP][0]) { case IP_VERSION(9, 4, 1): if (domain == SVM_RANGE_VRAM_DOMAIN) { - if (bo_adev == adev) { + if (bo_node == node) { mapping_flags |= coherent ? AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW; } else { mapping_flags |= coherent ? AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC; - if (amdgpu_xgmi_same_hive(adev, bo_adev)) + if (svm_nodes_in_same_hive(node, bo_node)) snoop = true; } } else { @@ -1180,15 +1179,15 @@ svm_range_get_pte_flags(struct amdgpu_device *adev, struct svm_range *prange, break; case IP_VERSION(9, 4, 2): if (domain == SVM_RANGE_VRAM_DOMAIN) { - if (bo_adev == adev) { + if (bo_node == node) { mapping_flags |= coherent ? AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW; - if (adev->gmc.xgmi.connected_to_cpu) + if (node->adev->gmc.xgmi.connected_to_cpu) snoop = true; } else { mapping_flags |= coherent ? AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC; - if (amdgpu_xgmi_same_hive(adev, bo_adev)) + if (svm_nodes_in_same_hive(node, bo_node)) snoop = true; } } else { @@ -1196,6 +1195,37 @@ svm_range_get_pte_flags(struct amdgpu_device *adev, struct svm_range *prange, AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC; } break; + case IP_VERSION(9, 4, 3): + mtype_local = amdgpu_mtype_local == 1 ? AMDGPU_VM_MTYPE_NC : + (amdgpu_mtype_local == 2 ? AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW); + snoop = true; + if (uncached) { + mapping_flags |= AMDGPU_VM_MTYPE_UC; + } else if (domain == SVM_RANGE_VRAM_DOMAIN) { + /* local HBM region close to partition */ + if (bo_node->adev == node->adev && + (!bo_node->xcp || !node->xcp || bo_node->xcp->mem_id == node->xcp->mem_id)) + mapping_flags |= mtype_local; + /* local HBM region far from partition or remote XGMI GPU */ + else if (svm_nodes_in_same_hive(bo_node, node)) + mapping_flags |= AMDGPU_VM_MTYPE_NC; + /* PCIe P2P */ + else + mapping_flags |= AMDGPU_VM_MTYPE_UC; + /* system memory accessed by the APU */ + } else if (node->adev->flags & AMD_IS_APU) { + /* On NUMA systems, locality is determined per-page + * in amdgpu_gmc_override_vm_pte_flags + */ + if (num_possible_nodes() <= 1) + mapping_flags |= mtype_local; + else + mapping_flags |= AMDGPU_VM_MTYPE_NC; + /* system memory accessed by the dGPU */ + } else { + mapping_flags |= AMDGPU_VM_MTYPE_UC; + } + break; default: mapping_flags |= coherent ? AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC; @@ -1212,7 +1242,7 @@ svm_range_get_pte_flags(struct amdgpu_device *adev, struct svm_range *prange, pte_flags |= (domain == SVM_RANGE_VRAM_DOMAIN) ? 0 : AMDGPU_PTE_SYSTEM; pte_flags |= snoop ? AMDGPU_PTE_SNOOPED : 0; - pte_flags |= amdgpu_gem_va_map_flags(adev, mapping_flags); + pte_flags |= amdgpu_gem_va_map_flags(node->adev, mapping_flags); return pte_flags; } @@ -1319,7 +1349,7 @@ svm_range_map_to_gpu(struct kfd_process_device *pdd, struct svm_range *prange, pr_debug("Mapping range [0x%lx 0x%llx] on domain: %s\n", last_start, prange->start + i, last_domain ? "GPU" : "CPU"); - pte_flags = svm_range_get_pte_flags(adev, prange, last_domain); + pte_flags = svm_range_get_pte_flags(pdd->dev, prange, last_domain); if (readonly) pte_flags &= ~AMDGPU_PTE_WRITEABLE; @@ -1328,6 +1358,10 @@ svm_range_map_to_gpu(struct kfd_process_device *pdd, struct svm_range *prange, (last_domain == SVM_RANGE_VRAM_DOMAIN) ? 1 : 0, pte_flags); + /* For dGPU mode, we use same vm_manager to allocate VRAM for + * different memory partition based on fpfn/lpfn, we should use + * same vm_manager.vram_base_offset regardless memory partition. + */ r = amdgpu_vm_update_range(adev, vm, false, false, flush_tlb, NULL, last_start, prange->start + i, pte_flags, @@ -1365,16 +1399,14 @@ svm_range_map_to_gpus(struct svm_range *prange, unsigned long offset, unsigned long *bitmap, bool wait, bool flush_tlb) { struct kfd_process_device *pdd; - struct amdgpu_device *bo_adev; + struct amdgpu_device *bo_adev = NULL; struct kfd_process *p; struct dma_fence *fence = NULL; uint32_t gpuidx; int r = 0; if (prange->svm_bo && prange->ttm_res) - bo_adev = amdgpu_ttm_adev(prange->svm_bo->bo->tbo.bdev); - else - bo_adev = NULL; + bo_adev = prange->svm_bo->node->adev; p = container_of(prange->svms, struct kfd_process, svms); for_each_set_bit(gpuidx, bitmap, MAX_GPU_INSTANCE) { @@ -1522,48 +1554,54 @@ static int svm_range_validate_and_map(struct mm_struct *mm, struct svm_range *prange, int32_t gpuidx, bool intr, bool wait, bool flush_tlb) { - struct svm_validate_context ctx; + struct svm_validate_context *ctx; unsigned long start, end, addr; struct kfd_process *p; void *owner; int32_t idx; int r = 0; - ctx.process = container_of(prange->svms, struct kfd_process, svms); - ctx.prange = prange; - ctx.intr = intr; + ctx = kzalloc(sizeof(struct svm_validate_context), GFP_KERNEL); + if (!ctx) + return -ENOMEM; + ctx->process = container_of(prange->svms, struct kfd_process, svms); + ctx->prange = prange; + ctx->intr = intr; if (gpuidx < MAX_GPU_INSTANCE) { - bitmap_zero(ctx.bitmap, MAX_GPU_INSTANCE); - bitmap_set(ctx.bitmap, gpuidx, 1); - } else if (ctx.process->xnack_enabled) { - bitmap_copy(ctx.bitmap, prange->bitmap_aip, MAX_GPU_INSTANCE); + bitmap_zero(ctx->bitmap, MAX_GPU_INSTANCE); + bitmap_set(ctx->bitmap, gpuidx, 1); + } else if (ctx->process->xnack_enabled) { + bitmap_copy(ctx->bitmap, prange->bitmap_aip, MAX_GPU_INSTANCE); /* If prefetch range to GPU, or GPU retry fault migrate range to * GPU, which has ACCESS attribute to the range, create mapping * on that GPU. */ if (prange->actual_loc) { - gpuidx = kfd_process_gpuidx_from_gpuid(ctx.process, + gpuidx = kfd_process_gpuidx_from_gpuid(ctx->process, prange->actual_loc); if (gpuidx < 0) { WARN_ONCE(1, "failed get device by id 0x%x\n", prange->actual_loc); - return -EINVAL; + r = -EINVAL; + goto free_ctx; } if (test_bit(gpuidx, prange->bitmap_access)) - bitmap_set(ctx.bitmap, gpuidx, 1); + bitmap_set(ctx->bitmap, gpuidx, 1); } } else { - bitmap_or(ctx.bitmap, prange->bitmap_access, + bitmap_or(ctx->bitmap, prange->bitmap_access, prange->bitmap_aip, MAX_GPU_INSTANCE); } - if (bitmap_empty(ctx.bitmap, MAX_GPU_INSTANCE)) { - if (!prange->mapped_to_gpu) - return 0; + if (bitmap_empty(ctx->bitmap, MAX_GPU_INSTANCE)) { + if (!prange->mapped_to_gpu) { + r = 0; + goto free_ctx; + } - bitmap_copy(ctx.bitmap, prange->bitmap_access, MAX_GPU_INSTANCE); + bitmap_copy(ctx->bitmap, prange->bitmap_access, MAX_GPU_INSTANCE); } if (prange->actual_loc && !prange->ttm_res) { @@ -1571,15 +1609,16 @@ static int svm_range_validate_and_map(struct mm_struct *mm, * svm_migrate_ram_to_vram after allocating a BO. */ WARN_ONCE(1, "VRAM BO missing during validation\n"); - return -EINVAL; + r = -EINVAL; + goto free_ctx; } - svm_range_reserve_bos(&ctx); + svm_range_reserve_bos(ctx); p = container_of(prange->svms, struct kfd_process, svms); - owner = kfd_svm_page_owner(p, find_first_bit(ctx.bitmap, + owner = kfd_svm_page_owner(p, find_first_bit(ctx->bitmap, MAX_GPU_INSTANCE)); - for_each_set_bit(idx, ctx.bitmap, MAX_GPU_INSTANCE) { + for_each_set_bit(idx, ctx->bitmap, MAX_GPU_INSTANCE) { if (kfd_svm_page_owner(p, idx) != owner) { owner = NULL; break; @@ -1616,7 +1655,7 @@ static int svm_range_validate_and_map(struct mm_struct *mm, } offset = (addr - start) >> PAGE_SHIFT; - r = svm_range_dma_map(prange, ctx.bitmap, offset, npages, + r = svm_range_dma_map(prange, ctx->bitmap, offset, npages, hmm_range->hmm_pfns); if (r) { pr_debug("failed %d to dma map range\n", r); @@ -1636,7 +1675,7 @@ static int svm_range_validate_and_map(struct mm_struct *mm, } r = svm_range_map_to_gpus(prange, offset, npages, readonly, - ctx.bitmap, wait, flush_tlb); + ctx->bitmap, wait, flush_tlb); unlock_out: svm_range_unlock(prange); @@ -1650,11 +1689,15 @@ unlock_out: } unreserve_out: - svm_range_unreserve_bos(&ctx); + svm_range_unreserve_bos(ctx); + prange->is_error_flag = !!r; if (!r) prange->validate_timestamp = ktime_get_boottime(); +free_ctx: + kfree(ctx); + return r; } @@ -1783,6 +1826,7 @@ out_reschedule: * @mm: current process mm_struct * @start: starting process queue number * @last: last process queue number + * @event: mmu notifier event when range is evicted or migrated * * Stop all queues of the process to ensure GPU doesn't access the memory, then * return to let CPU evict the buffer and proceed CPU pagetable update. @@ -1906,14 +1950,23 @@ void svm_range_set_max_pages(struct amdgpu_device *adev) { uint64_t max_pages; uint64_t pages, _pages; + uint64_t min_pages = 0; + int i, id; + + for (i = 0; i < adev->kfd.dev->num_nodes; i++) { + if (adev->kfd.dev->nodes[i]->xcp) + id = adev->kfd.dev->nodes[i]->xcp->id; + else + id = -1; + pages = KFD_XCP_MEMORY_SIZE(adev, id) >> 17; + pages = clamp(pages, 1ULL << 9, 1ULL << 18); + pages = rounddown_pow_of_two(pages); + min_pages = min_not_zero(min_pages, pages); + } - /* 1/32 VRAM size in pages */ - pages = adev->gmc.real_vram_size >> 17; - pages = clamp(pages, 1ULL << 9, 1ULL << 18); - pages = rounddown_pow_of_two(pages); do { max_pages = READ_ONCE(max_svm_range_pages); - _pages = min_not_zero(max_pages, pages); + _pages = min_not_zero(max_pages, min_pages); } while (cmpxchg(&max_svm_range_pages, max_pages, _pages) != max_pages); } @@ -2507,29 +2560,31 @@ svm_range_from_addr(struct svm_range_list *svms, unsigned long addr, */ static int32_t svm_range_best_restore_location(struct svm_range *prange, - struct amdgpu_device *adev, + struct kfd_node *node, int32_t *gpuidx) { - struct amdgpu_device *bo_adev, *preferred_adev; + struct kfd_node *bo_node, *preferred_node; struct kfd_process *p; uint32_t gpuid; int r; p = container_of(prange->svms, struct kfd_process, svms); - r = kfd_process_gpuid_from_adev(p, adev, &gpuid, gpuidx); + r = kfd_process_gpuid_from_node(p, node, &gpuid, gpuidx); if (r < 0) { pr_debug("failed to get gpuid from kgd\n"); return -1; } + if (node->adev->gmc.is_app_apu) + return 0; + if (prange->preferred_loc == gpuid || prange->preferred_loc == KFD_IOCTL_SVM_LOCATION_SYSMEM) { return prange->preferred_loc; } else if (prange->preferred_loc != KFD_IOCTL_SVM_LOCATION_UNDEFINED) { - preferred_adev = svm_range_get_adev_by_id(prange, - prange->preferred_loc); - if (amdgpu_xgmi_same_hive(adev, preferred_adev)) + preferred_node = svm_range_get_node_by_id(prange, prange->preferred_loc); + if (preferred_node && svm_nodes_in_same_hive(node, preferred_node)) return prange->preferred_loc; /* fall through */ } @@ -2541,8 +2596,8 @@ svm_range_best_restore_location(struct svm_range *prange, if (!prange->actual_loc) return 0; - bo_adev = svm_range_get_adev_by_id(prange, prange->actual_loc); - if (amdgpu_xgmi_same_hive(adev, bo_adev)) + bo_node = svm_range_get_node_by_id(prange, prange->actual_loc); + if (bo_node && svm_nodes_in_same_hive(node, bo_node)) return prange->actual_loc; else return 0; @@ -2659,7 +2714,7 @@ svm_range_check_vm_userptr(struct kfd_process *p, uint64_t start, uint64_t last, } static struct -svm_range *svm_range_create_unregistered_range(struct amdgpu_device *adev, +svm_range *svm_range_create_unregistered_range(struct kfd_node *node, struct kfd_process *p, struct mm_struct *mm, int64_t addr) @@ -2694,7 +2749,7 @@ svm_range *svm_range_create_unregistered_range(struct amdgpu_device *adev, pr_debug("Failed to create prange in address [0x%llx]\n", addr); return NULL; } - if (kfd_process_gpuid_from_adev(p, adev, &gpuid, &gpuidx)) { + if (kfd_process_gpuid_from_node(p, node, &gpuid, &gpuidx)) { pr_debug("failed to get gpuid from kgd\n"); svm_range_free(prange, true); return NULL; @@ -2748,7 +2803,7 @@ static bool svm_range_skip_recover(struct svm_range *prange) } static void -svm_range_count_fault(struct amdgpu_device *adev, struct kfd_process *p, +svm_range_count_fault(struct kfd_node *node, struct kfd_process *p, int32_t gpuidx) { struct kfd_process_device *pdd; @@ -2761,7 +2816,7 @@ svm_range_count_fault(struct amdgpu_device *adev, struct kfd_process *p, uint32_t gpuid; int r; - r = kfd_process_gpuid_from_adev(p, adev, &gpuid, &gpuidx); + r = kfd_process_gpuid_from_node(p, node, &gpuid, &gpuidx); if (r < 0) return; } @@ -2789,6 +2844,7 @@ svm_fault_allowed(struct vm_area_struct *vma, bool write_fault) int svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid, + uint32_t vmid, uint32_t node_id, uint64_t addr, bool write_fault) { struct mm_struct *mm = NULL; @@ -2796,6 +2852,7 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid, struct svm_range *prange; struct kfd_process *p; ktime_t timestamp = ktime_get_boottime(); + struct kfd_node *node; int32_t best_loc; int32_t gpuidx = MAX_GPU_INSTANCE; bool write_locked = false; @@ -2803,7 +2860,7 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid, bool migration = false; int r = 0; - if (!KFD_IS_SVM_API_SUPPORTED(adev->kfd.dev)) { + if (!KFD_IS_SVM_API_SUPPORTED(adev)) { pr_debug("device does not support SVM\n"); return -EFAULT; } @@ -2839,6 +2896,13 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid, goto out; } + node = kfd_node_by_irq_ids(adev, node_id, vmid); + if (!node) { + pr_debug("kfd node does not exist node_id: %d, vmid: %d\n", node_id, + vmid); + r = -EFAULT; + goto out; + } mmap_read_lock(mm); retry_write_locked: mutex_lock(&svms->lock); @@ -2857,7 +2921,7 @@ retry_write_locked: write_locked = true; goto retry_write_locked; } - prange = svm_range_create_unregistered_range(adev, p, mm, addr); + prange = svm_range_create_unregistered_range(node, p, mm, addr); if (!prange) { pr_debug("failed to create unregistered range svms 0x%p address [0x%llx]\n", svms, addr); @@ -2872,7 +2936,7 @@ retry_write_locked: mutex_lock(&prange->migrate_mutex); if (svm_range_skip_recover(prange)) { - amdgpu_gmc_filter_faults_remove(adev, addr, pasid); + amdgpu_gmc_filter_faults_remove(node->adev, addr, pasid); r = 0; goto out_unlock_range; } @@ -2903,7 +2967,7 @@ retry_write_locked: goto out_unlock_range; } - best_loc = svm_range_best_restore_location(prange, adev, &gpuidx); + best_loc = svm_range_best_restore_location(prange, node, &gpuidx); if (best_loc == -1) { pr_debug("svms %p failed get best restore loc [0x%lx 0x%lx]\n", svms, prange->start, prange->last); @@ -2915,7 +2979,7 @@ retry_write_locked: svms, prange->start, prange->last, best_loc, prange->actual_loc); - kfd_smi_event_page_fault_start(adev->kfd.dev, p->lead_thread->pid, addr, + kfd_smi_event_page_fault_start(node, p->lead_thread->pid, addr, write_fault, timestamp); if (prange->actual_loc != best_loc) { @@ -2953,7 +3017,7 @@ retry_write_locked: pr_debug("failed %d to map svms 0x%p [0x%lx 0x%lx] to gpus\n", r, svms, prange->start, prange->last); - kfd_smi_event_page_fault_end(adev->kfd.dev, p->lead_thread->pid, addr, + kfd_smi_event_page_fault_end(node, p->lead_thread->pid, addr, migration); out_unlock_range: @@ -2962,7 +3026,7 @@ out_unlock_svms: mutex_unlock(&svms->lock); mmap_read_unlock(mm); - svm_range_count_fault(adev, p, gpuidx); + svm_range_count_fault(node, p, gpuidx); mmput(mm); out: @@ -2970,7 +3034,7 @@ out: if (r == -EAGAIN) { pr_debug("recover vm fault later\n"); - amdgpu_gmc_filter_faults_remove(adev, addr, pasid); + amdgpu_gmc_filter_faults_remove(node->adev, addr, pasid); r = 0; } return r; @@ -2994,10 +3058,10 @@ svm_range_switch_xnack_reserve_mem(struct kfd_process *p, bool xnack_enabled) size = (pchild->last - pchild->start + 1) << PAGE_SHIFT; if (xnack_enabled) { amdgpu_amdkfd_unreserve_mem_limit(NULL, size, - KFD_IOC_ALLOC_MEM_FLAGS_USERPTR); + KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0); } else { r = amdgpu_amdkfd_reserve_mem_limit(NULL, size, - KFD_IOC_ALLOC_MEM_FLAGS_USERPTR); + KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0); if (r) goto out_unlock; reserved_size += size; @@ -3007,10 +3071,10 @@ svm_range_switch_xnack_reserve_mem(struct kfd_process *p, bool xnack_enabled) size = (prange->last - prange->start + 1) << PAGE_SHIFT; if (xnack_enabled) { amdgpu_amdkfd_unreserve_mem_limit(NULL, size, - KFD_IOC_ALLOC_MEM_FLAGS_USERPTR); + KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0); } else { r = amdgpu_amdkfd_reserve_mem_limit(NULL, size, - KFD_IOC_ALLOC_MEM_FLAGS_USERPTR); + KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0); if (r) goto out_unlock; reserved_size += size; @@ -3023,7 +3087,7 @@ out_unlock: if (r) amdgpu_amdkfd_unreserve_mem_limit(NULL, reserved_size, - KFD_IOC_ALLOC_MEM_FLAGS_USERPTR); + KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0); else /* Change xnack mode must be inside svms lock, to avoid race with * svm_range_deferred_list_work unreserve memory in parallel. @@ -3081,7 +3145,7 @@ int svm_range_list_init(struct kfd_process *p) spin_lock_init(&svms->deferred_list_lock); for (i = 0; i < p->n_pdds; i++) - if (KFD_IS_SVM_API_SUPPORTED(p->pdds[i]->dev)) + if (KFD_IS_SVM_API_SUPPORTED(p->pdds[i]->dev->adev)) bitmap_set(svms->bitmap_supported, i, 1); return 0; @@ -3212,7 +3276,7 @@ svm_range_best_prefetch_location(struct svm_range *prange) DECLARE_BITMAP(bitmap, MAX_GPU_INSTANCE); uint32_t best_loc = prange->prefetch_loc; struct kfd_process_device *pdd; - struct amdgpu_device *bo_adev; + struct kfd_node *bo_node; struct kfd_process *p; uint32_t gpuidx; @@ -3221,9 +3285,14 @@ svm_range_best_prefetch_location(struct svm_range *prange) if (!best_loc || best_loc == KFD_IOCTL_SVM_LOCATION_UNDEFINED) goto out; - bo_adev = svm_range_get_adev_by_id(prange, best_loc); - if (!bo_adev) { - WARN_ONCE(1, "failed to get device by id 0x%x\n", best_loc); + bo_node = svm_range_get_node_by_id(prange, best_loc); + if (!bo_node) { + WARN_ONCE(1, "failed to get valid kfd node at id%x\n", best_loc); + best_loc = 0; + goto out; + } + + if (bo_node->adev->gmc.is_app_apu) { best_loc = 0; goto out; } @@ -3241,10 +3310,10 @@ svm_range_best_prefetch_location(struct svm_range *prange) continue; } - if (pdd->dev->adev == bo_adev) + if (pdd->dev->adev == bo_node->adev) continue; - if (!amdgpu_xgmi_same_hive(pdd->dev->adev, bo_adev)) { + if (!svm_nodes_in_same_hive(pdd->dev, bo_node)) { best_loc = 0; break; } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.h b/drivers/gpu/drm/amd/amdkfd/kfd_svm.h index 7a33b93f9df6..21b14510882b 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.h @@ -48,6 +48,7 @@ struct svm_range_bo { struct work_struct eviction_work; uint32_t evicting; struct work_struct release_work; + struct kfd_node *node; }; enum svm_work_list_ops { @@ -133,6 +134,7 @@ struct svm_range { DECLARE_BITMAP(bitmap_aip, MAX_GPU_INSTANCE); bool validated_once; bool mapped_to_gpu; + bool is_error_flag; }; static inline void svm_range_lock(struct svm_range *prange) @@ -163,16 +165,17 @@ int svm_ioctl(struct kfd_process *p, enum kfd_ioctl_svm_op op, uint64_t start, struct svm_range *svm_range_from_addr(struct svm_range_list *svms, unsigned long addr, struct svm_range **parent); -struct amdgpu_device *svm_range_get_adev_by_id(struct svm_range *prange, - uint32_t id); -int svm_range_vram_node_new(struct amdgpu_device *adev, - struct svm_range *prange, bool clear); +struct kfd_node *svm_range_get_node_by_id(struct svm_range *prange, + uint32_t gpu_id); +int svm_range_vram_node_new(struct kfd_node *node, struct svm_range *prange, + bool clear); void svm_range_vram_node_free(struct svm_range *prange); int svm_range_split_by_granularity(struct kfd_process *p, struct mm_struct *mm, unsigned long addr, struct svm_range *parent, struct svm_range *prange); -int svm_range_restore_pages(struct amdgpu_device *adev, - unsigned int pasid, uint64_t addr, bool write_fault); +int svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid, + uint32_t vmid, uint32_t node_id, uint64_t addr, + bool write_fault); int svm_range_schedule_evict_svm_bo(struct amdgpu_amdkfd_fence *fence); void svm_range_add_list_work(struct svm_range_list *svms, struct svm_range *prange, struct mm_struct *mm, @@ -192,13 +195,14 @@ int kfd_criu_restore_svm(struct kfd_process *p, uint64_t max_priv_data_size); int kfd_criu_resume_svm(struct kfd_process *p); struct kfd_process_device * -svm_range_get_pdd_by_adev(struct svm_range *prange, struct amdgpu_device *adev); +svm_range_get_pdd_by_node(struct svm_range *prange, struct kfd_node *node); void svm_range_list_lock_and_flush_work(struct svm_range_list *svms, struct mm_struct *mm); /* SVM API and HMM page migration work together, device memory type * is initialized to not 0 when page migration register device memory. */ -#define KFD_IS_SVM_API_SUPPORTED(dev) ((dev)->pgmap.type != 0) +#define KFD_IS_SVM_API_SUPPORTED(adev) ((adev)->kfd.pgmap.type != 0 ||\ + (adev)->gmc.is_app_apu) void svm_range_bo_unref_async(struct svm_range_bo *svm_bo); @@ -219,8 +223,9 @@ static inline void svm_range_list_fini(struct kfd_process *p) } static inline int svm_range_restore_pages(struct amdgpu_device *adev, - unsigned int pasid, uint64_t addr, - bool write_fault) + unsigned int pasid, + uint32_t client_id, uint32_t node_id, + uint64_t addr, bool write_fault) { return -EFAULT; } @@ -261,6 +266,10 @@ static inline int kfd_criu_resume_svm(struct kfd_process *p) return 0; } +static inline void svm_range_set_max_pages(struct amdgpu_device *adev) +{ +} + #define KFD_IS_SVM_API_SUPPORTED(dev) false #endif /* IS_ENABLED(CONFIG_HSA_AMD_SVM) */ diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c index 8e4124dcb6e4..90b86a6ac7bd 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c @@ -96,7 +96,7 @@ struct kfd_topology_device *kfd_topology_device_by_id(uint32_t gpu_id) return ret; } -struct kfd_dev *kfd_device_by_id(uint32_t gpu_id) +struct kfd_node *kfd_device_by_id(uint32_t gpu_id) { struct kfd_topology_device *top_dev; @@ -107,10 +107,10 @@ struct kfd_dev *kfd_device_by_id(uint32_t gpu_id) return top_dev->gpu; } -struct kfd_dev *kfd_device_by_pci_dev(const struct pci_dev *pdev) +struct kfd_node *kfd_device_by_pci_dev(const struct pci_dev *pdev) { struct kfd_topology_device *top_dev; - struct kfd_dev *device = NULL; + struct kfd_node *device = NULL; down_read(&topology_lock); @@ -125,24 +125,6 @@ struct kfd_dev *kfd_device_by_pci_dev(const struct pci_dev *pdev) return device; } -struct kfd_dev *kfd_device_by_adev(const struct amdgpu_device *adev) -{ - struct kfd_topology_device *top_dev; - struct kfd_dev *device = NULL; - - down_read(&topology_lock); - - list_for_each_entry(top_dev, &topology_device_list, list) - if (top_dev->gpu && top_dev->gpu->adev == adev) { - device = top_dev->gpu; - break; - } - - up_read(&topology_lock); - - return device; -} - /* Called with write topology_lock acquired */ static void kfd_release_topology_device(struct kfd_topology_device *dev) { @@ -468,7 +450,8 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr, sysfs_show_32bit_prop(buffer, offs, "cpu_cores_count", dev->node_props.cpu_cores_count); sysfs_show_32bit_prop(buffer, offs, "simd_count", - dev->gpu ? dev->node_props.simd_count : 0); + dev->gpu ? (dev->node_props.simd_count * + NUM_XCC(dev->gpu->xcc_mask)) : 0); sysfs_show_32bit_prop(buffer, offs, "mem_banks_count", dev->node_props.mem_banks_count); sysfs_show_32bit_prop(buffer, offs, "caches_count", @@ -492,7 +475,8 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr, sysfs_show_32bit_prop(buffer, offs, "wave_front_size", dev->node_props.wave_front_size); sysfs_show_32bit_prop(buffer, offs, "array_count", - dev->node_props.array_count); + dev->gpu ? (dev->node_props.array_count * + NUM_XCC(dev->gpu->xcc_mask)) : 0); sysfs_show_32bit_prop(buffer, offs, "simd_arrays_per_engine", dev->node_props.simd_arrays_per_engine); sysfs_show_32bit_prop(buffer, offs, "cu_per_simd_array", @@ -526,7 +510,7 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr, if (dev->gpu) { log_max_watch_addr = - __ilog2_u32(dev->gpu->device_info.num_of_watch_points); + __ilog2_u32(dev->gpu->kfd->device_info.num_of_watch_points); if (log_max_watch_addr) { dev->node_props.capability |= @@ -548,14 +532,17 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr, sysfs_show_64bit_prop(buffer, offs, "local_mem_size", 0ULL); sysfs_show_32bit_prop(buffer, offs, "fw_version", - dev->gpu->mec_fw_version); + dev->gpu->kfd->mec_fw_version); sysfs_show_32bit_prop(buffer, offs, "capability", dev->node_props.capability); + sysfs_show_64bit_prop(buffer, offs, "debug_prop", + dev->node_props.debug_prop); sysfs_show_32bit_prop(buffer, offs, "sdma_fw_version", - dev->gpu->sdma_fw_version); + dev->gpu->kfd->sdma_fw_version); sysfs_show_64bit_prop(buffer, offs, "unique_id", dev->gpu->adev->unique_id); - + sysfs_show_32bit_prop(buffer, offs, "num_xcc", + NUM_XCC(dev->gpu->xcc_mask)); } return sysfs_show_32bit_prop(buffer, offs, "max_engine_clk_ccompute", @@ -1157,10 +1144,10 @@ void kfd_topology_shutdown(void) up_write(&topology_lock); } -static uint32_t kfd_generate_gpu_id(struct kfd_dev *gpu) +static uint32_t kfd_generate_gpu_id(struct kfd_node *gpu) { uint32_t hashout; - uint32_t buf[7]; + uint32_t buf[8]; uint64_t local_mem_size; int i; @@ -1177,8 +1164,9 @@ static uint32_t kfd_generate_gpu_id(struct kfd_dev *gpu) buf[4] = gpu->adev->pdev->bus->number; buf[5] = lower_32_bits(local_mem_size); buf[6] = upper_32_bits(local_mem_size); + buf[7] = (ffs(gpu->xcc_mask) - 1) | (NUM_XCC(gpu->xcc_mask) << 16); - for (i = 0, hashout = 0; i < 7; i++) + for (i = 0, hashout = 0; i < 8; i++) hashout ^= hash_32(buf[i], KFD_GPU_ID_HASH_WIDTH); return hashout; @@ -1188,7 +1176,7 @@ static uint32_t kfd_generate_gpu_id(struct kfd_dev *gpu) * list then return NULL. This means a new topology device has to * be created for this GPU. */ -static struct kfd_topology_device *kfd_assign_gpu(struct kfd_dev *gpu) +static struct kfd_topology_device *kfd_assign_gpu(struct kfd_node *gpu) { struct kfd_topology_device *dev; struct kfd_topology_device *out_dev = NULL; @@ -1201,7 +1189,7 @@ static struct kfd_topology_device *kfd_assign_gpu(struct kfd_dev *gpu) /* Discrete GPUs need their own topology device list * entries. Don't assign them to CPU/APU nodes. */ - if (!gpu->use_iommu_v2 && + if (!gpu->kfd->use_iommu_v2 && dev->node_props.cpu_cores_count) continue; @@ -1248,7 +1236,8 @@ static void kfd_fill_mem_clk_max_info(struct kfd_topology_device *dev) * for APUs - If CRAT from ACPI reports more than one bank, then * all the banks will report the same mem_clk_max information */ - amdgpu_amdkfd_get_local_mem_info(dev->gpu->adev, &local_mem_info); + amdgpu_amdkfd_get_local_mem_info(dev->gpu->adev, &local_mem_info, + dev->gpu->xcp); list_for_each_entry(mem, &dev->mem_props, list) mem->mem_clk_max = local_mem_info.mem_clk_max; @@ -1275,7 +1264,7 @@ static void kfd_set_iolink_no_atomics(struct kfd_topology_device *dev, CRAT_IOLINK_FLAGS_NO_ATOMICS_64_BIT; /* set gpu (dev) flags. */ } else { - if (!dev->gpu->pci_atomic_requested || + if (!dev->gpu->kfd->pci_atomic_requested || dev->gpu->adev->asic_type == CHIP_HAWAII) link->flags |= CRAT_IOLINK_FLAGS_NO_ATOMICS_32_BIT | CRAT_IOLINK_FLAGS_NO_ATOMICS_64_BIT; @@ -1323,10 +1312,16 @@ static void kfd_fill_iolink_non_crat_info(struct kfd_topology_device *dev) continue; /* Include the CPU peer in GPU hive if connected over xGMI. */ - if (!peer_dev->gpu && !peer_dev->node_props.hive_id && - dev->node_props.hive_id && - dev->gpu->adev->gmc.xgmi.connected_to_cpu) + if (!peer_dev->gpu && + link->iolink_type == CRAT_IOLINK_TYPE_XGMI) { + /* + * If the GPU is not part of a GPU hive, use its pci + * device location as the hive ID to bind with the CPU. + */ + if (!dev->node_props.hive_id) + dev->node_props.hive_id = pci_dev_id(dev->gpu->adev->pdev); peer_dev->node_props.hive_id = dev->node_props.hive_id; + } list_for_each_entry(inbound_link, &peer_dev->io_link_props, list) { @@ -1569,8 +1564,8 @@ static int kfd_dev_create_p2p_links(void) if (dev == new_dev) break; if (!dev->gpu || !dev->gpu->adev || - (dev->gpu->hive_id && - dev->gpu->hive_id == new_dev->gpu->hive_id)) + (dev->gpu->kfd->hive_id && + dev->gpu->kfd->hive_id == new_dev->gpu->kfd->hive_id)) goto next; /* check if node(s) is/are peer accessible in one direction or bi-direction */ @@ -1590,7 +1585,6 @@ out: return ret; } - /* Helper function. See kfd_fill_gpu_cache_info for parameter description */ static int fill_in_l1_pcache(struct kfd_cache_properties **props_ext, struct kfd_gpu_cache_info *pcache_info, @@ -1723,7 +1717,7 @@ static int fill_in_l2_l3_pcache(struct kfd_cache_properties **props_ext, /* kfd_fill_cache_non_crat_info - Fill GPU cache info using kfd_gpu_cache_info * tables */ -static void kfd_fill_cache_non_crat_info(struct kfd_topology_device *dev, struct kfd_dev *kdev) +static void kfd_fill_cache_non_crat_info(struct kfd_topology_device *dev, struct kfd_node *kdev) { struct kfd_gpu_cache_info *pcache_info = NULL; int i, j, k; @@ -1805,7 +1799,7 @@ static void kfd_fill_cache_non_crat_info(struct kfd_topology_device *dev, struct pr_debug("Added [%d] GPU cache entries\n", num_of_entries); } -static int kfd_topology_add_device_locked(struct kfd_dev *gpu, uint32_t gpu_id, +static int kfd_topology_add_device_locked(struct kfd_node *gpu, uint32_t gpu_id, struct kfd_topology_device **dev) { int proximity_domain = ++topology_crat_proximity_domain; @@ -1865,7 +1859,103 @@ err: return res; } -int kfd_topology_add_device(struct kfd_dev *gpu) +static void kfd_topology_set_dbg_firmware_support(struct kfd_topology_device *dev) +{ + bool firmware_supported = true; + + if (KFD_GC_VERSION(dev->gpu) >= IP_VERSION(11, 0, 0) && + KFD_GC_VERSION(dev->gpu) < IP_VERSION(12, 0, 0)) { + uint32_t mes_api_rev = (dev->gpu->adev->mes.sched_version & + AMDGPU_MES_API_VERSION_MASK) >> + AMDGPU_MES_API_VERSION_SHIFT; + uint32_t mes_rev = dev->gpu->adev->mes.sched_version & + AMDGPU_MES_VERSION_MASK; + + firmware_supported = (mes_api_rev >= 14) && (mes_rev >= 64); + goto out; + } + + /* + * Note: Any unlisted devices here are assumed to support exception handling. + * Add additional checks here as needed. + */ + switch (KFD_GC_VERSION(dev->gpu)) { + case IP_VERSION(9, 0, 1): + firmware_supported = dev->gpu->kfd->mec_fw_version >= 459 + 32768; + break; + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 2, 1): + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 3, 0): + case IP_VERSION(9, 4, 0): + firmware_supported = dev->gpu->kfd->mec_fw_version >= 459; + break; + case IP_VERSION(9, 4, 1): + firmware_supported = dev->gpu->kfd->mec_fw_version >= 60; + break; + case IP_VERSION(9, 4, 2): + firmware_supported = dev->gpu->kfd->mec_fw_version >= 51; + break; + case IP_VERSION(10, 1, 10): + case IP_VERSION(10, 1, 2): + case IP_VERSION(10, 1, 1): + firmware_supported = dev->gpu->kfd->mec_fw_version >= 144; + break; + case IP_VERSION(10, 3, 0): + case IP_VERSION(10, 3, 2): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 4): + case IP_VERSION(10, 3, 5): + firmware_supported = dev->gpu->kfd->mec_fw_version >= 89; + break; + case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 3, 3): + firmware_supported = false; + break; + default: + break; + } + +out: + if (firmware_supported) + dev->node_props.capability |= HSA_CAP_TRAP_DEBUG_FIRMWARE_SUPPORTED; +} + +static void kfd_topology_set_capabilities(struct kfd_topology_device *dev) +{ + dev->node_props.capability |= ((HSA_CAP_DOORBELL_TYPE_2_0 << + HSA_CAP_DOORBELL_TYPE_TOTALBITS_SHIFT) & + HSA_CAP_DOORBELL_TYPE_TOTALBITS_MASK); + + dev->node_props.capability |= HSA_CAP_TRAP_DEBUG_SUPPORT | + HSA_CAP_TRAP_DEBUG_WAVE_LAUNCH_TRAP_OVERRIDE_SUPPORTED | + HSA_CAP_TRAP_DEBUG_WAVE_LAUNCH_MODE_SUPPORTED; + + if (KFD_GC_VERSION(dev->gpu) < IP_VERSION(10, 0, 0)) { + dev->node_props.debug_prop |= HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX9 | + HSA_DBG_WATCH_ADDR_MASK_HI_BIT; + + if (KFD_GC_VERSION(dev->gpu) < IP_VERSION(9, 4, 2)) + dev->node_props.debug_prop |= + HSA_DBG_DISPATCH_INFO_ALWAYS_VALID; + else + dev->node_props.capability |= + HSA_CAP_TRAP_DEBUG_PRECISE_MEMORY_OPERATIONS_SUPPORTED; + } else { + dev->node_props.debug_prop |= HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX10 | + HSA_DBG_WATCH_ADDR_MASK_HI_BIT; + + if (KFD_GC_VERSION(dev->gpu) < IP_VERSION(11, 0, 0)) + dev->node_props.debug_prop |= HSA_DBG_DISPATCH_INFO_ALWAYS_VALID; + else + dev->node_props.capability |= + HSA_CAP_TRAP_DEBUG_PRECISE_MEMORY_OPERATIONS_SUPPORTED; + } + + kfd_topology_set_dbg_firmware_support(dev); +} + +int kfd_topology_add_device(struct kfd_node *gpu) { uint32_t gpu_id; struct kfd_topology_device *dev; @@ -1916,28 +2006,37 @@ int kfd_topology_add_device(struct kfd_dev *gpu) dev->node_props.simd_arrays_per_engine = cu_info.num_shader_arrays_per_engine; - dev->node_props.gfx_target_version = gpu->device_info.gfx_target_version; + dev->node_props.gfx_target_version = + gpu->kfd->device_info.gfx_target_version; dev->node_props.vendor_id = gpu->adev->pdev->vendor; dev->node_props.device_id = gpu->adev->pdev->device; dev->node_props.capability |= ((dev->gpu->adev->rev_id << HSA_CAP_ASIC_REVISION_SHIFT) & HSA_CAP_ASIC_REVISION_MASK); + dev->node_props.location_id = pci_dev_id(gpu->adev->pdev); + if (KFD_GC_VERSION(dev->gpu->kfd) == IP_VERSION(9, 4, 3)) + dev->node_props.location_id |= dev->gpu->node_id; + dev->node_props.domain = pci_domain_nr(gpu->adev->pdev->bus); dev->node_props.max_engine_clk_fcompute = amdgpu_amdkfd_get_max_engine_clock_in_mhz(dev->gpu->adev); dev->node_props.max_engine_clk_ccompute = cpufreq_quick_get_max(0) / 1000; - dev->node_props.drm_render_minor = - gpu->shared_resources.drm_render_minor; - dev->node_props.hive_id = gpu->hive_id; + if (gpu->xcp) + dev->node_props.drm_render_minor = gpu->xcp->ddev->render->index; + else + dev->node_props.drm_render_minor = + gpu->kfd->shared_resources.drm_render_minor; + + dev->node_props.hive_id = gpu->kfd->hive_id; dev->node_props.num_sdma_engines = kfd_get_num_sdma_engines(gpu); dev->node_props.num_sdma_xgmi_engines = kfd_get_num_xgmi_sdma_engines(gpu); dev->node_props.num_sdma_queues_per_engine = - gpu->device_info.num_sdma_queues_per_engine - - gpu->device_info.num_reserved_sdma_queues_per_engine; + gpu->kfd->device_info.num_sdma_queues_per_engine - + gpu->kfd->device_info.num_reserved_sdma_queues_per_engine; dev->node_props.num_gws = (dev->gpu->gws && dev->gpu->dqm->sched_policy != KFD_SCHED_POLICY_NO_HWS) ? dev->gpu->adev->gds.gws_size : 0; @@ -1966,20 +2065,18 @@ int kfd_topology_add_device(struct kfd_dev *gpu) HSA_CAP_DOORBELL_TYPE_TOTALBITS_MASK); break; default: - if (KFD_GC_VERSION(dev->gpu) >= IP_VERSION(9, 0, 1)) - dev->node_props.capability |= ((HSA_CAP_DOORBELL_TYPE_2_0 << - HSA_CAP_DOORBELL_TYPE_TOTALBITS_SHIFT) & - HSA_CAP_DOORBELL_TYPE_TOTALBITS_MASK); - else + if (KFD_GC_VERSION(dev->gpu) < IP_VERSION(9, 0, 1)) WARN(1, "Unexpected ASIC family %u", dev->gpu->adev->asic_type); + else + kfd_topology_set_capabilities(dev); } /* * Overwrite ATS capability according to needs_iommu_device to fix * potential missing corresponding bit in CRAT of BIOS. */ - if (dev->gpu->use_iommu_v2) + if (dev->gpu->kfd->use_iommu_v2) dev->node_props.capability |= HSA_CAP_ATS_PRESENT; else dev->node_props.capability &= ~HSA_CAP_ATS_PRESENT; @@ -2007,7 +2104,7 @@ int kfd_topology_add_device(struct kfd_dev *gpu) dev->node_props.capability |= (dev->gpu->adev->ras_enabled != 0) ? HSA_CAP_RASEVENTNOTIFY : 0; - if (KFD_IS_SVM_API_SUPPORTED(dev->gpu->adev->kfd.dev)) + if (KFD_IS_SVM_API_SUPPORTED(dev->gpu->adev)) dev->node_props.capability |= HSA_CAP_SVMAPI_SUPPORTED; kfd_debug_print_topology(); @@ -2079,7 +2176,7 @@ static void kfd_topology_update_io_links(int proximity_domain) } } -int kfd_topology_remove_device(struct kfd_dev *gpu) +int kfd_topology_remove_device(struct kfd_node *gpu) { struct kfd_topology_device *dev, *tmp; uint32_t gpu_id; @@ -2119,7 +2216,7 @@ int kfd_topology_remove_device(struct kfd_dev *gpu) * Return - 0: On success (@kdev will be NULL for non GPU nodes) * -1: If end of list */ -int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_dev **kdev) +int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_node **kdev) { struct kfd_topology_device *top_dev; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.h b/drivers/gpu/drm/amd/amdkfd/kfd_topology.h index fca30d00a9bb..cba2cd5ed9d1 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.h @@ -31,6 +31,11 @@ #define KFD_TOPOLOGY_PUBLIC_NAME_SIZE 32 +#define HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX9 6 +#define HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX10 7 +#define HSA_DBG_WATCH_ADDR_MASK_HI_BIT \ + (29 << HSA_DBG_WATCH_ADDR_MASK_HI_BIT_SHIFT) + struct kfd_node_properties { uint64_t hive_id; uint32_t cpu_cores_count; @@ -42,6 +47,7 @@ struct kfd_node_properties { uint32_t cpu_core_id_base; uint32_t simd_id_base; uint32_t capability; + uint64_t debug_prop; uint32_t max_waves_per_simd; uint32_t lds_size_in_kb; uint32_t gds_size_in_kb; @@ -75,7 +81,7 @@ struct kfd_mem_properties { uint32_t flags; uint32_t width; uint32_t mem_clk_max; - struct kfd_dev *gpu; + struct kfd_node *gpu; struct kobject *kobj; struct attribute attr; }; @@ -93,7 +99,7 @@ struct kfd_cache_properties { uint32_t cache_latency; uint32_t cache_type; uint8_t sibling_map[CACHE_SIBLINGMAP_SIZE]; - struct kfd_dev *gpu; + struct kfd_node *gpu; struct kobject *kobj; struct attribute attr; uint32_t sibling_map_size; @@ -113,7 +119,7 @@ struct kfd_iolink_properties { uint32_t max_bandwidth; uint32_t rec_transfer_size; uint32_t flags; - struct kfd_dev *gpu; + struct kfd_node *gpu; struct kobject *kobj; struct attribute attr; }; @@ -135,7 +141,7 @@ struct kfd_topology_device { struct list_head io_link_props; struct list_head p2p_link_props; struct list_head perf_props; - struct kfd_dev *gpu; + struct kfd_node *gpu; struct kobject *kobj_node; struct kobject *kobj_mem; struct kobject *kobj_cache; diff --git a/drivers/gpu/drm/amd/amdxcp/Makefile b/drivers/gpu/drm/amd/amdxcp/Makefile new file mode 100644 index 000000000000..870501a4bb8c --- /dev/null +++ b/drivers/gpu/drm/amd/amdxcp/Makefile @@ -0,0 +1,25 @@ +# +# Copyright 2023 Advanced Micro Devices, Inc. +# +# Permission is hereby granted, free of charge, to any person obtaining a +# copy of this software and associated documentation files (the "Software"), +# to deal in the Software without restriction, including without limitation +# the rights to use, copy, modify, merge, publish, distribute, sublicense, +# and/or sell copies of the Software, and to permit persons to whom the +# Software is furnished to do so, subject to the following conditions: +# +# The above copyright notice and this permission notice shall be included in +# all copies or substantial portions of the Software. +# +# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL +# THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR +# OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, +# ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR +# OTHER DEALINGS IN THE SOFTWARE. +# + +amdxcp-y := amdgpu_xcp_drv.o + +obj-$(CONFIG_DRM_AMDGPU) += amdxcp.o diff --git a/drivers/gpu/drm/amd/amdxcp/amdgpu_xcp_drv.c b/drivers/gpu/drm/amd/amdxcp/amdgpu_xcp_drv.c new file mode 100644 index 000000000000..353597fc908d --- /dev/null +++ b/drivers/gpu/drm/amd/amdxcp/amdgpu_xcp_drv.c @@ -0,0 +1,110 @@ +/* + * Copyright 2023 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#include <linux/init.h> +#include <linux/module.h> +#include <linux/platform_device.h> + +#include <drm/drm_drv.h> + +#include "amdgpu_xcp_drv.h" + +#define MAX_XCP_PLATFORM_DEVICE 64 + +struct xcp_device { + struct drm_device drm; + struct platform_device *pdev; +}; + +static const struct drm_driver amdgpu_xcp_driver = { + .driver_features = DRIVER_GEM | DRIVER_RENDER, + .name = "amdgpu_xcp_drv", + .major = 1, + .minor = 0, +}; + +static int pdev_num; +static struct xcp_device *xcp_dev[MAX_XCP_PLATFORM_DEVICE]; + +int amdgpu_xcp_drm_dev_alloc(struct drm_device **ddev) +{ + struct platform_device *pdev; + struct xcp_device *pxcp_dev; + int ret; + + if (pdev_num >= MAX_XCP_PLATFORM_DEVICE) + return -ENODEV; + + pdev = platform_device_register_simple("amdgpu_xcp", pdev_num, NULL, 0); + if (IS_ERR(pdev)) + return PTR_ERR(pdev); + + if (!devres_open_group(&pdev->dev, NULL, GFP_KERNEL)) { + ret = -ENOMEM; + goto out_unregister; + } + + pxcp_dev = devm_drm_dev_alloc(&pdev->dev, &amdgpu_xcp_driver, struct xcp_device, drm); + if (IS_ERR(pxcp_dev)) { + ret = PTR_ERR(pxcp_dev); + goto out_devres; + } + + xcp_dev[pdev_num] = pxcp_dev; + xcp_dev[pdev_num]->pdev = pdev; + *ddev = &pxcp_dev->drm; + pdev_num++; + + return 0; + +out_devres: + devres_release_group(&pdev->dev, NULL); +out_unregister: + platform_device_unregister(pdev); + + return ret; +} +EXPORT_SYMBOL(amdgpu_xcp_drm_dev_alloc); + +void amdgpu_xcp_drv_release(void) +{ + for (--pdev_num; pdev_num >= 0; --pdev_num) { + devres_release_group(&xcp_dev[pdev_num]->pdev->dev, NULL); + platform_device_unregister(xcp_dev[pdev_num]->pdev); + xcp_dev[pdev_num]->pdev = NULL; + xcp_dev[pdev_num] = NULL; + } + pdev_num = 0; +} +EXPORT_SYMBOL(amdgpu_xcp_drv_release); + +static void __exit amdgpu_xcp_drv_exit(void) +{ + amdgpu_xcp_drv_release(); +} + +module_exit(amdgpu_xcp_drv_exit); + +MODULE_AUTHOR("AMD linux driver team"); +MODULE_DESCRIPTION("AMD XCP PLATFORM DEVICES"); +MODULE_LICENSE("GPL and additional rights"); diff --git a/drivers/gpu/drm/amd/amdxcp/amdgpu_xcp_drv.h b/drivers/gpu/drm/amd/amdxcp/amdgpu_xcp_drv.h new file mode 100644 index 000000000000..c1c4b679bf95 --- /dev/null +++ b/drivers/gpu/drm/amd/amdxcp/amdgpu_xcp_drv.h @@ -0,0 +1,29 @@ +/* + * Copyright 2023 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef _AMDGPU_XCP_DRV_H_ +#define _AMDGPU_XCP_DRV_H_ + +int amdgpu_xcp_drm_dev_alloc(struct drm_device **ddev); +void amdgpu_xcp_drv_release(void); +#endif /* _AMDGPU_XCP_DRV_H_ */ diff --git a/drivers/gpu/drm/amd/display/Kconfig b/drivers/gpu/drm/amd/display/Kconfig index 2d8e55e29637..bf0a655d009e 100644 --- a/drivers/gpu/drm/amd/display/Kconfig +++ b/drivers/gpu/drm/amd/display/Kconfig @@ -8,7 +8,7 @@ config DRM_AMD_DC depends on BROKEN || !CC_IS_CLANG || X86_64 || SPARC64 || ARM64 select SND_HDA_COMPONENT if SND_HDA_CORE # !CC_IS_CLANG: https://github.com/ClangBuiltLinux/linux/issues/1752 - select DRM_AMD_DC_FP if (X86 || (PPC64 && ALTIVEC) || (ARM64 && KERNEL_MODE_NEON && !CC_IS_CLANG)) + select DRM_AMD_DC_FP if (X86 || LOONGARCH || (PPC64 && ALTIVEC) || (ARM64 && KERNEL_MODE_NEON && !CC_IS_CLANG)) help Choose this option if you want to use the new display engine support for AMDGPU. This adds required support for Vega and @@ -42,16 +42,13 @@ config DEBUG_KERNEL_DC Choose this option if you want to hit kdgb_break in assert. config DRM_AMD_SECURE_DISPLAY - bool "Enable secure display support" - depends on DEBUG_FS - depends on DRM_AMD_DC_FP - help - Choose this option if you want to - support secure display - - This option enables the calculation - of crc of specific region via debugfs. - Cooperate with specific DMCU FW. + bool "Enable secure display support" + depends on DEBUG_FS + depends on DRM_AMD_DC_FP + help + Choose this option if you want to support secure display + This option enables the calculation of crc of specific region via + debugfs. Cooperate with specific DMCU FW. endmenu diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index 8b4b186c57f5..514f6785a020 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -365,6 +365,14 @@ static inline void reverse_planes_order(struct dc_surface_update *array_of_surfa * adjustments and preparation before calling it. This function is a wrapper * for the dc_update_planes_and_stream that does any required configuration * before passing control to DC. + * + * @dc: Display Core control structure + * @update_type: specify whether it is FULL/MEDIUM/FAST update + * @planes_count: planes count to update + * @stream: stream state + * @stream_update: stream update + * @array_of_surface_update: dc surface update pointer + * */ static inline bool update_planes_and_stream_adapter(struct dc *dc, int update_type, @@ -1646,11 +1654,6 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) if (amdgpu_dc_feature_mask & DC_DISABLE_LTTPR_DP2_0) init_data.flags.allow_lttpr_non_transparent_mode.bits.DP2_0 = true; - /* Disable SubVP + DRR config by default */ - init_data.flags.disable_subvp_drr = true; - if (amdgpu_dc_feature_mask & DC_ENABLE_SUBVP_DRR) - init_data.flags.disable_subvp_drr = false; - init_data.flags.seamless_boot_edp_requested = false; if (check_seamless_boot_capability(adev)) { @@ -1672,9 +1675,11 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) adev->dm.dc = dc_create(&init_data); if (adev->dm.dc) { - DRM_INFO("Display Core initialized with v%s!\n", DC_VER); + DRM_INFO("Display Core v%s initialized on %s\n", DC_VER, + dce_version_to_string(adev->dm.dc->ctx->dce_version)); } else { - DRM_INFO("Display Core failed to initialize with v%s!\n", DC_VER); + DRM_INFO("Display Core v%s failed to initialize on %s\n", DC_VER, + dce_version_to_string(adev->dm.dc->ctx->dce_version)); goto error; } @@ -1776,12 +1781,6 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) dc_init_callbacks(adev->dm.dc, &init_params); } -#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY) - adev->dm.secure_display_ctxs = amdgpu_dm_crtc_secure_display_create_contexts(adev); - if (!adev->dm.secure_display_ctxs) { - DRM_ERROR("amdgpu: failed to initialize secure_display_ctxs.\n"); - } -#endif if (dc_is_dmub_outbox_supported(adev->dm.dc)) { init_completion(&adev->dm.dmub_aux_transfer_done); adev->dm.dmub_notify = kzalloc(sizeof(struct dmub_notification), GFP_KERNEL); @@ -1840,6 +1839,11 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) goto error; } +#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY) + adev->dm.secure_display_ctxs = amdgpu_dm_crtc_secure_display_create_contexts(adev); + if (!adev->dm.secure_display_ctxs) + DRM_ERROR("amdgpu: failed to initialize secure display contexts.\n"); +#endif DRM_DEBUG_DRIVER("KMS initialized.\n"); @@ -2479,20 +2483,25 @@ static void dm_gpureset_toggle_interrupts(struct amdgpu_device *adev, if (acrtc && state->stream_status[i].plane_count != 0) { irq_source = IRQ_TYPE_PFLIP + acrtc->otg_inst; rc = dc_interrupt_set(adev->dm.dc, irq_source, enable) ? 0 : -EBUSY; - DRM_DEBUG_VBL("crtc %d - vupdate irq %sabling: r=%d\n", - acrtc->crtc_id, enable ? "en" : "dis", rc); if (rc) DRM_WARN("Failed to %s pflip interrupts\n", enable ? "enable" : "disable"); if (enable) { - rc = amdgpu_dm_crtc_enable_vblank(&acrtc->base); - if (rc) - DRM_WARN("Failed to enable vblank interrupts\n"); - } else { - amdgpu_dm_crtc_disable_vblank(&acrtc->base); - } + if (amdgpu_dm_crtc_vrr_active(to_dm_crtc_state(acrtc->base.state))) + rc = amdgpu_dm_crtc_set_vupdate_irq(&acrtc->base, true); + } else + rc = amdgpu_dm_crtc_set_vupdate_irq(&acrtc->base, false); + + if (rc) + DRM_WARN("Failed to %sable vupdate interrupt\n", enable ? "en" : "dis"); + irq_source = IRQ_TYPE_VBLANK + acrtc->otg_inst; + /* During gpu-reset we disable and then enable vblank irq, so + * don't use amdgpu_irq_get/put() to avoid refcount change. + */ + if (!dc_interrupt_set(adev->dm.dc, irq_source, enable)) + DRM_WARN("Failed to %sable vblank interrupt\n", enable ? "en" : "dis"); } } @@ -2852,7 +2861,7 @@ static int dm_resume(void *handle) * this is the case when traversing through already created * MST connectors, should be skipped */ - if (aconnector->dc_link->type == dc_connection_mst_branch) + if (aconnector && aconnector->mst_root) continue; mutex_lock(&aconnector->hpd_lock); @@ -3258,6 +3267,7 @@ static void dm_handle_mst_sideband_msg(struct amdgpu_dm_connector *aconnector) while (dret == dpcd_bytes_to_read && process_count < max_process_count) { + u8 ack[DP_PSR_ERROR_STATUS - DP_SINK_COUNT_ESI] = {}; u8 retry; dret = 0; @@ -3266,28 +3276,29 @@ static void dm_handle_mst_sideband_msg(struct amdgpu_dm_connector *aconnector) DRM_DEBUG_DRIVER("ESI %02x %02x %02x\n", esi[0], esi[1], esi[2]); /* handle HPD short pulse irq */ if (aconnector->mst_mgr.mst_state) - drm_dp_mst_hpd_irq( - &aconnector->mst_mgr, - esi, - &new_irq_handled); + drm_dp_mst_hpd_irq_handle_event(&aconnector->mst_mgr, + esi, + ack, + &new_irq_handled); if (new_irq_handled) { /* ACK at DPCD to notify down stream */ - const int ack_dpcd_bytes_to_write = - dpcd_bytes_to_read - 1; - for (retry = 0; retry < 3; retry++) { - u8 wret; - - wret = drm_dp_dpcd_write( - &aconnector->dm_dp_aux.aux, - dpcd_addr + 1, - &esi[1], - ack_dpcd_bytes_to_write); - if (wret == ack_dpcd_bytes_to_write) + ssize_t wret; + + wret = drm_dp_dpcd_writeb(&aconnector->dm_dp_aux.aux, + dpcd_addr + 1, + ack[1]); + if (wret == 1) break; } + if (retry == 3) { + DRM_ERROR("Failed to ack MST event.\n"); + return; + } + + drm_dp_mst_hpd_irq_send_new_request(&aconnector->mst_mgr); /* check if there is new irq to be handled */ dret = drm_dp_dpcd_read( &aconnector->dm_dp_aux.aux, @@ -5326,21 +5337,44 @@ get_aspect_ratio(const struct drm_display_mode *mode_in) } static enum dc_color_space -get_output_color_space(const struct dc_crtc_timing *dc_crtc_timing) +get_output_color_space(const struct dc_crtc_timing *dc_crtc_timing, + const struct drm_connector_state *connector_state) { enum dc_color_space color_space = COLOR_SPACE_SRGB; - switch (dc_crtc_timing->pixel_encoding) { - case PIXEL_ENCODING_YCBCR422: - case PIXEL_ENCODING_YCBCR444: - case PIXEL_ENCODING_YCBCR420: - { + switch (connector_state->colorspace) { + case DRM_MODE_COLORIMETRY_BT601_YCC: + if (dc_crtc_timing->flags.Y_ONLY) + color_space = COLOR_SPACE_YCBCR601_LIMITED; + else + color_space = COLOR_SPACE_YCBCR601; + break; + case DRM_MODE_COLORIMETRY_BT709_YCC: + if (dc_crtc_timing->flags.Y_ONLY) + color_space = COLOR_SPACE_YCBCR709_LIMITED; + else + color_space = COLOR_SPACE_YCBCR709; + break; + case DRM_MODE_COLORIMETRY_OPRGB: + color_space = COLOR_SPACE_ADOBERGB; + break; + case DRM_MODE_COLORIMETRY_BT2020_RGB: + case DRM_MODE_COLORIMETRY_BT2020_YCC: + if (dc_crtc_timing->pixel_encoding == PIXEL_ENCODING_RGB) + color_space = COLOR_SPACE_2020_RGB_FULLRANGE; + else + color_space = COLOR_SPACE_2020_YCBCR; + break; + case DRM_MODE_COLORIMETRY_DEFAULT: // ITU601 + default: + if (dc_crtc_timing->pixel_encoding == PIXEL_ENCODING_RGB) { + color_space = COLOR_SPACE_SRGB; /* * 27030khz is the separation point between HDTV and SDTV * according to HDMI spec, we use YCbCr709 and YCbCr601 * respectively */ - if (dc_crtc_timing->pix_clk_100hz > 270300) { + } else if (dc_crtc_timing->pix_clk_100hz > 270300) { if (dc_crtc_timing->flags.Y_ONLY) color_space = COLOR_SPACE_YCBCR709_LIMITED; @@ -5353,15 +5387,6 @@ get_output_color_space(const struct dc_crtc_timing *dc_crtc_timing) else color_space = COLOR_SPACE_YCBCR601; } - - } - break; - case PIXEL_ENCODING_RGB: - color_space = COLOR_SPACE_SRGB; - break; - - default: - WARN_ON(1); break; } @@ -5500,7 +5525,7 @@ static void fill_stream_properties_from_drm_display_mode( } } - stream->output_color_space = get_output_color_space(timing_out); + stream->output_color_space = get_output_color_space(timing_out, connector_state); } static void fill_audio_info(struct audio_info *audio_info, @@ -5942,15 +5967,14 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector, { struct drm_display_mode *preferred_mode = NULL; struct drm_connector *drm_connector; - const struct drm_connector_state *con_state = - dm_state ? &dm_state->base : NULL; + const struct drm_connector_state *con_state = &dm_state->base; struct dc_stream_state *stream = NULL; struct drm_display_mode mode; struct drm_display_mode saved_mode; struct drm_display_mode *freesync_mode = NULL; bool native_mode_found = false; bool recalculate_timing = false; - bool scale = dm_state ? (dm_state->scaling != RMX_OFF) : false; + bool scale = dm_state->scaling != RMX_OFF; int mode_refresh; int preferred_refresh = 0; enum color_transfer_func tf = TRANSFER_FUNC_UNKNOWN; @@ -6013,8 +6037,7 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector, */ DRM_DEBUG_DRIVER("No preferred mode found\n"); } else { - recalculate_timing = amdgpu_freesync_vid_mode && - is_freesync_video_mode(&mode, aconnector); + recalculate_timing = is_freesync_video_mode(&mode, aconnector); if (recalculate_timing) { freesync_mode = get_highest_refresh_rate_mode(aconnector, false); drm_mode_copy(&saved_mode, &mode); @@ -6029,7 +6052,7 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector, if (recalculate_timing) drm_mode_set_crtcinfo(&saved_mode, 0); - else if (!dm_state) + else drm_mode_set_crtcinfo(&mode, 0); /* @@ -6342,6 +6365,31 @@ amdgpu_dm_connector_late_register(struct drm_connector *connector) return 0; } +static void amdgpu_dm_connector_funcs_force(struct drm_connector *connector) +{ + struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector); + struct dc_link *dc_link = aconnector->dc_link; + struct dc_sink *dc_em_sink = aconnector->dc_em_sink; + struct edid *edid; + + if (!connector->edid_override) + return; + + drm_edid_override_connector_update(&aconnector->base); + edid = aconnector->base.edid_blob_ptr->data; + aconnector->edid = edid; + + /* Update emulated (virtual) sink's EDID */ + if (dc_em_sink && dc_link) { + memset(&dc_em_sink->edid_caps, 0, sizeof(struct dc_edid_caps)); + memmove(dc_em_sink->dc_edid.raw_edid, edid, (edid->extensions + 1) * EDID_LENGTH); + dm_helpers_parse_edid_caps( + dc_link, + &dc_em_sink->dc_edid, + &dc_em_sink->edid_caps); + } +} + static const struct drm_connector_funcs amdgpu_dm_connector_funcs = { .reset = amdgpu_dm_connector_funcs_reset, .detect = amdgpu_dm_connector_detect, @@ -6352,7 +6400,8 @@ static const struct drm_connector_funcs amdgpu_dm_connector_funcs = { .atomic_set_property = amdgpu_dm_connector_atomic_set_property, .atomic_get_property = amdgpu_dm_connector_atomic_get_property, .late_register = amdgpu_dm_connector_late_register, - .early_unregister = amdgpu_dm_connector_unregister + .early_unregister = amdgpu_dm_connector_unregister, + .force = amdgpu_dm_connector_funcs_force }; static int get_modes(struct drm_connector *connector) @@ -6369,11 +6418,19 @@ static void create_eml_sink(struct amdgpu_dm_connector *aconnector) struct edid *edid; if (!aconnector->base.edid_blob_ptr) { - DRM_ERROR("No EDID firmware found on connector: %s ,forcing to OFF!\n", - aconnector->base.name); + /* if connector->edid_override valid, pass + * it to edid_override to edid_blob_ptr + */ - aconnector->base.force = DRM_FORCE_OFF; - return; + drm_edid_override_connector_update(&aconnector->base); + + if (!aconnector->base.edid_blob_ptr) { + DRM_ERROR("No EDID firmware found on connector: %s ,forcing to OFF!\n", + aconnector->base.name); + + aconnector->base.force = DRM_FORCE_OFF; + return; + } } edid = (struct edid *) aconnector->base.edid_blob_ptr->data; @@ -6558,7 +6615,9 @@ enum drm_mode_status amdgpu_dm_connector_mode_valid(struct drm_connector *connec goto fail; } - stream = create_validate_stream_for_sink(aconnector, mode, NULL, NULL); + stream = create_validate_stream_for_sink(aconnector, mode, + to_dm_connector_state(connector->state), + NULL); if (stream) { dc_stream_release(stream); result = MODE_OK; @@ -6652,6 +6711,14 @@ amdgpu_dm_connector_atomic_check(struct drm_connector *conn, if (!crtc) return 0; + if (new_con_state->colorspace != old_con_state->colorspace) { + new_crtc_state = drm_atomic_get_crtc_state(state, crtc); + if (IS_ERR(new_crtc_state)) + return PTR_ERR(new_crtc_state); + + new_crtc_state->mode_changed = true; + } + if (!drm_connector_atomic_hdr_metadata_equal(old_con_state, new_con_state)) { struct dc_info_packet hdr_infopacket; @@ -6674,7 +6741,7 @@ amdgpu_dm_connector_atomic_check(struct drm_connector *conn, * set is permissible, however. So only force a * modeset if we're entering or exiting HDR. */ - new_crtc_state->mode_changed = + new_crtc_state->mode_changed = new_crtc_state->mode_changed || !old_con_state->hdr_output_metadata || !new_con_state->hdr_output_metadata; } @@ -6737,7 +6804,7 @@ static int dm_encoder_helper_atomic_check(struct drm_encoder *encoder, int clock, bpp = 0; bool is_y420 = false; - if (!aconnector->mst_output_port || !aconnector->dc_sink) + if (!aconnector->mst_output_port) return 0; mst_port = aconnector->mst_output_port; @@ -7163,7 +7230,7 @@ static void amdgpu_dm_connector_add_freesync_modes(struct drm_connector *connect struct amdgpu_dm_connector *amdgpu_dm_connector = to_amdgpu_dm_connector(connector); - if (!(amdgpu_freesync_vid_mode && edid)) + if (!edid) return; if (amdgpu_dm_connector->max_vfreq - amdgpu_dm_connector->min_vfreq > 10) @@ -7191,7 +7258,13 @@ static int amdgpu_dm_connector_get_modes(struct drm_connector *connector) drm_add_modes_noedid(connector, 1920, 1080); } else { amdgpu_dm_connector_ddc_get_modes(connector, edid); - amdgpu_dm_connector_add_common_modes(encoder, connector); + /* most eDP supports only timings from its edid, + * usually only detailed timings are available + * from eDP edid. timings which are not from edid + * may damage eDP + */ + if (connector->connector_type != DRM_MODE_CONNECTOR_eDP) + amdgpu_dm_connector_add_common_modes(encoder, connector); amdgpu_dm_connector_add_freesync_modes(connector, edid); } amdgpu_dm_fbc_init(connector); @@ -7199,6 +7272,12 @@ static int amdgpu_dm_connector_get_modes(struct drm_connector *connector) return amdgpu_dm_connector->num_modes; } +static const u32 supported_colorspaces = + BIT(DRM_MODE_COLORIMETRY_BT709_YCC) | + BIT(DRM_MODE_COLORIMETRY_OPRGB) | + BIT(DRM_MODE_COLORIMETRY_BT2020_RGB) | + BIT(DRM_MODE_COLORIMETRY_BT2020_YCC); + void amdgpu_dm_connector_init_helper(struct amdgpu_display_manager *dm, struct amdgpu_dm_connector *aconnector, int connector_type, @@ -7279,6 +7358,15 @@ void amdgpu_dm_connector_init_helper(struct amdgpu_display_manager *dm, adev->mode_info.abm_level_property, 0); } + if (connector_type == DRM_MODE_CONNECTOR_HDMIA) { + if (!drm_mode_create_hdmi_colorspace_property(&aconnector->base, supported_colorspaces)) + drm_connector_attach_colorspace_property(&aconnector->base); + } else if (connector_type == DRM_MODE_CONNECTOR_DisplayPort || + connector_type == DRM_MODE_CONNECTOR_eDP) { + if (!drm_mode_create_dp_colorspace_property(&aconnector->base, supported_colorspaces)) + drm_connector_attach_colorspace_property(&aconnector->base); + } + if (connector_type == DRM_MODE_CONNECTOR_HDMIA || connector_type == DRM_MODE_CONNECTOR_DisplayPort || connector_type == DRM_MODE_CONNECTOR_eDP) { @@ -8193,6 +8281,12 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state, if (acrtc_state->abm_level != dm_old_crtc_state->abm_level) bundle->stream_update.abm_level = &acrtc_state->abm_level; + mutex_lock(&dm->dc_lock); + if ((acrtc_state->update_type > UPDATE_TYPE_FAST) && + acrtc_state->stream->link->psr_settings.psr_allow_active) + amdgpu_dm_psr_disable(acrtc_state->stream); + mutex_unlock(&dm->dc_lock); + /* * If FreeSync state on the stream has changed then we need to * re-adjust the min/max bounds now that DC doesn't handle this @@ -8206,10 +8300,6 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state, spin_unlock_irqrestore(&pcrtc->dev->event_lock, flags); } mutex_lock(&dm->dc_lock); - if ((acrtc_state->update_type > UPDATE_TYPE_FAST) && - acrtc_state->stream->link->psr_settings.psr_allow_active) - amdgpu_dm_psr_disable(acrtc_state->stream); - update_planes_and_stream_adapter(dm->dc, acrtc_state->update_type, planes_count, @@ -8871,10 +8961,17 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) drm_atomic_helper_cleanup_planes(dev, state); - /* return the stolen vga memory back to VRAM */ - if (!adev->mman.keep_stolen_vga_memory) - amdgpu_bo_free_kernel(&adev->mman.stolen_vga_memory, NULL, NULL); - amdgpu_bo_free_kernel(&adev->mman.stolen_extended_memory, NULL, NULL); + /* Don't free the memory if we are hitting this as part of suspend. + * This way we don't free any memory during suspend; see + * amdgpu_bo_free_kernel(). The memory will be freed in the first + * non-suspend modeset or when the driver is torn down. + */ + if (!adev->in_suspend) { + /* return the stolen vga memory back to VRAM */ + if (!adev->mman.keep_stolen_vga_memory) + amdgpu_bo_free_kernel(&adev->mman.stolen_vga_memory, NULL, NULL); + amdgpu_bo_free_kernel(&adev->mman.stolen_extended_memory, NULL, NULL); + } /* * Finally, drop a runtime PM reference for each newly disabled CRTC, @@ -9208,8 +9305,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm, * TODO: Refactor this function to allow this check to work * in all conditions. */ - if (amdgpu_freesync_vid_mode && - dm_new_crtc_state->stream && + if (dm_new_crtc_state->stream && is_timing_unchanged_for_freesync(new_crtc_state, old_crtc_state)) goto skip_modeset; @@ -9251,7 +9347,9 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm, } /* Now check if we should set freesync video mode */ - if (amdgpu_freesync_vid_mode && dm_new_crtc_state->stream && + if (dm_new_crtc_state->stream && + dc_is_stream_unchanged(new_stream, dm_old_crtc_state->stream) && + dc_is_stream_scaling_unchanged(new_stream, dm_old_crtc_state->stream) && is_timing_unchanged_for_freesync(new_crtc_state, old_crtc_state)) { new_crtc_state->mode_changed = false; @@ -9263,7 +9361,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm, set_freesync_fixed_config(dm_new_crtc_state); goto skip_modeset; - } else if (amdgpu_freesync_vid_mode && aconnector && + } else if (aconnector && is_freesync_video_mode(&new_crtc_state->mode, aconnector)) { struct drm_display_mode *high_mode; @@ -10323,7 +10421,7 @@ static bool dm_edid_parser_send_cea(struct amdgpu_display_manager *dm, input->cea_total_length = total_length; memcpy(input->payload, data, length); - res = dc_dmub_srv_cmd_with_reply_data(dm->dc->ctx->dmub_srv, &cmd); + res = dm_execute_dmub_cmd(dm->dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY); if (!res) { DRM_ERROR("EDID CEA parser failed\n"); return false; @@ -10773,3 +10871,13 @@ bool check_seamless_boot_capability(struct amdgpu_device *adev) return false; } + +bool dm_execute_dmub_cmd(const struct dc_context *ctx, union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type) +{ + return dc_dmub_srv_cmd_run(ctx->dmub_srv, cmd, wait_type); +} + +bool dm_execute_dmub_cmd_list(const struct dc_context *ctx, unsigned int count, union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type) +{ + return dc_dmub_srv_cmd_run_list(ctx->dmub_srv, count, cmd, wait_type); +} diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h index 2e2413fd73a4..4561f55afa99 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h @@ -661,10 +661,6 @@ struct amdgpu_dm_connector { struct mutex hpd_lock; bool fake_enable; -#ifdef CONFIG_DEBUG_FS - uint32_t debugfs_dpcd_address; - uint32_t debugfs_dpcd_size; -#endif bool force_yuv420_output; struct dsc_preferred_settings dsc_settings; union dp_downstream_port_present mst_downstream_port_present; diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c index 27711743c22c..0802f8e8fac5 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c @@ -83,12 +83,15 @@ const char *const *amdgpu_dm_crtc_get_crc_sources(struct drm_crtc *crtc, } #ifdef CONFIG_DRM_AMD_SECURE_DISPLAY -static void amdgpu_dm_set_crc_window_default(struct drm_crtc *crtc) +static void amdgpu_dm_set_crc_window_default(struct drm_crtc *crtc, struct dc_stream_state *stream) { struct drm_device *drm_dev = crtc->dev; + struct amdgpu_display_manager *dm = &drm_to_adev(drm_dev)->dm; struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc); + bool was_activated; spin_lock_irq(&drm_dev->event_lock); + was_activated = acrtc->dm_irq_params.window_param.activated; acrtc->dm_irq_params.window_param.x_start = 0; acrtc->dm_irq_params.window_param.y_start = 0; acrtc->dm_irq_params.window_param.x_end = 0; @@ -97,6 +100,14 @@ static void amdgpu_dm_set_crc_window_default(struct drm_crtc *crtc) acrtc->dm_irq_params.window_param.update_win = false; acrtc->dm_irq_params.window_param.skip_frame_cnt = 0; spin_unlock_irq(&drm_dev->event_lock); + + /* Disable secure_display if it was enabled */ + if (was_activated) { + /* stop ROI update on this crtc */ + flush_work(&dm->secure_display_ctxs[crtc->index].notify_ta_work); + flush_work(&dm->secure_display_ctxs[crtc->index].forward_roi_work); + dc_stream_forward_crc_window(stream, NULL, true); + } } static void amdgpu_dm_crtc_notify_ta_to_read(struct work_struct *work) @@ -204,9 +215,6 @@ int amdgpu_dm_crtc_configure_crc_source(struct drm_crtc *crtc, struct dm_crtc_state *dm_crtc_state, enum amdgpu_dm_pipe_crc_source source) { -#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY) - int i; -#endif struct amdgpu_device *adev = drm_to_adev(crtc->dev); struct dc_stream_state *stream_state = dm_crtc_state->stream; bool enable = amdgpu_dm_is_valid_crc_source(source); @@ -220,19 +228,6 @@ int amdgpu_dm_crtc_configure_crc_source(struct drm_crtc *crtc, /* Enable or disable CRTC CRC generation */ if (dm_is_crc_source_crtc(source) || source == AMDGPU_DM_PIPE_CRC_SOURCE_NONE) { -#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY) - /* Disable secure_display if it was enabled */ - if (!enable) { - for (i = 0; i < adev->mode_info.num_crtc; i++) { - if (adev->dm.secure_display_ctxs[i].crtc == crtc) { - /* stop ROI update on this crtc */ - flush_work(&adev->dm.secure_display_ctxs[i].notify_ta_work); - flush_work(&adev->dm.secure_display_ctxs[i].forward_roi_work); - dc_stream_forward_crc_window(stream_state, NULL, true); - } - } - } -#endif if (!dc_stream_configure_crc(stream_state->ctx->dc, stream_state, NULL, enable, enable)) { ret = -EINVAL; @@ -363,7 +358,7 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name) #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY) /* Reset secure_display when we change crc source from debugfs */ - amdgpu_dm_set_crc_window_default(crtc); + amdgpu_dm_set_crc_window_default(crtc, crtc_state->stream); #endif if (amdgpu_dm_crtc_configure_crc_source(crtc, crtc_state, source)) { diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h index 935adca6f048..748e80ef40d0 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h @@ -100,7 +100,7 @@ struct secure_display_context *amdgpu_dm_crtc_secure_display_create_contexts( #else #define amdgpu_dm_crc_window_is_activated(x) #define amdgpu_dm_crtc_handle_crc_window_irq(x) -#define amdgpu_dm_crtc_secure_display_create_contexts() +#define amdgpu_dm_crtc_secure_display_create_contexts(x) #endif #endif /* AMD_DAL_DEV_AMDGPU_DM_AMDGPU_DM_CRC_H_ */ diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c index e3762e806617..440fc0869a34 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c @@ -146,7 +146,6 @@ static void vblank_control_worker(struct work_struct *work) static inline int dm_set_vblank(struct drm_crtc *crtc, bool enable) { - enum dc_irq_source irq_source; struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc); struct amdgpu_device *adev = drm_to_adev(crtc->dev); struct dm_crtc_state *acrtc_state = to_dm_crtc_state(crtc->state); @@ -169,18 +168,9 @@ static inline int dm_set_vblank(struct drm_crtc *crtc, bool enable) if (rc) return rc; - if (amdgpu_in_reset(adev)) { - irq_source = IRQ_TYPE_VBLANK + acrtc->otg_inst; - /* During gpu-reset we disable and then enable vblank irq, so - * don't use amdgpu_irq_get/put() to avoid refcount change. - */ - if (!dc_interrupt_set(adev->dm.dc, irq_source, enable)) - rc = -EBUSY; - } else { - rc = (enable) - ? amdgpu_irq_get(adev, &adev->crtc_irq, acrtc->crtc_id) - : amdgpu_irq_put(adev, &adev->crtc_irq, acrtc->crtc_id); - } + rc = (enable) + ? amdgpu_irq_get(adev, &adev->crtc_irq, acrtc->crtc_id) + : amdgpu_irq_put(adev, &adev->crtc_irq, acrtc->crtc_id); if (rc) return rc; diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c index 827fcb4fb3b3..5ea3284b2b77 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c @@ -907,6 +907,61 @@ unlock: DEFINE_SHOW_ATTRIBUTE(amdgpu_current_bpc); /* + * Returns the current colorspace for the crtc. + * Example usage: cat /sys/kernel/debug/dri/0/crtc-0/amdgpu_current_colorspace + */ +static int amdgpu_current_colorspace_show(struct seq_file *m, void *data) +{ + struct drm_crtc *crtc = m->private; + struct drm_device *dev = crtc->dev; + struct dm_crtc_state *dm_crtc_state = NULL; + int res = -ENODEV; + + mutex_lock(&dev->mode_config.mutex); + drm_modeset_lock(&crtc->mutex, NULL); + if (crtc->state == NULL) + goto unlock; + + dm_crtc_state = to_dm_crtc_state(crtc->state); + if (dm_crtc_state->stream == NULL) + goto unlock; + + switch (dm_crtc_state->stream->output_color_space) { + case COLOR_SPACE_SRGB: + seq_printf(m, "sRGB"); + break; + case COLOR_SPACE_YCBCR601: + case COLOR_SPACE_YCBCR601_LIMITED: + seq_printf(m, "BT601_YCC"); + break; + case COLOR_SPACE_YCBCR709: + case COLOR_SPACE_YCBCR709_LIMITED: + seq_printf(m, "BT709_YCC"); + break; + case COLOR_SPACE_ADOBERGB: + seq_printf(m, "opRGB"); + break; + case COLOR_SPACE_2020_RGB_FULLRANGE: + seq_printf(m, "BT2020_RGB"); + break; + case COLOR_SPACE_2020_YCBCR: + seq_printf(m, "BT2020_YCC"); + break; + default: + goto unlock; + } + res = 0; + +unlock: + drm_modeset_unlock(&crtc->mutex); + mutex_unlock(&dev->mode_config.mutex); + + return res; +} +DEFINE_SHOW_ATTRIBUTE(amdgpu_current_colorspace); + + +/* * Example usage: * Disable dsc passthrough, i.e.,: have dsc decoding at converver, not external RX * echo 1 /sys/kernel/debug/dri/0/DP-1/dsc_disable_passthrough @@ -1039,88 +1094,6 @@ static ssize_t dp_sdp_message_debugfs_write(struct file *f, const char __user *b return write_size; } -static ssize_t dp_dpcd_address_write(struct file *f, const char __user *buf, - size_t size, loff_t *pos) -{ - int r; - struct amdgpu_dm_connector *connector = file_inode(f)->i_private; - - if (size < sizeof(connector->debugfs_dpcd_address)) - return -EINVAL; - - r = copy_from_user(&connector->debugfs_dpcd_address, - buf, sizeof(connector->debugfs_dpcd_address)); - - return size - r; -} - -static ssize_t dp_dpcd_size_write(struct file *f, const char __user *buf, - size_t size, loff_t *pos) -{ - int r; - struct amdgpu_dm_connector *connector = file_inode(f)->i_private; - - if (size < sizeof(connector->debugfs_dpcd_size)) - return -EINVAL; - - r = copy_from_user(&connector->debugfs_dpcd_size, - buf, sizeof(connector->debugfs_dpcd_size)); - - if (connector->debugfs_dpcd_size > 256) - connector->debugfs_dpcd_size = 0; - - return size - r; -} - -static ssize_t dp_dpcd_data_write(struct file *f, const char __user *buf, - size_t size, loff_t *pos) -{ - int r; - char *data; - struct amdgpu_dm_connector *connector = file_inode(f)->i_private; - struct dc_link *link = connector->dc_link; - uint32_t write_size = connector->debugfs_dpcd_size; - - if (!write_size || size < write_size) - return -EINVAL; - - data = kzalloc(write_size, GFP_KERNEL); - if (!data) - return 0; - - r = copy_from_user(data, buf, write_size); - - dm_helpers_dp_write_dpcd(link->ctx, link, - connector->debugfs_dpcd_address, data, write_size - r); - kfree(data); - return write_size - r; -} - -static ssize_t dp_dpcd_data_read(struct file *f, char __user *buf, - size_t size, loff_t *pos) -{ - int r; - char *data; - struct amdgpu_dm_connector *connector = file_inode(f)->i_private; - struct dc_link *link = connector->dc_link; - uint32_t read_size = connector->debugfs_dpcd_size; - - if (!read_size || size < read_size) - return 0; - - data = kzalloc(read_size, GFP_KERNEL); - if (!data) - return 0; - - dm_helpers_dp_read_dpcd(link->ctx, link, - connector->debugfs_dpcd_address, data, read_size); - - r = copy_to_user(buf, data, read_size); - - kfree(data); - return read_size - r; -} - /* function: Read link's DSC & FEC capabilities * * @@ -2682,25 +2655,6 @@ static const struct file_operations sdp_message_fops = { .llseek = default_llseek }; -static const struct file_operations dp_dpcd_address_debugfs_fops = { - .owner = THIS_MODULE, - .write = dp_dpcd_address_write, - .llseek = default_llseek -}; - -static const struct file_operations dp_dpcd_size_debugfs_fops = { - .owner = THIS_MODULE, - .write = dp_dpcd_size_write, - .llseek = default_llseek -}; - -static const struct file_operations dp_dpcd_data_debugfs_fops = { - .owner = THIS_MODULE, - .read = dp_dpcd_data_read, - .write = dp_dpcd_data_write, - .llseek = default_llseek -}; - static const struct file_operations dp_max_bpc_debugfs_fops = { .owner = THIS_MODULE, .read = dp_max_bpc_read, @@ -2724,9 +2678,6 @@ static const struct { {"test_pattern", &dp_phy_test_pattern_fops}, {"hdcp_sink_capability", &hdcp_sink_capability_fops}, {"sdp_message", &sdp_message_fops}, - {"aux_dpcd_address", &dp_dpcd_address_debugfs_fops}, - {"aux_dpcd_size", &dp_dpcd_size_debugfs_fops}, - {"aux_dpcd_data", &dp_dpcd_data_debugfs_fops}, {"dsc_clock_en", &dp_dsc_clock_en_debugfs_fops}, {"dsc_slice_width", &dp_dsc_slice_width_debugfs_fops}, {"dsc_slice_height", &dp_dsc_slice_height_debugfs_fops}, @@ -2809,6 +2760,32 @@ static int psr_read_residency(void *data, u64 *val) return 0; } +/* read allow_edp_hotplug_detection */ +static int allow_edp_hotplug_detection_get(void *data, u64 *val) +{ + struct amdgpu_dm_connector *aconnector = data; + struct drm_connector *connector = &aconnector->base; + struct drm_device *dev = connector->dev; + struct amdgpu_device *adev = drm_to_adev(dev); + + *val = adev->dm.dc->config.allow_edp_hotplug_detection; + + return 0; +} + +/* set allow_edp_hotplug_detection */ +static int allow_edp_hotplug_detection_set(void *data, u64 val) +{ + struct amdgpu_dm_connector *aconnector = data; + struct drm_connector *connector = &aconnector->base; + struct drm_device *dev = connector->dev; + struct amdgpu_device *adev = drm_to_adev(dev); + + adev->dm.dc->config.allow_edp_hotplug_detection = (uint32_t) val; + + return 0; +} + /* * Set dmcub trace event IRQ enable or disable. * Usage to enable dmcub trace event IRQ: echo 1 > /sys/kernel/debug/dri/0/amdgpu_dm_dmcub_trace_event_en @@ -2847,6 +2824,10 @@ DEFINE_DEBUGFS_ATTRIBUTE(psr_fops, psr_get, NULL, "%llu\n"); DEFINE_DEBUGFS_ATTRIBUTE(psr_residency_fops, psr_read_residency, NULL, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(allow_edp_hotplug_detection_fops, + allow_edp_hotplug_detection_get, + allow_edp_hotplug_detection_set, "%llu\n"); + DEFINE_SHOW_ATTRIBUTE(current_backlight); DEFINE_SHOW_ATTRIBUTE(target_backlight); @@ -3017,6 +2998,8 @@ void connector_debugfs_init(struct amdgpu_dm_connector *connector) &target_backlight_fops); debugfs_create_file("ilr_setting", 0644, dir, connector, &edp_ilr_debugfs_fops); + debugfs_create_file("allow_edp_hotplug_detection", 0644, dir, connector, + &allow_edp_hotplug_detection_fops); } for (i = 0; i < ARRAY_SIZE(connector_debugfs_entries); i++) { @@ -3025,9 +3008,6 @@ void connector_debugfs_init(struct amdgpu_dm_connector *connector) connector_debugfs_entries[i].fops); } - connector->debugfs_dpcd_address = 0; - connector->debugfs_dpcd_size = 0; - if (connector->base.connector_type == DRM_MODE_CONNECTOR_HDMIA) { for (i = 0; i < ARRAY_SIZE(hdmi_debugfs_entries); i++) { debugfs_create_file(hdmi_debugfs_entries[i].name, @@ -3246,6 +3226,8 @@ void crtc_debugfs_init(struct drm_crtc *crtc) #endif debugfs_create_file("amdgpu_current_bpc", 0644, crtc->debugfs_entry, crtc, &amdgpu_current_bpc_fops); + debugfs_create_file("amdgpu_current_colorspace", 0644, crtc->debugfs_entry, + crtc, &amdgpu_current_colorspace_fops); } /* diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c index c6ce2b7123b7..cd20cfc04996 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c @@ -44,9 +44,6 @@ #include "dm_helpers.h" #include "ddc_service_types.h" -/* MST Dock */ -static const uint8_t SYNAPTICS_DEVICE_ID[] = "SYNA"; - /* dm_helpers_parse_edid_caps * * Parse edid caps @@ -702,6 +699,9 @@ static void apply_synaptics_fifo_reset_wa(struct drm_dp_aux *aux) DC_LOG_DC("Done apply_synaptics_fifo_reset_wa\n"); } +/* MST Dock */ +static const uint8_t SYNAPTICS_DEVICE_ID[] = "SYNA"; + static uint8_t write_dsc_enable_synaptics_non_virtual_dpcd_mst( struct drm_dp_aux *aux, const struct dc_stream_state *stream, @@ -885,10 +885,34 @@ enum dc_edid_status dm_helpers_read_local_edid( DRM_ERROR("EDID err: %d, on connector: %s", edid_status, aconnector->base.name); + if (link->aux_mode) { + union test_request test_request = {0}; + union test_response test_response = {0}; - /* DP Compliance Test 4.2.2.3 */ - if (link->aux_mode) - drm_dp_send_real_edid_checksum(&aconnector->dm_dp_aux.aux, sink->dc_edid.raw_edid[sink->dc_edid.length-1]); + dm_helpers_dp_read_dpcd(ctx, + link, + DP_TEST_REQUEST, + &test_request.raw, + sizeof(union test_request)); + + if (!test_request.bits.EDID_READ) + return edid_status; + + test_response.bits.EDID_CHECKSUM_WRITE = 1; + + dm_helpers_dp_write_dpcd(ctx, + link, + DP_TEST_EDID_CHECKSUM, + &sink->dc_edid.raw_edid[sink->dc_edid.length-1], + 1); + + dm_helpers_dp_write_dpcd(ctx, + link, + DP_TEST_RESPONSE, + &test_response.raw, + sizeof(test_response)); + + } return edid_status; } diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c index 810ab682f424..46d0a8f57e55 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c @@ -45,8 +45,7 @@ #endif #include "dc/dcn20/dcn20_resource.h" -bool is_timing_changed(struct dc_stream_state *cur_stream, - struct dc_stream_state *new_stream); + #define PEAK_FACTOR_X1000 1006 static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux, @@ -1422,7 +1421,7 @@ int pre_validate_dsc(struct drm_atomic_state *state, struct dc_stream_state *stream = dm_state->context->streams[i]; if (local_dc_state->streams[i] && - is_timing_changed(stream, local_dc_state->streams[i])) { + dc_is_timing_changed(stream, local_dc_state->streams[i])) { DRM_INFO_ONCE("crtc[%d] needs mode_changed\n", i); } else { int ind = find_crtc_index_in_state_by_stream(state, stream); diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/dc_fpu.c b/drivers/gpu/drm/amd/display/amdgpu_dm/dc_fpu.c index c42aa947c969..172aa10a8800 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/dc_fpu.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/dc_fpu.c @@ -33,6 +33,8 @@ #include <asm/cputable.h> #elif defined(CONFIG_ARM64) #include <asm/neon.h> +#elif defined(CONFIG_LOONGARCH) +#include <asm/fpu.h> #endif /** @@ -88,7 +90,7 @@ void dc_fpu_begin(const char *function_name, const int line) *pcpu += 1; if (*pcpu == 1) { -#if defined(CONFIG_X86) +#if defined(CONFIG_X86) || defined(CONFIG_LOONGARCH) migrate_disable(); kernel_fpu_begin(); #elif defined(CONFIG_PPC64) @@ -128,7 +130,7 @@ void dc_fpu_end(const char *function_name, const int line) pcpu = get_cpu_ptr(&fpu_recursion_depth); *pcpu -= 1; if (*pcpu <= 0) { -#if defined(CONFIG_X86) +#if defined(CONFIG_X86) || defined(CONFIG_LOONGARCH) kernel_fpu_end(); migrate_enable(); #elif defined(CONFIG_PPC64) diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table2.c b/drivers/gpu/drm/amd/display/dc/bios/command_table2.c index 1ef9e4053bb7..90a02d7bd3da 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/command_table2.c +++ b/drivers/gpu/drm/amd/display/dc/bios/command_table2.c @@ -123,9 +123,7 @@ static void encoder_control_dmcub( sizeof(cmd.digx_encoder_control.header); cmd.digx_encoder_control.encoder_control.dig.stream_param = *dig; - dc_dmub_srv_cmd_queue(dmcub, &cmd); - dc_dmub_srv_cmd_execute(dmcub); - dc_dmub_srv_wait_idle(dmcub); + dm_execute_dmub_cmd(dmcub->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } static enum bp_result encoder_control_digx_v1_5( @@ -261,9 +259,7 @@ static void transmitter_control_dmcub( sizeof(cmd.dig1_transmitter_control.header); cmd.dig1_transmitter_control.transmitter_control.dig = *dig; - dc_dmub_srv_cmd_queue(dmcub, &cmd); - dc_dmub_srv_cmd_execute(dmcub); - dc_dmub_srv_wait_idle(dmcub); + dm_execute_dmub_cmd(dmcub->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } static enum bp_result transmitter_control_v1_6( @@ -325,9 +321,7 @@ static void transmitter_control_dmcub_v1_7( sizeof(cmd.dig1_transmitter_control.header); cmd.dig1_transmitter_control.transmitter_control.dig_v1_7 = *dig; - dc_dmub_srv_cmd_queue(dmcub, &cmd); - dc_dmub_srv_cmd_execute(dmcub); - dc_dmub_srv_wait_idle(dmcub); + dm_execute_dmub_cmd(dmcub->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } static enum bp_result transmitter_control_v1_7( @@ -435,9 +429,7 @@ static void set_pixel_clock_dmcub( sizeof(cmd.set_pixel_clock.header); cmd.set_pixel_clock.pixel_clock.clk = *clk; - dc_dmub_srv_cmd_queue(dmcub, &cmd); - dc_dmub_srv_cmd_execute(dmcub); - dc_dmub_srv_wait_idle(dmcub); + dm_execute_dmub_cmd(dmcub->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } static enum bp_result set_pixel_clock_v7( @@ -804,9 +796,7 @@ static void enable_disp_power_gating_dmcub( sizeof(cmd.enable_disp_power_gating.header); cmd.enable_disp_power_gating.power_gating.pwr = *pwr; - dc_dmub_srv_cmd_queue(dmcub, &cmd); - dc_dmub_srv_cmd_execute(dmcub); - dc_dmub_srv_wait_idle(dmcub); + dm_execute_dmub_cmd(dmcub->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } static enum bp_result enable_disp_power_gating_v2_1( @@ -1016,10 +1006,7 @@ static void enable_lvtma_control_dmcub( panel_instance; cmd.lvtma_control.data.bypass_panel_control_wait = bypass_panel_control_wait; - dc_dmub_srv_cmd_queue(dmcub, &cmd); - dc_dmub_srv_cmd_execute(dmcub); - dc_dmub_srv_wait_idle(dmcub); - + dm_execute_dmub_cmd(dmcub->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } static enum bp_result enable_lvtma_control( diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce112/dce112_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce112/dce112_clk_mgr.c index 934e6423dc1a..1f36ad8a7de4 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce112/dce112_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce112/dce112_clk_mgr.c @@ -111,12 +111,10 @@ int dce112_set_clock(struct clk_mgr *clk_mgr_base, int requested_clk_khz) bp->funcs->set_dce_clock(bp, &dce_clk_params); - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) { - if (clk_mgr_dce->dfs_bypass_disp_clk != actual_clock) - dmcu->funcs->set_psr_wait_loop(dmcu, - actual_clock / 1000 / 7); - } + if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) { + if (clk_mgr_dce->dfs_bypass_disp_clk != actual_clock) + dmcu->funcs->set_psr_wait_loop(dmcu, + actual_clock / 1000 / 7); } clk_mgr_dce->dfs_bypass_disp_clk = actual_clock; @@ -153,12 +151,10 @@ int dce112_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_clk_khz) clk_mgr->cur_min_clks_state = DM_PP_CLOCKS_STATE_NOMINAL; - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) { - if (clk_mgr->dfs_bypass_disp_clk != actual_clock) - dmcu->funcs->set_psr_wait_loop(dmcu, - actual_clock / 1000 / 7); - } + if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) { + if (clk_mgr->dfs_bypass_disp_clk != actual_clock) + dmcu->funcs->set_psr_wait_loop(dmcu, + actual_clock / 1000 / 7); } clk_mgr->dfs_bypass_disp_clk = actual_clock; diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.c index 450eaead4f20..89b79dd39628 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.c @@ -135,12 +135,10 @@ int rv1_vbios_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_di VBIOSSMC_MSG_SetDispclkFreq, khz_to_mhz_ceil(requested_dispclk_khz)); - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) { - if (clk_mgr->dfs_bypass_disp_clk != actual_dispclk_set_mhz) - dmcu->funcs->set_psr_wait_loop(dmcu, - actual_dispclk_set_mhz / 7); - } + if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) { + if (clk_mgr->dfs_bypass_disp_clk != actual_dispclk_set_mhz) + dmcu->funcs->set_psr_wait_loop(dmcu, + actual_dispclk_set_mhz / 7); } return actual_dispclk_set_mhz * 1000; diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c index 650f3b4b562e..c435f7632e8e 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c @@ -531,6 +531,11 @@ void dcn20_clk_mgr_construct( struct pp_smu_funcs *pp_smu, struct dccg *dccg) { + int dprefclk_did; + int target_div; + uint32_t pll_req_reg; + struct fixed31_32 pll_req; + clk_mgr->base.ctx = ctx; clk_mgr->pp_smu = pp_smu; clk_mgr->base.funcs = &dcn2_funcs; @@ -547,42 +552,34 @@ void dcn20_clk_mgr_construct( clk_mgr->base.dprefclk_khz = 700000; // 700 MHz planned if VCO is 3.85 GHz, will be retrieved - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) { - dcn2_funcs.update_clocks = dcn2_update_clocks_fpga; - clk_mgr->base.dentist_vco_freq_khz = 3850000; + /* DFS Slice 2 should be used for DPREFCLK */ + dprefclk_did = REG_READ(CLK3_CLK2_DFS_CNTL); + /* Convert DPREFCLK DFS Slice DID to actual divider */ + target_div = dentist_get_divider_from_did(dprefclk_did); + /* get FbMult value */ + pll_req_reg = REG_READ(CLK3_CLK_PLL_REQ); - } else { - /* DFS Slice 2 should be used for DPREFCLK */ - int dprefclk_did = REG_READ(CLK3_CLK2_DFS_CNTL); - /* Convert DPREFCLK DFS Slice DID to actual divider*/ - int target_div = dentist_get_divider_from_did(dprefclk_did); - - /* get FbMult value */ - uint32_t pll_req_reg = REG_READ(CLK3_CLK_PLL_REQ); - struct fixed31_32 pll_req; - - /* set up a fixed-point number - * this works because the int part is on the right edge of the register - * and the frac part is on the left edge - */ + /* set up a fixed-point number + * this works because the int part is on the right edge of the register + * and the frac part is on the left edge + */ - pll_req = dc_fixpt_from_int(pll_req_reg & clk_mgr->clk_mgr_mask->FbMult_int); - pll_req.value |= pll_req_reg & clk_mgr->clk_mgr_mask->FbMult_frac; + pll_req = dc_fixpt_from_int(pll_req_reg & clk_mgr->clk_mgr_mask->FbMult_int); + pll_req.value |= pll_req_reg & clk_mgr->clk_mgr_mask->FbMult_frac; - /* multiply by REFCLK period */ - pll_req = dc_fixpt_mul_int(pll_req, 100000); + /* multiply by REFCLK period */ + pll_req = dc_fixpt_mul_int(pll_req, 100000); - /* integer part is now VCO frequency in kHz */ - clk_mgr->base.dentist_vco_freq_khz = dc_fixpt_floor(pll_req); + /* integer part is now VCO frequency in kHz */ + clk_mgr->base.dentist_vco_freq_khz = dc_fixpt_floor(pll_req); - /* in case we don't get a value from the register, use default */ - if (clk_mgr->base.dentist_vco_freq_khz == 0) - clk_mgr->base.dentist_vco_freq_khz = 3850000; + /* in case we don't get a value from the register, use default */ + if (clk_mgr->base.dentist_vco_freq_khz == 0) + clk_mgr->base.dentist_vco_freq_khz = 3850000; - /* Calculate the DPREFCLK in kHz.*/ - clk_mgr->base.dprefclk_khz = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR - * clk_mgr->base.dentist_vco_freq_khz) / target_div; - } + /* Calculate the DPREFCLK in kHz.*/ + clk_mgr->base.dprefclk_khz = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR + * clk_mgr->base.dentist_vco_freq_khz) / target_div; //Integrated_info table does not exist on dGPU projects so should not be referenced //anywhere in code for dGPUs. //Also there is no plan for now that DFS BYPASS will be used on NV10/12/14. @@ -590,4 +587,3 @@ void dcn20_clk_mgr_construct( dce_clock_read_ss_info(clk_mgr); } - diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c index 811720749faf..694fe4271b4d 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c @@ -190,23 +190,17 @@ void dcn201_clk_mgr_construct(struct dc_context *ctx, clk_mgr->dprefclk_ss_divider = 1000; clk_mgr->ss_on_dprefclk = false; - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) { - dcn201_funcs.update_clocks = dcn2_update_clocks_fpga; - clk_mgr->base.dprefclk_khz = 600000; - clk_mgr->base.dentist_vco_freq_khz = 3000000; - } else { - clk_mgr->base.dprefclk_khz = REG_READ(CLK4_CLK2_CURRENT_CNT); - clk_mgr->base.dprefclk_khz *= 100; + clk_mgr->base.dprefclk_khz = REG_READ(CLK4_CLK2_CURRENT_CNT); + clk_mgr->base.dprefclk_khz *= 100; - if (clk_mgr->base.dprefclk_khz == 0) - clk_mgr->base.dprefclk_khz = 600000; + if (clk_mgr->base.dprefclk_khz == 0) + clk_mgr->base.dprefclk_khz = 600000; - REG_GET(CLK4_CLK_PLL_REQ, FbMult_int, &clk_mgr->base.dentist_vco_freq_khz); - clk_mgr->base.dentist_vco_freq_khz *= 100000; + REG_GET(CLK4_CLK_PLL_REQ, FbMult_int, &clk_mgr->base.dentist_vco_freq_khz); + clk_mgr->base.dentist_vco_freq_khz *= 100000; - if (clk_mgr->base.dentist_vco_freq_khz == 0) - clk_mgr->base.dentist_vco_freq_khz = 3000000; - } + if (clk_mgr->base.dentist_vco_freq_khz == 0) + clk_mgr->base.dentist_vco_freq_khz = 3000000; if (!debug->disable_dfs_bypass && bp->integrated_info) if (bp->integrated_info->gpu_cap_info & DFS_BYPASS_ENABLE) diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.c index bd9fd0b54f46..0c6a4ab72b1d 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.c @@ -705,6 +705,7 @@ void rn_clk_mgr_construct( struct dpm_clocks clock_table = { 0 }; enum pp_smu_status status = 0; int is_green_sardine = 0; + struct clk_log_info log_info = {0}; #if defined(CONFIG_DRM_AMD_DC_FP) is_green_sardine = ASICREV_IS_GREEN_SARDINE(ctx->asic_id.hw_internal_rev); @@ -725,48 +726,41 @@ void rn_clk_mgr_construct( clk_mgr->smu_ver = rn_vbios_smu_get_smu_version(clk_mgr); - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) { - dcn21_funcs.update_clocks = dcn2_update_clocks_fpga; + clk_mgr->periodic_retraining_disabled = rn_vbios_smu_is_periodic_retraining_disabled(clk_mgr); + + /* SMU Version 55.51.0 and up no longer have an issue + * that needs to limit minimum dispclk */ + if (clk_mgr->smu_ver >= SMU_VER_55_51_0) + debug->min_disp_clk_khz = 0; + + /* TODO: Check we get what we expect during bringup */ + clk_mgr->base.dentist_vco_freq_khz = get_vco_frequency_from_reg(clk_mgr); + + /* in case we don't get a value from the register, use default */ + if (clk_mgr->base.dentist_vco_freq_khz == 0) clk_mgr->base.dentist_vco_freq_khz = 3600000; - } else { - struct clk_log_info log_info = {0}; - - clk_mgr->periodic_retraining_disabled = rn_vbios_smu_is_periodic_retraining_disabled(clk_mgr); - - /* SMU Version 55.51.0 and up no longer have an issue - * that needs to limit minimum dispclk */ - if (clk_mgr->smu_ver >= SMU_VER_55_51_0) - debug->min_disp_clk_khz = 0; - - /* TODO: Check we get what we expect during bringup */ - clk_mgr->base.dentist_vco_freq_khz = get_vco_frequency_from_reg(clk_mgr); - - /* in case we don't get a value from the register, use default */ - if (clk_mgr->base.dentist_vco_freq_khz == 0) - clk_mgr->base.dentist_vco_freq_khz = 3600000; - - if (ctx->dc_bios->integrated_info->memory_type == LpDdr4MemType) { - if (clk_mgr->periodic_retraining_disabled) { - rn_bw_params.wm_table = lpddr4_wm_table_with_disabled_ppt; - } else { - if (is_green_sardine) - rn_bw_params.wm_table = lpddr4_wm_table_gs; - else - rn_bw_params.wm_table = lpddr4_wm_table_rn; - } + + if (ctx->dc_bios->integrated_info->memory_type == LpDdr4MemType) { + if (clk_mgr->periodic_retraining_disabled) { + rn_bw_params.wm_table = lpddr4_wm_table_with_disabled_ppt; } else { if (is_green_sardine) - rn_bw_params.wm_table = ddr4_wm_table_gs; - else { - if (ctx->dc->config.is_single_rank_dimm) - rn_bw_params.wm_table = ddr4_1R_wm_table_rn; - else - rn_bw_params.wm_table = ddr4_wm_table_rn; - } + rn_bw_params.wm_table = lpddr4_wm_table_gs; + else + rn_bw_params.wm_table = lpddr4_wm_table_rn; + } + } else { + if (is_green_sardine) + rn_bw_params.wm_table = ddr4_wm_table_gs; + else { + if (ctx->dc->config.is_single_rank_dimm) + rn_bw_params.wm_table = ddr4_1R_wm_table_rn; + else + rn_bw_params.wm_table = ddr4_wm_table_rn; } - /* Saved clocks configured at boot for debug purposes */ - rn_dump_clk_registers(&clk_mgr->base.boot_snapshot, &clk_mgr->base, &log_info); } + /* Saved clocks configured at boot for debug purposes */ + rn_dump_clk_registers(&clk_mgr->base.boot_snapshot, &clk_mgr->base, &log_info); clk_mgr->base.dprefclk_khz = 600000; dce_clock_read_ss_info(clk_mgr); @@ -786,9 +780,8 @@ void rn_clk_mgr_construct( } } - if (!IS_FPGA_MAXIMUS_DC(ctx->dce_environment) && clk_mgr->smu_ver >= 0x00371500) { - /* enable powerfeatures when displaycount goes to 0 */ + /* enable powerfeatures when displaycount goes to 0 */ + if (clk_mgr->smu_ver >= 0x00371500) rn_vbios_smu_enable_48mhz_tmdp_refclk_pwrdwn(clk_mgr, !debug->disable_48mhz_pwrdwn); - } } diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.c index 27fbe906682f..8c9d45e5b13b 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.c @@ -147,17 +147,14 @@ int rn_vbios_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dis VBIOSSMC_MSG_SetDispclkFreq, khz_to_mhz_ceil(requested_dispclk_khz)); - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) { - if (clk_mgr->dfs_bypass_disp_clk != actual_dispclk_set_mhz) - dmcu->funcs->set_psr_wait_loop(dmcu, - actual_dispclk_set_mhz / 7); - } + if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) { + if (clk_mgr->dfs_bypass_disp_clk != actual_dispclk_set_mhz) + dmcu->funcs->set_psr_wait_loop(dmcu, + actual_dispclk_set_mhz / 7); } // pmfw always set clock more than or equal requested clock - if (!IS_DIAG_DC(dc->ctx->dce_environment)) - ASSERT(actual_dispclk_set_mhz >= khz_to_mhz_ceil(requested_dispclk_khz)); + ASSERT(actual_dispclk_set_mhz >= khz_to_mhz_ceil(requested_dispclk_khz)); return actual_dispclk_set_mhz * 1000; } @@ -221,15 +218,13 @@ void rn_vbios_smu_set_phyclk(struct clk_mgr_internal *clk_mgr, int requested_phy int rn_vbios_smu_set_dppclk(struct clk_mgr_internal *clk_mgr, int requested_dpp_khz) { int actual_dppclk_set_mhz = -1; - struct dc *dc = clk_mgr->base.ctx->dc; actual_dppclk_set_mhz = rn_vbios_smu_send_msg_with_param( clk_mgr, VBIOSSMC_MSG_SetDppclkFreq, khz_to_mhz_ceil(requested_dpp_khz)); - if (!IS_DIAG_DC(dc->ctx->dce_environment)) - ASSERT(actual_dppclk_set_mhz >= khz_to_mhz_ceil(requested_dpp_khz)); + ASSERT(actual_dppclk_set_mhz >= khz_to_mhz_ceil(requested_dpp_khz)); return actual_dppclk_set_mhz * 1000; } diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.c index 694a9d3d92ae..3271c8c7905d 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.c @@ -206,7 +206,6 @@ static void dcn3_update_clocks(struct clk_mgr *clk_mgr_base, bool force_reset = false; bool update_uclk = false; bool p_state_change_support; - int total_plane_count; if (dc->work_arounds.skip_clock_update || !clk_mgr->smu_present) return; @@ -247,8 +246,7 @@ static void dcn3_update_clocks(struct clk_mgr *clk_mgr_base, clk_mgr_base->clks.socclk_khz = new_clocks->socclk_khz; clk_mgr_base->clks.prev_p_state_change_support = clk_mgr_base->clks.p_state_change_support; - total_plane_count = clk_mgr_helper_get_active_plane_cnt(dc, context); - p_state_change_support = new_clocks->p_state_change_support || (total_plane_count == 0); + p_state_change_support = new_clocks->p_state_change_support; // invalidate the current P-State forced min in certain dc_mode_softmax situations if (dc->clk_mgr->dc_mode_softmax_enabled && safe_to_lower && !p_state_change_support) { @@ -523,6 +521,8 @@ void dcn3_clk_mgr_construct( struct pp_smu_funcs *pp_smu, struct dccg *dccg) { + struct clk_state_registers_and_bypass s = { 0 }; + clk_mgr->base.ctx = ctx; clk_mgr->base.funcs = &dcn3_funcs; clk_mgr->regs = &clk_mgr_regs; @@ -539,27 +539,19 @@ void dcn3_clk_mgr_construct( clk_mgr->base.dprefclk_khz = 730000; // 700 MHz planned if VCO is 3.85 GHz, will be retrieved - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) { - clk_mgr->base.funcs = &dcn3_fpga_funcs; - clk_mgr->base.dentist_vco_freq_khz = 3650000; - - } else { - struct clk_state_registers_and_bypass s = { 0 }; + /* integer part is now VCO frequency in kHz */ + clk_mgr->base.dentist_vco_freq_khz = dcn30_get_vco_frequency_from_reg(clk_mgr); - /* integer part is now VCO frequency in kHz */ - clk_mgr->base.dentist_vco_freq_khz = dcn30_get_vco_frequency_from_reg(clk_mgr); - - /* in case we don't get a value from the register, use default */ - if (clk_mgr->base.dentist_vco_freq_khz == 0) - clk_mgr->base.dentist_vco_freq_khz = 3650000; - /* Convert dprefclk units from MHz to KHz */ - /* Value already divided by 10, some resolution lost */ + /* in case we don't get a value from the register, use default */ + if (clk_mgr->base.dentist_vco_freq_khz == 0) + clk_mgr->base.dentist_vco_freq_khz = 3650000; + /* Convert dprefclk units from MHz to KHz */ + /* Value already divided by 10, some resolution lost */ - /*TODO: uncomment assert once dcn3_dump_clk_registers is implemented */ - //ASSERT(s.dprefclk != 0); - if (s.dprefclk != 0) - clk_mgr->base.dprefclk_khz = s.dprefclk * 1000; - } + /*TODO: uncomment assert once dcn3_dump_clk_registers is implemented */ + //ASSERT(s.dprefclk != 0); + if (s.dprefclk != 0) + clk_mgr->base.dprefclk_khz = s.dprefclk * 1000; clk_mgr->dfs_bypass_enabled = false; diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr_smu_msg.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr_smu_msg.c index 1fbf1c105dc1..bdbf18306698 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr_smu_msg.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr_smu_msg.c @@ -312,6 +312,9 @@ void dcn30_smu_set_display_refresh_from_mall(struct clk_mgr_internal *clk_mgr, b /* bits 8:7 for cache timer scale, bits 6:1 for cache timer delay, bit 0 = 1 for enable, = 0 for disable */ uint32_t param = (cache_timer_scale << 7) | (cache_timer_delay << 1) | (enable ? 1 : 0); + smu_print("SMU Set display refresh from mall: enable = %d, cache_timer_delay = %d, cache_timer_scale = %d\n", + enable, cache_timer_delay, cache_timer_scale); + dcn30_smu_send_msg_with_param(clk_mgr, DALSMC_MSG_SetDisplayRefreshFromMall, param, NULL); } diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c index 01383aac6b41..a5489fe6875f 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c @@ -117,7 +117,7 @@ static void vg_update_clocks(struct clk_mgr *clk_mgr_base, display_count = vg_get_active_display_cnt_wa(dc, context); /* if we can go lower, go lower */ - if (display_count == 0 && !IS_DIAG_DC(dc->ctx->dce_environment)) { + if (display_count == 0) { union display_idle_optimization_u idle_info = { 0 }; idle_info.idle_info.df_request_disabled = 1; @@ -151,10 +151,8 @@ static void vg_update_clocks(struct clk_mgr *clk_mgr_base, } // workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow. - if (!IS_DIAG_DC(dc->ctx->dce_environment)) { - if (new_clocks->dppclk_khz < 100000) - new_clocks->dppclk_khz = 100000; - } + if (new_clocks->dppclk_khz < 100000) + new_clocks->dppclk_khz = 100000; if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->base.clks.dppclk_khz)) { if (clk_mgr->base.clks.dppclk_khz > new_clocks->dppclk_khz) @@ -664,6 +662,7 @@ void vg_clk_mgr_construct( struct dccg *dccg) { struct smu_dpm_clks smu_dpm_clks = { 0 }; + struct clk_log_info log_info = {0}; clk_mgr->base.base.ctx = ctx; clk_mgr->base.base.funcs = &vg_funcs; @@ -703,32 +702,25 @@ void vg_clk_mgr_construct( ASSERT(smu_dpm_clks.dpm_clks); - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) { - vg_funcs.update_clocks = dcn2_update_clocks_fpga; - clk_mgr->base.base.dentist_vco_freq_khz = 3600000; - } else { - struct clk_log_info log_info = {0}; + clk_mgr->base.smu_ver = dcn301_smu_get_smu_version(&clk_mgr->base); - clk_mgr->base.smu_ver = dcn301_smu_get_smu_version(&clk_mgr->base); + if (clk_mgr->base.smu_ver) + clk_mgr->base.smu_present = true; - if (clk_mgr->base.smu_ver) - clk_mgr->base.smu_present = true; + /* TODO: Check we get what we expect during bringup */ + clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base); - /* TODO: Check we get what we expect during bringup */ - clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base); - - /* in case we don't get a value from the register, use default */ - if (clk_mgr->base.base.dentist_vco_freq_khz == 0) - clk_mgr->base.base.dentist_vco_freq_khz = 3600000; + /* in case we don't get a value from the register, use default */ + if (clk_mgr->base.base.dentist_vco_freq_khz == 0) + clk_mgr->base.base.dentist_vco_freq_khz = 3600000; - if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) { - vg_bw_params.wm_table = lpddr5_wm_table; - } else { - vg_bw_params.wm_table = ddr4_wm_table; - } - /* Saved clocks configured at boot for debug purposes */ - vg_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, &clk_mgr->base.base, &log_info); + if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) { + vg_bw_params.wm_table = lpddr5_wm_table; + } else { + vg_bw_params.wm_table = ddr4_wm_table; } + /* Saved clocks configured at boot for debug purposes */ + vg_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, &clk_mgr->base.base, &log_info); clk_mgr->base.base.dprefclk_khz = 600000; dce_clock_read_ss_info(&clk_mgr->base); @@ -746,12 +738,6 @@ void vg_clk_mgr_construct( if (smu_dpm_clks.dpm_clks && smu_dpm_clks.mc_address.quad_part != 0) dm_helpers_free_gpu_mem(clk_mgr->base.base.ctx, DC_MEM_ALLOC_TYPE_FRAME_BUFFER, smu_dpm_clks.dpm_clks); -/* - if (!IS_FPGA_MAXIMUS_DC(ctx->dce_environment) && clk_mgr->base.smu_ver) { - enable powerfeatures when displaycount goes to 0 - dcn301_smu_enable_phy_refclk_pwrdwn(clk_mgr, !debug->disable_48mhz_pwrdwn); - } -*/ } void vg_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr_int) diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_clk_mgr.c index f9e2e0c3095e..7ccd96959256 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_clk_mgr.c @@ -205,10 +205,8 @@ void dcn31_update_clocks(struct clk_mgr *clk_mgr_base, } // workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow. - if (!IS_DIAG_DC(dc->ctx->dce_environment)) { - if (new_clocks->dppclk_khz < 100000) - new_clocks->dppclk_khz = 100000; - } + if (new_clocks->dppclk_khz < 100000) + new_clocks->dppclk_khz = 100000; if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->base.clks.dppclk_khz)) { if (clk_mgr->base.clks.dppclk_khz > new_clocks->dppclk_khz) @@ -250,9 +248,7 @@ void dcn31_update_clocks(struct clk_mgr *clk_mgr_base, cmd.notify_clocks.clocks.dispclk_khz = clk_mgr_base->clks.dispclk_khz; cmd.notify_clocks.clocks.dppclk_khz = clk_mgr_base->clks.dppclk_khz; - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } static int get_vco_frequency_from_reg(struct clk_mgr_internal *clk_mgr) @@ -674,6 +670,7 @@ void dcn31_clk_mgr_construct( struct dccg *dccg) { struct dcn31_smu_dpm_clks smu_dpm_clks = { 0 }; + struct clk_log_info log_info = {0}; clk_mgr->base.base.ctx = ctx; clk_mgr->base.base.funcs = &dcn31_funcs; @@ -713,29 +710,22 @@ void dcn31_clk_mgr_construct( ASSERT(smu_dpm_clks.dpm_clks); - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) { - clk_mgr->base.base.funcs = &dcn3_fpga_funcs; - } else { - struct clk_log_info log_info = {0}; - - clk_mgr->base.smu_ver = dcn31_smu_get_smu_version(&clk_mgr->base); + clk_mgr->base.smu_ver = dcn31_smu_get_smu_version(&clk_mgr->base); - if (clk_mgr->base.smu_ver) - clk_mgr->base.smu_present = true; + if (clk_mgr->base.smu_ver) + clk_mgr->base.smu_present = true; - /* TODO: Check we get what we expect during bringup */ - clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base); - - if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) { - dcn31_bw_params.wm_table = lpddr5_wm_table; - } else { - dcn31_bw_params.wm_table = ddr5_wm_table; - } - /* Saved clocks configured at boot for debug purposes */ - dcn31_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, - &clk_mgr->base.base, &log_info); + /* TODO: Check we get what we expect during bringup */ + clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base); + if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) { + dcn31_bw_params.wm_table = lpddr5_wm_table; + } else { + dcn31_bw_params.wm_table = ddr5_wm_table; } + /* Saved clocks configured at boot for debug purposes */ + dcn31_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, + &clk_mgr->base.base, &log_info); clk_mgr->base.base.dprefclk_khz = 600000; clk_mgr->base.base.clks.ref_dtbclk_khz = 600000; diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_smu.c index 0827c7df2855..32279c5db724 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_smu.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_smu.c @@ -130,7 +130,7 @@ static int dcn31_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr, if (result == VBIOSSMC_Result_Failed) { if (msg_id == VBIOSSMC_MSG_TransferTableDram2Smu && param == TABLE_WATERMARKS) - DC_LOG_WARNING("Watermarks table not configured properly by SMU"); + DC_LOG_DEBUG("Watermarks table not configured properly by SMU"); else ASSERT(0); REG_WRITE(MP1_SMN_C2PMSG_91, VBIOSSMC_Result_OK); diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_clk_mgr.c index 5cb44f838bde..2f7c8996b19d 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_clk_mgr.c @@ -241,10 +241,8 @@ void dcn314_update_clocks(struct clk_mgr *clk_mgr_base, } // workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow. - if (!IS_DIAG_DC(dc->ctx->dce_environment)) { - if (new_clocks->dppclk_khz < 100000) - new_clocks->dppclk_khz = 100000; - } + if (new_clocks->dppclk_khz < 100000) + new_clocks->dppclk_khz = 100000; if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->base.clks.dppclk_khz)) { if (clk_mgr->base.clks.dppclk_khz > new_clocks->dppclk_khz) @@ -286,9 +284,7 @@ void dcn314_update_clocks(struct clk_mgr *clk_mgr_base, cmd.notify_clocks.clocks.dispclk_khz = clk_mgr_base->clks.dispclk_khz; cmd.notify_clocks.clocks.dppclk_khz = clk_mgr_base->clks.dppclk_khz; - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } static int get_vco_frequency_from_reg(struct clk_mgr_internal *clk_mgr) @@ -405,32 +401,32 @@ static struct wm_table lpddr5_wm_table = { .wm_inst = WM_A, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.65333, - .sr_exit_time_us = 16.5, - .sr_enter_plus_exit_time_us = 18.5, + .sr_exit_time_us = 30.0, + .sr_enter_plus_exit_time_us = 32.0, .valid = true, }, { .wm_inst = WM_B, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.65333, - .sr_exit_time_us = 16.5, - .sr_enter_plus_exit_time_us = 18.5, + .sr_exit_time_us = 30.0, + .sr_enter_plus_exit_time_us = 32.0, .valid = true, }, { .wm_inst = WM_C, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.65333, - .sr_exit_time_us = 16.5, - .sr_enter_plus_exit_time_us = 18.5, + .sr_exit_time_us = 30.0, + .sr_enter_plus_exit_time_us = 32.0, .valid = true, }, { .wm_inst = WM_D, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.65333, - .sr_exit_time_us = 16.5, - .sr_enter_plus_exit_time_us = 18.5, + .sr_exit_time_us = 30.0, + .sr_enter_plus_exit_time_us = 32.0, .valid = true, }, } @@ -726,6 +722,7 @@ void dcn314_clk_mgr_construct( struct dccg *dccg) { struct dcn314_smu_dpm_clks smu_dpm_clks = { 0 }; + struct clk_log_info log_info = {0}; clk_mgr->base.base.ctx = ctx; clk_mgr->base.base.funcs = &dcn314_funcs; @@ -765,29 +762,22 @@ void dcn314_clk_mgr_construct( ASSERT(smu_dpm_clks.dpm_clks); - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) { - clk_mgr->base.base.funcs = &dcn3_fpga_funcs; - } else { - struct clk_log_info log_info = {0}; - - clk_mgr->base.smu_ver = dcn314_smu_get_smu_version(&clk_mgr->base); + clk_mgr->base.smu_ver = dcn314_smu_get_smu_version(&clk_mgr->base); - if (clk_mgr->base.smu_ver) - clk_mgr->base.smu_present = true; + if (clk_mgr->base.smu_ver) + clk_mgr->base.smu_present = true; - /* TODO: Check we get what we expect during bringup */ - clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base); + /* TODO: Check we get what we expect during bringup */ + clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base); - if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) - dcn314_bw_params.wm_table = lpddr5_wm_table; - else - dcn314_bw_params.wm_table = ddr5_wm_table; + if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) + dcn314_bw_params.wm_table = lpddr5_wm_table; + else + dcn314_bw_params.wm_table = ddr5_wm_table; - /* Saved clocks configured at boot for debug purposes */ - dcn314_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, - &clk_mgr->base.base, &log_info); - - } + /* Saved clocks configured at boot for debug purposes */ + dcn314_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, + &clk_mgr->base.base, &log_info); clk_mgr->base.base.dprefclk_khz = 600000; clk_mgr->base.base.clks.ref_dtbclk_khz = 600000; diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_smu.c index 0765334f0825..07baa10a8647 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_smu.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_smu.c @@ -145,7 +145,7 @@ static int dcn314_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr, if (result == VBIOSSMC_Result_Failed) { if (msg_id == VBIOSSMC_MSG_TransferTableDram2Smu && param == TABLE_WATERMARKS) - DC_LOG_WARNING("Watermarks table not configured properly by SMU"); + DC_LOG_DEBUG("Watermarks table not configured properly by SMU"); else if (msg_id == VBIOSSMC_MSG_SetHardMinDcfclkByFreq || msg_id == VBIOSSMC_MSG_SetMinDeepSleepDcfclk) DC_LOG_WARNING("DCFCLK_DPM is not enabled by BIOS"); diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c index b737cbc468f5..b2c4f97afc8b 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c @@ -184,12 +184,10 @@ static void dcn315_update_clocks(struct clk_mgr *clk_mgr_base, } // workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow. - if (!IS_DIAG_DC(dc->ctx->dce_environment)) { - if (new_clocks->dppclk_khz < MIN_DPP_DISP_CLK) - new_clocks->dppclk_khz = MIN_DPP_DISP_CLK; - if (new_clocks->dispclk_khz < MIN_DPP_DISP_CLK) - new_clocks->dispclk_khz = MIN_DPP_DISP_CLK; - } + if (new_clocks->dppclk_khz < MIN_DPP_DISP_CLK) + new_clocks->dppclk_khz = MIN_DPP_DISP_CLK; + if (new_clocks->dispclk_khz < MIN_DPP_DISP_CLK) + new_clocks->dispclk_khz = MIN_DPP_DISP_CLK; if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->base.clks.dppclk_khz)) { if (clk_mgr->base.clks.dppclk_khz > new_clocks->dppclk_khz) @@ -234,9 +232,7 @@ static void dcn315_update_clocks(struct clk_mgr *clk_mgr_base, cmd.notify_clocks.clocks.dispclk_khz = clk_mgr_base->clks.dispclk_khz; cmd.notify_clocks.clocks.dppclk_khz = clk_mgr_base->clks.dppclk_khz; - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } static void dcn315_dump_clk_registers(struct clk_state_registers_and_bypass *regs_and_bypass, @@ -602,6 +598,7 @@ void dcn315_clk_mgr_construct( struct dccg *dccg) { struct dcn315_smu_dpm_clks smu_dpm_clks = { 0 }; + struct clk_log_info log_info = {0}; clk_mgr->base.base.ctx = ctx; clk_mgr->base.base.funcs = &dcn315_funcs; @@ -641,26 +638,19 @@ void dcn315_clk_mgr_construct( ASSERT(smu_dpm_clks.dpm_clks); - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) { - clk_mgr->base.base.funcs = &dcn3_fpga_funcs; - } else { - struct clk_log_info log_info = {0}; - - clk_mgr->base.smu_ver = dcn315_smu_get_smu_version(&clk_mgr->base); + clk_mgr->base.smu_ver = dcn315_smu_get_smu_version(&clk_mgr->base); - if (clk_mgr->base.smu_ver > 0) - clk_mgr->base.smu_present = true; - - if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) { - dcn315_bw_params.wm_table = lpddr5_wm_table; - } else { - dcn315_bw_params.wm_table = ddr5_wm_table; - } - /* Saved clocks configured at boot for debug purposes */ - dcn315_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, - &clk_mgr->base.base, &log_info); + if (clk_mgr->base.smu_ver > 0) + clk_mgr->base.smu_present = true; + if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) { + dcn315_bw_params.wm_table = lpddr5_wm_table; + } else { + dcn315_bw_params.wm_table = ddr5_wm_table; } + /* Saved clocks configured at boot for debug purposes */ + dcn315_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, + &clk_mgr->base.base, &log_info); clk_mgr->base.base.dprefclk_khz = 600000; clk_mgr->base.base.dprefclk_khz = dcn315_smu_get_dpref_clk(&clk_mgr->base); diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c index 93db4dbee713..d7de756301cf 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c @@ -207,12 +207,10 @@ static void dcn316_update_clocks(struct clk_mgr *clk_mgr_base, } // workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow. - if (!IS_DIAG_DC(dc->ctx->dce_environment)) { - if (new_clocks->dppclk_khz < 100000) - new_clocks->dppclk_khz = 100000; - if (new_clocks->dispclk_khz < 100000) - new_clocks->dispclk_khz = 100000; - } + if (new_clocks->dppclk_khz < 100000) + new_clocks->dppclk_khz = 100000; + if (new_clocks->dispclk_khz < 100000) + new_clocks->dispclk_khz = 100000; if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->base.clks.dppclk_khz)) { if (clk_mgr->base.clks.dppclk_khz > new_clocks->dppclk_khz) @@ -254,9 +252,7 @@ static void dcn316_update_clocks(struct clk_mgr *clk_mgr_base, cmd.notify_clocks.clocks.dispclk_khz = clk_mgr_base->clks.dispclk_khz; cmd.notify_clocks.clocks.dppclk_khz = clk_mgr_base->clks.dppclk_khz; - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } static void dcn316_dump_clk_registers(struct clk_state_registers_and_bypass *regs_and_bypass, @@ -618,6 +614,7 @@ void dcn316_clk_mgr_construct( struct dccg *dccg) { struct dcn316_smu_dpm_clks smu_dpm_clks = { 0 }; + struct clk_log_info log_info = {0}; clk_mgr->base.base.ctx = ctx; clk_mgr->base.base.funcs = &dcn316_funcs; @@ -657,35 +654,27 @@ void dcn316_clk_mgr_construct( ASSERT(smu_dpm_clks.dpm_clks); - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) { - clk_mgr->base.base.funcs = &dcn3_fpga_funcs; - clk_mgr->base.base.dentist_vco_freq_khz = 2500000; - } else { - struct clk_log_info log_info = {0}; - - clk_mgr->base.smu_ver = dcn316_smu_get_smu_version(&clk_mgr->base); + clk_mgr->base.smu_ver = dcn316_smu_get_smu_version(&clk_mgr->base); - if (clk_mgr->base.smu_ver > 0) - clk_mgr->base.smu_present = true; + if (clk_mgr->base.smu_ver > 0) + clk_mgr->base.smu_present = true; - // Skip this for now as it did not work on DCN315, renable during bring up - clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base); + // Skip this for now as it did not work on DCN315, renable during bring up + clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base); - /* in case we don't get a value from the register, use default */ - if (clk_mgr->base.base.dentist_vco_freq_khz == 0) - clk_mgr->base.base.dentist_vco_freq_khz = 2500000; /* 2400MHz */ + /* in case we don't get a value from the register, use default */ + if (clk_mgr->base.base.dentist_vco_freq_khz == 0) + clk_mgr->base.base.dentist_vco_freq_khz = 2500000; /* 2400MHz */ - if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) { - dcn316_bw_params.wm_table = lpddr5_wm_table; - } else { - dcn316_bw_params.wm_table = ddr4_wm_table; - } - /* Saved clocks configured at boot for debug purposes */ - dcn316_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, - &clk_mgr->base.base, &log_info); - + if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) { + dcn316_bw_params.wm_table = lpddr5_wm_table; + } else { + dcn316_bw_params.wm_table = ddr4_wm_table; } + /* Saved clocks configured at boot for debug purposes */ + dcn316_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, + &clk_mgr->base.base, &log_info); clk_mgr->base.base.dprefclk_khz = 600000; clk_mgr->base.base.dprefclk_khz = dcn316_smu_get_dpref_clk(&clk_mgr->base); diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr.c index 8d9444db092a..6a811755e2e6 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr.c @@ -182,23 +182,32 @@ void dcn32_init_clocks(struct clk_mgr *clk_mgr_base) dcn32_init_single_clock(clk_mgr, PPCLK_DCFCLK, &clk_mgr_base->bw_params->clk_table.entries[0].dcfclk_mhz, &num_entries_per_clk->num_dcfclk_levels); + clk_mgr_base->bw_params->dc_mode_limit.dcfclk_mhz = dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_DCFCLK); /* SOCCLK */ dcn32_init_single_clock(clk_mgr, PPCLK_SOCCLK, &clk_mgr_base->bw_params->clk_table.entries[0].socclk_mhz, &num_entries_per_clk->num_socclk_levels); + clk_mgr_base->bw_params->dc_mode_limit.socclk_mhz = dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_SOCCLK); /* DTBCLK */ - if (!clk_mgr->base.ctx->dc->debug.disable_dtb_ref_clk_switch) + if (!clk_mgr->base.ctx->dc->debug.disable_dtb_ref_clk_switch) { dcn32_init_single_clock(clk_mgr, PPCLK_DTBCLK, &clk_mgr_base->bw_params->clk_table.entries[0].dtbclk_mhz, &num_entries_per_clk->num_dtbclk_levels); + clk_mgr_base->bw_params->dc_mode_limit.dtbclk_mhz = + dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_DTBCLK); + } /* DISPCLK */ dcn32_init_single_clock(clk_mgr, PPCLK_DISPCLK, &clk_mgr_base->bw_params->clk_table.entries[0].dispclk_mhz, &num_entries_per_clk->num_dispclk_levels); num_levels = num_entries_per_clk->num_dispclk_levels; + clk_mgr_base->bw_params->dc_mode_limit.dispclk_mhz = dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_DISPCLK); + //HW recommends limit of 1950 MHz in display clock for all DCN3.2.x + if (clk_mgr_base->bw_params->dc_mode_limit.dispclk_mhz > 1950) + clk_mgr_base->bw_params->dc_mode_limit.dispclk_mhz = 1950; if (num_entries_per_clk->num_dcfclk_levels && num_entries_per_clk->num_dtbclk_levels && @@ -233,6 +242,32 @@ void dcn32_init_clocks(struct clk_mgr *clk_mgr_base) DC_FP_END(); } +static void dcn32_update_clocks_update_dtb_dto(struct clk_mgr_internal *clk_mgr, + struct dc_state *context, + int ref_dtbclk_khz) +{ + struct dccg *dccg = clk_mgr->dccg; + uint32_t tg_mask = 0; + int i; + + for (i = 0; i < clk_mgr->base.ctx->dc->res_pool->pipe_count; i++) { + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; + struct dtbclk_dto_params dto_params = {0}; + + /* use mask to program DTO once per tg */ + if (pipe_ctx->stream_res.tg && + !(tg_mask & (1 << pipe_ctx->stream_res.tg->inst))) { + tg_mask |= (1 << pipe_ctx->stream_res.tg->inst); + + dto_params.otg_inst = pipe_ctx->stream_res.tg->inst; + dto_params.ref_dtbclk_khz = ref_dtbclk_khz; + + dccg->funcs->set_dtbclk_dto(clk_mgr->dccg, &dto_params); + //dccg->funcs->set_audio_dtbclk_dto(clk_mgr->dccg, &dto_params); + } + } +} + /* Since DPPCLK request to PMFW needs to be exact (due to DPP DTO programming), * update DPPCLK to be the exact frequency that will be set after the DPPCLK * divider is updated. This will prevent rounding issues that could cause DPP @@ -433,10 +468,6 @@ static void dcn32_update_clocks(struct clk_mgr *clk_mgr_base, bool update_uclk = false, update_fclk = false; bool p_state_change_support; bool fclk_p_state_change_support; - int total_plane_count; - - if (dc->work_arounds.skip_clock_update) - return; if (clk_mgr_base->clks.dispclk_khz == 0 || (dc->debug.force_clock_mode & 0x1)) { @@ -462,10 +493,10 @@ static void dcn32_update_clocks(struct clk_mgr *clk_mgr_base, clk_mgr_base->clks.fclk_prev_p_state_change_support = clk_mgr_base->clks.fclk_p_state_change_support; - total_plane_count = clk_mgr_helper_get_active_plane_cnt(dc, context); - fclk_p_state_change_support = new_clocks->fclk_p_state_change_support || (total_plane_count == 0); + fclk_p_state_change_support = new_clocks->fclk_p_state_change_support; - if (should_update_pstate_support(safe_to_lower, fclk_p_state_change_support, clk_mgr_base->clks.fclk_p_state_change_support)) { + if (should_update_pstate_support(safe_to_lower, fclk_p_state_change_support, clk_mgr_base->clks.fclk_p_state_change_support) && + !dc->work_arounds.clock_update_disable_mask.fclk) { clk_mgr_base->clks.fclk_p_state_change_support = fclk_p_state_change_support; /* To enable FCLK P-state switching, send FCLK_PSTATE_SUPPORTED message to PMFW */ @@ -479,12 +510,14 @@ static void dcn32_update_clocks(struct clk_mgr *clk_mgr_base, new_clocks->dcfclk_khz = (new_clocks->dcfclk_khz > (dc->debug.force_min_dcfclk_mhz * 1000)) ? new_clocks->dcfclk_khz : (dc->debug.force_min_dcfclk_mhz * 1000); - if (should_set_clock(safe_to_lower, new_clocks->dcfclk_khz, clk_mgr_base->clks.dcfclk_khz)) { + if (should_set_clock(safe_to_lower, new_clocks->dcfclk_khz, clk_mgr_base->clks.dcfclk_khz) && + !dc->work_arounds.clock_update_disable_mask.dcfclk) { clk_mgr_base->clks.dcfclk_khz = new_clocks->dcfclk_khz; dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_DCFCLK, khz_to_mhz_ceil(clk_mgr_base->clks.dcfclk_khz)); } - if (should_set_clock(safe_to_lower, new_clocks->dcfclk_deep_sleep_khz, clk_mgr_base->clks.dcfclk_deep_sleep_khz)) { + if (should_set_clock(safe_to_lower, new_clocks->dcfclk_deep_sleep_khz, clk_mgr_base->clks.dcfclk_deep_sleep_khz) && + !dc->work_arounds.clock_update_disable_mask.dcfclk_ds) { clk_mgr_base->clks.dcfclk_deep_sleep_khz = new_clocks->dcfclk_deep_sleep_khz; dcn30_smu_set_min_deep_sleep_dcef_clk(clk_mgr, khz_to_mhz_ceil(clk_mgr_base->clks.dcfclk_deep_sleep_khz)); } @@ -502,9 +535,9 @@ static void dcn32_update_clocks(struct clk_mgr *clk_mgr_base, dcn32_smu_send_cab_for_uclk_message(clk_mgr, clk_mgr_base->clks.num_ways); } - - p_state_change_support = new_clocks->p_state_change_support || (total_plane_count == 0); - if (should_update_pstate_support(safe_to_lower, p_state_change_support, clk_mgr_base->clks.p_state_change_support)) { + p_state_change_support = new_clocks->p_state_change_support; + if (should_update_pstate_support(safe_to_lower, p_state_change_support, clk_mgr_base->clks.p_state_change_support) && + !dc->work_arounds.clock_update_disable_mask.uclk) { clk_mgr_base->clks.p_state_change_support = p_state_change_support; /* to disable P-State switching, set UCLK min = max */ @@ -518,20 +551,23 @@ static void dcn32_update_clocks(struct clk_mgr *clk_mgr_base, update_fclk = true; } - if (clk_mgr_base->ctx->dce_version != DCN_VERSION_3_21 && !clk_mgr_base->clks.fclk_p_state_change_support && update_fclk) { + if (clk_mgr_base->ctx->dce_version != DCN_VERSION_3_21 && !clk_mgr_base->clks.fclk_p_state_change_support && update_fclk && + !dc->work_arounds.clock_update_disable_mask.fclk) { /* Handle code for sending a message to PMFW that FCLK P-state change is not supported */ dcn32_smu_send_fclk_pstate_message(clk_mgr, FCLK_PSTATE_NOTSUPPORTED); } /* Always update saved value, even if new value not set due to P-State switching unsupported */ - if (should_set_clock(safe_to_lower, new_clocks->dramclk_khz, clk_mgr_base->clks.dramclk_khz)) { + if (should_set_clock(safe_to_lower, new_clocks->dramclk_khz, clk_mgr_base->clks.dramclk_khz) && + !dc->work_arounds.clock_update_disable_mask.uclk) { clk_mgr_base->clks.dramclk_khz = new_clocks->dramclk_khz; update_uclk = true; } /* set UCLK to requested value if P-State switching is supported, or to re-enable P-State switching */ if (clk_mgr_base->clks.p_state_change_support && - (update_uclk || !clk_mgr_base->clks.prev_p_state_change_support)) + (update_uclk || !clk_mgr_base->clks.prev_p_state_change_support) && + !dc->work_arounds.clock_update_disable_mask.uclk) dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK, khz_to_mhz_ceil(clk_mgr_base->clks.dramclk_khz)); if (clk_mgr_base->clks.num_ways != new_clocks->num_ways && @@ -570,6 +606,7 @@ static void dcn32_update_clocks(struct clk_mgr *clk_mgr_base, /* DCCG requires KHz precision for DTBCLK */ clk_mgr_base->clks.ref_dtbclk_khz = dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_DTBCLK, khz_to_mhz_ceil(new_clocks->ref_dtbclk_khz)); + dcn32_update_clocks_update_dtb_dto(clk_mgr, context, clk_mgr_base->clks.ref_dtbclk_khz); } if (dc->config.forced_clocks == false || (force_reset && safe_to_lower)) { @@ -789,6 +826,7 @@ static void dcn32_get_memclk_states_from_smu(struct clk_mgr *clk_mgr_base) dcn32_init_single_clock(clk_mgr, PPCLK_UCLK, &clk_mgr_base->bw_params->clk_table.entries[0].memclk_mhz, &num_entries_per_clk->num_memclk_levels); + clk_mgr_base->bw_params->dc_mode_limit.memclk_mhz = dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_UCLK); /* memclk must have at least one level */ num_entries_per_clk->num_memclk_levels = num_entries_per_clk->num_memclk_levels ? num_entries_per_clk->num_memclk_levels : 1; @@ -796,6 +834,7 @@ static void dcn32_get_memclk_states_from_smu(struct clk_mgr *clk_mgr_base) dcn32_init_single_clock(clk_mgr, PPCLK_FCLK, &clk_mgr_base->bw_params->clk_table.entries[0].fclk_mhz, &num_entries_per_clk->num_fclk_levels); + clk_mgr_base->bw_params->dc_mode_limit.fclk_mhz = dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_FCLK); if (num_entries_per_clk->num_memclk_levels >= num_entries_per_clk->num_fclk_levels) { num_levels = num_entries_per_clk->num_memclk_levels; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c index 52564b93f7eb..dd3a9d06c6e2 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c @@ -515,8 +515,7 @@ dc_stream_forward_dmub_crc_window(struct dc_dmub_srv *dmub_srv, cmd.secure_display.roi_info.y_end = rect->y + rect->height; } - dc_dmub_srv_cmd_queue(dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dmub_srv); + dm_execute_dmub_cmd(dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_NO_WAIT); } static inline void @@ -858,7 +857,6 @@ static bool dc_construct_ctx(struct dc *dc, const struct dc_init_data *init_params) { struct dc_context *dc_ctx; - enum dce_version dc_version = DCE_VERSION_UNKNOWN; dc_ctx = kzalloc(sizeof(*dc_ctx), GFP_KERNEL); if (!dc_ctx) @@ -876,8 +874,7 @@ static bool dc_construct_ctx(struct dc *dc, /* Create logger */ - dc_version = resource_parse_asic_id(init_params->asic_id); - dc_ctx->dce_version = dc_version; + dc_ctx->dce_version = resource_parse_asic_id(init_params->asic_id); dc_ctx->perf_trace = dc_perf_trace_create(); if (!dc_ctx->perf_trace) { @@ -1120,6 +1117,33 @@ static void phantom_pipe_blank( hws->funcs.wait_for_blank_complete(opp); } +static void dc_update_viusal_confirm_color(struct dc *dc, struct dc_state *context, struct pipe_ctx *pipe_ctx) +{ + if (dc->ctx->dce_version >= DCN_VERSION_1_0) { + memset(&pipe_ctx->visual_confirm_color, 0, sizeof(struct tg_color)); + + if (dc->debug.visual_confirm == VISUAL_CONFIRM_HDR) + get_hdr_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color)); + else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SURFACE) + get_surface_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color)); + else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SWIZZLE) + get_surface_tile_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color)); + else { + if (dc->ctx->dce_version < DCN_VERSION_2_0) + color_space_to_black_color( + dc, pipe_ctx->stream->output_color_space, &(pipe_ctx->visual_confirm_color)); + } + if (dc->ctx->dce_version >= DCN_VERSION_2_0) { + if (dc->debug.visual_confirm == VISUAL_CONFIRM_MPCTREE) + get_mpctree_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color)); + else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SUBVP) + get_subvp_visual_confirm_color(dc, context, pipe_ctx, &(pipe_ctx->visual_confirm_color)); + else if (dc->debug.visual_confirm == VISUAL_CONFIRM_MCLK_SWITCH) + get_mclk_switch_visual_confirm_color(dc, context, pipe_ctx, &(pipe_ctx->visual_confirm_color)); + } + } +} + static void disable_dangling_plane(struct dc *dc, struct dc_state *context) { int i, j; @@ -1190,6 +1214,9 @@ static void disable_dangling_plane(struct dc *dc, struct dc_state *context) dc_rem_all_planes_for_stream(dc, old_stream, dangling_context); disable_all_writeback_pipes_for_stream(dc, old_stream, dangling_context); + if (pipe->stream && pipe->plane_state) + dc_update_viusal_confirm_color(dc, context, pipe); + if (dc->hwss.apply_ctx_for_surface) { apply_ctx_interdependent_lock(dc, dc->current_state, old_stream, true); dc->hwss.apply_ctx_for_surface(dc, old_stream, 0, dangling_context); @@ -1981,6 +2008,9 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c return result; } +static bool commit_minimal_transition_state(struct dc *dc, + struct dc_state *transition_base_context); + /** * dc_commit_streams - Commit current stream state * @@ -2002,6 +2032,8 @@ enum dc_status dc_commit_streams(struct dc *dc, struct dc_state *context; enum dc_status res = DC_OK; struct dc_validation_set set[MAX_STREAMS] = {0}; + struct pipe_ctx *pipe; + bool handle_exit_odm2to1 = false; if (dc->ctx->dce_environment == DCE_ENV_VIRTUAL_HW) return res; @@ -2026,6 +2058,22 @@ enum dc_status dc_commit_streams(struct dc *dc, } } + /* Check for case where we are going from odm 2:1 to max + * pipe scenario. For these cases, we will call + * commit_minimal_transition_state() to exit out of odm 2:1 + * first before processing new streams + */ + if (stream_count == dc->res_pool->pipe_count) { + for (i = 0; i < dc->res_pool->pipe_count; i++) { + pipe = &dc->current_state->res_ctx.pipe_ctx[i]; + if (pipe->next_odm_pipe) + handle_exit_odm2to1 = true; + } + } + + if (handle_exit_odm2to1) + res = commit_minimal_transition_state(dc, dc->current_state); + context = dc_create_state(dc); if (!context) goto context_alloc_fail; @@ -2483,9 +2531,6 @@ static enum surface_update_type det_surface_update(const struct dc *dc, enum surface_update_type overall_type = UPDATE_TYPE_FAST; union surface_update_flags *update_flags = &u->surface->update_flags; - if (u->flip_addr) - update_flags->bits.addr_update = 1; - if (!is_surface_in_context(context, u->surface) || u->surface->force_full_update) { update_flags->raw = 0xFFFFFFFF; return UPDATE_TYPE_FULL; @@ -2544,15 +2589,19 @@ static enum surface_update_type det_surface_update(const struct dc *dc, elevate_update_type(&overall_type, type); } - if (update_flags->bits.input_csc_change - || update_flags->bits.coeff_reduction_change - || update_flags->bits.lut_3d - || update_flags->bits.gamma_change - || update_flags->bits.gamut_remap_change) { + if (update_flags->bits.lut_3d) { type = UPDATE_TYPE_FULL; elevate_update_type(&overall_type, type); } + if (dc->debug.enable_legacy_fast_update && + (update_flags->bits.gamma_change || + update_flags->bits.gamut_remap_change || + update_flags->bits.input_csc_change || + update_flags->bits.coeff_reduction_change)) { + type = UPDATE_TYPE_FULL; + elevate_update_type(&overall_type, type); + } return overall_type; } @@ -2585,7 +2634,7 @@ static enum surface_update_type check_update_surfaces_for_stream( stream_update->integer_scaling_update) su_flags->bits.scaling = 1; - if (stream_update->out_transfer_func) + if (dc->debug.enable_legacy_fast_update && stream_update->out_transfer_func) su_flags->bits.out_tf = 1; if (stream_update->abm_level) @@ -2605,14 +2654,23 @@ static enum surface_update_type check_update_surfaces_for_stream( if (stream_update->mst_bw_update) su_flags->bits.mst_bw = 1; - if (stream_update->crtc_timing_adjust && dc_extended_blank_supported(dc)) - su_flags->bits.crtc_timing_adjust = 1; + + if (stream_update->stream && stream_update->stream->freesync_on_desktop && + (stream_update->vrr_infopacket || stream_update->allow_freesync || + stream_update->vrr_active_variable || stream_update->vrr_active_fixed)) + su_flags->bits.fams_changed = 1; if (su_flags->raw != 0) overall_type = UPDATE_TYPE_FULL; if (stream_update->output_csc_transform || stream_update->output_color_space) su_flags->bits.out_csc = 1; + + /* Output transfer function changes do not require bandwidth recalculation, + * so don't trigger a full update + */ + if (!dc->debug.enable_legacy_fast_update && stream_update->out_transfer_func) + su_flags->bits.out_tf = 1; } for (i = 0 ; i < surface_count; i++) { @@ -2965,6 +3023,9 @@ static void copy_stream_update_to_stream(struct dc *dc, if (update->vrr_active_variable) stream->vrr_active_variable = *update->vrr_active_variable; + if (update->vrr_active_fixed) + stream->vrr_active_fixed = *update->vrr_active_fixed; + if (update->crtc_timing_adjust) stream->adjust = *update->crtc_timing_adjust; @@ -3269,6 +3330,13 @@ static void commit_planes_do_stream_update(struct dc *dc, dc->hwss.prepare_bandwidth(dc, dc->current_state); dc->link_srv->set_dpms_on(dc->current_state, pipe_ctx); } + } else if (pipe_ctx->stream->link->wa_flags.blank_stream_on_ocs_change && stream_update->output_color_space + && !stream->dpms_off && dc_is_dp_signal(pipe_ctx->stream->signal)) { + /* + * Workaround for firmware issue in some receivers where they don't pick up + * correct output color space unless DP link is disabled/re-enabled + */ + dc->link_srv->set_dpms_on(dc->current_state, pipe_ctx); } if (stream_update->abm_level && pipe_ctx->stream_res.abm) { @@ -3309,7 +3377,6 @@ void dc_dmub_update_dirty_rect(struct dc *dc, struct dc_state *context) { union dmub_rb_cmd cmd; - struct dc_context *dc_ctx = dc->ctx; struct dmub_cmd_update_dirty_rect_data *update_dirty_rect; unsigned int i, j; unsigned int panel_inst = 0; @@ -3350,10 +3417,166 @@ void dc_dmub_update_dirty_rect(struct dc *dc, update_dirty_rect->panel_inst = panel_inst; update_dirty_rect->pipe_idx = j; - dc_dmub_srv_cmd_queue(dc_ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc_ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_NO_WAIT); + } + } +} + +static void build_dmub_update_dirty_rect( + struct dc *dc, + int surface_count, + struct dc_stream_state *stream, + struct dc_surface_update *srf_updates, + struct dc_state *context, + struct dc_dmub_cmd dc_dmub_cmd[], + unsigned int *dmub_cmd_count) +{ + union dmub_rb_cmd cmd; + struct dmub_cmd_update_dirty_rect_data *update_dirty_rect; + unsigned int i, j; + unsigned int panel_inst = 0; + + if (!dc_dmub_should_send_dirty_rect_cmd(dc, stream)) + return; + + if (!dc_get_edp_link_panel_inst(dc, stream->link, &panel_inst)) + return; + + memset(&cmd, 0x0, sizeof(cmd)); + cmd.update_dirty_rect.header.type = DMUB_CMD__UPDATE_DIRTY_RECT; + cmd.update_dirty_rect.header.sub_type = 0; + cmd.update_dirty_rect.header.payload_bytes = + sizeof(cmd.update_dirty_rect) - + sizeof(cmd.update_dirty_rect.header); + update_dirty_rect = &cmd.update_dirty_rect.update_dirty_rect_data; + for (i = 0; i < surface_count; i++) { + struct dc_plane_state *plane_state = srf_updates[i].surface; + const struct dc_flip_addrs *flip_addr = srf_updates[i].flip_addr; + + if (!srf_updates[i].surface || !flip_addr) + continue; + /* Do not send in immediate flip mode */ + if (srf_updates[i].surface->flip_immediate) + continue; + update_dirty_rect->cmd_version = DMUB_CMD_PSR_CONTROL_VERSION_1; + update_dirty_rect->dirty_rect_count = flip_addr->dirty_rect_count; + memcpy(update_dirty_rect->src_dirty_rects, flip_addr->dirty_rects, + sizeof(flip_addr->dirty_rects)); + for (j = 0; j < dc->res_pool->pipe_count; j++) { + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j]; + + if (pipe_ctx->stream != stream) + continue; + if (pipe_ctx->plane_state != plane_state) + continue; + update_dirty_rect->panel_inst = panel_inst; + update_dirty_rect->pipe_idx = j; + dc_dmub_cmd[*dmub_cmd_count].dmub_cmd = cmd; + dc_dmub_cmd[*dmub_cmd_count].wait_type = DM_DMUB_WAIT_TYPE_NO_WAIT; + (*dmub_cmd_count)++; + } + } +} + + +/** + * build_dmub_cmd_list() - Build an array of DMCUB commands to be sent to DMCUB + * + * @dc: Current DC state + * @srf_updates: Array of surface updates + * @surface_count: Number of surfaces that have an updated + * @stream: Corresponding stream to be updated in the current flip + * @context: New DC state to be programmed + * + * @dc_dmub_cmd: Array of DMCUB commands to be sent to DMCUB + * @dmub_cmd_count: Count indicating the number of DMCUB commands in dc_dmub_cmd array + * + * This function builds an array of DMCUB commands to be sent to DMCUB. This function is required + * to build an array of commands and have them sent while the OTG lock is acquired. + * + * Return: void + */ +static void build_dmub_cmd_list(struct dc *dc, + struct dc_surface_update *srf_updates, + int surface_count, + struct dc_stream_state *stream, + struct dc_state *context, + struct dc_dmub_cmd dc_dmub_cmd[], + unsigned int *dmub_cmd_count) +{ + // Initialize cmd count to 0 + *dmub_cmd_count = 0; + build_dmub_update_dirty_rect(dc, surface_count, stream, srf_updates, context, dc_dmub_cmd, dmub_cmd_count); +} + +static void commit_planes_for_stream_fast(struct dc *dc, + struct dc_surface_update *srf_updates, + int surface_count, + struct dc_stream_state *stream, + struct dc_stream_update *stream_update, + enum surface_update_type update_type, + struct dc_state *context) +{ + int i, j; + struct pipe_ctx *top_pipe_to_program = NULL; + dc_z10_restore(dc); + + for (j = 0; j < dc->res_pool->pipe_count; j++) { + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j]; + + if (!pipe_ctx->top_pipe && + !pipe_ctx->prev_odm_pipe && + pipe_ctx->stream && + pipe_ctx->stream == stream) { + top_pipe_to_program = pipe_ctx; + } + } + + if (dc->debug.visual_confirm) { + for (i = 0; i < dc->res_pool->pipe_count; i++) { + struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; + + if (pipe->stream && pipe->plane_state) + dc_update_viusal_confirm_color(dc, context, pipe); + } + } + + for (i = 0; i < surface_count; i++) { + struct dc_plane_state *plane_state = srf_updates[i].surface; + /*set logical flag for lock/unlock use*/ + for (j = 0; j < dc->res_pool->pipe_count; j++) { + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j]; + + if (!pipe_ctx->plane_state) + continue; + if (should_update_pipe_for_plane(context, pipe_ctx, plane_state)) + continue; + pipe_ctx->plane_state->triplebuffer_flips = false; + if (update_type == UPDATE_TYPE_FAST && + dc->hwss.program_triplebuffer && + !pipe_ctx->plane_state->flip_immediate && dc->debug.enable_tri_buf) { + /*triple buffer for VUpdate only*/ + pipe_ctx->plane_state->triplebuffer_flips = true; + } } } + + build_dmub_cmd_list(dc, + srf_updates, + surface_count, + stream, + context, + context->dc_dmub_cmd, + &(context->dmub_cmd_count)); + hwss_build_fast_sequence(dc, + context->dc_dmub_cmd, + context->dmub_cmd_count, + context->block_sequence, + &(context->block_sequence_steps), + top_pipe_to_program); + hwss_execute_sequence(dc, + context->block_sequence, + context->block_sequence_steps); } static void commit_planes_for_stream(struct dc *dc, @@ -3393,21 +3616,6 @@ static void commit_planes_for_stream(struct dc *dc, } } - if (get_seamless_boot_stream_count(context) > 0 && surface_count > 0) { - /* Optimize seamless boot flag keeps clocks and watermarks high until - * first flip. After first flip, optimization is required to lower - * bandwidth. Important to note that it is expected UEFI will - * only light up a single display on POST, therefore we only expect - * one stream with seamless boot flag set. - */ - if (stream->apply_seamless_boot_optimization) { - stream->apply_seamless_boot_optimization = false; - - if (get_seamless_boot_stream_count(context) == 0) - dc->optimized_required = true; - } - } - if (update_type == UPDATE_TYPE_FULL) { dc_allow_idle_optimizations(dc, false); @@ -3449,6 +3657,14 @@ static void commit_planes_for_stream(struct dc *dc, } } + if (dc->debug.visual_confirm) + for (i = 0; i < dc->res_pool->pipe_count; i++) { + struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; + + if (pipe->stream && pipe->plane_state) + dc_update_viusal_confirm_color(dc, context, pipe); + } + if (stream->test_pattern.type != DP_TEST_PATTERN_VIDEO_MODE) { struct pipe_ctx *mpcc_pipe; struct pipe_ctx *odm_pipe; @@ -3532,43 +3748,40 @@ static void commit_planes_for_stream(struct dc *dc, for (j = 0; j < dc->res_pool->pipe_count; j++) { struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j]; - if (dc->debug.visual_confirm == VISUAL_CONFIRM_SUBVP && + if ((dc->debug.visual_confirm == VISUAL_CONFIRM_SUBVP || + dc->debug.visual_confirm == VISUAL_CONFIRM_MCLK_SWITCH) && pipe_ctx->stream && pipe_ctx->plane_state) { - /* Only update visual confirm for SUBVP here. + /* Only update visual confirm for SUBVP and Mclk switching here. * The bar appears on all pipes, so we need to update the bar on all displays, * so the information doesn't get stale. */ - struct mpcc_blnd_cfg blnd_cfg = { 0 }; - - dc->hwss.update_visual_confirm_color(dc, pipe_ctx, &blnd_cfg.black_color, + dc->hwss.update_visual_confirm_color(dc, pipe_ctx, pipe_ctx->plane_res.hubp->inst); } } } - if (!IS_DIAG_DC(dc->ctx->dce_environment)) { - for (i = 0; i < surface_count; i++) { - struct dc_plane_state *plane_state = srf_updates[i].surface; - /*set logical flag for lock/unlock use*/ - for (j = 0; j < dc->res_pool->pipe_count; j++) { - struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j]; - if (!pipe_ctx->plane_state) - continue; - if (should_update_pipe_for_plane(context, pipe_ctx, plane_state)) - continue; - pipe_ctx->plane_state->triplebuffer_flips = false; - if (update_type == UPDATE_TYPE_FAST && - dc->hwss.program_triplebuffer != NULL && - !pipe_ctx->plane_state->flip_immediate && dc->debug.enable_tri_buf) { - /*triple buffer for VUpdate only*/ - pipe_ctx->plane_state->triplebuffer_flips = true; - } - } - if (update_type == UPDATE_TYPE_FULL) { - /* force vsync flip when reconfiguring pipes to prevent underflow */ - plane_state->flip_immediate = false; + for (i = 0; i < surface_count; i++) { + struct dc_plane_state *plane_state = srf_updates[i].surface; + /*set logical flag for lock/unlock use*/ + for (j = 0; j < dc->res_pool->pipe_count; j++) { + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j]; + if (!pipe_ctx->plane_state) + continue; + if (should_update_pipe_for_plane(context, pipe_ctx, plane_state)) + continue; + pipe_ctx->plane_state->triplebuffer_flips = false; + if (update_type == UPDATE_TYPE_FAST && + dc->hwss.program_triplebuffer != NULL && + !pipe_ctx->plane_state->flip_immediate && dc->debug.enable_tri_buf) { + /*triple buffer for VUpdate only*/ + pipe_ctx->plane_state->triplebuffer_flips = true; } } + if (update_type == UPDATE_TYPE_FULL) { + /* force vsync flip when reconfiguring pipes to prevent underflow */ + plane_state->flip_immediate = false; + } } // Update Type FULL, Surface updates @@ -3872,6 +4085,7 @@ static bool commit_minimal_transition_state(struct dc *dc, unsigned int i, j; unsigned int pipe_in_use = 0; bool subvp_in_use = false; + bool odm_in_use = false; if (!transition_context) return false; @@ -3900,6 +4114,18 @@ static bool commit_minimal_transition_state(struct dc *dc, } } + /* If ODM is enabled and we are adding or removing planes from any ODM + * pipe, we must use the minimal transition. + */ + for (i = 0; i < dc->res_pool->pipe_count; i++) { + struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i]; + + if (pipe->stream && pipe->next_odm_pipe) { + odm_in_use = true; + break; + } + } + /* When the OS add a new surface if we have been used all of pipes with odm combine * and mpc split feature, it need use commit_minimal_transition_state to transition safely. * After OS exit MPO, it will back to use odm and mpc split with all of pipes, we need @@ -3908,7 +4134,7 @@ static bool commit_minimal_transition_state(struct dc *dc, * Reduce the scenarios to use dc_commit_state_no_check in the stage of flip. Especially * enter/exit MPO when DCN still have enough resources. */ - if (pipe_in_use != dc->res_pool->pipe_count && !subvp_in_use) { + if (pipe_in_use != dc->res_pool->pipe_count && !subvp_in_use && !odm_in_use) { dc_release_state(transition_context); return true; } @@ -3972,6 +4198,41 @@ static bool commit_minimal_transition_state(struct dc *dc, return true; } +/** + * update_seamless_boot_flags() - Helper function for updating seamless boot flags + * + * @dc: Current DC state + * @context: New DC state to be programmed + * @surface_count: Number of surfaces that have an updated + * @stream: Corresponding stream to be updated in the current flip + * + * Updating seamless boot flags do not need to be part of the commit sequence. This + * helper function will update the seamless boot flags on each flip (if required) + * outside of the HW commit sequence (fast or slow). + * + * Return: void + */ +static void update_seamless_boot_flags(struct dc *dc, + struct dc_state *context, + int surface_count, + struct dc_stream_state *stream) +{ + if (get_seamless_boot_stream_count(context) > 0 && surface_count > 0) { + /* Optimize seamless boot flag keeps clocks and watermarks high until + * first flip. After first flip, optimization is required to lower + * bandwidth. Important to note that it is expected UEFI will + * only light up a single display on POST, therefore we only expect + * one stream with seamless boot flag set. + */ + if (stream->apply_seamless_boot_optimization) { + stream->apply_seamless_boot_optimization = false; + + if (get_seamless_boot_stream_count(context) == 0) + dc->optimized_required = true; + } + } +} + bool dc_update_planes_and_stream(struct dc *dc, struct dc_surface_update *srf_updates, int surface_count, struct dc_stream_state *stream, @@ -4038,14 +4299,25 @@ bool dc_update_planes_and_stream(struct dc *dc, update_type = UPDATE_TYPE_FULL; } - commit_planes_for_stream( - dc, - srf_updates, - surface_count, - stream, - stream_update, - update_type, - context); + update_seamless_boot_flags(dc, context, surface_count, stream); + if (!dc->debug.enable_legacy_fast_update && update_type == UPDATE_TYPE_FAST) { + commit_planes_for_stream_fast(dc, + srf_updates, + surface_count, + stream, + stream_update, + update_type, + context); + } else { + commit_planes_for_stream( + dc, + srf_updates, + surface_count, + stream, + stream_update, + update_type, + context); + } if (dc->current_state != context) { @@ -4170,7 +4442,17 @@ void dc_commit_updates_for_stream(struct dc *dc, TRACE_DC_PIPE_STATE(pipe_ctx, i, MAX_PIPES); - commit_planes_for_stream( + update_seamless_boot_flags(dc, context, surface_count, stream); + if (!dc->debug.enable_legacy_fast_update && update_type == UPDATE_TYPE_FAST) { + commit_planes_for_stream_fast(dc, + srf_updates, + surface_count, + stream, + stream_update, + update_type, + context); + } else { + commit_planes_for_stream( dc, srf_updates, surface_count, @@ -4178,6 +4460,7 @@ void dc_commit_updates_for_stream(struct dc *dc, stream_update, update_type, context); + } /*update current_State*/ if (dc->current_state != context) { @@ -4264,9 +4547,6 @@ void dc_set_power_state( dc_z10_restore(dc); - if (dc->ctx->dmub_srv) - dc_dmub_srv_wait_phy_init(dc->ctx->dmub_srv); - dc->hwss.init_hw(dc); if (dc->hwss.init_sys_ctx != NULL && @@ -4606,7 +4886,6 @@ bool dc_process_dmub_aux_transfer_async(struct dc *dc, { uint8_t action; union dmub_rb_cmd cmd = {0}; - struct dc_dmub_srv *dmub_srv = dc->ctx->dmub_srv; ASSERT(payload->length <= 16); @@ -4654,9 +4933,7 @@ bool dc_process_dmub_aux_transfer_async(struct dc *dc, ); } - dc_dmub_srv_cmd_queue(dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dmub_srv); - dc_dmub_srv_wait_idle(dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); return true; } @@ -4700,7 +4977,6 @@ bool dc_process_dmub_set_config_async(struct dc *dc, struct dmub_notification *notify) { union dmub_rb_cmd cmd = {0}; - struct dc_dmub_srv *dmub_srv = dc->ctx->dmub_srv; bool is_cmd_complete = true; /* prepare SET_CONFIG command */ @@ -4711,7 +4987,7 @@ bool dc_process_dmub_set_config_async(struct dc *dc, cmd.set_config_access.set_config_control.cmd_pkt.msg_type = payload->msg_type; cmd.set_config_access.set_config_control.cmd_pkt.msg_data = payload->msg_data; - if (!dc_dmub_srv_cmd_with_reply_data(dmub_srv, &cmd)) { + if (!dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)) { /* command is not processed by dmub */ notify->sc_status = SET_CONFIG_UNKNOWN_ERROR; return is_cmd_complete; @@ -4746,7 +5022,6 @@ enum dc_status dc_process_dmub_set_mst_slots(const struct dc *dc, uint8_t *mst_slots_in_use) { union dmub_rb_cmd cmd = {0}; - struct dc_dmub_srv *dmub_srv = dc->ctx->dmub_srv; /* prepare MST_ALLOC_SLOTS command */ cmd.set_mst_alloc_slots.header.type = DMUB_CMD__DPIA; @@ -4755,7 +5030,7 @@ enum dc_status dc_process_dmub_set_mst_slots(const struct dc *dc, cmd.set_mst_alloc_slots.mst_slots_control.instance = dc->links[link_index]->ddc_hw_inst; cmd.set_mst_alloc_slots.mst_slots_control.mst_alloc_slots = mst_alloc_slots; - if (!dc_dmub_srv_cmd_with_reply_data(dmub_srv, &cmd)) + if (!dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)) /* command is not processed by dmub */ return DC_ERROR_UNEXPECTED; @@ -4789,19 +5064,28 @@ void dc_process_dmub_dpia_hpd_int_enable(const struct dc *dc, uint32_t hpd_int_enable) { union dmub_rb_cmd cmd = {0}; - struct dc_dmub_srv *dmub_srv = dc->ctx->dmub_srv; cmd.dpia_hpd_int_enable.header.type = DMUB_CMD__DPIA_HPD_INT_ENABLE; cmd.dpia_hpd_int_enable.enable = hpd_int_enable; - dc_dmub_srv_cmd_queue(dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dmub_srv); - dc_dmub_srv_wait_idle(dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); DC_LOG_DEBUG("%s: hpd_int_enable(%d)\n", __func__, hpd_int_enable); } /** + * dc_print_dmub_diagnostic_data - Print DMUB diagnostic data for debugging + * + * @dc: [in] dc structure + * + * + */ +void dc_print_dmub_diagnostic_data(const struct dc *dc) +{ + dc_dmub_srv_log_diagnostic_data(dc->ctx->dmub_srv); +} + +/** * dc_disable_accelerated_mode - disable accelerated mode * @dc: dc structure */ @@ -4860,21 +5144,3 @@ void dc_notify_vsync_int_state(struct dc *dc, struct dc_stream_state *stream, bo if (pipe->stream_res.abm && pipe->stream_res.abm->funcs->set_abm_pause) pipe->stream_res.abm->funcs->set_abm_pause(pipe->stream_res.abm, !enable, i, pipe->stream_res.tg->inst); } - -/** - * dc_extended_blank_supported - Decide whether extended blank is supported - * - * @dc: [in] Current DC state - * - * Extended blank is a freesync optimization feature to be enabled in the - * future. During the extra vblank period gained from freesync, we have the - * ability to enter z9/z10. - * - * Return: - * Indicate whether extended blank is supported (%true or %false) - */ -bool dc_extended_blank_supported(struct dc *dc) -{ - return dc->debug.extended_blank_optimization && !dc->debug.disable_z10 - && dc->caps.zstate_support && dc->caps.is_apu; -} diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c index 2acbf692193f..d7d00fefaab9 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c @@ -27,6 +27,8 @@ #include "core_types.h" #include "timing_generator.h" #include "hw_sequencer.h" +#include "hw_sequencer_private.h" +#include "basics/dc_common.h" #define NUM_ELEMENTS(a) (sizeof(a) / sizeof((a)[0])) @@ -421,6 +423,7 @@ void get_hdr_visual_confirm_color( void get_subvp_visual_confirm_color( struct dc *dc, + struct dc_state *context, struct pipe_ctx *pipe_ctx, struct tg_color *color) { @@ -428,15 +431,17 @@ void get_subvp_visual_confirm_color( bool enable_subvp = false; int i; - if (!dc->ctx || !dc->ctx->dmub_srv || !pipe_ctx) + if (!dc->ctx || !dc->ctx->dmub_srv || !pipe_ctx || !context) return; for (i = 0; i < dc->res_pool->pipe_count; i++) { - struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i]; + struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; if (pipe->stream && pipe->stream->mall_stream_config.paired_stream && pipe->stream->mall_stream_config.type == SUBVP_MAIN) { /* SubVP enable - red */ + color->color_g_y = 0; + color->color_b_cb = 0; color->color_r_cr = color_value; enable_subvp = true; @@ -448,12 +453,374 @@ void get_subvp_visual_confirm_color( if (enable_subvp && pipe_ctx->stream->mall_stream_config.type == SUBVP_NONE) { color->color_r_cr = 0; - if (pipe_ctx->stream->ignore_msa_timing_param == 1) + if (pipe_ctx->stream->allow_freesync == 1) { /* SubVP enable and DRR on - green */ + color->color_b_cb = 0; color->color_g_y = color_value; - else + } else { /* SubVP enable and No DRR - blue */ + color->color_g_y = 0; color->color_b_cb = color_value; + } + } +} + +void hwss_build_fast_sequence(struct dc *dc, + struct dc_dmub_cmd *dc_dmub_cmd, + unsigned int dmub_cmd_count, + struct block_sequence block_sequence[], + int *num_steps, + struct pipe_ctx *pipe_ctx) +{ + struct dc_plane_state *plane = pipe_ctx->plane_state; + struct dc_stream_state *stream = pipe_ctx->stream; + struct dce_hwseq *hws = dc->hwseq; + struct pipe_ctx *current_pipe = NULL; + struct pipe_ctx *current_mpc_pipe = NULL; + unsigned int i = 0; + + *num_steps = 0; // Initialize to 0 + + if (!plane || !stream) + return; + + if (dc->hwss.subvp_pipe_control_lock_fast) { + block_sequence[*num_steps].params.subvp_pipe_control_lock_fast_params.dc = dc; + block_sequence[*num_steps].params.subvp_pipe_control_lock_fast_params.lock = true; + block_sequence[*num_steps].params.subvp_pipe_control_lock_fast_params.pipe_ctx = pipe_ctx; + block_sequence[*num_steps].func = DMUB_SUBVP_PIPE_CONTROL_LOCK_FAST; + (*num_steps)++; + } + if (dc->hwss.pipe_control_lock) { + block_sequence[*num_steps].params.pipe_control_lock_params.dc = dc; + block_sequence[*num_steps].params.pipe_control_lock_params.lock = true; + block_sequence[*num_steps].params.pipe_control_lock_params.pipe_ctx = pipe_ctx; + block_sequence[*num_steps].func = OPTC_PIPE_CONTROL_LOCK; + (*num_steps)++; + } + + for (i = 0; i < dmub_cmd_count; i++) { + block_sequence[*num_steps].params.send_dmcub_cmd_params.ctx = dc->ctx; + block_sequence[*num_steps].params.send_dmcub_cmd_params.cmd = &(dc_dmub_cmd[i].dmub_cmd); + block_sequence[*num_steps].params.send_dmcub_cmd_params.wait_type = dc_dmub_cmd[i].wait_type; + block_sequence[*num_steps].func = DMUB_SEND_DMCUB_CMD; + (*num_steps)++; + } + + current_pipe = pipe_ctx; + while (current_pipe) { + current_mpc_pipe = current_pipe; + while (current_mpc_pipe) { + if (dc->hwss.set_flip_control_gsl && current_mpc_pipe->plane_state && current_mpc_pipe->plane_state->update_flags.raw) { + block_sequence[*num_steps].params.set_flip_control_gsl_params.pipe_ctx = current_mpc_pipe; + block_sequence[*num_steps].params.set_flip_control_gsl_params.flip_immediate = current_mpc_pipe->plane_state->flip_immediate; + block_sequence[*num_steps].func = HUBP_SET_FLIP_CONTROL_GSL; + (*num_steps)++; + } + if (dc->hwss.program_triplebuffer && dc->debug.enable_tri_buf && current_mpc_pipe->plane_state->update_flags.raw) { + block_sequence[*num_steps].params.program_triplebuffer_params.dc = dc; + block_sequence[*num_steps].params.program_triplebuffer_params.pipe_ctx = current_mpc_pipe; + block_sequence[*num_steps].params.program_triplebuffer_params.enableTripleBuffer = current_mpc_pipe->plane_state->triplebuffer_flips; + block_sequence[*num_steps].func = HUBP_PROGRAM_TRIPLEBUFFER; + (*num_steps)++; + } + if (dc->hwss.update_plane_addr && current_mpc_pipe->plane_state->update_flags.bits.addr_update) { + block_sequence[*num_steps].params.update_plane_addr_params.dc = dc; + block_sequence[*num_steps].params.update_plane_addr_params.pipe_ctx = current_mpc_pipe; + block_sequence[*num_steps].func = HUBP_UPDATE_PLANE_ADDR; + (*num_steps)++; + } + + if (hws->funcs.set_input_transfer_func && current_mpc_pipe->plane_state->update_flags.bits.gamma_change) { + block_sequence[*num_steps].params.set_input_transfer_func_params.dc = dc; + block_sequence[*num_steps].params.set_input_transfer_func_params.pipe_ctx = current_mpc_pipe; + block_sequence[*num_steps].params.set_input_transfer_func_params.plane_state = current_mpc_pipe->plane_state; + block_sequence[*num_steps].func = DPP_SET_INPUT_TRANSFER_FUNC; + (*num_steps)++; + } + + if (dc->hwss.program_gamut_remap && current_mpc_pipe->plane_state->update_flags.bits.gamut_remap_change) { + block_sequence[*num_steps].params.program_gamut_remap_params.pipe_ctx = current_mpc_pipe; + block_sequence[*num_steps].func = DPP_PROGRAM_GAMUT_REMAP; + (*num_steps)++; + } + if (current_mpc_pipe->plane_state->update_flags.bits.input_csc_change) { + block_sequence[*num_steps].params.setup_dpp_params.pipe_ctx = current_mpc_pipe; + block_sequence[*num_steps].func = DPP_SETUP_DPP; + (*num_steps)++; + } + if (current_mpc_pipe->plane_state->update_flags.bits.coeff_reduction_change) { + block_sequence[*num_steps].params.program_bias_and_scale_params.pipe_ctx = current_mpc_pipe; + block_sequence[*num_steps].func = DPP_PROGRAM_BIAS_AND_SCALE; + (*num_steps)++; + } + if (hws->funcs.set_output_transfer_func && current_mpc_pipe->stream->update_flags.bits.out_tf) { + block_sequence[*num_steps].params.set_output_transfer_func_params.dc = dc; + block_sequence[*num_steps].params.set_output_transfer_func_params.pipe_ctx = current_mpc_pipe; + block_sequence[*num_steps].params.set_output_transfer_func_params.stream = current_mpc_pipe->stream; + block_sequence[*num_steps].func = DPP_SET_OUTPUT_TRANSFER_FUNC; + (*num_steps)++; + } + + if (current_mpc_pipe->stream->update_flags.bits.out_csc) { + block_sequence[*num_steps].params.power_on_mpc_mem_pwr_params.mpc = dc->res_pool->mpc; + block_sequence[*num_steps].params.power_on_mpc_mem_pwr_params.mpcc_id = current_mpc_pipe->plane_res.hubp->inst; + block_sequence[*num_steps].params.power_on_mpc_mem_pwr_params.power_on = true; + block_sequence[*num_steps].func = MPC_POWER_ON_MPC_MEM_PWR; + (*num_steps)++; + + if (current_mpc_pipe->stream->csc_color_matrix.enable_adjustment == true) { + block_sequence[*num_steps].params.set_output_csc_params.mpc = dc->res_pool->mpc; + block_sequence[*num_steps].params.set_output_csc_params.opp_id = current_mpc_pipe->stream_res.opp->inst; + block_sequence[*num_steps].params.set_output_csc_params.regval = current_mpc_pipe->stream->csc_color_matrix.matrix; + block_sequence[*num_steps].params.set_output_csc_params.ocsc_mode = MPC_OUTPUT_CSC_COEF_A; + block_sequence[*num_steps].func = MPC_SET_OUTPUT_CSC; + (*num_steps)++; + } else { + block_sequence[*num_steps].params.set_ocsc_default_params.mpc = dc->res_pool->mpc; + block_sequence[*num_steps].params.set_ocsc_default_params.opp_id = current_mpc_pipe->stream_res.opp->inst; + block_sequence[*num_steps].params.set_ocsc_default_params.color_space = current_mpc_pipe->stream->output_color_space; + block_sequence[*num_steps].params.set_ocsc_default_params.ocsc_mode = MPC_OUTPUT_CSC_COEF_A; + block_sequence[*num_steps].func = MPC_SET_OCSC_DEFAULT; + (*num_steps)++; + } + } + current_mpc_pipe = current_mpc_pipe->bottom_pipe; + } + current_pipe = current_pipe->next_odm_pipe; + } + + if (dc->hwss.pipe_control_lock) { + block_sequence[*num_steps].params.pipe_control_lock_params.dc = dc; + block_sequence[*num_steps].params.pipe_control_lock_params.lock = false; + block_sequence[*num_steps].params.pipe_control_lock_params.pipe_ctx = pipe_ctx; + block_sequence[*num_steps].func = OPTC_PIPE_CONTROL_LOCK; + (*num_steps)++; + } + if (dc->hwss.subvp_pipe_control_lock_fast) { + block_sequence[*num_steps].params.subvp_pipe_control_lock_fast_params.dc = dc; + block_sequence[*num_steps].params.subvp_pipe_control_lock_fast_params.lock = false; + block_sequence[*num_steps].params.subvp_pipe_control_lock_fast_params.pipe_ctx = pipe_ctx; + block_sequence[*num_steps].func = DMUB_SUBVP_PIPE_CONTROL_LOCK_FAST; + (*num_steps)++; + } + + current_pipe = pipe_ctx; + while (current_pipe) { + current_mpc_pipe = current_pipe; + + while (current_mpc_pipe) { + if (!current_mpc_pipe->bottom_pipe && !pipe_ctx->next_odm_pipe && + current_mpc_pipe->stream && current_mpc_pipe->plane_state && + current_mpc_pipe->plane_state->update_flags.bits.addr_update && + !current_mpc_pipe->plane_state->skip_manual_trigger) { + block_sequence[*num_steps].params.program_manual_trigger_params.pipe_ctx = current_mpc_pipe; + block_sequence[*num_steps].func = OPTC_PROGRAM_MANUAL_TRIGGER; + (*num_steps)++; + } + current_mpc_pipe = current_mpc_pipe->bottom_pipe; + } + current_pipe = current_pipe->next_odm_pipe; + } +} + +void hwss_execute_sequence(struct dc *dc, + struct block_sequence block_sequence[], + int num_steps) +{ + unsigned int i; + union block_sequence_params *params; + struct dce_hwseq *hws = dc->hwseq; + + for (i = 0; i < num_steps; i++) { + params = &(block_sequence[i].params); + switch (block_sequence[i].func) { + + case DMUB_SUBVP_PIPE_CONTROL_LOCK_FAST: + dc->hwss.subvp_pipe_control_lock_fast(params); + break; + case OPTC_PIPE_CONTROL_LOCK: + dc->hwss.pipe_control_lock(params->pipe_control_lock_params.dc, + params->pipe_control_lock_params.pipe_ctx, + params->pipe_control_lock_params.lock); + break; + case HUBP_SET_FLIP_CONTROL_GSL: + dc->hwss.set_flip_control_gsl(params->set_flip_control_gsl_params.pipe_ctx, + params->set_flip_control_gsl_params.flip_immediate); + break; + case HUBP_PROGRAM_TRIPLEBUFFER: + dc->hwss.program_triplebuffer(params->program_triplebuffer_params.dc, + params->program_triplebuffer_params.pipe_ctx, + params->program_triplebuffer_params.enableTripleBuffer); + break; + case HUBP_UPDATE_PLANE_ADDR: + dc->hwss.update_plane_addr(params->update_plane_addr_params.dc, + params->update_plane_addr_params.pipe_ctx); + break; + case DPP_SET_INPUT_TRANSFER_FUNC: + hws->funcs.set_input_transfer_func(params->set_input_transfer_func_params.dc, + params->set_input_transfer_func_params.pipe_ctx, + params->set_input_transfer_func_params.plane_state); + break; + case DPP_PROGRAM_GAMUT_REMAP: + dc->hwss.program_gamut_remap(params->program_gamut_remap_params.pipe_ctx); + break; + case DPP_SETUP_DPP: + hwss_setup_dpp(params); + break; + case DPP_PROGRAM_BIAS_AND_SCALE: + hwss_program_bias_and_scale(params); + break; + case OPTC_PROGRAM_MANUAL_TRIGGER: + hwss_program_manual_trigger(params); + break; + case DPP_SET_OUTPUT_TRANSFER_FUNC: + hws->funcs.set_output_transfer_func(params->set_output_transfer_func_params.dc, + params->set_output_transfer_func_params.pipe_ctx, + params->set_output_transfer_func_params.stream); + break; + case MPC_UPDATE_VISUAL_CONFIRM: + dc->hwss.update_visual_confirm_color(params->update_visual_confirm_params.dc, + params->update_visual_confirm_params.pipe_ctx, + params->update_visual_confirm_params.mpcc_id); + break; + case MPC_POWER_ON_MPC_MEM_PWR: + hwss_power_on_mpc_mem_pwr(params); + break; + case MPC_SET_OUTPUT_CSC: + hwss_set_output_csc(params); + break; + case MPC_SET_OCSC_DEFAULT: + hwss_set_ocsc_default(params); + break; + case DMUB_SEND_DMCUB_CMD: + hwss_send_dmcub_cmd(params); + break; + default: + ASSERT(false); + break; + } + } +} + +void hwss_send_dmcub_cmd(union block_sequence_params *params) +{ + struct dc_context *ctx = params->send_dmcub_cmd_params.ctx; + union dmub_rb_cmd *cmd = params->send_dmcub_cmd_params.cmd; + enum dm_dmub_wait_type wait_type = params->send_dmcub_cmd_params.wait_type; + + dm_execute_dmub_cmd(ctx, cmd, wait_type); +} + +void hwss_program_manual_trigger(union block_sequence_params *params) +{ + struct pipe_ctx *pipe_ctx = params->program_manual_trigger_params.pipe_ctx; + + if (pipe_ctx->stream_res.tg->funcs->program_manual_trigger) + pipe_ctx->stream_res.tg->funcs->program_manual_trigger(pipe_ctx->stream_res.tg); +} + +void hwss_setup_dpp(union block_sequence_params *params) +{ + struct pipe_ctx *pipe_ctx = params->setup_dpp_params.pipe_ctx; + struct dpp *dpp = pipe_ctx->plane_res.dpp; + struct dc_plane_state *plane_state = pipe_ctx->plane_state; + + if (dpp && dpp->funcs->dpp_setup) { + // program the input csc + dpp->funcs->dpp_setup(dpp, + plane_state->format, + EXPANSION_MODE_ZERO, + plane_state->input_csc_color_matrix, + plane_state->color_space, + NULL); + } +} + +void hwss_program_bias_and_scale(union block_sequence_params *params) +{ + struct pipe_ctx *pipe_ctx = params->program_bias_and_scale_params.pipe_ctx; + struct dpp *dpp = pipe_ctx->plane_res.dpp; + struct dc_plane_state *plane_state = pipe_ctx->plane_state; + struct dc_bias_and_scale bns_params = {0}; + + //TODO :for CNVC set scale and bias registers if necessary + build_prescale_params(&bns_params, plane_state); + if (dpp->funcs->dpp_program_bias_and_scale) + dpp->funcs->dpp_program_bias_and_scale(dpp, &bns_params); +} + +void hwss_power_on_mpc_mem_pwr(union block_sequence_params *params) +{ + struct mpc *mpc = params->power_on_mpc_mem_pwr_params.mpc; + int mpcc_id = params->power_on_mpc_mem_pwr_params.mpcc_id; + bool power_on = params->power_on_mpc_mem_pwr_params.power_on; + + if (mpc->funcs->power_on_mpc_mem_pwr) + mpc->funcs->power_on_mpc_mem_pwr(mpc, mpcc_id, power_on); +} + +void hwss_set_output_csc(union block_sequence_params *params) +{ + struct mpc *mpc = params->set_output_csc_params.mpc; + int opp_id = params->set_output_csc_params.opp_id; + const uint16_t *matrix = params->set_output_csc_params.regval; + enum mpc_output_csc_mode ocsc_mode = params->set_output_csc_params.ocsc_mode; + + if (mpc->funcs->set_output_csc != NULL) + mpc->funcs->set_output_csc(mpc, + opp_id, + matrix, + ocsc_mode); +} + +void hwss_set_ocsc_default(union block_sequence_params *params) +{ + struct mpc *mpc = params->set_ocsc_default_params.mpc; + int opp_id = params->set_ocsc_default_params.opp_id; + enum dc_color_space colorspace = params->set_ocsc_default_params.color_space; + enum mpc_output_csc_mode ocsc_mode = params->set_ocsc_default_params.ocsc_mode; + + if (mpc->funcs->set_ocsc_default != NULL) + mpc->funcs->set_ocsc_default(mpc, + opp_id, + colorspace, + ocsc_mode); +} + +void get_mclk_switch_visual_confirm_color( + struct dc *dc, + struct dc_state *context, + struct pipe_ctx *pipe_ctx, + struct tg_color *color) +{ + uint32_t color_value = MAX_TG_COLOR_VALUE; + struct vba_vars_st *vba = &context->bw_ctx.dml.vba; + + if (!dc->ctx || !dc->ctx->dmub_srv || !pipe_ctx || !vba || !context) + return; + + if (vba->DRAMClockChangeSupport[vba->VoltageLevel][vba->maxMpcComb] != + dm_dram_clock_change_unsupported) { + /* MCLK switching is supported */ + if (!pipe_ctx->has_vactive_margin) { + /* In Vblank - yellow */ + color->color_r_cr = color_value; + color->color_g_y = color_value; + + if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) { + /* FPO + Vblank - cyan */ + color->color_r_cr = 0; + color->color_g_y = color_value; + color->color_b_cb = color_value; + } + } else { + /* In Vactive - pink */ + color->color_r_cr = color_value; + color->color_b_cb = color_value; + } + /* SubVP */ + get_subvp_visual_confirm_color(dc, context, pipe_ctx, color); } } diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c index 117d80cb36fb..2f3d9a698486 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c @@ -69,6 +69,7 @@ #include "../dcn32/dcn32_resource.h" #include "../dcn321/dcn321_resource.h" + #define DC_LOGGER_INIT(logger) enum dce_version resource_parse_asic_id(struct hw_asic_id asic_id) @@ -1446,6 +1447,26 @@ static int acquire_first_split_pipe( split_pipe->stream = stream; return i; + } else if (split_pipe->prev_odm_pipe && + split_pipe->prev_odm_pipe->plane_state == split_pipe->plane_state) { + split_pipe->prev_odm_pipe->next_odm_pipe = split_pipe->next_odm_pipe; + if (split_pipe->next_odm_pipe) + split_pipe->next_odm_pipe->prev_odm_pipe = split_pipe->prev_odm_pipe; + + if (split_pipe->prev_odm_pipe->plane_state) + resource_build_scaling_params(split_pipe->prev_odm_pipe); + + memset(split_pipe, 0, sizeof(*split_pipe)); + split_pipe->stream_res.tg = pool->timing_generators[i]; + split_pipe->plane_res.hubp = pool->hubps[i]; + split_pipe->plane_res.ipp = pool->ipps[i]; + split_pipe->plane_res.dpp = pool->dpps[i]; + split_pipe->stream_res.opp = pool->opps[i]; + split_pipe->plane_res.mpcc_inst = pool->dpps[i]->inst; + split_pipe->pipe_idx = i; + + split_pipe->stream = stream; + return i; } } return -1; @@ -1858,7 +1879,7 @@ bool dc_add_all_planes_for_stream( return add_all_planes_for_stream(dc, stream, &set, 1, context); } -bool is_timing_changed(struct dc_stream_state *cur_stream, +bool dc_is_timing_changed(struct dc_stream_state *cur_stream, struct dc_stream_state *new_stream) { if (cur_stream == NULL) @@ -1883,7 +1904,7 @@ static bool are_stream_backends_same( if (stream_a == NULL || stream_b == NULL) return false; - if (is_timing_changed(stream_a, stream_b)) + if (dc_is_timing_changed(stream_a, stream_b)) return false; if (stream_a->signal != stream_b->signal) @@ -3014,23 +3035,29 @@ static void set_avi_info_frame( hdmi_info.bits.S0_S1 = scan_type; /* C0, C1 : Colorimetry */ - if (color_space == COLOR_SPACE_YCBCR709 || - color_space == COLOR_SPACE_YCBCR709_LIMITED) + switch (color_space) { + case COLOR_SPACE_YCBCR709: + case COLOR_SPACE_YCBCR709_LIMITED: hdmi_info.bits.C0_C1 = COLORIMETRY_ITU709; - else if (color_space == COLOR_SPACE_YCBCR601 || - color_space == COLOR_SPACE_YCBCR601_LIMITED) + break; + case COLOR_SPACE_YCBCR601: + case COLOR_SPACE_YCBCR601_LIMITED: hdmi_info.bits.C0_C1 = COLORIMETRY_ITU601; - else { - hdmi_info.bits.C0_C1 = COLORIMETRY_NO_DATA; - } - if (color_space == COLOR_SPACE_2020_RGB_FULLRANGE || - color_space == COLOR_SPACE_2020_RGB_LIMITEDRANGE || - color_space == COLOR_SPACE_2020_YCBCR) { + break; + case COLOR_SPACE_2020_RGB_FULLRANGE: + case COLOR_SPACE_2020_RGB_LIMITEDRANGE: + case COLOR_SPACE_2020_YCBCR: hdmi_info.bits.EC0_EC2 = COLORIMETRYEX_BT2020RGBYCBCR; hdmi_info.bits.C0_C1 = COLORIMETRY_EXTENDED; - } else if (color_space == COLOR_SPACE_ADOBERGB) { + break; + case COLOR_SPACE_ADOBERGB: hdmi_info.bits.EC0_EC2 = COLORIMETRYEX_ADOBERGB; hdmi_info.bits.C0_C1 = COLORIMETRY_EXTENDED; + break; + case COLOR_SPACE_SRGB: + default: + hdmi_info.bits.C0_C1 = COLORIMETRY_NO_DATA; + break; } if (pixel_encoding && color_space == COLOR_SPACE_2020_YCBCR && @@ -3508,7 +3535,7 @@ bool pipe_need_reprogram( if (pipe_ctx_old->stream_res.stream_enc != pipe_ctx->stream_res.stream_enc) return true; - if (is_timing_changed(pipe_ctx_old->stream, pipe_ctx->stream)) + if (dc_is_timing_changed(pipe_ctx_old->stream, pipe_ctx->stream)) return true; if (pipe_ctx_old->stream->dpms_off != pipe_ctx->stream->dpms_off) diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c index 72b261ad9587..6e11d2b701f8 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c @@ -276,8 +276,8 @@ static void program_cursor_attributes( } dc->hwss.set_cursor_attribute(pipe_ctx); - - dc_send_update_cursor_info_to_dmu(pipe_ctx, i); + if (dc->ctx->dmub_srv) + dc_send_update_cursor_info_to_dmu(pipe_ctx, i); if (dc->hwss.set_cursor_sdr_white_level) dc->hwss.set_cursor_sdr_white_level(pipe_ctx); } @@ -396,8 +396,8 @@ static void program_cursor_position( } dc->hwss.set_cursor_position(pipe_ctx); - - dc_send_update_cursor_info_to_dmu(pipe_ctx, i); + if (dc->ctx->dmub_srv) + dc_send_update_cursor_info_to_dmu(pipe_ctx, i); } if (pipe_to_program) @@ -490,25 +490,6 @@ bool dc_stream_add_writeback(struct dc *dc, struct dwbc *dwb = dc->res_pool->dwbc[wb_info->dwb_pipe_inst]; dwb->otg_inst = stream_status->primary_otg_inst; } - if (IS_DIAG_DC(dc->ctx->dce_environment)) { - if (!dc->hwss.update_bandwidth(dc, dc->current_state)) { - dm_error("DC: update_bandwidth failed!\n"); - return false; - } - - /* enable writeback */ - if (dc->hwss.enable_writeback) { - struct dwbc *dwb = dc->res_pool->dwbc[wb_info->dwb_pipe_inst]; - - if (dwb->funcs->is_enabled(dwb)) { - /* writeback pipe already enabled, only need to update */ - dc->hwss.update_writeback(dc, wb_info, dc->current_state); - } else { - /* Enable writeback pipe from scratch*/ - dc->hwss.enable_writeback(dc, wb_info, dc->current_state); - } - } - } return true; } @@ -553,17 +534,6 @@ bool dc_stream_remove_writeback(struct dc *dc, } stream->num_wb_info = j; - if (IS_DIAG_DC(dc->ctx->dce_environment)) { - /* recalculate and apply DML parameters */ - if (!dc->hwss.update_bandwidth(dc, dc->current_state)) { - dm_error("DC: update_bandwidth failed!\n"); - return false; - } - - /* disable writeback */ - if (dc->hwss.disable_writeback) - dc->hwss.disable_writeback(dc, dwb_pipe_inst); - } return true; } diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h index 30f0ba05a6e6..26d05e225088 100644 --- a/drivers/gpu/drm/amd/display/dc/dc.h +++ b/drivers/gpu/drm/amd/display/dc/dc.h @@ -1,5 +1,5 @@ /* - * Copyright 2012-14 Advanced Micro Devices, Inc. + * Copyright 2012-2023 Advanced Micro Devices, Inc. * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), @@ -45,7 +45,7 @@ struct aux_payload; struct set_config_cmd_payload; struct dmub_notification; -#define DC_VER "3.2.230" +#define DC_VER "3.2.239" #define MAX_SURFACES 3 #define MAX_PLANES 6 @@ -60,7 +60,9 @@ struct dc_versions { }; enum dp_protocol_version { - DP_VERSION_1_4, + DP_VERSION_1_4 = 0, + DP_VERSION_2_1, + DP_VERSION_UNKNOWN, }; enum dc_plane_type { @@ -209,6 +211,8 @@ struct dc_color_caps { struct dc_dmub_caps { bool psr; bool mclk_sw; + bool subvp_psr; + bool gecc_enable; }; struct dc_caps { @@ -262,6 +266,7 @@ struct dc_caps { uint16_t subvp_pstate_allow_width_us; uint16_t subvp_vertical_int_margin_us; bool seamless_odm; + uint32_t max_v_total; uint8_t subvp_drr_vblank_start_margin_us; }; @@ -270,8 +275,13 @@ struct dc_bug_wa { bool dedcn20_305_wa; bool skip_clock_update; bool lt_early_cr_pattern; + struct { + uint8_t uclk : 1; + uint8_t fclk : 1; + uint8_t dcfclk : 1; + uint8_t dcfclk_ds: 1; + } clock_update_disable_mask; }; - struct dc_dcc_surface_param { struct dc_size surface_size; enum surface_pixel_format format; @@ -419,6 +429,7 @@ enum visual_confirm { VISUAL_CONFIRM_FAMS = 7, VISUAL_CONFIRM_SWIZZLE = 9, VISUAL_CONFIRM_SUBVP = 14, + VISUAL_CONFIRM_MCLK_SWITCH = 16, }; enum dc_psr_power_opts { @@ -698,6 +709,8 @@ struct dc_virtual_addr_space_config { struct dc_bounding_box_overrides { int sr_exit_time_ns; int sr_enter_plus_exit_time_ns; + int sr_exit_z8_time_ns; + int sr_enter_plus_exit_z8_time_ns; int urgent_latency_ns; int percent_of_ideal_drambw; int dram_clock_change_latency_ns; @@ -767,6 +780,8 @@ struct dc_debug_options { int sr_enter_plus_exit_time_dpm0_ns; int sr_exit_time_ns; int sr_enter_plus_exit_time_ns; + int sr_exit_z8_time_ns; + int sr_enter_plus_exit_z8_time_ns; int urgent_latency_ns; uint32_t underflow_assert_delay_us; int percent_of_ideal_drambw; @@ -855,7 +870,6 @@ struct dc_debug_options { bool force_usr_allow; /* uses value at boot and disables switch */ bool disable_dtb_ref_clk_switch; - uint32_t fixed_vs_aux_delay_config_wa; bool extended_blank_optimization; union aux_wake_wa_options aux_wake_wa; uint32_t mst_start_top_delay; @@ -879,6 +893,14 @@ struct dc_debug_options { uint32_t fpo_vactive_margin_us; bool disable_fpo_vactive; bool disable_boot_optimizations; + bool override_odm_optimization; + bool minimize_dispclk_using_odm; + bool disable_subvp_high_refresh; + bool disable_dp_plus_plus_wa; + uint32_t fpo_vactive_min_active_margin_us; + uint32_t fpo_vactive_max_blank_us; + bool enable_legacy_fast_update; + bool disable_dc_mode_overwrite; }; struct gpu_info_soc_bounding_box_v1_0; @@ -1502,6 +1524,7 @@ struct dc_link { /* Forced DPIA into TBT3 compatibility mode. */ bool dpia_forced_tbt3_mode; bool dongle_mode_timing_override; + bool blank_stream_on_ocs_change; } wa_flags; struct link_mst_stream_allocation_table mst_stream_alloc_table; @@ -2126,8 +2149,6 @@ struct dc_sink_init_data { bool converter_disable_audio; }; -bool dc_extended_blank_supported(struct dc *dc); - struct dc_sink *dc_sink_create(const struct dc_sink_init_data *init_params); /* Newer interfaces */ @@ -2220,10 +2241,15 @@ enum dc_status dc_process_dmub_set_mst_slots(const struct dc *dc, void dc_process_dmub_dpia_hpd_int_enable(const struct dc *dc, uint32_t hpd_int_enable); +void dc_print_dmub_diagnostic_data(const struct dc *dc); + /* DSC Interfaces */ #include "dc_dsc.h" /* Disable acc mode Interfaces */ void dc_disable_accelerated_mode(struct dc *dc); +bool dc_is_timing_changed(struct dc_stream_state *cur_stream, + struct dc_stream_state *new_stream); + #endif /* DC_INTERFACE_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c index a9b9490a532c..c52c40b16387 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c +++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c @@ -65,47 +65,6 @@ void dc_dmub_srv_destroy(struct dc_dmub_srv **dmub_srv) } } -void dc_dmub_srv_cmd_queue(struct dc_dmub_srv *dc_dmub_srv, - union dmub_rb_cmd *cmd) -{ - struct dmub_srv *dmub = dc_dmub_srv->dmub; - struct dc_context *dc_ctx = dc_dmub_srv->ctx; - enum dmub_status status; - - status = dmub_srv_cmd_queue(dmub, cmd); - if (status == DMUB_STATUS_OK) - return; - - if (status != DMUB_STATUS_QUEUE_FULL) - goto error; - - /* Execute and wait for queue to become empty again. */ - dc_dmub_srv_cmd_execute(dc_dmub_srv); - dc_dmub_srv_wait_idle(dc_dmub_srv); - - /* Requeue the command. */ - status = dmub_srv_cmd_queue(dmub, cmd); - if (status == DMUB_STATUS_OK) - return; - -error: - DC_ERROR("Error queuing DMUB command: status=%d\n", status); - dc_dmub_srv_log_diagnostic_data(dc_dmub_srv); -} - -void dc_dmub_srv_cmd_execute(struct dc_dmub_srv *dc_dmub_srv) -{ - struct dmub_srv *dmub = dc_dmub_srv->dmub; - struct dc_context *dc_ctx = dc_dmub_srv->ctx; - enum dmub_status status; - - status = dmub_srv_cmd_execute(dmub); - if (status != DMUB_STATUS_OK) { - DC_ERROR("Error starting DMUB execution: status=%d\n", status); - dc_dmub_srv_log_diagnostic_data(dc_dmub_srv); - } -} - void dc_dmub_srv_wait_idle(struct dc_dmub_srv *dc_dmub_srv) { struct dmub_srv *dmub = dc_dmub_srv->dmub; @@ -159,50 +118,89 @@ void dc_dmub_srv_send_inbox0_cmd(struct dc_dmub_srv *dmub_srv, } } -bool dc_dmub_srv_cmd_with_reply_data(struct dc_dmub_srv *dc_dmub_srv, union dmub_rb_cmd *cmd) +bool dc_dmub_srv_cmd_run(struct dc_dmub_srv *dc_dmub_srv, union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type) +{ + return dc_dmub_srv_cmd_run_list(dc_dmub_srv, 1, cmd, wait_type); +} + +bool dc_dmub_srv_cmd_run_list(struct dc_dmub_srv *dc_dmub_srv, unsigned int count, union dmub_rb_cmd *cmd_list, enum dm_dmub_wait_type wait_type) { + struct dc_context *dc_ctx; struct dmub_srv *dmub; enum dmub_status status; + int i; if (!dc_dmub_srv || !dc_dmub_srv->dmub) return false; + dc_ctx = dc_dmub_srv->ctx; dmub = dc_dmub_srv->dmub; - status = dmub_srv_cmd_with_reply_data(dmub, cmd); + for (i = 0 ; i < count; i++) { + // Queue command + status = dmub_srv_cmd_queue(dmub, &cmd_list[i]); + + if (status == DMUB_STATUS_QUEUE_FULL) { + /* Execute and wait for queue to become empty again. */ + dmub_srv_cmd_execute(dmub); + dmub_srv_wait_for_idle(dmub, 100000); + + /* Requeue the command. */ + status = dmub_srv_cmd_queue(dmub, &cmd_list[i]); + } + + if (status != DMUB_STATUS_OK) { + DC_ERROR("Error queueing DMUB command: status=%d\n", status); + dc_dmub_srv_log_diagnostic_data(dc_dmub_srv); + return false; + } + } + + status = dmub_srv_cmd_execute(dmub); if (status != DMUB_STATUS_OK) { - DC_LOG_DEBUG("No reply for DMUB command: status=%d\n", status); + DC_ERROR("Error starting DMUB execution: status=%d\n", status); + dc_dmub_srv_log_diagnostic_data(dc_dmub_srv); return false; } + // Wait for DMUB to process command + if (wait_type != DM_DMUB_WAIT_TYPE_NO_WAIT) { + status = dmub_srv_wait_for_idle(dmub, 100000); + + if (status != DMUB_STATUS_OK) { + DC_LOG_DEBUG("No reply for DMUB command: status=%d\n", status); + dc_dmub_srv_log_diagnostic_data(dc_dmub_srv); + return false; + } + + // Copy data back from ring buffer into command + if (wait_type == DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY) + dmub_rb_get_return_data(&dmub->inbox1_rb, cmd_list); + } + return true; } -void dc_dmub_srv_wait_phy_init(struct dc_dmub_srv *dc_dmub_srv) +bool dc_dmub_srv_optimized_init_done(struct dc_dmub_srv *dc_dmub_srv) { - struct dmub_srv *dmub = dc_dmub_srv->dmub; - struct dc_context *dc_ctx = dc_dmub_srv->ctx; + struct dmub_srv *dmub; + struct dc_context *dc_ctx; + union dmub_fw_boot_status boot_status; enum dmub_status status; - for (;;) { - /* Wait up to a second for PHY init. */ - status = dmub_srv_wait_for_phy_init(dmub, 1000000); - if (status == DMUB_STATUS_OK) - /* Initialization OK */ - break; - - DC_ERROR("DMCUB PHY init failed: status=%d\n", status); - ASSERT(0); + if (!dc_dmub_srv || !dc_dmub_srv->dmub) + return false; - if (status != DMUB_STATUS_TIMEOUT) - /* - * Server likely initialized or we don't have - * DMCUB HW support - this won't end. - */ - break; + dmub = dc_dmub_srv->dmub; + dc_ctx = dc_dmub_srv->ctx; - /* Continue spinning so we don't hang the ASIC. */ + status = dmub_srv_get_fw_boot_status(dmub, &boot_status); + if (status != DMUB_STATUS_OK) { + DC_ERROR("Error querying DMUB boot status: error=%d\n", status); + return false; } + + return boot_status.bits.optimized_init_done; } bool dc_dmub_srv_notify_stream_mask(struct dc_dmub_srv *dc_dmub_srv, @@ -267,9 +265,7 @@ void dc_dmub_srv_drr_update_cmd(struct dc *dc, uint32_t tg_inst, uint32_t vtotal cmd.drr_update.header.payload_bytes = sizeof(cmd.drr_update) - sizeof(cmd.drr_update.header); // Send the command to the DMCUB. - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } void dc_dmub_srv_set_drr_manual_trigger_cmd(struct dc *dc, uint32_t tg_inst) @@ -283,9 +279,7 @@ void dc_dmub_srv_set_drr_manual_trigger_cmd(struct dc *dc, uint32_t tg_inst) cmd.drr_update.header.payload_bytes = sizeof(cmd.drr_update) - sizeof(cmd.drr_update.header); // Send the command to the DMCUB. - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } static uint8_t dc_dmub_srv_get_pipes_for_stream(struct dc *dc, struct dc_stream_state *stream) @@ -378,21 +372,14 @@ bool dc_dmub_srv_p_state_delegate(struct dc *dc, bool should_manage_pstate, stru sizeof(cmd.fw_assisted_mclk_switch) - sizeof(cmd.fw_assisted_mclk_switch.header); // Send the command to the DMCUB. - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); return true; } -void dc_dmub_srv_query_caps_cmd(struct dmub_srv *dmub) +void dc_dmub_srv_query_caps_cmd(struct dc_dmub_srv *dc_dmub_srv) { union dmub_rb_cmd cmd = { 0 }; - enum dmub_status status; - - if (!dmub) { - return; - } memset(&cmd, 0, sizeof(cmd)); @@ -402,15 +389,10 @@ void dc_dmub_srv_query_caps_cmd(struct dmub_srv *dmub) cmd.query_feature_caps.header.ret_status = 1; cmd.query_feature_caps.header.payload_bytes = sizeof(struct dmub_cmd_query_feature_caps_data); - /* Send command to fw */ - status = dmub_srv_cmd_with_reply_data(dmub, &cmd); - - ASSERT(status == DMUB_STATUS_OK); - /* If command was processed, copy feature caps to dmub srv */ - if (status == DMUB_STATUS_OK && + if (dm_execute_dmub_cmd(dc_dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY) && cmd.query_feature_caps.header.ret_status == 0) { - memcpy(&dmub->feature_caps, + memcpy(&dc_dmub_srv->dmub->feature_caps, &cmd.query_feature_caps.query_feature_caps_data, sizeof(struct dmub_feature_caps)); } @@ -419,7 +401,6 @@ void dc_dmub_srv_query_caps_cmd(struct dmub_srv *dmub) void dc_dmub_srv_get_visual_confirm_color_cmd(struct dc *dc, struct pipe_ctx *pipe_ctx) { union dmub_rb_cmd cmd = { 0 }; - enum dmub_status status; unsigned int panel_inst = 0; dc_get_edp_link_panel_inst(dc, pipe_ctx->stream->link, &panel_inst); @@ -433,13 +414,8 @@ void dc_dmub_srv_get_visual_confirm_color_cmd(struct dc *dc, struct pipe_ctx *pi cmd.visual_confirm_color.header.payload_bytes = sizeof(struct dmub_cmd_visual_confirm_color_data); cmd.visual_confirm_color.visual_confirm_color_data.visual_confirm_color.panel_inst = panel_inst; - // Send command to fw - status = dmub_srv_cmd_with_reply_data(dc->ctx->dmub_srv->dmub, &cmd); - - ASSERT(status == DMUB_STATUS_OK); - // If command was processed, copy feature caps to dmub srv - if (status == DMUB_STATUS_OK && + if (dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY) && cmd.visual_confirm_color.header.ret_status == 0) { memcpy(&dc->ctx->dmub_srv->dmub->visual_confirm_color, &cmd.visual_confirm_color.visual_confirm_color_data, @@ -797,9 +773,8 @@ void dc_dmub_setup_subvp_dmub_command(struct dc *dc, cmd.fw_assisted_mclk_switch_v2.config_data.watermark_a_cache = wm_val_refclk < 0xFFFF ? wm_val_refclk : 0xFFFF; } - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } bool dc_dmub_srv_get_diagnostic_data(struct dc_dmub_srv *dc_dmub_srv, struct dmub_diagnostic_data *diag_data) @@ -823,74 +798,40 @@ void dc_dmub_srv_log_diagnostic_data(struct dc_dmub_srv *dc_dmub_srv) return; } - DC_LOG_DEBUG( - "DMCUB STATE\n" - " dmcub_version : %08x\n" - " scratch [0] : %08x\n" - " scratch [1] : %08x\n" - " scratch [2] : %08x\n" - " scratch [3] : %08x\n" - " scratch [4] : %08x\n" - " scratch [5] : %08x\n" - " scratch [6] : %08x\n" - " scratch [7] : %08x\n" - " scratch [8] : %08x\n" - " scratch [9] : %08x\n" - " scratch [10] : %08x\n" - " scratch [11] : %08x\n" - " scratch [12] : %08x\n" - " scratch [13] : %08x\n" - " scratch [14] : %08x\n" - " scratch [15] : %08x\n" - " pc : %08x\n" - " unk_fault_addr : %08x\n" - " inst_fault_addr : %08x\n" - " data_fault_addr : %08x\n" - " inbox1_rptr : %08x\n" - " inbox1_wptr : %08x\n" - " inbox1_size : %08x\n" - " inbox0_rptr : %08x\n" - " inbox0_wptr : %08x\n" - " inbox0_size : %08x\n" - " is_enabled : %d\n" - " is_soft_reset : %d\n" - " is_secure_reset : %d\n" - " is_traceport_en : %d\n" - " is_cw0_en : %d\n" - " is_cw6_en : %d\n", - diag_data.dmcub_version, - diag_data.scratch[0], - diag_data.scratch[1], - diag_data.scratch[2], - diag_data.scratch[3], - diag_data.scratch[4], - diag_data.scratch[5], - diag_data.scratch[6], - diag_data.scratch[7], - diag_data.scratch[8], - diag_data.scratch[9], - diag_data.scratch[10], - diag_data.scratch[11], - diag_data.scratch[12], - diag_data.scratch[13], - diag_data.scratch[14], - diag_data.scratch[15], - diag_data.pc, - diag_data.undefined_address_fault_addr, - diag_data.inst_fetch_fault_addr, - diag_data.data_write_fault_addr, - diag_data.inbox1_rptr, - diag_data.inbox1_wptr, - diag_data.inbox1_size, - diag_data.inbox0_rptr, - diag_data.inbox0_wptr, - diag_data.inbox0_size, - diag_data.is_dmcub_enabled, - diag_data.is_dmcub_soft_reset, - diag_data.is_dmcub_secure_reset, - diag_data.is_traceport_en, - diag_data.is_cw0_enabled, - diag_data.is_cw6_enabled); + DC_LOG_DEBUG("DMCUB STATE:"); + DC_LOG_DEBUG(" dmcub_version : %08x", diag_data.dmcub_version); + DC_LOG_DEBUG(" scratch [0] : %08x", diag_data.scratch[0]); + DC_LOG_DEBUG(" scratch [1] : %08x", diag_data.scratch[1]); + DC_LOG_DEBUG(" scratch [2] : %08x", diag_data.scratch[2]); + DC_LOG_DEBUG(" scratch [3] : %08x", diag_data.scratch[3]); + DC_LOG_DEBUG(" scratch [4] : %08x", diag_data.scratch[4]); + DC_LOG_DEBUG(" scratch [5] : %08x", diag_data.scratch[5]); + DC_LOG_DEBUG(" scratch [6] : %08x", diag_data.scratch[6]); + DC_LOG_DEBUG(" scratch [7] : %08x", diag_data.scratch[7]); + DC_LOG_DEBUG(" scratch [8] : %08x", diag_data.scratch[8]); + DC_LOG_DEBUG(" scratch [9] : %08x", diag_data.scratch[9]); + DC_LOG_DEBUG(" scratch [10] : %08x", diag_data.scratch[10]); + DC_LOG_DEBUG(" scratch [11] : %08x", diag_data.scratch[11]); + DC_LOG_DEBUG(" scratch [12] : %08x", diag_data.scratch[12]); + DC_LOG_DEBUG(" scratch [13] : %08x", diag_data.scratch[13]); + DC_LOG_DEBUG(" scratch [14] : %08x", diag_data.scratch[14]); + DC_LOG_DEBUG(" scratch [15] : %08x", diag_data.scratch[15]); + DC_LOG_DEBUG(" pc : %08x", diag_data.pc); + DC_LOG_DEBUG(" unk_fault_addr : %08x", diag_data.undefined_address_fault_addr); + DC_LOG_DEBUG(" inst_fault_addr : %08x", diag_data.inst_fetch_fault_addr); + DC_LOG_DEBUG(" data_fault_addr : %08x", diag_data.data_write_fault_addr); + DC_LOG_DEBUG(" inbox1_rptr : %08x", diag_data.inbox1_rptr); + DC_LOG_DEBUG(" inbox1_wptr : %08x", diag_data.inbox1_wptr); + DC_LOG_DEBUG(" inbox1_size : %08x", diag_data.inbox1_size); + DC_LOG_DEBUG(" inbox0_rptr : %08x", diag_data.inbox0_rptr); + DC_LOG_DEBUG(" inbox0_wptr : %08x", diag_data.inbox0_wptr); + DC_LOG_DEBUG(" inbox0_size : %08x", diag_data.inbox0_size); + DC_LOG_DEBUG(" is_enabled : %d", diag_data.is_dmcub_enabled); + DC_LOG_DEBUG(" is_soft_reset : %d", diag_data.is_dmcub_soft_reset); + DC_LOG_DEBUG(" is_secure_reset : %d", diag_data.is_dmcub_secure_reset); + DC_LOG_DEBUG(" is_traceport_en : %d", diag_data.is_traceport_en); + DC_LOG_DEBUG(" is_cw0_en : %d", diag_data.is_cw0_enabled); + DC_LOG_DEBUG(" is_cw6_en : %d", diag_data.is_cw6_enabled); } static bool dc_can_pipe_disable_cursor(struct pipe_ctx *pipe_ctx) @@ -982,14 +923,6 @@ static void dc_build_cursor_update_payload0( payload->panel_inst = panel_inst; } -static void dc_send_cmd_to_dmu(struct dc_dmub_srv *dmub_srv, - union dmub_rb_cmd *cmd) -{ - dc_dmub_srv_cmd_queue(dmub_srv, cmd); - dc_dmub_srv_cmd_execute(dmub_srv); - dc_dmub_srv_wait_idle(dmub_srv); -} - static void dc_build_cursor_position_update_payload0( struct dmub_cmd_update_cursor_payload0 *pl, const uint8_t p_idx, const struct hubp *hubp, const struct dpp *dpp) @@ -1032,9 +965,11 @@ static void dc_build_cursor_attribute_update_payload1( void dc_send_update_cursor_info_to_dmu( struct pipe_ctx *pCtx, uint8_t pipe_idx) { - union dmub_rb_cmd cmd = { 0 }; - union dmub_cmd_update_cursor_info_data *update_cursor_info = - &cmd.update_cursor_info.update_cursor_info_data; + union dmub_rb_cmd cmd[2]; + union dmub_cmd_update_cursor_info_data *update_cursor_info_0 = + &cmd[0].update_cursor_info.update_cursor_info_data; + + memset(cmd, 0, sizeof(cmd)); if (!dc_dmub_should_update_cursor_data(pCtx)) return; @@ -1051,31 +986,28 @@ void dc_send_update_cursor_info_to_dmu( { /* Build Payload#0 Header */ - cmd.update_cursor_info.header.type = DMUB_CMD__UPDATE_CURSOR_INFO; - cmd.update_cursor_info.header.payload_bytes = - sizeof(cmd.update_cursor_info.update_cursor_info_data); - cmd.update_cursor_info.header.multi_cmd_pending = 1; /* To combine multi dmu cmd, 1st cmd */ + cmd[0].update_cursor_info.header.type = DMUB_CMD__UPDATE_CURSOR_INFO; + cmd[0].update_cursor_info.header.payload_bytes = + sizeof(cmd[0].update_cursor_info.update_cursor_info_data); + cmd[0].update_cursor_info.header.multi_cmd_pending = 1; //To combine multi dmu cmd, 1st cmd /* Prepare Payload */ - dc_build_cursor_update_payload0(pCtx, pipe_idx, &update_cursor_info->payload0); + dc_build_cursor_update_payload0(pCtx, pipe_idx, &update_cursor_info_0->payload0); - dc_build_cursor_position_update_payload0(&update_cursor_info->payload0, pipe_idx, + dc_build_cursor_position_update_payload0(&update_cursor_info_0->payload0, pipe_idx, pCtx->plane_res.hubp, pCtx->plane_res.dpp); - /* Send update_curosr_info to queue */ - dc_dmub_srv_cmd_queue(pCtx->stream->ctx->dmub_srv, &cmd); - } + } { /* Build Payload#1 Header */ - memset(update_cursor_info, 0, sizeof(union dmub_cmd_update_cursor_info_data)); - cmd.update_cursor_info.header.type = DMUB_CMD__UPDATE_CURSOR_INFO; - cmd.update_cursor_info.header.payload_bytes = sizeof(struct cursor_attributes_cfg); - cmd.update_cursor_info.header.multi_cmd_pending = 0; /* Indicate it's the last command. */ + cmd[1].update_cursor_info.header.type = DMUB_CMD__UPDATE_CURSOR_INFO; + cmd[1].update_cursor_info.header.payload_bytes = sizeof(struct cursor_attributes_cfg); + cmd[1].update_cursor_info.header.multi_cmd_pending = 0; //Indicate it's the last command. dc_build_cursor_attribute_update_payload1( - &cmd.update_cursor_info.update_cursor_info_data.payload1.attribute_cfg, + &cmd[1].update_cursor_info.update_cursor_info_data.payload1.attribute_cfg, pipe_idx, pCtx->plane_res.hubp, pCtx->plane_res.dpp); /* Combine 2nd cmds update_curosr_info to DMU */ - dc_send_cmd_to_dmu(pCtx->stream->ctx->dmub_srv, &cmd); + dm_execute_dmub_cmd_list(pCtx->stream->ctx, 2, cmd, DM_DMUB_WAIT_TYPE_WAIT); } } diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h index d34f5563df2e..a5196a9292b3 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h +++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h @@ -26,7 +26,7 @@ #ifndef _DMUB_DC_SRV_H_ #define _DMUB_DC_SRV_H_ -#include "os_types.h" +#include "dm_services_types.h" #include "dmub/dmub_srv.h" struct dmub_srv; @@ -52,16 +52,13 @@ struct dc_dmub_srv { void *dm; }; -void dc_dmub_srv_cmd_queue(struct dc_dmub_srv *dc_dmub_srv, - union dmub_rb_cmd *cmd); - -void dc_dmub_srv_cmd_execute(struct dc_dmub_srv *dc_dmub_srv); - void dc_dmub_srv_wait_idle(struct dc_dmub_srv *dc_dmub_srv); -void dc_dmub_srv_wait_phy_init(struct dc_dmub_srv *dc_dmub_srv); +bool dc_dmub_srv_optimized_init_done(struct dc_dmub_srv *dc_dmub_srv); + +bool dc_dmub_srv_cmd_run(struct dc_dmub_srv *dc_dmub_srv, union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type); -bool dc_dmub_srv_cmd_with_reply_data(struct dc_dmub_srv *dc_dmub_srv, union dmub_rb_cmd *cmd); +bool dc_dmub_srv_cmd_run_list(struct dc_dmub_srv *dc_dmub_srv, unsigned int count, union dmub_rb_cmd *cmd_list, enum dm_dmub_wait_type wait_type); bool dc_dmub_srv_notify_stream_mask(struct dc_dmub_srv *dc_dmub_srv, unsigned int stream_mask); @@ -77,7 +74,7 @@ void dc_dmub_srv_drr_update_cmd(struct dc *dc, uint32_t tg_inst, uint32_t vtotal void dc_dmub_srv_set_drr_manual_trigger_cmd(struct dc *dc, uint32_t tg_inst); bool dc_dmub_srv_p_state_delegate(struct dc *dc, bool enable_pstate, struct dc_state *context); -void dc_dmub_srv_query_caps_cmd(struct dmub_srv *dmub); +void dc_dmub_srv_query_caps_cmd(struct dc_dmub_srv *dc_dmub_srv); void dc_dmub_srv_get_visual_confirm_color_cmd(struct dc *dc, struct pipe_ctx *pipe_ctx); void dc_dmub_srv_clear_inbox0_ack(struct dc_dmub_srv *dmub_srv); void dc_dmub_srv_wait_for_inbox0_ack(struct dc_dmub_srv *dmub_srv); diff --git a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h index 49aab1924665..55139d7bf422 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h @@ -61,7 +61,7 @@ enum dc_link_rate { */ LINK_RATE_UHBR10 = 1000, // UHBR10 - 10.0 Gbps/Lane LINK_RATE_UHBR13_5 = 1350, // UHBR13.5 - 13.5 Gbps/Lane - LINK_RATE_UHBR20 = 2000, // UHBR10 - 20.0 Gbps/Lane + LINK_RATE_UHBR20 = 2000, // UHBR20 - 20.0 Gbps/Lane }; enum dc_link_spread { @@ -566,6 +566,12 @@ struct dpcd_amd_device_id { uint8_t dal_version_byte2; }; +struct target_luminance_value { + uint8_t byte0; + uint8_t byte1; + uint8_t byte2; +}; + struct dpcd_source_backlight_set { struct { uint8_t byte0; @@ -1225,6 +1231,7 @@ struct dpcd_caps { union dp_main_line_channel_coding_cap channel_coding_cap; union dp_sink_video_fallback_formats fallback_formats; union dp_fec_capability1 fec_cap1; + bool panel_luminance_control; union dp_cable_id cable_id; uint8_t edp_rev; union edp_alpm_caps alpm_caps; diff --git a/drivers/gpu/drm/amd/display/dc/dc_dsc.h b/drivers/gpu/drm/amd/display/dc/dc_dsc.h index 0e92a322c2ed..9491b76d61f5 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_dsc.h +++ b/drivers/gpu/drm/amd/display/dc/dc_dsc.h @@ -58,6 +58,7 @@ struct dc_dsc_config_options { uint32_t dsc_min_slice_height_override; uint32_t max_target_bpp_limit_override_x16; uint32_t slice_height_granularity; + uint32_t dsc_force_odm_hslice_override; }; bool dc_dsc_parse_dsc_dpcd(const struct dc *dc, diff --git a/drivers/gpu/drm/amd/display/dc/dc_helper.c b/drivers/gpu/drm/amd/display/dc/dc_helper.c index f43cce16bb6c..3907eeff560c 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_helper.c +++ b/drivers/gpu/drm/amd/display/dc/dc_helper.c @@ -41,19 +41,13 @@ static inline void submit_dmub_read_modify_write( const struct dc_context *ctx) { struct dmub_rb_cmd_read_modify_write *cmd_buf = &offload->cmd_data.read_modify_write; - bool gather = false; offload->should_burst_write = (offload->same_addr_count == (DMUB_READ_MODIFY_WRITE_SEQ__MAX - 1)); cmd_buf->header.payload_bytes = sizeof(struct dmub_cmd_read_modify_write_sequence) * offload->reg_seq_count; - gather = ctx->dmub_srv->reg_helper_offload.gather_in_progress; - ctx->dmub_srv->reg_helper_offload.gather_in_progress = false; - - dc_dmub_srv_cmd_queue(ctx->dmub_srv, &offload->cmd_data); - - ctx->dmub_srv->reg_helper_offload.gather_in_progress = gather; + dm_execute_dmub_cmd(ctx, &offload->cmd_data, DM_DMUB_WAIT_TYPE_NO_WAIT); memset(cmd_buf, 0, sizeof(*cmd_buf)); @@ -66,17 +60,11 @@ static inline void submit_dmub_burst_write( const struct dc_context *ctx) { struct dmub_rb_cmd_burst_write *cmd_buf = &offload->cmd_data.burst_write; - bool gather = false; cmd_buf->header.payload_bytes = sizeof(uint32_t) * offload->reg_seq_count; - gather = ctx->dmub_srv->reg_helper_offload.gather_in_progress; - ctx->dmub_srv->reg_helper_offload.gather_in_progress = false; - - dc_dmub_srv_cmd_queue(ctx->dmub_srv, &offload->cmd_data); - - ctx->dmub_srv->reg_helper_offload.gather_in_progress = gather; + dm_execute_dmub_cmd(ctx, &offload->cmd_data, DM_DMUB_WAIT_TYPE_NO_WAIT); memset(cmd_buf, 0, sizeof(*cmd_buf)); @@ -88,17 +76,11 @@ static inline void submit_dmub_reg_wait( const struct dc_context *ctx) { struct dmub_rb_cmd_reg_wait *cmd_buf = &offload->cmd_data.reg_wait; - bool gather = false; - - gather = ctx->dmub_srv->reg_helper_offload.gather_in_progress; - ctx->dmub_srv->reg_helper_offload.gather_in_progress = false; - dc_dmub_srv_cmd_queue(ctx->dmub_srv, &offload->cmd_data); + dm_execute_dmub_cmd(ctx, &offload->cmd_data, DM_DMUB_WAIT_TYPE_NO_WAIT); memset(cmd_buf, 0, sizeof(*cmd_buf)); offload->reg_seq_count = 0; - - ctx->dmub_srv->reg_helper_offload.gather_in_progress = gather; } struct dc_reg_value_masks { @@ -151,7 +133,6 @@ static void dmub_flush_buffer_execute( const struct dc_context *ctx) { submit_dmub_read_modify_write(offload, ctx); - dc_dmub_srv_cmd_execute(ctx->dmub_srv); } static void dmub_flush_burst_write_buffer_execute( @@ -159,7 +140,6 @@ static void dmub_flush_burst_write_buffer_execute( const struct dc_context *ctx) { submit_dmub_burst_write(offload, ctx); - dc_dmub_srv_cmd_execute(ctx->dmub_srv); } static bool dmub_reg_value_burst_set_pack(const struct dc_context *ctx, uint32_t addr, @@ -484,8 +464,7 @@ void generic_reg_wait(const struct dc_context *ctx, field_value = get_reg_field_value_ex(reg_val, mask, shift); if (field_value == condition_value) { - if (i * delay_between_poll_us > 1000 && - !IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) + if (i * delay_between_poll_us > 1000) DC_LOG_DC("REG_WAIT taking a while: %dms in %s line:%d\n", delay_between_poll_us * i / 1000, func_name, line); @@ -497,8 +476,7 @@ void generic_reg_wait(const struct dc_context *ctx, delay_between_poll_us, time_out_num_tries, func_name, line); - if (!IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) - BREAK_TO_DEBUGGER(); + BREAK_TO_DEBUGGER(); } void generic_write_indirect_reg(const struct dc_context *ctx, @@ -691,8 +669,6 @@ void reg_sequence_start_execute(const struct dc_context *ctx) default: return; } - - dc_dmub_srv_cmd_execute(ctx->dmub_srv); } } @@ -712,3 +688,59 @@ void reg_sequence_wait_done(const struct dc_context *ctx) dc_dmub_srv_wait_idle(ctx->dmub_srv); } } + +char *dce_version_to_string(const int version) +{ + switch (version) { + case DCE_VERSION_8_0: + return "DCE 8.0"; + case DCE_VERSION_8_1: + return "DCE 8.1"; + case DCE_VERSION_8_3: + return "DCE 8.3"; + case DCE_VERSION_10_0: + return "DCE 10.0"; + case DCE_VERSION_11_0: + return "DCE 11.0"; + case DCE_VERSION_11_2: + return "DCE 11.2"; + case DCE_VERSION_11_22: + return "DCE 11.22"; + case DCE_VERSION_12_0: + return "DCE 12.0"; + case DCE_VERSION_12_1: + return "DCE 12.1"; + case DCN_VERSION_1_0: + return "DCN 1.0"; + case DCN_VERSION_1_01: + return "DCN 1.0.1"; + case DCN_VERSION_2_0: + return "DCN 2.0"; + case DCN_VERSION_2_1: + return "DCN 2.1"; + case DCN_VERSION_2_01: + return "DCN 2.0.1"; + case DCN_VERSION_3_0: + return "DCN 3.0"; + case DCN_VERSION_3_01: + return "DCN 3.0.1"; + case DCN_VERSION_3_02: + return "DCN 3.0.2"; + case DCN_VERSION_3_03: + return "DCN 3.0.3"; + case DCN_VERSION_3_1: + return "DCN 3.1"; + case DCN_VERSION_3_14: + return "DCN 3.1.4"; + case DCN_VERSION_3_15: + return "DCN 3.1.5"; + case DCN_VERSION_3_16: + return "DCN 3.1.6"; + case DCN_VERSION_3_2: + return "DCN 3.2"; + case DCN_VERSION_3_21: + return "DCN 3.2.1"; + default: + return "Unknown"; + } +} diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h index 25284006019c..3697ea1d14c1 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_stream.h +++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h @@ -131,6 +131,7 @@ union stream_update_flags { uint32_t dsc_changed : 1; uint32_t mst_bw : 1; uint32_t crtc_timing_adjust : 1; + uint32_t fams_changed : 1; } bits; uint32_t raw; @@ -171,6 +172,10 @@ struct mall_temp_config { bool is_phantom_plane[MAX_PIPES]; }; +struct dc_stream_debug_options { + char force_odm_combine_segments; +}; + struct dc_stream_state { // sink is deprecated, new code should not reference // this pointer @@ -181,6 +186,7 @@ struct dc_stream_state { * a stream via the volatile dc_state rather than the static dc_link. */ struct link_encoder *link_enc; + struct dc_stream_debug_options debug; struct dc_panel_patch sink_patches; union display_content_support content_support; struct dc_crtc_timing timing; @@ -227,6 +233,7 @@ struct dc_stream_state { */ bool vrr_active_variable; bool freesync_on_desktop; + bool vrr_active_fixed; bool converter_disable_audio; uint8_t qs_bit; @@ -295,6 +302,7 @@ struct dc_stream_state { bool vblank_synchronized; bool fpo_in_use; struct mall_stream_config mall_stream_config; + bool skip_edp_power_down; }; #define ABM_LEVEL_IMMEDIATE_DISABLE 255 @@ -320,6 +328,7 @@ struct dc_stream_update { bool integer_scaling_update; bool *allow_freesync; bool *vrr_active_variable; + bool *vrr_active_fixed; struct colorspace_transform *gamut_remap; enum dc_color_space *output_color_space; diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h index 45ab48fe5d00..0ce7728a5a4b 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_types.h @@ -69,13 +69,6 @@ enum dce_environment { DCE_ENV_VIRTUAL_HW }; -/* Note: use these macro definitions instead of direct comparison! */ -#define IS_FPGA_MAXIMUS_DC(dce_environment) \ - (dce_environment == DCE_ENV_FPGA_MAXIMUS) - -#define IS_DIAG_DC(dce_environment) \ - (IS_FPGA_MAXIMUS_DC(dce_environment) || (dce_environment == DCE_ENV_DIAG)) - struct dc_perf_trace { unsigned long read_count; unsigned long write_count; @@ -83,7 +76,7 @@ struct dc_perf_trace { unsigned long last_entry_write; }; -#define MAX_SURFACE_NUM 4 +#define MAX_SURFACE_NUM 6 #define NUM_PIXEL_FORMATS 10 enum tiling_mode { @@ -196,6 +189,7 @@ struct dc_panel_patch { unsigned int disable_fams; unsigned int skip_avmute; unsigned int mst_start_top_delay; + unsigned int delay_disable_aux_intercept_ms; }; struct dc_edid_caps { @@ -603,6 +597,7 @@ enum dc_psr_state { PSR_STATE4b_FULL_FRAME, PSR_STATE4c_FULL_FRAME, PSR_STATE4_FULL_FRAME_POWERUP, + PSR_STATE4_FULL_FRAME_HW_LOCK, PSR_STATE5, PSR_STATE5a, PSR_STATE5b, diff --git a/drivers/gpu/drm/amd/display/dc/dce/Makefile b/drivers/gpu/drm/amd/display/dc/dce/Makefile index 0d7db132a20f..01490c9ba958 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/Makefile +++ b/drivers/gpu/drm/amd/display/dc/dce/Makefile @@ -29,7 +29,7 @@ DCE = dce_audio.o dce_stream_encoder.o dce_link_encoder.o dce_hwseq.o \ dce_mem_input.o dce_clock_source.o dce_scl_filters.o dce_transform.o \ dce_opp.o dce_dmcu.o dce_abm.o dce_ipp.o dce_aux.o \ -dce_i2c.o dce_i2c_hw.o dce_i2c_sw.o dmub_psr.o dmub_abm.o dce_panel_cntl.o \ +dce_i2c.o dce_i2c_hw.o dce_i2c_sw.o dmub_psr.o dmub_abm.o dmub_abm_lcd.o dce_panel_cntl.o \ dmub_hw_lock_mgr.o dmub_outbox.o AMD_DAL_DCE = $(addprefix $(AMDDALPATH)/dc/dce/,$(DCE)) diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c index 462c7a3ec3cc..ed8936405dfa 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c @@ -920,25 +920,6 @@ static bool dce112_program_pix_clk( struct dce110_clk_src *clk_src = TO_DCE110_CLK_SRC(clock_source); struct bp_pixel_clock_parameters bp_pc_params = {0}; - if (IS_FPGA_MAXIMUS_DC(clock_source->ctx->dce_environment)) { - unsigned int inst = pix_clk_params->controller_id - CONTROLLER_ID_D0; - unsigned dp_dto_ref_100hz = 7000000; - unsigned clock_100hz = pll_settings->actual_pix_clk_100hz; - - /* Set DTO values: phase = target clock, modulo = reference clock */ - REG_WRITE(PHASE[inst], clock_100hz); - REG_WRITE(MODULO[inst], dp_dto_ref_100hz); - - /* Enable DTO */ - if (clk_src->cs_mask->PIPE0_DTO_SRC_SEL) - REG_UPDATE_2(PIXEL_RATE_CNTL[inst], - DP_DTO0_ENABLE, 1, - PIPE0_DTO_SRC_SEL, 1); - else - REG_UPDATE(PIXEL_RATE_CNTL[inst], - DP_DTO0_ENABLE, 1); - return true; - } /* First disable SS * ATOMBIOS will enable by default SS on PLL for DP, * do not disable it here @@ -1015,25 +996,6 @@ static bool dcn31_program_pix_clk( REG_UPDATE(PIXEL_RATE_CNTL[inst], DP_DTO0_ENABLE, 1); } else { - if (IS_FPGA_MAXIMUS_DC(clock_source->ctx->dce_environment)) { - unsigned int inst = pix_clk_params->controller_id - CONTROLLER_ID_D0; - unsigned dp_dto_ref_100hz = 7000000; - unsigned clock_100hz = pll_settings->actual_pix_clk_100hz; - - /* Set DTO values: phase = target clock, modulo = reference clock */ - REG_WRITE(PHASE[inst], clock_100hz); - REG_WRITE(MODULO[inst], dp_dto_ref_100hz); - - /* Enable DTO */ - if (clk_src->cs_mask->PIPE0_DTO_SRC_SEL) - REG_UPDATE_2(PIXEL_RATE_CNTL[inst], - DP_DTO0_ENABLE, 1, - PIPE0_DTO_SRC_SEL, 1); - else - REG_UPDATE(PIXEL_RATE_CNTL[inst], - DP_DTO0_ENABLE, 1); - return true; - } if (clk_src->cs_mask->PIPE0_DTO_SRC_SEL) REG_UPDATE(PIXEL_RATE_CNTL[inst], diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c b/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c index e74266cc0098..63009db8b5a7 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c @@ -1093,11 +1093,9 @@ static void dcn21_dmcu_construct( dce_dmcu_construct(dmcu_dce, ctx, regs, dmcu_shift, dmcu_mask); - if (!IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) { - psp_version = dm_read_reg(ctx, mmMP0_SMN_C2PMSG_58); - dmcu_dce->base.auto_load_dmcu = ((psp_version & 0x00FF00FF) > 0x00110029); - dmcu_dce->base.psp_version = psp_version; - } + psp_version = dm_read_reg(ctx, mmMP0_SMN_C2PMSG_58); + dmcu_dce->base.auto_load_dmcu = ((psp_version & 0x00FF00FF) > 0x00110029); + dmcu_dce->base.psp_version = psp_version; } struct dmcu *dce_dmcu_create( diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c index 9fc48208c2e4..2fb9572ce25d 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c @@ -24,212 +24,139 @@ */ #include "dmub_abm.h" -#include "dce_abm.h" +#include "dmub_abm_lcd.h" #include "dc.h" -#include "dc_dmub_srv.h" -#include "dmub/dmub_srv.h" #include "core_types.h" -#include "dm_services.h" -#include "reg_helper.h" -#include "fixed31_32.h" - -#include "atom.h" #define TO_DMUB_ABM(abm)\ container_of(abm, struct dce_abm, base) -#define REG(reg) \ - (dce_abm->regs->reg) - -#undef FN -#define FN(reg_name, field_name) \ - dce_abm->abm_shift->field_name, dce_abm->abm_mask->field_name - -#define CTX \ - dce_abm->base.ctx - -#define DISABLE_ABM_IMMEDIATELY 255 - - +#define ABM_FEATURE_NO_SUPPORT 0 +#define ABM_LCD_SUPPORT 1 -static void dmub_abm_enable_fractional_pwm(struct dc_context *dc) +static unsigned int abm_feature_support(struct abm *abm, unsigned int panel_inst) { - union dmub_rb_cmd cmd; - uint32_t fractional_pwm = (dc->dc->config.disable_fractional_pwm == false) ? 1 : 0; - uint32_t edp_id_count = dc->dc_edp_id_count; + struct dc_context *dc = abm->ctx; + struct dc_link *edp_links[MAX_NUM_EDP]; int i; - uint8_t panel_mask = 0; - - for (i = 0; i < edp_id_count; i++) - panel_mask |= 0x01 << i; - - memset(&cmd, 0, sizeof(cmd)); - cmd.abm_set_pwm_frac.header.type = DMUB_CMD__ABM; - cmd.abm_set_pwm_frac.header.sub_type = DMUB_CMD__ABM_SET_PWM_FRAC; - cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.fractional_pwm = fractional_pwm; - cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1; - cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.panel_mask = panel_mask; - cmd.abm_set_pwm_frac.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_pwm_frac_data); - - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); -} - -static void dmub_abm_init(struct abm *abm, uint32_t backlight) -{ - struct dce_abm *dce_abm = TO_DMUB_ABM(abm); - - REG_WRITE(DC_ABM1_HG_SAMPLE_RATE, 0x3); - REG_WRITE(DC_ABM1_HG_SAMPLE_RATE, 0x1); - REG_WRITE(DC_ABM1_LS_SAMPLE_RATE, 0x3); - REG_WRITE(DC_ABM1_LS_SAMPLE_RATE, 0x1); - REG_WRITE(BL1_PWM_BL_UPDATE_SAMPLE_RATE, 0x1); - - REG_SET_3(DC_ABM1_HG_MISC_CTRL, 0, - ABM1_HG_NUM_OF_BINS_SEL, 0, - ABM1_HG_VMAX_SEL, 1, - ABM1_HG_BIN_BITWIDTH_SIZE_SEL, 0); - - REG_SET_3(DC_ABM1_IPCSC_COEFF_SEL, 0, - ABM1_IPCSC_COEFF_SEL_R, 2, - ABM1_IPCSC_COEFF_SEL_G, 4, - ABM1_IPCSC_COEFF_SEL_B, 2); - - REG_UPDATE(BL1_PWM_CURRENT_ABM_LEVEL, - BL1_PWM_CURRENT_ABM_LEVEL, backlight); - - REG_UPDATE(BL1_PWM_TARGET_ABM_LEVEL, - BL1_PWM_TARGET_ABM_LEVEL, backlight); + int edp_num; + unsigned int ret = ABM_FEATURE_NO_SUPPORT; - REG_UPDATE(BL1_PWM_USER_LEVEL, - BL1_PWM_USER_LEVEL, backlight); + dc_get_edp_links(dc->dc, edp_links, &edp_num); - REG_UPDATE_2(DC_ABM1_LS_MIN_MAX_PIXEL_VALUE_THRES, - ABM1_LS_MIN_PIXEL_VALUE_THRES, 0, - ABM1_LS_MAX_PIXEL_VALUE_THRES, 1000); + for (i = 0; i < edp_num; i++) { + if (panel_inst == i) + break; + } - REG_SET_3(DC_ABM1_HGLS_REG_READ_PROGRESS, 0, - ABM1_HG_REG_READ_MISSED_FRAME_CLEAR, 1, - ABM1_LS_REG_READ_MISSED_FRAME_CLEAR, 1, - ABM1_BL_REG_READ_MISSED_FRAME_CLEAR, 1); + if (i < edp_num) { + ret = ABM_LCD_SUPPORT; + } - dmub_abm_enable_fractional_pwm(abm->ctx); + return ret; } -static unsigned int dmub_abm_get_current_backlight(struct abm *abm) +static void dmub_abm_init_ex(struct abm *abm, uint32_t backlight) { - struct dce_abm *dce_abm = TO_DMUB_ABM(abm); - unsigned int backlight = REG_READ(BL1_PWM_CURRENT_ABM_LEVEL); - - /* return backlight in hardware format which is unsigned 17 bits, with - * 1 bit integer and 16 bit fractional - */ - return backlight; + dmub_abm_init(abm, backlight); } -static unsigned int dmub_abm_get_target_backlight(struct abm *abm) +static unsigned int dmub_abm_get_current_backlight_ex(struct abm *abm) { - struct dce_abm *dce_abm = TO_DMUB_ABM(abm); - unsigned int backlight = REG_READ(BL1_PWM_TARGET_ABM_LEVEL); + return dmub_abm_get_current_backlight(abm); +} - /* return backlight in hardware format which is unsigned 17 bits, with - * 1 bit integer and 16 bit fractional - */ - return backlight; +static unsigned int dmub_abm_get_target_backlight_ex(struct abm *abm) +{ + return dmub_abm_get_target_backlight(abm); } -static bool dmub_abm_set_level(struct abm *abm, uint32_t level) +static bool dmub_abm_set_level_ex(struct abm *abm, uint32_t level) { - union dmub_rb_cmd cmd; - struct dc_context *dc = abm->ctx; - struct dc_link *edp_links[MAX_NUM_EDP]; - int i; - int edp_num; - uint8_t panel_mask = 0; + bool ret = false; + unsigned int feature_support, i; + uint8_t panel_mask0 = 0; - dc_get_edp_links(dc->dc, edp_links, &edp_num); + for (i = 0; i < MAX_NUM_EDP; i++) { + feature_support = abm_feature_support(abm, i); - for (i = 0; i < edp_num; i++) { - if (edp_links[i]->link_status.link_active) - panel_mask |= (0x01 << i); + if (feature_support == ABM_LCD_SUPPORT) + panel_mask0 |= (0x01 << i); } - memset(&cmd, 0, sizeof(cmd)); - cmd.abm_set_level.header.type = DMUB_CMD__ABM; - cmd.abm_set_level.header.sub_type = DMUB_CMD__ABM_SET_LEVEL; - cmd.abm_set_level.abm_set_level_data.level = level; - cmd.abm_set_level.abm_set_level_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1; - cmd.abm_set_level.abm_set_level_data.panel_mask = panel_mask; - cmd.abm_set_level.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_level_data); + if (panel_mask0) + ret = dmub_abm_set_level(abm, level, panel_mask0); - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); - - return true; + return ret; } -static bool dmub_abm_init_config(struct abm *abm, +static bool dmub_abm_init_config_ex(struct abm *abm, const char *src, unsigned int bytes, unsigned int inst) { - union dmub_rb_cmd cmd; - struct dc_context *dc = abm->ctx; - uint8_t panel_mask = 0x01 << inst; + unsigned int feature_support; + + feature_support = abm_feature_support(abm, inst); + + if (feature_support == ABM_LCD_SUPPORT) + dmub_abm_init_config(abm, src, bytes, inst); + + return true; +} - // TODO: Optimize by only reading back final 4 bytes - dmub_flush_buffer_mem(&dc->dmub_srv->dmub->scratch_mem_fb); +static bool dmub_abm_set_pause_ex(struct abm *abm, bool pause, unsigned int panel_inst, unsigned int stream_inst) +{ + bool ret = false; + unsigned int feature_support; - // Copy iramtable into cw7 - memcpy(dc->dmub_srv->dmub->scratch_mem_fb.cpu_addr, (void *)src, bytes); + feature_support = abm_feature_support(abm, panel_inst); - memset(&cmd, 0, sizeof(cmd)); - // Fw will copy from cw7 to fw_state - cmd.abm_init_config.header.type = DMUB_CMD__ABM; - cmd.abm_init_config.header.sub_type = DMUB_CMD__ABM_INIT_CONFIG; - cmd.abm_init_config.abm_init_config_data.src.quad_part = dc->dmub_srv->dmub->scratch_mem_fb.gpu_addr; - cmd.abm_init_config.abm_init_config_data.bytes = bytes; - cmd.abm_init_config.abm_init_config_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1; - cmd.abm_init_config.abm_init_config_data.panel_mask = panel_mask; + if (feature_support == ABM_LCD_SUPPORT) + ret = dmub_abm_set_pause(abm, pause, panel_inst, stream_inst); - cmd.abm_init_config.header.payload_bytes = sizeof(struct dmub_cmd_abm_init_config_data); + return ret; +} - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); +static bool dmub_abm_set_pipe_ex(struct abm *abm, uint32_t otg_inst, uint32_t option, uint32_t panel_inst) +{ + bool ret = false; + unsigned int feature_support; - return true; + feature_support = abm_feature_support(abm, panel_inst); + + if (feature_support == ABM_LCD_SUPPORT) + ret = dmub_abm_set_pipe(abm, otg_inst, option, panel_inst); + + return ret; } -static bool dmub_abm_set_pause(struct abm *abm, bool pause, unsigned int panel_inst, unsigned int stream_inst) +static bool dmub_abm_set_backlight_level_pwm_ex(struct abm *abm, + unsigned int backlight_pwm_u16_16, + unsigned int frame_ramp, + unsigned int controller_id, + unsigned int panel_inst) { - union dmub_rb_cmd cmd; - struct dc_context *dc = abm->ctx; - uint8_t panel_mask = 0x01 << panel_inst; + bool ret = false; + unsigned int feature_support; - memset(&cmd, 0, sizeof(cmd)); - cmd.abm_pause.header.type = DMUB_CMD__ABM; - cmd.abm_pause.header.sub_type = DMUB_CMD__ABM_PAUSE; - cmd.abm_pause.abm_pause_data.enable = pause; - cmd.abm_pause.abm_pause_data.panel_mask = panel_mask; - cmd.abm_set_level.header.payload_bytes = sizeof(struct dmub_cmd_abm_pause_data); + feature_support = abm_feature_support(abm, panel_inst); - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); + if (feature_support == ABM_LCD_SUPPORT) + ret = dmub_abm_set_backlight_level(abm, backlight_pwm_u16_16, frame_ramp, panel_inst); - return true; + return ret; } static const struct abm_funcs abm_funcs = { - .abm_init = dmub_abm_init, - .set_abm_level = dmub_abm_set_level, - .get_current_backlight = dmub_abm_get_current_backlight, - .get_target_backlight = dmub_abm_get_target_backlight, - .init_abm_config = dmub_abm_init_config, - .set_abm_pause = dmub_abm_set_pause, + .abm_init = dmub_abm_init_ex, + .set_abm_level = dmub_abm_set_level_ex, + .get_current_backlight = dmub_abm_get_current_backlight_ex, + .get_target_backlight = dmub_abm_get_target_backlight_ex, + .init_abm_config = dmub_abm_init_config_ex, + .set_abm_pause = dmub_abm_set_pause_ex, + .set_pipe_ex = dmub_abm_set_pipe_ex, + .set_backlight_level_pwm = dmub_abm_set_backlight_level_pwm_ex, }; static void dmub_abm_construct( @@ -256,16 +183,19 @@ struct abm *dmub_abm_create( const struct dce_abm_shift *abm_shift, const struct dce_abm_mask *abm_mask) { - struct dce_abm *abm_dce = kzalloc(sizeof(*abm_dce), GFP_KERNEL); + if (ctx->dc->caps.dmcub_support) { + struct dce_abm *abm_dce = kzalloc(sizeof(*abm_dce), GFP_KERNEL); - if (abm_dce == NULL) { - BREAK_TO_DEBUGGER(); - return NULL; - } + if (abm_dce == NULL) { + BREAK_TO_DEBUGGER(); + return NULL; + } - dmub_abm_construct(abm_dce, ctx, regs, abm_shift, abm_mask); + dmub_abm_construct(abm_dce, ctx, regs, abm_shift, abm_mask); - return &abm_dce->base; + return &abm_dce->base; + } + return NULL; } void dmub_abm_destroy(struct abm **abm) diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c new file mode 100644 index 000000000000..39da73eba86e --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c @@ -0,0 +1,252 @@ +/* + * Copyright 2019 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: AMD + * + */ + +#include "dmub_abm.h" +#include "dmub_abm_lcd.h" +#include "dce_abm.h" +#include "dc.h" +#include "dc_dmub_srv.h" +#include "dmub/dmub_srv.h" +#include "core_types.h" +#include "dm_services.h" +#include "reg_helper.h" +#include "fixed31_32.h" + +#ifdef _WIN32 +#include "atombios.h" +#else +#include "atom.h" +#endif + +#define TO_DMUB_ABM(abm)\ + container_of(abm, struct dce_abm, base) + +#define REG(reg) \ + (dce_abm->regs->reg) + +#undef FN +#define FN(reg_name, field_name) \ + dce_abm->abm_shift->field_name, dce_abm->abm_mask->field_name + +#define CTX \ + dce_abm->base.ctx + +#define DISABLE_ABM_IMMEDIATELY 255 + + + +static void dmub_abm_enable_fractional_pwm(struct dc_context *dc) +{ + union dmub_rb_cmd cmd; + uint32_t fractional_pwm = (dc->dc->config.disable_fractional_pwm == false) ? 1 : 0; + uint32_t edp_id_count = dc->dc_edp_id_count; + int i; + uint8_t panel_mask = 0; + + for (i = 0; i < edp_id_count; i++) + panel_mask |= 0x01 << i; + + memset(&cmd, 0, sizeof(cmd)); + cmd.abm_set_pwm_frac.header.type = DMUB_CMD__ABM; + cmd.abm_set_pwm_frac.header.sub_type = DMUB_CMD__ABM_SET_PWM_FRAC; + cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.fractional_pwm = fractional_pwm; + cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1; + cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.panel_mask = panel_mask; + cmd.abm_set_pwm_frac.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_pwm_frac_data); + + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); +} + +void dmub_abm_init(struct abm *abm, uint32_t backlight) +{ + struct dce_abm *dce_abm = TO_DMUB_ABM(abm); + + REG_WRITE(DC_ABM1_HG_SAMPLE_RATE, 0x3); + REG_WRITE(DC_ABM1_HG_SAMPLE_RATE, 0x1); + REG_WRITE(DC_ABM1_LS_SAMPLE_RATE, 0x3); + REG_WRITE(DC_ABM1_LS_SAMPLE_RATE, 0x1); + REG_WRITE(BL1_PWM_BL_UPDATE_SAMPLE_RATE, 0x1); + + REG_SET_3(DC_ABM1_HG_MISC_CTRL, 0, + ABM1_HG_NUM_OF_BINS_SEL, 0, + ABM1_HG_VMAX_SEL, 1, + ABM1_HG_BIN_BITWIDTH_SIZE_SEL, 0); + + REG_SET_3(DC_ABM1_IPCSC_COEFF_SEL, 0, + ABM1_IPCSC_COEFF_SEL_R, 2, + ABM1_IPCSC_COEFF_SEL_G, 4, + ABM1_IPCSC_COEFF_SEL_B, 2); + + REG_UPDATE(BL1_PWM_CURRENT_ABM_LEVEL, + BL1_PWM_CURRENT_ABM_LEVEL, backlight); + + REG_UPDATE(BL1_PWM_TARGET_ABM_LEVEL, + BL1_PWM_TARGET_ABM_LEVEL, backlight); + + REG_UPDATE(BL1_PWM_USER_LEVEL, + BL1_PWM_USER_LEVEL, backlight); + + REG_UPDATE_2(DC_ABM1_LS_MIN_MAX_PIXEL_VALUE_THRES, + ABM1_LS_MIN_PIXEL_VALUE_THRES, 0, + ABM1_LS_MAX_PIXEL_VALUE_THRES, 1000); + + REG_SET_3(DC_ABM1_HGLS_REG_READ_PROGRESS, 0, + ABM1_HG_REG_READ_MISSED_FRAME_CLEAR, 1, + ABM1_LS_REG_READ_MISSED_FRAME_CLEAR, 1, + ABM1_BL_REG_READ_MISSED_FRAME_CLEAR, 1); + + dmub_abm_enable_fractional_pwm(abm->ctx); +} + +unsigned int dmub_abm_get_current_backlight(struct abm *abm) +{ + struct dce_abm *dce_abm = TO_DMUB_ABM(abm); + unsigned int backlight = REG_READ(BL1_PWM_CURRENT_ABM_LEVEL); + + /* return backlight in hardware format which is unsigned 17 bits, with + * 1 bit integer and 16 bit fractional + */ + return backlight; +} + +unsigned int dmub_abm_get_target_backlight(struct abm *abm) +{ + struct dce_abm *dce_abm = TO_DMUB_ABM(abm); + unsigned int backlight = REG_READ(BL1_PWM_TARGET_ABM_LEVEL); + + /* return backlight in hardware format which is unsigned 17 bits, with + * 1 bit integer and 16 bit fractional + */ + return backlight; +} + +bool dmub_abm_set_level(struct abm *abm, uint32_t level, uint8_t panel_mask) +{ + union dmub_rb_cmd cmd; + struct dc_context *dc = abm->ctx; + + memset(&cmd, 0, sizeof(cmd)); + cmd.abm_set_level.header.type = DMUB_CMD__ABM; + cmd.abm_set_level.header.sub_type = DMUB_CMD__ABM_SET_LEVEL; + cmd.abm_set_level.abm_set_level_data.level = level; + cmd.abm_set_level.abm_set_level_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1; + cmd.abm_set_level.abm_set_level_data.panel_mask = panel_mask; + cmd.abm_set_level.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_level_data); + + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); + + return true; +} + +void dmub_abm_init_config(struct abm *abm, + const char *src, + unsigned int bytes, + unsigned int inst) +{ + union dmub_rb_cmd cmd; + struct dc_context *dc = abm->ctx; + uint8_t panel_mask = 0x01 << inst; + + // TODO: Optimize by only reading back final 4 bytes + dmub_flush_buffer_mem(&dc->dmub_srv->dmub->scratch_mem_fb); + + // Copy iramtable into cw7 + memcpy(dc->dmub_srv->dmub->scratch_mem_fb.cpu_addr, (void *)src, bytes); + + memset(&cmd, 0, sizeof(cmd)); + // Fw will copy from cw7 to fw_state + cmd.abm_init_config.header.type = DMUB_CMD__ABM; + cmd.abm_init_config.header.sub_type = DMUB_CMD__ABM_INIT_CONFIG; + cmd.abm_init_config.abm_init_config_data.src.quad_part = dc->dmub_srv->dmub->scratch_mem_fb.gpu_addr; + cmd.abm_init_config.abm_init_config_data.bytes = bytes; + cmd.abm_init_config.abm_init_config_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1; + cmd.abm_init_config.abm_init_config_data.panel_mask = panel_mask; + + cmd.abm_init_config.header.payload_bytes = sizeof(struct dmub_cmd_abm_init_config_data); + + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); + +} + +bool dmub_abm_set_pause(struct abm *abm, bool pause, unsigned int panel_inst, unsigned int stream_inst) +{ + union dmub_rb_cmd cmd; + struct dc_context *dc = abm->ctx; + uint8_t panel_mask = 0x01 << panel_inst; + + memset(&cmd, 0, sizeof(cmd)); + cmd.abm_pause.header.type = DMUB_CMD__ABM; + cmd.abm_pause.header.sub_type = DMUB_CMD__ABM_PAUSE; + cmd.abm_pause.abm_pause_data.enable = pause; + cmd.abm_pause.abm_pause_data.panel_mask = panel_mask; + cmd.abm_set_level.header.payload_bytes = sizeof(struct dmub_cmd_abm_pause_data); + + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); + + return true; +} + +bool dmub_abm_set_pipe(struct abm *abm, uint32_t otg_inst, uint32_t option, uint32_t panel_inst) +{ + union dmub_rb_cmd cmd; + struct dc_context *dc = abm->ctx; + uint32_t ramping_boundary = 0xFFFF; + + memset(&cmd, 0, sizeof(cmd)); + cmd.abm_set_pipe.header.type = DMUB_CMD__ABM; + cmd.abm_set_pipe.header.sub_type = DMUB_CMD__ABM_SET_PIPE; + cmd.abm_set_pipe.abm_set_pipe_data.otg_inst = otg_inst; + cmd.abm_set_pipe.abm_set_pipe_data.set_pipe_option = option; + cmd.abm_set_pipe.abm_set_pipe_data.panel_inst = panel_inst; + cmd.abm_set_pipe.abm_set_pipe_data.ramping_boundary = ramping_boundary; + cmd.abm_set_pipe.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_pipe_data); + + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); + + return true; +} + +bool dmub_abm_set_backlight_level(struct abm *abm, + unsigned int backlight_pwm_u16_16, + unsigned int frame_ramp, + unsigned int panel_inst) +{ + union dmub_rb_cmd cmd; + struct dc_context *dc = abm->ctx; + + memset(&cmd, 0, sizeof(cmd)); + cmd.abm_set_backlight.header.type = DMUB_CMD__ABM; + cmd.abm_set_backlight.header.sub_type = DMUB_CMD__ABM_SET_BACKLIGHT; + cmd.abm_set_backlight.abm_set_backlight_data.frame_ramp = frame_ramp; + cmd.abm_set_backlight.abm_set_backlight_data.backlight_user_level = backlight_pwm_u16_16; + cmd.abm_set_backlight.abm_set_backlight_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1; + cmd.abm_set_backlight.abm_set_backlight_data.panel_mask = (0x01 << panel_inst); + cmd.abm_set_backlight.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_backlight_data); + + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); + + return true; +} + diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.h b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.h new file mode 100644 index 000000000000..00b4e268768e --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.h @@ -0,0 +1,46 @@ +/* + * Copyright 2019 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: AMD + * + */ + +#ifndef __DMUB_ABM_LCD_H__ +#define __DMUB_ABM_LCD_H__ + +#include "abm.h" + +void dmub_abm_init(struct abm *abm, uint32_t backlight); +bool dmub_abm_set_level(struct abm *abm, uint32_t level, uint8_t panel_mask); +unsigned int dmub_abm_get_current_backlight(struct abm *abm); +unsigned int dmub_abm_get_target_backlight(struct abm *abm); +void dmub_abm_init_config(struct abm *abm, + const char *src, + unsigned int bytes, + unsigned int inst); + +bool dmub_abm_set_pause(struct abm *abm, bool pause, unsigned int panel_inst, unsigned int stream_inst); +bool dmub_abm_set_pipe(struct abm *abm, uint32_t otg_inst, uint32_t option, uint32_t panel_inst); +bool dmub_abm_set_backlight_level(struct abm *abm, + unsigned int backlight_pwm_u16_16, + unsigned int frame_ramp, + unsigned int panel_inst); +#endif diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c index 3f32e9c3fbaf..2aa0e01a6891 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c @@ -47,9 +47,7 @@ void dmub_hw_lock_mgr_cmd(struct dc_dmub_srv *dmub_srv, if (!lock) cmd.lock_hw.lock_hw_data.should_release = 1; - dc_dmub_srv_cmd_queue(dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dmub_srv); - dc_dmub_srv_wait_idle(dmub_srv); + dm_execute_dmub_cmd(dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } void dmub_hw_lock_mgr_inbox0_cmd(struct dc_dmub_srv *dmub_srv, diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_outbox.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_outbox.c index fff1d07d865d..d8009b2dc56a 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dmub_outbox.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_outbox.c @@ -48,7 +48,5 @@ void dmub_enable_outbox_notification(struct dc_dmub_srv *dmub_srv) sizeof(cmd.outbox1_enable.header); cmd.outbox1_enable.enable = true; - dc_dmub_srv_cmd_queue(dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dmub_srv); - dc_dmub_srv_wait_idle(dmub_srv); + dm_execute_dmub_cmd(dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c index 9705d8f88382..0f24b6fbd220 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c @@ -87,6 +87,8 @@ static enum dc_psr_state convert_psr_state(uint32_t raw_state) state = PSR_STATE4c_FULL_FRAME; else if (raw_state == 0x4E) state = PSR_STATE4_FULL_FRAME_POWERUP; + else if (raw_state == 0x4F) + state = PSR_STATE4_FULL_FRAME_HW_LOCK; else if (raw_state == 0x60) state = PSR_STATE_HWLOCK_MGR; else if (raw_state == 0x61) @@ -168,9 +170,7 @@ static bool dmub_psr_set_version(struct dmub_psr *dmub, struct dc_stream_state * cmd.psr_set_version.psr_set_version_data.panel_inst = panel_inst; cmd.psr_set_version.header.payload_bytes = sizeof(struct dmub_cmd_psr_set_version_data); - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); return true; } @@ -198,9 +198,7 @@ static void dmub_psr_enable(struct dmub_psr *dmub, bool enable, bool wait, uint8 cmd.psr_enable.header.payload_bytes = 0; // Send header only - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); + dm_execute_dmub_cmd(dc->dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); /* Below loops 1000 x 500us = 500 ms. * Exit PSR may need to wait 1-2 frames to power up. Timeout after at @@ -248,9 +246,7 @@ static void dmub_psr_set_level(struct dmub_psr *dmub, uint16_t psr_level, uint8_ cmd.psr_set_level.psr_set_level_data.psr_level = psr_level; cmd.psr_set_level.psr_set_level_data.cmd_version = DMUB_CMD_PSR_CONTROL_VERSION_1; cmd.psr_set_level.psr_set_level_data.panel_inst = panel_inst; - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } /* @@ -269,9 +265,7 @@ static void dmub_psr_set_sink_vtotal_in_psr_active(struct dmub_psr *dmub, cmd.psr_set_vtotal.psr_set_vtotal_data.psr_vtotal_idle = psr_vtotal_idle; cmd.psr_set_vtotal.psr_set_vtotal_data.psr_vtotal_su = psr_vtotal_su; - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } /* @@ -290,9 +284,7 @@ static void dmub_psr_set_power_opt(struct dmub_psr *dmub, unsigned int power_opt cmd.psr_set_power_opt.psr_set_power_opt_data.power_opt = power_opt; cmd.psr_set_power_opt.psr_set_power_opt_data.panel_inst = panel_inst; - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } /* @@ -422,9 +414,7 @@ static bool dmub_psr_copy_settings(struct dmub_psr *dmub, copy_settings_data->relock_delay_frame_cnt = 2; copy_settings_data->dsc_slice_height = psr_context->dsc_slice_height; - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); return true; } @@ -445,9 +435,7 @@ static void dmub_psr_force_static(struct dmub_psr *dmub, uint8_t panel_inst) cmd.psr_force_static.header.sub_type = DMUB_CMD__PSR_FORCE_STATIC; cmd.psr_enable.header.payload_bytes = 0; - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } /* diff --git a/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c b/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c index 54805802cbd5..42e9b6a529f6 100644 --- a/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c @@ -401,6 +401,10 @@ static const struct dc_plane_cap plane_cap = { } }; +static const struct dc_debug_options debug_defaults = { + .enable_legacy_fast_update = true, +}; + #define CTX ctx #define REG(reg) mm ## reg @@ -1071,6 +1075,7 @@ static bool dce100_resource_construct( dc->caps.dual_link_dvi = true; dc->caps.disable_dp_clk_share = true; dc->caps.extended_aux_timeout_support = false; + dc->debug = debug_defaults; for (i = 0; i < pool->base.pipe_count; i++) { pool->base.timing_generators[i] = diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c index 8d2460d06bce..6c9ca43d1040 100644 --- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c @@ -209,9 +209,6 @@ static bool dce110_enable_display_power_gating( struct dc_context *ctx = dc->ctx; unsigned int underlay_idx = dc->res_pool->underlay_pipe_index; - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) - return true; - if (power_gating == PIPE_GATING_CONTROL_INIT) cntl = ASIC_PIPE_INIT; else if (power_gating == PIPE_GATING_CONTROL_ENABLE) @@ -1219,7 +1216,8 @@ void dce110_blank_stream(struct pipe_ctx *pipe_ctx) struct dce_hwseq *hws = link->dc->hwseq; if (link->local_sink && link->local_sink->sink_signal == SIGNAL_TYPE_EDP) { - hws->funcs.edp_backlight_control(link, false); + if (!stream->skip_edp_power_down) + hws->funcs.edp_backlight_control(link, false); link->dc->hwss.set_abm_immediate_disable(pipe_ctx); } @@ -2291,6 +2289,11 @@ enum dc_status dce110_apply_ctx_to_hw( if (DC_OK != status) return status; + +#ifdef CONFIG_DRM_AMD_DC_FP + if (hws->funcs.resync_fifo_dccg_dio) + hws->funcs.resync_fifo_dccg_dio(hws, dc, context); +#endif } if (dc->fbc_compressor) diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c index a4a45a6ce61e..46eca5a21e1c 100644 --- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c @@ -424,6 +424,10 @@ static const struct dc_plane_cap plane_cap = { 64 }; +static const struct dc_debug_options debug_defaults = { + .enable_legacy_fast_update = true, +}; + static const struct dc_plane_cap underlay_plane_cap = { .type = DC_PLANE_TYPE_DCE_UNDERLAY, .per_pixel_alpha = 1, @@ -1368,6 +1372,7 @@ static bool dce110_resource_construct( dc->caps.min_horizontal_blanking_period = 80; dc->caps.is_apu = true; dc->caps.extended_aux_timeout_support = false; + dc->debug = debug_defaults; /************************************************* * Create resources * diff --git a/drivers/gpu/drm/amd/display/dc/dce112/dce112_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce112/dce112_hw_sequencer.c index 19873ee1f78d..690caaaff019 100644 --- a/drivers/gpu/drm/amd/display/dc/dce112/dce112_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dce112/dce112_hw_sequencer.c @@ -120,9 +120,6 @@ static bool dce112_enable_display_power_gating( enum bp_pipe_control_action cntl; struct dc_context *ctx = dc->ctx; - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) - return true; - if (power_gating == PIPE_GATING_CONTROL_INIT) cntl = ASIC_PIPE_INIT; else if (power_gating == PIPE_GATING_CONTROL_ENABLE) diff --git a/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c b/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c index e179e80667d1..808855886183 100644 --- a/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c @@ -429,6 +429,10 @@ static const struct dc_plane_cap plane_cap = { 64 }; +static const struct dc_debug_options debug_defaults = { + .enable_legacy_fast_update = true, +}; + #define CTX ctx #define REG(reg) mm ## reg @@ -1239,6 +1243,7 @@ static bool dce112_resource_construct( dc->caps.min_horizontal_blanking_period = 80; dc->caps.dual_link_dvi = true; dc->caps.extended_aux_timeout_support = false; + dc->debug = debug_defaults; /************************************************* * Create resources * diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c index d4afe6c824d2..45e08c4d5861 100644 --- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c @@ -159,9 +159,6 @@ static bool dce120_enable_display_power_gating( enum bp_pipe_control_action cntl; struct dc_context *ctx = dc->ctx; - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) - return true; - if (power_gating == PIPE_GATING_CONTROL_INIT) cntl = ASIC_PIPE_INIT; else if (power_gating == PIPE_GATING_CONTROL_ENABLE) diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c b/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c index af631085e88c..18c5a86d2d61 100644 --- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c @@ -526,6 +526,7 @@ static const struct dc_plane_cap plane_cap = { static const struct dc_debug_options debug_defaults = { .disable_clock_gate = true, + .enable_legacy_fast_update = true, }; static struct clock_source *dce120_clock_source_create( diff --git a/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c b/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c index 5825e6f412bd..3935fd455f0f 100644 --- a/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c @@ -418,6 +418,10 @@ static const struct dc_plane_cap plane_cap = { } }; +static const struct dc_debug_options debug_defaults = { + .enable_legacy_fast_update = true, +}; + static const struct dce_dmcu_registers dmcu_regs = { DMCU_DCE80_REG_LIST() }; @@ -969,6 +973,7 @@ static bool dce80_construct( dc->caps.min_horizontal_blanking_period = 80; dc->caps.dual_link_dvi = true; dc->caps.extended_aux_timeout_support = false; + dc->debug = debug_defaults; /************************************************* * Create resources * @@ -1369,6 +1374,7 @@ static bool dce83_construct( dc->caps.max_cursor_size = 128; dc->caps.min_horizontal_blanking_period = 80; dc->caps.is_apu = true; + dc->debug = debug_defaults; /************************************************* * Create resources * diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h index 0b17c2993ca5..09784222cc03 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h @@ -690,6 +690,8 @@ struct dcn_hubp_state { uint32_t primary_surface_addr_hi; uint32_t primary_meta_addr_lo; uint32_t primary_meta_addr_hi; + uint32_t uclk_pstate_force; + uint32_t hubp_cntl; }; struct dcn10_hubp { diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c index 1c3b6f25a782..20a1582be0b1 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c @@ -1012,31 +1012,29 @@ static void dcn10_reset_back_end_for_pipe( return; } - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - link = pipe_ctx->stream->link; - /* DPMS may already disable or */ - /* dpms_off status is incorrect due to fastboot - * feature. When system resume from S4 with second - * screen only, the dpms_off would be true but - * VBIOS lit up eDP, so check link status too. - */ - if (!pipe_ctx->stream->dpms_off || link->link_status.link_active) - dc->link_srv->set_dpms_off(pipe_ctx); - else if (pipe_ctx->stream_res.audio) - dc->hwss.disable_audio_stream(pipe_ctx); - - if (pipe_ctx->stream_res.audio) { - /*disable az_endpoint*/ - pipe_ctx->stream_res.audio->funcs->az_disable(pipe_ctx->stream_res.audio); - - /*free audio*/ - if (dc->caps.dynamic_audio == true) { - /*we have to dynamic arbitrate the audio endpoints*/ - /*we free the resource, need reset is_audio_acquired*/ - update_audio_usage(&dc->current_state->res_ctx, dc->res_pool, - pipe_ctx->stream_res.audio, false); - pipe_ctx->stream_res.audio = NULL; - } + link = pipe_ctx->stream->link; + /* DPMS may already disable or */ + /* dpms_off status is incorrect due to fastboot + * feature. When system resume from S4 with second + * screen only, the dpms_off would be true but + * VBIOS lit up eDP, so check link status too. + */ + if (!pipe_ctx->stream->dpms_off || link->link_status.link_active) + dc->link_srv->set_dpms_off(pipe_ctx); + else if (pipe_ctx->stream_res.audio) + dc->hwss.disable_audio_stream(pipe_ctx); + + if (pipe_ctx->stream_res.audio) { + /*disable az_endpoint*/ + pipe_ctx->stream_res.audio->funcs->az_disable(pipe_ctx->stream_res.audio); + + /*free audio*/ + if (dc->caps.dynamic_audio == true) { + /*we have to dynamic arbitrate the audio endpoints*/ + /*we free the resource, need reset is_audio_acquired*/ + update_audio_usage(&dc->current_state->res_ctx, dc->res_pool, + pipe_ctx->stream_res.audio, false); + pipe_ctx->stream_res.audio = NULL; } } @@ -1499,54 +1497,32 @@ void dcn10_init_hw(struct dc *dc) if (dc->res_pool->dccg && dc->res_pool->dccg->funcs->dccg_init) dc->res_pool->dccg->funcs->dccg_init(res_pool->dccg); - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - - REG_WRITE(REFCLK_CNTL, 0); - REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1); - REG_WRITE(DIO_MEM_PWR_CTRL, 0); - - if (!dc->debug.disable_clock_gate) { - /* enable all DCN clock gating */ - REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0); - - REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0); - - REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0); - } - - //Enable ability to power gate / don't force power on permanently - if (hws->funcs.enable_power_gating_plane) - hws->funcs.enable_power_gating_plane(hws, true); - - return; - } - if (!dcb->funcs->is_accelerated_mode(dcb)) hws->funcs.disable_vga(dc->hwseq); - hws->funcs.bios_golden_init(dc); + if (!dc_dmub_srv_optimized_init_done(dc->ctx->dmub_srv)) + hws->funcs.bios_golden_init(dc); + if (dc->ctx->dc_bios->fw_info_valid) { res_pool->ref_clocks.xtalin_clock_inKhz = dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency; - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - if (res_pool->dccg && res_pool->hubbub) { + if (res_pool->dccg && res_pool->hubbub) { - (res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg, - dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency, - &res_pool->ref_clocks.dccg_ref_clock_inKhz); + (res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg, + dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency, + &res_pool->ref_clocks.dccg_ref_clock_inKhz); - (res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub, - res_pool->ref_clocks.dccg_ref_clock_inKhz, - &res_pool->ref_clocks.dchub_ref_clock_inKhz); - } else { - // Not all ASICs have DCCG sw component - res_pool->ref_clocks.dccg_ref_clock_inKhz = - res_pool->ref_clocks.xtalin_clock_inKhz; - res_pool->ref_clocks.dchub_ref_clock_inKhz = - res_pool->ref_clocks.xtalin_clock_inKhz; - } + (res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub, + res_pool->ref_clocks.dccg_ref_clock_inKhz, + &res_pool->ref_clocks.dchub_ref_clock_inKhz); + } else { + // Not all ASICs have DCCG sw component + res_pool->ref_clocks.dccg_ref_clock_inKhz = + res_pool->ref_clocks.xtalin_clock_inKhz; + res_pool->ref_clocks.dchub_ref_clock_inKhz = + res_pool->ref_clocks.xtalin_clock_inKhz; } } else ASSERT_CRITICAL(false); @@ -1923,6 +1899,11 @@ void dcn10_pipe_control_lock( * * TODO: Optimize cursor programming to be once per frame before VUPDATE * to avoid the need for this workaround. + * + * @dc: Current DC state + * @pipe_ctx: Pipe_ctx pointer for delayed cursor update + * + * Return: void */ static void delay_cursor_until_vupdate(struct dc *dc, struct pipe_ctx *pipe_ctx) { @@ -2600,23 +2581,15 @@ static void dcn10_update_dpp(struct dpp *dpp, struct dc_plane_state *plane_state dpp->funcs->dpp_program_bias_and_scale(dpp, &bns_params); } -void dcn10_update_visual_confirm_color(struct dc *dc, struct pipe_ctx *pipe_ctx, struct tg_color *color, int mpcc_id) +void dcn10_update_visual_confirm_color(struct dc *dc, + struct pipe_ctx *pipe_ctx, + int mpcc_id) { struct mpc *mpc = dc->res_pool->mpc; - if (dc->debug.visual_confirm == VISUAL_CONFIRM_HDR) - get_hdr_visual_confirm_color(pipe_ctx, color); - else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SURFACE) - get_surface_visual_confirm_color(pipe_ctx, color); - else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SWIZZLE) - get_surface_tile_visual_confirm_color(pipe_ctx, color); - else - color_space_to_black_color( - dc, pipe_ctx->stream->output_color_space, color); - if (mpc->funcs->set_bg_color) { - memcpy(&pipe_ctx->plane_state->visual_confirm_color, color, sizeof(struct tg_color)); - mpc->funcs->set_bg_color(mpc, color, mpcc_id); + memcpy(&pipe_ctx->plane_state->visual_confirm_color, &(pipe_ctx->visual_confirm_color), sizeof(struct tg_color)); + mpc->funcs->set_bg_color(mpc, &(pipe_ctx->visual_confirm_color), mpcc_id); } } @@ -2669,7 +2642,7 @@ void dcn10_update_mpcc(struct dc *dc, struct pipe_ctx *pipe_ctx) /* If there is no full update, don't need to touch MPC tree*/ if (!pipe_ctx->plane_state->update_flags.bits.full_update) { mpc->funcs->update_blending(mpc, &blnd_cfg, mpcc_id); - dc->hwss.update_visual_confirm_color(dc, pipe_ctx, &blnd_cfg.black_color, mpcc_id); + dc->hwss.update_visual_confirm_color(dc, pipe_ctx, mpcc_id); return; } @@ -2691,7 +2664,7 @@ void dcn10_update_mpcc(struct dc *dc, struct pipe_ctx *pipe_ctx) NULL, hubp->inst, mpcc_id); - dc->hwss.update_visual_confirm_color(dc, pipe_ctx, &blnd_cfg.black_color, mpcc_id); + dc->hwss.update_visual_confirm_color(dc, pipe_ctx, mpcc_id); ASSERT(new_mpcc != NULL); hubp->opp_id = pipe_ctx->stream_res.opp->inst; @@ -3076,15 +3049,13 @@ void dcn10_prepare_bandwidth( if (dc->debug.sanity_checks) hws->funcs.verify_allow_pstate_change_high(dc); - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - if (context->stream_count == 0) - context->bw_ctx.bw.dcn.clk.phyclk_khz = 0; + if (context->stream_count == 0) + context->bw_ctx.bw.dcn.clk.phyclk_khz = 0; - dc->clk_mgr->funcs->update_clocks( - dc->clk_mgr, - context, - false); - } + dc->clk_mgr->funcs->update_clocks( + dc->clk_mgr, + context, + false); dc->wm_optimized_required = hubbub->funcs->program_watermarks(hubbub, &context->bw_ctx.bw.dcn.watermarks, @@ -3116,15 +3087,13 @@ void dcn10_optimize_bandwidth( if (dc->debug.sanity_checks) hws->funcs.verify_allow_pstate_change_high(dc); - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - if (context->stream_count == 0) - context->bw_ctx.bw.dcn.clk.phyclk_khz = 0; + if (context->stream_count == 0) + context->bw_ctx.bw.dcn.clk.phyclk_khz = 0; - dc->clk_mgr->funcs->update_clocks( - dc->clk_mgr, - context, - true); - } + dc->clk_mgr->funcs->update_clocks( + dc->clk_mgr, + context, + true); hubbub->funcs->program_watermarks(hubbub, &context->bw_ctx.bw.dcn.watermarks, diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h index 0ef7bf7ddb75..ef6d56da417c 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h @@ -202,7 +202,6 @@ void dcn10_get_dcc_en_bits(struct dc *dc, int *dcc_en_bits); void dcn10_update_visual_confirm_color( struct dc *dc, struct pipe_ctx *pipe_ctx, - struct tg_color *color, int mpcc_id); #endif /* __DC_HWSS_DCN10_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c index a0f8e31d2adc..46a2ebcabd1a 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c @@ -45,7 +45,8 @@ #include "dcn10_cm_common.h" #include "clk_mgr.h" -unsigned int snprintf_count(char *pBuf, unsigned int bufSize, char *fmt, ...) +__printf(3, 4) +unsigned int snprintf_count(char *pbuf, unsigned int bufsize, char *fmt, ...) { int ret_vsnprintf; unsigned int chars_printed; @@ -53,15 +54,15 @@ unsigned int snprintf_count(char *pBuf, unsigned int bufSize, char *fmt, ...) va_list args; va_start(args, fmt); - ret_vsnprintf = vsnprintf(pBuf, bufSize, fmt, args); + ret_vsnprintf = vsnprintf(pbuf, bufsize, fmt, args); va_end(args); if (ret_vsnprintf > 0) { - if (ret_vsnprintf < bufSize) + if (ret_vsnprintf < bufsize) chars_printed = ret_vsnprintf; else - chars_printed = bufSize - 1; + chars_printed = bufsize - 1; } else chars_printed = 0; diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_opp.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_opp.c index 41cec7acf51f..0dec57679269 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_opp.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_opp.c @@ -37,14 +37,14 @@ #define CTX \ oppn10->base.ctx - -/************* FORMATTER ************/ - /** - * set_truncation + * opp1_set_truncation(): * 1) set truncation depth: 0 for 18 bpp or 1 for 24 bpp * 2) enable truncation * 3) HW remove 12bit FMT support for DCE11 power saving reason. + * + * @oppn10: output_pixel_processor struct instance for dcn10. + * @params: pointer to bit_depth_reduction_params. */ static void opp1_set_truncation( struct dcn10_opp *oppn10, @@ -149,11 +149,12 @@ void opp1_program_bit_depth_reduction( } /** - * set_pixel_encoding - * - * Set Pixel Encoding + * opp1_set_pixel_encoding(): * 0: RGB 4:4:4 or YCbCr 4:4:4 or YOnly * 1: YCbCr 4:2:2 + * + * @oppn10: output_pixel_processor struct instance for dcn10. + * @params: pointer to clamping_and_pixel_encoding_params. */ static void opp1_set_pixel_encoding( struct dcn10_opp *oppn10, @@ -180,13 +181,16 @@ static void opp1_set_pixel_encoding( } /** - * Set Clamping + * opp1_set_clamping(): * 1) Set clamping format based on bpc - 0 for 6bpc (No clamping) * 1 for 8 bpc * 2 for 10 bpc * 3 for 12 bpc * 7 for programable * 2) Enable clamp if Limited range requested + * + * @oppn10: output_pixel_processor struct instance for dcn10. + * @params: pointer to clamping_and_pixel_encoding_params. */ static void opp1_set_clamping( struct dcn10_opp *oppn10, diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c index c9e53dc49c92..0e8f4f36c87c 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c @@ -42,11 +42,13 @@ #define STATIC_SCREEN_EVENT_MASK_RANGETIMING_DOUBLE_BUFFER_UPDATE_EN 0x100 /** -* apply_front_porch_workaround TODO FPGA still need? -* -* This is a workaround for a bug that has existed since R5xx and has not been -* fixed keep Front porch at minimum 2 for Interlaced mode or 1 for progressive. -*/ + * apply_front_porch_workaround() - This is a workaround for a bug that has + * existed since R5xx and has not been fixed + * keep Front porch at minimum 2 for Interlaced + * mode or 1 for progressive. + * + * @timing: Timing parameters used to configure DCN blocks. + */ static void apply_front_porch_workaround(struct dc_crtc_timing *timing) { if (timing->flags.INTERLACE == 1) { @@ -133,9 +135,20 @@ void optc1_setup_vertical_interrupt2( } /** - * program_timing_generator used by mode timing set - * Program CRTC Timing Registers - OTG_H_*, OTG_V_*, Pixel repetition. - * Including SYNC. Call BIOS command table to program Timings. + * optc1_program_timing() - used by mode timing set Program + * CRTC Timing Registers - OTG_H_*, + * OTG_V_*, Pixel repetition. + * Including SYNC. Call BIOS command table to program Timings. + * + * @optc: timing_generator instance. + * @dc_crtc_timing: Timing parameters used to configure DCN blocks. + * @vready_offset: Vready's starting position. + * @vstartup_start: Vstartup period. + * @vupdate_offset: Vupdate starting position. + * @vupdate_width: Vupdate duration. + * @signal: DC signal types. + * @use_vbios: to program timings from BIOS command table. + * */ void optc1_program_timing( struct timing_generator *optc, @@ -385,6 +398,9 @@ void optc1_set_blank_data_double_buffer(struct timing_generator *optc, bool enab * Sets double buffer point for V_TOTAL, H_TOTAL, VTOTAL_MIN, * VTOTAL_MAX, VTOTAL_MIN_SEL and VTOTAL_MAX_SEL registers. * + * @optc: timing_generator instance. + * @enable: Enable DRR double buffering control if true, disable otherwise. + * * Options: any time, start of frame, dp start of frame (range timing) */ void optc1_set_timing_double_buffer(struct timing_generator *optc, bool enable) @@ -397,8 +413,9 @@ void optc1_set_timing_double_buffer(struct timing_generator *optc, bool enable) } /** - * unblank_crtc - * Call ASIC Control Object to UnBlank CRTC. + * optc1_unblank_crtc() - Call ASIC Control Object to UnBlank CRTC. + * + * @optc: timing_generator instance. */ static void optc1_unblank_crtc(struct timing_generator *optc) { @@ -419,8 +436,9 @@ static void optc1_unblank_crtc(struct timing_generator *optc) } /** - * blank_crtc - * Call ASIC Control Object to Blank CRTC. + * optc1_blank_crtc() - Call ASIC Control Object to Blank CRTC. + * + * @optc: timing_generator instance. */ static void optc1_blank_crtc(struct timing_generator *optc) @@ -493,8 +511,9 @@ void optc1_enable_optc_clock(struct timing_generator *optc, bool enable) } /** - * Enable CRTC - * Enable CRTC - call ASIC Control Object to enable Timing generator. + * optc1_enable_crtc() - Enable CRTC - call ASIC Control Object to enable Timing generator. + * + * @optc: timing_generator instance. */ static bool optc1_enable_crtc(struct timing_generator *optc) { @@ -653,11 +672,9 @@ void optc1_lock(struct timing_generator *optc) REG_SET(OTG_MASTER_UPDATE_LOCK, 0, OTG_MASTER_UPDATE_LOCK, 1); - /* Should be fast, status does not update on maximus */ - if (optc->ctx->dce_environment != DCE_ENV_FPGA_MAXIMUS) - REG_WAIT(OTG_MASTER_UPDATE_LOCK, - UPDATE_LOCK_STATUS, 1, - 1, 10); + REG_WAIT(OTG_MASTER_UPDATE_LOCK, + UPDATE_LOCK_STATUS, 1, + 1, 10); TRACE_OPTC_LOCK_UNLOCK_STATE(optc1, optc->inst, true); } @@ -892,15 +909,11 @@ static void optc1_program_manual_trigger(struct timing_generator *optc) MANUAL_FLOW_CONTROL, 0); } - /** - ***************************************************************************** - * Function: set_drr + * optc1_set_drr() - Program dynamic refresh rate registers m_OTGx_OTG_V_TOTAL_*. * - * @brief - * Program dynamic refresh rate registers m_OTGx_OTG_V_TOTAL_*. - * - ***************************************************************************** + * @optc: timing_generator instance. + * @params: parameters used for Dynamic Refresh Rate. */ void optc1_set_drr( struct timing_generator *optc, @@ -932,19 +945,10 @@ void optc1_set_drr( OTG_FORCE_LOCK_ON_EVENT, 0, OTG_SET_V_TOTAL_MIN_MASK_EN, 0, OTG_SET_V_TOTAL_MIN_MASK, 0); - - // Setup manual flow control for EOF via TRIG_A - optc->funcs->setup_manual_trigger(optc); - - } else { - REG_UPDATE_4(OTG_V_TOTAL_CONTROL, - OTG_SET_V_TOTAL_MIN_MASK, 0, - OTG_V_TOTAL_MIN_SEL, 0, - OTG_V_TOTAL_MAX_SEL, 0, - OTG_FORCE_LOCK_ON_EVENT, 0); - - optc->funcs->set_vtotal_min_max(optc, 0, 0); } + + // Setup manual flow control for EOF via TRIG_A + optc->funcs->setup_manual_trigger(optc); } void optc1_set_vtotal_min_max(struct timing_generator *optc, int vtotal_min, int vtotal_max) diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c index 21ec1ba5ed75..4b02f8443534 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c @@ -553,6 +553,7 @@ static const struct dc_debug_options debug_defaults_drv = { .recovery_enabled = false, /*enable this by default after testing.*/ .max_downscale_src_width = 3840, .underflow_assert_delay_us = 0xFFFFFFFF, + .enable_legacy_fast_update = true, }; static const struct dc_debug_options debug_defaults_diags = { @@ -886,13 +887,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn10_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hwseq = dcn10_hwseq_create, -}; - static void dcn10_clock_source_destroy(struct clock_source **clk_src) { kfree(TO_DCE110_CLK_SRC(*clk_src)); @@ -1651,9 +1645,8 @@ static bool dcn10_resource_construct( } if (!resource_construct(num_virtual_links, dc, &pool->base, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) - goto fail; + &res_create_funcs)) + goto fail; dcn10_hw_sequencer_construct(dc); dc->caps.max_planes = pool->base.pipe_count; diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dccg.h b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dccg.h index 7bdc146f7cb5..c8602bcfa393 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dccg.h +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dccg.h @@ -208,7 +208,9 @@ #define DCCG314_REG_FIELD_LIST(type) \ type DSCCLK3_DTO_PHASE;\ type DSCCLK3_DTO_MODULO;\ - type DSCCLK3_DTO_ENABLE; + type DSCCLK3_DTO_ENABLE;\ + type DENTIST_DISPCLK_RDIVIDER;\ + type DENTIST_DISPCLK_WDIVIDER; #define DCCG32_REG_FIELD_LIST(type) \ type DPSTREAMCLK0_EN;\ diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dsc.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dsc.c index 5bd698cd6d20..5eebe7f03ddc 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dsc.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dsc.c @@ -30,22 +30,13 @@ #include "dsc/dscc_types.h" #include "dsc/rc_calc.h" -static void dsc_log_pps(struct display_stream_compressor *dsc, struct drm_dsc_config *pps); -static bool dsc_prepare_config(const struct dsc_config *dsc_cfg, struct dsc_reg_values *dsc_reg_vals, - struct dsc_optc_config *dsc_optc_cfg); -static void dsc_init_reg_values(struct dsc_reg_values *reg_vals); -static void dsc_update_from_dsc_parameters(struct dsc_reg_values *reg_vals, const struct dsc_parameters *dsc_params); static void dsc_write_to_registers(struct display_stream_compressor *dsc, const struct dsc_reg_values *reg_vals); -static enum dsc_pixel_format dsc_dc_pixel_encoding_to_dsc_pixel_format(enum dc_pixel_encoding dc_pix_enc, bool is_ycbcr422_simple); -static enum dsc_bits_per_comp dsc_dc_color_depth_to_dsc_bits_per_comp(enum dc_color_depth); /* Object I/F functions */ -static void dsc2_get_enc_caps(struct dsc_enc_caps *dsc_enc_caps, int pixel_clock_100Hz); static void dsc2_read_state(struct display_stream_compressor *dsc, struct dcn_dsc_state *s); static bool dsc2_validate_stream(struct display_stream_compressor *dsc, const struct dsc_config *dsc_cfg); static void dsc2_set_config(struct display_stream_compressor *dsc, const struct dsc_config *dsc_cfg, struct dsc_optc_config *dsc_optc_cfg); -static bool dsc2_get_packed_pps(struct display_stream_compressor *dsc, const struct dsc_config *dsc_cfg, uint8_t *dsc_packed_pps); static void dsc2_enable(struct display_stream_compressor *dsc, int opp_pipe); static void dsc2_disable(struct display_stream_compressor *dsc); static void dsc2_disconnect(struct display_stream_compressor *dsc); @@ -108,7 +99,7 @@ void dsc2_construct(struct dcn20_dsc *dsc, /* This returns the capabilities for a single DSC encoder engine. Number of slices and total throughput * can be doubled, tripled etc. by using additional DSC engines. */ -static void dsc2_get_enc_caps(struct dsc_enc_caps *dsc_enc_caps, int pixel_clock_100Hz) +void dsc2_get_enc_caps(struct dsc_enc_caps *dsc_enc_caps, int pixel_clock_100Hz) { dsc_enc_caps->dsc_version = 0x21; /* v1.2 - DP spec defined it in reverse order and we kept it */ @@ -184,7 +175,7 @@ static bool dsc2_validate_stream(struct display_stream_compressor *dsc, const st } -static void dsc_config_log(struct display_stream_compressor *dsc, const struct dsc_config *config) +void dsc_config_log(struct display_stream_compressor *dsc, const struct dsc_config *config) { DC_LOG_DSC("\tnum_slices_h %d", config->dc_dsc_cfg.num_slices_h); DC_LOG_DSC("\tnum_slices_v %d", config->dc_dsc_cfg.num_slices_v); @@ -211,7 +202,7 @@ static void dsc2_set_config(struct display_stream_compressor *dsc, const struct } -static bool dsc2_get_packed_pps(struct display_stream_compressor *dsc, const struct dsc_config *dsc_cfg, uint8_t *dsc_packed_pps) +bool dsc2_get_packed_pps(struct display_stream_compressor *dsc, const struct dsc_config *dsc_cfg, uint8_t *dsc_packed_pps) { bool is_config_ok; struct dsc_reg_values dsc_reg_vals; @@ -291,7 +282,7 @@ static void dsc2_disconnect(struct display_stream_compressor *dsc) } /* This module's internal functions */ -static void dsc_log_pps(struct display_stream_compressor *dsc, struct drm_dsc_config *pps) +void dsc_log_pps(struct display_stream_compressor *dsc, struct drm_dsc_config *pps) { int i; int bits_per_pixel = pps->bits_per_pixel; @@ -345,7 +336,7 @@ static void dsc_log_pps(struct display_stream_compressor *dsc, struct drm_dsc_co } } -static void dsc_override_rc_params(struct rc_params *rc, const struct dc_dsc_rc_params_override *override) +void dsc_override_rc_params(struct rc_params *rc, const struct dc_dsc_rc_params_override *override) { uint8_t i; @@ -372,7 +363,7 @@ static void dsc_override_rc_params(struct rc_params *rc, const struct dc_dsc_rc_ rc->flatness_det_thresh = override->flatness_det_thresh; } -static bool dsc_prepare_config(const struct dsc_config *dsc_cfg, struct dsc_reg_values *dsc_reg_vals, +bool dsc_prepare_config(const struct dsc_config *dsc_cfg, struct dsc_reg_values *dsc_reg_vals, struct dsc_optc_config *dsc_optc_cfg) { struct dsc_parameters dsc_params; @@ -463,7 +454,7 @@ static bool dsc_prepare_config(const struct dsc_config *dsc_cfg, struct dsc_reg_ } -static enum dsc_pixel_format dsc_dc_pixel_encoding_to_dsc_pixel_format(enum dc_pixel_encoding dc_pix_enc, bool is_ycbcr422_simple) +enum dsc_pixel_format dsc_dc_pixel_encoding_to_dsc_pixel_format(enum dc_pixel_encoding dc_pix_enc, bool is_ycbcr422_simple) { enum dsc_pixel_format dsc_pix_fmt = DSC_PIXFMT_UNKNOWN; @@ -495,7 +486,7 @@ static enum dsc_pixel_format dsc_dc_pixel_encoding_to_dsc_pixel_format(enum dc_p } -static enum dsc_bits_per_comp dsc_dc_color_depth_to_dsc_bits_per_comp(enum dc_color_depth dc_color_depth) +enum dsc_bits_per_comp dsc_dc_color_depth_to_dsc_bits_per_comp(enum dc_color_depth dc_color_depth) { enum dsc_bits_per_comp bpc = DSC_BPC_UNKNOWN; @@ -518,7 +509,7 @@ static enum dsc_bits_per_comp dsc_dc_color_depth_to_dsc_bits_per_comp(enum dc_co } -static void dsc_init_reg_values(struct dsc_reg_values *reg_vals) +void dsc_init_reg_values(struct dsc_reg_values *reg_vals) { int i; @@ -574,7 +565,7 @@ static void dsc_init_reg_values(struct dsc_reg_values *reg_vals) * This is required because dscc_compute_dsc_parameters returns a modified PPS, which in turn * affects non-PPS register values. */ -static void dsc_update_from_dsc_parameters(struct dsc_reg_values *reg_vals, const struct dsc_parameters *dsc_params) +void dsc_update_from_dsc_parameters(struct dsc_reg_values *reg_vals, const struct dsc_parameters *dsc_params) { int i; diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dsc.h b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dsc.h index 7ce64a3c1b02..ba869387c3c5 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dsc.h +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dsc.h @@ -549,6 +549,27 @@ struct dcn20_dsc { int max_image_width; }; +void dsc_config_log(struct display_stream_compressor *dsc, + const struct dsc_config *config); + +void dsc_log_pps(struct display_stream_compressor *dsc, + struct drm_dsc_config *pps); + +void dsc_override_rc_params(struct rc_params *rc, + const struct dc_dsc_rc_params_override *override); + +bool dsc_prepare_config(const struct dsc_config *dsc_cfg, + struct dsc_reg_values *dsc_reg_vals, + struct dsc_optc_config *dsc_optc_cfg); + +enum dsc_pixel_format dsc_dc_pixel_encoding_to_dsc_pixel_format(enum dc_pixel_encoding dc_pix_enc, + bool is_ycbcr422_simple); + +enum dsc_bits_per_comp dsc_dc_color_depth_to_dsc_bits_per_comp(enum dc_color_depth dc_color_depth); + +void dsc_init_reg_values(struct dsc_reg_values *reg_vals); + +void dsc_update_from_dsc_parameters(struct dsc_reg_values *reg_vals, const struct dsc_parameters *dsc_params); void dsc2_construct(struct dcn20_dsc *dsc, struct dc_context *ctx, @@ -557,5 +578,12 @@ void dsc2_construct(struct dcn20_dsc *dsc, const struct dcn20_dsc_shift *dsc_shift, const struct dcn20_dsc_mask *dsc_mask); +void dsc2_get_enc_caps(struct dsc_enc_caps *dsc_enc_caps, + int pixel_clock_100Hz); + +bool dsc2_get_packed_pps(struct display_stream_compressor *dsc, + const struct dsc_config *dsc_cfg, + uint8_t *dsc_packed_pps); + #endif diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c index 24bd93219936..6eebcb22e317 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c @@ -623,6 +623,17 @@ void hubbub2_read_state(struct hubbub *hubbub, struct dcn_hubbub_state *hubbub_s REG_GET(DCN_VM_FAULT_STATUS, DCN_VM_ERROR_VMID, &hubbub_state->vm_error_vmid); REG_GET(DCN_VM_FAULT_STATUS, DCN_VM_ERROR_PIPE, &hubbub_state->vm_error_pipe); } + + if (REG(DCHUBBUB_TEST_DEBUG_INDEX) && REG(DCHUBBUB_TEST_DEBUG_DATA)) { + REG_WRITE(DCHUBBUB_TEST_DEBUG_INDEX, 0x6); + hubbub_state->test_debug_data = REG_READ(DCHUBBUB_TEST_DEBUG_DATA); + } + + if (REG(DCHUBBUB_ARB_WATERMARK_CHANGE_CNTL)) + hubbub_state->watermark_change_cntl = REG_READ(DCHUBBUB_ARB_WATERMARK_CHANGE_CNTL); + + if (REG(DCHUBBUB_ARB_DRAM_STATE_CNTL)) + hubbub_state->dram_state_cntl = REG_READ(DCHUBBUB_ARB_DRAM_STATE_CNTL); } static const struct hubbub_funcs hubbub2_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c index 422fbf79da64..eaf9e9ccad2a 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c @@ -313,6 +313,10 @@ void dcn20_init_blank( } opp = dc->res_pool->opps[opp_id_src0]; + /* don't override the blank pattern if already enabled with the correct one. */ + if (opp->funcs->dpg_is_blanked && opp->funcs->dpg_is_blanked(opp)) + return; + if (num_opps == 2) { otg_active_width = otg_active_width / 2; @@ -1357,6 +1361,7 @@ static void dcn20_detect_pipe_changes(struct pipe_ctx *old_pipe, struct pipe_ctx new_pipe->update_flags.bits.dppclk = 1; new_pipe->update_flags.bits.hubp_interdependent = 1; new_pipe->update_flags.bits.hubp_rq_dlg_ttu = 1; + new_pipe->update_flags.bits.unbounded_req = 1; new_pipe->update_flags.bits.gamut_remap = 1; new_pipe->update_flags.bits.scaler = 1; new_pipe->update_flags.bits.viewport = 1; @@ -1500,6 +1505,9 @@ static void dcn20_detect_pipe_changes(struct pipe_ctx *old_pipe, struct pipe_ctx memcmp(&old_pipe->rq_regs, &new_pipe->rq_regs, sizeof(old_pipe->rq_regs))) new_pipe->update_flags.bits.hubp_rq_dlg_ttu = 1; } + + if (old_pipe->unbounded_req != new_pipe->unbounded_req) + new_pipe->update_flags.bits.unbounded_req = 1; } static void dcn20_update_dchubp_dpp( @@ -1533,10 +1541,11 @@ static void dcn20_update_dchubp_dpp( &pipe_ctx->ttu_regs, &pipe_ctx->rq_regs, &pipe_ctx->pipe_dlg_param); - - if (hubp->funcs->set_unbounded_requesting) - hubp->funcs->set_unbounded_requesting(hubp, pipe_ctx->unbounded_req); } + + if (pipe_ctx->update_flags.bits.unbounded_req && hubp->funcs->set_unbounded_requesting) + hubp->funcs->set_unbounded_requesting(hubp, pipe_ctx->unbounded_req); + if (pipe_ctx->update_flags.bits.hubp_interdependent) hubp->funcs->hubp_setup_interdependent( hubp, @@ -2113,20 +2122,11 @@ void dcn20_optimize_bandwidth( if (hubbub->funcs->program_compbuf_size) hubbub->funcs->program_compbuf_size(hubbub, context->bw_ctx.bw.dcn.compbuf_size_kb, true); - if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) { - dc_dmub_srv_p_state_delegate(dc, - true, context); - context->bw_ctx.bw.dcn.clk.p_state_change_support = true; - dc->clk_mgr->clks.fw_based_mclk_switching = true; - } else { - dc->clk_mgr->clks.fw_based_mclk_switching = false; - } - dc->clk_mgr->funcs->update_clocks( dc->clk_mgr, context, true); - if (dc_extended_blank_supported(dc) && context->bw_ctx.bw.dcn.clk.zstate_support == DCN_ZSTATE_SUPPORT_ALLOW) { + if (context->bw_ctx.bw.dcn.clk.zstate_support == DCN_ZSTATE_SUPPORT_ALLOW) { for (i = 0; i < dc->res_pool->pipe_count; ++i) { struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; @@ -2134,7 +2134,7 @@ void dcn20_optimize_bandwidth( && pipe_ctx->stream->adjust.v_total_min == pipe_ctx->stream->adjust.v_total_max && pipe_ctx->stream->adjust.v_total_max > pipe_ctx->stream->timing.v_total) pipe_ctx->plane_res.hubp->funcs->program_extended_blank(pipe_ctx->plane_res.hubp, - pipe_ctx->dlg_regs.optimized_min_dst_y_next_start); + pipe_ctx->dlg_regs.min_dst_y_next_start); } } } @@ -2471,36 +2471,31 @@ static void dcn20_reset_back_end_for_pipe( return; } - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - /* DPMS may already disable or */ - /* dpms_off status is incorrect due to fastboot - * feature. When system resume from S4 with second - * screen only, the dpms_off would be true but - * VBIOS lit up eDP, so check link status too. - */ - if (!pipe_ctx->stream->dpms_off || link->link_status.link_active) - dc->link_srv->set_dpms_off(pipe_ctx); - else if (pipe_ctx->stream_res.audio) - dc->hwss.disable_audio_stream(pipe_ctx); - - /* free acquired resources */ - if (pipe_ctx->stream_res.audio) { - /*disable az_endpoint*/ - pipe_ctx->stream_res.audio->funcs->az_disable(pipe_ctx->stream_res.audio); - - /*free audio*/ - if (dc->caps.dynamic_audio == true) { - /*we have to dynamic arbitrate the audio endpoints*/ - /*we free the resource, need reset is_audio_acquired*/ - update_audio_usage(&dc->current_state->res_ctx, dc->res_pool, - pipe_ctx->stream_res.audio, false); - pipe_ctx->stream_res.audio = NULL; - } + /* DPMS may already disable or */ + /* dpms_off status is incorrect due to fastboot + * feature. When system resume from S4 with second + * screen only, the dpms_off would be true but + * VBIOS lit up eDP, so check link status too. + */ + if (!pipe_ctx->stream->dpms_off || link->link_status.link_active) + dc->link_srv->set_dpms_off(pipe_ctx); + else if (pipe_ctx->stream_res.audio) + dc->hwss.disable_audio_stream(pipe_ctx); + + /* free acquired resources */ + if (pipe_ctx->stream_res.audio) { + /*disable az_endpoint*/ + pipe_ctx->stream_res.audio->funcs->az_disable(pipe_ctx->stream_res.audio); + + /*free audio*/ + if (dc->caps.dynamic_audio == true) { + /*we have to dynamic arbitrate the audio endpoints*/ + /*we free the resource, need reset is_audio_acquired*/ + update_audio_usage(&dc->current_state->res_ctx, dc->res_pool, + pipe_ctx->stream_res.audio, false); + pipe_ctx->stream_res.audio = NULL; } } - else if (pipe_ctx->stream_res.dsc) { - dc->link_srv->set_dsc_enable(pipe_ctx, false); - } /* by upper caller loop, parent pipe: pipe0, will be reset last. * back end share by all pipes and will be disable only when disable @@ -2576,28 +2571,6 @@ void dcn20_reset_hw_ctx_wrap( } } -void dcn20_update_visual_confirm_color(struct dc *dc, struct pipe_ctx *pipe_ctx, struct tg_color *color, int mpcc_id) -{ - struct mpc *mpc = dc->res_pool->mpc; - - // input to MPCC is always RGB, by default leave black_color at 0 - if (dc->debug.visual_confirm == VISUAL_CONFIRM_HDR) - get_hdr_visual_confirm_color(pipe_ctx, color); - else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SURFACE) - get_surface_visual_confirm_color(pipe_ctx, color); - else if (dc->debug.visual_confirm == VISUAL_CONFIRM_MPCTREE) - get_mpctree_visual_confirm_color(pipe_ctx, color); - else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SWIZZLE) - get_surface_tile_visual_confirm_color(pipe_ctx, color); - else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SUBVP) - get_subvp_visual_confirm_color(dc, pipe_ctx, color); - - if (mpc->funcs->set_bg_color) { - memcpy(&pipe_ctx->plane_state->visual_confirm_color, color, sizeof(struct tg_color)); - mpc->funcs->set_bg_color(mpc, color, mpcc_id); - } -} - void dcn20_update_mpcc(struct dc *dc, struct pipe_ctx *pipe_ctx) { struct hubp *hubp = pipe_ctx->plane_res.hubp; @@ -2653,7 +2626,7 @@ void dcn20_update_mpcc(struct dc *dc, struct pipe_ctx *pipe_ctx) if (!pipe_ctx->plane_state->update_flags.bits.full_update && !pipe_ctx->update_flags.bits.mpcc) { mpc->funcs->update_blending(mpc, &blnd_cfg, mpcc_id); - dc->hwss.update_visual_confirm_color(dc, pipe_ctx, &blnd_cfg.black_color, mpcc_id); + dc->hwss.update_visual_confirm_color(dc, pipe_ctx, mpcc_id); return; } @@ -2675,7 +2648,7 @@ void dcn20_update_mpcc(struct dc *dc, struct pipe_ctx *pipe_ctx) NULL, hubp->inst, mpcc_id); - dc->hwss.update_visual_confirm_color(dc, pipe_ctx, &blnd_cfg.black_color, mpcc_id); + dc->hwss.update_visual_confirm_color(dc, pipe_ctx, mpcc_id); ASSERT(new_mpcc != NULL); hubp->opp_id = pipe_ctx->stream_res.opp->inst; diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.h b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.h index 33a36c02b2f8..01901b08644c 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.h +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.h @@ -150,10 +150,5 @@ void dcn20_set_disp_pattern_generator(const struct dc *dc, const struct tg_color *solid_color, int width, int height, int offset); -void dcn20_update_visual_confirm_color(struct dc *dc, - struct pipe_ctx *pipe_ctx, - struct tg_color *color, - int mpcc_id); - #endif /* __DC_HWSS_DCN20_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c index 7c5817c426fa..e4b44e691ce6 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c @@ -102,7 +102,7 @@ static const struct hw_sequencer_funcs dcn20_funcs = { .disable_link_output = dce110_disable_link_output, .set_disp_pattern_generator = dcn20_set_disp_pattern_generator, .get_dcc_en_bits = dcn10_get_dcc_en_bits, - .update_visual_confirm_color = dcn20_update_visual_confirm_color + .update_visual_confirm_color = dcn10_update_visual_confirm_color, }; static const struct hwseq_private_funcs dcn20_private_funcs = { @@ -145,8 +145,4 @@ void dcn20_hw_sequencer_construct(struct dc *dc) dc->hwss = dcn20_funcs; dc->hwseq->funcs = dcn20_private_funcs; - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - dc->hwss.init_hw = dcn20_fpga_init_hw; - dc->hwseq->funcs.init_pipes = NULL; - } } diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_optc.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_optc.c index a08c335b7383..58bdbd859bf9 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_optc.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_optc.c @@ -38,8 +38,12 @@ optc1->tg_shift->field_name, optc1->tg_mask->field_name /** - * Enable CRTC - * Enable CRTC - call ASIC Control Object to enable Timing generator. + * optc2_enable_crtc() - Enable CRTC - call ASIC Control Object to enable Timing generator. + * + * @optc: timing_generator instance. + * + * Return: If CRTC is enabled, return true. + * */ bool optc2_enable_crtc(struct timing_generator *optc) { @@ -73,15 +77,18 @@ bool optc2_enable_crtc(struct timing_generator *optc) } /** - *For the below, I'm not sure how your GSL parameters are stored in your env, - * so I will assume a gsl_params struct for now + * optc2_set_gsl() - Assign OTG to GSL groups, + * set one of the OTGs to be master & rest are slaves + * + * @optc: timing_generator instance. + * @params: pointer to gsl_params */ void optc2_set_gsl(struct timing_generator *optc, const struct gsl_params *params) { struct optc *optc1 = DCN10TG_FROM_TG(optc); -/** +/* * There are (MAX_OPTC+1)/2 gsl groups available for use. * In each group (assign an OTG to a group by setting OTG_GSLX_EN = 1, * set one of the OTGs to be the master (OTG_GSL_MASTER_EN = 1) and the rest are slaves. @@ -391,10 +398,9 @@ void optc2_triplebuffer_lock(struct timing_generator *optc) REG_SET(OTG_MASTER_UPDATE_LOCK, 0, OTG_MASTER_UPDATE_LOCK, 1); - if (optc->ctx->dce_environment != DCE_ENV_FPGA_MAXIMUS) - REG_WAIT(OTG_MASTER_UPDATE_LOCK, - UPDATE_LOCK_STATUS, 1, - 1, 10); + REG_WAIT(OTG_MASTER_UPDATE_LOCK, + UPDATE_LOCK_STATUS, 1, + 1, 10); } void optc2_triplebuffer_unlock(struct timing_generator *optc) @@ -456,6 +462,16 @@ void optc2_setup_manual_trigger(struct timing_generator *optc) { struct optc *optc1 = DCN10TG_FROM_TG(optc); + /* Set the min/max selectors unconditionally so that + * DMCUB fw may change OTG timings when necessary + * TODO: Remove the w/a after fixing the issue in DMCUB firmware + */ + REG_UPDATE_4(OTG_V_TOTAL_CONTROL, + OTG_V_TOTAL_MIN_SEL, 1, + OTG_V_TOTAL_MAX_SEL, 1, + OTG_FORCE_LOCK_ON_EVENT, 0, + OTG_SET_V_TOTAL_MIN_MASK, (1 << 1)); /* TRIGA */ + REG_SET_8(OTG_TRIGA_CNTL, 0, OTG_TRIGA_SOURCE_SELECT, 21, OTG_TRIGA_SOURCE_PIPE_SELECT, optc->inst, diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c index 1d8c5805ef20..4cc8de2627ce 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c @@ -722,22 +722,7 @@ static const struct dc_debug_options debug_defaults_drv = { .scl_reset_length10 = true, .sanity_checks = false, .underflow_assert_delay_us = 0xFFFFFFFF, -}; - -static const struct dc_debug_options debug_defaults_diags = { - .disable_dmcu = false, - .force_abm_enable = false, - .timing_trace = true, - .clock_trace = true, - .disable_dpp_power_gate = true, - .disable_hubp_power_gate = true, - .disable_clock_gate = true, - .disable_pplib_clock_request = true, - .disable_pplib_wm_range = true, - .disable_stutter = true, - .scl_reset_length10 = true, - .underflow_assert_delay_us = 0xFFFFFFFF, - .enable_tri_buf = true, + .enable_legacy_fast_update = true, }; void dcn20_dpp_destroy(struct dpp **dpp) @@ -1066,13 +1051,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn20_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hwseq = dcn20_hwseq_create, -}; - static void dcn20_pp_smu_destroy(struct pp_smu_funcs **pp_smu); void dcn20_clock_source_destroy(struct clock_source **clk_src) @@ -2488,15 +2466,9 @@ static bool dcn20_resource_construct( dc->caps.dp_hdmi21_pcon_support = true; - if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) { + if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) dc->debug = debug_defaults_drv; - } else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) { - pool->base.pipe_count = 4; - pool->base.mpcc_count = pool->base.pipe_count; - dc->debug = debug_defaults_diags; - } else { - dc->debug = debug_defaults_diags; - } + //dcn2.0x dc->work_arounds.dedcn20_305_wa = true; @@ -2734,9 +2706,8 @@ static bool dcn20_resource_construct( } if (!resource_construct(num_virtual_links, dc, &pool->base, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) - goto create_fail; + &res_create_funcs)) + goto create_fail; dcn20_hw_sequencer_construct(dc); diff --git a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_hwseq.c index 1aeb04fbd89d..9e027db6d752 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_hwseq.c @@ -231,52 +231,39 @@ void dcn201_init_hw(struct dc *dc) if (dc->clk_mgr && dc->clk_mgr->funcs->init_clocks) dc->clk_mgr->funcs->init_clocks(dc->clk_mgr); - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - REG_WRITE(RBBMIF_TIMEOUT_DIS, 0xFFFFFFFF); - REG_WRITE(RBBMIF_TIMEOUT_DIS_2, 0xFFFFFFFF); - - hws->funcs.dccg_init(hws); - - REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_REFDIV, 2); - REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1); - REG_WRITE(REFCLK_CNTL, 0); - } else { - hws->funcs.bios_golden_init(dc); - - if (dc->ctx->dc_bios->fw_info_valid) { - res_pool->ref_clocks.xtalin_clock_inKhz = - dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency; - - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - if (res_pool->dccg && res_pool->hubbub) { - (res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg, - dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency, - &res_pool->ref_clocks.dccg_ref_clock_inKhz); - - (res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub, - res_pool->ref_clocks.dccg_ref_clock_inKhz, - &res_pool->ref_clocks.dchub_ref_clock_inKhz); - } else { - res_pool->ref_clocks.dccg_ref_clock_inKhz = - res_pool->ref_clocks.xtalin_clock_inKhz; - res_pool->ref_clocks.dchub_ref_clock_inKhz = - res_pool->ref_clocks.xtalin_clock_inKhz; - } - } - } else - ASSERT_CRITICAL(false); - for (i = 0; i < dc->link_count; i++) { - /* Power up AND update implementation according to the - * required signal (which may be different from the - * default signal on connector). - */ - struct dc_link *link = dc->links[i]; - - link->link_enc->funcs->hw_init(link->link_enc); + hws->funcs.bios_golden_init(dc); + + if (dc->ctx->dc_bios->fw_info_valid) { + res_pool->ref_clocks.xtalin_clock_inKhz = + dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency; + + if (res_pool->dccg && res_pool->hubbub) { + (res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg, + dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency, + &res_pool->ref_clocks.dccg_ref_clock_inKhz); + + (res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub, + res_pool->ref_clocks.dccg_ref_clock_inKhz, + &res_pool->ref_clocks.dchub_ref_clock_inKhz); + } else { + res_pool->ref_clocks.dccg_ref_clock_inKhz = + res_pool->ref_clocks.xtalin_clock_inKhz; + res_pool->ref_clocks.dchub_ref_clock_inKhz = + res_pool->ref_clocks.xtalin_clock_inKhz; } - if (hws->fb_offset.quad_part == 0) - read_mmhub_vm_setup(hws); + } else + ASSERT_CRITICAL(false); + for (i = 0; i < dc->link_count; i++) { + /* Power up AND update implementation according to the + * required signal (which may be different from the + * default signal on connector). + */ + struct dc_link *link = dc->links[i]; + + link->link_enc->funcs->hw_init(link->link_enc); } + if (hws->fb_offset.quad_part == 0) + read_mmhub_vm_setup(hws); /* Blank pixel data with OPP DPG */ for (i = 0; i < res_pool->timing_generator_count; i++) { @@ -362,10 +349,6 @@ void dcn201_init_hw(struct dc *dc) tg->funcs->tg_init(tg); } - /* end of FPGA. Below if real ASIC */ - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) - return; - for (i = 0; i < res_pool->audio_count; i++) { struct audio *audio = res_pool->audios[i]; @@ -496,7 +479,7 @@ void dcn201_update_mpcc(struct dc *dc, struct pipe_ctx *pipe_ctx) /* If there is no full update, don't need to touch MPC tree*/ if (!pipe_ctx->plane_state->update_flags.bits.full_update) { - dc->hwss.update_visual_confirm_color(dc, pipe_ctx, &blnd_cfg.black_color, mpcc_id); + dc->hwss.update_visual_confirm_color(dc, pipe_ctx, mpcc_id); mpc->funcs->update_blending(mpc, &blnd_cfg, mpcc_id); return; } @@ -521,7 +504,7 @@ void dcn201_update_mpcc(struct dc *dc, struct pipe_ctx *pipe_ctx) dc->res_pool->mpc, mpcc_id); /* Call MPC to insert new plane */ - dc->hwss.update_visual_confirm_color(dc, pipe_ctx, &blnd_cfg.black_color, mpcc_id); + dc->hwss.update_visual_confirm_color(dc, pipe_ctx, mpcc_id); new_mpcc = mpc->funcs->insert_plane(dc->res_pool->mpc, mpc_tree_params, &blnd_cfg, diff --git a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_init.c b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_init.c index 9c16633e473a..92dd4cddbab8 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_init.c +++ b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_init.c @@ -91,7 +91,7 @@ static const struct hw_sequencer_funcs dcn201_funcs = { .enable_dp_link_output = dce110_enable_dp_link_output, .disable_link_output = dce110_disable_link_output, .set_disp_pattern_generator = dcn20_set_disp_pattern_generator, - .update_visual_confirm_color = dcn20_update_visual_confirm_color, + .update_visual_confirm_color = dcn10_update_visual_confirm_color, }; static const struct hwseq_private_funcs dcn201_private_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_optc.c b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_optc.c index 730875dfd8b4..70fcbec03fb6 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_optc.c +++ b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_optc.c @@ -55,10 +55,9 @@ static void optc201_triplebuffer_lock(struct timing_generator *optc) REG_SET(OTG_MASTER_UPDATE_LOCK, 0, OTG_MASTER_UPDATE_LOCK, 1); - if (optc->ctx->dce_environment != DCE_ENV_FPGA_MAXIMUS) - REG_WAIT(OTG_MASTER_UPDATE_LOCK, - UPDATE_LOCK_STATUS, 1, - 1, 10); + REG_WAIT(OTG_MASTER_UPDATE_LOCK, + UPDATE_LOCK_STATUS, 1, + 1, 10); } static void optc201_triplebuffer_unlock(struct timing_generator *optc) diff --git a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_resource.c b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_resource.c index 6ea70da28aaa..fdba8a9f5c30 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_resource.c @@ -613,6 +613,7 @@ static const struct dc_debug_options debug_defaults_drv = { .sanity_checks = false, .underflow_assert_delay_us = 0xFFFFFFFF, .enable_tri_buf = false, + .enable_legacy_fast_update = true, }; static void dcn201_dpp_destroy(struct dpp **dpp) @@ -896,13 +897,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn201_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hwseq = dcn201_hwseq_create, -}; - static void dcn201_clock_source_destroy(struct clock_source **clk_src) { kfree(TO_DCE110_CLK_SRC(*clk_src)); @@ -1272,9 +1266,8 @@ static bool dcn201_resource_construct( } if (!resource_construct(num_virtual_links, dc, &pool->base, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) - goto create_fail; + &res_create_funcs)) + goto create_fail; dcn201_hw_sequencer_construct(dc); diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubp.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubp.c index 58e459c7e7d3..f976fac8dc3f 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubp.c +++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubp.c @@ -667,7 +667,6 @@ static void program_surface_flip_and_addr(struct hubp *hubp, struct surface_flip static void dmcub_PLAT_54186_wa(struct hubp *hubp, struct surface_flip_registers *flip_regs) { - struct dc_dmub_srv *dmcub = hubp->ctx->dmub_srv; struct dcn21_hubp *hubp21 = TO_DCN21_HUBP(hubp); union dmub_rb_cmd cmd; @@ -690,11 +689,7 @@ static void dmcub_PLAT_54186_wa(struct hubp *hubp, cmd.PLAT_54186_wa.flip.flip_params.vmid = flip_regs->vmid; PERF_TRACE(); // TODO: remove after performance is stable. - dc_dmub_srv_cmd_queue(dmcub, &cmd); - PERF_TRACE(); // TODO: remove after performance is stable. - dc_dmub_srv_cmd_execute(dmcub); - PERF_TRACE(); // TODO: remove after performance is stable. - dc_dmub_srv_wait_idle(dmcub); + dm_execute_dmub_cmd(hubp->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); PERF_TRACE(); // TODO: remove after performance is stable. } diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hwseq.c index 2a182c2f57d6..43463d08f21b 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hwseq.c @@ -152,13 +152,28 @@ static bool dmub_abm_set_pipe(struct abm *abm, uint32_t otg_inst, uint32_t optio cmd.abm_set_pipe.abm_set_pipe_data.ramping_boundary = ramping_boundary; cmd.abm_set_pipe.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_pipe_data); - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); return true; } +static void dmub_abm_set_backlight(struct dc_context *dc, uint32_t backlight_pwm_u16_16, + uint32_t frame_ramp, uint32_t panel_inst) +{ + union dmub_rb_cmd cmd; + + memset(&cmd, 0, sizeof(cmd)); + cmd.abm_set_backlight.header.type = DMUB_CMD__ABM; + cmd.abm_set_backlight.header.sub_type = DMUB_CMD__ABM_SET_BACKLIGHT; + cmd.abm_set_backlight.abm_set_backlight_data.frame_ramp = frame_ramp; + cmd.abm_set_backlight.abm_set_backlight_data.backlight_user_level = backlight_pwm_u16_16; + cmd.abm_set_backlight.abm_set_backlight_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1; + cmd.abm_set_backlight.abm_set_backlight_data.panel_mask = (0x01 << panel_inst); + cmd.abm_set_backlight.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_backlight_data); + + dm_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT); +} + void dcn21_set_abm_immediate_disable(struct pipe_ctx *pipe_ctx) { struct abm *abm = pipe_ctx->stream_res.abm; @@ -173,8 +188,12 @@ void dcn21_set_abm_immediate_disable(struct pipe_ctx *pipe_ctx) } if (abm && panel_cntl) { - dmub_abm_set_pipe(abm, otg_inst, SET_ABM_PIPE_IMMEDIATELY_DISABLE, - panel_cntl->inst); + if (abm->funcs && abm->funcs->set_pipe_ex) { + abm->funcs->set_pipe_ex(abm, otg_inst, SET_ABM_PIPE_IMMEDIATELY_DISABLE, + panel_cntl->inst); + } else { + dmub_abm_set_pipe(abm, otg_inst, SET_ABM_PIPE_IMMEDIATELY_DISABLE, panel_cntl->inst); + } panel_cntl->funcs->store_backlight_level(panel_cntl); } } @@ -191,18 +210,21 @@ void dcn21_set_pipe(struct pipe_ctx *pipe_ctx) return; } - if (abm && panel_cntl) - dmub_abm_set_pipe(abm, otg_inst, SET_ABM_PIPE_NORMAL, panel_cntl->inst); + if (abm && panel_cntl) { + if (abm->funcs && abm->funcs->set_pipe_ex) { + abm->funcs->set_pipe_ex(abm, otg_inst, SET_ABM_PIPE_NORMAL, panel_cntl->inst); + } else { + dmub_abm_set_pipe(abm, otg_inst, SET_ABM_PIPE_NORMAL, panel_cntl->inst); + } + } } bool dcn21_set_backlight_level(struct pipe_ctx *pipe_ctx, uint32_t backlight_pwm_u16_16, uint32_t frame_ramp) { - union dmub_rb_cmd cmd; struct dc_context *dc = pipe_ctx->stream->ctx; struct abm *abm = pipe_ctx->stream_res.abm; - uint32_t otg_inst = pipe_ctx->stream_res.tg->inst; struct panel_cntl *panel_cntl = pipe_ctx->stream->link->panel_cntl; if (dc->dc->res_pool->dmcu) { @@ -210,21 +232,23 @@ bool dcn21_set_backlight_level(struct pipe_ctx *pipe_ctx, return true; } - if (abm && panel_cntl) - dmub_abm_set_pipe(abm, otg_inst, SET_ABM_PIPE_NORMAL, panel_cntl->inst); + if (abm != NULL) { + uint32_t otg_inst = pipe_ctx->stream_res.tg->inst; - memset(&cmd, 0, sizeof(cmd)); - cmd.abm_set_backlight.header.type = DMUB_CMD__ABM; - cmd.abm_set_backlight.header.sub_type = DMUB_CMD__ABM_SET_BACKLIGHT; - cmd.abm_set_backlight.abm_set_backlight_data.frame_ramp = frame_ramp; - cmd.abm_set_backlight.abm_set_backlight_data.backlight_user_level = backlight_pwm_u16_16; - cmd.abm_set_backlight.abm_set_backlight_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1; - cmd.abm_set_backlight.abm_set_backlight_data.panel_mask = (0x01 << panel_cntl->inst); - cmd.abm_set_backlight.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_backlight_data); + if (abm && panel_cntl) { + if (abm->funcs && abm->funcs->set_pipe_ex) { + abm->funcs->set_pipe_ex(abm, otg_inst, SET_ABM_PIPE_NORMAL, panel_cntl->inst); + } else { + dmub_abm_set_pipe(abm, otg_inst, SET_ABM_PIPE_NORMAL, panel_cntl->inst); + } + } + } - dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->dmub_srv); - dc_dmub_srv_wait_idle(dc->dmub_srv); + if (abm && abm->funcs && abm->funcs->set_backlight_level_pwm) + abm->funcs->set_backlight_level_pwm(abm, backlight_pwm_u16_16, + frame_ramp, 0, panel_cntl->inst); + else + dmub_abm_set_backlight(dc, backlight_pwm_u16_16, frame_ramp, panel_cntl->inst); return true; } diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c index fe1a8e2e08ef..f024157bd6eb 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c +++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c @@ -106,7 +106,7 @@ static const struct hw_sequencer_funcs dcn21_funcs = { .is_abm_supported = dcn21_is_abm_supported, .set_disp_pattern_generator = dcn20_set_disp_pattern_generator, .get_dcc_en_bits = dcn10_get_dcc_en_bits, - .update_visual_confirm_color = dcn20_update_visual_confirm_color, + .update_visual_confirm_color = dcn10_update_visual_confirm_color, }; static const struct hwseq_private_funcs dcn21_private_funcs = { @@ -151,8 +151,4 @@ void dcn21_hw_sequencer_construct(struct dc *dc) dc->hwss = dcn21_funcs; dc->hwseq->funcs = dcn21_private_funcs; - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - dc->hwss.init_hw = dcn20_fpga_init_hw; - dc->hwseq->funcs.init_pipes = NULL; - } } diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c index 19aaa557b2db..d693ea42d033 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c @@ -653,22 +653,7 @@ static const struct dc_debug_options debug_defaults_drv = { .usbc_combo_phy_reset_wa = true, .dmub_command_table = true, .use_max_lb = true, -}; - -static const struct dc_debug_options debug_defaults_diags = { - .disable_dmcu = false, - .force_abm_enable = false, - .timing_trace = true, - .clock_trace = true, - .disable_dpp_power_gate = true, - .disable_hubp_power_gate = true, - .disable_clock_gate = true, - .disable_pplib_clock_request = true, - .disable_pplib_wm_range = true, - .disable_stutter = true, - .disable_48mhz_pwrdwn = true, - .enable_tri_buf = true, - .use_max_lb = true + .enable_legacy_fast_update = true, }; static const struct dc_panel_config panel_config_defaults = { @@ -1219,13 +1204,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn21_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hwseq = dcn21_hwseq_create, -}; - static const struct encoder_feature_support link_enc_feature = { .max_hdmi_deep_color = COLOR_DEPTH_121212, .max_hdmi_pixel_clock = 600000, @@ -1503,11 +1481,6 @@ static bool dcn21_resource_construct( if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) dc->debug = debug_defaults_drv; - else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) { - pool->base.pipe_count = 4; - dc->debug = debug_defaults_diags; - } else - dc->debug = debug_defaults_diags; // Init the vm_helper if (dc->vm_helper) @@ -1721,9 +1694,8 @@ static bool dcn21_resource_construct( } if (!resource_construct(num_virtual_links, dc, &pool->base, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) - goto create_fail; + &res_create_funcs)) + goto create_fail; dcn21_hw_sequencer_construct(dc); diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/Makefile b/drivers/gpu/drm/amd/display/dc/dcn30/Makefile index b7c2ae9ddfda..4a3e9e47b6b6 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn30/Makefile +++ b/drivers/gpu/drm/amd/display/dc/dcn30/Makefile @@ -1,16 +1,16 @@ -# +# # Copyright 2020 Advanced Micro Devices, Inc. -# +# # Permission is hereby granted, free of charge, to any person obtaining a # copy of this software and associated documentation files (the "Software"), # to deal in the Software without restriction, including without limitation # the rights to use, copy, modify, merge, publish, distribute, sublicense, # and/or sell copies of the Software, and to permit persons to whom the # Software is furnished to do so, subject to the following conditions: -# +# # The above copyright notice and this permission notice shall be included in # all copies or substantial portions of the Software. -# +# # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR # IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, # FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL @@ -18,17 +18,31 @@ # OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, # ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR # OTHER DEALINGS IN THE SOFTWARE. -# +# # Authors: AMD -# -# +# +# + +DCN30 := \ + dcn30_init.o \ + dcn30_hubbub.o \ + dcn30_hubp.o \ + dcn30_dpp.o \ + dcn30_optc.o \ + dcn30_dccg.o \ + dcn30_hwseq.o \ + dcn30_mpc.o dcn30_vpg.o \ + dcn30_afmt.o \ + dcn30_dio_stream_encoder.o \ + dcn30_dwb.o \ + dcn30_dpp_cm.o \ + dcn30_dwb_cm.o \ + dcn30_cm_common.o \ + dcn30_mmhubbub.o \ + dcn30_resource.o \ + dcn30_dio_link_encoder.o -DCN30 = dcn30_init.o dcn30_hubbub.o dcn30_hubp.o dcn30_dpp.o dcn30_optc.o \ - dcn30_dccg.o dcn30_hwseq.o dcn30_mpc.o dcn30_vpg.o \ - dcn30_afmt.o dcn30_dio_stream_encoder.o dcn30_dwb.o \ - dcn30_dpp_cm.o dcn30_dwb_cm.o dcn30_cm_common.o dcn30_mmhubbub.o \ - dcn30_dio_link_encoder.o dcn30_resource.o AMD_DAL_DCN30 = $(addprefix $(AMDDALPATH)/dc/dcn30/,$(DCN30)) diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_stream_encoder.c index 9d08127d209b..005dbe099a7a 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_stream_encoder.c +++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_stream_encoder.c @@ -436,6 +436,21 @@ void enc3_stream_encoder_update_dp_info_packets( &info_frame->vsc, true); } + /* TODO: VSC SDP at packetIndex 1 should be retricted only if PSR-SU on. + * There should have another Infopacket type (e.g. vsc_psrsu) for PSR_SU. + * In addition, currently the driver check the valid bit then update and + * send the corresponding Infopacket. For PSR-SU, the SDP only be sent + * while entering PSR-SU mode. So we need another parameter(e.g. send) + * in dc_info_packet to indicate which infopacket should be enabled by + * default here. + */ + if (info_frame->vsc.valid) { + enc->vpg->funcs->update_generic_info_packet( + enc->vpg, + 1, /* packetIndex */ + &info_frame->vsc, + true); + } /* TODO: VSC SDP at packetIndex 1 should be restricted only if PSR-SU on. * There should have another Infopacket type (e.g. vsc_psrsu) for PSR_SU. * In addition, currently the driver check the valid bit then update and diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.c index e46bbe7ddcc9..2861d974fcf6 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.c +++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.c @@ -449,6 +449,12 @@ void hubp3_read_state(struct hubp *hubp) SWATH_HEIGHT_C, &rq_regs->rq_regs_c.swath_height, PTE_ROW_HEIGHT_LINEAR_C, &rq_regs->rq_regs_c.pte_row_height_linear); + if (REG(UCLK_PSTATE_FORCE)) + s->uclk_pstate_force = REG_READ(UCLK_PSTATE_FORCE); + + if (REG(DCHUBP_CNTL)) + s->hubp_cntl = REG_READ(DCHUBP_CNTL); + } void hubp3_setup( diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.c index 8263a07f265f..b9753867d97b 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.c @@ -330,10 +330,6 @@ void dcn30_enable_writeback( DC_LOG_DWB("%s dwb_pipe_inst = %d, mpcc_inst = %d",\ __func__, wb_info->dwb_pipe_inst,\ wb_info->mpcc_inst); - if (IS_DIAG_DC(dc->ctx->dce_environment)) { - /*till diags switch to warmup interface*/ - dcn30_mmhubbub_warmup(dc, 1, wb_info); - } /* Update writeback pipe */ dcn30_set_writeback(dc, wb_info, context); @@ -447,28 +443,6 @@ void dcn30_init_hw(struct dc *dc) if (res_pool->dccg->funcs->dccg_init) res_pool->dccg->funcs->dccg_init(res_pool->dccg); - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - - REG_WRITE(REFCLK_CNTL, 0); - REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1); - REG_WRITE(DIO_MEM_PWR_CTRL, 0); - - if (!dc->debug.disable_clock_gate) { - /* enable all DCN clock gating */ - REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0); - - REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0); - - REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0); - } - - //Enable ability to power gate / don't force power on permanently - if (hws->funcs.enable_power_gating_plane) - hws->funcs.enable_power_gating_plane(hws, true); - - return; - } - if (!dcb->funcs->is_accelerated_mode(dcb)) { hws->funcs.bios_golden_init(dc); hws->funcs.disable_vga(dc->hwseq); @@ -491,23 +465,21 @@ void dcn30_init_hw(struct dc *dc) res_pool->ref_clocks.xtalin_clock_inKhz = dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency; - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - if (res_pool->dccg && res_pool->hubbub) { + if (res_pool->dccg && res_pool->hubbub) { - (res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg, - dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency, - &res_pool->ref_clocks.dccg_ref_clock_inKhz); + (res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg, + dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency, + &res_pool->ref_clocks.dccg_ref_clock_inKhz); - (res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub, - res_pool->ref_clocks.dccg_ref_clock_inKhz, - &res_pool->ref_clocks.dchub_ref_clock_inKhz); - } else { - // Not all ASICs have DCCG sw component - res_pool->ref_clocks.dccg_ref_clock_inKhz = - res_pool->ref_clocks.xtalin_clock_inKhz; - res_pool->ref_clocks.dchub_ref_clock_inKhz = - res_pool->ref_clocks.xtalin_clock_inKhz; - } + (res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub, + res_pool->ref_clocks.dccg_ref_clock_inKhz, + &res_pool->ref_clocks.dchub_ref_clock_inKhz); + } else { + // Not all ASICs have DCCG sw component + res_pool->ref_clocks.dccg_ref_clock_inKhz = + res_pool->ref_clocks.xtalin_clock_inKhz; + res_pool->ref_clocks.dchub_ref_clock_inKhz = + res_pool->ref_clocks.xtalin_clock_inKhz; } } else ASSERT_CRITICAL(false); @@ -632,7 +604,7 @@ void dcn30_init_hw(struct dc *dc) dc->res_pool->hubbub->funcs->init_crb(dc->res_pool->hubbub); // Get DMCUB capabilities - dc_dmub_srv_query_caps_cmd(dc->ctx->dmub_srv->dmub); + dc_dmub_srv_query_caps_cmd(dc->ctx->dmub_srv); dc->caps.dmub_caps.psr = dc->ctx->dmub_srv->dmub->feature_caps.psr; dc->caps.dmub_caps.mclk_sw = dc->ctx->dmub_srv->dmub->feature_caps.fw_assisted_mclk_switch; } @@ -736,8 +708,7 @@ bool dcn30_apply_idle_power_optimizations(struct dc *dc, bool enable) cmd.mall.header.sub_type = DMUB_CMD__MALL_ACTION_NO_DF_REQ; cmd.mall.header.payload_bytes = sizeof(cmd.mall) - sizeof(cmd.mall.header); - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_NO_WAIT); return true; } @@ -859,9 +830,7 @@ bool dcn30_apply_idle_power_optimizations(struct dc *dc, bool enable) cmd.mall.cursor_height = cursor_attr.height; cmd.mall.cursor_pitch = cursor_attr.pitch; - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); /* Use copied cursor, and it's okay to not switch back */ cursor_attr.address.quad_part = cmd.mall.cursor_copy_dst.quad_part; @@ -877,8 +846,7 @@ bool dcn30_apply_idle_power_optimizations(struct dc *dc, bool enable) cmd.mall.tmr_scale = tmr_scale; cmd.mall.debug_bits = dc->debug.mall_error_as_fatal; - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_NO_WAIT); return true; } @@ -895,9 +863,7 @@ bool dcn30_apply_idle_power_optimizations(struct dc *dc, bool enable) cmd.mall.header.payload_bytes = sizeof(cmd.mall) - sizeof(cmd.mall.header); - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); return true; } @@ -983,36 +949,13 @@ void dcn30_set_disp_pattern_generator(const struct dc *dc, } void dcn30_prepare_bandwidth(struct dc *dc, - struct dc_state *context) + struct dc_state *context) { - bool p_state_change_support = context->bw_ctx.bw.dcn.clk.p_state_change_support; - /* Any transition into an FPO config should disable MCLK switching first to avoid - * driver and FW P-State synchronization issues. - */ - if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching || dc->clk_mgr->clks.fw_based_mclk_switching) { - dc->optimized_required = true; - context->bw_ctx.bw.dcn.clk.p_state_change_support = false; - } - if (dc->clk_mgr->dc_mode_softmax_enabled) if (dc->clk_mgr->clks.dramclk_khz <= dc->clk_mgr->bw_params->dc_mode_softmax_memclk * 1000 && context->bw_ctx.bw.dcn.clk.dramclk_khz > dc->clk_mgr->bw_params->dc_mode_softmax_memclk * 1000) dc->clk_mgr->funcs->set_max_memclk(dc->clk_mgr, dc->clk_mgr->bw_params->clk_table.entries[dc->clk_mgr->bw_params->clk_table.num_entries - 1].memclk_mhz); dcn20_prepare_bandwidth(dc, context); - /* - * enabled -> enabled: do not disable - * enabled -> disabled: disable - * disabled -> enabled: don't care - * disabled -> disabled: don't care - */ - if (!context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) - dc_dmub_srv_p_state_delegate(dc, false, context); - - if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching || dc->clk_mgr->clks.fw_based_mclk_switching) { - /* After disabling P-State, restore the original value to ensure we get the correct P-State - * on the next optimize. */ - context->bw_ctx.bw.dcn.clk.p_state_change_support = p_state_change_support; - } } diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_init.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_init.c index 3216d10c58ba..3d19acaa12f3 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_init.c +++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_init.c @@ -106,7 +106,7 @@ static const struct hw_sequencer_funcs dcn30_funcs = { .disable_link_output = dce110_disable_link_output, .set_disp_pattern_generator = dcn30_set_disp_pattern_generator, .get_dcc_en_bits = dcn10_get_dcc_en_bits, - .update_visual_confirm_color = dcn20_update_visual_confirm_color, + .update_visual_confirm_color = dcn10_update_visual_confirm_color, .is_abm_supported = dcn21_is_abm_supported }; @@ -151,8 +151,4 @@ void dcn30_hw_sequencer_construct(struct dc *dc) dc->hwss = dcn30_funcs; dc->hwseq->funcs = dcn30_private_funcs; - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - dc->hwss.init_hw = dcn20_fpga_init_hw; - dc->hwseq->funcs.init_pipes = NULL; - } } diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_optc.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_optc.c index c95f000b63b2..dfb8f62765f2 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_optc.c +++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_optc.c @@ -55,10 +55,9 @@ void optc3_triplebuffer_lock(struct timing_generator *optc) REG_SET(OTG_MASTER_UPDATE_LOCK, 0, OTG_MASTER_UPDATE_LOCK, 1); - if (optc->ctx->dce_environment != DCE_ENV_FPGA_MAXIMUS) - REG_WAIT(OTG_MASTER_UPDATE_LOCK, - UPDATE_LOCK_STATUS, 1, - 1, 10); + REG_WAIT(OTG_MASTER_UPDATE_LOCK, + UPDATE_LOCK_STATUS, 1, + 1, 10); TRACE_OPTC_LOCK_UNLOCK_STATE(optc1, optc->inst, true); } @@ -280,6 +279,9 @@ static void optc3_set_odm_combine(struct timing_generator *optc, int *opp_id, in * Sets double buffer point for V_TOTAL, H_TOTAL, VTOTAL_MIN, * VTOTAL_MAX, VTOTAL_MIN_SEL and VTOTAL_MAX_SEL registers. * + * @optc: timing_generator instance. + * @enable: Enable DRR double buffering control if true, disable otherwise. + * * Options: any time, start of frame, dp start of frame (range timing) */ static void optc3_set_timing_double_buffer(struct timing_generator *optc, bool enable) @@ -301,7 +303,12 @@ static void optc3_wait_drr_doublebuffer_pending_clear(struct timing_generator *o void optc3_set_vtotal_min_max(struct timing_generator *optc, int vtotal_min, int vtotal_max) { - optc1_set_vtotal_min_max(optc, vtotal_min, vtotal_max); + struct dc *dc = optc->ctx->dc; + + if (dc->caps.dmub_caps.mclk_sw && !dc->debug.disable_fams) + dc_dmub_srv_drr_update_cmd(dc, optc->inst, vtotal_min, vtotal_max); + else + optc1_set_vtotal_min_max(optc, vtotal_min, vtotal_max); } void optc3_tg_init(struct timing_generator *optc) diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c index 67a34cda3774..1a0284a068b2 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c @@ -728,24 +728,6 @@ static const struct dc_debug_options debug_defaults_drv = { .exit_idle_opt_for_cursor_updates = true }; -static const struct dc_debug_options debug_defaults_diags = { - .disable_dmcu = true, //No dmcu on DCN30 - .force_abm_enable = false, - .timing_trace = true, - .clock_trace = true, - .disable_dpp_power_gate = true, - .disable_hubp_power_gate = true, - .disable_clock_gate = true, - .disable_pplib_clock_request = true, - .disable_pplib_wm_range = true, - .disable_stutter = false, - .scl_reset_length10 = true, - .dwb_fi_phase = -1, // -1 = disable - .dmub_command_table = true, - .enable_tri_buf = true, - .use_max_lb = true -}; - static const struct dc_panel_config panel_config_defaults = { .psr = { .disable_psr = false, @@ -1076,13 +1058,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn30_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hwseq = dcn30_hwseq_create, -}; - static void dcn30_resource_destruct(struct dcn30_resource_pool *pool) { unsigned int i; @@ -2353,6 +2328,7 @@ static bool dcn30_resource_construct( dc->caps.color.mpc.ocsc = 1; dc->caps.dp_hdmi21_pcon_support = true; + dc->caps.max_v_total = (1 << 15) - 1; /* read VBIOS LTTPR caps */ { @@ -2376,10 +2352,7 @@ static bool dcn30_resource_construct( if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) dc->debug = debug_defaults_drv; - else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) { - dc->debug = debug_defaults_diags; - } else - dc->debug = debug_defaults_diags; + // Init the vm_helper if (dc->vm_helper) vm_helper_init(dc->vm_helper, 16); @@ -2577,8 +2550,7 @@ static bool dcn30_resource_construct( /* Audio, Stream Encoders including DIG and virtual, MPC 3D LUTs */ if (!resource_construct(num_virtual_links, dc, &pool->base, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) + &res_create_funcs)) goto create_fail; /* HW Sequencer and Plane caps */ diff --git a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_init.c b/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_init.c index 6192851c59ed..257df8660b4c 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_init.c +++ b/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_init.c @@ -107,7 +107,7 @@ static const struct hw_sequencer_funcs dcn301_funcs = { .get_dcc_en_bits = dcn10_get_dcc_en_bits, .optimize_pwr_state = dcn21_optimize_pwr_state, .exit_optimized_pwr_state = dcn21_exit_optimized_pwr_state, - .update_visual_confirm_color = dcn20_update_visual_confirm_color, + .update_visual_confirm_color = dcn10_update_visual_confirm_color, }; static const struct hwseq_private_funcs dcn301_private_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_resource.c b/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_resource.c index 5ac2a272c380..3485fbb1093e 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_resource.c @@ -702,23 +702,6 @@ static const struct dc_debug_options debug_defaults_drv = { .exit_idle_opt_for_cursor_updates = true }; -static const struct dc_debug_options debug_defaults_diags = { - .disable_dmcu = true, - .force_abm_enable = false, - .timing_trace = true, - .clock_trace = true, - .disable_dpp_power_gate = false, - .disable_hubp_power_gate = false, - .disable_clock_gate = true, - .disable_pplib_clock_request = true, - .disable_pplib_wm_range = true, - .disable_stutter = true, - .scl_reset_length10 = true, - .dwb_fi_phase = -1, // -1 = disable - .dmub_command_table = true, - .use_max_lb = false, -}; - static void dcn301_dpp_destroy(struct dpp **dpp) { kfree(TO_DCN20_DPP(*dpp)); @@ -1047,13 +1030,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn301_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hwseq = dcn301_hwseq_create, -}; - static void dcn301_destruct(struct dcn301_resource_pool *pool) { unsigned int i; @@ -1513,10 +1489,7 @@ static bool dcn301_resource_construct( if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) dc->debug = debug_defaults_drv; - else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) { - dc->debug = debug_defaults_diags; - } else - dc->debug = debug_defaults_diags; + // Init the vm_helper if (dc->vm_helper) vm_helper_init(dc->vm_helper, 16); @@ -1710,9 +1683,8 @@ static bool dcn301_resource_construct( /* Audio, Stream Encoders including HPO and virtual, MPC 3D LUTs */ if (!resource_construct(num_virtual_links, dc, &pool->base, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) - goto create_fail; + &res_create_funcs)) + goto create_fail; /* HW Sequencer and Plane caps */ dcn301_hw_sequencer_construct(dc); diff --git a/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.c b/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.c index 9f93c43115ba..7dc065ea247a 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.c @@ -98,24 +98,6 @@ static const struct dc_debug_options debug_defaults_drv = { .exit_idle_opt_for_cursor_updates = true }; -static const struct dc_debug_options debug_defaults_diags = { - .disable_dmcu = true, - .force_abm_enable = false, - .timing_trace = true, - .clock_trace = true, - .disable_dpp_power_gate = true, - .disable_hubp_power_gate = true, - .disable_clock_gate = true, - .disable_pplib_clock_request = true, - .disable_pplib_wm_range = true, - .disable_stutter = false, - .scl_reset_length10 = true, - .dwb_fi_phase = -1, // -1 = disable - .dmub_command_table = true, - .enable_tri_buf = true, - .use_max_lb = true -}; - static const struct dc_panel_config panel_config_defaults = { .psr = { .disable_psr = false, @@ -954,13 +936,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn302_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hwseq = dcn302_hwseq_create, -}; - static bool is_soc_bounding_box_valid(struct dc *dc) { uint32_t hw_internal_rev = dc->ctx->asic_id.hw_internal_rev; @@ -1252,6 +1227,7 @@ static bool dcn302_resource_construct( dc->caps.force_dp_tps4_for_cp2520 = true; dc->caps.extended_aux_timeout_support = true; dc->caps.dmcub_support = true; + dc->caps.max_v_total = (1 << 15) - 1; /* Color pipeline capabilities */ dc->caps.color.dpp.dcn_arch = 1; @@ -1309,8 +1285,6 @@ static bool dcn302_resource_construct( if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) dc->debug = debug_defaults_drv; - else - dc->debug = debug_defaults_diags; // Init the vm_helper if (dc->vm_helper) @@ -1489,8 +1463,7 @@ static bool dcn302_resource_construct( /* Audio, Stream Encoders including HPO and virtual, MPC 3D LUTs */ if (!resource_construct(num_virtual_links, dc, pool, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) + &res_create_funcs)) goto create_fail; /* HW Sequencer and Plane caps */ diff --git a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_resource.c b/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_resource.c index 7f72ef882ca4..6d9761395288 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_resource.c @@ -81,23 +81,6 @@ static const struct dc_debug_options debug_defaults_drv = { .disable_idle_power_optimizations = false, }; -static const struct dc_debug_options debug_defaults_diags = { - .disable_dmcu = true, - .force_abm_enable = false, - .timing_trace = true, - .clock_trace = true, - .disable_dpp_power_gate = true, - .disable_hubp_power_gate = true, - .disable_clock_gate = true, - .disable_pplib_clock_request = true, - .disable_pplib_wm_range = true, - .disable_stutter = false, - .scl_reset_length10 = true, - .dwb_fi_phase = -1, // -1 = disable - .dmub_command_table = true, - .enable_tri_buf = true, -}; - static const struct dc_panel_config panel_config_defaults = { .psr = { .disable_psr = false, @@ -881,13 +864,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn303_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hwseq = dcn303_hwseq_create, -}; - static bool is_soc_bounding_box_valid(struct dc *dc) { uint32_t hw_internal_rev = dc->ctx->asic_id.hw_internal_rev; @@ -1176,6 +1152,7 @@ static bool dcn303_resource_construct( dc->caps.force_dp_tps4_for_cp2520 = true; dc->caps.extended_aux_timeout_support = true; dc->caps.dmcub_support = true; + dc->caps.max_v_total = (1 << 15) - 1; /* Color pipeline capabilities */ dc->caps.color.dpp.dcn_arch = 1; @@ -1232,8 +1209,6 @@ static bool dcn303_resource_construct( if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) dc->debug = debug_defaults_drv; - else - dc->debug = debug_defaults_diags; // Init the vm_helper if (dc->vm_helper) @@ -1400,8 +1375,7 @@ static bool dcn303_resource_construct( /* Audio, Stream Encoders including HPO and virtual, MPC 3D LUTs */ if (!resource_construct(num_virtual_links, dc, pool, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) + &res_create_funcs)) goto create_fail; /* HW Sequencer and Plane caps */ diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dccg.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dccg.c index 4c2fdfea162f..65c1d754e2d6 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dccg.c +++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dccg.c @@ -47,6 +47,14 @@ void dccg31_update_dpp_dto(struct dccg *dccg, int dpp_inst, int req_dppclk) { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); + if (dccg->dpp_clock_gated[dpp_inst]) { + /* + * Do not update the DPPCLK DTO if the clock is stopped. + * It is treated the same as if the pipe itself were in PG. + */ + return; + } + if (dccg->ref_dppclk && req_dppclk) { int ref_dppclk = dccg->ref_dppclk; int modulo, phase; diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dio_link_encoder.c index 745a5d187a98..bd62502380d8 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dio_link_encoder.c +++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dio_link_encoder.c @@ -117,7 +117,6 @@ static bool query_dp_alt_from_dmub(struct link_encoder *enc, union dmub_rb_cmd *cmd) { struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc); - struct dc_dmub_srv *dc_dmub_srv = enc->ctx->dmub_srv; memset(cmd, 0, sizeof(*cmd)); cmd->query_dp_alt.header.type = DMUB_CMD__VBIOS; @@ -126,7 +125,7 @@ static bool query_dp_alt_from_dmub(struct link_encoder *enc, cmd->query_dp_alt.header.payload_bytes = sizeof(cmd->query_dp_alt.data); cmd->query_dp_alt.data.phy_id = phy_id_from_transmitter(enc10->base.transmitter); - if (!dc_dmub_srv_cmd_with_reply_data(dc_dmub_srv, cmd)) + if (!dm_execute_dmub_cmd(enc->ctx, cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)) return false; return true; @@ -425,7 +424,6 @@ static bool link_dpia_control(struct dc_context *dc_ctx, struct dmub_cmd_dig_dpia_control_data *dpia_control) { union dmub_rb_cmd cmd; - struct dc_dmub_srv *dmub = dc_ctx->dmub_srv; memset(&cmd, 0, sizeof(cmd)); @@ -438,9 +436,7 @@ static bool link_dpia_control(struct dc_context *dc_ctx, cmd.dig1_dpia_control.dpia_control = *dpia_control; - dc_dmub_srv_cmd_queue(dmub, &cmd); - dc_dmub_srv_cmd_execute(dmub); - dc_dmub_srv_wait_idle(dmub); + dm_execute_dmub_cmd(dc_ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); return true; } diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hubbub.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hubbub.c index 7e7cd5b64e6a..7445ed27852a 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hubbub.c +++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hubbub.c @@ -103,6 +103,7 @@ static void dcn31_program_det_size(struct hubbub *hubbub, int hubp_inst, unsigne default: break; } + DC_LOG_DEBUG("Set DET%d to %d segments\n", hubp_inst, det_size_segments); /* Should never be hit, if it is we have an erroneous hw config*/ ASSERT(hubbub2->det0_size + hubbub2->det1_size + hubbub2->det2_size + hubbub2->det3_size + hubbub2->compbuf_size_segments <= hubbub2->crb_size_segs); diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.c index 62ce36c75c4d..2a7f47642a44 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.c @@ -117,28 +117,6 @@ void dcn31_init_hw(struct dc *dc) if (dc->clk_mgr && dc->clk_mgr->funcs->init_clocks) dc->clk_mgr->funcs->init_clocks(dc->clk_mgr); - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - - REG_WRITE(REFCLK_CNTL, 0); - REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1); - REG_WRITE(DIO_MEM_PWR_CTRL, 0); - - if (!dc->debug.disable_clock_gate) { - /* enable all DCN clock gating */ - REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0); - - REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0); - - REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0); - } - - //Enable ability to power gate / don't force power on permanently - if (hws->funcs.enable_power_gating_plane) - hws->funcs.enable_power_gating_plane(hws, true); - - return; - } - if (!dcb->funcs->is_accelerated_mode(dcb)) { hws->funcs.bios_golden_init(dc); if (hws->funcs.disable_vga) @@ -154,23 +132,21 @@ void dcn31_init_hw(struct dc *dc) res_pool->ref_clocks.xtalin_clock_inKhz = dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency; - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - if (res_pool->dccg && res_pool->hubbub) { - - (res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg, - dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency, - &res_pool->ref_clocks.dccg_ref_clock_inKhz); - - (res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub, - res_pool->ref_clocks.dccg_ref_clock_inKhz, - &res_pool->ref_clocks.dchub_ref_clock_inKhz); - } else { - // Not all ASICs have DCCG sw component - res_pool->ref_clocks.dccg_ref_clock_inKhz = - res_pool->ref_clocks.xtalin_clock_inKhz; - res_pool->ref_clocks.dchub_ref_clock_inKhz = - res_pool->ref_clocks.xtalin_clock_inKhz; - } + if (res_pool->dccg && res_pool->hubbub) { + + (res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg, + dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency, + &res_pool->ref_clocks.dccg_ref_clock_inKhz); + + (res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub, + res_pool->ref_clocks.dccg_ref_clock_inKhz, + &res_pool->ref_clocks.dchub_ref_clock_inKhz); + } else { + // Not all ASICs have DCCG sw component + res_pool->ref_clocks.dccg_ref_clock_inKhz = + res_pool->ref_clocks.xtalin_clock_inKhz; + res_pool->ref_clocks.dchub_ref_clock_inKhz = + res_pool->ref_clocks.xtalin_clock_inKhz; } } else ASSERT_CRITICAL(false); @@ -197,10 +173,6 @@ void dcn31_init_hw(struct dc *dc) } } - /* Enables outbox notifications for usb4 dpia */ - if (dc->res_pool->usb4_dpia_count) - dmub_enable_outbox_notification(dc->ctx->dmub_srv); - /* we want to turn off all dp displays before doing detection */ dc->link_srv->blank_all_dp_displays(dc); @@ -297,8 +269,9 @@ void dcn31_init_hw(struct dc *dc) #endif // Get DMCUB capabilities - dc_dmub_srv_query_caps_cmd(dc->ctx->dmub_srv->dmub); + dc_dmub_srv_query_caps_cmd(dc->ctx->dmub_srv); dc->caps.dmub_caps.psr = dc->ctx->dmub_srv->dmub->feature_caps.psr; + dc->caps.dmub_caps.mclk_sw = dc->ctx->dmub_srv->dmub->feature_caps.fw_assisted_mclk_switch; } void dcn31_dsc_pg_control( @@ -442,9 +415,7 @@ void dcn31_z10_save_init(struct dc *dc) cmd.dcn_restore.header.type = DMUB_CMD__IDLE_OPT; cmd.dcn_restore.header.sub_type = DMUB_CMD__IDLE_OPT_DCN_SAVE_INIT; - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } void dcn31_z10_restore(const struct dc *dc) @@ -462,9 +433,7 @@ void dcn31_z10_restore(const struct dc *dc) cmd.dcn_restore.header.type = DMUB_CMD__IDLE_OPT; cmd.dcn_restore.header.sub_type = DMUB_CMD__IDLE_OPT_DCN_RESTORE; - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); } void dcn31_hubp_pg_control(struct dce_hwseq *hws, unsigned int hubp_inst, bool power_on) @@ -560,35 +529,31 @@ static void dcn31_reset_back_end_for_pipe( pipe_ctx->stream_res.tg->funcs->set_drr( pipe_ctx->stream_res.tg, NULL); - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - link = pipe_ctx->stream->link; - /* DPMS may already disable or */ - /* dpms_off status is incorrect due to fastboot - * feature. When system resume from S4 with second - * screen only, the dpms_off would be true but - * VBIOS lit up eDP, so check link status too. - */ - if (!pipe_ctx->stream->dpms_off || link->link_status.link_active) - dc->link_srv->set_dpms_off(pipe_ctx); - else if (pipe_ctx->stream_res.audio) - dc->hwss.disable_audio_stream(pipe_ctx); - - /* free acquired resources */ - if (pipe_ctx->stream_res.audio) { - /*disable az_endpoint*/ - pipe_ctx->stream_res.audio->funcs->az_disable(pipe_ctx->stream_res.audio); - - /*free audio*/ - if (dc->caps.dynamic_audio == true) { - /*we have to dynamic arbitrate the audio endpoints*/ - /*we free the resource, need reset is_audio_acquired*/ - update_audio_usage(&dc->current_state->res_ctx, dc->res_pool, - pipe_ctx->stream_res.audio, false); - pipe_ctx->stream_res.audio = NULL; - } + link = pipe_ctx->stream->link; + /* DPMS may already disable or */ + /* dpms_off status is incorrect due to fastboot + * feature. When system resume from S4 with second + * screen only, the dpms_off would be true but + * VBIOS lit up eDP, so check link status too. + */ + if (!pipe_ctx->stream->dpms_off || link->link_status.link_active) + dc->link_srv->set_dpms_off(pipe_ctx); + else if (pipe_ctx->stream_res.audio) + dc->hwss.disable_audio_stream(pipe_ctx); + + /* free acquired resources */ + if (pipe_ctx->stream_res.audio) { + /*disable az_endpoint*/ + pipe_ctx->stream_res.audio->funcs->az_disable(pipe_ctx->stream_res.audio); + + /*free audio*/ + if (dc->caps.dynamic_audio == true) { + /*we have to dynamic arbitrate the audio endpoints*/ + /*we free the resource, need reset is_audio_acquired*/ + update_audio_usage(&dc->current_state->res_ctx, dc->res_pool, + pipe_ctx->stream_res.audio, false); + pipe_ctx->stream_res.audio = NULL; } - } else if (pipe_ctx->stream_res.dsc) { - dc->link_srv->set_dsc_enable(pipe_ctx, false); } pipe_ctx->stream = NULL; diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_init.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_init.c index 3a32810bbe38..fc25cc300a17 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_init.c +++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_init.c @@ -58,6 +58,7 @@ static const struct hw_sequencer_funcs dcn31_funcs = { .enable_audio_stream = dce110_enable_audio_stream, .disable_audio_stream = dce110_disable_audio_stream, .disable_plane = dcn20_disable_plane, + .disable_pixel_data = dcn20_disable_pixel_data, .pipe_control_lock = dcn20_pipe_control_lock, .interdependent_update_lock = dcn10_lock_all_pipes, .cursor_lock = dcn10_cursor_lock, @@ -109,7 +110,7 @@ static const struct hw_sequencer_funcs dcn31_funcs = { .set_disp_pattern_generator = dcn30_set_disp_pattern_generator, .optimize_pwr_state = dcn21_optimize_pwr_state, .exit_optimized_pwr_state = dcn21_exit_optimized_pwr_state, - .update_visual_confirm_color = dcn20_update_visual_confirm_color, + .update_visual_confirm_color = dcn10_update_visual_confirm_color, }; static const struct hwseq_private_funcs dcn31_private_funcs = { @@ -153,8 +154,4 @@ void dcn31_hw_sequencer_construct(struct dc *dc) dc->hwss = dcn31_funcs; dc->hwseq->funcs = dcn31_private_funcs; - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - dc->hwss.init_hw = dcn20_fpga_init_hw; - dc->hwseq->funcs.init_pipes = NULL; - } } diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c index 11ea9d13e312..217acd4e292a 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c +++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c @@ -52,7 +52,7 @@ static bool dcn31_query_backlight_info(struct panel_cntl *panel_cntl, union dmub cmd->panel_cntl.header.payload_bytes = sizeof(cmd->panel_cntl.data); cmd->panel_cntl.data.inst = dcn31_panel_cntl->base.inst; - return dc_dmub_srv_cmd_with_reply_data(dc_dmub_srv, cmd); + return dm_execute_dmub_cmd(dc_dmub_srv->ctx, cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY); } static uint32_t dcn31_get_16_bit_backlight_from_pwm(struct panel_cntl *panel_cntl) @@ -85,7 +85,7 @@ static uint32_t dcn31_panel_cntl_hw_init(struct panel_cntl *panel_cntl) panel_cntl->stored_backlight_registers.LVTMA_PWRSEQ_REF_DIV_BL_PWM_REF_DIV; cmd.panel_cntl.data.bl_pwm_ref_div2 = panel_cntl->stored_backlight_registers.PANEL_PWRSEQ_REF_DIV2; - if (!dc_dmub_srv_cmd_with_reply_data(dc_dmub_srv, &cmd)) + if (!dm_execute_dmub_cmd(dc_dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)) return 0; panel_cntl->stored_backlight_registers.BL_PWM_CNTL = cmd.panel_cntl.data.bl_pwm_cntl; diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.c index ff8cd5076434..fc33b5fcabe1 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.c @@ -887,28 +887,11 @@ static const struct dc_debug_options debug_defaults_drv = { } }, .disable_z10 = true, + .enable_legacy_fast_update = true, .enable_z9_disable_interface = true, /* Allow support for the PMFW interface for disable Z9*/ .dml_hostvm_override = DML_HOSTVM_OVERRIDE_FALSE, }; -static const struct dc_debug_options debug_defaults_diags = { - .disable_dmcu = true, - .force_abm_enable = false, - .timing_trace = true, - .clock_trace = true, - .disable_dpp_power_gate = true, - .disable_hubp_power_gate = true, - .disable_clock_gate = true, - .disable_pplib_clock_request = true, - .disable_pplib_wm_range = true, - .disable_stutter = false, - .scl_reset_length10 = true, - .dwb_fi_phase = -1, // -1 = disable - .dmub_command_table = true, - .enable_tri_buf = true, - .use_max_lb = true -}; - static const struct dc_panel_config panel_config_defaults = { .psr = { .disable_psr = false, @@ -1341,13 +1324,6 @@ static struct dce_hwseq *dcn31_hwseq_create( hws->regs = &hwseq_reg; hws->shifts = &hwseq_shift; hws->masks = &hwseq_mask; - /* DCN3.1 FPGA Workaround - * Need to enable HPO DP Stream Encoder before setting OTG master enable. - * To do so, move calling function enable_stream_timing to only be done AFTER calling - * function core_link_enable_stream - */ - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) - hws->wa.dp_hpo_and_otg_sequence = true; } return hws; } @@ -1360,15 +1336,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn31_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hpo_dp_stream_encoder = dcn31_hpo_dp_stream_encoder_create, - .create_hpo_dp_link_encoder = dcn31_hpo_dp_link_encoder_create, - .create_hwseq = dcn31_hwseq_create, -}; - static void dcn31_resource_destruct(struct dcn31_resource_pool *pool) { unsigned int i; @@ -1988,10 +1955,7 @@ static bool dcn31_resource_construct( if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) dc->debug = debug_defaults_drv; - else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) { - dc->debug = debug_defaults_diags; - } else - dc->debug = debug_defaults_diags; + // Init the vm_helper if (dc->vm_helper) vm_helper_init(dc->vm_helper, 16); @@ -2195,9 +2159,8 @@ static bool dcn31_resource_construct( /* Audio, Stream Encoders including HPO and virtual, MPC 3D LUTs */ if (!resource_construct(num_virtual_links, dc, &pool->base, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) - goto create_fail; + &res_create_funcs)) + goto create_fail; /* HW Sequencer and Plane caps */ dcn31_hw_sequencer_construct(dc); diff --git a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_dccg.c b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_dccg.c index de7bfba2c179..cf23d7bc560a 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_dccg.c +++ b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_dccg.c @@ -45,6 +45,16 @@ #define DC_LOGGER \ dccg->ctx->logger +static void dccg314_trigger_dio_fifo_resync( + struct dccg *dccg) +{ + struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); + uint32_t dispclk_rdivider_value = 0; + + REG_GET(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_RDIVIDER, &dispclk_rdivider_value); + REG_UPDATE(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, dispclk_rdivider_value); +} + static void dccg314_get_pixel_rate_div( struct dccg *dccg, uint32_t otg_inst, @@ -322,6 +332,9 @@ static void dccg314_dpp_root_clock_control( { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); + if (dccg->dpp_clock_gated[dpp_inst] == clock_on) + return; + if (clock_on) { /* turn off the DTO and leave phase/modulo at max */ REG_UPDATE(DPPCLK_DTO_CTRL, DPPCLK_DTO_ENABLE[dpp_inst], 0); @@ -335,6 +348,8 @@ static void dccg314_dpp_root_clock_control( DPPCLK0_DTO_PHASE, 0, DPPCLK0_DTO_MODULO, 1); } + + dccg->dpp_clock_gated[dpp_inst] = !clock_on; } static const struct dccg_funcs dccg314_funcs = { @@ -357,6 +372,7 @@ static const struct dccg_funcs dccg314_funcs = { .disable_dsc = dccg31_disable_dscclk, .enable_dsc = dccg31_enable_dscclk, .set_pixel_rate_div = dccg314_set_pixel_rate_div, + .trigger_dio_fifo_resync = dccg314_trigger_dio_fifo_resync, .set_valid_pixel_rate = dccg314_set_valid_pixel_rate, }; diff --git a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_dccg.h b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_dccg.h index 90687a9e8fdd..8e07d3151f91 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_dccg.h +++ b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_dccg.h @@ -192,7 +192,10 @@ DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYBSYMCLK_GATE_DISABLE, mask_sh),\ DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYCSYMCLK_GATE_DISABLE, mask_sh),\ DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYDSYMCLK_GATE_DISABLE, mask_sh),\ - DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYESYMCLK_GATE_DISABLE, mask_sh) + DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYESYMCLK_GATE_DISABLE, mask_sh),\ + DCCG_SF(HDMISTREAMCLK0_DTO_PARAM, HDMISTREAMCLK0_DTO_MODULO, mask_sh),\ + DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_RDIVIDER, mask_sh),\ + DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, mask_sh) struct dccg *dccg314_create( struct dc_context *ctx, diff --git a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_hwseq.c index cc3fe9cac5b5..7a43f8868500 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_hwseq.c @@ -337,14 +337,13 @@ void dcn314_enable_power_gating_plane(struct dce_hwseq *hws, bool enable) REG_SET(DC_IP_REQUEST_CNTL, 0, IP_REQUEST_EN, 0); } -unsigned int dcn314_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsigned int *k1_div, unsigned int *k2_div) +void dcn314_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsigned int *k1_div, unsigned int *k2_div) { struct dc_stream_state *stream = pipe_ctx->stream; - unsigned int odm_combine_factor = 0; bool two_pix_per_container = false; two_pix_per_container = optc2_is_two_pixels_per_containter(&stream->timing); - odm_combine_factor = get_odm_config(pipe_ctx, NULL); + get_odm_config(pipe_ctx, NULL); if (stream->ctx->dc->link_srv->dp_is_128b_132b_signal(pipe_ctx)) { *k1_div = PIXEL_RATE_DIV_BY_1; @@ -362,15 +361,11 @@ unsigned int dcn314_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsig } else { *k1_div = PIXEL_RATE_DIV_BY_1; *k2_div = PIXEL_RATE_DIV_BY_4; - if (odm_combine_factor == 2) - *k2_div = PIXEL_RATE_DIV_BY_2; } } if ((*k1_div == PIXEL_RATE_DIV_NA) && (*k2_div == PIXEL_RATE_DIV_NA)) ASSERT(false); - - return odm_combine_factor; } void dcn314_set_pixels_per_cycle(struct pipe_ctx *pipe_ctx) @@ -390,6 +385,35 @@ void dcn314_set_pixels_per_cycle(struct pipe_ctx *pipe_ctx) pix_per_cycle); } +void dcn314_resync_fifo_dccg_dio(struct dce_hwseq *hws, struct dc *dc, struct dc_state *context) +{ + unsigned int i; + struct pipe_ctx *pipe = NULL; + bool otg_disabled[MAX_PIPES] = {false}; + + for (i = 0; i < dc->res_pool->pipe_count; i++) { + pipe = &dc->current_state->res_ctx.pipe_ctx[i]; + + if (pipe->top_pipe || pipe->prev_odm_pipe) + continue; + + if (pipe->stream && (pipe->stream->dpms_off || dc_is_virtual_signal(pipe->stream->signal))) { + pipe->stream_res.tg->funcs->disable_crtc(pipe->stream_res.tg); + reset_sync_context_for_pipe(dc, context, i); + otg_disabled[i] = true; + } + } + + hws->ctx->dc->res_pool->dccg->funcs->trigger_dio_fifo_resync(hws->ctx->dc->res_pool->dccg); + + for (i = 0; i < dc->res_pool->pipe_count; i++) { + pipe = &dc->current_state->res_ctx.pipe_ctx[i]; + + if (otg_disabled[i]) + pipe->stream_res.tg->funcs->enable_crtc(pipe->stream_res.tg); + } +} + void dcn314_dpp_root_clock_control(struct dce_hwseq *hws, unsigned int dpp_inst, bool clock_on) { if (!hws->ctx->dc->debug.root_clock_optimization.bits.dpp) @@ -417,9 +441,7 @@ void dcn314_hubp_pg_control(struct dce_hwseq *hws, unsigned int hubp_inst, bool cmd.domain_control.data.inst = hubp_inst; cmd.domain_control.data.power_gate = !power_on; - dc_dmub_srv_cmd_queue(ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(ctx->dmub_srv); - dc_dmub_srv_wait_idle(ctx->dmub_srv); + dm_execute_dmub_cmd(ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); PERF_TRACE(); } diff --git a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_hwseq.h b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_hwseq.h index 6d0b62503caa..96035c75e0df 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_hwseq.h +++ b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_hwseq.h @@ -37,10 +37,12 @@ void dcn314_dsc_pg_control(struct dce_hwseq *hws, unsigned int dsc_inst, bool po void dcn314_enable_power_gating_plane(struct dce_hwseq *hws, bool enable); -unsigned int dcn314_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsigned int *k1_div, unsigned int *k2_div); +void dcn314_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsigned int *k1_div, unsigned int *k2_div); void dcn314_set_pixels_per_cycle(struct pipe_ctx *pipe_ctx); +void dcn314_resync_fifo_dccg_dio(struct dce_hwseq *hws, struct dc *dc, struct dc_state *context); + void dcn314_hubp_pg_control(struct dce_hwseq *hws, unsigned int hubp_inst, bool power_on); void dcn314_dpp_root_clock_control(struct dce_hwseq *hws, unsigned int dpp_inst, bool clock_on); diff --git a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_init.c b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_init.c index a588f46b166f..86d6a514dec0 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_init.c +++ b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_init.c @@ -60,6 +60,7 @@ static const struct hw_sequencer_funcs dcn314_funcs = { .enable_audio_stream = dce110_enable_audio_stream, .disable_audio_stream = dce110_disable_audio_stream, .disable_plane = dcn20_disable_plane, + .disable_pixel_data = dcn20_disable_pixel_data, .pipe_control_lock = dcn20_pipe_control_lock, .interdependent_update_lock = dcn10_lock_all_pipes, .cursor_lock = dcn10_cursor_lock, @@ -111,7 +112,7 @@ static const struct hw_sequencer_funcs dcn314_funcs = { .set_disp_pattern_generator = dcn30_set_disp_pattern_generator, .optimize_pwr_state = dcn21_optimize_pwr_state, .exit_optimized_pwr_state = dcn21_exit_optimized_pwr_state, - .update_visual_confirm_color = dcn20_update_visual_confirm_color, + .update_visual_confirm_color = dcn10_update_visual_confirm_color, }; static const struct hwseq_private_funcs dcn314_private_funcs = { @@ -151,6 +152,7 @@ static const struct hwseq_private_funcs dcn314_private_funcs = { .setup_hpo_hw_control = dcn31_setup_hpo_hw_control, .calculate_dccg_k1_k2_values = dcn314_calculate_dccg_k1_k2_values, .set_pixels_per_cycle = dcn314_set_pixels_per_cycle, + .resync_fifo_dccg_dio = dcn314_resync_fifo_dccg_dio, }; void dcn314_hw_sequencer_construct(struct dc *dc) @@ -158,8 +160,4 @@ void dcn314_hw_sequencer_construct(struct dc *dc) dc->hwss = dcn314_funcs; dc->hwseq->funcs = dcn314_private_funcs; - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - dc->hwss.init_hw = dcn20_fpga_init_hw; - dc->hwseq->funcs.init_pipes = NULL; - } } diff --git a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_resource.c b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_resource.c index abeeede38fb3..a840b008d660 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_resource.c @@ -117,23 +117,6 @@ #define regBIF_BX2_BIOS_SCRATCH_6 0x003e #define regBIF_BX2_BIOS_SCRATCH_6_BASE_IDX 1 -struct IP_BASE_INSTANCE { - unsigned int segment[MAX_SEGMENT]; -}; - -struct IP_BASE { - struct IP_BASE_INSTANCE instance[MAX_INSTANCE]; -}; - -static const struct IP_BASE DCN_BASE = { { { { 0x00000012, 0x000000C0, 0x000034C0, 0x00009000, 0x02403C00, 0, 0, 0 } }, - { { 0, 0, 0, 0, 0, 0, 0, 0 } }, - { { 0, 0, 0, 0, 0, 0, 0, 0 } }, - { { 0, 0, 0, 0, 0, 0, 0, 0 } }, - { { 0, 0, 0, 0, 0, 0, 0, 0 } }, - { { 0, 0, 0, 0, 0, 0, 0, 0 } }, - { { 0, 0, 0, 0, 0, 0, 0, 0 } } } }; - - #define DC_LOGGER_INIT(logger) enum dcn31_clk_src_array_id { @@ -891,8 +874,8 @@ static const struct dc_debug_options debug_defaults_drv = { .force_abm_enable = false, .timing_trace = false, .clock_trace = true, - .disable_dpp_power_gate = true, - .disable_hubp_power_gate = true, + .disable_dpp_power_gate = false, + .disable_hubp_power_gate = false, .disable_pplib_clock_request = false, .pipe_split_policy = MPC_SPLIT_DYNAMIC, .force_single_disp_pipe_split = false, @@ -921,6 +904,22 @@ static const struct dc_debug_options debug_defaults_drv = { .afmt = true, } }, + + .root_clock_optimization = { + .bits = { + .dpp = true, + .dsc = false, + .hdmistream = false, + .hdmichar = false, + .dpstream = false, + .symclk32_se = false, + .symclk32_le = false, + .symclk_fe = false, + .physymclk = false, + .dpiasymclk = false, + } + }, + .seamless_boot_odm_combine = true }; @@ -1375,13 +1374,6 @@ static struct dce_hwseq *dcn314_hwseq_create( hws->regs = &hwseq_reg; hws->shifts = &hwseq_shift; hws->masks = &hwseq_mask; - /* DCN3.1 FPGA Workaround - * Need to enable HPO DP Stream Encoder before setting OTG master enable. - * To do so, move calling function enable_stream_timing to only be done AFTER calling - * function core_link_enable_stream - */ - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) - hws->wa.dp_hpo_and_otg_sequence = true; } return hws; } @@ -1394,15 +1386,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn314_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hpo_dp_stream_encoder = dcn31_hpo_dp_stream_encoder_create, - .create_hpo_dp_link_encoder = dcn31_hpo_dp_link_encoder_create, - .create_hwseq = dcn314_hwseq_create, -}; - static void dcn314_resource_destruct(struct dcn314_resource_pool *pool) { unsigned int i; @@ -1900,6 +1883,13 @@ static bool dcn314_resource_construct( /* Use pipe context based otg sync logic */ dc->config.use_pipe_ctx_sync_logic = true; + /* Disable pipe power gating when unsupported */ + if (ctx->asic_id.hw_internal_rev == 0x01 || + ctx->asic_id.hw_internal_rev == 0x80) { + dc->debug.disable_dpp_power_gate = true; + dc->debug.disable_hubp_power_gate = true; + } + /* read VBIOS LTTPR caps */ { if (ctx->dc_bios->funcs->get_lttpr_caps) { @@ -2101,8 +2091,7 @@ static bool dcn314_resource_construct( /* Audio, Stream Encoders including HPO and virtual, MPC 3D LUTs */ if (!resource_construct(num_virtual_links, dc, &pool->base, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) + &res_create_funcs)) goto create_fail; /* HW Sequencer and Plane caps */ diff --git a/drivers/gpu/drm/amd/display/dc/dcn315/dcn315_resource.c b/drivers/gpu/drm/amd/display/dc/dcn315/dcn315_resource.c index 41c972c8eb19..f1153941907e 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn315/dcn315_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn315/dcn315_resource.c @@ -136,6 +136,9 @@ #define DCN3_15_MAX_DET_SIZE 384 #define DCN3_15_CRB_SEGMENT_SIZE_KB 64 +#define DCN3_15_MAX_DET_SEGS (DCN3_15_MAX_DET_SIZE / DCN3_15_CRB_SEGMENT_SIZE_KB) +/* Minimum 2 extra segments need to be in compbuf and claimable to guarantee seamless mpo transitions */ +#define MIN_RESERVED_DET_SEGS 2 enum dcn31_clk_src_array_id { DCN31_CLK_SRC_PLL0, @@ -884,27 +887,10 @@ static const struct dc_debug_options debug_defaults_drv = { .afmt = true, } }, + .enable_legacy_fast_update = true, .psr_power_use_phy_fsm = 0, }; -static const struct dc_debug_options debug_defaults_diags = { - .disable_dmcu = true, - .force_abm_enable = false, - .timing_trace = true, - .clock_trace = true, - .disable_dpp_power_gate = true, - .disable_hubp_power_gate = true, - .disable_clock_gate = true, - .disable_pplib_clock_request = true, - .disable_pplib_wm_range = true, - .disable_stutter = false, - .scl_reset_length10 = true, - .dwb_fi_phase = -1, // -1 = disable - .dmub_command_table = true, - .enable_tri_buf = true, - .use_max_lb = true -}; - static const struct dc_panel_config panel_config_defaults = { .psr = { .disable_psr = false, @@ -1339,13 +1325,6 @@ static struct dce_hwseq *dcn31_hwseq_create( hws->regs = &hwseq_reg; hws->shifts = &hwseq_shift; hws->masks = &hwseq_mask; - /* DCN3.1 FPGA Workaround - * Need to enable HPO DP Stream Encoder before setting OTG master enable. - * To do so, move calling function enable_stream_timing to only be done AFTER calling - * function core_link_enable_stream - */ - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) - hws->wa.dp_hpo_and_otg_sequence = true; } return hws; } @@ -1358,15 +1337,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn31_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hpo_dp_stream_encoder = dcn31_hpo_dp_stream_encoder_create, - .create_hpo_dp_link_encoder = dcn31_hpo_dp_link_encoder_create, - .create_hwseq = dcn31_hwseq_create, -}; - static void dcn315_resource_destruct(struct dcn315_resource_pool *pool) { unsigned int i; @@ -1636,21 +1606,69 @@ static bool is_dual_plane(enum surface_pixel_format format) return format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN || format == SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA; } +static int source_format_to_bpp (enum source_format_class SourcePixelFormat) +{ + if (SourcePixelFormat == dm_444_64) + return 8; + else if (SourcePixelFormat == dm_444_16 || SourcePixelFormat == dm_444_16) + return 2; + else if (SourcePixelFormat == dm_444_8) + return 1; + else if (SourcePixelFormat == dm_rgbe_alpha) + return 5; + else if (SourcePixelFormat == dm_420_8) + return 3; + else if (SourcePixelFormat == dm_420_12) + return 6; + else + return 4; +} + +static bool allow_pixel_rate_crb(struct dc *dc, struct dc_state *context) +{ + int i; + struct resource_context *res_ctx = &context->res_ctx; + + /*Don't apply for single stream*/ + if (context->stream_count < 2) + return false; + + for (i = 0; i < dc->res_pool->pipe_count; i++) { + if (!res_ctx->pipe_ctx[i].stream) + continue; + + /*Don't apply if scaling*/ + if (res_ctx->pipe_ctx[i].stream->src.width != res_ctx->pipe_ctx[i].stream->dst.width || + res_ctx->pipe_ctx[i].stream->src.height != res_ctx->pipe_ctx[i].stream->dst.height || + (res_ctx->pipe_ctx[i].plane_state && (res_ctx->pipe_ctx[i].plane_state->src_rect.width + != res_ctx->pipe_ctx[i].plane_state->dst_rect.width || + res_ctx->pipe_ctx[i].plane_state->src_rect.height + != res_ctx->pipe_ctx[i].plane_state->dst_rect.height))) + return false; + /*Don't apply if MPO to avoid transition issues*/ + if (res_ctx->pipe_ctx[i].top_pipe && res_ctx->pipe_ctx[i].top_pipe->plane_state != res_ctx->pipe_ctx[i].plane_state) + return false; + } + return true; +} + static int dcn315_populate_dml_pipes_from_context( struct dc *dc, struct dc_state *context, display_e2e_pipe_params_st *pipes, bool fast_validate) { - int i, pipe_cnt; + int i, pipe_cnt, crb_idx, crb_pipes; struct resource_context *res_ctx = &context->res_ctx; struct pipe_ctx *pipe; const int max_usable_det = context->bw_ctx.dml.ip.config_return_buffer_size_in_kbytes - DCN3_15_MIN_COMPBUF_SIZE_KB; + int remaining_det_segs = max_usable_det / DCN3_15_CRB_SEGMENT_SIZE_KB; + bool pixel_rate_crb = allow_pixel_rate_crb(dc, context); DC_FP_START(); dcn31x_populate_dml_pipes_from_context(dc, context, pipes, fast_validate); DC_FP_END(); - for (i = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) { + for (i = 0, pipe_cnt = 0, crb_pipes = 0; i < dc->res_pool->pipe_count; i++) { struct dc_crtc_timing *timing; if (!res_ctx->pipe_ctx[i].stream) @@ -1671,6 +1689,28 @@ static int dcn315_populate_dml_pipes_from_context( pipes[pipe_cnt].dout.dsc_input_bpc = 0; DC_FP_START(); dcn31_zero_pipe_dcc_fraction(pipes, pipe_cnt); + if (pixel_rate_crb && !pipe->top_pipe && !pipe->prev_odm_pipe) { + int bpp = source_format_to_bpp(pipes[pipe_cnt].pipe.src.source_format); + /* Ceil to crb segment size */ + int approx_det_segs_required_for_pstate = dcn_get_approx_det_segs_required_for_pstate( + &context->bw_ctx.dml.soc, timing->pix_clk_100hz, bpp, DCN3_15_CRB_SEGMENT_SIZE_KB); + + if (approx_det_segs_required_for_pstate <= 2 * DCN3_15_MAX_DET_SEGS) { + bool split_required = approx_det_segs_required_for_pstate > DCN3_15_MAX_DET_SEGS; + split_required = split_required || timing->pix_clk_100hz >= dcn_get_max_non_odm_pix_rate_100hz(&dc->dml.soc); + split_required = split_required || (pipe->plane_state && pipe->plane_state->src_rect.width > 5120); + + /* Minimum 2 segments to allow mpc/odm combine if its used later */ + if (approx_det_segs_required_for_pstate < 2) + approx_det_segs_required_for_pstate = 2; + if (split_required) + approx_det_segs_required_for_pstate += approx_det_segs_required_for_pstate % 2; + pipes[pipe_cnt].pipe.src.det_size_override = approx_det_segs_required_for_pstate; + remaining_det_segs -= approx_det_segs_required_for_pstate; + } else + remaining_det_segs = -1; + crb_pipes++; + } DC_FP_END(); if (pipes[pipe_cnt].dout.dsc_enable) { @@ -1689,16 +1729,55 @@ static int dcn315_populate_dml_pipes_from_context( break; } } - pipe_cnt++; } + /* Spread remaining unreserved crb evenly among all pipes*/ + if (pixel_rate_crb) { + for (i = 0, pipe_cnt = 0, crb_idx = 0; i < dc->res_pool->pipe_count; i++) { + pipe = &res_ctx->pipe_ctx[i]; + if (!pipe->stream) + continue; + + /* Do not use asymetric crb if not enough for pstate support */ + if (remaining_det_segs < 0) { + pipes[pipe_cnt].pipe.src.det_size_override = 0; + pipe_cnt++; + continue; + } + + if (!pipe->top_pipe && !pipe->prev_odm_pipe) { + bool split_required = pipe->stream->timing.pix_clk_100hz >= dcn_get_max_non_odm_pix_rate_100hz(&dc->dml.soc) + || (pipe->plane_state && pipe->plane_state->src_rect.width > 5120); + + if (remaining_det_segs > MIN_RESERVED_DET_SEGS) + pipes[pipe_cnt].pipe.src.det_size_override += (remaining_det_segs - MIN_RESERVED_DET_SEGS) / crb_pipes + + (crb_idx < (remaining_det_segs - MIN_RESERVED_DET_SEGS) % crb_pipes ? 1 : 0); + if (pipes[pipe_cnt].pipe.src.det_size_override > 2 * DCN3_15_MAX_DET_SEGS) { + /* Clamp to 2 pipe split max det segments */ + remaining_det_segs += pipes[pipe_cnt].pipe.src.det_size_override - 2 * (DCN3_15_MAX_DET_SEGS); + pipes[pipe_cnt].pipe.src.det_size_override = 2 * DCN3_15_MAX_DET_SEGS; + } + if (pipes[pipe_cnt].pipe.src.det_size_override > DCN3_15_MAX_DET_SEGS || split_required) { + /* If we are splitting we must have an even number of segments */ + remaining_det_segs += pipes[pipe_cnt].pipe.src.det_size_override % 2; + pipes[pipe_cnt].pipe.src.det_size_override -= pipes[pipe_cnt].pipe.src.det_size_override % 2; + } + /* Convert segments into size for DML use */ + pipes[pipe_cnt].pipe.src.det_size_override *= DCN3_15_CRB_SEGMENT_SIZE_KB; + + crb_idx++; + } + pipe_cnt++; + } + } + if (pipe_cnt) context->bw_ctx.dml.ip.det_buffer_size_kbytes = (max_usable_det / DCN3_15_CRB_SEGMENT_SIZE_KB / pipe_cnt) * DCN3_15_CRB_SEGMENT_SIZE_KB; if (context->bw_ctx.dml.ip.det_buffer_size_kbytes > DCN3_15_MAX_DET_SIZE) context->bw_ctx.dml.ip.det_buffer_size_kbytes = DCN3_15_MAX_DET_SIZE; - ASSERT(context->bw_ctx.dml.ip.det_buffer_size_kbytes >= DCN3_15_DEFAULT_DET_SIZE); + dc->config.enable_4to1MPC = false; if (pipe_cnt == 1 && pipe->plane_state && !dc->debug.disable_z9_mpc) { if (is_dual_plane(pipe->plane_state->format) @@ -1845,10 +1924,7 @@ static bool dcn315_resource_construct( if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) dc->debug = debug_defaults_drv; - else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) { - dc->debug = debug_defaults_diags; - } else - dc->debug = debug_defaults_diags; + // Init the vm_helper if (dc->vm_helper) vm_helper_init(dc->vm_helper, 16); @@ -2029,9 +2105,8 @@ static bool dcn315_resource_construct( /* Audio, Stream Encoders including HPO and virtual, MPC 3D LUTs */ if (!resource_construct(num_virtual_links, dc, &pool->base, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) - goto create_fail; + &res_create_funcs)) + goto create_fail; /* HW Sequencer and Plane caps */ dcn31_hw_sequencer_construct(dc); diff --git a/drivers/gpu/drm/amd/display/dc/dcn316/dcn316_resource.c b/drivers/gpu/drm/amd/display/dc/dcn316/dcn316_resource.c index 9ead347a33e9..707cf28bbceb 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn316/dcn316_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn316/dcn316_resource.c @@ -884,24 +884,7 @@ static const struct dc_debug_options debug_defaults_drv = { .afmt = true, } }, -}; - -static const struct dc_debug_options debug_defaults_diags = { - .disable_dmcu = true, - .force_abm_enable = false, - .timing_trace = true, - .clock_trace = true, - .disable_dpp_power_gate = true, - .disable_hubp_power_gate = true, - .disable_clock_gate = true, - .disable_pplib_clock_request = true, - .disable_pplib_wm_range = true, - .disable_stutter = false, - .scl_reset_length10 = true, - .dwb_fi_phase = -1, // -1 = disable - .dmub_command_table = true, - .enable_tri_buf = true, - .use_max_lb = true + .enable_legacy_fast_update = true, }; static const struct dc_panel_config panel_config_defaults = { @@ -1340,13 +1323,6 @@ static struct dce_hwseq *dcn31_hwseq_create( hws->regs = &hwseq_reg; hws->shifts = &hwseq_shift; hws->masks = &hwseq_mask; - /* DCN3.1 FPGA Workaround - * Need to enable HPO DP Stream Encoder before setting OTG master enable. - * To do so, move calling function enable_stream_timing to only be done AFTER calling - * function core_link_enable_stream - */ - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) - hws->wa.dp_hpo_and_otg_sequence = true; } return hws; } @@ -1359,15 +1335,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn31_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hpo_dp_stream_encoder = dcn31_hpo_dp_stream_encoder_create, - .create_hpo_dp_link_encoder = dcn31_hpo_dp_link_encoder_create, - .create_hwseq = dcn31_hwseq_create, -}; - static void dcn316_resource_destruct(struct dcn316_resource_pool *pool) { unsigned int i; @@ -1844,10 +1811,7 @@ static bool dcn316_resource_construct( if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) dc->debug = debug_defaults_drv; - else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) { - dc->debug = debug_defaults_diags; - } else - dc->debug = debug_defaults_diags; + // Init the vm_helper if (dc->vm_helper) vm_helper_init(dc->vm_helper, 16); @@ -2028,9 +1992,8 @@ static bool dcn316_resource_construct( /* Audio, Stream Encoders including HPO and virtual, MPC 3D LUTs */ if (!resource_construct(num_virtual_links, dc, &pool->base, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) - goto create_fail; + &res_create_funcs)) + goto create_fail; /* HW Sequencer and Plane caps */ dcn31_hw_sequencer_construct(dc); diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_dccg.c b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_dccg.c index ffbb739d85b6..11e28e056cf7 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_dccg.c +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_dccg.c @@ -42,18 +42,14 @@ #define DC_LOGGER \ dccg->ctx->logger -/* This function is a workaround for writing to OTG_PIXEL_RATE_DIV - * without the probability of causing a DIG FIFO error. - */ -static void dccg32_wait_for_dentist_change_done( +static void dccg32_trigger_dio_fifo_resync( struct dccg *dccg) { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); + uint32_t dispclk_rdivider_value = 0; - uint32_t dentist_dispclk_value = REG_READ(DENTIST_DISPCLK_CNTL); - - REG_WRITE(DENTIST_DISPCLK_CNTL, dentist_dispclk_value); - REG_WAIT(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, 1, 50, 2000); + REG_GET(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_RDIVIDER, &dispclk_rdivider_value); + REG_UPDATE(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, dispclk_rdivider_value); } static void dccg32_get_pixel_rate_div( @@ -124,29 +120,21 @@ static void dccg32_set_pixel_rate_div( REG_UPDATE_2(OTG_PIXEL_RATE_DIV, OTG0_PIXEL_RATE_DIVK1, k1, OTG0_PIXEL_RATE_DIVK2, k2); - - dccg32_wait_for_dentist_change_done(dccg); break; case 1: REG_UPDATE_2(OTG_PIXEL_RATE_DIV, OTG1_PIXEL_RATE_DIVK1, k1, OTG1_PIXEL_RATE_DIVK2, k2); - - dccg32_wait_for_dentist_change_done(dccg); break; case 2: REG_UPDATE_2(OTG_PIXEL_RATE_DIV, OTG2_PIXEL_RATE_DIVK1, k1, OTG2_PIXEL_RATE_DIVK2, k2); - - dccg32_wait_for_dentist_change_done(dccg); break; case 3: REG_UPDATE_2(OTG_PIXEL_RATE_DIV, OTG3_PIXEL_RATE_DIVK1, k1, OTG3_PIXEL_RATE_DIVK2, k2); - - dccg32_wait_for_dentist_change_done(dccg); break; default: BREAK_TO_DEBUGGER(); @@ -352,6 +340,7 @@ static const struct dccg_funcs dccg32_funcs = { .otg_add_pixel = dccg32_otg_add_pixel, .otg_drop_pixel = dccg32_otg_drop_pixel, .set_pixel_rate_div = dccg32_set_pixel_rate_div, + .trigger_dio_fifo_resync = dccg32_trigger_dio_fifo_resync, }; struct dccg *dccg32_create( diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_dccg.h b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_dccg.h index 8071ab98d708..cf5508718122 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_dccg.h +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_dccg.h @@ -112,8 +112,9 @@ DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P3_EN, mask_sh),\ DCCG_SF(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO_SEL, mask_sh),\ DCCG_SF(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO0_SOURCE_SEL, mask_sh),\ - DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, mask_sh) - + DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, mask_sh),\ + DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_RDIVIDER, mask_sh),\ + DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, mask_sh) struct dccg *dccg32_create( struct dc_context *ctx, diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hubbub.c b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hubbub.c index eb08ccc38e79..a18b9c0c5709 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hubbub.c +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hubbub.c @@ -42,8 +42,8 @@ hubbub2->shifts->field_name, hubbub2->masks->field_name /** - * @DCN32_CRB_SEGMENT_SIZE_KB: Maximum Configurable Return Buffer size for - * DCN32 + * DCN32_CRB_SEGMENT_SIZE_KB: Maximum Configurable Return Buffer size for + * DCN32 */ #define DCN32_CRB_SEGMENT_SIZE_KB 64 diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hwseq.c index 1f5ee5cde6e1..c586468872e2 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hwseq.c @@ -274,8 +274,7 @@ bool dcn32_apply_idle_power_optimizations(struct dc *dc, bool enable) cmd.cab.header.sub_type = DMUB_CMD__CAB_NO_DCN_REQ; cmd.cab.header.payload_bytes = sizeof(cmd.cab) - sizeof(cmd.cab.header); - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_NO_WAIT); return true; } @@ -309,8 +308,7 @@ bool dcn32_apply_idle_power_optimizations(struct dc *dc, bool enable) cmd.cab.header.payload_bytes = sizeof(cmd.cab) - sizeof(cmd.cab.header); cmd.cab.cab_alloc_ways = ways; - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_NO_WAIT); return true; } @@ -326,9 +324,7 @@ bool dcn32_apply_idle_power_optimizations(struct dc *dc, bool enable) cmd.cab.header.payload_bytes = sizeof(cmd.cab) - sizeof(cmd.cab.header); - dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd); - dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv); - dc_dmub_srv_wait_idle(dc->ctx->dmub_srv); + dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); return true; } @@ -413,6 +409,30 @@ void dcn32_subvp_pipe_control_lock(struct dc *dc, } } +void dcn32_subvp_pipe_control_lock_fast(union block_sequence_params *params) +{ + struct dc *dc = params->subvp_pipe_control_lock_fast_params.dc; + bool lock = params->subvp_pipe_control_lock_fast_params.lock; + struct pipe_ctx *pipe_ctx = params->subvp_pipe_control_lock_fast_params.pipe_ctx; + bool subvp_immediate_flip = false; + + if (pipe_ctx && pipe_ctx->stream && pipe_ctx->plane_state) { + if (pipe_ctx->stream->mall_stream_config.type == SUBVP_MAIN && + pipe_ctx->plane_state->flip_immediate) + subvp_immediate_flip = true; + } + + // Don't need to lock for DRR VSYNC flips -- FW will wait for DRR pending update cleared. + if (subvp_immediate_flip) { + union dmub_inbox0_cmd_lock_hw hw_lock_cmd = { 0 }; + + hw_lock_cmd.bits.command_code = DMUB_INBOX0_CMD__HW_LOCK; + hw_lock_cmd.bits.hw_lock_client = HW_LOCK_CLIENT_DRIVER; + hw_lock_cmd.bits.lock = lock; + hw_lock_cmd.bits.should_release = !lock; + dmub_hw_lock_mgr_inbox0_cmd(dc->ctx->dmub_srv, hw_lock_cmd); + } +} bool dcn32_set_mpc_shaper_3dlut( struct pipe_ctx *pipe_ctx, const struct dc_stream_state *stream) @@ -587,8 +607,8 @@ void dcn32_update_force_pstate(struct dc *dc, struct dc_state *context) struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; struct hubp *hubp = pipe->plane_res.hubp; - if (!pipe->stream || (pipe->stream && !(pipe->stream->mall_stream_config.type == SUBVP_MAIN || - pipe->stream->fpo_in_use))) { + if (!pipe->stream || !(pipe->stream->mall_stream_config.type == SUBVP_MAIN || + pipe->stream->fpo_in_use)) { if (hubp && hubp->funcs->hubp_update_force_pstate_disallow) hubp->funcs->hubp_update_force_pstate_disallow(hubp, false); } @@ -596,7 +616,7 @@ void dcn32_update_force_pstate(struct dc *dc, struct dc_state *context) /* Today only FPO uses cursor P-State force. Only clear cursor P-State force * if it's not FPO. */ - if (!pipe->stream || (pipe->stream && !pipe->stream->fpo_in_use)) { + if (!pipe->stream || !pipe->stream->fpo_in_use) { if (hubp && hubp->funcs->hubp_update_force_cursor_pstate_disallow) hubp->funcs->hubp_update_force_cursor_pstate_disallow(hubp, false); } @@ -721,6 +741,9 @@ static void dcn32_initialize_min_clocks(struct dc *dc) clocks->socclk_khz = dc->clk_mgr->bw_params->clk_table.entries[0].socclk_mhz * 1000; clocks->dramclk_khz = dc->clk_mgr->bw_params->clk_table.entries[0].memclk_mhz * 1000; clocks->dppclk_khz = dc->clk_mgr->bw_params->clk_table.entries[0].dppclk_mhz * 1000; + clocks->ref_dtbclk_khz = dc->clk_mgr->bw_params->clk_table.entries[0].dtbclk_mhz * 1000; + clocks->fclk_p_state_change_support = true; + clocks->p_state_change_support = true; if (dc->debug.disable_boot_optimizations) { clocks->dispclk_khz = dc->clk_mgr->bw_params->clk_table.entries[0].dispclk_mhz * 1000; } else { @@ -730,9 +753,6 @@ static void dcn32_initialize_min_clocks(struct dc *dc) * freq to ensure that the timing is valid and unchanged. */ clocks->dispclk_khz = dc->clk_mgr->funcs->get_dispclk_from_dentist(dc->clk_mgr); - clocks->ref_dtbclk_khz = dc->clk_mgr->bw_params->clk_table.entries[0].dtbclk_mhz * 1000; - clocks->fclk_p_state_change_support = true; - clocks->p_state_change_support = true; } dc->clk_mgr->funcs->update_clocks( @@ -946,8 +966,10 @@ void dcn32_init_hw(struct dc *dc) // Get DMCUB capabilities if (dc->ctx->dmub_srv) { - dc_dmub_srv_query_caps_cmd(dc->ctx->dmub_srv->dmub); + dc_dmub_srv_query_caps_cmd(dc->ctx->dmub_srv); dc->caps.dmub_caps.psr = dc->ctx->dmub_srv->dmub->feature_caps.psr; + dc->caps.dmub_caps.subvp_psr = dc->ctx->dmub_srv->dmub->feature_caps.subvp_psr_support; + dc->caps.dmub_caps.gecc_enable = dc->ctx->dmub_srv->dmub->feature_caps.gecc_enable; dc->caps.dmub_caps.mclk_sw = dc->ctx->dmub_srv->dmub->feature_caps.fw_assisted_mclk_switch; } } @@ -1119,10 +1141,9 @@ void dcn32_update_odm(struct dc *dc, struct dc_state *context, struct pipe_ctx * } } -unsigned int dcn32_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsigned int *k1_div, unsigned int *k2_div) +void dcn32_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsigned int *k1_div, unsigned int *k2_div) { struct dc_stream_state *stream = pipe_ctx->stream; - unsigned int odm_combine_factor = 0; bool two_pix_per_container = false; // For phantom pipes, use the same programming as the main pipes @@ -1130,7 +1151,6 @@ unsigned int dcn32_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsign stream = pipe_ctx->stream->mall_stream_config.paired_stream; } two_pix_per_container = optc2_is_two_pixels_per_containter(&stream->timing); - odm_combine_factor = get_odm_config(pipe_ctx, NULL); if (stream->ctx->dc->link_srv->dp_is_128b_132b_signal(pipe_ctx)) { *k1_div = PIXEL_RATE_DIV_BY_1; @@ -1148,15 +1168,13 @@ unsigned int dcn32_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsign } else { *k1_div = PIXEL_RATE_DIV_BY_1; *k2_div = PIXEL_RATE_DIV_BY_4; - if ((odm_combine_factor == 2) || dcn32_is_dp_dig_pixel_rate_div_policy(pipe_ctx)) + if (dcn32_is_dp_dig_pixel_rate_div_policy(pipe_ctx)) *k2_div = PIXEL_RATE_DIV_BY_2; } } if ((*k1_div == PIXEL_RATE_DIV_NA) && (*k2_div == PIXEL_RATE_DIV_NA)) ASSERT(false); - - return odm_combine_factor; } void dcn32_set_pixels_per_cycle(struct pipe_ctx *pipe_ctx) @@ -1177,6 +1195,36 @@ void dcn32_set_pixels_per_cycle(struct pipe_ctx *pipe_ctx) pix_per_cycle); } +void dcn32_resync_fifo_dccg_dio(struct dce_hwseq *hws, struct dc *dc, struct dc_state *context) +{ + unsigned int i; + struct pipe_ctx *pipe = NULL; + bool otg_disabled[MAX_PIPES] = {false}; + + for (i = 0; i < dc->res_pool->pipe_count; i++) { + pipe = &dc->current_state->res_ctx.pipe_ctx[i]; + + if (pipe->top_pipe || pipe->prev_odm_pipe) + continue; + + if (pipe->stream && (pipe->stream->dpms_off || dc_is_virtual_signal(pipe->stream->signal)) + && pipe->stream->mall_stream_config.type != SUBVP_PHANTOM) { + pipe->stream_res.tg->funcs->disable_crtc(pipe->stream_res.tg); + reset_sync_context_for_pipe(dc, context, i); + otg_disabled[i] = true; + } + } + + hws->ctx->dc->res_pool->dccg->funcs->trigger_dio_fifo_resync(hws->ctx->dc->res_pool->dccg); + + for (i = 0; i < dc->res_pool->pipe_count; i++) { + pipe = &dc->current_state->res_ctx.pipe_ctx[i]; + + if (otg_disabled[i]) + pipe->stream_res.tg->funcs->enable_crtc(pipe->stream_res.tg); + } +} + void dcn32_unblank_stream(struct pipe_ctx *pipe_ctx, struct dc_link_settings *link_settings) { diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hwseq.h b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hwseq.h index 6694c1d14aa3..bf9bffabe0c0 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hwseq.h +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hwseq.h @@ -71,10 +71,12 @@ void dcn32_update_force_pstate(struct dc *dc, struct dc_state *context); void dcn32_update_odm(struct dc *dc, struct dc_state *context, struct pipe_ctx *pipe_ctx); -unsigned int dcn32_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsigned int *k1_div, unsigned int *k2_div); +void dcn32_calculate_dccg_k1_k2_values(struct pipe_ctx *pipe_ctx, unsigned int *k1_div, unsigned int *k2_div); void dcn32_set_pixels_per_cycle(struct pipe_ctx *pipe_ctx); +void dcn32_resync_fifo_dccg_dio(struct dce_hwseq *hws, struct dc *dc, struct dc_state *context); + void dcn32_subvp_pipe_control_lock(struct dc *dc, struct dc_state *context, bool lock, @@ -82,6 +84,8 @@ void dcn32_subvp_pipe_control_lock(struct dc *dc, struct pipe_ctx *top_pipe_to_program, bool subvp_prev_use); +void dcn32_subvp_pipe_control_lock_fast(union block_sequence_params *params); + void dcn32_unblank_stream(struct pipe_ctx *pipe_ctx, struct dc_link_settings *link_settings); diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_init.c b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_init.c index 8085f2acb1a9..c2490e16a66a 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_init.c +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_init.c @@ -109,7 +109,8 @@ static const struct hw_sequencer_funcs dcn32_funcs = { .commit_subvp_config = dcn32_commit_subvp_config, .enable_phantom_streams = dcn32_enable_phantom_streams, .subvp_pipe_control_lock = dcn32_subvp_pipe_control_lock, - .update_visual_confirm_color = dcn20_update_visual_confirm_color, + .update_visual_confirm_color = dcn10_update_visual_confirm_color, + .subvp_pipe_control_lock_fast = dcn32_subvp_pipe_control_lock_fast, .update_phantom_vp_position = dcn32_update_phantom_vp_position, .update_dsc_pg = dcn32_update_dsc_pg, .apply_update_flags_for_phantom = dcn32_apply_update_flags_for_phantom, @@ -153,6 +154,7 @@ static const struct hwseq_private_funcs dcn32_private_funcs = { .update_mall_sel = dcn32_update_mall_sel, .calculate_dccg_k1_k2_values = dcn32_calculate_dccg_k1_k2_values, .set_pixels_per_cycle = dcn32_set_pixels_per_cycle, + .resync_fifo_dccg_dio = dcn32_resync_fifo_dccg_dio, .is_dp_dig_pixel_rate_div_policy = dcn32_is_dp_dig_pixel_rate_div_policy, }; @@ -161,8 +163,4 @@ void dcn32_hw_sequencer_init_functions(struct dc *dc) dc->hwss = dcn32_funcs; dc->hwseq->funcs = dcn32_private_funcs; - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - dc->hwss.init_hw = dcn20_fpga_init_hw; - dc->hwseq->funcs.init_pipes = NULL; - } } diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_optc.c b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_optc.c index 2ee798965bc2..8abb94f60078 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_optc.c +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_optc.c @@ -98,7 +98,7 @@ static void optc32_set_odm_combine(struct timing_generator *optc, int *opp_id, i optc1->opp_count = opp_cnt; } -static void optc32_set_h_timing_div_manual_mode(struct timing_generator *optc, bool manual_mode) +void optc32_set_h_timing_div_manual_mode(struct timing_generator *optc, bool manual_mode) { struct optc *optc1 = DCN10TG_FROM_TG(optc); @@ -106,8 +106,11 @@ static void optc32_set_h_timing_div_manual_mode(struct timing_generator *optc, b OTG_H_TIMING_DIV_MODE_MANUAL, manual_mode ? 1 : 0); } /** - * Enable CRTC - * Enable CRTC - call ASIC Control Object to enable Timing generator. + * optc32_enable_crtc() - Enable CRTC - call ASIC Control Object to enable Timing generator. + * + * @optc: timing_generator instance. + * + * Return: If CRTC is enabled, return true. */ static bool optc32_enable_crtc(struct timing_generator *optc) { @@ -245,16 +248,9 @@ static void optc32_set_drr( } optc->funcs->set_vtotal_min_max(optc, params->vertical_total_min - 1, params->vertical_total_max - 1); - optc32_setup_manual_trigger(optc); - } else { - REG_UPDATE_4(OTG_V_TOTAL_CONTROL, - OTG_SET_V_TOTAL_MIN_MASK, 0, - OTG_V_TOTAL_MIN_SEL, 0, - OTG_V_TOTAL_MAX_SEL, 0, - OTG_FORCE_LOCK_ON_EVENT, 0); - - optc->funcs->set_vtotal_min_max(optc, 0, 0); } + + optc32_setup_manual_trigger(optc); } static struct timing_generator_funcs dcn32_tg_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_optc.h b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_optc.h index b92ba8c75694..abf0121a1006 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_optc.h +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_optc.h @@ -179,5 +179,6 @@ SF(OTG0_OTG_DRR_CONTROL, OTG_V_TOTAL_LAST_USED_BY_DRR, mask_sh) void dcn32_timing_generator_init(struct optc *optc1); +void optc32_set_h_timing_div_manual_mode(struct timing_generator *optc, bool manual_mode); #endif /* __DC_OPTC_DCN32_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource.c b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource.c index 22dd1ebea618..19f134caa8ad 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource.c @@ -726,28 +726,12 @@ static const struct dc_debug_options debug_defaults_drv = { .override_dispclk_programming = true, .disable_fpo_optimizations = false, .fpo_vactive_margin_us = 2000, // 2000us - .disable_fpo_vactive = true, + .disable_fpo_vactive = false, .disable_boot_optimizations = false, -}; - -static const struct dc_debug_options debug_defaults_diags = { - .disable_dmcu = true, - .force_abm_enable = false, - .timing_trace = true, - .clock_trace = true, - .disable_dpp_power_gate = true, - .disable_hubp_power_gate = true, - .disable_dsc_power_gate = true, - .disable_clock_gate = true, - .disable_pplib_clock_request = true, - .disable_pplib_wm_range = true, - .disable_stutter = false, - .scl_reset_length10 = true, - .dwb_fi_phase = -1, // -1 = disable - .dmub_command_table = true, - .enable_tri_buf = true, - .use_max_lb = true, - .force_disable_subvp = true + .disable_subvp_high_refresh = false, + .disable_dp_plus_plus_wa = true, + .fpo_vactive_min_active_margin_us = 200, + .fpo_vactive_max_blank_us = 1000, }; static struct dce_aux *dcn32_aux_engine_create( @@ -1353,15 +1337,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn32_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hpo_dp_stream_encoder = dcn32_hpo_dp_stream_encoder_create, - .create_hpo_dp_link_encoder = dcn32_hpo_dp_link_encoder_create, - .create_hwseq = dcn32_hwseq_create, -}; - static void dcn32_resource_destruct(struct dcn32_resource_pool *pool) { unsigned int i; @@ -1888,6 +1863,8 @@ bool dcn32_validate_bandwidth(struct dc *dc, dc->res_pool->funcs->calculate_wm_and_dlg(dc, context, pipes, pipe_cnt, vlevel); + dcn32_override_min_req_memclk(dc, context); + BW_VAL_TRACE_END_WATERMARKS(); goto validate_out; @@ -2198,6 +2175,7 @@ static bool dcn32_resource_construct( dc->caps.extended_aux_timeout_support = true; dc->caps.dmcub_support = true; dc->caps.seamless_odm = true; + dc->caps.max_v_total = (1 << 15) - 1; /* Color pipeline capabilities */ dc->caps.color.dpp.dcn_arch = 1; @@ -2254,10 +2232,7 @@ static bool dcn32_resource_construct( if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) dc->debug = debug_defaults_drv; - else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) { - dc->debug = debug_defaults_diags; - } else - dc->debug = debug_defaults_diags; + // Init the vm_helper if (dc->vm_helper) vm_helper_init(dc->vm_helper, 16); @@ -2313,8 +2288,7 @@ static bool dcn32_resource_construct( } /* DML */ - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) - dml_init_instance(&dc->dml, &dcn3_2_soc, &dcn3_2_ip, DML_PROJECT_DCN32); + dml_init_instance(&dc->dml, &dcn3_2_soc, &dcn3_2_ip, DML_PROJECT_DCN32); /* IRQ Service */ init_data.ctx = dc->ctx; @@ -2451,9 +2425,8 @@ static bool dcn32_resource_construct( /* Audio, HWSeq, Stream Encoders including HPO and virtual, MPC 3D LUTs */ if (!resource_construct(num_virtual_links, dc, &pool->base, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) - goto create_fail; + &res_create_funcs)) + goto create_fail; /* HW Sequencer init functions and Plane caps */ dcn32_hw_sequencer_init_functions(dc); diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource.h b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource.h index 3937dbc1e552..81e443170829 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource.h +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource.h @@ -37,9 +37,10 @@ #define DCN3_2_MBLK_WIDTH 128 #define DCN3_2_MBLK_HEIGHT_4BPE 128 #define DCN3_2_MBLK_HEIGHT_8BPE 64 -#define DCN3_2_VMIN_DISPCLK_HZ 717000000 #define DCN3_2_DCFCLK_DS_INIT_KHZ 10000 // Choose 10Mhz for init DCFCLK DS freq -#define DCN3_2_MIN_ACTIVE_SWITCH_MARGIN_FPO_US 100 // Only allow FPO + Vactive if active margin >= 100 +#define SUBVP_HIGH_REFRESH_LIST_LEN 3 +#define DCN3_2_MAX_SUBVP_PIXEL_RATE_MHZ 1800 +#define DCN3_2_VMIN_DISPCLK_HZ 717000000 #define TO_DCN32_RES_POOL(pool)\ container_of(pool, struct dcn32_resource_pool, base) @@ -47,6 +48,15 @@ extern struct _vcs_dpi_ip_params_st dcn3_2_ip; extern struct _vcs_dpi_soc_bounding_box_st dcn3_2_soc; +struct subvp_high_refresh_list { + int min_refresh; + int max_refresh; + struct resolution { + int width; + int height; + } res[SUBVP_HIGH_REFRESH_LIST_LEN]; +}; + struct dcn32_resource_pool { struct resource_pool base; }; @@ -151,10 +161,18 @@ struct dc_stream_state *dcn32_can_support_mclk_switch_using_fw_based_vblank_stre bool dcn32_allow_subvp_with_active_margin(struct pipe_ctx *pipe); +bool dcn32_allow_subvp_high_refresh_rate(struct dc *dc, struct dc_state *context, struct pipe_ctx *pipe); + unsigned int dcn32_calc_num_avail_chans_for_mall(struct dc *dc, int num_chans); double dcn32_determine_max_vratio_prefetch(struct dc *dc, struct dc_state *context); +bool dcn32_check_native_scaling_for_res(struct pipe_ctx *pipe, unsigned int width, unsigned int height); + +bool dcn32_subvp_drr_admissable(struct dc *dc, struct dc_state *context); + +bool dcn32_subvp_vblank_admissable(struct dc *dc, struct dc_state *context, int vlevel); + /* definitions for run time init of reg offsets */ /* CLK SRC */ diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource_helpers.c b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource_helpers.c index eeca16faf31a..a9c41ef0751f 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource_helpers.c +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource_helpers.c @@ -94,18 +94,15 @@ uint32_t dcn32_helper_calculate_mall_bytes_for_cursor( } /** - * ******************************************************************************************** - * dcn32_helper_calculate_num_ways_for_subvp: Calculate number of ways needed for SubVP + * dcn32_helper_calculate_num_ways_for_subvp(): Calculate number of ways needed for SubVP * * Gets total allocation required for the phantom viewport calculated by DML in bytes and * converts to number of cache ways. * - * @param [in] dc: current dc state - * @param [in] context: new dc state + * @dc: current dc state + * @context: new dc state * - * @return: number of ways required for SubVP - * - * ******************************************************************************************** + * Return: number of ways required for SubVP */ uint32_t dcn32_helper_calculate_num_ways_for_subvp( struct dc *dc, @@ -258,11 +255,8 @@ bool dcn32_is_psr_capable(struct pipe_ctx *pipe) return psr_capable; } -#define DCN3_2_NEW_DET_OVERRIDE_MIN_MULTIPLIER 7 - /** - * ******************************************************************************************* - * dcn32_determine_det_override: Determine DET allocation for each pipe + * dcn32_determine_det_override(): Determine DET allocation for each pipe * * This function determines how much DET to allocate for each pipe. The total number of * DET segments will be split equally among each of the streams, and after that the DET @@ -271,6 +265,7 @@ bool dcn32_is_psr_capable(struct pipe_ctx *pipe) * If there is a plane that's driven by more than 1 pipe (i.e. pipe split), then the * number of DET for that given plane will be split among the pipes driving that plane. * + * * High level algorithm: * 1. Split total DET among number of streams * 2. For each stream, split DET among the planes @@ -278,25 +273,11 @@ bool dcn32_is_psr_capable(struct pipe_ctx *pipe) * among those pipes. * 4. Assign the DET override to the DML pipes. * - * Special cases: - * - * For two displays that have a large difference in pixel rate, we may experience - * underflow on the larger display when we divide the DET equally. For this, we - * will implement a modified algorithm to assign more DET to larger display. - * - * 1. Calculate difference in pixel rates ( multiplier ) between two displays - * 2. If the multiplier exceeds DCN3_2_NEW_DET_OVERRIDE_MIN_MULTIPLIER, then - * implement the modified DET override algorithm. - * 3. Assign smaller DET size for lower pixel display and higher DET size for - * higher pixel display + * @dc: Current DC state + * @context: New DC state to be programmed + * @pipes: Array of DML pipes * - * @param [in]: dc: Current DC state - * @param [in]: context: New DC state to be programmed - * @param [in]: pipes: Array of DML pipes - * - * @return: void - * - * ******************************************************************************************* + * Return: void */ void dcn32_determine_det_override(struct dc *dc, struct dc_state *context, @@ -309,31 +290,10 @@ void dcn32_determine_det_override(struct dc *dc, struct dc_plane_state *current_plane = NULL; uint8_t stream_count = 0; - int phy_pix_clk_mult, lower_mode_stream_index; - int phy_pix_clk[MAX_PIPES] = {0}; - bool use_new_det_override_algorithm = false; - for (i = 0; i < context->stream_count; i++) { /* Don't count SubVP streams for DET allocation */ - if (context->streams[i]->mall_stream_config.type != SUBVP_PHANTOM) { - phy_pix_clk[i] = context->streams[i]->phy_pix_clk; + if (context->streams[i]->mall_stream_config.type != SUBVP_PHANTOM) stream_count++; - } - } - - /* Check for special case with two displays, one with much higher pixel rate */ - if (stream_count == 2) { - ASSERT((phy_pix_clk[0] > 0) && (phy_pix_clk[1] > 0)); - if (phy_pix_clk[0] < phy_pix_clk[1]) { - lower_mode_stream_index = 0; - phy_pix_clk_mult = phy_pix_clk[1] / phy_pix_clk[0]; - } else { - lower_mode_stream_index = 1; - phy_pix_clk_mult = phy_pix_clk[0] / phy_pix_clk[1]; - } - - if (phy_pix_clk_mult >= DCN3_2_NEW_DET_OVERRIDE_MIN_MULTIPLIER) - use_new_det_override_algorithm = true; } if (stream_count > 0) { @@ -342,13 +302,6 @@ void dcn32_determine_det_override(struct dc *dc, if (context->streams[i]->mall_stream_config.type == SUBVP_PHANTOM) continue; - if (use_new_det_override_algorithm) { - if (i == lower_mode_stream_index) - stream_segments = 4; - else - stream_segments = 14; - } - if (context->stream_status[i].plane_count > 0) plane_segments = stream_segments / context->stream_status[i].plane_count; else @@ -432,8 +385,7 @@ void dcn32_set_det_allocations(struct dc *dc, struct dc_state *context, } /** - * ******************************************************************************************* - * dcn32_save_mall_state: Save MALL (SubVP) state for fast validation cases + * dcn32_save_mall_state(): Save MALL (SubVP) state for fast validation cases * * This function saves the MALL (SubVP) case for fast validation cases. For fast validation, * there are situations where a shallow copy of the dc->current_state is created for the @@ -446,13 +398,11 @@ void dcn32_set_det_allocations(struct dc *dc, struct dc_state *context, * NOTE: This function ONLY works if the streams are not moved to a different pipe in the * validation. We don't expect this to happen in fast_validation=1 cases. * - * @param [in]: dc: Current DC state - * @param [in]: context: New DC state to be programmed - * @param [out]: temp_config: struct used to cache the existing MALL state - * - * @return: void + * @dc: Current DC state + * @context: New DC state to be programmed + * @temp_config: struct used to cache the existing MALL state * - * ******************************************************************************************* + * Return: void */ void dcn32_save_mall_state(struct dc *dc, struct dc_state *context, @@ -472,18 +422,15 @@ void dcn32_save_mall_state(struct dc *dc, } /** - * ******************************************************************************************* - * dcn32_restore_mall_state: Restore MALL (SubVP) state for fast validation cases + * dcn32_restore_mall_state(): Restore MALL (SubVP) state for fast validation cases * * Restore the MALL state based on the previously saved state from dcn32_save_mall_state * - * @param [in]: dc: Current DC state - * @param [in/out]: context: New DC state to be programmed, restore MALL state into here - * @param [in]: temp_config: struct that has the cached MALL state + * @dc: Current DC state + * @context: New DC state to be programmed, restore MALL state into here + * @temp_config: struct that has the cached MALL state * - * @return: void - * - * ******************************************************************************************* + * Return: void */ void dcn32_restore_mall_state(struct dc *dc, struct dc_state *context, @@ -588,10 +535,11 @@ static int get_refresh_rate(struct dc_stream_state *fpo_candidate_stream) } /** - * dcn32_can_support_mclk_switch_using_fw_based_vblank_stretch - Determines if config can support FPO + * dcn32_can_support_mclk_switch_using_fw_based_vblank_stretch() - Determines if config can + * support FPO * - * @param [in]: dc - current dc state - * @param [in]: context - new dc state + * @dc: current dc state + * @context: new dc state * * Return: Pointer to FPO stream candidate if config can support FPO, otherwise NULL */ @@ -626,7 +574,7 @@ struct dc_stream_state *dcn32_can_support_mclk_switch_using_fw_based_vblank_stre DC_FP_END(); DC_FP_START(); - is_fpo_vactive = dcn32_find_vactive_pipe(dc, context, DCN3_2_MIN_ACTIVE_SWITCH_MARGIN_FPO_US); + is_fpo_vactive = dcn32_find_vactive_pipe(dc, context, dc->debug.fpo_vactive_min_active_margin_us); DC_FP_END(); if (!is_fpo_vactive || dc->debug.disable_fpo_vactive) return NULL; @@ -656,3 +604,120 @@ struct dc_stream_state *dcn32_can_support_mclk_switch_using_fw_based_vblank_stre return fpo_candidate_stream; } + +bool dcn32_check_native_scaling_for_res(struct pipe_ctx *pipe, unsigned int width, unsigned int height) +{ + bool is_native_scaling = false; + + if (pipe->stream->timing.h_addressable == width && + pipe->stream->timing.v_addressable == height && + pipe->plane_state->src_rect.width == width && + pipe->plane_state->src_rect.height == height && + pipe->plane_state->dst_rect.width == width && + pipe->plane_state->dst_rect.height == height) + is_native_scaling = true; + + return is_native_scaling; +} + +/** + * dcn32_subvp_drr_admissable() - Determine if SubVP + DRR config is admissible + * + * @dc: Current DC state + * @context: New DC state to be programmed + * + * SubVP + DRR is admissible under the following conditions: + * - Config must have 2 displays (i.e., 2 non-phantom master pipes) + * - One display is SubVP + * - Other display must have Freesync enabled + * - The potential DRR display must not be PSR capable + * + * Return: True if admissible, false otherwise + */ +bool dcn32_subvp_drr_admissable(struct dc *dc, struct dc_state *context) +{ + bool result = false; + uint32_t i; + uint8_t subvp_count = 0; + uint8_t non_subvp_pipes = 0; + bool drr_pipe_found = false; + bool drr_psr_capable = false; + + for (i = 0; i < dc->res_pool->pipe_count; i++) { + struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; + + if (!pipe->stream) + continue; + + if (pipe->plane_state && !pipe->top_pipe) { + if (pipe->stream->mall_stream_config.type == SUBVP_MAIN) + subvp_count++; + if (pipe->stream->mall_stream_config.type == SUBVP_NONE) { + non_subvp_pipes++; + drr_psr_capable = (drr_psr_capable || dcn32_is_psr_capable(pipe)); + if (pipe->stream->ignore_msa_timing_param && + (pipe->stream->allow_freesync || pipe->stream->vrr_active_variable)) { + drr_pipe_found = true; + } + } + } + } + + if (subvp_count == 1 && non_subvp_pipes == 1 && drr_pipe_found && !drr_psr_capable) + result = true; + + return result; +} + +/** + * dcn32_subvp_vblank_admissable() - Determine if SubVP + Vblank config is admissible + * + * @dc: Current DC state + * @context: New DC state to be programmed + * @vlevel: Voltage level calculated by DML + * + * SubVP + Vblank is admissible under the following conditions: + * - Config must have 2 displays (i.e., 2 non-phantom master pipes) + * - One display is SubVP + * - Other display must not have Freesync capability + * - DML must have output DRAM clock change support as SubVP + Vblank + * - The potential vblank display must not be PSR capable + * + * Return: True if admissible, false otherwise + */ +bool dcn32_subvp_vblank_admissable(struct dc *dc, struct dc_state *context, int vlevel) +{ + bool result = false; + uint32_t i; + uint8_t subvp_count = 0; + uint8_t non_subvp_pipes = 0; + bool drr_pipe_found = false; + struct vba_vars_st *vba = &context->bw_ctx.dml.vba; + bool vblank_psr_capable = false; + + for (i = 0; i < dc->res_pool->pipe_count; i++) { + struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; + + if (!pipe->stream) + continue; + + if (pipe->plane_state && !pipe->top_pipe) { + if (pipe->stream->mall_stream_config.type == SUBVP_MAIN) + subvp_count++; + if (pipe->stream->mall_stream_config.type == SUBVP_NONE) { + non_subvp_pipes++; + vblank_psr_capable = (vblank_psr_capable || dcn32_is_psr_capable(pipe)); + if (pipe->stream->ignore_msa_timing_param && + (pipe->stream->allow_freesync || pipe->stream->vrr_active_variable)) { + drr_pipe_found = true; + } + } + } + } + + if (subvp_count == 1 && non_subvp_pipes == 1 && !drr_pipe_found && !vblank_psr_capable && + vba->DRAMClockChangeSupport[vlevel][vba->maxMpcComb] == dm_dram_clock_change_vblank_w_mall_sub_vp) + result = true; + + return result; +} diff --git a/drivers/gpu/drm/amd/display/dc/dcn321/dcn321_resource.c b/drivers/gpu/drm/amd/display/dc/dcn321/dcn321_resource.c index a60ddb343d13..ea204742ad35 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn321/dcn321_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn321/dcn321_resource.c @@ -725,31 +725,13 @@ static const struct dc_debug_options debug_defaults_drv = { .override_dispclk_programming = true, .disable_fpo_optimizations = false, .fpo_vactive_margin_us = 2000, // 2000us - .disable_fpo_vactive = true, + .disable_fpo_vactive = false, .disable_boot_optimizations = false, + .disable_subvp_high_refresh = false, + .fpo_vactive_min_active_margin_us = 200, + .fpo_vactive_max_blank_us = 1000, }; -static const struct dc_debug_options debug_defaults_diags = { - .disable_dmcu = true, - .force_abm_enable = false, - .timing_trace = true, - .clock_trace = true, - .disable_dpp_power_gate = true, - .disable_hubp_power_gate = true, - .disable_dsc_power_gate = true, - .disable_clock_gate = true, - .disable_pplib_clock_request = true, - .disable_pplib_wm_range = true, - .disable_stutter = false, - .scl_reset_length10 = true, - .dwb_fi_phase = -1, // -1 = disable - .dmub_command_table = true, - .enable_tri_buf = true, - .use_max_lb = true, - .force_disable_subvp = true, -}; - - static struct dce_aux *dcn321_aux_engine_create( struct dc_context *ctx, uint32_t inst) @@ -1340,15 +1322,6 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dcn321_hwseq_create, }; -static const struct resource_create_funcs res_create_maximus_funcs = { - .read_dce_straps = NULL, - .create_audio = NULL, - .create_stream_encoder = NULL, - .create_hpo_dp_stream_encoder = dcn321_hpo_dp_stream_encoder_create, - .create_hpo_dp_link_encoder = dcn321_hpo_dp_link_encoder_create, - .create_hwseq = dcn321_hwseq_create, -}; - static void dcn321_resource_destruct(struct dcn321_resource_pool *pool) { unsigned int i; @@ -1735,9 +1708,9 @@ static bool dcn321_resource_construct( dc->caps.subvp_pstate_allow_width_us = 20; dc->caps.subvp_vertical_int_margin_us = 30; dc->caps.subvp_drr_vblank_start_margin_us = 100; // 100us margin - dc->caps.max_slave_planes = 1; - dc->caps.max_slave_yuv_planes = 1; - dc->caps.max_slave_rgb_planes = 1; + dc->caps.max_slave_planes = 2; + dc->caps.max_slave_yuv_planes = 2; + dc->caps.max_slave_rgb_planes = 2; dc->caps.post_blend_color_processing = true; dc->caps.force_dp_tps4_for_cp2520 = true; dc->caps.dp_hpo = true; @@ -1745,6 +1718,7 @@ static bool dcn321_resource_construct( dc->caps.edp_dsc_support = true; dc->caps.extended_aux_timeout_support = true; dc->caps.dmcub_support = true; + dc->caps.max_v_total = (1 << 15) - 1; /* Color pipeline capabilities */ dc->caps.color.dpp.dcn_arch = 1; @@ -1798,10 +1772,7 @@ static bool dcn321_resource_construct( if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) dc->debug = debug_defaults_drv; - else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) { - dc->debug = debug_defaults_diags; - } else - dc->debug = debug_defaults_diags; + // Init the vm_helper if (dc->vm_helper) vm_helper_init(dc->vm_helper, 16); @@ -1857,8 +1828,7 @@ static bool dcn321_resource_construct( } /* DML */ - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) - dml_init_instance(&dc->dml, &dcn3_21_soc, &dcn3_21_ip, DML_PROJECT_DCN32); + dml_init_instance(&dc->dml, &dcn3_21_soc, &dcn3_21_ip, DML_PROJECT_DCN32); /* IRQ Service */ init_data.ctx = dc->ctx; @@ -1990,9 +1960,8 @@ static bool dcn321_resource_construct( /* Audio, HWSeq, Stream Encoders including HPO and virtual, MPC 3D LUTs */ if (!resource_construct(num_virtual_links, dc, &pool->base, - (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ? - &res_create_funcs : &res_create_maximus_funcs))) - goto create_fail; + &res_create_funcs)) + goto create_fail; /* HW Sequencer init functions and Plane caps */ dcn32_hw_sequencer_init_functions(dc); diff --git a/drivers/gpu/drm/amd/display/dc/dm_services.h b/drivers/gpu/drm/amd/display/dc/dm_services.h index 9a3f2a44f882..d0eed3b4771e 100644 --- a/drivers/gpu/drm/amd/display/dc/dm_services.h +++ b/drivers/gpu/drm/amd/display/dc/dm_services.h @@ -40,6 +40,7 @@ struct dmub_srv; struct dc_dmub_srv; +union dmub_rb_cmd; irq_handler_idx dm_register_interrupt( struct dc_context *ctx, @@ -274,6 +275,12 @@ void dm_perf_trace_timestamp(const char *func_name, unsigned int line, struct dc #define PERF_TRACE_CTX(__CTX) dm_perf_trace_timestamp(__func__, __LINE__, __CTX) /* + * DMUB Interfaces + */ +bool dm_execute_dmub_cmd(const struct dc_context *ctx, union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type); +bool dm_execute_dmub_cmd_list(const struct dc_context *ctx, unsigned int count, union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type); + +/* * Debug and verification hooks */ @@ -285,4 +292,6 @@ void dm_dtn_log_append_v(struct dc_context *ctx, void dm_dtn_log_end(struct dc_context *ctx, struct dc_log_buffer_ctx *log_ctx); +char *dce_version_to_string(const int version); + #endif /* __DM_SERVICES_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/dm_services_types.h b/drivers/gpu/drm/amd/display/dc/dm_services_types.h index b52ba6ffabe1..facf269c4326 100644 --- a/drivers/gpu/drm/amd/display/dc/dm_services_types.h +++ b/drivers/gpu/drm/amd/display/dc/dm_services_types.h @@ -269,4 +269,10 @@ struct dtn_min_clk_info { uint32_t min_memory_clock_khz; }; +enum dm_dmub_wait_type { + DM_DMUB_WAIT_TYPE_NO_WAIT, + DM_DMUB_WAIT_TYPE_WAIT, + DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY, +}; + #endif diff --git a/drivers/gpu/drm/amd/display/dc/dml/Makefile b/drivers/gpu/drm/amd/display/dc/dml/Makefile index 01db035589c5..77cf5545c94c 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/Makefile +++ b/drivers/gpu/drm/amd/display/dc/dml/Makefile @@ -38,6 +38,11 @@ ifdef CONFIG_ARM64 dml_rcflags := -mgeneral-regs-only endif +ifdef CONFIG_LOONGARCH +dml_ccflags := -mfpu=64 +dml_rcflags := -msoft-float +endif + ifdef CONFIG_CC_IS_GCC ifneq ($(call gcc-min-version, 70100),y) IS_OLD_GCC = 1 diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c index f1c1a4b5fcac..8ae5ddbd1b27 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c @@ -948,10 +948,10 @@ static enum dcn_zstate_support_state decide_zstate_support(struct dc *dc, struc { int plane_count; int i; - unsigned int optimized_min_dst_y_next_start_us; + unsigned int min_dst_y_next_start_us; plane_count = 0; - optimized_min_dst_y_next_start_us = 0; + min_dst_y_next_start_us = 0; for (i = 0; i < dc->res_pool->pipe_count; i++) { if (context->res_ctx.pipe_ctx[i].plane_state) plane_count++; @@ -973,19 +973,18 @@ static enum dcn_zstate_support_state decide_zstate_support(struct dc *dc, struc else if (context->stream_count == 1 && context->streams[0]->signal == SIGNAL_TYPE_EDP) { struct dc_link *link = context->streams[0]->sink->link; struct dc_stream_status *stream_status = &context->stream_status[0]; + struct dc_stream_state *current_stream = context->streams[0]; int minmum_z8_residency = dc->debug.minimum_z8_residency_time > 0 ? dc->debug.minimum_z8_residency_time : 1000; bool allow_z8 = context->bw_ctx.dml.vba.StutterPeriod > (double)minmum_z8_residency; bool is_pwrseq0 = link->link_index == 0; + bool isFreesyncVideo; - if (dc_extended_blank_supported(dc)) { - for (i = 0; i < dc->res_pool->pipe_count; i++) { - if (context->res_ctx.pipe_ctx[i].stream == context->streams[0] - && context->res_ctx.pipe_ctx[i].stream->adjust.v_total_min == context->res_ctx.pipe_ctx[i].stream->adjust.v_total_max - && context->res_ctx.pipe_ctx[i].stream->adjust.v_total_min > context->res_ctx.pipe_ctx[i].stream->timing.v_total) { - optimized_min_dst_y_next_start_us = - context->res_ctx.pipe_ctx[i].dlg_regs.optimized_min_dst_y_next_start_us; - break; - } + isFreesyncVideo = current_stream->adjust.v_total_min == current_stream->adjust.v_total_max; + isFreesyncVideo = isFreesyncVideo && current_stream->timing.v_total < current_stream->adjust.v_total_min; + for (i = 0; i < dc->res_pool->pipe_count; i++) { + if (context->res_ctx.pipe_ctx[i].stream == current_stream && isFreesyncVideo) { + min_dst_y_next_start_us = context->res_ctx.pipe_ctx[i].dlg_regs.min_dst_y_next_start_us; + break; } } @@ -993,7 +992,7 @@ static enum dcn_zstate_support_state decide_zstate_support(struct dc *dc, struc if (stream_status->plane_count > 1) return DCN_ZSTATE_SUPPORT_DISALLOW; - if (is_pwrseq0 && (context->bw_ctx.dml.vba.StutterPeriod > 5000.0 || optimized_min_dst_y_next_start_us > 5000)) + if (is_pwrseq0 && (context->bw_ctx.dml.vba.StutterPeriod > 5000.0 || min_dst_y_next_start_us > 5000)) return DCN_ZSTATE_SUPPORT_ALLOW; else if (is_pwrseq0 && link->psr_settings.psr_version == DC_PSR_VERSION_1 && !link->panel_config.psr.disable_psr) return allow_z8 ? DCN_ZSTATE_SUPPORT_ALLOW_Z8_Z10_ONLY : DCN_ZSTATE_SUPPORT_ALLOW_Z10_ONLY; @@ -1043,7 +1042,7 @@ void dcn20_calculate_dlg_params(struct dc *dc, int pipe_cnt, int vlevel) { - int i, pipe_idx; + int i, pipe_idx, active_hubp_count = 0; dc_assert_fp_enabled(); @@ -1079,6 +1078,8 @@ void dcn20_calculate_dlg_params(struct dc *dc, for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) { if (!context->res_ctx.pipe_ctx[i].stream) continue; + if (context->res_ctx.pipe_ctx[i].plane_state) + active_hubp_count++; pipes[pipe_idx].pipe.dest.vstartup_start = get_vstartup(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx); pipes[pipe_idx].pipe.dest.vupdate_offset = get_vupdate_offset(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx); pipes[pipe_idx].pipe.dest.vupdate_width = get_vupdate_width(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx); @@ -1098,13 +1099,13 @@ void dcn20_calculate_dlg_params(struct dc *dc, context->res_ctx.pipe_ctx[i].plane_res.bw.dppclk_khz = pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000; context->res_ctx.pipe_ctx[i].pipe_dlg_param = pipes[pipe_idx].pipe.dest; - if (context->res_ctx.pipe_ctx[i].stream->adaptive_sync_infopacket.valid) - dcn20_adjust_freesync_v_startup( - &context->res_ctx.pipe_ctx[i].stream->timing, - &context->res_ctx.pipe_ctx[i].pipe_dlg_param.vstartup_start); pipe_idx++; } + /* If DCN isn't making memory requests we can allow pstate change */ + if (!active_hubp_count) { + context->bw_ctx.bw.dcn.clk.p_state_change_support = true; + } /*save a original dppclock copy*/ context->bw_ctx.bw.dcn.clk.bw_dppclk_khz = context->bw_ctx.bw.dcn.clk.dppclk_khz; context->bw_ctx.bw.dcn.clk.bw_dispclk_khz = context->bw_ctx.bw.dcn.clk.dispclk_khz; @@ -1885,6 +1886,17 @@ void dcn20_patch_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_st dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0; } + if ((int)(bb->sr_exit_z8_time_us * 1000) + != dc->bb_overrides.sr_exit_z8_time_ns + && dc->bb_overrides.sr_exit_z8_time_ns) { + bb->sr_exit_z8_time_us = dc->bb_overrides.sr_exit_z8_time_ns / 1000.0; + } + + if ((int)(bb->sr_enter_plus_exit_z8_time_us * 1000) + != dc->bb_overrides.sr_enter_plus_exit_z8_time_ns + && dc->bb_overrides.sr_enter_plus_exit_z8_time_ns) { + bb->sr_enter_plus_exit_z8_time_us = dc->bb_overrides.sr_enter_plus_exit_z8_time_ns / 1000.0; + } if ((int)(bb->urgent_latency_us * 1000) != dc->bb_overrides.urgent_latency_ns && dc->bb_overrides.urgent_latency_ns) { bb->urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0; @@ -1915,6 +1927,7 @@ static bool dcn20_validate_bandwidth_internal(struct dc *dc, struct dc_state *co int vlevel = 0; int pipe_split_from[MAX_PIPES]; int pipe_cnt = 0; + int i = 0; display_e2e_pipe_params_st *pipes = kzalloc(dc->res_pool->pipe_count * sizeof(display_e2e_pipe_params_st), GFP_ATOMIC); DC_LOGGER_INIT(dc->ctx->logger); @@ -1938,6 +1951,15 @@ static bool dcn20_validate_bandwidth_internal(struct dc *dc, struct dc_state *co dcn20_calculate_wm(dc, context, pipes, &pipe_cnt, pipe_split_from, vlevel, fast_validate); dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel); + for (i = 0; i < dc->res_pool->pipe_count; i++) { + if (!context->res_ctx.pipe_ctx[i].stream) + continue; + if (context->res_ctx.pipe_ctx[i].stream->adaptive_sync_infopacket.valid) + dcn20_adjust_freesync_v_startup( + &context->res_ctx.pipe_ctx[i].stream->timing, + &context->res_ctx.pipe_ctx[i].pipe_dlg_param.vstartup_start); + } + BW_VAL_TRACE_END_WATERMARKS(); goto validate_out; @@ -2210,6 +2232,7 @@ bool dcn21_validate_bandwidth_fp(struct dc *dc, int vlevel = 0; int pipe_split_from[MAX_PIPES]; int pipe_cnt = 0; + int i = 0; display_e2e_pipe_params_st *pipes = kzalloc(dc->res_pool->pipe_count * sizeof(display_e2e_pipe_params_st), GFP_ATOMIC); DC_LOGGER_INIT(dc->ctx->logger); @@ -2238,6 +2261,15 @@ bool dcn21_validate_bandwidth_fp(struct dc *dc, dcn21_calculate_wm(dc, context, pipes, &pipe_cnt, pipe_split_from, vlevel, fast_validate); dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel); + for (i = 0; i < dc->res_pool->pipe_count; i++) { + if (!context->res_ctx.pipe_ctx[i].stream) + continue; + if (context->res_ctx.pipe_ctx[i].stream->adaptive_sync_infopacket.valid) + dcn20_adjust_freesync_v_startup( + &context->res_ctx.pipe_ctx[i].stream->timing, + &context->res_ctx.pipe_ctx[i].pipe_dlg_param.vstartup_start); + } + BW_VAL_TRACE_END_WATERMARKS(); goto validate_out; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_mode_vba_21.c b/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_mode_vba_21.c index b7c2844d0cbe..f294f2f8c75b 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_mode_vba_21.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_mode_vba_21.c @@ -810,7 +810,7 @@ static bool CalculatePrefetchSchedule( *swath_width_chroma_ub = dml_ceil(SwathWidthY / 2 - 1, myPipe->BlockWidth256BytesC) + myPipe->BlockWidth256BytesC; } else { *swath_width_luma_ub = dml_ceil(SwathWidthY - 1, myPipe->BlockHeight256BytesY) + myPipe->BlockHeight256BytesY; - if (myPipe->BlockWidth256BytesC > 0) + if (myPipe->BlockHeight256BytesC > 0) *swath_width_chroma_ub = dml_ceil(SwathWidthY / 2 - 1, myPipe->BlockHeight256BytesC) + myPipe->BlockHeight256BytesC; } diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.c index a352c703e258..ccb4ad78f667 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.c @@ -674,10 +674,19 @@ void dcn30_fpu_update_bw_bounding_box(struct dc *dc, } /** - * Finds dummy_latency_index when MCLK switching using firmware based - * vblank stretch is enabled. This function will iterate through the - * table of dummy pstate latencies until the lowest value that allows + * dcn30_find_dummy_latency_index_for_fw_based_mclk_switch() - Finds + * dummy_latency_index when MCLK switching using firmware based vblank stretch + * is enabled. This function will iterate through the table of dummy pstate + * latencies until the lowest value that allows * dm_allow_self_refresh_and_mclk_switch to happen is found + * + * @dc: Current DC state + * @context: new dc state + * @pipes: DML pipe params + * @pipe_cnt: number of DML pipes + * @vlevel: Voltage level calculated by DML + * + * Return: lowest dummy_latency_index value */ int dcn30_find_dummy_latency_index_for_fw_based_mclk_switch(struct dc *dc, struct dc_state *context, diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c index 7d0626e42ea6..9af1a43c042b 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c @@ -4939,8 +4939,8 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l } v->TotImmediateFlipBytes = 0.0; for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) { - v->TotImmediateFlipBytes = v->TotImmediateFlipBytes + v->NoOfDPP[i][j][k] * v->PDEAndMetaPTEBytesPerFrame[i][j][k] - + v->MetaRowBytes[i][j][k] + v->DPTEBytesPerRow[i][j][k]; + v->TotImmediateFlipBytes = v->TotImmediateFlipBytes + v->NoOfDPP[i][j][k] * (v->PDEAndMetaPTEBytesPerFrame[i][j][k] + + v->MetaRowBytes[i][j][k] + v->DPTEBytesPerRow[i][j][k]); } for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) { @@ -5130,7 +5130,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l ViewportExceedsSurface = true; if (v->SourcePixelFormat[k] != dm_444_64 && v->SourcePixelFormat[k] != dm_444_32 && v->SourcePixelFormat[k] != dm_444_16 - && v->SourcePixelFormat[k] != dm_444_8 && v->SourcePixelFormat[k] != dm_rgbe) { + && v->SourcePixelFormat[k] != dm_444_16 && v->SourcePixelFormat[k] != dm_444_8 && v->SourcePixelFormat[k] != dm_rgbe) { if (v->ViewportWidthChroma[k] > v->SurfaceWidthC[k] || v->ViewportHeightChroma[k] > v->SurfaceHeightC[k]) { ViewportExceedsSurface = true; } diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c index cd3cfcb2a2b0..0497a5d74a62 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c @@ -980,7 +980,7 @@ static void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib, unsigned int vstartup_start = 0; unsigned int dst_x_after_scaler = 0; - unsigned int dst_y_after_scaler = 0; + int dst_y_after_scaler = 0; double line_wait = 0; double dst_y_prefetch = 0; double dst_y_per_vm_vblank = 0; @@ -1171,6 +1171,8 @@ static void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib, dst_x_after_scaler = get_dst_x_after_scaler(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); dst_y_after_scaler = get_dst_y_after_scaler(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); + if (dst_y_after_scaler < 0) + dst_y_after_scaler = 0; // do some adjustment on the dst_after scaler to account for odm combine mode dml_print("DML_DLG: %s: input dst_x_after_scaler = %d\n", diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c index 422f17aefd4a..6ce90678b33c 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c @@ -333,45 +333,43 @@ void dcn301_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_param memcpy(s, dcn3_01_soc.clock_limits, sizeof(dcn3_01_soc.clock_limits)); /* Default clock levels are used for diags, which may lead to overclocking. */ - if (!IS_DIAG_DC(dc->ctx->dce_environment)) { - dcn3_01_ip.max_num_otg = pool->base.res_cap->num_timing_generator; - dcn3_01_ip.max_num_dpp = pool->base.pipe_count; - dcn3_01_soc.num_chans = bw_params->num_channels; - - ASSERT(clk_table->num_entries); - for (i = 0; i < clk_table->num_entries; i++) { - /* loop backwards*/ - for (closest_clk_lvl = 0, j = dcn3_01_soc.num_states - 1; j >= 0; j--) { - if ((unsigned int) dcn3_01_soc.clock_limits[j].dcfclk_mhz <= clk_table->entries[i].dcfclk_mhz) { - closest_clk_lvl = j; - break; - } + dcn3_01_ip.max_num_otg = pool->base.res_cap->num_timing_generator; + dcn3_01_ip.max_num_dpp = pool->base.pipe_count; + dcn3_01_soc.num_chans = bw_params->num_channels; + + ASSERT(clk_table->num_entries); + for (i = 0; i < clk_table->num_entries; i++) { + /* loop backwards*/ + for (closest_clk_lvl = 0, j = dcn3_01_soc.num_states - 1; j >= 0; j--) { + if ((unsigned int) dcn3_01_soc.clock_limits[j].dcfclk_mhz <= clk_table->entries[i].dcfclk_mhz) { + closest_clk_lvl = j; + break; } - - s[i].state = i; - s[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz; - s[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz; - s[i].socclk_mhz = clk_table->entries[i].socclk_mhz; - s[i].dram_speed_mts = clk_table->entries[i].memclk_mhz * 2; - - s[i].dispclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dispclk_mhz; - s[i].dppclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dppclk_mhz; - s[i].dram_bw_per_chan_gbps = - dcn3_01_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps; - s[i].dscclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dscclk_mhz; - s[i].dtbclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dtbclk_mhz; - s[i].phyclk_d18_mhz = - dcn3_01_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz; - s[i].phyclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].phyclk_mhz; } - if (clk_table->num_entries) { - dcn3_01_soc.num_states = clk_table->num_entries; - /* duplicate last level */ - s[dcn3_01_soc.num_states] = - dcn3_01_soc.clock_limits[dcn3_01_soc.num_states - 1]; - s[dcn3_01_soc.num_states].state = dcn3_01_soc.num_states; - } + s[i].state = i; + s[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz; + s[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz; + s[i].socclk_mhz = clk_table->entries[i].socclk_mhz; + s[i].dram_speed_mts = clk_table->entries[i].memclk_mhz * 2; + + s[i].dispclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dispclk_mhz; + s[i].dppclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dppclk_mhz; + s[i].dram_bw_per_chan_gbps = + dcn3_01_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps; + s[i].dscclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dscclk_mhz; + s[i].dtbclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dtbclk_mhz; + s[i].phyclk_d18_mhz = + dcn3_01_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz; + s[i].phyclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].phyclk_mhz; + } + + if (clk_table->num_entries) { + dcn3_01_soc.num_states = clk_table->num_entries; + /* duplicate last level */ + s[dcn3_01_soc.num_states] = + dcn3_01_soc.clock_limits[dcn3_01_soc.num_states - 1]; + s[dcn3_01_soc.num_states].state = dcn3_01_soc.num_states; } memcpy(dcn3_01_soc.clock_limits, s, sizeof(dcn3_01_soc.clock_limits)); diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c index 59836570603a..deb6d162a2d5 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c @@ -483,7 +483,7 @@ void dcn31_calculate_wm_and_dlg_fp( int pipe_cnt, int vlevel) { - int i, pipe_idx, active_hubp_count = 0; + int i, pipe_idx, total_det = 0, active_hubp_count = 0; double dcfclk = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb]; dc_assert_fp_enabled(); @@ -563,6 +563,18 @@ void dcn31_calculate_wm_and_dlg_fp( if (context->res_ctx.pipe_ctx[i].stream) context->res_ctx.pipe_ctx[i].plane_res.bw.dppclk_khz = 0; } + for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) { + if (!context->res_ctx.pipe_ctx[i].stream) + continue; + + context->res_ctx.pipe_ctx[i].det_buffer_size_kb = + get_det_buffer_size_kbytes(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx); + if (context->res_ctx.pipe_ctx[i].det_buffer_size_kb > 384) + context->res_ctx.pipe_ctx[i].det_buffer_size_kb /= 2; + total_det += context->res_ctx.pipe_ctx[i].det_buffer_size_kb; + pipe_idx++; + } + context->bw_ctx.bw.dcn.compbuf_size_kb = context->bw_ctx.dml.ip.config_return_buffer_size_in_kbytes - total_det; } void dcn31_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params) @@ -570,6 +582,7 @@ void dcn31_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params struct _vcs_dpi_voltage_scaling_st *s = dc->scratch.update_bw_bounding_box.clock_limits; struct clk_limit_table *clk_table = &bw_params->clk_table; unsigned int i, closest_clk_lvl; + int max_dispclk_mhz = 0, max_dppclk_mhz = 0; int j; dc_assert_fp_enabled(); @@ -577,59 +590,55 @@ void dcn31_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params memcpy(s, dcn3_1_soc.clock_limits, sizeof(dcn3_1_soc.clock_limits)); // Default clock levels are used for diags, which may lead to overclocking. - if (!IS_DIAG_DC(dc->ctx->dce_environment)) { - int max_dispclk_mhz = 0, max_dppclk_mhz = 0; + dcn3_1_ip.max_num_otg = dc->res_pool->res_cap->num_timing_generator; + dcn3_1_ip.max_num_dpp = dc->res_pool->pipe_count; + dcn3_1_soc.num_chans = bw_params->num_channels; - dcn3_1_ip.max_num_otg = dc->res_pool->res_cap->num_timing_generator; - dcn3_1_ip.max_num_dpp = dc->res_pool->pipe_count; - dcn3_1_soc.num_chans = bw_params->num_channels; + ASSERT(clk_table->num_entries); - ASSERT(clk_table->num_entries); + /* Prepass to find max clocks independent of voltage level. */ + for (i = 0; i < clk_table->num_entries; ++i) { + if (clk_table->entries[i].dispclk_mhz > max_dispclk_mhz) + max_dispclk_mhz = clk_table->entries[i].dispclk_mhz; + if (clk_table->entries[i].dppclk_mhz > max_dppclk_mhz) + max_dppclk_mhz = clk_table->entries[i].dppclk_mhz; + } - /* Prepass to find max clocks independent of voltage level. */ - for (i = 0; i < clk_table->num_entries; ++i) { - if (clk_table->entries[i].dispclk_mhz > max_dispclk_mhz) - max_dispclk_mhz = clk_table->entries[i].dispclk_mhz; - if (clk_table->entries[i].dppclk_mhz > max_dppclk_mhz) - max_dppclk_mhz = clk_table->entries[i].dppclk_mhz; + for (i = 0; i < clk_table->num_entries; i++) { + /* loop backwards*/ + for (closest_clk_lvl = 0, j = dcn3_1_soc.num_states - 1; j >= 0; j--) { + if ((unsigned int) dcn3_1_soc.clock_limits[j].dcfclk_mhz <= clk_table->entries[i].dcfclk_mhz) { + closest_clk_lvl = j; + break; + } } - for (i = 0; i < clk_table->num_entries; i++) { - /* loop backwards*/ - for (closest_clk_lvl = 0, j = dcn3_1_soc.num_states - 1; j >= 0; j--) { - if ((unsigned int) dcn3_1_soc.clock_limits[j].dcfclk_mhz <= clk_table->entries[i].dcfclk_mhz) { - closest_clk_lvl = j; - break; - } - } + s[i].state = i; - s[i].state = i; - - /* Clocks dependent on voltage level. */ - s[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz; - s[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz; - s[i].socclk_mhz = clk_table->entries[i].socclk_mhz; - s[i].dram_speed_mts = clk_table->entries[i].memclk_mhz * - 2 * clk_table->entries[i].wck_ratio; - - /* Clocks independent of voltage level. */ - s[i].dispclk_mhz = max_dispclk_mhz ? max_dispclk_mhz : - dcn3_1_soc.clock_limits[closest_clk_lvl].dispclk_mhz; - - s[i].dppclk_mhz = max_dppclk_mhz ? max_dppclk_mhz : - dcn3_1_soc.clock_limits[closest_clk_lvl].dppclk_mhz; - - s[i].dram_bw_per_chan_gbps = - dcn3_1_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps; - s[i].dscclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].dscclk_mhz; - s[i].dtbclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].dtbclk_mhz; - s[i].phyclk_d18_mhz = - dcn3_1_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz; - s[i].phyclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].phyclk_mhz; - } - if (clk_table->num_entries) { - dcn3_1_soc.num_states = clk_table->num_entries; - } + /* Clocks dependent on voltage level. */ + s[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz; + s[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz; + s[i].socclk_mhz = clk_table->entries[i].socclk_mhz; + s[i].dram_speed_mts = clk_table->entries[i].memclk_mhz * + 2 * clk_table->entries[i].wck_ratio; + + /* Clocks independent of voltage level. */ + s[i].dispclk_mhz = max_dispclk_mhz ? max_dispclk_mhz : + dcn3_1_soc.clock_limits[closest_clk_lvl].dispclk_mhz; + + s[i].dppclk_mhz = max_dppclk_mhz ? max_dppclk_mhz : + dcn3_1_soc.clock_limits[closest_clk_lvl].dppclk_mhz; + + s[i].dram_bw_per_chan_gbps = + dcn3_1_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps; + s[i].dscclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].dscclk_mhz; + s[i].dtbclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].dtbclk_mhz; + s[i].phyclk_d18_mhz = + dcn3_1_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz; + s[i].phyclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].phyclk_mhz; + } + if (clk_table->num_entries) { + dcn3_1_soc.num_states = clk_table->num_entries; } memcpy(dcn3_1_soc.clock_limits, s, sizeof(dcn3_1_soc.clock_limits)); @@ -643,10 +652,7 @@ void dcn31_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params dcn3_1_soc.dram_clock_change_latency_us = dc->debug.dram_clock_change_latency_ns / 1000; } - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) - dml_init_instance(&dc->dml, &dcn3_1_soc, &dcn3_1_ip, DML_PROJECT_DCN31); - else - dml_init_instance(&dc->dml, &dcn3_1_soc, &dcn3_1_ip, DML_PROJECT_DCN31_FPGA); + dml_init_instance(&dc->dml, &dcn3_1_soc, &dcn3_1_ip, DML_PROJECT_DCN31); } void dcn315_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params) @@ -707,10 +713,7 @@ void dcn315_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_param dcn3_15_soc.dram_clock_change_latency_us = dc->debug.dram_clock_change_latency_ns / 1000; } - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) - dml_init_instance(&dc->dml, &dcn3_15_soc, &dcn3_15_ip, DML_PROJECT_DCN315); - else - dml_init_instance(&dc->dml, &dcn3_15_soc, &dcn3_15_ip, DML_PROJECT_DCN31_FPGA); + dml_init_instance(&dc->dml, &dcn3_15_soc, &dcn3_15_ip, DML_PROJECT_DCN315); } void dcn316_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params) @@ -726,71 +729,68 @@ void dcn316_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_param memcpy(s, dcn3_16_soc.clock_limits, sizeof(dcn3_16_soc.clock_limits)); // Default clock levels are used for diags, which may lead to overclocking. - if (!IS_DIAG_DC(dc->ctx->dce_environment)) { + dcn3_16_ip.max_num_otg = dc->res_pool->res_cap->num_timing_generator; + dcn3_16_ip.max_num_dpp = dc->res_pool->pipe_count; + dcn3_16_soc.num_chans = bw_params->num_channels; - dcn3_16_ip.max_num_otg = dc->res_pool->res_cap->num_timing_generator; - dcn3_16_ip.max_num_dpp = dc->res_pool->pipe_count; - dcn3_16_soc.num_chans = bw_params->num_channels; - - ASSERT(clk_table->num_entries); + ASSERT(clk_table->num_entries); - /* Prepass to find max clocks independent of voltage level. */ - for (i = 0; i < clk_table->num_entries; ++i) { - if (clk_table->entries[i].dispclk_mhz > max_dispclk_mhz) - max_dispclk_mhz = clk_table->entries[i].dispclk_mhz; - if (clk_table->entries[i].dppclk_mhz > max_dppclk_mhz) - max_dppclk_mhz = clk_table->entries[i].dppclk_mhz; - } + /* Prepass to find max clocks independent of voltage level. */ + for (i = 0; i < clk_table->num_entries; ++i) { + if (clk_table->entries[i].dispclk_mhz > max_dispclk_mhz) + max_dispclk_mhz = clk_table->entries[i].dispclk_mhz; + if (clk_table->entries[i].dppclk_mhz > max_dppclk_mhz) + max_dppclk_mhz = clk_table->entries[i].dppclk_mhz; + } - for (i = 0; i < clk_table->num_entries; i++) { - /* loop backwards*/ - for (closest_clk_lvl = 0, j = dcn3_16_soc.num_states - 1; j >= 0; j--) { - if ((unsigned int) dcn3_16_soc.clock_limits[j].dcfclk_mhz <= - clk_table->entries[i].dcfclk_mhz) { - closest_clk_lvl = j; - break; - } - } - // Ported from DCN315 - if (clk_table->num_entries == 1) { - /*smu gives one DPM level, let's take the highest one*/ - closest_clk_lvl = dcn3_16_soc.num_states - 1; + for (i = 0; i < clk_table->num_entries; i++) { + /* loop backwards*/ + for (closest_clk_lvl = 0, j = dcn3_16_soc.num_states - 1; j >= 0; j--) { + if ((unsigned int) dcn3_16_soc.clock_limits[j].dcfclk_mhz <= + clk_table->entries[i].dcfclk_mhz) { + closest_clk_lvl = j; + break; } + } + // Ported from DCN315 + if (clk_table->num_entries == 1) { + /*smu gives one DPM level, let's take the highest one*/ + closest_clk_lvl = dcn3_16_soc.num_states - 1; + } - s[i].state = i; + s[i].state = i; - /* Clocks dependent on voltage level. */ - s[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz; - if (clk_table->num_entries == 1 && - s[i].dcfclk_mhz < - dcn3_16_soc.clock_limits[closest_clk_lvl].dcfclk_mhz) { - /*SMU fix not released yet*/ - s[i].dcfclk_mhz = - dcn3_16_soc.clock_limits[closest_clk_lvl].dcfclk_mhz; - } - s[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz; - s[i].socclk_mhz = clk_table->entries[i].socclk_mhz; - s[i].dram_speed_mts = clk_table->entries[i].memclk_mhz * - 2 * clk_table->entries[i].wck_ratio; - - /* Clocks independent of voltage level. */ - s[i].dispclk_mhz = max_dispclk_mhz ? max_dispclk_mhz : - dcn3_16_soc.clock_limits[closest_clk_lvl].dispclk_mhz; - - s[i].dppclk_mhz = max_dppclk_mhz ? max_dppclk_mhz : - dcn3_16_soc.clock_limits[closest_clk_lvl].dppclk_mhz; - - s[i].dram_bw_per_chan_gbps = - dcn3_16_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps; - s[i].dscclk_mhz = dcn3_16_soc.clock_limits[closest_clk_lvl].dscclk_mhz; - s[i].dtbclk_mhz = dcn3_16_soc.clock_limits[closest_clk_lvl].dtbclk_mhz; - s[i].phyclk_d18_mhz = - dcn3_16_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz; - s[i].phyclk_mhz = dcn3_16_soc.clock_limits[closest_clk_lvl].phyclk_mhz; - } - if (clk_table->num_entries) { - dcn3_16_soc.num_states = clk_table->num_entries; + /* Clocks dependent on voltage level. */ + s[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz; + if (clk_table->num_entries == 1 && + s[i].dcfclk_mhz < + dcn3_16_soc.clock_limits[closest_clk_lvl].dcfclk_mhz) { + /*SMU fix not released yet*/ + s[i].dcfclk_mhz = + dcn3_16_soc.clock_limits[closest_clk_lvl].dcfclk_mhz; } + s[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz; + s[i].socclk_mhz = clk_table->entries[i].socclk_mhz; + s[i].dram_speed_mts = clk_table->entries[i].memclk_mhz * + 2 * clk_table->entries[i].wck_ratio; + + /* Clocks independent of voltage level. */ + s[i].dispclk_mhz = max_dispclk_mhz ? max_dispclk_mhz : + dcn3_16_soc.clock_limits[closest_clk_lvl].dispclk_mhz; + + s[i].dppclk_mhz = max_dppclk_mhz ? max_dppclk_mhz : + dcn3_16_soc.clock_limits[closest_clk_lvl].dppclk_mhz; + + s[i].dram_bw_per_chan_gbps = + dcn3_16_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps; + s[i].dscclk_mhz = dcn3_16_soc.clock_limits[closest_clk_lvl].dscclk_mhz; + s[i].dtbclk_mhz = dcn3_16_soc.clock_limits[closest_clk_lvl].dtbclk_mhz; + s[i].phyclk_d18_mhz = + dcn3_16_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz; + s[i].phyclk_mhz = dcn3_16_soc.clock_limits[closest_clk_lvl].phyclk_mhz; + } + if (clk_table->num_entries) { + dcn3_16_soc.num_states = clk_table->num_entries; } memcpy(dcn3_16_soc.clock_limits, s, sizeof(dcn3_16_soc.clock_limits)); @@ -805,13 +805,21 @@ void dcn316_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_param dcn3_16_soc.dram_clock_change_latency_us = dc->debug.dram_clock_change_latency_ns / 1000; } - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) - dml_init_instance(&dc->dml, &dcn3_16_soc, &dcn3_16_ip, DML_PROJECT_DCN31); - else - dml_init_instance(&dc->dml, &dcn3_16_soc, &dcn3_16_ip, DML_PROJECT_DCN31_FPGA); + dml_init_instance(&dc->dml, &dcn3_16_soc, &dcn3_16_ip, DML_PROJECT_DCN31); } int dcn_get_max_non_odm_pix_rate_100hz(struct _vcs_dpi_soc_bounding_box_st *soc) { return soc->clock_limits[0].dispclk_mhz * 10000.0 / (1.0 + soc->dcn_downspread_percent / 100.0); } + +int dcn_get_approx_det_segs_required_for_pstate( + struct _vcs_dpi_soc_bounding_box_st *soc, + int pix_clk_100hz, int bpp, int seg_size_kb) +{ + /* Roughly calculate required crb to hide latency. In practice there is slightly + * more buffer available for latency hiding + */ + return (int)(soc->dram_clock_change_latency_us * pix_clk_100hz * bpp + / 10240000 + seg_size_kb - 1) / seg_size_kb; +} diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.h index 687d3522cc33..8f9c8faed260 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.h +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.h @@ -47,6 +47,9 @@ void dcn31_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params void dcn315_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params); void dcn316_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params); int dcn_get_max_non_odm_pix_rate_100hz(struct _vcs_dpi_soc_bounding_box_st *soc); +int dcn_get_approx_det_segs_required_for_pstate( + struct _vcs_dpi_soc_bounding_box_st *soc, + int pix_clk_100hz, int bpp, int seg_size_kb); int dcn31x_populate_dml_pipes_from_context(struct dc *dc, struct dc_state *context, diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c index bd674dc30df3..43016c462251 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c @@ -532,7 +532,8 @@ static void CalculateStutterEfficiency( static void CalculateSwathAndDETConfiguration( bool ForceSingleDPP, int NumberOfActivePlanes, - unsigned int DETBufferSizeInKByte, + bool DETSharedByAllDPP, + unsigned int DETBufferSizeInKByte[], double MaximumSwathWidthLuma[], double MaximumSwathWidthChroma[], enum scan_direction_class SourceScan[], @@ -3118,7 +3119,7 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman v->SurfaceWidthC[k], v->SurfaceHeightY[k], v->SurfaceHeightC[k], - v->DETBufferSizeInKByte[0] * 1024, + v->DETBufferSizeInKByte[k] * 1024, v->BlockHeight256BytesY[k], v->BlockHeight256BytesC[k], v->SurfaceTiling[k], @@ -3313,7 +3314,8 @@ static void DisplayPipeConfiguration(struct display_mode_lib *mode_lib) CalculateSwathAndDETConfiguration( false, v->NumberOfActivePlanes, - v->DETBufferSizeInKByte[0], + mode_lib->project == DML_PROJECT_DCN315 && v->DETSizeOverride[0], + v->DETBufferSizeInKByte, dummy1, dummy2, v->SourceScan, @@ -3779,14 +3781,16 @@ static noinline void CalculatePrefetchSchedulePerPlane( &v->VReadyOffsetPix[k]); } -static void PatchDETBufferSizeInKByte(unsigned int NumberOfActivePlanes, int NoOfDPPThisState[], unsigned int config_return_buffer_size_in_kbytes, unsigned int *DETBufferSizeInKByte) +static void PatchDETBufferSizeInKByte(unsigned int NumberOfActivePlanes, int NoOfDPPThisState[], unsigned int config_return_buffer_size_in_kbytes, unsigned int DETBufferSizeInKByte[]) { int i, total_pipes = 0; for (i = 0; i < NumberOfActivePlanes; i++) total_pipes += NoOfDPPThisState[i]; - *DETBufferSizeInKByte = ((config_return_buffer_size_in_kbytes - DCN3_15_MIN_COMPBUF_SIZE_KB) / 64 / total_pipes) * 64; - if (*DETBufferSizeInKByte > DCN3_15_MAX_DET_SIZE) - *DETBufferSizeInKByte = DCN3_15_MAX_DET_SIZE; + DETBufferSizeInKByte[0] = ((config_return_buffer_size_in_kbytes - DCN3_15_MIN_COMPBUF_SIZE_KB) / 64 / total_pipes) * 64; + if (DETBufferSizeInKByte[0] > DCN3_15_MAX_DET_SIZE) + DETBufferSizeInKByte[0] = DCN3_15_MAX_DET_SIZE; + for (i = 1; i < NumberOfActivePlanes; i++) + DETBufferSizeInKByte[i] = DETBufferSizeInKByte[0]; } @@ -4026,7 +4030,8 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l CalculateSwathAndDETConfiguration( true, v->NumberOfActivePlanes, - v->DETBufferSizeInKByte[0], + mode_lib->project == DML_PROJECT_DCN315 && v->DETSizeOverride[0], + v->DETBufferSizeInKByte, v->MaximumSwathWidthLuma, v->MaximumSwathWidthChroma, v->SourceScan, @@ -4166,6 +4171,10 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l || (v->PlaneRequiredDISPCLK > v->MaxDispclkRoundedDownToDFSGranularity)) { v->DISPCLK_DPPCLK_Support[i][j] = false; } + if (mode_lib->project == DML_PROJECT_DCN315 && v->DETSizeOverride[k] > DCN3_15_MAX_DET_SIZE && v->NoOfDPP[i][j][k] < 2) { + v->MPCCombine[i][j][k] = true; + v->NoOfDPP[i][j][k] = 2; + } } v->TotalNumberOfActiveDPP[i][j] = 0; v->TotalNumberOfSingleDPPPlanes[i][j] = 0; @@ -4642,12 +4651,13 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l v->ODMCombineEnableThisState[k] = v->ODMCombineEnablePerState[i][k]; } - if (v->NumberOfActivePlanes > 1 && mode_lib->project == DML_PROJECT_DCN315) - PatchDETBufferSizeInKByte(v->NumberOfActivePlanes, v->NoOfDPPThisState, v->ip.config_return_buffer_size_in_kbytes, &v->DETBufferSizeInKByte[0]); + if (v->NumberOfActivePlanes > 1 && mode_lib->project == DML_PROJECT_DCN315 && !v->DETSizeOverride[0]) + PatchDETBufferSizeInKByte(v->NumberOfActivePlanes, v->NoOfDPPThisState, v->ip.config_return_buffer_size_in_kbytes, v->DETBufferSizeInKByte); CalculateSwathAndDETConfiguration( false, v->NumberOfActivePlanes, - v->DETBufferSizeInKByte[0], + mode_lib->project == DML_PROJECT_DCN315 && v->DETSizeOverride[0], + v->DETBufferSizeInKByte, v->MaximumSwathWidthLuma, v->MaximumSwathWidthChroma, v->SourceScan, @@ -5274,8 +5284,8 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l v->TotImmediateFlipBytes = 0.0; for (k = 0; k < v->NumberOfActivePlanes; k++) { v->TotImmediateFlipBytes = v->TotImmediateFlipBytes - + v->NoOfDPP[i][j][k] * v->PDEAndMetaPTEBytesPerFrame[i][j][k] + v->MetaRowBytes[i][j][k] - + v->DPTEBytesPerRow[i][j][k]; + + v->NoOfDPP[i][j][k] * (v->PDEAndMetaPTEBytesPerFrame[i][j][k] + v->MetaRowBytes[i][j][k] + + v->DPTEBytesPerRow[i][j][k]); } for (k = 0; k < v->NumberOfActivePlanes; k++) { @@ -6611,7 +6621,8 @@ static void CalculateStutterEfficiency( static void CalculateSwathAndDETConfiguration( bool ForceSingleDPP, int NumberOfActivePlanes, - unsigned int DETBufferSizeInKByte, + bool DETSharedByAllDPP, + unsigned int DETBufferSizeInKByteA[], double MaximumSwathWidthLuma[], double MaximumSwathWidthChroma[], enum scan_direction_class SourceScan[], @@ -6695,6 +6706,10 @@ static void CalculateSwathAndDETConfiguration( *ViewportSizeSupport = true; for (k = 0; k < NumberOfActivePlanes; ++k) { + unsigned int DETBufferSizeInKByte = DETBufferSizeInKByteA[k]; + + if (DETSharedByAllDPP && DPPPerPlane[k]) + DETBufferSizeInKByte /= DPPPerPlane[k]; if ((SourcePixelFormat[k] == dm_444_64 || SourcePixelFormat[k] == dm_444_32 || SourcePixelFormat[k] == dm_444_16 || SourcePixelFormat[k] == dm_mono_16 || SourcePixelFormat[k] == dm_mono_8 || SourcePixelFormat[k] == dm_rgbe)) { if (SurfaceTiling[k] == dm_sw_linear @@ -7017,7 +7032,7 @@ static double CalculateUrgentLatency( return ret; } -static void UseMinimumDCFCLK( +static noinline_for_stack void UseMinimumDCFCLK( struct display_mode_lib *mode_lib, int MaxPrefetchMode, int ReorderingBytes) diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_rq_dlg_calc_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_rq_dlg_calc_31.c index 2244e4fb8c96..4113ce79c4af 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_rq_dlg_calc_31.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_rq_dlg_calc_31.c @@ -987,8 +987,7 @@ static void dml_rq_dlg_get_dlg_params( dlg_vblank_start = interlaced ? (vblank_start / 2) : vblank_start; disp_dlg_regs->min_dst_y_next_start = (unsigned int) (((double) dlg_vblank_start) * dml_pow(2, 2)); - disp_dlg_regs->optimized_min_dst_y_next_start_us = 0; - disp_dlg_regs->optimized_min_dst_y_next_start = disp_dlg_regs->min_dst_y_next_start; + disp_dlg_regs->min_dst_y_next_start_us = 0; ASSERT(disp_dlg_regs->min_dst_y_next_start < (unsigned int)dml_pow(2, 18)); dml_print("DML_DLG: %s: min_ttu_vblank (us) = %3.2f\n", __func__, min_ttu_vblank); @@ -1433,14 +1432,6 @@ static void dml_rq_dlg_get_dlg_params( dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_flip = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_flip); dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_flip = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_flip); - // hack for FPGA - if (mode_lib->project == DML_PROJECT_DCN31_FPGA) { - if (disp_dlg_regs->vratio_prefetch >= (unsigned int) dml_pow(2, 22)) { - disp_dlg_regs->vratio_prefetch = (unsigned int) dml_pow(2, 22) - 1; - dml_print("vratio_prefetch exceed the max value, the register field is [21:0]\n"); - } - } - disp_dlg_regs->refcyc_per_pte_group_vblank_l = (unsigned int) (dst_y_per_row_vblank * (double) htotal * ref_freq_to_pix_freq / (double) dpte_groups_per_row_ub_l); ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_l < (unsigned int)dml_pow(2, 13)); diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.c index 9e54e3d0eb78..c9afddd11589 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.c @@ -33,7 +33,7 @@ #include "dml/display_mode_vba.h" struct _vcs_dpi_ip_params_st dcn3_14_ip = { - .VBlankNomDefaultUS = 668, + .VBlankNomDefaultUS = 800, .gpuvm_enable = 1, .gpuvm_max_page_table_levels = 1, .hostvm_enable = 1, @@ -190,8 +190,7 @@ void dcn314_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p dc_assert_fp_enabled(); // Default clock levels are used for diags, which may lead to overclocking. - if (!IS_DIAG_DC(dc->ctx->dce_environment) && dc->config.use_default_clock_table == false) { - + if (dc->config.use_default_clock_table == false) { dcn3_14_ip.max_num_otg = dc->res_pool->res_cap->num_timing_generator; dcn3_14_ip.max_num_dpp = dc->res_pool->pipe_count; @@ -266,11 +265,7 @@ void dcn314_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p } dcn20_patch_bounding_box(dc, &dcn3_14_soc); - - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) - dml_init_instance(&dc->dml, &dcn3_14_soc, &dcn3_14_ip, DML_PROJECT_DCN314); - else - dml_init_instance(&dc->dml, &dcn3_14_soc, &dcn3_14_ip, DML_PROJECT_DCN31_FPGA); + dml_init_instance(&dc->dml, &dcn3_14_soc, &dcn3_14_ip, DML_PROJECT_DCN314); } static bool is_dual_plane(enum surface_pixel_format format) @@ -286,6 +281,7 @@ int dcn314_populate_dml_pipes_from_context_fpu(struct dc *dc, struct dc_state *c struct resource_context *res_ctx = &context->res_ctx; struct pipe_ctx *pipe; bool upscaled = false; + const unsigned int max_allowed_vblank_nom = 1023; dc_assert_fp_enabled(); @@ -299,9 +295,11 @@ int dcn314_populate_dml_pipes_from_context_fpu(struct dc *dc, struct dc_state *c pipe = &res_ctx->pipe_ctx[i]; timing = &pipe->stream->timing; - if (dc_extended_blank_supported(dc) && pipe->stream->adjust.v_total_max == pipe->stream->adjust.v_total_min - && pipe->stream->adjust.v_total_min > timing->v_total) - pipes[pipe_cnt].pipe.dest.vtotal = pipe->stream->adjust.v_total_min; + pipes[pipe_cnt].pipe.dest.vtotal = pipe->stream->adjust.v_total_min; + pipes[pipe_cnt].pipe.dest.vblank_nom = timing->v_total - pipes[pipe_cnt].pipe.dest.vactive; + pipes[pipe_cnt].pipe.dest.vblank_nom = min(pipes[pipe_cnt].pipe.dest.vblank_nom, dcn3_14_ip.VBlankNomDefaultUS); + pipes[pipe_cnt].pipe.dest.vblank_nom = max(pipes[pipe_cnt].pipe.dest.vblank_nom, timing->v_sync_width); + pipes[pipe_cnt].pipe.dest.vblank_nom = min(pipes[pipe_cnt].pipe.dest.vblank_nom, max_allowed_vblank_nom); if (pipe->plane_state && (pipe->plane_state->src_rect.height < pipe->plane_state->dst_rect.height || @@ -323,8 +321,6 @@ int dcn314_populate_dml_pipes_from_context_fpu(struct dc *dc, struct dc_state *c pipes[pipe_cnt].pipe.src.dcc_fraction_of_zs_req_luma = 0; pipes[pipe_cnt].pipe.src.dcc_fraction_of_zs_req_chroma = 0; pipes[pipe_cnt].pipe.dest.vfront_porch = timing->v_front_porch; - pipes[pipe_cnt].pipe.dest.vblank_nom = - dcn3_14_ip.VBlankNomDefaultUS / (timing->h_total / (timing->pix_clk_100hz / 10000.0)); pipes[pipe_cnt].pipe.src.dcc_rate = 3; pipes[pipe_cnt].dout.dsc_input_bpc = 0; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c index 7eb2173b7691..9010c47476e9 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c @@ -5371,8 +5371,8 @@ void dml314_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_ v->TotImmediateFlipBytes = 0.0; for (k = 0; k < v->NumberOfActivePlanes; k++) { v->TotImmediateFlipBytes = v->TotImmediateFlipBytes - + v->NoOfDPP[i][j][k] * v->PDEAndMetaPTEBytesPerFrame[i][j][k] + v->MetaRowBytes[i][j][k] - + v->DPTEBytesPerRow[i][j][k]; + + v->NoOfDPP[i][j][k] * (v->PDEAndMetaPTEBytesPerFrame[i][j][k] + v->MetaRowBytes[i][j][k] + + v->DPTEBytesPerRow[i][j][k]); } for (k = 0; k < v->NumberOfActivePlanes; k++) { @@ -5557,6 +5557,65 @@ void dml314_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_ } } } + for (i = v->soc.num_states; i >= 0; i--) { + for (j = 0; j < 2; j++) { + enum dm_validation_status status = DML_VALIDATION_OK; + + if (!v->ScaleRatioAndTapsSupport) { + status = DML_FAIL_SCALE_RATIO_TAP; + } else if (!v->SourceFormatPixelAndScanSupport) { + status = DML_FAIL_SOURCE_PIXEL_FORMAT; + } else if (!v->ViewportSizeSupport[i][j]) { + status = DML_FAIL_VIEWPORT_SIZE; + } else if (P2IWith420) { + status = DML_FAIL_P2I_WITH_420; + } else if (DSCOnlyIfNecessaryWithBPP) { + status = DML_FAIL_DSC_ONLY_IF_NECESSARY_WITH_BPP; + } else if (DSC422NativeNotSupported) { + status = DML_FAIL_NOT_DSC422_NATIVE; + } else if (!v->ODMCombine4To1SupportCheckOK[i]) { + status = DML_FAIL_ODM_COMBINE4TO1; + } else if (v->NotEnoughDSCUnits[i]) { + status = DML_FAIL_NOT_ENOUGH_DSC; + } else if (!v->ROBSupport[i][j]) { + status = DML_FAIL_REORDERING_BUFFER; + } else if (!v->DISPCLK_DPPCLK_Support[i][j]) { + status = DML_FAIL_DISPCLK_DPPCLK; + } else if (!v->TotalAvailablePipesSupport[i][j]) { + status = DML_FAIL_TOTAL_AVAILABLE_PIPES; + } else if (!EnoughWritebackUnits) { + status = DML_FAIL_ENOUGH_WRITEBACK_UNITS; + } else if (!v->WritebackLatencySupport) { + status = DML_FAIL_WRITEBACK_LATENCY; + } else if (!v->WritebackScaleRatioAndTapsSupport) { + status = DML_FAIL_WRITEBACK_SCALE_RATIO_TAP; + } else if (!v->CursorSupport) { + status = DML_FAIL_CURSOR_SUPPORT; + } else if (!v->PitchSupport) { + status = DML_FAIL_PITCH_SUPPORT; + } else if (ViewportExceedsSurface) { + status = DML_FAIL_VIEWPORT_EXCEEDS_SURFACE; + } else if (!v->PrefetchSupported[i][j]) { + status = DML_FAIL_PREFETCH_SUPPORT; + } else if (!v->DynamicMetadataSupported[i][j]) { + status = DML_FAIL_DYNAMIC_METADATA; + } else if (!v->TotalVerticalActiveBandwidthSupport[i][j]) { + status = DML_FAIL_TOTAL_V_ACTIVE_BW; + } else if (!v->VRatioInPrefetchSupported[i][j]) { + status = DML_FAIL_V_RATIO_PREFETCH; + } else if (!v->PTEBufferSizeNotExceeded[i][j]) { + status = DML_FAIL_PTE_BUFFER_SIZE; + } else if (v->NonsupportedDSCInputBPC) { + status = DML_FAIL_DSC_INPUT_BPC; + } else if ((v->HostVMEnable + && !v->ImmediateFlipSupportedForState[i][j])) { + status = DML_FAIL_HOST_VM_IMMEDIATE_FLIP; + } else if (FMTBufferExceeded) { + status = DML_FAIL_FMT_BUFFER_EXCEEDED; + } + mode_lib->vba.ValidationStatus[i] = status; + } + } { unsigned int MaximumMPCCombine = 0; @@ -7061,7 +7120,7 @@ static double CalculateUrgentLatency( return ret; } -static void UseMinimumDCFCLK( +static noinline_for_stack void UseMinimumDCFCLK( struct display_mode_lib *mode_lib, int MaxPrefetchMode, int ReorderingBytes) diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_rq_dlg_calc_314.c b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_rq_dlg_calc_314.c index ea4eb66066c4..b3e8dc08030c 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_rq_dlg_calc_314.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_rq_dlg_calc_314.c @@ -951,7 +951,6 @@ static void dml_rq_dlg_get_dlg_params( { const display_pipe_source_params_st *src = &e2e_pipe_param[pipe_idx].pipe.src; const display_pipe_dest_params_st *dst = &e2e_pipe_param[pipe_idx].pipe.dest; - const display_output_params_st *dout = &e2e_pipe_param[pipe_idx].dout; const display_clocks_and_cfg_st *clks = &e2e_pipe_param[pipe_idx].clks_cfg; const scaler_ratio_depth_st *scl = &e2e_pipe_param[pipe_idx].pipe.scale_ratio_depth; const scaler_taps_st *taps = &e2e_pipe_param[pipe_idx].pipe.scale_taps; @@ -1000,8 +999,6 @@ static void dml_rq_dlg_get_dlg_params( unsigned int vupdate_width; unsigned int vready_offset; - unsigned int dispclk_delay_subtotal; - unsigned int vstartup_start; unsigned int dst_x_after_scaler; unsigned int dst_y_after_scaler; @@ -1051,7 +1048,6 @@ static void dml_rq_dlg_get_dlg_params( float vba__refcyc_per_req_delivery_pre_l = get_refcyc_per_req_delivery_pre_l_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz; // From VBA float vba__refcyc_per_req_delivery_l = get_refcyc_per_req_delivery_l_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz; // From VBA - int blank_lines = 0; memset(disp_dlg_regs, 0, sizeof(*disp_dlg_regs)); memset(disp_ttu_regs, 0, sizeof(*disp_ttu_regs)); @@ -1075,17 +1071,10 @@ static void dml_rq_dlg_get_dlg_params( min_ttu_vblank = get_min_ttu_vblank_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); // From VBA dlg_vblank_start = interlaced ? (vblank_start / 2) : vblank_start; - disp_dlg_regs->optimized_min_dst_y_next_start = disp_dlg_regs->min_dst_y_next_start; - disp_dlg_regs->optimized_min_dst_y_next_start_us = 0; - disp_dlg_regs->min_dst_y_next_start = (unsigned int) (((double) dlg_vblank_start) * dml_pow(2, 2)); - blank_lines = (dst->vblank_end + dst->vtotal_min - dst->vblank_start - dst->vstartup_start - 1); - if (blank_lines < 0) - blank_lines = 0; - if (blank_lines != 0) { - disp_dlg_regs->optimized_min_dst_y_next_start = vba__min_dst_y_next_start; - disp_dlg_regs->optimized_min_dst_y_next_start_us = (disp_dlg_regs->optimized_min_dst_y_next_start * dst->hactive) / (unsigned int) dst->pixel_rate_mhz; - disp_dlg_regs->min_dst_y_next_start = disp_dlg_regs->optimized_min_dst_y_next_start; - } + disp_dlg_regs->min_dst_y_next_start_us = + (vba__min_dst_y_next_start * dst->hactive) / (unsigned int) dst->pixel_rate_mhz; + disp_dlg_regs->min_dst_y_next_start = vba__min_dst_y_next_start * dml_pow(2, 2); + ASSERT(disp_dlg_regs->min_dst_y_next_start < (unsigned int)dml_pow(2, 18)); dml_print("DML_DLG: %s: min_ttu_vblank (us) = %3.2f\n", __func__, min_ttu_vblank); @@ -1127,13 +1116,6 @@ static void dml_rq_dlg_get_dlg_params( vupdate_offset = dst->vupdate_offset; vupdate_width = dst->vupdate_width; vready_offset = dst->vready_offset; - dispclk_delay_subtotal = mode_lib->ip.dispclk_delay_subtotal; - - if (dout->dsc_enable) { - double dsc_delay = get_dsc_delay(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); // FROM VBA - - dispclk_delay_subtotal += dsc_delay; - } vstartup_start = dst->vstartup_start; if (interlaced) { @@ -1538,14 +1520,6 @@ static void dml_rq_dlg_get_dlg_params( dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_flip = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_flip); dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_flip = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_flip); - // hack for FPGA - if (mode_lib->project == DML_PROJECT_DCN31_FPGA) { - if (disp_dlg_regs->vratio_prefetch >= (unsigned int) dml_pow(2, 22)) { - disp_dlg_regs->vratio_prefetch = (unsigned int) dml_pow(2, 22) - 1; - dml_print("vratio_prefetch exceed the max value, the register field is [21:0]\n"); - } - } - disp_dlg_regs->refcyc_per_pte_group_vblank_l = (unsigned int) (dst_y_per_row_vblank * (double) htotal * ref_freq_to_pix_freq / (double) dpte_groups_per_row_ub_l); ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_l < (unsigned int)dml_pow(2, 13)); diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c index 47beb4ea779d..e2bb2b9971f3 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c @@ -35,6 +35,15 @@ #define DC_LOGGER_INIT(logger) +static const struct subvp_high_refresh_list subvp_high_refresh_list = { + .min_refresh = 120, + .max_refresh = 165, + .res = { + {.width = 3840, .height = 2160, }, + {.width = 3440, .height = 1440, }, + {.width = 2560, .height = 1440, }}, +}; + struct _vcs_dpi_ip_params_st dcn3_2_ip = { .gpuvm_enable = 0, .gpuvm_max_page_table_levels = 4, @@ -138,7 +147,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_2_soc = { .urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096, .urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096, .urgent_out_of_order_return_per_channel_vm_only_bytes = 4096, - .pct_ideal_sdp_bw_after_urgent = 100.0, + .pct_ideal_sdp_bw_after_urgent = 90.0, .pct_ideal_fabric_bw_after_urgent = 67.0, .pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 20.0, .pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0, // N/A, for now keep as is until DML implemented @@ -670,7 +679,6 @@ static bool dcn32_assign_subvp_pipe(struct dc *dc, unsigned int max_frame_time = 0; bool valid_assignment_found = false; unsigned int free_pipes = dcn32_get_num_free_pipes(dc, context); - bool current_assignment_freesync = false; struct vba_vars_st *vba = &context->bw_ctx.dml.vba; for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) { @@ -692,8 +700,12 @@ static bool dcn32_assign_subvp_pipe(struct dc *dc, * to combine this with SubVP can cause issues with the scheduling). * - Not TMZ surface */ - if (pipe->plane_state && !pipe->top_pipe && !dcn32_is_center_timing(pipe) && !dcn32_is_psr_capable(pipe) && - pipe->stream->mall_stream_config.type == SUBVP_NONE && refresh_rate < 120 && !pipe->plane_state->address.tmz_surface && + if (pipe->plane_state && !pipe->top_pipe && !dcn32_is_center_timing(pipe) && + !(pipe->stream->timing.pix_clk_100hz / 10000 > DCN3_2_MAX_SUBVP_PIXEL_RATE_MHZ) && + (!dcn32_is_psr_capable(pipe) || (context->stream_count == 1 && dc->caps.dmub_caps.subvp_psr)) && + pipe->stream->mall_stream_config.type == SUBVP_NONE && + (refresh_rate < 120 || dcn32_allow_subvp_high_refresh_rate(dc, context, pipe)) && + !pipe->plane_state->address.tmz_surface && (vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] <= 0 || (vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] > 0 && dcn32_allow_subvp_with_active_margin(pipe)))) { @@ -707,19 +719,10 @@ static bool dcn32_assign_subvp_pipe(struct dc *dc, struct dc_stream_state *stream = pipe->stream; unsigned int frame_us = (stream->timing.v_total * stream->timing.h_total / (double)(stream->timing.pix_clk_100hz * 100)) * 1000000; - if (frame_us > max_frame_time && !stream->ignore_msa_timing_param) { + if (frame_us > max_frame_time) { *index = i; max_frame_time = frame_us; valid_assignment_found = true; - current_assignment_freesync = false; - /* For the 2-Freesync display case, still choose the one with the - * longest frame time - */ - } else if (stream->ignore_msa_timing_param && (!valid_assignment_found || - (current_assignment_freesync && frame_us > max_frame_time))) { - *index = i; - valid_assignment_found = true; - current_assignment_freesync = true; } } } @@ -851,10 +854,9 @@ static bool subvp_subvp_schedulable(struct dc *dc, struct dc_state *context) } /** - * subvp_drr_schedulable - Determine if SubVP + DRR config is schedulable + * subvp_drr_schedulable() - Determine if SubVP + DRR config is schedulable * @dc: current dc state * @context: new dc state - * @drr_pipe: DRR pipe_ctx for the SubVP + DRR config * * High level algorithm: * 1. Get timing for SubVP pipe, phantom pipe, and DRR pipe @@ -865,11 +867,12 @@ static bool subvp_subvp_schedulable(struct dc *dc, struct dc_state *context) * * Return: True if the SubVP + DRR config is schedulable, false otherwise */ -static bool subvp_drr_schedulable(struct dc *dc, struct dc_state *context, struct pipe_ctx *drr_pipe) +static bool subvp_drr_schedulable(struct dc *dc, struct dc_state *context) { bool schedulable = false; uint32_t i; struct pipe_ctx *pipe = NULL; + struct pipe_ctx *drr_pipe = NULL; struct dc_crtc_timing *main_timing = NULL; struct dc_crtc_timing *phantom_timing = NULL; struct dc_crtc_timing *drr_timing = NULL; @@ -880,10 +883,6 @@ static bool subvp_drr_schedulable(struct dc *dc, struct dc_state *context, struc int16_t stretched_drr_us = 0; int16_t drr_stretched_vblank_us = 0; int16_t max_vblank_mallregion = 0; - const struct dc_config *config = &dc->config; - - if (config->disable_subvp_drr) - return false; // Find SubVP pipe for (i = 0; i < dc->res_pool->pipe_count; i++) { @@ -899,6 +898,19 @@ static bool subvp_drr_schedulable(struct dc *dc, struct dc_state *context, struc break; } + // Find the DRR pipe + for (i = 0; i < dc->res_pool->pipe_count; i++) { + drr_pipe = &context->res_ctx.pipe_ctx[i]; + + // We check for master pipe only + if (!drr_pipe->stream || !drr_pipe->plane_state || drr_pipe->top_pipe || drr_pipe->prev_odm_pipe) + continue; + + if (drr_pipe->stream->mall_stream_config.type == SUBVP_NONE && drr_pipe->stream->ignore_msa_timing_param && + (drr_pipe->stream->allow_freesync || drr_pipe->stream->vrr_active_variable)) + break; + } + main_timing = &pipe->stream->timing; phantom_timing = &pipe->stream->mall_stream_config.paired_stream->timing; drr_timing = &drr_pipe->stream->timing; @@ -984,13 +996,7 @@ static bool subvp_vblank_schedulable(struct dc *dc, struct dc_state *context) if (!subvp_pipe && pipe->stream->mall_stream_config.type == SUBVP_MAIN) subvp_pipe = pipe; } - // Use ignore_msa_timing_param and VRR active, or Freesync flag to identify as DRR On - if (found && context->res_ctx.pipe_ctx[vblank_index].stream->ignore_msa_timing_param && - (context->res_ctx.pipe_ctx[vblank_index].stream->allow_freesync || - context->res_ctx.pipe_ctx[vblank_index].stream->vrr_active_variable)) { - // SUBVP + DRR case -- only allowed if run through DRR validation path - schedulable = false; - } else if (found) { + if (found) { main_timing = &subvp_pipe->stream->timing; phantom_timing = &subvp_pipe->stream->mall_stream_config.paired_stream->timing; vblank_timing = &context->res_ctx.pipe_ctx[vblank_index].stream->timing; @@ -1020,6 +1026,53 @@ static bool subvp_vblank_schedulable(struct dc *dc, struct dc_state *context) } /** + * subvp_subvp_admissable() - Determine if subvp + subvp config is admissible + * + * @dc: Current DC state + * @context: New DC state to be programmed + * + * SubVP + SubVP is admissible under the following conditions: + * - All SubVP pipes are < 120Hz OR + * - All SubVP pipes are >= 120hz + * + * Return: True if admissible, false otherwise + */ +static bool subvp_subvp_admissable(struct dc *dc, + struct dc_state *context) +{ + bool result = false; + uint32_t i; + uint8_t subvp_count = 0; + uint32_t min_refresh = subvp_high_refresh_list.min_refresh, max_refresh = 0; + uint32_t refresh_rate = 0; + + for (i = 0; i < dc->res_pool->pipe_count; i++) { + struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; + + if (!pipe->stream) + continue; + + if (pipe->plane_state && !pipe->top_pipe && + pipe->stream->mall_stream_config.type == SUBVP_MAIN) { + refresh_rate = (pipe->stream->timing.pix_clk_100hz * 100 + + pipe->stream->timing.v_total * pipe->stream->timing.h_total - 1) + / (double)(pipe->stream->timing.v_total * pipe->stream->timing.h_total); + if (refresh_rate < min_refresh) + min_refresh = refresh_rate; + if (refresh_rate > max_refresh) + max_refresh = refresh_rate; + subvp_count++; + } + } + + if (subvp_count == 2 && ((min_refresh < 120 && max_refresh < 120) || + (min_refresh >= 120 && max_refresh >= 120))) + result = true; + + return result; +} + +/** * subvp_validate_static_schedulability - Check which SubVP case is calculated * and handle static analysis based on the case. * @dc: current dc state @@ -1037,11 +1090,12 @@ static bool subvp_validate_static_schedulability(struct dc *dc, struct dc_state *context, int vlevel) { - bool schedulable = true; // true by default for single display case + bool schedulable = false; struct vba_vars_st *vba = &context->bw_ctx.dml.vba; uint32_t i, pipe_idx; uint8_t subvp_count = 0; uint8_t vactive_count = 0; + uint8_t non_subvp_pipes = 0; for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) { struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; @@ -1049,14 +1103,18 @@ static bool subvp_validate_static_schedulability(struct dc *dc, if (!pipe->stream) continue; - if (pipe->plane_state && !pipe->top_pipe && - pipe->stream->mall_stream_config.type == SUBVP_MAIN) - subvp_count++; + if (pipe->plane_state && !pipe->top_pipe) { + if (pipe->stream->mall_stream_config.type == SUBVP_MAIN) + subvp_count++; + if (pipe->stream->mall_stream_config.type == SUBVP_NONE) { + non_subvp_pipes++; + } + } // Count how many planes that aren't SubVP/phantom are capable of VACTIVE // switching (SubVP + VACTIVE unsupported). In situations where we force // SubVP for a VACTIVE plane, we don't want to increment the vactive_count. - if (vba->ActiveDRAMClockChangeLatencyMargin[vba->pipe_plane[pipe_idx]] > 0 && + if (vba->ActiveDRAMClockChangeLatencyMarginPerState[vlevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] > 0 && pipe->stream->mall_stream_config.type == SUBVP_NONE) { vactive_count++; } @@ -1065,13 +1123,14 @@ static bool subvp_validate_static_schedulability(struct dc *dc, if (subvp_count == 2) { // Static schedulability check for SubVP + SubVP case - schedulable = subvp_subvp_schedulable(dc, context); - } else if (vba->DRAMClockChangeSupport[vlevel][vba->maxMpcComb] == dm_dram_clock_change_vblank_w_mall_sub_vp) { - // Static schedulability check for SubVP + VBLANK case. Also handle the case where - // DML outputs SubVP + VBLANK + VACTIVE (DML will report as SubVP + VBLANK) - if (vactive_count > 0) - schedulable = false; - else + schedulable = subvp_subvp_admissable(dc, context) && subvp_subvp_schedulable(dc, context); + } else if (subvp_count == 1 && non_subvp_pipes == 0) { + // Single SubVP configs will be supported by default as long as it's suppported by DML + schedulable = true; + } else if (subvp_count == 1 && non_subvp_pipes == 1) { + if (dcn32_subvp_drr_admissable(dc, context)) + schedulable = subvp_drr_schedulable(dc, context); + else if (dcn32_subvp_vblank_admissable(dc, context, vlevel)) schedulable = subvp_vblank_schedulable(dc, context); } else if (vba->DRAMClockChangeSupport[vlevel][vba->maxMpcComb] == dm_dram_clock_change_vactive_w_mall_sub_vp && vactive_count > 0) { @@ -1095,10 +1154,6 @@ static void dcn32_full_validate_bw_helper(struct dc *dc, unsigned int dc_pipe_idx = 0; int i = 0; bool found_supported_config = false; - struct pipe_ctx *pipe = NULL; - uint32_t non_subvp_pipes = 0; - bool drr_pipe_found = false; - uint32_t drr_pipe_index = 0; dc_assert_fp_enabled(); @@ -1129,7 +1184,7 @@ static void dcn32_full_validate_bw_helper(struct dc *dc, * 4. Display configuration passes validation * 5. (Config doesn't support MCLK in VACTIVE/VBLANK || dc->debug.force_subvp_mclk_switch) */ - if (!dc->debug.force_disable_subvp && dcn32_all_pipes_have_stream_and_plane(dc, context) && + if (!dc->debug.force_disable_subvp && !dc->caps.dmub_caps.gecc_enable && dcn32_all_pipes_have_stream_and_plane(dc, context) && !dcn32_mpo_in_use(context) && !dcn32_any_surfaces_rotated(dc, context) && (*vlevel == context->bw_ctx.dml.soc.num_states || vba->DRAMClockChangeSupport[*vlevel][vba->maxMpcComb] == dm_dram_clock_change_unsupported || @@ -1188,31 +1243,12 @@ static void dcn32_full_validate_bw_helper(struct dc *dc, } } - if (*vlevel < context->bw_ctx.dml.soc.num_states && - vba->DRAMClockChangeSupport[*vlevel][vba->maxMpcComb] != dm_dram_clock_change_unsupported - && subvp_validate_static_schedulability(dc, context, *vlevel)) { + if (*vlevel < context->bw_ctx.dml.soc.num_states + && subvp_validate_static_schedulability(dc, context, *vlevel)) found_supported_config = true; - } else if (*vlevel < context->bw_ctx.dml.soc.num_states) { - /* Case where 1 SubVP is added, and DML reports MCLK unsupported or DRR is allowed. - * This handles the case for SubVP + DRR, where the DRR display does not support MCLK - * switch at it's native refresh rate / timing, or DRR is allowed for the non-subvp - * display. - */ - for (i = 0; i < dc->res_pool->pipe_count; i++) { - pipe = &context->res_ctx.pipe_ctx[i]; - if (pipe->stream && pipe->plane_state && !pipe->top_pipe && - pipe->stream->mall_stream_config.type == SUBVP_NONE) { - non_subvp_pipes++; - // Use ignore_msa_timing_param flag to identify as DRR - if (pipe->stream->ignore_msa_timing_param && pipe->stream->allow_freesync) { - drr_pipe_found = true; - drr_pipe_index = i; - } - } - } - // If there is only 1 remaining non SubVP pipe that is DRR, check static - // schedulability for SubVP + DRR. - if (non_subvp_pipes == 1 && drr_pipe_found) { + if (found_supported_config) { + // For SubVP + DRR cases, we can force the lowest vlevel that supports the mode + if (dcn32_subvp_drr_admissable(dc, context) && subvp_drr_schedulable(dc, context)) { /* find lowest vlevel that supports the config */ for (i = *vlevel; i >= 0; i--) { if (vba->ModeSupport[i][vba->maxMpcComb]) { @@ -1221,9 +1257,6 @@ static void dcn32_full_validate_bw_helper(struct dc *dc, break; } } - - found_supported_config = subvp_drr_schedulable(dc, context, - &context->res_ctx.pipe_ctx[drr_pipe_index]); } } } @@ -1315,6 +1348,7 @@ static void dcn32_calculate_dlg_params(struct dc *dc, struct dc_state *context, int i, pipe_idx, active_hubp_count = 0; bool usr_retraining_support = false; bool unbounded_req_enabled = false; + struct vba_vars_st *vba = &context->bw_ctx.dml.vba; dc_assert_fp_enabled(); @@ -1396,6 +1430,11 @@ static void dcn32_calculate_dlg_params(struct dc *dc, struct dc_state *context, context->res_ctx.pipe_ctx[i].surface_size_in_mall_bytes = get_surface_size_in_mall(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx); + if (vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] > 0) + context->res_ctx.pipe_ctx[i].has_vactive_margin = true; + else + context->res_ctx.pipe_ctx[i].has_vactive_margin = false; + /* MALL Allocation Sizes */ /* count from active, top pipes per plane only */ if (context->res_ctx.pipe_ctx[i].stream && context->res_ctx.pipe_ctx[i].plane_state && @@ -1432,6 +1471,7 @@ static void dcn32_calculate_dlg_params(struct dc *dc, struct dc_state *context, context->bw_ctx.bw.dcn.clk.dramclk_khz = 0; context->bw_ctx.bw.dcn.clk.fclk_khz = 0; context->bw_ctx.bw.dcn.clk.p_state_change_support = true; + context->bw_ctx.bw.dcn.clk.fclk_p_state_change_support = true; } /*save a original dppclock copy*/ context->bw_ctx.bw.dcn.clk.bw_dppclk_khz = context->bw_ctx.bw.dcn.clk.dppclk_khz; @@ -2005,6 +2045,7 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context, maxMpcComb = context->bw_ctx.dml.vba.maxMpcComb; dcfclk_from_fw_based_mclk_switching = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb]; pstate_en = true; + context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][maxMpcComb] = dm_dram_clock_change_vblank; } else { /* Restore FCLK latency and re-run validation to go back to original validation * output if we find that enabling FPO does not give us any benefit (i.e. lower @@ -2062,6 +2103,7 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context, * sr_enter_exit/sr_exit should be lower than used for DRAM (TBD after bringup or later, use as decided in Clk Mgr) */ + /* if (dcn3_2_soc.num_states > 2) { vlevel_temp = 0; dcfclk = dc->clk_mgr->bw_params->clk_table.entries[0].dcfclk_mhz; @@ -2088,6 +2130,7 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context, context->bw_ctx.bw.dcn.watermarks.d.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000; context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.fclk_pstate_change_ns = get_fclk_watermark(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000; context->bw_ctx.bw.dcn.watermarks.d.usr_retraining_ns = get_usr_retraining_watermark(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000; + */ /* Set C, for Dummy P-State: * All clocks min. @@ -2189,6 +2232,9 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context, context->bw_ctx.bw.dcn.watermarks.a.usr_retraining_ns = get_usr_retraining_watermark(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000; } + /* Make set D = set A since we do not optimized watermarks for MALL */ + context->bw_ctx.bw.dcn.watermarks.d = context->bw_ctx.bw.dcn.watermarks.a; + for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) { if (!context->res_ctx.pipe_ctx[i].stream) continue; @@ -2303,14 +2349,48 @@ void dcn32_patch_dpm_table(struct clk_bw_params *bw_params) bw_params->clk_table.entries[0].memclk_mhz = dcn3_2_soc.clock_limits[0].dram_speed_mts / 16; } -static int build_synthetic_soc_states(struct clk_bw_params *bw_params, +/* + * override_max_clk_values - Overwrite the max clock frequencies with the max DC mode timings + * Input: + * max_clk_limit - struct containing the desired clock timings + * Output: + * curr_clk_limit - struct containing the timings that need to be overwritten + * Return: 0 upon success, non-zero for failure + */ +static int override_max_clk_values(struct clk_limit_table_entry *max_clk_limit, + struct clk_limit_table_entry *curr_clk_limit) +{ + if (NULL == max_clk_limit || NULL == curr_clk_limit) + return -1; //invalid parameters + + //only overwrite if desired max clock frequency is initialized + if (max_clk_limit->dcfclk_mhz != 0) + curr_clk_limit->dcfclk_mhz = max_clk_limit->dcfclk_mhz; + + if (max_clk_limit->fclk_mhz != 0) + curr_clk_limit->fclk_mhz = max_clk_limit->fclk_mhz; + + if (max_clk_limit->memclk_mhz != 0) + curr_clk_limit->memclk_mhz = max_clk_limit->memclk_mhz; + + if (max_clk_limit->socclk_mhz != 0) + curr_clk_limit->socclk_mhz = max_clk_limit->socclk_mhz; + + if (max_clk_limit->dtbclk_mhz != 0) + curr_clk_limit->dtbclk_mhz = max_clk_limit->dtbclk_mhz; + + if (max_clk_limit->dispclk_mhz != 0) + curr_clk_limit->dispclk_mhz = max_clk_limit->dispclk_mhz; + + return 0; +} + +static int build_synthetic_soc_states(bool disable_dc_mode_overwrite, struct clk_bw_params *bw_params, struct _vcs_dpi_voltage_scaling_st *table, unsigned int *num_entries) { int i, j; struct _vcs_dpi_voltage_scaling_st entry = {0}; - - unsigned int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, - max_phyclk_mhz = 0, max_dtbclk_mhz = 0, max_fclk_mhz = 0, max_uclk_mhz = 0; + struct clk_limit_table_entry max_clk_data = {0}; unsigned int min_dcfclk_mhz = 199, min_fclk_mhz = 299; @@ -2321,51 +2401,76 @@ static int build_synthetic_soc_states(struct clk_bw_params *bw_params, unsigned int num_fclk_dpms = 0; unsigned int num_dcfclk_dpms = 0; - for (i = 0; i < MAX_NUM_DPM_LVL; i++) { - if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz) - max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz; - if (bw_params->clk_table.entries[i].fclk_mhz > max_fclk_mhz) - max_fclk_mhz = bw_params->clk_table.entries[i].fclk_mhz; - if (bw_params->clk_table.entries[i].memclk_mhz > max_uclk_mhz) - max_uclk_mhz = bw_params->clk_table.entries[i].memclk_mhz; - if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz) - max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz; - if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz) - max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz; - if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz) - max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz; - if (bw_params->clk_table.entries[i].dtbclk_mhz > max_dtbclk_mhz) - max_dtbclk_mhz = bw_params->clk_table.entries[i].dtbclk_mhz; + unsigned int num_dc_uclk_dpms = 0; + unsigned int num_dc_fclk_dpms = 0; + unsigned int num_dc_dcfclk_dpms = 0; - if (bw_params->clk_table.entries[i].memclk_mhz > 0) + for (i = 0; i < MAX_NUM_DPM_LVL; i++) { + if (bw_params->clk_table.entries[i].dcfclk_mhz > max_clk_data.dcfclk_mhz) + max_clk_data.dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz; + if (bw_params->clk_table.entries[i].fclk_mhz > max_clk_data.fclk_mhz) + max_clk_data.fclk_mhz = bw_params->clk_table.entries[i].fclk_mhz; + if (bw_params->clk_table.entries[i].memclk_mhz > max_clk_data.memclk_mhz) + max_clk_data.memclk_mhz = bw_params->clk_table.entries[i].memclk_mhz; + if (bw_params->clk_table.entries[i].dispclk_mhz > max_clk_data.dispclk_mhz) + max_clk_data.dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz; + if (bw_params->clk_table.entries[i].dppclk_mhz > max_clk_data.dppclk_mhz) + max_clk_data.dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz; + if (bw_params->clk_table.entries[i].phyclk_mhz > max_clk_data.phyclk_mhz) + max_clk_data.phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz; + if (bw_params->clk_table.entries[i].dtbclk_mhz > max_clk_data.dtbclk_mhz) + max_clk_data.dtbclk_mhz = bw_params->clk_table.entries[i].dtbclk_mhz; + + if (bw_params->clk_table.entries[i].memclk_mhz > 0) { num_uclk_dpms++; - if (bw_params->clk_table.entries[i].fclk_mhz > 0) + if (bw_params->clk_table.entries[i].memclk_mhz <= bw_params->dc_mode_limit.memclk_mhz) + num_dc_uclk_dpms++; + } + if (bw_params->clk_table.entries[i].fclk_mhz > 0) { num_fclk_dpms++; - if (bw_params->clk_table.entries[i].dcfclk_mhz > 0) + if (bw_params->clk_table.entries[i].fclk_mhz <= bw_params->dc_mode_limit.fclk_mhz) + num_dc_fclk_dpms++; + } + if (bw_params->clk_table.entries[i].dcfclk_mhz > 0) { num_dcfclk_dpms++; + if (bw_params->clk_table.entries[i].dcfclk_mhz <= bw_params->dc_mode_limit.dcfclk_mhz) + num_dc_dcfclk_dpms++; + } + } + + if (!disable_dc_mode_overwrite) { + //Overwrite max frequencies with max DC mode frequencies for DC mode systems + override_max_clk_values(&bw_params->dc_mode_limit, &max_clk_data); + num_uclk_dpms = num_dc_uclk_dpms; + num_fclk_dpms = num_dc_fclk_dpms; + num_dcfclk_dpms = num_dc_dcfclk_dpms; + bw_params->clk_table.num_entries_per_clk.num_memclk_levels = num_uclk_dpms; + bw_params->clk_table.num_entries_per_clk.num_fclk_levels = num_fclk_dpms; } if (num_dcfclk_dpms > 0 && bw_params->clk_table.entries[0].fclk_mhz > min_fclk_mhz) min_fclk_mhz = bw_params->clk_table.entries[0].fclk_mhz; - if (!max_dcfclk_mhz || !max_dispclk_mhz || !max_dtbclk_mhz) + if (!max_clk_data.dcfclk_mhz || !max_clk_data.dispclk_mhz || !max_clk_data.dtbclk_mhz) return -1; - if (max_dppclk_mhz == 0) - max_dppclk_mhz = max_dispclk_mhz; + if (max_clk_data.dppclk_mhz == 0) + max_clk_data.dppclk_mhz = max_clk_data.dispclk_mhz; - if (max_fclk_mhz == 0) - max_fclk_mhz = max_dcfclk_mhz * dcn3_2_soc.pct_ideal_sdp_bw_after_urgent / dcn3_2_soc.pct_ideal_fabric_bw_after_urgent; + if (max_clk_data.fclk_mhz == 0) + max_clk_data.fclk_mhz = max_clk_data.dcfclk_mhz * + dcn3_2_soc.pct_ideal_sdp_bw_after_urgent / + dcn3_2_soc.pct_ideal_fabric_bw_after_urgent; - if (max_phyclk_mhz == 0) - max_phyclk_mhz = dcn3_2_soc.clock_limits[0].phyclk_mhz; + if (max_clk_data.phyclk_mhz == 0) + max_clk_data.phyclk_mhz = dcn3_2_soc.clock_limits[0].phyclk_mhz; *num_entries = 0; - entry.dispclk_mhz = max_dispclk_mhz; - entry.dscclk_mhz = max_dispclk_mhz / 3; - entry.dppclk_mhz = max_dppclk_mhz; - entry.dtbclk_mhz = max_dtbclk_mhz; - entry.phyclk_mhz = max_phyclk_mhz; + entry.dispclk_mhz = max_clk_data.dispclk_mhz; + entry.dscclk_mhz = max_clk_data.dispclk_mhz / 3; + entry.dppclk_mhz = max_clk_data.dppclk_mhz; + entry.dtbclk_mhz = max_clk_data.dtbclk_mhz; + entry.phyclk_mhz = max_clk_data.phyclk_mhz; entry.phyclk_d18_mhz = dcn3_2_soc.clock_limits[0].phyclk_d18_mhz; entry.phyclk_d32_mhz = dcn3_2_soc.clock_limits[0].phyclk_d32_mhz; @@ -2379,7 +2484,7 @@ static int build_synthetic_soc_states(struct clk_bw_params *bw_params, } // Insert the max DCFCLK - entry.dcfclk_mhz = max_dcfclk_mhz; + entry.dcfclk_mhz = max_clk_data.dcfclk_mhz; entry.fabricclk_mhz = 0; entry.dram_speed_mts = 0; @@ -2407,7 +2512,7 @@ static int build_synthetic_soc_states(struct clk_bw_params *bw_params, // If FCLK fine grained, only insert max else { entry.dcfclk_mhz = 0; - entry.fabricclk_mhz = max_fclk_mhz; + entry.fabricclk_mhz = max_clk_data.fclk_mhz; entry.dram_speed_mts = 0; insert_entry_into_table_sorted(table, num_entries, &entry); @@ -2419,9 +2524,9 @@ static int build_synthetic_soc_states(struct clk_bw_params *bw_params, // Remove states that require higher clocks than are supported for (i = *num_entries - 1; i >= 0 ; i--) { - if (table[i].dcfclk_mhz > max_dcfclk_mhz || - table[i].fabricclk_mhz > max_fclk_mhz || - table[i].dram_speed_mts > max_uclk_mhz * 16) + if (table[i].dcfclk_mhz > max_clk_data.dcfclk_mhz || + table[i].fabricclk_mhz > max_clk_data.fclk_mhz || + table[i].dram_speed_mts > max_clk_data.memclk_mhz * 16) remove_entry_from_table_at_index(table, num_entries, i); } @@ -2508,80 +2613,78 @@ void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_pa { dc_assert_fp_enabled(); - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - /* Overrides from dc->config options */ - dcn3_2_ip.clamp_min_dcfclk = dc->config.clamp_min_dcfclk; + /* Overrides from dc->config options */ + dcn3_2_ip.clamp_min_dcfclk = dc->config.clamp_min_dcfclk; - /* Override from passed dc->bb_overrides if available*/ - if ((int)(dcn3_2_soc.sr_exit_time_us * 1000) != dc->bb_overrides.sr_exit_time_ns - && dc->bb_overrides.sr_exit_time_ns) { - dcn3_2_soc.sr_exit_time_us = dc->bb_overrides.sr_exit_time_ns / 1000.0; - } - - if ((int)(dcn3_2_soc.sr_enter_plus_exit_time_us * 1000) - != dc->bb_overrides.sr_enter_plus_exit_time_ns - && dc->bb_overrides.sr_enter_plus_exit_time_ns) { - dcn3_2_soc.sr_enter_plus_exit_time_us = - dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0; - } + /* Override from passed dc->bb_overrides if available*/ + if ((int)(dcn3_2_soc.sr_exit_time_us * 1000) != dc->bb_overrides.sr_exit_time_ns + && dc->bb_overrides.sr_exit_time_ns) { + dcn3_2_soc.sr_exit_time_us = dc->bb_overrides.sr_exit_time_ns / 1000.0; + } - if ((int)(dcn3_2_soc.urgent_latency_us * 1000) != dc->bb_overrides.urgent_latency_ns - && dc->bb_overrides.urgent_latency_ns) { - dcn3_2_soc.urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0; - dcn3_2_soc.urgent_latency_pixel_data_only_us = dc->bb_overrides.urgent_latency_ns / 1000.0; - } + if ((int)(dcn3_2_soc.sr_enter_plus_exit_time_us * 1000) + != dc->bb_overrides.sr_enter_plus_exit_time_ns + && dc->bb_overrides.sr_enter_plus_exit_time_ns) { + dcn3_2_soc.sr_enter_plus_exit_time_us = + dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0; + } - if ((int)(dcn3_2_soc.dram_clock_change_latency_us * 1000) - != dc->bb_overrides.dram_clock_change_latency_ns - && dc->bb_overrides.dram_clock_change_latency_ns) { - dcn3_2_soc.dram_clock_change_latency_us = - dc->bb_overrides.dram_clock_change_latency_ns / 1000.0; - } + if ((int)(dcn3_2_soc.urgent_latency_us * 1000) != dc->bb_overrides.urgent_latency_ns + && dc->bb_overrides.urgent_latency_ns) { + dcn3_2_soc.urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0; + dcn3_2_soc.urgent_latency_pixel_data_only_us = dc->bb_overrides.urgent_latency_ns / 1000.0; + } - if ((int)(dcn3_2_soc.fclk_change_latency_us * 1000) - != dc->bb_overrides.fclk_clock_change_latency_ns - && dc->bb_overrides.fclk_clock_change_latency_ns) { - dcn3_2_soc.fclk_change_latency_us = - dc->bb_overrides.fclk_clock_change_latency_ns / 1000; - } + if ((int)(dcn3_2_soc.dram_clock_change_latency_us * 1000) + != dc->bb_overrides.dram_clock_change_latency_ns + && dc->bb_overrides.dram_clock_change_latency_ns) { + dcn3_2_soc.dram_clock_change_latency_us = + dc->bb_overrides.dram_clock_change_latency_ns / 1000.0; + } - if ((int)(dcn3_2_soc.dummy_pstate_latency_us * 1000) - != dc->bb_overrides.dummy_clock_change_latency_ns - && dc->bb_overrides.dummy_clock_change_latency_ns) { - dcn3_2_soc.dummy_pstate_latency_us = - dc->bb_overrides.dummy_clock_change_latency_ns / 1000.0; - } + if ((int)(dcn3_2_soc.fclk_change_latency_us * 1000) + != dc->bb_overrides.fclk_clock_change_latency_ns + && dc->bb_overrides.fclk_clock_change_latency_ns) { + dcn3_2_soc.fclk_change_latency_us = + dc->bb_overrides.fclk_clock_change_latency_ns / 1000; + } - /* Override from VBIOS if VBIOS bb_info available */ - if (dc->ctx->dc_bios->funcs->get_soc_bb_info) { - struct bp_soc_bb_info bb_info = {0}; + if ((int)(dcn3_2_soc.dummy_pstate_latency_us * 1000) + != dc->bb_overrides.dummy_clock_change_latency_ns + && dc->bb_overrides.dummy_clock_change_latency_ns) { + dcn3_2_soc.dummy_pstate_latency_us = + dc->bb_overrides.dummy_clock_change_latency_ns / 1000.0; + } - if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) { - if (bb_info.dram_clock_change_latency_100ns > 0) - dcn3_2_soc.dram_clock_change_latency_us = - bb_info.dram_clock_change_latency_100ns * 10; + /* Override from VBIOS if VBIOS bb_info available */ + if (dc->ctx->dc_bios->funcs->get_soc_bb_info) { + struct bp_soc_bb_info bb_info = {0}; - if (bb_info.dram_sr_enter_exit_latency_100ns > 0) - dcn3_2_soc.sr_enter_plus_exit_time_us = - bb_info.dram_sr_enter_exit_latency_100ns * 10; + if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) { + if (bb_info.dram_clock_change_latency_100ns > 0) + dcn3_2_soc.dram_clock_change_latency_us = + bb_info.dram_clock_change_latency_100ns * 10; - if (bb_info.dram_sr_exit_latency_100ns > 0) - dcn3_2_soc.sr_exit_time_us = - bb_info.dram_sr_exit_latency_100ns * 10; - } - } + if (bb_info.dram_sr_enter_exit_latency_100ns > 0) + dcn3_2_soc.sr_enter_plus_exit_time_us = + bb_info.dram_sr_enter_exit_latency_100ns * 10; - /* Override from VBIOS for num_chan */ - if (dc->ctx->dc_bios->vram_info.num_chans) { - dcn3_2_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans; - dcn3_2_soc.mall_allocated_for_dcn_mbytes = (double)(dcn32_calc_num_avail_chans_for_mall(dc, - dc->ctx->dc_bios->vram_info.num_chans) * dc->caps.mall_size_per_mem_channel); + if (bb_info.dram_sr_exit_latency_100ns > 0) + dcn3_2_soc.sr_exit_time_us = + bb_info.dram_sr_exit_latency_100ns * 10; } + } - if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes) - dcn3_2_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes; + /* Override from VBIOS for num_chan */ + if (dc->ctx->dc_bios->vram_info.num_chans) { + dcn3_2_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans; + dcn3_2_soc.mall_allocated_for_dcn_mbytes = (double)(dcn32_calc_num_avail_chans_for_mall(dc, + dc->ctx->dc_bios->vram_info.num_chans) * dc->caps.mall_size_per_mem_channel); } + if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes) + dcn3_2_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes; + /* DML DSC delay factor workaround */ dcn3_2_ip.dsc_delay_factor_wa = dc->debug.dsc_delay_factor_wa_x1000 / 1000.0; @@ -2592,7 +2695,7 @@ void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_pa dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0; /* Overrides Clock levelsfrom CLK Mgr table entries as reported by PM FW */ - if ((!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) && (bw_params->clk_table.entries[0].memclk_mhz)) { + if (bw_params->clk_table.entries[0].memclk_mhz) { if (dc->debug.use_legacy_soc_bb_mechanism) { unsigned int i = 0, j = 0, num_states = 0; @@ -2736,7 +2839,8 @@ void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_pa dcn3_2_soc.clock_limits[i].phyclk_d32_mhz = dcn3_2_soc.clock_limits[0].phyclk_d32_mhz; } } else { - build_synthetic_soc_states(bw_params, dcn3_2_soc.clock_limits, &dcn3_2_soc.num_states); + build_synthetic_soc_states(dc->debug.disable_dc_mode_overwrite, bw_params, + dcn3_2_soc.clock_limits, &dcn3_2_soc.num_states); } /* Re-init DML with updated bb */ @@ -2783,15 +2887,76 @@ bool dcn32_allow_subvp_with_active_margin(struct pipe_ctx *pipe) } /** - * ******************************************************************************************* - * dcn32_determine_max_vratio_prefetch: Determine max Vratio for prefetch by driver policy + * dcn32_allow_subvp_high_refresh_rate: Determine if the high refresh rate config will allow subvp * - * @param [in]: dc: Current DC state - * @param [in]: context: New DC state to be programmed + * @dc: Current DC state + * @context: New DC state to be programmed + * @pipe: Pipe to be considered for use in subvp * - * @return: Max vratio for prefetch + * On high refresh rate display configs, we will allow subvp under the following conditions: + * 1. Resolution is 3840x2160, 3440x1440, or 2560x1440 + * 2. Refresh rate is between 120hz - 165hz + * 3. No scaling + * 4. Freesync is inactive + * 5. For single display cases, freesync must be disabled * - * ******************************************************************************************* + * Return: True if pipe can be used for subvp, false otherwise + */ +bool dcn32_allow_subvp_high_refresh_rate(struct dc *dc, struct dc_state *context, struct pipe_ctx *pipe) +{ + bool allow = false; + uint32_t refresh_rate = 0; + uint32_t subvp_min_refresh = subvp_high_refresh_list.min_refresh; + uint32_t subvp_max_refresh = subvp_high_refresh_list.max_refresh; + uint32_t min_refresh = subvp_max_refresh; + uint32_t i; + + /* Only allow SubVP on high refresh displays if all connected displays + * are considered "high refresh" (i.e. >= 120hz). We do not want to + * allow combinations such as 120hz (SubVP) + 60hz (SubVP). + */ + for (i = 0; i < dc->res_pool->pipe_count; i++) { + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; + + if (!pipe_ctx->stream) + continue; + refresh_rate = (pipe_ctx->stream->timing.pix_clk_100hz * 100 + + pipe_ctx->stream->timing.v_total * pipe_ctx->stream->timing.h_total - 1) + / (double)(pipe_ctx->stream->timing.v_total * pipe_ctx->stream->timing.h_total); + + if (refresh_rate < min_refresh) + min_refresh = refresh_rate; + } + + if (!dc->debug.disable_subvp_high_refresh && min_refresh >= subvp_min_refresh && pipe->stream && + pipe->plane_state && !(pipe->stream->vrr_active_variable || pipe->stream->vrr_active_fixed)) { + refresh_rate = (pipe->stream->timing.pix_clk_100hz * 100 + + pipe->stream->timing.v_total * pipe->stream->timing.h_total - 1) + / (double)(pipe->stream->timing.v_total * pipe->stream->timing.h_total); + if (refresh_rate >= subvp_min_refresh && refresh_rate <= subvp_max_refresh) { + for (i = 0; i < SUBVP_HIGH_REFRESH_LIST_LEN; i++) { + uint32_t width = subvp_high_refresh_list.res[i].width; + uint32_t height = subvp_high_refresh_list.res[i].height; + + if (dcn32_check_native_scaling_for_res(pipe, width, height)) { + if ((context->stream_count == 1 && !pipe->stream->allow_freesync) || context->stream_count > 1) { + allow = true; + break; + } + } + } + } + } + return allow; +} + +/** + * dcn32_determine_max_vratio_prefetch: Determine max Vratio for prefetch by driver policy + * + * @dc: Current DC state + * @context: New DC state to be programmed + * + * Return: Max vratio for prefetch */ double dcn32_determine_max_vratio_prefetch(struct dc *dc, struct dc_state *context) { @@ -2821,9 +2986,9 @@ double dcn32_determine_max_vratio_prefetch(struct dc *dc, struct dc_state *conte * ActiveMargin <= 0 to be the FPO stream candidate if found. * * - * @param [in]: dc - current dc state - * @param [in]: context - new dc state - * @param [out]: fpo_candidate_stream - pointer to FPO stream candidate if one is found + * @dc: current dc state + * @context: new dc state + * @fpo_candidate_stream: pointer to FPO stream candidate if one is found * * Return: void */ @@ -2849,10 +3014,9 @@ void dcn32_assign_fpo_vactive_candidate(struct dc *dc, const struct dc_state *co /** * dcn32_find_vactive_pipe - Determines if the config has a pipe that can switch in VACTIVE * - * @param [in]: dc - current dc state - * @param [in]: context - new dc state - * @param [in]: vactive_margin_req_us - The vactive marign required for a vactive pipe to be - * considered "found" + * @dc: current dc state + * @context: new dc state + * @vactive_margin_req_us: The vactive marign required for a vactive pipe to be considered "found" * * Return: True if VACTIVE display is found, false otherwise */ @@ -2861,6 +3025,7 @@ bool dcn32_find_vactive_pipe(struct dc *dc, const struct dc_state *context, uint unsigned int i, pipe_idx; const struct vba_vars_st *vba = &context->bw_ctx.dml.vba; bool vactive_found = false; + unsigned int blank_us = 0; for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) { const struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; @@ -2868,7 +3033,10 @@ bool dcn32_find_vactive_pipe(struct dc *dc, const struct dc_state *context, uint if (!pipe->stream) continue; - if (vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] >= vactive_margin_req_us) { + blank_us = ((pipe->stream->timing.v_total - pipe->stream->timing.v_addressable) * pipe->stream->timing.h_total / + (double)(pipe->stream->timing.pix_clk_100hz * 100)) * 1000000; + if (vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] >= vactive_margin_req_us && + !(pipe->stream->vrr_active_variable || pipe->stream->vrr_active_fixed) && blank_us < dc->debug.fpo_vactive_max_blank_us) { vactive_found = true; break; } @@ -2882,3 +3050,18 @@ void dcn32_set_clock_limits(const struct _vcs_dpi_soc_bounding_box_st *soc_bb) dc_assert_fp_enabled(); dcn3_2_soc.clock_limits[0].dcfclk_mhz = 1200.0; } + +void dcn32_override_min_req_memclk(struct dc *dc, struct dc_state *context) +{ + // WA: restrict FPO and SubVP to use first non-strobe mode (DCN32 BW issue) + if ((context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching || dcn32_subvp_in_use(dc, context)) && + dc->dml.soc.num_chans <= 8) { + int num_mclk_levels = dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels; + + if (context->bw_ctx.dml.vba.DRAMSpeed <= dc->clk_mgr->bw_params->clk_table.entries[0].memclk_mhz * 16 && + num_mclk_levels > 1) { + context->bw_ctx.dml.vba.DRAMSpeed = dc->clk_mgr->bw_params->clk_table.entries[1].memclk_mhz * 16; + context->bw_ctx.bw.dcn.clk.dramclk_khz = context->bw_ctx.dml.vba.DRAMSpeed * 1000 / 16; + } + } +} diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.h index dcf512cd3072..a4206b71d650 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.h +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.h @@ -80,6 +80,8 @@ void dcn32_assign_fpo_vactive_candidate(struct dc *dc, const struct dc_state *co bool dcn32_find_vactive_pipe(struct dc *dc, const struct dc_state *context, uint32_t vactive_margin_req); +void dcn32_override_min_req_memclk(struct dc *dc, struct dc_state *context); + void dcn32_set_clock_limits(const struct _vcs_dpi_soc_bounding_box_st *soc_bb); #endif diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.c index d75248b6cae9..cbdfb762c10c 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.c @@ -811,7 +811,7 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman v->SwathHeightC[k], TWait, (v->DRAMSpeedPerState[mode_lib->vba.VoltageLevel] <= MEM_STROBE_FREQ_MHZ || - v->DCFCLKPerState[mode_lib->vba.VoltageLevel] <= MIN_DCFCLK_FREQ_MHZ) ? + v->DCFCLKPerState[mode_lib->vba.VoltageLevel] <= DCFCLK_FREQ_EXTRA_PREFETCH_REQ_MHZ) ? mode_lib->vba.ip.min_prefetch_in_strobe_us : 0, /* Output */ &v->DSTXAfterScaler[k], @@ -2323,10 +2323,14 @@ void dml32_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l mode_lib->vba.LinkCapacitySupport[i] = true; for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) { if (mode_lib->vba.BlendingAndTiming[k] == k - && (mode_lib->vba.Output[k] == dm_dp || mode_lib->vba.Output[k] == dm_dp2p0 - || mode_lib->vba.Output[k] == dm_edp - || mode_lib->vba.Output[k] == dm_hdmi) - && mode_lib->vba.OutputBppPerState[i][k] == 0) { + && (mode_lib->vba.Output[k] == dm_dp || mode_lib->vba.Output[k] == dm_dp2p0 + || mode_lib->vba.Output[k] == dm_edp + || mode_lib->vba.Output[k] == dm_hdmi) + && mode_lib->vba.OutputBppPerState[i][k] == 0 && + (mode_lib->vba.UsesMALLForPStateChange[k] != dm_use_mall_pstate_change_phantom_pipe)) { + /* Phantom pipes don't consider DSC in DML, so it could fail link check. + * However, we don't care about the link for phantom pipes. + */ mode_lib->vba.LinkCapacitySupport[i] = false; } } @@ -3311,7 +3315,7 @@ void dml32_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l v->swath_width_chroma_ub_this_state[k], v->SwathHeightYThisState[k], v->SwathHeightCThisState[k], v->TWait, - (v->DRAMSpeedPerState[i] <= MEM_STROBE_FREQ_MHZ || v->DCFCLKState[i][j] <= MIN_DCFCLK_FREQ_MHZ) ? + (v->DRAMSpeedPerState[i] <= MEM_STROBE_FREQ_MHZ || v->DCFCLKState[i][j] <= DCFCLK_FREQ_EXTRA_PREFETCH_REQ_MHZ) ? mode_lib->vba.ip.min_prefetch_in_strobe_us : 0, /* Output */ diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.h b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.h index d98e36a9a09c..c4745d63039b 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.h +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.h @@ -53,7 +53,7 @@ #define BPP_BLENDED_PIPE 0xffffffff #define MEM_STROBE_FREQ_MHZ 1600 -#define MIN_DCFCLK_FREQ_MHZ 200 +#define DCFCLK_FREQ_EXTRA_PREFETCH_REQ_MHZ 300 #define MEM_STROBE_MAX_DELIVERY_TIME_US 60.0 struct display_mode_lib; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c index 61cc4904ade4..a50e7f4dce42 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c @@ -1595,7 +1595,6 @@ double dml32_TruncToValidBPP( unsigned int NonDSCBPP0; unsigned int NonDSCBPP1; unsigned int NonDSCBPP2; - unsigned int NonDSCBPP3; if (Format == dm_420) { NonDSCBPP0 = 12; @@ -1604,10 +1603,9 @@ double dml32_TruncToValidBPP( MinDSCBPP = 6; MaxDSCBPP = 1.5 * DSCInputBitPerComponent - 1 / 16; } else if (Format == dm_444) { - NonDSCBPP0 = 18; - NonDSCBPP1 = 24; - NonDSCBPP2 = 30; - NonDSCBPP3 = 36; + NonDSCBPP0 = 24; + NonDSCBPP1 = 30; + NonDSCBPP2 = 36; MinDSCBPP = 8; MaxDSCBPP = 3 * DSCInputBitPerComponent - 1.0 / 16; } else { @@ -1661,9 +1659,7 @@ double dml32_TruncToValidBPP( else return dml_floor(16.0 * MaxLinkBPP, 1.0) / 16.0; } else { - if (MaxLinkBPP >= NonDSCBPP3) - return NonDSCBPP3; - else if (MaxLinkBPP >= NonDSCBPP2) + if (MaxLinkBPP >= NonDSCBPP2) return NonDSCBPP2; else if (MaxLinkBPP >= NonDSCBPP1) return NonDSCBPP1; @@ -1674,7 +1670,7 @@ double dml32_TruncToValidBPP( } } else { if (!((DSCEnable == false && (DesiredBPP == NonDSCBPP2 || DesiredBPP == NonDSCBPP1 || - DesiredBPP == NonDSCBPP0 || DesiredBPP == NonDSCBPP3)) || + DesiredBPP <= NonDSCBPP0)) || (DSCEnable && DesiredBPP >= MinDSCBPP && DesiredBPP <= MaxDSCBPP))) return BPP_INVALID; else @@ -4342,7 +4338,7 @@ void dml32_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport( + v->WritebackChunkSize * 1024.0 / 32.0 / SOCCLK; } if (v->USRRetrainingRequiredFinal) - v->Watermark.WritebackUrgentWatermark = v->Watermark.WritebackUrgentWatermark + v->Watermark.WritebackDRAMClockChangeWatermark = v->Watermark.WritebackDRAMClockChangeWatermark + mmSOCParameters.USRRetrainingLatency; if (TotalActiveWriteback <= 1) { @@ -4660,6 +4656,10 @@ void dml32_CalculateMinAndMaxPrefetchMode( } else if (AllowForPStateChangeOrStutterInVBlankFinal == dm_prefetch_support_uclk_fclk_and_stutter) { *MinPrefetchMode = 0; *MaxPrefetchMode = 0; + } else if (AllowForPStateChangeOrStutterInVBlankFinal == + dm_prefetch_support_uclk_fclk_and_stutter_if_possible) { + *MinPrefetchMode = 0; + *MaxPrefetchMode = 3; } else { *MinPrefetchMode = 0; *MaxPrefetchMode = 3; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_rq_dlg_calc_32.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_rq_dlg_calc_32.c index 395ae8761980..9ba6cb67655f 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_rq_dlg_calc_32.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_rq_dlg_calc_32.c @@ -116,7 +116,7 @@ void dml32_rq_dlg_get_rq_reg(display_rq_regs_st *rq_regs, else rq_regs->rq_regs_l.min_meta_chunk_size = dml_log2(min_meta_chunk_bytes) - 6 + 1; - if (min_meta_chunk_bytes == 0) + if (p1_min_meta_chunk_bytes == 0) rq_regs->rq_regs_c.min_meta_chunk_size = 0; else rq_regs->rq_regs_c.min_meta_chunk_size = dml_log2(p1_min_meta_chunk_bytes) - 6 + 1; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c index 342a1bcb4927..f0683fd9d3f0 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c @@ -252,14 +252,48 @@ static void remove_entry_from_table_at_index(struct _vcs_dpi_voltage_scaling_st memset(&table[--(*num_entries)], 0, sizeof(struct _vcs_dpi_voltage_scaling_st)); } -static int build_synthetic_soc_states(struct clk_bw_params *bw_params, +/* + * override_max_clk_values - Overwrite the max clock frequencies with the max DC mode timings + * Input: + * max_clk_limit - struct containing the desired clock timings + * Output: + * curr_clk_limit - struct containing the timings that need to be overwritten + * Return: 0 upon success, non-zero for failure + */ +static int override_max_clk_values(struct clk_limit_table_entry *max_clk_limit, + struct clk_limit_table_entry *curr_clk_limit) +{ + if (NULL == max_clk_limit || NULL == curr_clk_limit) + return -1; //invalid parameters + + //only overwrite if desired max clock frequency is initialized + if (max_clk_limit->dcfclk_mhz != 0) + curr_clk_limit->dcfclk_mhz = max_clk_limit->dcfclk_mhz; + + if (max_clk_limit->fclk_mhz != 0) + curr_clk_limit->fclk_mhz = max_clk_limit->fclk_mhz; + + if (max_clk_limit->memclk_mhz != 0) + curr_clk_limit->memclk_mhz = max_clk_limit->memclk_mhz; + + if (max_clk_limit->socclk_mhz != 0) + curr_clk_limit->socclk_mhz = max_clk_limit->socclk_mhz; + + if (max_clk_limit->dtbclk_mhz != 0) + curr_clk_limit->dtbclk_mhz = max_clk_limit->dtbclk_mhz; + + if (max_clk_limit->dispclk_mhz != 0) + curr_clk_limit->dispclk_mhz = max_clk_limit->dispclk_mhz; + + return 0; +} + +static int build_synthetic_soc_states(bool disable_dc_mode_overwrite, struct clk_bw_params *bw_params, struct _vcs_dpi_voltage_scaling_st *table, unsigned int *num_entries) { int i, j; struct _vcs_dpi_voltage_scaling_st entry = {0}; - - unsigned int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, - max_phyclk_mhz = 0, max_dtbclk_mhz = 0, max_fclk_mhz = 0, max_uclk_mhz = 0; + struct clk_limit_table_entry max_clk_data = {0}; unsigned int min_dcfclk_mhz = 199, min_fclk_mhz = 299; @@ -270,53 +304,78 @@ static int build_synthetic_soc_states(struct clk_bw_params *bw_params, unsigned int num_fclk_dpms = 0; unsigned int num_dcfclk_dpms = 0; + unsigned int num_dc_uclk_dpms = 0; + unsigned int num_dc_fclk_dpms = 0; + unsigned int num_dc_dcfclk_dpms = 0; + for (i = 0; i < MAX_NUM_DPM_LVL; i++) { - if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz) - max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz; - if (bw_params->clk_table.entries[i].fclk_mhz > max_fclk_mhz) - max_fclk_mhz = bw_params->clk_table.entries[i].fclk_mhz; - if (bw_params->clk_table.entries[i].memclk_mhz > max_uclk_mhz) - max_uclk_mhz = bw_params->clk_table.entries[i].memclk_mhz; - if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz) - max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz; - if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz) - max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz; - if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz) - max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz; - if (bw_params->clk_table.entries[i].dtbclk_mhz > max_dtbclk_mhz) - max_dtbclk_mhz = bw_params->clk_table.entries[i].dtbclk_mhz; - - if (bw_params->clk_table.entries[i].memclk_mhz > 0) + if (bw_params->clk_table.entries[i].dcfclk_mhz > max_clk_data.dcfclk_mhz) + max_clk_data.dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz; + if (bw_params->clk_table.entries[i].fclk_mhz > max_clk_data.fclk_mhz) + max_clk_data.fclk_mhz = bw_params->clk_table.entries[i].fclk_mhz; + if (bw_params->clk_table.entries[i].memclk_mhz > max_clk_data.memclk_mhz) + max_clk_data.memclk_mhz = bw_params->clk_table.entries[i].memclk_mhz; + if (bw_params->clk_table.entries[i].dispclk_mhz > max_clk_data.dispclk_mhz) + max_clk_data.dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz; + if (bw_params->clk_table.entries[i].dppclk_mhz > max_clk_data.dppclk_mhz) + max_clk_data.dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz; + if (bw_params->clk_table.entries[i].phyclk_mhz > max_clk_data.phyclk_mhz) + max_clk_data.phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz; + if (bw_params->clk_table.entries[i].dtbclk_mhz > max_clk_data.dtbclk_mhz) + max_clk_data.dtbclk_mhz = bw_params->clk_table.entries[i].dtbclk_mhz; + + if (bw_params->clk_table.entries[i].memclk_mhz > 0) { num_uclk_dpms++; - if (bw_params->clk_table.entries[i].fclk_mhz > 0) + if (bw_params->clk_table.entries[i].memclk_mhz <= bw_params->dc_mode_limit.memclk_mhz) + num_dc_uclk_dpms++; + } + if (bw_params->clk_table.entries[i].fclk_mhz > 0) { num_fclk_dpms++; - if (bw_params->clk_table.entries[i].dcfclk_mhz > 0) + if (bw_params->clk_table.entries[i].fclk_mhz <= bw_params->dc_mode_limit.fclk_mhz) + num_dc_fclk_dpms++; + } + if (bw_params->clk_table.entries[i].dcfclk_mhz > 0) { num_dcfclk_dpms++; + if (bw_params->clk_table.entries[i].dcfclk_mhz <= bw_params->dc_mode_limit.dcfclk_mhz) + num_dc_dcfclk_dpms++; + } + } + + if (!disable_dc_mode_overwrite) { + //Overwrite max frequencies with max DC mode frequencies for DC mode systems + override_max_clk_values(&bw_params->dc_mode_limit, &max_clk_data); + num_uclk_dpms = num_dc_uclk_dpms; + num_fclk_dpms = num_dc_fclk_dpms; + num_dcfclk_dpms = num_dc_dcfclk_dpms; + bw_params->clk_table.num_entries_per_clk.num_memclk_levels = num_uclk_dpms; + bw_params->clk_table.num_entries_per_clk.num_fclk_levels = num_fclk_dpms; } if (num_dcfclk_dpms > 0 && bw_params->clk_table.entries[0].fclk_mhz > min_fclk_mhz) min_fclk_mhz = bw_params->clk_table.entries[0].fclk_mhz; - if (!max_dcfclk_mhz || !max_dispclk_mhz || !max_dtbclk_mhz) + if (!max_clk_data.dcfclk_mhz || !max_clk_data.dispclk_mhz || !max_clk_data.dtbclk_mhz) return -1; - if (max_dppclk_mhz == 0) - max_dppclk_mhz = max_dispclk_mhz; + if (max_clk_data.dppclk_mhz == 0) + max_clk_data.dppclk_mhz = max_clk_data.dispclk_mhz; - if (max_fclk_mhz == 0) - max_fclk_mhz = max_dcfclk_mhz * dcn3_21_soc.pct_ideal_sdp_bw_after_urgent / dcn3_21_soc.pct_ideal_fabric_bw_after_urgent; + if (max_clk_data.fclk_mhz == 0) + max_clk_data.fclk_mhz = max_clk_data.dcfclk_mhz * + dcn3_2_soc.pct_ideal_sdp_bw_after_urgent / + dcn3_2_soc.pct_ideal_fabric_bw_after_urgent; - if (max_phyclk_mhz == 0) - max_phyclk_mhz = dcn3_21_soc.clock_limits[0].phyclk_mhz; + if (max_clk_data.phyclk_mhz == 0) + max_clk_data.phyclk_mhz = dcn3_2_soc.clock_limits[0].phyclk_mhz; *num_entries = 0; - entry.dispclk_mhz = max_dispclk_mhz; - entry.dscclk_mhz = max_dispclk_mhz / 3; - entry.dppclk_mhz = max_dppclk_mhz; - entry.dtbclk_mhz = max_dtbclk_mhz; - entry.phyclk_mhz = max_phyclk_mhz; - entry.phyclk_d18_mhz = dcn3_21_soc.clock_limits[0].phyclk_d18_mhz; - entry.phyclk_d32_mhz = dcn3_21_soc.clock_limits[0].phyclk_d32_mhz; + entry.dispclk_mhz = max_clk_data.dispclk_mhz; + entry.dscclk_mhz = max_clk_data.dispclk_mhz / 3; + entry.dppclk_mhz = max_clk_data.dppclk_mhz; + entry.dtbclk_mhz = max_clk_data.dtbclk_mhz; + entry.phyclk_mhz = max_clk_data.phyclk_mhz; + entry.phyclk_d18_mhz = dcn3_2_soc.clock_limits[0].phyclk_d18_mhz; + entry.phyclk_d32_mhz = dcn3_2_soc.clock_limits[0].phyclk_d32_mhz; // Insert all the DCFCLK STAs for (i = 0; i < num_dcfclk_stas; i++) { @@ -328,7 +387,7 @@ static int build_synthetic_soc_states(struct clk_bw_params *bw_params, } // Insert the max DCFCLK - entry.dcfclk_mhz = max_dcfclk_mhz; + entry.dcfclk_mhz = max_clk_data.dcfclk_mhz; entry.fabricclk_mhz = 0; entry.dram_speed_mts = 0; @@ -356,7 +415,7 @@ static int build_synthetic_soc_states(struct clk_bw_params *bw_params, // If FCLK fine grained, only insert max else { entry.dcfclk_mhz = 0; - entry.fabricclk_mhz = max_fclk_mhz; + entry.fabricclk_mhz = max_clk_data.fclk_mhz; entry.dram_speed_mts = 0; dcn321_insert_entry_into_table_sorted(table, num_entries, &entry); @@ -368,9 +427,9 @@ static int build_synthetic_soc_states(struct clk_bw_params *bw_params, // Remove states that require higher clocks than are supported for (i = *num_entries - 1; i >= 0 ; i--) { - if (table[i].dcfclk_mhz > max_dcfclk_mhz || - table[i].fabricclk_mhz > max_fclk_mhz || - table[i].dram_speed_mts > max_uclk_mhz * 16) + if (table[i].dcfclk_mhz > max_clk_data.dcfclk_mhz || + table[i].fabricclk_mhz > max_clk_data.fclk_mhz || + table[i].dram_speed_mts > max_clk_data.memclk_mhz * 16) remove_entry_from_table_at_index(table, num_entries, i); } @@ -471,80 +530,78 @@ static void dcn321_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts, void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_params) { dc_assert_fp_enabled(); - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - /* Overrides from dc->config options */ - dcn3_21_ip.clamp_min_dcfclk = dc->config.clamp_min_dcfclk; - - /* Override from passed dc->bb_overrides if available*/ - if ((int)(dcn3_21_soc.sr_exit_time_us * 1000) != dc->bb_overrides.sr_exit_time_ns - && dc->bb_overrides.sr_exit_time_ns) { - dcn3_21_soc.sr_exit_time_us = dc->bb_overrides.sr_exit_time_ns / 1000.0; - } + /* Overrides from dc->config options */ + dcn3_21_ip.clamp_min_dcfclk = dc->config.clamp_min_dcfclk; - if ((int)(dcn3_21_soc.sr_enter_plus_exit_time_us * 1000) - != dc->bb_overrides.sr_enter_plus_exit_time_ns - && dc->bb_overrides.sr_enter_plus_exit_time_ns) { - dcn3_21_soc.sr_enter_plus_exit_time_us = - dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0; - } + /* Override from passed dc->bb_overrides if available*/ + if ((int)(dcn3_21_soc.sr_exit_time_us * 1000) != dc->bb_overrides.sr_exit_time_ns + && dc->bb_overrides.sr_exit_time_ns) { + dcn3_21_soc.sr_exit_time_us = dc->bb_overrides.sr_exit_time_ns / 1000.0; + } - if ((int)(dcn3_21_soc.urgent_latency_us * 1000) != dc->bb_overrides.urgent_latency_ns - && dc->bb_overrides.urgent_latency_ns) { - dcn3_21_soc.urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0; - dcn3_21_soc.urgent_latency_pixel_data_only_us = dc->bb_overrides.urgent_latency_ns / 1000.0; - } + if ((int)(dcn3_21_soc.sr_enter_plus_exit_time_us * 1000) + != dc->bb_overrides.sr_enter_plus_exit_time_ns + && dc->bb_overrides.sr_enter_plus_exit_time_ns) { + dcn3_21_soc.sr_enter_plus_exit_time_us = + dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0; + } - if ((int)(dcn3_21_soc.dram_clock_change_latency_us * 1000) - != dc->bb_overrides.dram_clock_change_latency_ns - && dc->bb_overrides.dram_clock_change_latency_ns) { - dcn3_21_soc.dram_clock_change_latency_us = - dc->bb_overrides.dram_clock_change_latency_ns / 1000.0; - } + if ((int)(dcn3_21_soc.urgent_latency_us * 1000) != dc->bb_overrides.urgent_latency_ns + && dc->bb_overrides.urgent_latency_ns) { + dcn3_21_soc.urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0; + dcn3_21_soc.urgent_latency_pixel_data_only_us = dc->bb_overrides.urgent_latency_ns / 1000.0; + } - if ((int)(dcn3_21_soc.fclk_change_latency_us * 1000) - != dc->bb_overrides.fclk_clock_change_latency_ns - && dc->bb_overrides.fclk_clock_change_latency_ns) { - dcn3_21_soc.fclk_change_latency_us = - dc->bb_overrides.fclk_clock_change_latency_ns / 1000; - } + if ((int)(dcn3_21_soc.dram_clock_change_latency_us * 1000) + != dc->bb_overrides.dram_clock_change_latency_ns + && dc->bb_overrides.dram_clock_change_latency_ns) { + dcn3_21_soc.dram_clock_change_latency_us = + dc->bb_overrides.dram_clock_change_latency_ns / 1000.0; + } - if ((int)(dcn3_21_soc.dummy_pstate_latency_us * 1000) - != dc->bb_overrides.dummy_clock_change_latency_ns - && dc->bb_overrides.dummy_clock_change_latency_ns) { - dcn3_21_soc.dummy_pstate_latency_us = - dc->bb_overrides.dummy_clock_change_latency_ns / 1000.0; - } + if ((int)(dcn3_21_soc.fclk_change_latency_us * 1000) + != dc->bb_overrides.fclk_clock_change_latency_ns + && dc->bb_overrides.fclk_clock_change_latency_ns) { + dcn3_21_soc.fclk_change_latency_us = + dc->bb_overrides.fclk_clock_change_latency_ns / 1000; + } - /* Override from VBIOS if VBIOS bb_info available */ - if (dc->ctx->dc_bios->funcs->get_soc_bb_info) { - struct bp_soc_bb_info bb_info = {0}; + if ((int)(dcn3_21_soc.dummy_pstate_latency_us * 1000) + != dc->bb_overrides.dummy_clock_change_latency_ns + && dc->bb_overrides.dummy_clock_change_latency_ns) { + dcn3_21_soc.dummy_pstate_latency_us = + dc->bb_overrides.dummy_clock_change_latency_ns / 1000.0; + } - if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) { - if (bb_info.dram_clock_change_latency_100ns > 0) - dcn3_21_soc.dram_clock_change_latency_us = - bb_info.dram_clock_change_latency_100ns * 10; + /* Override from VBIOS if VBIOS bb_info available */ + if (dc->ctx->dc_bios->funcs->get_soc_bb_info) { + struct bp_soc_bb_info bb_info = {0}; - if (bb_info.dram_sr_enter_exit_latency_100ns > 0) - dcn3_21_soc.sr_enter_plus_exit_time_us = - bb_info.dram_sr_enter_exit_latency_100ns * 10; + if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) { + if (bb_info.dram_clock_change_latency_100ns > 0) + dcn3_21_soc.dram_clock_change_latency_us = + bb_info.dram_clock_change_latency_100ns * 10; - if (bb_info.dram_sr_exit_latency_100ns > 0) - dcn3_21_soc.sr_exit_time_us = - bb_info.dram_sr_exit_latency_100ns * 10; - } - } + if (bb_info.dram_sr_enter_exit_latency_100ns > 0) + dcn3_21_soc.sr_enter_plus_exit_time_us = + bb_info.dram_sr_enter_exit_latency_100ns * 10; - /* Override from VBIOS for num_chan */ - if (dc->ctx->dc_bios->vram_info.num_chans) { - dcn3_21_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans; - dcn3_21_soc.mall_allocated_for_dcn_mbytes = (double)(dcn32_calc_num_avail_chans_for_mall(dc, - dc->ctx->dc_bios->vram_info.num_chans) * dc->caps.mall_size_per_mem_channel); + if (bb_info.dram_sr_exit_latency_100ns > 0) + dcn3_21_soc.sr_exit_time_us = + bb_info.dram_sr_exit_latency_100ns * 10; } + } - if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes) - dcn3_21_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes; + /* Override from VBIOS for num_chan */ + if (dc->ctx->dc_bios->vram_info.num_chans) { + dcn3_21_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans; + dcn3_21_soc.mall_allocated_for_dcn_mbytes = (double)(dcn32_calc_num_avail_chans_for_mall(dc, + dc->ctx->dc_bios->vram_info.num_chans) * dc->caps.mall_size_per_mem_channel); } + if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes) + dcn3_21_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes; + /* DML DSC delay factor workaround */ dcn3_21_ip.dsc_delay_factor_wa = dc->debug.dsc_delay_factor_wa_x1000 / 1000.0; @@ -555,150 +612,149 @@ void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0; /* Overrides Clock levelsfrom CLK Mgr table entries as reported by PM FW */ - if ((!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) && (bw_params->clk_table.entries[0].memclk_mhz)) { - if (dc->debug.use_legacy_soc_bb_mechanism) { - unsigned int i = 0, j = 0, num_states = 0; - - unsigned int dcfclk_mhz[DC__VOLTAGE_STATES] = {0}; - unsigned int dram_speed_mts[DC__VOLTAGE_STATES] = {0}; - unsigned int optimal_uclk_for_dcfclk_sta_targets[DC__VOLTAGE_STATES] = {0}; - unsigned int optimal_dcfclk_for_uclk[DC__VOLTAGE_STATES] = {0}; - - unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {615, 906, 1324, 1564}; - unsigned int num_dcfclk_sta_targets = 4, num_uclk_states = 0; - unsigned int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, max_phyclk_mhz = 0; - - for (i = 0; i < MAX_NUM_DPM_LVL; i++) { - if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz) - max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz; - if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz) - max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz; - if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz) - max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz; - if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz) - max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz; - } - if (!max_dcfclk_mhz) - max_dcfclk_mhz = dcn3_21_soc.clock_limits[0].dcfclk_mhz; - if (!max_dispclk_mhz) - max_dispclk_mhz = dcn3_21_soc.clock_limits[0].dispclk_mhz; - if (!max_dppclk_mhz) - max_dppclk_mhz = dcn3_21_soc.clock_limits[0].dppclk_mhz; - if (!max_phyclk_mhz) - max_phyclk_mhz = dcn3_21_soc.clock_limits[0].phyclk_mhz; - - if (max_dcfclk_mhz > dcfclk_sta_targets[num_dcfclk_sta_targets-1]) { - // If max DCFCLK is greater than the max DCFCLK STA target, insert into the DCFCLK STA target array - dcfclk_sta_targets[num_dcfclk_sta_targets] = max_dcfclk_mhz; - num_dcfclk_sta_targets++; - } else if (max_dcfclk_mhz < dcfclk_sta_targets[num_dcfclk_sta_targets-1]) { - // If max DCFCLK is less than the max DCFCLK STA target, cap values and remove duplicates - for (i = 0; i < num_dcfclk_sta_targets; i++) { - if (dcfclk_sta_targets[i] > max_dcfclk_mhz) { - dcfclk_sta_targets[i] = max_dcfclk_mhz; - break; - } + if (dc->debug.use_legacy_soc_bb_mechanism) { + unsigned int i = 0, j = 0, num_states = 0; + + unsigned int dcfclk_mhz[DC__VOLTAGE_STATES] = {0}; + unsigned int dram_speed_mts[DC__VOLTAGE_STATES] = {0}; + unsigned int optimal_uclk_for_dcfclk_sta_targets[DC__VOLTAGE_STATES] = {0}; + unsigned int optimal_dcfclk_for_uclk[DC__VOLTAGE_STATES] = {0}; + + unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {615, 906, 1324, 1564}; + unsigned int num_dcfclk_sta_targets = 4, num_uclk_states = 0; + unsigned int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, max_phyclk_mhz = 0; + + for (i = 0; i < MAX_NUM_DPM_LVL; i++) { + if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz) + max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz; + if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz) + max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz; + if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz) + max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz; + if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz) + max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz; + } + if (!max_dcfclk_mhz) + max_dcfclk_mhz = dcn3_21_soc.clock_limits[0].dcfclk_mhz; + if (!max_dispclk_mhz) + max_dispclk_mhz = dcn3_21_soc.clock_limits[0].dispclk_mhz; + if (!max_dppclk_mhz) + max_dppclk_mhz = dcn3_21_soc.clock_limits[0].dppclk_mhz; + if (!max_phyclk_mhz) + max_phyclk_mhz = dcn3_21_soc.clock_limits[0].phyclk_mhz; + + if (max_dcfclk_mhz > dcfclk_sta_targets[num_dcfclk_sta_targets-1]) { + // If max DCFCLK is greater than the max DCFCLK STA target, insert into the DCFCLK STA target array + dcfclk_sta_targets[num_dcfclk_sta_targets] = max_dcfclk_mhz; + num_dcfclk_sta_targets++; + } else if (max_dcfclk_mhz < dcfclk_sta_targets[num_dcfclk_sta_targets-1]) { + // If max DCFCLK is less than the max DCFCLK STA target, cap values and remove duplicates + for (i = 0; i < num_dcfclk_sta_targets; i++) { + if (dcfclk_sta_targets[i] > max_dcfclk_mhz) { + dcfclk_sta_targets[i] = max_dcfclk_mhz; + break; } - // Update size of array since we "removed" duplicates - num_dcfclk_sta_targets = i + 1; } + // Update size of array since we "removed" duplicates + num_dcfclk_sta_targets = i + 1; + } - num_uclk_states = bw_params->clk_table.num_entries; + num_uclk_states = bw_params->clk_table.num_entries; - // Calculate optimal dcfclk for each uclk - for (i = 0; i < num_uclk_states; i++) { - dcn321_get_optimal_dcfclk_fclk_for_uclk(bw_params->clk_table.entries[i].memclk_mhz * 16, - &optimal_dcfclk_for_uclk[i], NULL); - if (optimal_dcfclk_for_uclk[i] < bw_params->clk_table.entries[0].dcfclk_mhz) { - optimal_dcfclk_for_uclk[i] = bw_params->clk_table.entries[0].dcfclk_mhz; - } + // Calculate optimal dcfclk for each uclk + for (i = 0; i < num_uclk_states; i++) { + dcn321_get_optimal_dcfclk_fclk_for_uclk(bw_params->clk_table.entries[i].memclk_mhz * 16, + &optimal_dcfclk_for_uclk[i], NULL); + if (optimal_dcfclk_for_uclk[i] < bw_params->clk_table.entries[0].dcfclk_mhz) { + optimal_dcfclk_for_uclk[i] = bw_params->clk_table.entries[0].dcfclk_mhz; } + } - // Calculate optimal uclk for each dcfclk sta target - for (i = 0; i < num_dcfclk_sta_targets; i++) { - for (j = 0; j < num_uclk_states; j++) { - if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) { - optimal_uclk_for_dcfclk_sta_targets[i] = - bw_params->clk_table.entries[j].memclk_mhz * 16; - break; - } + // Calculate optimal uclk for each dcfclk sta target + for (i = 0; i < num_dcfclk_sta_targets; i++) { + for (j = 0; j < num_uclk_states; j++) { + if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) { + optimal_uclk_for_dcfclk_sta_targets[i] = + bw_params->clk_table.entries[j].memclk_mhz * 16; + break; } } + } - i = 0; - j = 0; - // create the final dcfclk and uclk table - while (i < num_dcfclk_sta_targets && j < num_uclk_states && num_states < DC__VOLTAGE_STATES) { - if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j] && i < num_dcfclk_sta_targets) { - dcfclk_mhz[num_states] = dcfclk_sta_targets[i]; - dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++]; + i = 0; + j = 0; + // create the final dcfclk and uclk table + while (i < num_dcfclk_sta_targets && j < num_uclk_states && num_states < DC__VOLTAGE_STATES) { + if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j] && i < num_dcfclk_sta_targets) { + dcfclk_mhz[num_states] = dcfclk_sta_targets[i]; + dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++]; + } else { + if (j < num_uclk_states && optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) { + dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j]; + dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16; } else { - if (j < num_uclk_states && optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) { - dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j]; - dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16; - } else { - j = num_uclk_states; - } + j = num_uclk_states; } } + } - while (i < num_dcfclk_sta_targets && num_states < DC__VOLTAGE_STATES) { - dcfclk_mhz[num_states] = dcfclk_sta_targets[i]; - dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++]; - } + while (i < num_dcfclk_sta_targets && num_states < DC__VOLTAGE_STATES) { + dcfclk_mhz[num_states] = dcfclk_sta_targets[i]; + dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++]; + } - while (j < num_uclk_states && num_states < DC__VOLTAGE_STATES && - optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) { - dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j]; - dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16; - } + while (j < num_uclk_states && num_states < DC__VOLTAGE_STATES && + optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) { + dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j]; + dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16; + } - dcn3_21_soc.num_states = num_states; - for (i = 0; i < dcn3_21_soc.num_states; i++) { - dcn3_21_soc.clock_limits[i].state = i; - dcn3_21_soc.clock_limits[i].dcfclk_mhz = dcfclk_mhz[i]; - dcn3_21_soc.clock_limits[i].fabricclk_mhz = dcfclk_mhz[i]; - - /* Fill all states with max values of all these clocks */ - dcn3_21_soc.clock_limits[i].dispclk_mhz = max_dispclk_mhz; - dcn3_21_soc.clock_limits[i].dppclk_mhz = max_dppclk_mhz; - dcn3_21_soc.clock_limits[i].phyclk_mhz = max_phyclk_mhz; - dcn3_21_soc.clock_limits[i].dscclk_mhz = max_dispclk_mhz / 3; - - /* Populate from bw_params for DTBCLK, SOCCLK */ - if (i > 0) { - if (!bw_params->clk_table.entries[i].dtbclk_mhz) { - dcn3_21_soc.clock_limits[i].dtbclk_mhz = dcn3_21_soc.clock_limits[i-1].dtbclk_mhz; - } else { - dcn3_21_soc.clock_limits[i].dtbclk_mhz = bw_params->clk_table.entries[i].dtbclk_mhz; - } - } else if (bw_params->clk_table.entries[i].dtbclk_mhz) { + dcn3_21_soc.num_states = num_states; + for (i = 0; i < dcn3_21_soc.num_states; i++) { + dcn3_21_soc.clock_limits[i].state = i; + dcn3_21_soc.clock_limits[i].dcfclk_mhz = dcfclk_mhz[i]; + dcn3_21_soc.clock_limits[i].fabricclk_mhz = dcfclk_mhz[i]; + + /* Fill all states with max values of all these clocks */ + dcn3_21_soc.clock_limits[i].dispclk_mhz = max_dispclk_mhz; + dcn3_21_soc.clock_limits[i].dppclk_mhz = max_dppclk_mhz; + dcn3_21_soc.clock_limits[i].phyclk_mhz = max_phyclk_mhz; + dcn3_21_soc.clock_limits[i].dscclk_mhz = max_dispclk_mhz / 3; + + /* Populate from bw_params for DTBCLK, SOCCLK */ + if (i > 0) { + if (!bw_params->clk_table.entries[i].dtbclk_mhz) { + dcn3_21_soc.clock_limits[i].dtbclk_mhz = dcn3_21_soc.clock_limits[i-1].dtbclk_mhz; + } else { dcn3_21_soc.clock_limits[i].dtbclk_mhz = bw_params->clk_table.entries[i].dtbclk_mhz; } + } else if (bw_params->clk_table.entries[i].dtbclk_mhz) { + dcn3_21_soc.clock_limits[i].dtbclk_mhz = bw_params->clk_table.entries[i].dtbclk_mhz; + } - if (!bw_params->clk_table.entries[i].socclk_mhz && i > 0) - dcn3_21_soc.clock_limits[i].socclk_mhz = dcn3_21_soc.clock_limits[i-1].socclk_mhz; - else - dcn3_21_soc.clock_limits[i].socclk_mhz = bw_params->clk_table.entries[i].socclk_mhz; + if (!bw_params->clk_table.entries[i].socclk_mhz && i > 0) + dcn3_21_soc.clock_limits[i].socclk_mhz = dcn3_21_soc.clock_limits[i-1].socclk_mhz; + else + dcn3_21_soc.clock_limits[i].socclk_mhz = bw_params->clk_table.entries[i].socclk_mhz; - if (!dram_speed_mts[i] && i > 0) - dcn3_21_soc.clock_limits[i].dram_speed_mts = dcn3_21_soc.clock_limits[i-1].dram_speed_mts; - else - dcn3_21_soc.clock_limits[i].dram_speed_mts = dram_speed_mts[i]; + if (!dram_speed_mts[i] && i > 0) + dcn3_21_soc.clock_limits[i].dram_speed_mts = dcn3_21_soc.clock_limits[i-1].dram_speed_mts; + else + dcn3_21_soc.clock_limits[i].dram_speed_mts = dram_speed_mts[i]; - /* These clocks cannot come from bw_params, always fill from dcn3_21_soc[0] */ - /* PHYCLK_D18, PHYCLK_D32 */ - dcn3_21_soc.clock_limits[i].phyclk_d18_mhz = dcn3_21_soc.clock_limits[0].phyclk_d18_mhz; - dcn3_21_soc.clock_limits[i].phyclk_d32_mhz = dcn3_21_soc.clock_limits[0].phyclk_d32_mhz; - } - } else { - build_synthetic_soc_states(bw_params, dcn3_21_soc.clock_limits, &dcn3_21_soc.num_states); + /* These clocks cannot come from bw_params, always fill from dcn3_21_soc[0] */ + /* PHYCLK_D18, PHYCLK_D32 */ + dcn3_21_soc.clock_limits[i].phyclk_d18_mhz = dcn3_21_soc.clock_limits[0].phyclk_d18_mhz; + dcn3_21_soc.clock_limits[i].phyclk_d32_mhz = dcn3_21_soc.clock_limits[0].phyclk_d32_mhz; } - - /* Re-init DML with updated bb */ - dml_init_instance(&dc->dml, &dcn3_21_soc, &dcn3_21_ip, DML_PROJECT_DCN32); - if (dc->current_state) - dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_21_soc, &dcn3_21_ip, DML_PROJECT_DCN32); + } else { + build_synthetic_soc_states(dc->debug.disable_dc_mode_overwrite, bw_params, + dcn3_21_soc.clock_limits, &dcn3_21_soc.num_states); } + + /* Re-init DML with updated bb */ + dml_init_instance(&dc->dml, &dcn3_21_soc, &dcn3_21_ip, DML_PROJECT_DCN32); + if (dc->current_state) + dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_21_soc, &dcn3_21_ip, DML_PROJECT_DCN32); } diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h index 0bffae95f3a2..d5831a34f5a1 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h @@ -190,6 +190,14 @@ enum dm_validation_status { DML_FAIL_DSC_INPUT_BPC, DML_FAIL_PREFETCH_SUPPORT, DML_FAIL_V_RATIO_PREFETCH, + DML_FAIL_P2I_WITH_420, + DML_FAIL_DSC_ONLY_IF_NECESSARY_WITH_BPP, + DML_FAIL_NOT_DSC422_NATIVE, + DML_FAIL_ODM_COMBINE4TO1, + DML_FAIL_ENOUGH_WRITEBACK_UNITS, + DML_FAIL_VIEWPORT_EXCEEDS_SURFACE, + DML_FAIL_DYNAMIC_METADATA, + DML_FAIL_FMT_BUFFER_EXCEEDED, }; enum writeback_config { diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c index bdf3ac6cadd5..da0cfbb071e6 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c @@ -113,7 +113,6 @@ void dml_init_instance(struct display_mode_lib *lib, lib->funcs = dml30_funcs; break; case DML_PROJECT_DCN31: - case DML_PROJECT_DCN31_FPGA: case DML_PROJECT_DCN315: lib->funcs = dml31_funcs; break; diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h index a9d49ef58fb5..5edf69fa40d1 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h @@ -41,7 +41,6 @@ enum dml_project { DML_PROJECT_DCN30, DML_PROJECT_DCN31, DML_PROJECT_DCN315, - DML_PROJECT_DCN31_FPGA, DML_PROJECT_DCN314, DML_PROJECT_DCN32, }; diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h index 3c077164f362..ff0246a9458f 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h @@ -619,8 +619,7 @@ struct _vcs_dpi_display_dlg_regs_st { unsigned int refcyc_h_blank_end; unsigned int dlg_vblank_end; unsigned int min_dst_y_next_start; - unsigned int optimized_min_dst_y_next_start; - unsigned int optimized_min_dst_y_next_start_us; + unsigned int min_dst_y_next_start_us; unsigned int refcyc_per_htotal; unsigned int refcyc_x_after_scaler; unsigned int dst_y_after_scaler; diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c index f9653f511baa..9a3ded311195 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c @@ -571,6 +571,10 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib) mode_lib->vba.OutputLinkDPRate[mode_lib->vba.NumberOfActivePlanes] = dout->dp_rate; mode_lib->vba.ODMUse[mode_lib->vba.NumberOfActivePlanes] = dst->odm_combine_policy; mode_lib->vba.DETSizeOverride[mode_lib->vba.NumberOfActivePlanes] = src->det_size_override; + if (src->det_size_override) + mode_lib->vba.DETBufferSizeInKByte[mode_lib->vba.NumberOfActivePlanes] = src->det_size_override; + else + mode_lib->vba.DETBufferSizeInKByte[mode_lib->vba.NumberOfActivePlanes] = ip->det_buffer_size_kbytes; //TODO: Need to assign correct values to dp_multistream vars mode_lib->vba.OutputMultistreamEn[mode_lib->vba.NumberOfActiveSurfaces] = dout->dp_multistream_en; mode_lib->vba.OutputMultistreamId[mode_lib->vba.NumberOfActiveSurfaces] = dout->dp_multistream_id; @@ -785,6 +789,8 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib) mode_lib->vba.pipe_plane[k] = mode_lib->vba.NumberOfActivePlanes; mode_lib->vba.DPPPerPlane[mode_lib->vba.NumberOfActivePlanes]++; + if (src_k->det_size_override) + mode_lib->vba.DETBufferSizeInKByte[mode_lib->vba.NumberOfActivePlanes] = src_k->det_size_override; if (mode_lib->vba.SourceScan[mode_lib->vba.NumberOfActivePlanes] == dm_horz) { mode_lib->vba.ViewportWidth[mode_lib->vba.NumberOfActivePlanes] += @@ -927,18 +933,16 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib) } /** - * ******************************************************************************************** * cache_debug_params: Cache any params that needed to be maintained from the initial validation * for debug purposes. * * The DML getters can modify some of the VBA params that we are interested in (for example when * calculating with dummy p-state latency), so cache any params here that we want for debugging * - * @param [in] mode_lib: mode_lib input/output of validate call + * @mode_lib: mode_lib input/output of validate call * - * @return: void + * Return: void * - * ******************************************************************************************** */ static void cache_debug_params(struct display_mode_lib *mode_lib) { diff --git a/drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c b/drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c index 2bdc47615543..58dd62cce4bb 100644 --- a/drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c +++ b/drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c @@ -645,8 +645,6 @@ static int get_available_dsc_slices(union dsc_enc_slice_caps slice_caps, int *av { int idx = 0; - memset(available_slices, -1, MIN_AVAILABLE_SLICES_SIZE); - if (slice_caps.bits.NUM_SLICES_1) available_slices[idx++] = 1; @@ -700,7 +698,7 @@ static int inc_num_slices(union dsc_enc_slice_caps slice_caps, int num_slices) } } - if (new_num_slices == num_slices) // No biger number of slices found + if (new_num_slices == num_slices) // No bigger number of slices found new_num_slices++; return new_num_slices; @@ -952,6 +950,13 @@ static bool setup_dsc_config( else is_dsc_possible = false; } + // When we force 2:1 ODM, we can't have 1 slice to divide amongst 2 separate DSC instances + // need to enforce at minimum 2 horizontal slices + if (options->dsc_force_odm_hslice_override) { + num_slices_h = fit_num_slices_up(dsc_common_caps.slice_caps, 2); + if (num_slices_h == 0) + is_dsc_possible = false; + } if (!is_dsc_possible) goto done; @@ -1163,6 +1168,7 @@ void dc_dsc_policy_set_disable_dsc_stream_overhead(bool disable) void dc_dsc_get_default_config_option(const struct dc *dc, struct dc_dsc_config_options *options) { options->dsc_min_slice_height_override = dc->debug.dsc_min_slice_height_override; + options->dsc_force_odm_hslice_override = dc->debug.force_odm_combine; options->max_target_bpp_limit_override_x16 = 0; options->slice_height_granularity = 1; } diff --git a/drivers/gpu/drm/amd/display/dc/inc/core_types.h b/drivers/gpu/drm/amd/display/dc/inc/core_types.h index 2eb597a24425..034610b74a37 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/core_types.h +++ b/drivers/gpu/drm/amd/display/dc/inc/core_types.h @@ -37,6 +37,7 @@ #include "dwb.h" #include "mcif_wb.h" #include "panel_cntl.h" +#include "dmub/inc/dmub_cmd.h" #define MAX_CLOCK_SOURCES 7 #define MAX_SVP_PHANTOM_STREAMS 2 @@ -374,6 +375,7 @@ union pipe_update_flags { uint32_t viewport : 1; uint32_t plane_changed : 1; uint32_t det_size : 1; + uint32_t unbounded_req : 1; } bits; uint32_t raw; }; @@ -426,6 +428,8 @@ struct pipe_ctx { struct dwbc *dwbc; struct mcif_wb *mcif_wb; union pipe_update_flags update_flags; + struct tg_color visual_confirm_color; + bool has_vactive_margin; }; /* Data used for dynamic link encoder assignment. @@ -496,6 +500,11 @@ struct bw_context { struct display_mode_lib dml; }; +struct dc_dmub_cmd { + union dmub_rb_cmd dmub_cmd; + enum dm_dmub_wait_type wait_type; +}; + /** * struct dc_state - The full description of a state requested by users */ @@ -544,6 +553,11 @@ struct dc_state { */ struct bw_context bw_ctx; + struct block_sequence block_sequence[50]; + unsigned int block_sequence_steps; + struct dc_dmub_cmd dc_dmub_cmd[10]; + unsigned int dmub_cmd_count; + /** * @refcount: refcount reference * diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/abm.h b/drivers/gpu/drm/amd/display/dc/inc/hw/abm.h index ecb4191b6e64..d2190a3320f6 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/abm.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/abm.h @@ -55,6 +55,10 @@ struct abm_funcs { unsigned int bytes, unsigned int inst); bool (*set_abm_pause)(struct abm *abm, bool pause, unsigned int panel_inst, unsigned int otg_inst); + bool (*set_pipe_ex)(struct abm *abm, + unsigned int otg_inst, + unsigned int option, + unsigned int panel_inst); }; #endif diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr.h b/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr.h index bef843cc32a1..6faf40fa5c69 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr.h @@ -233,6 +233,7 @@ struct clk_bw_params { struct clk_limit_table clk_table; struct wm_table wm_table; struct dummy_pstate_entry dummy_pstate_table[4]; + struct clk_limit_table_entry dc_mode_limit; }; /* Public interfaces */ diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/dccg.h b/drivers/gpu/drm/amd/display/dc/inc/hw/dccg.h index ad6acd1b34e1..8dc804bbe98b 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/dccg.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/dccg.h @@ -68,6 +68,7 @@ struct dccg { const struct dccg_funcs *funcs; int pipe_dppclk_khz[MAX_PIPES]; int ref_dppclk; + bool dpp_clock_gated[MAX_PIPES]; //int dtbclk_khz[MAX_PIPES];/* TODO needs to be removed */ //int audio_dtbclk_khz;/* TODO needs to be removed */ //int ref_dtbclk_khz;/* TODO needs to be removed */ @@ -159,6 +160,9 @@ struct dccg_funcs { int otg_inst, int pixclk_khz); + void (*trigger_dio_fifo_resync)( + struct dccg *dccg); + void (*dpp_root_clock_control)( struct dccg *dccg, unsigned int dpp_inst, diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h b/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h index beb26dc8a07f..aaa293613846 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h @@ -111,6 +111,9 @@ struct dcn_hubbub_state { uint32_t vm_error_vmid; uint32_t vm_error_pipe; uint32_t vm_error_mode; + uint32_t test_debug_data; + uint32_t watermark_change_cntl; + uint32_t dram_state_cntl; }; struct hubbub_funcs { diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h index 88ac723d10aa..02ff99f7bec2 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h @@ -44,6 +44,138 @@ struct dc_virtual_addr_space_config; struct dpp; struct dce_hwseq; struct link_resource; +struct dc_dmub_cmd; + +struct subvp_pipe_control_lock_fast_params { + struct dc *dc; + bool lock; + struct pipe_ctx *pipe_ctx; +}; + +struct pipe_control_lock_params { + struct dc *dc; + struct pipe_ctx *pipe_ctx; + bool lock; +}; + +struct set_flip_control_gsl_params { + struct pipe_ctx *pipe_ctx; + bool flip_immediate; +}; + +struct program_triplebuffer_params { + const struct dc *dc; + struct pipe_ctx *pipe_ctx; + bool enableTripleBuffer; +}; + +struct update_plane_addr_params { + struct dc *dc; + struct pipe_ctx *pipe_ctx; +}; + +struct set_input_transfer_func_params { + struct dc *dc; + struct pipe_ctx *pipe_ctx; + struct dc_plane_state *plane_state; +}; + +struct program_gamut_remap_params { + struct pipe_ctx *pipe_ctx; +}; + +struct program_manual_trigger_params { + struct pipe_ctx *pipe_ctx; +}; + +struct send_dmcub_cmd_params { + struct dc_context *ctx; + union dmub_rb_cmd *cmd; + enum dm_dmub_wait_type wait_type; +}; + +struct setup_dpp_params { + struct pipe_ctx *pipe_ctx; +}; + +struct program_bias_and_scale_params { + struct pipe_ctx *pipe_ctx; +}; + +struct set_output_transfer_func_params { + struct dc *dc; + struct pipe_ctx *pipe_ctx; + const struct dc_stream_state *stream; +}; + +struct update_visual_confirm_params { + struct dc *dc; + struct pipe_ctx *pipe_ctx; + int mpcc_id; +}; + +struct power_on_mpc_mem_pwr_params { + struct mpc *mpc; + int mpcc_id; + bool power_on; +}; + +struct set_output_csc_params { + struct mpc *mpc; + int opp_id; + const uint16_t *regval; + enum mpc_output_csc_mode ocsc_mode; +}; + +struct set_ocsc_default_params { + struct mpc *mpc; + int opp_id; + enum dc_color_space color_space; + enum mpc_output_csc_mode ocsc_mode; +}; + +union block_sequence_params { + struct update_plane_addr_params update_plane_addr_params; + struct subvp_pipe_control_lock_fast_params subvp_pipe_control_lock_fast_params; + struct pipe_control_lock_params pipe_control_lock_params; + struct set_flip_control_gsl_params set_flip_control_gsl_params; + struct program_triplebuffer_params program_triplebuffer_params; + struct set_input_transfer_func_params set_input_transfer_func_params; + struct program_gamut_remap_params program_gamut_remap_params; + struct program_manual_trigger_params program_manual_trigger_params; + struct send_dmcub_cmd_params send_dmcub_cmd_params; + struct setup_dpp_params setup_dpp_params; + struct program_bias_and_scale_params program_bias_and_scale_params; + struct set_output_transfer_func_params set_output_transfer_func_params; + struct update_visual_confirm_params update_visual_confirm_params; + struct power_on_mpc_mem_pwr_params power_on_mpc_mem_pwr_params; + struct set_output_csc_params set_output_csc_params; + struct set_ocsc_default_params set_ocsc_default_params; +}; + +enum block_sequence_func { + DMUB_SUBVP_PIPE_CONTROL_LOCK_FAST = 0, + OPTC_PIPE_CONTROL_LOCK, + HUBP_SET_FLIP_CONTROL_GSL, + HUBP_PROGRAM_TRIPLEBUFFER, + HUBP_UPDATE_PLANE_ADDR, + DPP_SET_INPUT_TRANSFER_FUNC, + DPP_PROGRAM_GAMUT_REMAP, + OPTC_PROGRAM_MANUAL_TRIGGER, + DMUB_SEND_DMCUB_CMD, + DPP_SETUP_DPP, + DPP_PROGRAM_BIAS_AND_SCALE, + DPP_SET_OUTPUT_TRANSFER_FUNC, + MPC_UPDATE_VISUAL_CONFIRM, + MPC_POWER_ON_MPC_MEM_PWR, + MPC_SET_OUTPUT_CSC, + MPC_SET_OCSC_DEFAULT, +}; + +struct block_sequence { + union block_sequence_params params; + enum block_sequence_func func; +}; struct hw_sequencer_funcs { void (*hardware_release)(struct dc *dc); @@ -252,12 +384,12 @@ struct hw_sequencer_funcs { const struct tg_color *solid_color, int width, int height, int offset); + void (*subvp_pipe_control_lock_fast)(union block_sequence_params *params); void (*z10_restore)(const struct dc *dc); void (*z10_save_init)(struct dc *dc); void (*update_visual_confirm_color)(struct dc *dc, struct pipe_ctx *pipe_ctx, - struct tg_color *color, int mpcc_id); void (*update_phantom_vp_position)(struct dc *dc, @@ -294,6 +426,7 @@ void get_surface_visual_confirm_color( void get_subvp_visual_confirm_color( struct dc *dc, + struct dc_state *context, struct pipe_ctx *pipe_ctx, struct tg_color *color); @@ -306,4 +439,36 @@ void get_mpctree_visual_confirm_color( void get_surface_tile_visual_confirm_color( struct pipe_ctx *pipe_ctx, struct tg_color *color); + +void get_mclk_switch_visual_confirm_color( + struct dc *dc, + struct dc_state *context, + struct pipe_ctx *pipe_ctx, + struct tg_color *color); + +void hwss_execute_sequence(struct dc *dc, + struct block_sequence block_sequence[], + int num_steps); + +void hwss_build_fast_sequence(struct dc *dc, + struct dc_dmub_cmd *dc_dmub_cmd, + unsigned int dmub_cmd_count, + struct block_sequence block_sequence[], + int *num_steps, + struct pipe_ctx *pipe_ctx); + +void hwss_send_dmcub_cmd(union block_sequence_params *params); + +void hwss_program_manual_trigger(union block_sequence_params *params); + +void hwss_setup_dpp(union block_sequence_params *params); + +void hwss_program_bias_and_scale(union block_sequence_params *params); + +void hwss_power_on_mpc_mem_pwr(union block_sequence_params *params); + +void hwss_set_output_csc(union block_sequence_params *params); + +void hwss_set_ocsc_default(union block_sequence_params *params); + #endif /* __DC_HW_SEQUENCER_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer_private.h b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer_private.h index 4513544559be..a151865a3a20 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer_private.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer_private.h @@ -156,10 +156,12 @@ struct hwseq_private_funcs { void (*program_mall_pipe_config)(struct dc *dc, struct dc_state *context); void (*update_force_pstate)(struct dc *dc, struct dc_state *context); void (*update_mall_sel)(struct dc *dc, struct dc_state *context); - unsigned int (*calculate_dccg_k1_k2_values)(struct pipe_ctx *pipe_ctx, + void (*calculate_dccg_k1_k2_values)(struct pipe_ctx *pipe_ctx, unsigned int *k1_div, unsigned int *k2_div); void (*set_pixels_per_cycle)(struct pipe_ctx *pipe_ctx); + void (*resync_fifo_dccg_dio)(struct dce_hwseq *hws, struct dc *dc, + struct dc_state *context); bool (*is_dp_dig_pixel_rate_div_policy)(struct pipe_ctx *pipe_ctx); #endif }; diff --git a/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_hpo_dp.c b/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_hpo_dp.c index edd7d026a762..586fe25c1702 100644 --- a/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_hpo_dp.c +++ b/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_hpo_dp.c @@ -28,24 +28,6 @@ #include "dccg.h" #include "clk_mgr.h" -static enum phyd32clk_clock_source get_phyd32clk_src(struct dc_link *link) -{ - switch (link->link_enc->transmitter) { - case TRANSMITTER_UNIPHY_A: - return PHYD32CLKA; - case TRANSMITTER_UNIPHY_B: - return PHYD32CLKB; - case TRANSMITTER_UNIPHY_C: - return PHYD32CLKC; - case TRANSMITTER_UNIPHY_D: - return PHYD32CLKD; - case TRANSMITTER_UNIPHY_E: - return PHYD32CLKE; - default: - return PHYD32CLKA; - } -} - static void set_hpo_dp_throttled_vcp_size(struct pipe_ctx *pipe_ctx, struct fixed31_32 throttled_vcp_size) { @@ -120,81 +102,26 @@ static void setup_hpo_dp_stream_attribute(struct pipe_ctx *pipe_ctx) DPCD_SOURCE_SEQ_AFTER_DP_STREAM_ATTR); } -static void enable_hpo_dp_fpga_link_output(struct dc_link *link, - const struct link_resource *link_res, - enum signal_type signal, - enum clock_source_id clock_source, - const struct dc_link_settings *link_settings) -{ - const struct dc *dc = link->dc; - enum phyd32clk_clock_source phyd32clk = get_phyd32clk_src(link); - int phyd32clk_freq_khz = link_settings->link_rate == LINK_RATE_UHBR10 ? 312500 : - link_settings->link_rate == LINK_RATE_UHBR13_5 ? 412875 : - link_settings->link_rate == LINK_RATE_UHBR20 ? 625000 : 0; - - dm_set_phyd32clk(dc->ctx, phyd32clk_freq_khz); - dc->res_pool->dccg->funcs->set_physymclk( - dc->res_pool->dccg, - link->link_enc_hw_inst, - PHYSYMCLK_FORCE_SRC_PHYD32CLK, - true); - dc->res_pool->dccg->funcs->enable_symclk32_le( - dc->res_pool->dccg, - link_res->hpo_dp_link_enc->inst, - phyd32clk); - link_res->hpo_dp_link_enc->funcs->link_enable( - link_res->hpo_dp_link_enc, - link_settings->lane_count); - -} - static void enable_hpo_dp_link_output(struct dc_link *link, const struct link_resource *link_res, enum signal_type signal, enum clock_source_id clock_source, const struct dc_link_settings *link_settings) { - if (IS_FPGA_MAXIMUS_DC(link->dc->ctx->dce_environment)) - enable_hpo_dp_fpga_link_output(link, link_res, signal, - clock_source, link_settings); - else - link_res->hpo_dp_link_enc->funcs->enable_link_phy( - link_res->hpo_dp_link_enc, - link_settings, - link->link_enc->transmitter, - link->link_enc->hpd_source); -} - - -static void disable_hpo_dp_fpga_link_output(struct dc_link *link, - const struct link_resource *link_res, - enum signal_type signal) -{ - const struct dc *dc = link->dc; - - link_res->hpo_dp_link_enc->funcs->link_disable(link_res->hpo_dp_link_enc); - dc->res_pool->dccg->funcs->disable_symclk32_le( - dc->res_pool->dccg, - link_res->hpo_dp_link_enc->inst); - dc->res_pool->dccg->funcs->set_physymclk( - dc->res_pool->dccg, - link->link_enc_hw_inst, - PHYSYMCLK_FORCE_SRC_SYMCLK, - false); - dm_set_phyd32clk(dc->ctx, 0); + link_res->hpo_dp_link_enc->funcs->enable_link_phy( + link_res->hpo_dp_link_enc, + link_settings, + link->link_enc->transmitter, + link->link_enc->hpd_source); } static void disable_hpo_dp_link_output(struct dc_link *link, const struct link_resource *link_res, enum signal_type signal) { - if (IS_FPGA_MAXIMUS_DC(link->dc->ctx->dce_environment)) { - disable_hpo_dp_fpga_link_output(link, link_res, signal); - } else { link_res->hpo_dp_link_enc->funcs->link_disable(link_res->hpo_dp_link_enc); link_res->hpo_dp_link_enc->funcs->disable_link_phy( link_res->hpo_dp_link_enc, signal); - } } static void set_hpo_dp_link_test_pattern(struct dc_link *link, diff --git a/drivers/gpu/drm/amd/display/dc/link/link_detection.c b/drivers/gpu/drm/amd/display/dc/link/link_detection.c index a131e30fd7d6..8041b8369e45 100644 --- a/drivers/gpu/drm/amd/display/dc/link/link_detection.c +++ b/drivers/gpu/drm/amd/display/dc/link/link_detection.c @@ -593,6 +593,10 @@ static bool detect_dp(struct dc_link *link, /* DP SST branch */ link->type = dc_connection_sst_branch; } else { + if (link->dc->debug.disable_dp_plus_plus_wa && + link->link_enc->features.flags.bits.IS_UHBR20_CAPABLE) + return false; + /* DP passive dongles */ sink_caps->signal = dp_passive_dongle_detection(link->ddc, sink_caps, @@ -980,6 +984,11 @@ static bool detect_link_and_local_sink(struct dc_link *link, (link->dpcd_caps.dongle_type != DISPLAY_DONGLE_DP_HDMI_CONVERTER)) converter_disable_audio = true; + + /* limited link rate to HBR3 for DPIA until we implement USB4 V2 */ + if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA && + link->reported_link_cap.link_rate > LINK_RATE_HIGH3) + link->reported_link_cap.link_rate = LINK_RATE_HIGH3; break; } diff --git a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c index 2267fb097830..1a7b93e41e35 100644 --- a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c +++ b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c @@ -765,7 +765,7 @@ static bool dp_set_dsc_on_rx(struct pipe_ctx *pipe_ctx, bool enable) struct dc_stream_state *stream = pipe_ctx->stream; bool result = false; - if (dc_is_virtual_signal(stream->signal) || IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) + if (dc_is_virtual_signal(stream->signal)) result = true; else result = dm_helpers_dp_write_dsc_enable(dc->ctx, stream, enable); @@ -778,7 +778,6 @@ static bool dp_set_dsc_on_rx(struct pipe_ctx *pipe_ctx, bool enable) void link_set_dsc_on_stream(struct pipe_ctx *pipe_ctx, bool enable) { struct display_stream_compressor *dsc = pipe_ctx->stream_res.dsc; - struct dc *dc = pipe_ctx->stream->ctx->dc; struct dc_stream_state *stream = pipe_ctx->stream; struct pipe_ctx *odm_pipe; int opp_cnt = 1; @@ -816,8 +815,7 @@ void link_set_dsc_on_stream(struct pipe_ctx *pipe_ctx, bool enable) optc_dsc_mode = dsc_optc_cfg.is_pixel_format_444 ? OPTC_DSC_ENABLED_444 : OPTC_DSC_ENABLED_NATIVE_SUBSAMPLED; /* Enable DSC in encoder */ - if (dc_is_dp_signal(stream->signal) && !IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) - && !dp_is_128b_132b_signal(pipe_ctx)) { + if (dc_is_dp_signal(stream->signal) && !dp_is_128b_132b_signal(pipe_ctx)) { DC_LOG_DSC("Setting stream encoder DSC config for engine %d:", (int)pipe_ctx->stream_res.stream_enc->id); dsc_optc_config_log(dsc, &dsc_optc_cfg); pipe_ctx->stream_res.stream_enc->funcs->dp_set_dsc_config(pipe_ctx->stream_res.stream_enc, @@ -849,7 +847,7 @@ void link_set_dsc_on_stream(struct pipe_ctx *pipe_ctx, bool enable) false, NULL, true); - else if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { + else { pipe_ctx->stream_res.stream_enc->funcs->dp_set_dsc_config( pipe_ctx->stream_res.stream_enc, OPTC_DSC_DISABLED, 0, 0); @@ -2209,9 +2207,8 @@ static enum dc_status enable_link( * link settings. Need to call disable first before enabling at * new link settings. */ - if (link->link_status.link_active) { + if (link->link_status.link_active && !stream->skip_edp_power_down) disable_link(link, &pipe_ctx->link_res, pipe_ctx->stream->signal); - } switch (pipe_ctx->stream->signal) { case SIGNAL_TYPE_DISPLAY_PORT: @@ -2271,8 +2268,7 @@ void link_set_dpms_off(struct pipe_ctx *pipe_ctx) } } - if (!IS_DIAG_DC(dc->ctx->dce_environment) && - dc_is_virtual_signal(pipe_ctx->stream->signal)) + if (dc_is_virtual_signal(pipe_ctx->stream->signal)) return; if (!pipe_ctx->stream->sink->edid_caps.panel_patch.skip_avmute) { @@ -2330,7 +2326,9 @@ void link_set_dpms_off(struct pipe_ctx *pipe_ctx) dc->hwss.disable_stream(pipe_ctx); } else { dc->hwss.disable_stream(pipe_ctx); - disable_link(pipe_ctx->stream->link, &pipe_ctx->link_res, pipe_ctx->stream->signal); + if (!pipe_ctx->stream->skip_edp_power_down) { + disable_link(pipe_ctx->stream->link, &pipe_ctx->link_res, pipe_ctx->stream->signal); + } } if (pipe_ctx->stream->timing.flags.DSC) { @@ -2358,6 +2356,8 @@ void link_set_dpms_on( enum otg_out_mux_dest otg_out_dest = OUT_MUX_DIO; struct vpg *vpg = pipe_ctx->stream_res.stream_enc->vpg; const struct link_hwss *link_hwss = get_link_hwss(link, &pipe_ctx->link_res); + bool apply_edp_fast_boot_optimization = + pipe_ctx->stream->apply_edp_fast_boot_optimization; ASSERT(is_master_pipe_for_link(link, pipe_ctx)); @@ -2375,8 +2375,7 @@ void link_set_dpms_on( } } - if (!IS_DIAG_DC(dc->ctx->dce_environment) && - dc_is_virtual_signal(pipe_ctx->stream->signal)) + if (dc_is_virtual_signal(pipe_ctx->stream->signal)) return; link_enc = link_enc_cfg_get_link_enc(link); @@ -2402,138 +2401,126 @@ void link_set_dpms_on( link_hwss->setup_stream_attribute(pipe_ctx); - if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - bool apply_edp_fast_boot_optimization = - pipe_ctx->stream->apply_edp_fast_boot_optimization; - - pipe_ctx->stream->apply_edp_fast_boot_optimization = false; + pipe_ctx->stream->apply_edp_fast_boot_optimization = false; - // Enable VPG before building infoframe - if (vpg && vpg->funcs->vpg_poweron) - vpg->funcs->vpg_poweron(vpg); + // Enable VPG before building infoframe + if (vpg && vpg->funcs->vpg_poweron) + vpg->funcs->vpg_poweron(vpg); - resource_build_info_frame(pipe_ctx); - dc->hwss.update_info_frame(pipe_ctx); + resource_build_info_frame(pipe_ctx); + dc->hwss.update_info_frame(pipe_ctx); - if (dc_is_dp_signal(pipe_ctx->stream->signal)) - dp_trace_source_sequence(link, DPCD_SOURCE_SEQ_AFTER_UPDATE_INFO_FRAME); - - /* Do not touch link on seamless boot optimization. */ - if (pipe_ctx->stream->apply_seamless_boot_optimization) { - pipe_ctx->stream->dpms_off = false; + if (dc_is_dp_signal(pipe_ctx->stream->signal)) + dp_trace_source_sequence(link, DPCD_SOURCE_SEQ_AFTER_UPDATE_INFO_FRAME); - /* Still enable stream features & audio on seamless boot for DP external displays */ - if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT) { - enable_stream_features(pipe_ctx); - dc->hwss.enable_audio_stream(pipe_ctx); - } + /* Do not touch link on seamless boot optimization. */ + if (pipe_ctx->stream->apply_seamless_boot_optimization) { + pipe_ctx->stream->dpms_off = false; - update_psp_stream_config(pipe_ctx, false); - return; - } - - /* eDP lit up by bios already, no need to enable again. */ - if (pipe_ctx->stream->signal == SIGNAL_TYPE_EDP && - apply_edp_fast_boot_optimization && - !pipe_ctx->stream->timing.flags.DSC && - !pipe_ctx->next_odm_pipe) { - pipe_ctx->stream->dpms_off = false; - update_psp_stream_config(pipe_ctx, false); - return; + /* Still enable stream features & audio on seamless boot for DP external displays */ + if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT) { + enable_stream_features(pipe_ctx); + dc->hwss.enable_audio_stream(pipe_ctx); } - if (pipe_ctx->stream->dpms_off) - return; + update_psp_stream_config(pipe_ctx, false); + return; + } - /* Have to setup DSC before DIG FE and BE are connected (which happens before the - * link training). This is to make sure the bandwidth sent to DIG BE won't be - * bigger than what the link and/or DIG BE can handle. VBID[6]/CompressedStream_flag - * will be automatically set at a later time when the video is enabled - * (DP_VID_STREAM_EN = 1). - */ - if (pipe_ctx->stream->timing.flags.DSC) { - if (dc_is_dp_signal(pipe_ctx->stream->signal) || - dc_is_virtual_signal(pipe_ctx->stream->signal)) - link_set_dsc_enable(pipe_ctx, true); + /* eDP lit up by bios already, no need to enable again. */ + if (pipe_ctx->stream->signal == SIGNAL_TYPE_EDP && + apply_edp_fast_boot_optimization && + !pipe_ctx->stream->timing.flags.DSC && + !pipe_ctx->next_odm_pipe) { + pipe_ctx->stream->dpms_off = false; + update_psp_stream_config(pipe_ctx, false); + return; + } - } + if (pipe_ctx->stream->dpms_off) + return; - status = enable_link(state, pipe_ctx); + /* Have to setup DSC before DIG FE and BE are connected (which happens before the + * link training). This is to make sure the bandwidth sent to DIG BE won't be + * bigger than what the link and/or DIG BE can handle. VBID[6]/CompressedStream_flag + * will be automatically set at a later time when the video is enabled + * (DP_VID_STREAM_EN = 1). + */ + if (pipe_ctx->stream->timing.flags.DSC) { + if (dc_is_dp_signal(pipe_ctx->stream->signal) || + dc_is_virtual_signal(pipe_ctx->stream->signal)) + link_set_dsc_enable(pipe_ctx, true); - if (status != DC_OK) { - DC_LOG_WARNING("enabling link %u failed: %d\n", - pipe_ctx->stream->link->link_index, - status); + } - /* Abort stream enable *unless* the failure was due to - * DP link training - some DP monitors will recover and - * show the stream anyway. But MST displays can't proceed - * without link training. - */ - if (status != DC_FAIL_DP_LINK_TRAINING || - pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) { - if (false == stream->link->link_status.link_active) - disable_link(stream->link, &pipe_ctx->link_res, - pipe_ctx->stream->signal); - BREAK_TO_DEBUGGER(); - return; - } - } + status = enable_link(state, pipe_ctx); - /* turn off otg test pattern if enable */ - if (pipe_ctx->stream_res.tg->funcs->set_test_pattern) - pipe_ctx->stream_res.tg->funcs->set_test_pattern(pipe_ctx->stream_res.tg, - CONTROLLER_DP_TEST_PATTERN_VIDEOMODE, - COLOR_DEPTH_UNDEFINED); + if (status != DC_OK) { + DC_LOG_WARNING("enabling link %u failed: %d\n", + pipe_ctx->stream->link->link_index, + status); - /* This second call is needed to reconfigure the DIG - * as a workaround for the incorrect value being applied - * from transmitter control. + /* Abort stream enable *unless* the failure was due to + * DP link training - some DP monitors will recover and + * show the stream anyway. But MST displays can't proceed + * without link training. */ - if (!(dc_is_virtual_signal(pipe_ctx->stream->signal) || - dp_is_128b_132b_signal(pipe_ctx))) { - if (link_enc) - link_enc->funcs->setup( - link_enc, + if (status != DC_FAIL_DP_LINK_TRAINING || + pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) { + if (false == stream->link->link_status.link_active) + disable_link(stream->link, &pipe_ctx->link_res, pipe_ctx->stream->signal); - } + BREAK_TO_DEBUGGER(); + return; + } + } - dc->hwss.enable_stream(pipe_ctx); + /* turn off otg test pattern if enable */ + if (pipe_ctx->stream_res.tg->funcs->set_test_pattern) + pipe_ctx->stream_res.tg->funcs->set_test_pattern(pipe_ctx->stream_res.tg, + CONTROLLER_DP_TEST_PATTERN_VIDEOMODE, + COLOR_DEPTH_UNDEFINED); - /* Set DPS PPS SDP (AKA "info frames") */ - if (pipe_ctx->stream->timing.flags.DSC) { - if (dc_is_dp_signal(pipe_ctx->stream->signal) || - dc_is_virtual_signal(pipe_ctx->stream->signal)) { - dp_set_dsc_on_rx(pipe_ctx, true); - link_set_dsc_pps_packet(pipe_ctx, true, true); - } + /* This second call is needed to reconfigure the DIG + * as a workaround for the incorrect value being applied + * from transmitter control. + */ + if (!(dc_is_virtual_signal(pipe_ctx->stream->signal) || + dp_is_128b_132b_signal(pipe_ctx))) { + if (link_enc) + link_enc->funcs->setup( + link_enc, + pipe_ctx->stream->signal); } - if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) - allocate_mst_payload(pipe_ctx); - else if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT && - dp_is_128b_132b_signal(pipe_ctx)) - update_sst_payload(pipe_ctx, true); + dc->hwss.enable_stream(pipe_ctx); + + /* Set DPS PPS SDP (AKA "info frames") */ + if (pipe_ctx->stream->timing.flags.DSC) { + if (dc_is_dp_signal(pipe_ctx->stream->signal) || + dc_is_virtual_signal(pipe_ctx->stream->signal)) { + dp_set_dsc_on_rx(pipe_ctx, true); + link_set_dsc_pps_packet(pipe_ctx, true, true); + } + } - dc->hwss.unblank_stream(pipe_ctx, - &pipe_ctx->stream->link->cur_link_settings); + if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) + allocate_mst_payload(pipe_ctx); + else if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT && + dp_is_128b_132b_signal(pipe_ctx)) + update_sst_payload(pipe_ctx, true); - if (stream->sink_patches.delay_ignore_msa > 0) - msleep(stream->sink_patches.delay_ignore_msa); + dc->hwss.unblank_stream(pipe_ctx, + &pipe_ctx->stream->link->cur_link_settings); - if (dc_is_dp_signal(pipe_ctx->stream->signal)) - enable_stream_features(pipe_ctx); - update_psp_stream_config(pipe_ctx, false); + if (stream->sink_patches.delay_ignore_msa > 0) + msleep(stream->sink_patches.delay_ignore_msa); - dc->hwss.enable_audio_stream(pipe_ctx); + if (dc_is_dp_signal(pipe_ctx->stream->signal)) + enable_stream_features(pipe_ctx); + update_psp_stream_config(pipe_ctx, false); - } else { // if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) - if (dp_is_128b_132b_signal(pipe_ctx)) - dp_fpga_hpo_enable_link_and_stream(state, pipe_ctx); - if (dc_is_dp_signal(pipe_ctx->stream->signal) || - dc_is_virtual_signal(pipe_ctx->stream->signal)) - link_set_dsc_enable(pipe_ctx, true); - } + dc->hwss.enable_audio_stream(pipe_ctx); if (dc_is_hdmi_signal(pipe_ctx->stream->signal)) { set_avmute(pipe_ctx, false); diff --git a/drivers/gpu/drm/amd/display/dc/link/link_factory.c b/drivers/gpu/drm/amd/display/dc/link/link_factory.c index 1515c817f03b..ac1c3e2e7c1d 100644 --- a/drivers/gpu/drm/amd/display/dc/link/link_factory.c +++ b/drivers/gpu/drm/amd/display/dc/link/link_factory.c @@ -563,11 +563,9 @@ static bool construct_phy(struct dc_link *link, goto create_fail; } - /* TODO: #DAL3 Implement id to str function.*/ - LINK_INFO("Connector[%d] description:" - "signal %d\n", + LINK_INFO("Connector[%d] description: signal: %s\n", init_params->connector_index, - link->connector_signal); + signal_type_to_string(link->connector_signal)); ddc_service_init_data.ctx = link->ctx; ddc_service_init_data.id = link->link_id; diff --git a/drivers/gpu/drm/amd/display/dc/link/link_validation.c b/drivers/gpu/drm/amd/display/dc/link/link_validation.c index d4b7da526f0a..e8b2fc4002a5 100644 --- a/drivers/gpu/drm/amd/display/dc/link/link_validation.c +++ b/drivers/gpu/drm/amd/display/dc/link/link_validation.c @@ -359,5 +359,8 @@ bool link_validate_dpia_bandwidth(const struct dc_stream_state *stream, const un link[i] = stream[i].link; bw_needed[i] = dc_bandwidth_in_kbps_from_timing(&stream[i].timing); } + + ret = dpia_validate_usb4_bw(link, bw_needed, num_streams); + return ret; } diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c index ba98013fecd0..3a5e80b57711 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c @@ -326,8 +326,7 @@ bool dp_is_fec_supported(const struct dc_link *link) return (dc_is_dp_signal(link->connector_signal) && link_enc && link_enc->features.fec_supported && - link->dpcd_caps.fec_cap.bits.FEC_CAPABLE && - !IS_FPGA_MAXIMUS_DC(link->ctx->dce_environment)); + link->dpcd_caps.fec_cap.bits.FEC_CAPABLE); } bool dp_should_enable_fec(const struct dc_link *link) @@ -1043,9 +1042,7 @@ static enum dc_status wake_up_aux_channel(struct dc_link *link) DP_SET_POWER, &dpcd_power_state, sizeof(dpcd_power_state)); - if (status < 0) - DC_LOG_DC("%s: Failed to power up sink: %s\n", __func__, - dpcd_power_state == DP_SET_POWER_D0 ? "D0" : "D3"); + DC_LOG_DC("%s: Failed to power up sink\n", __func__); return DC_ERROR_UNEXPECTED; } @@ -1396,7 +1393,7 @@ static bool get_usbc_cable_id(struct dc_link *link, union dp_cable_id *cable_id) cmd.cable_id.header.payload_bytes = sizeof(cmd.cable_id.data); cmd.cable_id.data.input.phy_inst = resource_transmitter_to_phy_idx( link->dc, link->link_enc->transmitter); - if (dc_dmub_srv_cmd_with_reply_data(link->ctx->dmub_srv, &cmd) && + if (dm_execute_dmub_cmd(link->dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY) && cmd.cable_id.header.ret_status == 1) { cable_id->raw = cmd.cable_id.data.output_raw; DC_LOG_DC("usbc_cable_id = %d.\n", cable_id->raw); @@ -1452,7 +1449,8 @@ bool read_is_mst_supported(struct dc_link *link) */ static bool dpcd_read_sink_ext_caps(struct dc_link *link) { - uint8_t dpcd_data; + uint8_t dpcd_data = 0; + uint8_t edp_general_cap2 = 0; if (!link) return false; @@ -1461,6 +1459,12 @@ static bool dpcd_read_sink_ext_caps(struct dc_link *link) return false; link->dpcd_sink_ext_caps.raw = dpcd_data; + + if (core_link_read_dpcd(link, DP_EDP_GENERAL_CAP_2, &edp_general_cap2, 1) != DC_OK) + return false; + + link->dpcd_caps.panel_luminance_control = (edp_general_cap2 & DP_EDP_PANEL_LUMINANCE_CONTROL_CAPABLE) != 0; + return true; } @@ -1554,6 +1558,9 @@ static bool retrieve_link_cap(struct dc_link *link) int i; struct dp_sink_hw_fw_revision dp_hw_fw_revision; const uint32_t post_oui_delay = 30; // 30ms + bool is_fec_supported = false; + bool is_dsc_basic_supported = false; + bool is_dsc_passthrough_supported = false; memset(dpcd_data, '\0', sizeof(dpcd_data)); memset(&down_strm_port_count, @@ -1696,6 +1703,7 @@ static bool retrieve_link_cap(struct dc_link *link) /* TODO - decouple raw mst capability from policy decision */ link->dpcd_caps.is_mst_capable = read_is_mst_supported(link); + DC_LOG_DC("%s: MST_Support: %s\n", __func__, str_yes_no(link->dpcd_caps.is_mst_capable)); get_active_converter_info(ds_port.byte, link); @@ -1803,6 +1811,17 @@ static bool retrieve_link_cap(struct dc_link *link) DP_DSC_SUPPORT, link->dpcd_caps.dsc_caps.dsc_basic_caps.raw, sizeof(link->dpcd_caps.dsc_caps.dsc_basic_caps.raw)); + if (status == DC_OK) { + is_fec_supported = link->dpcd_caps.fec_cap.bits.FEC_CAPABLE; + is_dsc_basic_supported = link->dpcd_caps.dsc_caps.dsc_basic_caps.fields.dsc_support.DSC_SUPPORT; + is_dsc_passthrough_supported = link->dpcd_caps.dsc_caps.dsc_basic_caps.fields.dsc_support.DSC_PASSTHROUGH_SUPPORT; + DC_LOG_DC("%s: FEC_Sink_Support: %s\n", __func__, + str_yes_no(is_fec_supported)); + DC_LOG_DC("%s: DSC_Basic_Sink_Support: %s\n", __func__, + str_yes_no(is_dsc_basic_supported)); + DC_LOG_DC("%s: DSC_Passthrough_Sink_Support: %s\n", __func__, + str_yes_no(is_dsc_passthrough_supported)); + } if (link->dpcd_caps.dongle_type != DISPLAY_DONGLE_NONE) { status = core_link_read_dpcd( link, @@ -1931,6 +1950,9 @@ void detect_edp_sink_caps(struct dc_link *link) link_rate_in_khz = (supported_link_rates[entry+1] * 0x100 + supported_link_rates[entry]) * 200; + DC_LOG_DC("%s: eDP v1.4 supported sink rates: [%d] %d kHz\n", __func__, + entry / 2, link_rate_in_khz); + if (link_rate_in_khz != 0) { link_rate = linkRateInKHzToLinkRateMultiplier(link_rate_in_khz); link->dpcd_caps.edp_supported_link_rates[link->dpcd_caps.edp_supported_link_rates_count] = link_rate; diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.c index 4626fabc0a96..0bb749133909 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.c @@ -90,7 +90,7 @@ bool dpia_query_hpd_status(struct dc_link *link) cmd.query_hpd.data.ch_type = AUX_CHANNEL_DPIA; /* Return HPD status reported by DMUB if query successfully executed. */ - if (dc_dmub_srv_cmd_with_reply_data(dmub_srv, &cmd) && cmd.query_hpd.data.status == AUX_RET_SUCCESS) + if (dm_execute_dmub_cmd(dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY) && cmd.query_hpd.data.status == AUX_RET_SUCCESS) is_hpd_high = cmd.query_hpd.data.result; DC_LOG_DEBUG("%s: link(%d) dpia(%d) cmd_status(%d) result(%d)\n", diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c index 579fa222810d..e011df4bdaf2 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c @@ -1653,10 +1653,19 @@ bool perform_link_training_with_retries( break; } - DC_LOG_WARNING("%s: Link(%d) training attempt %u of %d failed @ rate(%d) x lane(%d) @ spread = %x : fail reason:(%d)\n", - __func__, link->link_index, (unsigned int)j + 1, attempts, - cur_link_settings.link_rate, cur_link_settings.lane_count, - cur_link_settings.link_spread, status); + if (j == (attempts - 1)) { + DC_LOG_WARNING( + "%s: Link(%d) training attempt %u of %d failed @ rate(%d) x lane(%d) @ spread = %x : fail reason:(%d)\n", + __func__, link->link_index, (unsigned int)j + 1, attempts, + cur_link_settings.link_rate, cur_link_settings.lane_count, + cur_link_settings.link_spread, status); + } else { + DC_LOG_HW_LINK_TRAINING( + "%s: Link(%d) training attempt %u of %d failed @ rate(%d) x lane(%d) @ spread = %x : fail reason:(%d)\n", + __func__, link->link_index, (unsigned int)j + 1, attempts, + cur_link_settings.link_rate, cur_link_settings.lane_count, + cur_link_settings.link_spread, status); + } dp_disable_link_phy(link, &pipe_ctx->link_res, signal); diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_128b_132b.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_128b_132b.c index 23d380f09a21..db87cfe37b5c 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_128b_132b.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_128b_132b.c @@ -211,11 +211,17 @@ enum link_training_result dp_perform_128b_132b_link_training( dpcd_set_link_settings(link, lt_settings); - if (result == LINK_TRAINING_SUCCESS) + if (result == LINK_TRAINING_SUCCESS) { result = dp_perform_128b_132b_channel_eq_done_sequence(link, link_res, lt_settings); + if (result == LINK_TRAINING_SUCCESS) + DC_LOG_HW_LINK_TRAINING("%s: Channel EQ done.\n", __func__); + } - if (result == LINK_TRAINING_SUCCESS) + if (result == LINK_TRAINING_SUCCESS) { result = dp_perform_128b_132b_cds_done_sequence(link, link_res, lt_settings); + if (result == LINK_TRAINING_SUCCESS) + DC_LOG_HW_LINK_TRAINING("%s: CDS done.\n", __func__); + } return result; } diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c index 3889ebb2256b..2b4c15b0b407 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c @@ -388,6 +388,8 @@ enum link_training_result dp_perform_8b_10b_link_training( link_res, lt_settings, repeater_id); + if (status == LINK_TRAINING_SUCCESS) + DC_LOG_HW_LINK_TRAINING("%s: Channel EQ done.\n", __func__); repeater_training_done(link, repeater_id); @@ -409,6 +411,8 @@ enum link_training_result dp_perform_8b_10b_link_training( link_res, lt_settings, DPRX); + if (status == LINK_TRAINING_SUCCESS) + DC_LOG_HW_LINK_TRAINING("%s: Channel EQ done.\n", __func__); } } diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_fixed_vs_pe_retimer.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_fixed_vs_pe_retimer.c index 5731c4b61f9f..15faaf645b14 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_fixed_vs_pe_retimer.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_fixed_vs_pe_retimer.c @@ -233,7 +233,7 @@ enum link_training_result dp_perform_fixed_vs_pe_training_sequence_legacy( link->dpcd_caps.lttpr_caps.phy_repeater_cnt); const uint8_t vendor_lttpr_write_data_intercept_en[4] = {0x1, 0x55, 0x63, 0x0}; const uint8_t vendor_lttpr_write_data_intercept_dis[4] = {0x1, 0x55, 0x63, 0x68}; - uint32_t pre_disable_intercept_delay_ms = link->dc->debug.fixed_vs_aux_delay_config_wa; + uint32_t pre_disable_intercept_delay_ms = 0; uint8_t vendor_lttpr_write_data_vs[4] = {0x1, 0x51, 0x63, 0x0}; uint8_t vendor_lttpr_write_data_pe[4] = {0x1, 0x52, 0x63, 0x0}; uint32_t vendor_lttpr_write_address = 0xF004F; @@ -244,6 +244,10 @@ enum link_training_result dp_perform_fixed_vs_pe_training_sequence_legacy( uint8_t toggle_rate; uint8_t rate; + if (link->local_sink) + pre_disable_intercept_delay_ms = + link->local_sink->edid_caps.panel_patch.delay_disable_aux_intercept_ms; + /* Only 8b/10b is supported */ ASSERT(link_dp_get_encoding_format(<_settings->link_settings) == DP_8b_10b_ENCODING); @@ -259,7 +263,7 @@ enum link_training_result dp_perform_fixed_vs_pe_training_sequence_legacy( /* Certain display and cable configuration require extra delay */ if (offset > 2) - pre_disable_intercept_delay_ms = link->dc->debug.fixed_vs_aux_delay_config_wa * 2; + pre_disable_intercept_delay_ms = pre_disable_intercept_delay_ms * 2; } /* Vendor specific: Reset lane settings */ @@ -380,7 +384,8 @@ enum link_training_result dp_perform_fixed_vs_pe_training_sequence_legacy( 0); /* Vendor specific: Disable intercept */ for (i = 0; i < max_vendor_dpcd_retries; i++) { - msleep(pre_disable_intercept_delay_ms); + if (pre_disable_intercept_delay_ms != 0) + msleep(pre_disable_intercept_delay_ms); dpcd_status = core_link_write_dpcd( link, vendor_lttpr_write_address, @@ -591,10 +596,9 @@ enum link_training_result dp_perform_fixed_vs_pe_training_sequence( const uint8_t vendor_lttpr_write_data_adicora_eq1[4] = {0x1, 0x55, 0x63, 0x2E}; const uint8_t vendor_lttpr_write_data_adicora_eq2[4] = {0x1, 0x55, 0x63, 0x01}; const uint8_t vendor_lttpr_write_data_adicora_eq3[4] = {0x1, 0x55, 0x63, 0x68}; - uint32_t pre_disable_intercept_delay_ms = link->dc->debug.fixed_vs_aux_delay_config_wa; uint8_t vendor_lttpr_write_data_vs[4] = {0x1, 0x51, 0x63, 0x0}; uint8_t vendor_lttpr_write_data_pe[4] = {0x1, 0x52, 0x63, 0x0}; - + uint32_t pre_disable_intercept_delay_ms = 0; uint32_t vendor_lttpr_write_address = 0xF004F; enum link_training_result status = LINK_TRAINING_SUCCESS; uint8_t lane = 0; @@ -603,6 +607,10 @@ enum link_training_result dp_perform_fixed_vs_pe_training_sequence( uint8_t toggle_rate; uint8_t rate; + if (link->local_sink) + pre_disable_intercept_delay_ms = + link->local_sink->edid_caps.panel_patch.delay_disable_aux_intercept_ms; + /* Only 8b/10b is supported */ ASSERT(link_dp_get_encoding_format(<_settings->link_settings) == DP_8b_10b_ENCODING); @@ -618,7 +626,7 @@ enum link_training_result dp_perform_fixed_vs_pe_training_sequence( /* Certain display and cable configuration require extra delay */ if (offset > 2) - pre_disable_intercept_delay_ms = link->dc->debug.fixed_vs_aux_delay_config_wa * 2; + pre_disable_intercept_delay_ms = pre_disable_intercept_delay_ms * 2; } /* Vendor specific: Reset lane settings */ @@ -739,7 +747,8 @@ enum link_training_result dp_perform_fixed_vs_pe_training_sequence( 0); /* Vendor specific: Disable intercept */ for (i = 0; i < max_vendor_dpcd_retries; i++) { - msleep(pre_disable_intercept_delay_ms); + if (pre_disable_intercept_delay_ms != 0) + msleep(pre_disable_intercept_delay_ms); dpcd_status = core_link_write_dpcd( link, vendor_lttpr_write_address, diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c index 8d1df863659c..2039a345f23a 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c @@ -165,14 +165,35 @@ bool edp_set_backlight_level_nits(struct dc_link *link, *(uint16_t *)&dpcd_backlight_set.backlight_transition_time_ms = (uint16_t)transition_time_in_ms; - if (core_link_write_dpcd(link, DP_SOURCE_BACKLIGHT_LEVEL, + if (!link->dpcd_caps.panel_luminance_control) { + if (core_link_write_dpcd(link, DP_SOURCE_BACKLIGHT_LEVEL, (uint8_t *)(&dpcd_backlight_set), sizeof(dpcd_backlight_set)) != DC_OK) - return false; + return false; - if (core_link_write_dpcd(link, DP_SOURCE_BACKLIGHT_CONTROL, + if (core_link_write_dpcd(link, DP_SOURCE_BACKLIGHT_CONTROL, &backlight_control, 1) != DC_OK) - return false; + return false; + } else { + const uint8_t backlight_enable = DP_EDP_PANEL_LUMINANCE_CONTROL_ENABLE; + struct target_luminance_value *target_luminance = NULL; + + //if target luminance value is greater than 24 bits, clip the value to 24 bits + if (backlight_millinits > 0xFFFFFF) + backlight_millinits = 0xFFFFFF; + + target_luminance = (struct target_luminance_value *)&backlight_millinits; + + if (core_link_write_dpcd(link, DP_EDP_BACKLIGHT_MODE_SET_REGISTER, + &backlight_enable, + sizeof(backlight_enable)) != DC_OK) + return false; + + if (core_link_write_dpcd(link, DP_EDP_PANEL_TARGET_LUMINANCE_VALUE, + (uint8_t *)(target_luminance), + sizeof(struct target_luminance_value)) != DC_OK) + return false; + } return true; } diff --git a/drivers/gpu/drm/amd/display/dmub/dmub_srv.h b/drivers/gpu/drm/amd/display/dmub/dmub_srv.h index ba1715e2d25a..7c9a2b34bd05 100644 --- a/drivers/gpu/drm/amd/display/dmub/dmub_srv.h +++ b/drivers/gpu/drm/amd/display/dmub/dmub_srv.h @@ -271,7 +271,7 @@ struct dmub_srv_hw_params { */ struct dmub_diagnostic_data { uint32_t dmcub_version; - uint32_t scratch[16]; + uint32_t scratch[17]; uint32_t pc; uint32_t undefined_address_fault_addr; uint32_t inst_fetch_fault_addr; @@ -282,6 +282,7 @@ struct dmub_diagnostic_data { uint32_t inbox0_rptr; uint32_t inbox0_wptr; uint32_t inbox0_size; + uint32_t gpint_datain0; uint8_t is_dmcub_enabled : 1; uint8_t is_dmcub_soft_reset : 1; uint8_t is_dmcub_secure_reset : 1; @@ -340,6 +341,8 @@ struct dmub_srv_hw_funcs { void (*setup_mailbox)(struct dmub_srv *dmub, const struct dmub_region *inbox1); + uint32_t (*get_inbox1_wptr)(struct dmub_srv *dmub); + uint32_t (*get_inbox1_rptr)(struct dmub_srv *dmub); void (*set_inbox1_wptr)(struct dmub_srv *dmub, uint32_t wptr_offset); @@ -366,7 +369,6 @@ struct dmub_srv_hw_funcs { bool (*is_hw_init)(struct dmub_srv *dmub); - bool (*is_phy_init)(struct dmub_srv *dmub); void (*enable_dmub_boot_options)(struct dmub_srv *dmub, const struct dmub_srv_hw_params *params); @@ -602,6 +604,18 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub, enum dmub_status dmub_srv_hw_reset(struct dmub_srv *dmub); /** + * dmub_srv_sync_inbox1() - sync sw state with hw state + * @dmub: the dmub service + * + * Sync sw state with hw state when resume from S0i3 + * + * Return: + * DMUB_STATUS_OK - success + * DMUB_STATUS_INVALID - unspecified error + */ +enum dmub_status dmub_srv_sync_inbox1(struct dmub_srv *dmub); + +/** * dmub_srv_cmd_queue() - queues a command to the DMUB * @dmub: the dmub service * @cmd: the command to queue diff --git a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h index 598fa1de54ce..af1f50742371 100644 --- a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h +++ b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h @@ -257,7 +257,9 @@ struct dmub_feature_caps { */ uint8_t psr; uint8_t fw_assisted_mclk_switch; - uint8_t reserved[6]; + uint8_t reserved[4]; + uint8_t subvp_psr_support; + uint8_t gecc_enable; }; struct dmub_visual_confirm_color { @@ -360,7 +362,7 @@ union dmub_fw_boot_status { uint32_t optimized_init_done : 1; /**< 1 if optimized init done */ uint32_t restore_required : 1; /**< 1 if driver should call restore */ uint32_t defer_load : 1; /**< 1 if VBIOS data is deferred programmed */ - uint32_t reserved : 1; + uint32_t fams_enabled : 1; /**< 1 if VBIOS data is deferred programmed */ uint32_t detection_required: 1; /**< if detection need to be triggered by driver */ uint32_t hw_power_init_done: 1; /**< 1 if hw power init is completed */ } bits; /**< status bits */ @@ -376,6 +378,7 @@ enum dmub_fw_boot_status_bit { DMUB_FW_BOOT_STATUS_BIT_OPTIMIZED_INIT_DONE = (1 << 2), /**< 1 if init done */ DMUB_FW_BOOT_STATUS_BIT_RESTORE_REQUIRED = (1 << 3), /**< 1 if driver should call restore */ DMUB_FW_BOOT_STATUS_BIT_DEFERRED_LOADED = (1 << 4), /**< 1 if VBIOS data is deferred programmed */ + DMUB_FW_BOOT_STATUS_BIT_FAMS_ENABLED = (1 << 5), /**< 1 if FAMS is enabled*/ DMUB_FW_BOOT_STATUS_BIT_DETECTION_REQUIRED = (1 << 6), /**< 1 if detection need to be triggered by driver*/ DMUB_FW_BOOT_STATUS_BIT_HW_POWER_INIT_DONE = (1 << 7), /**< 1 if hw power init is completed */ }; @@ -395,6 +398,12 @@ enum dmub_lvtma_status_bit { DMUB_LVTMA_STATUS_BIT_EDP_ON = (1 << 1), }; +enum dmub_ips_disable_type { + DMUB_IPS_DISABLE_IPS1 = 1, + DMUB_IPS_DISABLE_IPS2 = 2, + DMUB_IPS_DISABLE_IPS2_Z10 = 3, +}; + /** * union dmub_fw_boot_options - Boot option definitions for SCRATCH14 */ @@ -419,7 +428,10 @@ union dmub_fw_boot_options { uint32_t dpia_hpd_int_enable_supported: 1; /* 1 if dpia hpd int enable supported */ uint32_t usb4_dpia_bw_alloc_supported: 1; /* 1 if USB4 dpia BW allocation supported */ uint32_t disable_clk_ds: 1; /* 1 if disallow dispclk_ds and dppclk_ds*/ - uint32_t reserved : 14; /**< reserved */ + uint32_t disable_timeout_recovery : 1; /* 1 if timeout recovery should be disabled */ + uint32_t ips_pg_disable: 1; /* 1 to disable ONO domains power gating*/ + uint32_t ips_disable: 2; /* options to disable ips support*/ + uint32_t reserved : 10; /**< reserved */ } bits; /**< boot bits */ uint32_t all; /**< 32-bit access to bits */ }; @@ -988,16 +1000,25 @@ struct dmub_rb_cmd_mall { }; /** - * enum dmub_cmd_cab_type - TODO: + * enum dmub_cmd_cab_type - CAB command data. */ enum dmub_cmd_cab_type { + /** + * No idle optimizations (i.e. no CAB) + */ DMUB_CMD__CAB_NO_IDLE_OPTIMIZATION = 0, + /** + * No DCN requests for memory + */ DMUB_CMD__CAB_NO_DCN_REQ = 1, + /** + * Fit surfaces in CAB (i.e. CAB enable) + */ DMUB_CMD__CAB_DCN_SS_FIT_IN_CAB = 2, }; /** - * struct dmub_rb_cmd_cab_for_ss - TODO: + * struct dmub_rb_cmd_cab - CAB command data. */ struct dmub_rb_cmd_cab_for_ss { struct dmub_cmd_header header; @@ -1005,6 +1026,9 @@ struct dmub_rb_cmd_cab_for_ss { uint8_t debug_bits; /* debug bits */ }; +/** + * Enum for indicating which MCLK switch mode per pipe + */ enum mclk_switch_mode { NONE = 0, FPO = 1, @@ -1125,8 +1149,6 @@ struct dmub_rb_cmd_idle_opt_dcn_restore { */ struct dmub_dcn_notify_idle_cntl_data { uint8_t driver_idle; - uint8_t d3_entry; - uint8_t trigger; uint8_t pad[1]; }; @@ -3550,6 +3572,10 @@ union dmub_rb_cmd { * Definition of a DMUB_CMD__DPIA_HPD_INT_ENABLE command. */ struct dmub_rb_cmd_dpia_hpd_int_enable dpia_hpd_int_enable; + /** + * Definition of a DMUB_CMD__IDLE_OPT_DCN_NOTIFY_IDLE command. + */ + struct dmub_rb_cmd_idle_opt_dcn_notify_idle idle_opt_notify_idle; }; /** diff --git a/drivers/gpu/drm/amd/display/dmub/src/Makefile b/drivers/gpu/drm/amd/display/dmub/src/Makefile index 0589ad4778ee..caf095aca8f3 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/Makefile +++ b/drivers/gpu/drm/amd/display/dmub/src/Makefile @@ -22,7 +22,7 @@ DMUB = dmub_srv.o dmub_srv_stat.o dmub_reg.o dmub_dcn20.o dmub_dcn21.o DMUB += dmub_dcn30.o dmub_dcn301.o dmub_dcn302.o dmub_dcn303.o -DMUB += dmub_dcn31.o dmub_dcn315.o dmub_dcn316.o +DMUB += dmub_dcn31.o dmub_dcn314.o dmub_dcn315.o dmub_dcn316.o DMUB += dmub_dcn32.o AMD_DAL_DMUB = $(addprefix $(AMDDALPATH)/dmub/src/,$(DMUB)) diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.c index a6540e27044d..98dad0d47e72 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.c +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.c @@ -282,6 +282,11 @@ void dmub_dcn20_setup_mailbox(struct dmub_srv *dmub, REG_WRITE(DMCUB_INBOX1_SIZE, inbox1->top - inbox1->base); } +uint32_t dmub_dcn20_get_inbox1_wptr(struct dmub_srv *dmub) +{ + return REG_READ(DMCUB_INBOX1_WPTR); +} + uint32_t dmub_dcn20_get_inbox1_rptr(struct dmub_srv *dmub) { return REG_READ(DMCUB_INBOX1_RPTR); diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.h b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.h index c2e5831ac52c..1df128e57ed3 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.h +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn20.h @@ -202,6 +202,8 @@ void dmub_dcn20_setup_windows(struct dmub_srv *dmub, void dmub_dcn20_setup_mailbox(struct dmub_srv *dmub, const struct dmub_region *inbox1); +uint32_t dmub_dcn20_get_inbox1_wptr(struct dmub_srv *dmub); + uint32_t dmub_dcn20_get_inbox1_rptr(struct dmub_srv *dmub); void dmub_dcn20_set_inbox1_wptr(struct dmub_srv *dmub, uint32_t wptr_offset); diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn21.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn21.c index 51bb9bceb1b1..2d212bc974cc 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn21.c +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn21.c @@ -54,9 +54,3 @@ const struct dmub_srv_common_regs dmub_srv_dcn21_regs = { #undef DMUB_SF }; -/* Shared functions. */ - -bool dmub_dcn21_is_phy_init(struct dmub_srv *dmub) -{ - return REG_READ(DMCUB_SCRATCH10) == 0; -} diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn21.h b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn21.h index 6fd5b0cd4ef3..8c4033ae4007 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn21.h +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn21.h @@ -32,8 +32,4 @@ extern const struct dmub_srv_common_regs dmub_srv_dcn21_regs; -/* Hardware functions. */ - -bool dmub_dcn21_is_phy_init(struct dmub_srv *dmub); - #endif /* _DMUB_DCN21_H_ */ diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.c index c90b9ee42e12..ebf7aeec4029 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.c +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.c @@ -242,6 +242,11 @@ void dmub_dcn31_setup_mailbox(struct dmub_srv *dmub, REG_WRITE(DMCUB_INBOX1_SIZE, inbox1->top - inbox1->base); } +uint32_t dmub_dcn31_get_inbox1_wptr(struct dmub_srv *dmub) +{ + return REG_READ(DMCUB_INBOX1_WPTR); +} + uint32_t dmub_dcn31_get_inbox1_rptr(struct dmub_srv *dmub) { return REG_READ(DMCUB_INBOX1_RPTR); diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.h b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.h index f6db6f89d45d..7d5c10ee539b 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.h +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.h @@ -204,6 +204,8 @@ void dmub_dcn31_setup_windows(struct dmub_srv *dmub, void dmub_dcn31_setup_mailbox(struct dmub_srv *dmub, const struct dmub_region *inbox1); +uint32_t dmub_dcn31_get_inbox1_wptr(struct dmub_srv *dmub); + uint32_t dmub_dcn31_get_inbox1_rptr(struct dmub_srv *dmub); void dmub_dcn31_set_inbox1_wptr(struct dmub_srv *dmub, uint32_t wptr_offset); diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn314.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn314.c new file mode 100644 index 000000000000..48a06dbd9be7 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn314.c @@ -0,0 +1,62 @@ +/* + * Copyright 2021 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: AMD + * + */ + +#include "../dmub_srv.h" +#include "dmub_reg.h" +#include "dmub_dcn314.h" + +#include "dcn/dcn_3_1_4_offset.h" +#include "dcn/dcn_3_1_4_sh_mask.h" + +#define DCN_BASE__INST0_SEG0 0x00000012 +#define DCN_BASE__INST0_SEG1 0x000000C0 +#define DCN_BASE__INST0_SEG2 0x000034C0 +#define DCN_BASE__INST0_SEG3 0x00009000 +#define DCN_BASE__INST0_SEG4 0x02403C00 +#define DCN_BASE__INST0_SEG5 0 + +#define BASE_INNER(seg) DCN_BASE__INST0_SEG##seg +#define CTX dmub +#define REGS dmub->regs_dcn31 +#define REG_OFFSET_EXP(reg_name) (BASE(reg##reg_name##_BASE_IDX) + reg##reg_name) + +/* Registers. */ + +const struct dmub_srv_dcn31_regs dmub_srv_dcn314_regs = { +#define DMUB_SR(reg) REG_OFFSET_EXP(reg), + { + DMUB_DCN31_REGS() + DMCUB_INTERNAL_REGS() + }, +#undef DMUB_SR + +#define DMUB_SF(reg, field) FD_MASK(reg, field), + { DMUB_DCN31_FIELDS() }, +#undef DMUB_SF + +#define DMUB_SF(reg, field) FD_SHIFT(reg, field), + { DMUB_DCN31_FIELDS() }, +#undef DMUB_SF +}; diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn314.h b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn314.h new file mode 100644 index 000000000000..674267a2940e --- /dev/null +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn314.h @@ -0,0 +1,33 @@ +/* + * Copyright 2021 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: AMD + * + */ + +#ifndef _DMUB_DCN314_H_ +#define _DMUB_DCN314_H_ + +#include "dmub_dcn31.h" + +extern const struct dmub_srv_dcn31_regs dmub_srv_dcn314_regs; + +#endif /* _DMUB_DCN314_H_ */ diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn32.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn32.c index 9c20516be066..bf5994e292d9 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn32.c +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn32.c @@ -116,10 +116,6 @@ void dmub_dcn32_reset(struct dmub_srv *dmub) break; } - /* Clear the GPINT command manually so we don't reset again. */ - cmd.all = 0; - dmub->hw_funcs.set_gpint(dmub, cmd); - /* Force reset in case we timed out, DMCUB is likely hung. */ } @@ -133,6 +129,10 @@ void dmub_dcn32_reset(struct dmub_srv *dmub) REG_WRITE(DMCUB_OUTBOX0_RPTR, 0); REG_WRITE(DMCUB_OUTBOX0_WPTR, 0); REG_WRITE(DMCUB_SCRATCH0, 0); + + /* Clear the GPINT command manually so we don't reset again. */ + cmd.all = 0; + dmub->hw_funcs.set_gpint(dmub, cmd); } void dmub_dcn32_reset_release(struct dmub_srv *dmub) @@ -266,6 +266,11 @@ void dmub_dcn32_setup_mailbox(struct dmub_srv *dmub, REG_WRITE(DMCUB_INBOX1_SIZE, inbox1->top - inbox1->base); } +uint32_t dmub_dcn32_get_inbox1_wptr(struct dmub_srv *dmub) +{ + return REG_READ(DMCUB_INBOX1_WPTR); +} + uint32_t dmub_dcn32_get_inbox1_rptr(struct dmub_srv *dmub) { return REG_READ(DMCUB_INBOX1_RPTR); @@ -434,6 +439,7 @@ void dmub_dcn32_get_diagnostic_data(struct dmub_srv *dmub, struct dmub_diagnosti diag_data->scratch[13] = REG_READ(DMCUB_SCRATCH13); diag_data->scratch[14] = REG_READ(DMCUB_SCRATCH14); diag_data->scratch[15] = REG_READ(DMCUB_SCRATCH15); + diag_data->scratch[16] = REG_READ(DMCUB_SCRATCH16); diag_data->undefined_address_fault_addr = REG_READ(DMCUB_UNDEFINED_ADDRESS_FAULT_ADDR); diag_data->inst_fetch_fault_addr = REG_READ(DMCUB_INST_FETCH_FAULT_ADDR); @@ -464,6 +470,8 @@ void dmub_dcn32_get_diagnostic_data(struct dmub_srv *dmub, struct dmub_diagnosti REG_GET(DMCUB_REGION3_CW6_TOP_ADDRESS, DMCUB_REGION3_CW6_ENABLE, &is_cw6_enabled); diag_data->is_cw6_enabled = is_cw6_enabled; + + diag_data->gpint_datain0 = REG_READ(DMCUB_GPINT_DATAIN0); } void dmub_dcn32_configure_dmub_in_system_memory(struct dmub_srv *dmub) { diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn32.h b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn32.h index 7d1a6eb4d665..d58a1e4b9f1c 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn32.h +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn32.h @@ -107,6 +107,7 @@ struct dmub_srv; DMUB_SR(DMCUB_SCRATCH15) \ DMUB_SR(DMCUB_SCRATCH16) \ DMUB_SR(DMCUB_SCRATCH17) \ + DMUB_SR(DMCUB_GPINT_DATAIN0) \ DMUB_SR(DMCUB_GPINT_DATAIN1) \ DMUB_SR(DMCUB_GPINT_DATAOUT) \ DMUB_SR(CC_DC_PIPE_DIS) \ @@ -206,6 +207,8 @@ void dmub_dcn32_setup_windows(struct dmub_srv *dmub, void dmub_dcn32_setup_mailbox(struct dmub_srv *dmub, const struct dmub_region *inbox1); +uint32_t dmub_dcn32_get_inbox1_wptr(struct dmub_srv *dmub); + uint32_t dmub_dcn32_get_inbox1_rptr(struct dmub_srv *dmub); void dmub_dcn32_set_inbox1_wptr(struct dmub_srv *dmub, uint32_t wptr_offset); diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c index 92c18bfb98b3..9e9a6a44a7ac 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c @@ -32,6 +32,7 @@ #include "dmub_dcn302.h" #include "dmub_dcn303.h" #include "dmub_dcn31.h" +#include "dmub_dcn314.h" #include "dmub_dcn315.h" #include "dmub_dcn316.h" #include "dmub_dcn32.h" @@ -166,6 +167,7 @@ static bool dmub_srv_hw_setup(struct dmub_srv *dmub, enum dmub_asic asic) funcs->backdoor_load = dmub_dcn20_backdoor_load; funcs->setup_windows = dmub_dcn20_setup_windows; funcs->setup_mailbox = dmub_dcn20_setup_mailbox; + funcs->get_inbox1_wptr = dmub_dcn20_get_inbox1_wptr; funcs->get_inbox1_rptr = dmub_dcn20_get_inbox1_rptr; funcs->set_inbox1_wptr = dmub_dcn20_set_inbox1_wptr; funcs->is_supported = dmub_dcn20_is_supported; @@ -190,11 +192,9 @@ static bool dmub_srv_hw_setup(struct dmub_srv *dmub, enum dmub_asic asic) funcs->get_diagnostic_data = dmub_dcn20_get_diagnostic_data; - if (asic == DMUB_ASIC_DCN21) { + if (asic == DMUB_ASIC_DCN21) dmub->regs = &dmub_srv_dcn21_regs; - funcs->is_phy_init = dmub_dcn21_is_phy_init; - } if (asic == DMUB_ASIC_DCN30) { dmub->regs = &dmub_srv_dcn30_regs; @@ -226,7 +226,9 @@ static bool dmub_srv_hw_setup(struct dmub_srv *dmub, enum dmub_asic asic) case DMUB_ASIC_DCN314: case DMUB_ASIC_DCN315: case DMUB_ASIC_DCN316: - if (asic == DMUB_ASIC_DCN315) + if (asic == DMUB_ASIC_DCN314) + dmub->regs_dcn31 = &dmub_srv_dcn314_regs; + else if (asic == DMUB_ASIC_DCN315) dmub->regs_dcn31 = &dmub_srv_dcn315_regs; else if (asic == DMUB_ASIC_DCN316) dmub->regs_dcn31 = &dmub_srv_dcn316_regs; @@ -237,6 +239,7 @@ static bool dmub_srv_hw_setup(struct dmub_srv *dmub, enum dmub_asic asic) funcs->backdoor_load = dmub_dcn31_backdoor_load; funcs->setup_windows = dmub_dcn31_setup_windows; funcs->setup_mailbox = dmub_dcn31_setup_mailbox; + funcs->get_inbox1_wptr = dmub_dcn31_get_inbox1_wptr; funcs->get_inbox1_rptr = dmub_dcn31_get_inbox1_rptr; funcs->set_inbox1_wptr = dmub_dcn31_set_inbox1_wptr; funcs->setup_out_mailbox = dmub_dcn31_setup_out_mailbox; @@ -275,6 +278,7 @@ static bool dmub_srv_hw_setup(struct dmub_srv *dmub, enum dmub_asic asic) funcs->backdoor_load_zfb_mode = dmub_dcn32_backdoor_load_zfb_mode; funcs->setup_windows = dmub_dcn32_setup_windows; funcs->setup_mailbox = dmub_dcn32_setup_mailbox; + funcs->get_inbox1_wptr = dmub_dcn32_get_inbox1_wptr; funcs->get_inbox1_rptr = dmub_dcn32_get_inbox1_rptr; funcs->set_inbox1_wptr = dmub_dcn32_set_inbox1_wptr; funcs->setup_out_mailbox = dmub_dcn32_setup_out_mailbox; @@ -644,6 +648,20 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub, return DMUB_STATUS_OK; } +enum dmub_status dmub_srv_sync_inbox1(struct dmub_srv *dmub) +{ + if (!dmub->sw_init) + return DMUB_STATUS_INVALID; + + if (dmub->hw_funcs.get_inbox1_rptr && dmub->hw_funcs.get_inbox1_wptr) { + dmub->inbox1_rb.rptr = dmub->hw_funcs.get_inbox1_rptr(dmub); + dmub->inbox1_rb.wrpt = dmub->hw_funcs.get_inbox1_wptr(dmub); + dmub->inbox1_last_wptr = dmub->inbox1_rb.wrpt; + } + + return DMUB_STATUS_OK; +} + enum dmub_status dmub_srv_hw_reset(struct dmub_srv *dmub) { if (!dmub->sw_init) @@ -721,27 +739,6 @@ enum dmub_status dmub_srv_wait_for_auto_load(struct dmub_srv *dmub, return DMUB_STATUS_TIMEOUT; } -enum dmub_status dmub_srv_wait_for_phy_init(struct dmub_srv *dmub, - uint32_t timeout_us) -{ - uint32_t i = 0; - - if (!dmub->hw_init) - return DMUB_STATUS_INVALID; - - if (!dmub->hw_funcs.is_phy_init) - return DMUB_STATUS_OK; - - for (i = 0; i <= timeout_us; i += 10) { - if (dmub->hw_funcs.is_phy_init(dmub)) - return DMUB_STATUS_OK; - - udelay(10); - } - - return DMUB_STATUS_TIMEOUT; -} - enum dmub_status dmub_srv_wait_for_idle(struct dmub_srv *dmub, uint32_t timeout_us) { diff --git a/drivers/gpu/drm/amd/display/include/fixed31_32.h b/drivers/gpu/drm/amd/display/include/fixed31_32.h index ece97ae0e826..d4cf7ead1d87 100644 --- a/drivers/gpu/drm/amd/display/include/fixed31_32.h +++ b/drivers/gpu/drm/amd/display/include/fixed31_32.h @@ -525,7 +525,7 @@ static inline struct fixed31_32 dc_fixpt_truncate(struct fixed31_32 arg, unsigne if (negative) arg.value = -arg.value; - arg.value &= (~0LL) << (FIXED31_32_BITS_PER_FRACTIONAL_PART - frac_bits); + arg.value &= (~0ULL) << (FIXED31_32_BITS_PER_FRACTIONAL_PART - frac_bits); if (negative) arg.value = -arg.value; return arg; diff --git a/drivers/gpu/drm/amd/display/include/signal_types.h b/drivers/gpu/drm/amd/display/include/signal_types.h index 23a308c3eccb..325c5ba4c82a 100644 --- a/drivers/gpu/drm/amd/display/include/signal_types.h +++ b/drivers/gpu/drm/amd/display/include/signal_types.h @@ -44,6 +44,34 @@ enum signal_type { SIGNAL_TYPE_VIRTUAL = (1 << 9), /* Virtual Display */ }; +static inline const char *signal_type_to_string(const int type) +{ + switch (type) { + case SIGNAL_TYPE_NONE: + return "No signal"; + case SIGNAL_TYPE_DVI_SINGLE_LINK: + return "DVI: Single Link"; + case SIGNAL_TYPE_DVI_DUAL_LINK: + return "DVI: Dual Link"; + case SIGNAL_TYPE_HDMI_TYPE_A: + return "HDMI: TYPE A"; + case SIGNAL_TYPE_LVDS: + return "LVDS"; + case SIGNAL_TYPE_RGB: + return "RGB"; + case SIGNAL_TYPE_DISPLAY_PORT: + return "Display Port"; + case SIGNAL_TYPE_DISPLAY_PORT_MST: + return "Display Port: MST"; + case SIGNAL_TYPE_EDP: + return "Embedded Display Port"; + case SIGNAL_TYPE_VIRTUAL: + return "Virtual"; + default: + return "Unknown"; + } +} + /* help functions for signal types manipulation */ static inline bool dc_is_hdmi_tmds_signal(enum signal_type signal) { diff --git a/drivers/gpu/drm/amd/display/modules/freesync/freesync.c b/drivers/gpu/drm/amd/display/modules/freesync/freesync.c index 5c41a4751db4..dbd60811f95d 100644 --- a/drivers/gpu/drm/amd/display/modules/freesync/freesync.c +++ b/drivers/gpu/drm/amd/display/modules/freesync/freesync.c @@ -1,5 +1,5 @@ /* - * Copyright 2016 Advanced Micro Devices, Inc. + * Copyright 2016-2023 Advanced Micro Devices, Inc. * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), @@ -989,6 +989,7 @@ void mod_freesync_build_vrr_params(struct mod_freesync *mod_freesync, unsigned int refresh_range = 0; unsigned long long min_refresh_in_uhz = 0; unsigned long long max_refresh_in_uhz = 0; + unsigned long long min_hardware_refresh_in_uhz = 0; if (mod_freesync == NULL) return; @@ -999,7 +1000,13 @@ void mod_freesync_build_vrr_params(struct mod_freesync *mod_freesync, nominal_field_rate_in_uhz = mod_freesync_calc_nominal_field_rate(stream); - min_refresh_in_uhz = in_config->min_refresh_in_uhz; + if (stream->ctx->dc->caps.max_v_total != 0 && stream->timing.h_total != 0) { + min_hardware_refresh_in_uhz = div64_u64((stream->timing.pix_clk_100hz * 100000000ULL), + (stream->timing.h_total * stream->ctx->dc->caps.max_v_total)); + } + /* Limit minimum refresh rate to what can be supported by hardware */ + min_refresh_in_uhz = min_hardware_refresh_in_uhz > in_config->min_refresh_in_uhz ? + min_hardware_refresh_in_uhz : in_config->min_refresh_in_uhz; max_refresh_in_uhz = in_config->max_refresh_in_uhz; /* Full range may be larger than current video timing, so cap at nominal */ @@ -1137,10 +1144,6 @@ void mod_freesync_handle_preflip(struct mod_freesync *mod_freesync, if (in_out_vrr->supported && in_out_vrr->state == VRR_STATE_ACTIVE_VARIABLE) { - unsigned int oldest_index = plane->time.index + 1; - - if (oldest_index >= DC_PLANE_UPDATE_TIMES_MAX) - oldest_index = 0; last_render_time_in_us = curr_time_stamp_in_us - plane->time.prev_update_time_in_us; diff --git a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c index 51e76bce92ea..30349881a283 100644 --- a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c +++ b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c @@ -116,6 +116,27 @@ static const struct abm_parameters * const abm_settings[] = { abm_settings_config2, }; +static const struct dm_bl_data_point custom_backlight_curve0[] = { + {2, 14}, {4, 16}, {6, 18}, {8, 21}, {10, 23}, {12, 26}, {14, 29}, {16, 32}, {18, 35}, + {20, 38}, {22, 41}, {24, 44}, {26, 48}, {28, 52}, {30, 55}, {32, 59}, {34, 62}, + {36, 67}, {38, 71}, {40, 75}, {42, 80}, {44, 84}, {46, 88}, {48, 93}, {50, 98}, + {52, 103}, {54, 108}, {56, 113}, {58, 118}, {60, 123}, {62, 129}, {64, 135}, {66, 140}, + {68, 146}, {70, 152}, {72, 158}, {74, 164}, {76, 171}, {78, 177}, {80, 183}, {82, 190}, + {84, 197}, {86, 204}, {88, 211}, {90, 218}, {92, 225}, {94, 232}, {96, 240}, {98, 247}}; + +struct custom_backlight_profile { + uint8_t ac_level_percentage; + uint8_t dc_level_percentage; + uint8_t min_input_signal; + uint8_t max_input_signal; + uint8_t num_data_points; + const struct dm_bl_data_point *data_points; +}; + +static const struct custom_backlight_profile custom_backlight_profiles[] = { + {100, 32, 12, 255, ARRAY_SIZE(custom_backlight_curve0), custom_backlight_curve0}, +}; + #define NUM_AMBI_LEVEL 5 #define NUM_AGGR_LEVEL 4 #define NUM_POWER_FN_SEGS 8 @@ -944,3 +965,25 @@ bool psr_su_set_dsc_slice_height(struct dc *dc, struct dc_link *link, return true; } + +bool fill_custom_backlight_caps(unsigned int config_no, struct dm_acpi_atif_backlight_caps *caps) +{ + unsigned int data_points_size; + + if (config_no >= ARRAY_SIZE(custom_backlight_profiles)) + return false; + + data_points_size = custom_backlight_profiles[config_no].num_data_points + * sizeof(custom_backlight_profiles[config_no].data_points[0]); + + caps->size = sizeof(struct dm_acpi_atif_backlight_caps) - sizeof(caps->data_points) + data_points_size; + caps->flags = 0; + caps->error_code = 0; + caps->ac_level_percentage = custom_backlight_profiles[config_no].ac_level_percentage; + caps->dc_level_percentage = custom_backlight_profiles[config_no].dc_level_percentage; + caps->min_input_signal = custom_backlight_profiles[config_no].min_input_signal; + caps->max_input_signal = custom_backlight_profiles[config_no].max_input_signal; + caps->num_data_points = custom_backlight_profiles[config_no].num_data_points; + memcpy(caps->data_points, custom_backlight_profiles[config_no].data_points, data_points_size); + return true; +} diff --git a/drivers/gpu/drm/amd/display/modules/power/power_helpers.h b/drivers/gpu/drm/amd/display/modules/power/power_helpers.h index 1d3079e56799..ffc924c9991b 100644 --- a/drivers/gpu/drm/amd/display/modules/power/power_helpers.h +++ b/drivers/gpu/drm/amd/display/modules/power/power_helpers.h @@ -62,4 +62,7 @@ bool mod_power_only_edp(const struct dc_state *context, bool psr_su_set_dsc_slice_height(struct dc *dc, struct dc_link *link, struct dc_stream_state *stream, struct psr_config *config); + +bool fill_custom_backlight_caps(unsigned int config_no, + struct dm_acpi_atif_backlight_caps *caps); #endif /* MODULES_POWER_POWER_HELPERS_H_ */ diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h index e4a22c68517d..f175e65b853a 100644 --- a/drivers/gpu/drm/amd/include/amd_shared.h +++ b/drivers/gpu/drm/amd/include/amd_shared.h @@ -240,7 +240,6 @@ enum DC_FEATURE_MASK { DC_DISABLE_LTTPR_DP2_0 = (1 << 6), //0x40, disabled by default DC_PSR_ALLOW_SMU_OPT = (1 << 7), //0x80, disabled by default DC_PSR_ALLOW_MULTI_DISP_OPT = (1 << 8), //0x100, disabled by default - DC_ENABLE_SUBVP_DRR = (1 << 9), // 0x200, disabled by default }; enum DC_DEBUG_MASK { diff --git a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_1_0_offset.h b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_1_0_offset.h index 79c41004c0b6..4908044f7409 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_1_0_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_1_0_offset.h @@ -5194,6 +5194,20 @@ #define mmSPI_WCL_PIPE_PERCENT_CS6_BASE_IDX 0 #define mmSPI_WCL_PIPE_PERCENT_CS7 0x1f70 #define mmSPI_WCL_PIPE_PERCENT_CS7_BASE_IDX 0 +#define mmSPI_GDBG_WAVE_CNTL 0x1f71 +#define mmSPI_GDBG_WAVE_CNTL_BASE_IDX 0 +#define mmSPI_GDBG_TRAP_CONFIG 0x1f72 +#define mmSPI_GDBG_TRAP_CONFIG_BASE_IDX 0 +#define mmSPI_GDBG_TRAP_MASK 0x1f73 +#define mmSPI_GDBG_TRAP_MASK_BASE_IDX 0 +#define mmSPI_GDBG_WAVE_CNTL2 0x1f74 +#define mmSPI_GDBG_WAVE_CNTL2_BASE_IDX 0 +#define mmSPI_GDBG_WAVE_CNTL3 0x1f75 +#define mmSPI_GDBG_WAVE_CNTL3_BASE_IDX 0 +#define mmSPI_GDBG_TRAP_DATA0 0x1f78 +#define mmSPI_GDBG_TRAP_DATA0_BASE_IDX 0 +#define mmSPI_GDBG_TRAP_DATA1 0x1f79 +#define mmSPI_GDBG_TRAP_DATA1_BASE_IDX 0 #define mmSPI_COMPUTE_QUEUE_RESET 0x1f7b #define mmSPI_COMPUTE_QUEUE_RESET_BASE_IDX 0 #define mmSPI_RESOURCE_RESERVE_CU_0 0x1f7c diff --git a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_1_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_1_0_sh_mask.h index 52043e143067..9b7d219e7954 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_1_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_1_0_sh_mask.h @@ -19700,6 +19700,75 @@ //SPI_WCL_PIPE_PERCENT_CS7 #define SPI_WCL_PIPE_PERCENT_CS7__VALUE__SHIFT 0x0 #define SPI_WCL_PIPE_PERCENT_CS7__VALUE_MASK 0x7FL +//SPI_GDBG_WAVE_CNTL +#define SPI_GDBG_WAVE_CNTL__STALL_RA__SHIFT 0x0 +#define SPI_GDBG_WAVE_CNTL__STALL_VMID__SHIFT 0x1 +#define SPI_GDBG_WAVE_CNTL__STALL_RA_MASK 0x00000001L +#define SPI_GDBG_WAVE_CNTL__STALL_VMID_MASK 0x0001FFFEL +//SPI_GDBG_TRAP_CONFIG +#define SPI_GDBG_TRAP_CONFIG__ME_SEL__SHIFT 0x0 +#define SPI_GDBG_TRAP_CONFIG__PIPE_SEL__SHIFT 0x2 +#define SPI_GDBG_TRAP_CONFIG__QUEUE_SEL__SHIFT 0x4 +#define SPI_GDBG_TRAP_CONFIG__ME_MATCH__SHIFT 0x7 +#define SPI_GDBG_TRAP_CONFIG__PIPE_MATCH__SHIFT 0x8 +#define SPI_GDBG_TRAP_CONFIG__QUEUE_MATCH__SHIFT 0x9 +#define SPI_GDBG_TRAP_CONFIG__TRAP_EN__SHIFT 0xf +#define SPI_GDBG_TRAP_CONFIG__VMID_SEL__SHIFT 0x10 +#define SPI_GDBG_TRAP_CONFIG__ME_SEL_MASK 0x00000003L +#define SPI_GDBG_TRAP_CONFIG__PIPE_SEL_MASK 0x0000000CL +#define SPI_GDBG_TRAP_CONFIG__QUEUE_SEL_MASK 0x00000070L +#define SPI_GDBG_TRAP_CONFIG__ME_MATCH_MASK 0x00000080L +#define SPI_GDBG_TRAP_CONFIG__PIPE_MATCH_MASK 0x00000100L +#define SPI_GDBG_TRAP_CONFIG__QUEUE_MATCH_MASK 0x00000200L +#define SPI_GDBG_TRAP_CONFIG__TRAP_EN_MASK 0x00008000L +#define SPI_GDBG_TRAP_CONFIG__VMID_SEL_MASK 0xFFFF0000L +//SPI_GDBG_TRAP_MASK +#define SPI_GDBG_TRAP_MASK__EXCP_EN__SHIFT 0x0 +#define SPI_GDBG_TRAP_MASK__REPLACE__SHIFT 0x9 +#define SPI_GDBG_TRAP_MASK__EXCP_EN_MASK 0x01FFL +#define SPI_GDBG_TRAP_MASK__REPLACE_MASK 0x0200L +//SPI_GDBG_WAVE_CNTL2 +#define SPI_GDBG_WAVE_CNTL2__VMID_MASK__SHIFT 0x0 +#define SPI_GDBG_WAVE_CNTL2__MODE__SHIFT 0x10 +#define SPI_GDBG_WAVE_CNTL2__VMID_MASK_MASK 0x0000FFFFL +#define SPI_GDBG_WAVE_CNTL2__MODE_MASK 0x00030000L +//SPI_GDBG_WAVE_CNTL3 +#define SPI_GDBG_WAVE_CNTL3__STALL_PS__SHIFT 0x0 +#define SPI_GDBG_WAVE_CNTL3__STALL_VS__SHIFT 0x1 +#define SPI_GDBG_WAVE_CNTL3__STALL_GS__SHIFT 0x2 +#define SPI_GDBG_WAVE_CNTL3__STALL_HS__SHIFT 0x3 +#define SPI_GDBG_WAVE_CNTL3__STALL_CSG__SHIFT 0x4 +#define SPI_GDBG_WAVE_CNTL3__STALL_CS0__SHIFT 0x5 +#define SPI_GDBG_WAVE_CNTL3__STALL_CS1__SHIFT 0x6 +#define SPI_GDBG_WAVE_CNTL3__STALL_CS2__SHIFT 0x7 +#define SPI_GDBG_WAVE_CNTL3__STALL_CS3__SHIFT 0x8 +#define SPI_GDBG_WAVE_CNTL3__STALL_CS4__SHIFT 0x9 +#define SPI_GDBG_WAVE_CNTL3__STALL_CS5__SHIFT 0xa +#define SPI_GDBG_WAVE_CNTL3__STALL_CS6__SHIFT 0xb +#define SPI_GDBG_WAVE_CNTL3__STALL_CS7__SHIFT 0xc +#define SPI_GDBG_WAVE_CNTL3__STALL_DURATION__SHIFT 0xd +#define SPI_GDBG_WAVE_CNTL3__STALL_MULT__SHIFT 0x1c +#define SPI_GDBG_WAVE_CNTL3__STALL_PS_MASK 0x00000001L +#define SPI_GDBG_WAVE_CNTL3__STALL_VS_MASK 0x00000002L +#define SPI_GDBG_WAVE_CNTL3__STALL_GS_MASK 0x00000004L +#define SPI_GDBG_WAVE_CNTL3__STALL_HS_MASK 0x00000008L +#define SPI_GDBG_WAVE_CNTL3__STALL_CSG_MASK 0x00000010L +#define SPI_GDBG_WAVE_CNTL3__STALL_CS0_MASK 0x00000020L +#define SPI_GDBG_WAVE_CNTL3__STALL_CS1_MASK 0x00000040L +#define SPI_GDBG_WAVE_CNTL3__STALL_CS2_MASK 0x00000080L +#define SPI_GDBG_WAVE_CNTL3__STALL_CS3_MASK 0x00000100L +#define SPI_GDBG_WAVE_CNTL3__STALL_CS4_MASK 0x00000200L +#define SPI_GDBG_WAVE_CNTL3__STALL_CS5_MASK 0x00000400L +#define SPI_GDBG_WAVE_CNTL3__STALL_CS6_MASK 0x00000800L +#define SPI_GDBG_WAVE_CNTL3__STALL_CS7_MASK 0x00001000L +#define SPI_GDBG_WAVE_CNTL3__STALL_DURATION_MASK 0x0FFFE000L +#define SPI_GDBG_WAVE_CNTL3__STALL_MULT_MASK 0x10000000L +//SPI_GDBG_TRAP_DATA0 +#define SPI_GDBG_TRAP_DATA0__DATA__SHIFT 0x0 +#define SPI_GDBG_TRAP_DATA0__DATA_MASK 0xFFFFFFFFL +//SPI_GDBG_TRAP_DATA1 +#define SPI_GDBG_TRAP_DATA1__DATA__SHIFT 0x0 +#define SPI_GDBG_TRAP_DATA1__DATA_MASK 0xFFFFFFFFL //SPI_COMPUTE_QUEUE_RESET #define SPI_COMPUTE_QUEUE_RESET__RESET__SHIFT 0x0 #define SPI_COMPUTE_QUEUE_RESET__RESET_MASK 0x01L diff --git a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_3_0_offset.h b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_3_0_offset.h index a734abaa91a5..5e15ac14b63c 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_3_0_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_3_0_offset.h @@ -26,6 +26,8 @@ #define mmSQ_DEBUG_STS_GLOBAL_BASE_IDX 0 #define mmSQ_DEBUG_STS_GLOBAL2 0x10B0 #define mmSQ_DEBUG_STS_GLOBAL2_BASE_IDX 0 +#define mmSQ_DEBUG 0x10B1 +#define mmSQ_DEBUG_BASE_IDX 0 // addressBlock: gc_sdma0_sdma0dec // base address: 0x4980 @@ -4853,10 +4855,18 @@ #define mmSPI_WCL_PIPE_PERCENT_CS3_BASE_IDX 0 #define mmSPI_GDBG_WAVE_CNTL 0x1f71 #define mmSPI_GDBG_WAVE_CNTL_BASE_IDX 0 +#define mmSPI_GDBG_TRAP_CONFIG 0x1f72 +#define mmSPI_GDBG_TRAP_CONFIG_BASE_IDX 0 #define mmSPI_GDBG_TRAP_MASK 0x1f73 #define mmSPI_GDBG_TRAP_MASK_BASE_IDX 0 #define mmSPI_GDBG_WAVE_CNTL2 0x1f74 #define mmSPI_GDBG_WAVE_CNTL2_BASE_IDX 0 +#define mmSPI_GDBG_WAVE_CNTL3 0x1f75 +#define mmSPI_GDBG_WAVE_CNTL3_BASE_IDX 0 +#define mmSPI_GDBG_TRAP_DATA0 0x1f78 +#define mmSPI_GDBG_TRAP_DATA0_BASE_IDX 0 +#define mmSPI_GDBG_TRAP_DATA1 0x1f79 +#define mmSPI_GDBG_TRAP_DATA1_BASE_IDX 0 #define mmSPI_COMPUTE_QUEUE_RESET 0x1f7b #define mmSPI_COMPUTE_QUEUE_RESET_BASE_IDX 0 #define mmSPI_RESOURCE_RESERVE_CU_0 0x1f7c diff --git a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_3_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_3_0_sh_mask.h index d7a17bae2584..e4ecd6c2d20e 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_3_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_10_3_0_sh_mask.h @@ -47907,6 +47907,10 @@ // addressBlock: sqind +//SQ_DEBUG +#define SQ_DEBUG__SINGLE_MEMOP_MASK 0x00000001L +#define SQ_DEBUG__SINGLE_MEMOP__SHIFT 0x00000000 + //SQ_DEBUG_STS_GLOBAL #define SQ_DEBUG_STS_GLOBAL2__FIFO_LEVEL_GFX0_MASK 0x000000ffL #define SQ_DEBUG_STS_GLOBAL2__FIFO_LEVEL_GFX0__SHIFT 0x00000000 diff --git a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_11_0_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_11_0_0_sh_mask.h index 4f08f90856fc..3088a4a13cb5 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_11_0_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_11_0_0_sh_mask.h @@ -17216,11 +17216,15 @@ #define SPI_GDBG_PER_VMID_CNTL__TRAP_EN__SHIFT 0x3 #define SPI_GDBG_PER_VMID_CNTL__EXCP_EN__SHIFT 0x4 #define SPI_GDBG_PER_VMID_CNTL__EXCP_REPLACE__SHIFT 0xd +#define SPI_GDBG_PER_VMID_CNTL__TRAP_ON_START__SHIFT 0xe +#define SPI_GDBG_PER_VMID_CNTL__TRAP_ON_END__SHIFT 0xf #define SPI_GDBG_PER_VMID_CNTL__STALL_VMID_MASK 0x00000001L #define SPI_GDBG_PER_VMID_CNTL__LAUNCH_MODE_MASK 0x00000006L #define SPI_GDBG_PER_VMID_CNTL__TRAP_EN_MASK 0x00000008L #define SPI_GDBG_PER_VMID_CNTL__EXCP_EN_MASK 0x00001FF0L #define SPI_GDBG_PER_VMID_CNTL__EXCP_REPLACE_MASK 0x00002000L +#define SPI_GDBG_PER_VMID_CNTL__TRAP_ON_START_MASK 0x00004000L +#define SPI_GDBG_PER_VMID_CNTL__TRAP_ON_END_MASK 0x00008000L //SPI_COMPUTE_QUEUE_RESET #define SPI_COMPUTE_QUEUE_RESET__RESET__SHIFT 0x0 #define SPI_COMPUTE_QUEUE_RESET__RESET_MASK 0x01L diff --git a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_9_4_3_offset.h b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_9_4_3_offset.h index 3100de8b3881..393963502b7a 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_9_4_3_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_9_4_3_offset.h @@ -705,6 +705,46 @@ #define regSQC_ICACHE_UTCL1_STATUS_BASE_IDX 0 #define regSQC_DCACHE_UTCL1_STATUS 0x03d8 #define regSQC_DCACHE_UTCL1_STATUS_BASE_IDX 0 +#define regSQC_UE_EDC_LO 0x03d9 +#define regSQC_UE_EDC_LO_BASE_IDX 0 +#define regSQC_UE_EDC_HI 0x03da +#define regSQC_UE_EDC_HI_BASE_IDX 0 +#define regSQC_CE_EDC_LO 0x03db +#define regSQC_CE_EDC_LO_BASE_IDX 0 +#define regSQC_CE_EDC_HI 0x03dc +#define regSQC_CE_EDC_HI_BASE_IDX 0 +#define regSQ_UE_ERR_STATUS_LO 0x03dd +#define regSQ_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regSQ_UE_ERR_STATUS_HI 0x03de +#define regSQ_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regSQ_CE_ERR_STATUS_LO 0x03df +#define regSQ_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regSQ_CE_ERR_STATUS_HI 0x03e0 +#define regSQ_CE_ERR_STATUS_HI_BASE_IDX 0 +#define regLDS_UE_ERR_STATUS_LO 0x03e1 +#define regLDS_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regLDS_UE_ERR_STATUS_HI 0x03e2 +#define regLDS_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regLDS_CE_ERR_STATUS_LO 0x03e3 +#define regLDS_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regLDS_CE_ERR_STATUS_HI 0x03e4 +#define regLDS_CE_ERR_STATUS_HI_BASE_IDX 0 +#define regSP0_UE_ERR_STATUS_LO 0x03e5 +#define regSP0_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regSP0_UE_ERR_STATUS_HI 0x03e6 +#define regSP0_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regSP0_CE_ERR_STATUS_LO 0x03e7 +#define regSP0_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regSP0_CE_ERR_STATUS_HI 0x03e8 +#define regSP0_CE_ERR_STATUS_HI_BASE_IDX 0 +#define regSP1_UE_ERR_STATUS_LO 0x03e9 +#define regSP1_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regSP1_UE_ERR_STATUS_HI 0x03ea +#define regSP1_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regSP1_CE_ERR_STATUS_LO 0x03eb +#define regSP1_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regSP1_CE_ERR_STATUS_HI 0x03ec +#define regSP1_CE_ERR_STATUS_HI_BASE_IDX 0 // addressBlock: xcd0_gc_shsdec @@ -727,6 +767,14 @@ #define regSPI_DSM_CNTL2_BASE_IDX 0 #define regSPI_EDC_CNT 0x0445 #define regSPI_EDC_CNT_BASE_IDX 0 +#define regSPI_UE_ERR_STATUS_LO 0x0446 +#define regSPI_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regSPI_UE_ERR_STATUS_HI 0x0447 +#define regSPI_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regSPI_CE_ERR_STATUS_LO 0x0448 +#define regSPI_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regSPI_CE_ERR_STATUS_HI 0x0449 +#define regSPI_CE_ERR_STATUS_HI_BASE_IDX 0 #define regSPI_DEBUG_BUSY 0x0450 #define regSPI_DEBUG_BUSY_BASE_IDX 0 #define regSPI_CONFIG_PS_CU_EN 0x0452 @@ -871,6 +919,14 @@ #define regTD_STATUS_BASE_IDX 0 #define regTD_POWER_CNTL 0x052a #define regTD_POWER_CNTL_BASE_IDX 0 +#define regTD_UE_EDC_LO 0x052b +#define regTD_UE_EDC_LO_BASE_IDX 0 +#define regTD_UE_EDC_HI 0x052c +#define regTD_UE_EDC_HI_BASE_IDX 0 +#define regTD_CE_EDC_LO 0x052d +#define regTD_CE_EDC_LO_BASE_IDX 0 +#define regTD_CE_EDC_HI 0x052e +#define regTD_CE_EDC_HI_BASE_IDX 0 #define regTD_DSM_CNTL 0x052f #define regTD_DSM_CNTL_BASE_IDX 0 #define regTD_DSM_CNTL2 0x0530 @@ -893,6 +949,14 @@ #define regTA_DSM_CNTL_BASE_IDX 0 #define regTA_DSM_CNTL2 0x0585 #define regTA_DSM_CNTL2_BASE_IDX 0 +#define regTA_UE_EDC_LO 0x0587 +#define regTA_UE_EDC_LO_BASE_IDX 0 +#define regTA_UE_EDC_HI 0x0588 +#define regTA_UE_EDC_HI_BASE_IDX 0 +#define regTA_CE_EDC_LO 0x0589 +#define regTA_CE_EDC_LO_BASE_IDX 0 +#define regTA_CE_EDC_HI 0x058a +#define regTA_CE_EDC_HI_BASE_IDX 0 // addressBlock: xcd0_gc_gdsdec @@ -923,6 +987,14 @@ #define regGDS_DSM_CNTL2_BASE_IDX 0 #define regGDS_WD_GDS_CSB 0x05ce #define regGDS_WD_GDS_CSB_BASE_IDX 0 +#define regGDS_UE_ERR_STATUS_LO 0x05cf +#define regGDS_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regGDS_UE_ERR_STATUS_HI 0x05d0 +#define regGDS_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regGDS_CE_ERR_STATUS_LO 0x05d1 +#define regGDS_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regGDS_CE_ERR_STATUS_HI 0x05d2 +#define regGDS_CE_ERR_STATUS_HI_BASE_IDX 0 // addressBlock: xcd0_gc_rbdec @@ -1243,6 +1315,10 @@ #define regGCEA_MAM_CTRL_BASE_IDX 0 #define regGCEA_MAM_CTRL2 0x0702 #define regGCEA_MAM_CTRL2_BASE_IDX 0 +#define regGCEA_UE_ERR_STATUS_LO 0x0706 +#define regGCEA_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regGCEA_UE_ERR_STATUS_HI 0x0707 +#define regGCEA_UE_ERR_STATUS_HI_BASE_IDX 0 #define regGCEA_DSM_CNTL 0x0708 #define regGCEA_DSM_CNTL_BASE_IDX 0 #define regGCEA_DSM_CNTLA 0x0709 @@ -1277,6 +1353,10 @@ #define regGCEA_SDP_BACKDOOR_DATACREDITS1_BASE_IDX 0 #define regGCEA_SDP_BACKDOOR_MISCCREDITS 0x0719 #define regGCEA_SDP_BACKDOOR_MISCCREDITS_BASE_IDX 0 +#define regGCEA_CE_ERR_STATUS_LO 0x071b +#define regGCEA_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regGCEA_CE_ERR_STATUS_HI 0x071d +#define regGCEA_CE_ERR_STATUS_HI_BASE_IDX 0 #define regGCEA_SDP_ENABLE 0x071f #define regGCEA_SDP_ENABLE_BASE_IDX 0 @@ -1389,6 +1469,14 @@ #define regATC_L2_CNTL4_BASE_IDX 0 #define regATC_L2_MM_GROUP_RT_CLASSES 0x0816 #define regATC_L2_MM_GROUP_RT_CLASSES_BASE_IDX 0 +#define regATC_L2_UE_ERR_STATUS_LO 0x081a +#define regATC_L2_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regATC_L2_UE_ERR_STATUS_HI 0x081b +#define regATC_L2_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regATC_L2_CE_ERR_STATUS_LO 0x081c +#define regATC_L2_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regATC_L2_CE_ERR_STATUS_HI 0x081d +#define regATC_L2_CE_ERR_STATUS_HI_BASE_IDX 0 // addressBlock: xcd0_gc_utcl2_vml2pfdec @@ -1475,6 +1563,30 @@ #define regUTCL2_EDC_MODE_BASE_IDX 0 #define regUTCL2_EDC_CONFIG 0x084c #define regUTCL2_EDC_CONFIG_BASE_IDX 0 +#define regVML2_UE_ERR_STATUS_LO 0x084d +#define regVML2_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regVML2_WALKER_UE_ERR_STATUS_LO 0x084e +#define regVML2_WALKER_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regUTCL2_UE_ERR_STATUS_LO 0x084f +#define regUTCL2_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regVML2_UE_ERR_STATUS_HI 0x0850 +#define regVML2_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regVML2_WALKER_UE_ERR_STATUS_HI 0x0851 +#define regVML2_WALKER_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regUTCL2_UE_ERR_STATUS_HI 0x0852 +#define regUTCL2_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regVML2_CE_ERR_STATUS_LO 0x0853 +#define regVML2_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regVML2_WALKER_CE_ERR_STATUS_LO 0x0854 +#define regVML2_WALKER_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regUTCL2_CE_ERR_STATUS_LO 0x0855 +#define regUTCL2_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regVML2_CE_ERR_STATUS_HI 0x0856 +#define regVML2_CE_ERR_STATUS_HI_BASE_IDX 0 +#define regVML2_WALKER_CE_ERR_STATUS_HI 0x0857 +#define regVML2_WALKER_CE_ERR_STATUS_HI_BASE_IDX 0 +#define regUTCL2_CE_ERR_STATUS_HI 0x0858 +#define regUTCL2_CE_ERR_STATUS_HI_BASE_IDX 0 // addressBlock: xcd0_gc_utcl2_vml2vcdec @@ -2011,6 +2123,22 @@ #define regTC_CFG_L1_VOLATILE_BASE_IDX 0 #define regTC_CFG_L2_VOLATILE 0x0b23 #define regTC_CFG_L2_VOLATILE_BASE_IDX 0 +#define regTCP_UE_EDC_HI_REG 0x0b54 +#define regTCP_UE_EDC_HI_REG_BASE_IDX 0 +#define regTCP_UE_EDC_LO_REG 0x0b55 +#define regTCP_UE_EDC_LO_REG_BASE_IDX 0 +#define regTCP_CE_EDC_HI_REG 0x0b56 +#define regTCP_CE_EDC_HI_REG_BASE_IDX 0 +#define regTCP_CE_EDC_LO_REG 0x0b57 +#define regTCP_CE_EDC_LO_REG_BASE_IDX 0 +#define regTCI_UE_EDC_HI_REG 0x0b58 +#define regTCI_UE_EDC_HI_REG_BASE_IDX 0 +#define regTCI_UE_EDC_LO_REG 0x0b59 +#define regTCI_UE_EDC_LO_REG_BASE_IDX 0 +#define regTCI_CE_EDC_HI_REG 0x0b5a +#define regTCI_CE_EDC_HI_REG_BASE_IDX 0 +#define regTCI_CE_EDC_LO_REG 0x0b5b +#define regTCI_CE_EDC_LO_REG_BASE_IDX 0 #define regTCI_MISC 0x0b5c #define regTCI_MISC_BASE_IDX 0 #define regTCI_CNTL_3 0x0b5d @@ -2061,6 +2189,26 @@ #define regTCX_DSM_CNTL_BASE_IDX 0 #define regTCX_DSM_CNTL2 0x0bc8 #define regTCX_DSM_CNTL2_BASE_IDX 0 +#define regTCA_UE_ERR_STATUS_LO 0x0bc9 +#define regTCA_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regTCA_UE_ERR_STATUS_HI 0x0bca +#define regTCA_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regTCX_UE_ERR_STATUS_LO 0x0bcb +#define regTCX_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regTCX_UE_ERR_STATUS_HI 0x0bcc +#define regTCX_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regTCX_CE_ERR_STATUS_LO 0x0bcd +#define regTCX_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regTCX_CE_ERR_STATUS_HI 0x0bce +#define regTCX_CE_ERR_STATUS_HI_BASE_IDX 0 +#define regTCC_UE_ERR_STATUS_LO 0x0bcf +#define regTCC_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regTCC_UE_ERR_STATUS_HI 0x0bd0 +#define regTCC_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regTCC_CE_ERR_STATUS_LO 0x0bd1 +#define regTCC_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regTCC_CE_ERR_STATUS_HI 0x0bd2 +#define regTCC_CE_ERR_STATUS_HI_BASE_IDX 0 // addressBlock: xcd0_gc_shdec @@ -2905,6 +3053,30 @@ #define regCP_MEC2_F32_INT_DIS_BASE_IDX 0 #define regCP_VMID_STATUS 0x10bf #define regCP_VMID_STATUS_BASE_IDX 0 +#define regCPC_UE_ERR_STATUS_LO 0x10e0 +#define regCPC_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regCPC_UE_ERR_STATUS_HI 0x10e1 +#define regCPC_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regCPC_CE_ERR_STATUS_LO 0x10e2 +#define regCPC_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regCPC_CE_ERR_STATUS_HI 0x10e3 +#define regCPC_CE_ERR_STATUS_HI_BASE_IDX 0 +#define regCPF_UE_ERR_STATUS_LO 0x10e4 +#define regCPF_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regCPF_UE_ERR_STATUS_HI 0x10e5 +#define regCPF_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regCPF_CE_ERR_STATUS_LO 0x10e6 +#define regCPF_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regCPF_CE_ERR_STATUS_HI 0x10e7 +#define regCPF_CE_ERR_STATUS_HI_BASE_IDX 0 +#define regCPG_UE_ERR_STATUS_LO 0x10e8 +#define regCPG_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regCPG_UE_ERR_STATUS_HI 0x10e9 +#define regCPG_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regCPG_CE_ERR_STATUS_LO 0x10ea +#define regCPG_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regCPG_CE_ERR_STATUS_HI 0x10eb +#define regCPG_CE_ERR_STATUS_HI_BASE_IDX 0 // addressBlock: xcd0_gc_cppdec2 @@ -5364,6 +5536,18 @@ #define regSPI_WAVE_LIMIT_CNTL 0x2443 #define regSPI_WAVE_LIMIT_CNTL_BASE_IDX 1 +// addressBlock: xcd0_gc_gccanedec +// base address: 0x33d00 +#define regGC_CANE_ERR_STATUS 0x2f4d +#define regGC_CANE_ERR_STATUS_BASE_IDX 1 +#define regGC_CANE_UE_ERR_STATUS_LO 0x2f4e +#define regGC_CANE_UE_ERR_STATUS_LO_BASE_IDX 1 +#define regGC_CANE_UE_ERR_STATUS_HI 0x2f4f +#define regGC_CANE_UE_ERR_STATUS_HI_BASE_IDX 1 +#define regGC_CANE_CE_ERR_STATUS_LO 0x2f50 +#define regGC_CANE_CE_ERR_STATUS_LO_BASE_IDX 1 +#define regGC_CANE_CE_ERR_STATUS_HI 0x2f51 +#define regGC_CANE_CE_ERR_STATUS_HI_BASE_IDX 1 // addressBlock: xcd0_gc_perfddec // base address: 0x34000 @@ -6583,6 +6767,10 @@ #define regRLC_CAPTURE_GPU_CLOCK_COUNT_2_BASE_IDX 1 #define regRLC_CPG_STAT_INVAL 0x4d09 #define regRLC_CPG_STAT_INVAL_BASE_IDX 1 +#define regRLC_UE_ERR_STATUS_LOW 0x4d40 +#define regRLC_UE_ERR_STATUS_LOW_BASE_IDX 1 +#define regRLC_UE_ERR_STATUS_HIGH 0x4d41 +#define regRLC_UE_ERR_STATUS_HIGH_BASE_IDX 1 #define regRLC_DSM_CNTL 0x4d42 #define regRLC_DSM_CNTL_BASE_IDX 1 #define regRLC_DSM_CNTLA 0x4d43 @@ -6591,6 +6779,10 @@ #define regRLC_DSM_CNTL2_BASE_IDX 1 #define regRLC_DSM_CNTL2A 0x4d45 #define regRLC_DSM_CNTL2A_BASE_IDX 1 +#define regRLC_CE_ERR_STATUS_LOW 0x4d49 +#define regRLC_CE_ERR_STATUS_LOW_BASE_IDX 1 +#define regRLC_CE_ERR_STATUS_HIGH 0x4d4a +#define regRLC_CE_ERR_STATUS_HIGH_BASE_IDX 1 #define regRLC_RLCV_SPARE_INT 0x4f30 #define regRLC_RLCV_SPARE_INT_BASE_IDX 1 #define regRLC_SMU_CLK_REQ 0x4f97 diff --git a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_9_4_3_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_9_4_3_sh_mask.h index 84a75b58347f..2bd9f3f1026f 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_9_4_3_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_9_4_3_sh_mask.h @@ -4129,6 +4129,240 @@ #define SQC_DCACHE_UTCL1_STATUS__FAULT_DETECTED_MASK 0x00000001L #define SQC_DCACHE_UTCL1_STATUS__RETRY_DETECTED_MASK 0x00000002L #define SQC_DCACHE_UTCL1_STATUS__PRT_DETECTED_MASK 0x00000004L +//SQC_UE_EDC_LO +#define SQC_UE_EDC_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define SQC_UE_EDC_LO__ADDRESS_INFO_VALID_FLAG__SHIFT 0x1 +#define SQC_UE_EDC_LO__ADDRESS__SHIFT 0x2 +#define SQC_UE_EDC_LO__MEM_ID__SHIFT 0x18 +#define SQC_UE_EDC_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define SQC_UE_EDC_LO__ADDRESS_INFO_VALID_FLAG_MASK 0x00000002L +#define SQC_UE_EDC_LO__ADDRESS_MASK 0x00FFFFFCL +#define SQC_UE_EDC_LO__MEM_ID_MASK 0xFF000000L +//SQC_UE_EDC_HI +#define SQC_UE_EDC_HI__ECC__SHIFT 0x0 +#define SQC_UE_EDC_HI__PARITY__SHIFT 0x1 +#define SQC_UE_EDC_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define SQC_UE_EDC_HI__ERR_INFO__SHIFT 0x3 +#define SQC_UE_EDC_HI__UE_CNT__SHIFT 0x17 +#define SQC_UE_EDC_HI__FED_CNT__SHIFT 0x1a +#define SQC_UE_EDC_HI__ECC_MASK 0x00000001L +#define SQC_UE_EDC_HI__PARITY_MASK 0x00000002L +#define SQC_UE_EDC_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define SQC_UE_EDC_HI__ERR_INFO_MASK 0x007FFFF8L +#define SQC_UE_EDC_HI__UE_CNT_MASK 0x03800000L +#define SQC_UE_EDC_HI__FED_CNT_MASK 0x1C000000L +//SQC_CE_EDC_LO +#define SQC_CE_EDC_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define SQC_CE_EDC_LO__ADDRESS_INFO_VALID_FLAG__SHIFT 0x1 +#define SQC_CE_EDC_LO__ADDRESS__SHIFT 0x2 +#define SQC_CE_EDC_LO__MEM_ID__SHIFT 0x18 +#define SQC_CE_EDC_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define SQC_CE_EDC_LO__ADDRESS_INFO_VALID_FLAG_MASK 0x00000002L +#define SQC_CE_EDC_LO__ADDRESS_MASK 0x00FFFFFCL +#define SQC_CE_EDC_LO__MEM_ID_MASK 0xFF000000L +//SQC_CE_EDC_HI +#define SQC_CE_EDC_HI__ECC__SHIFT 0x0 +#define SQC_CE_EDC_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define SQC_CE_EDC_HI__ERR_INFO__SHIFT 0x3 +#define SQC_CE_EDC_HI__CE_CNT__SHIFT 0x17 +#define SQC_CE_EDC_HI__POSION__SHIFT 0x1a +#define SQC_CE_EDC_HI__ECC_MASK 0x00000001L +#define SQC_CE_EDC_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define SQC_CE_EDC_HI__ERR_INFO_MASK 0x007FFFF8L +#define SQC_CE_EDC_HI__CE_CNT_MASK 0x03800000L +#define SQC_CE_EDC_HI__POSION_MASK 0x04000000L +//SQ_UE_ERR_STATUS_LO +#define SQ_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define SQ_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define SQ_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define SQ_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define SQ_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define SQ_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define SQ_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define SQ_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//SQ_UE_ERR_STATUS_HI +#define SQ_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define SQ_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define SQ_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define SQ_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define SQ_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define SQ_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define SQ_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define SQ_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define SQ_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define SQ_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define SQ_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define SQ_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define SQ_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define SQ_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L +//SQ_CE_ERR_STATUS_LO +#define SQ_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define SQ_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define SQ_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define SQ_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define SQ_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define SQ_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define SQ_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define SQ_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//SQ_CE_ERR_STATUS_HI +#define SQ_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define SQ_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define SQ_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define SQ_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define SQ_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define SQ_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define SQ_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define SQ_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define SQ_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define SQ_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define SQ_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define SQ_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define SQ_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define SQ_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L +//LDS_UE_ERR_STATUS_LO +#define LDS_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define LDS_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define LDS_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define LDS_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define LDS_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define LDS_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define LDS_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define LDS_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//LDS_UE_ERR_STATUS_HI +#define LDS_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define LDS_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define LDS_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define LDS_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define LDS_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define LDS_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define LDS_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define LDS_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define LDS_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define LDS_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define LDS_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define LDS_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define LDS_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define LDS_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L +//LDS_CE_ERR_STATUS_LO +#define LDS_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define LDS_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define LDS_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define LDS_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define LDS_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define LDS_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define LDS_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define LDS_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//LDS_CE_ERR_STATUS_HI +#define LDS_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define LDS_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define LDS_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define LDS_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define LDS_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define LDS_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define LDS_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define LDS_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define LDS_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define LDS_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define LDS_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define LDS_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define LDS_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define LDS_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L +//SP0_UE_ERR_STATUS_LO +#define SP0_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define SP0_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define SP0_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define SP0_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define SP0_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define SP0_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define SP0_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define SP0_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//SP0_UE_ERR_STATUS_HI +#define SP0_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define SP0_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define SP0_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define SP0_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define SP0_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define SP0_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define SP0_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define SP0_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define SP0_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define SP0_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define SP0_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define SP0_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define SP0_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define SP0_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L +//SP0_CE_ERR_STATUS_LO +#define SP0_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define SP0_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define SP0_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define SP0_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define SP0_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define SP0_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define SP0_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define SP0_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//SP0_CE_ERR_STATUS_HI +#define SP0_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define SP0_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define SP0_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define SP0_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define SP0_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define SP0_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define SP0_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define SP0_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define SP0_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define SP0_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define SP0_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define SP0_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define SP0_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define SP0_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L +//SP1_UE_ERR_STATUS_LO +#define SP1_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define SP1_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define SP1_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define SP1_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define SP1_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define SP1_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define SP1_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define SP1_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//SP1_UE_ERR_STATUS_HI +#define SP1_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define SP1_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define SP1_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define SP1_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define SP1_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define SP1_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define SP1_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define SP1_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define SP1_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define SP1_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define SP1_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define SP1_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define SP1_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define SP1_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L +//SP1_CE_ERR_STATUS_LO +#define SP1_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define SP1_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define SP1_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define SP1_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define SP1_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define SP1_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define SP1_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define SP1_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//SP1_CE_ERR_STATUS_HI +#define SP1_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define SP1_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define SP1_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define SP1_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define SP1_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define SP1_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define SP1_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define SP1_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define SP1_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define SP1_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define SP1_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define SP1_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define SP1_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define SP1_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L // addressBlock: xcd0_gc_shsdec @@ -4235,6 +4469,54 @@ #define SPI_EDC_CNT__SPI_LIFE_CNT_SEC_COUNT_MASK 0x00030000L #define SPI_EDC_CNT__SPI_LIFE_CNT_DED_COUNT_MASK 0x000C0000L #define SPI_EDC_CNT__UNUSED_MASK 0xFFF00000L +//SPI_UE_ERR_STATUS_LO +#define SPI_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define SPI_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define SPI_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define SPI_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define SPI_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define SPI_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define SPI_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define SPI_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//SPI_UE_ERR_STATUS_HI +#define SPI_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define SPI_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define SPI_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define SPI_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define SPI_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define SPI_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define SPI_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define SPI_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define SPI_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define SPI_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define SPI_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define SPI_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define SPI_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define SPI_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L +//SPI_CE_ERR_STATUS_LO +#define SPI_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define SPI_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define SPI_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define SPI_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define SPI_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define SPI_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define SPI_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define SPI_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//SPI_CE_ERR_STATUS_HI +#define SPI_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define SPI_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define SPI_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define SPI_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define SPI_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define SPI_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define SPI_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define SPI_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define SPI_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define SPI_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define SPI_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define SPI_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define SPI_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define SPI_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L //SPI_DEBUG_BUSY #define SPI_DEBUG_BUSY__HS_BUSY__SHIFT 0x0 #define SPI_DEBUG_BUSY__GS_BUSY__SHIFT 0x1 @@ -4622,6 +4904,48 @@ #define TD_POWER_CNTL__MGCG_OUTPUTSTAGE_MASK 0x00000002L #define TD_POWER_CNTL__MID0_THREAD_DATA_MASK 0x00000004L #define TD_POWER_CNTL__MID2_ACCUM_DATA_MASK 0x00000008L +//TD_UE_EDC_LO +#define TD_UE_EDC_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define TD_UE_EDC_LO__ADDRESS_INFO_VALID_FLAG__SHIFT 0x1 +#define TD_UE_EDC_LO__ADDRESS__SHIFT 0x2 +#define TD_UE_EDC_LO__MEM_ID__SHIFT 0x18 +#define TD_UE_EDC_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define TD_UE_EDC_LO__ADDRESS_INFO_VALID_FLAG_MASK 0x00000002L +#define TD_UE_EDC_LO__ADDRESS_MASK 0x00FFFFFCL +#define TD_UE_EDC_LO__MEM_ID_MASK 0xFF000000L +//TD_UE_EDC_HI +#define TD_UE_EDC_HI__ECC__SHIFT 0x0 +#define TD_UE_EDC_HI__PARITY__SHIFT 0x1 +#define TD_UE_EDC_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define TD_UE_EDC_HI__ERR_INFO__SHIFT 0x3 +#define TD_UE_EDC_HI__UE_CNT__SHIFT 0x17 +#define TD_UE_EDC_HI__FED_CNT__SHIFT 0x1a +#define TD_UE_EDC_HI__ECC_MASK 0x00000001L +#define TD_UE_EDC_HI__PARITY_MASK 0x00000002L +#define TD_UE_EDC_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define TD_UE_EDC_HI__ERR_INFO_MASK 0x007FFFF8L +#define TD_UE_EDC_HI__UE_CNT_MASK 0x03800000L +#define TD_UE_EDC_HI__FED_CNT_MASK 0x1C000000L +//TD_CE_EDC_LO +#define TD_CE_EDC_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define TD_CE_EDC_LO__ADDRESS_INFO_VALID_FLAG__SHIFT 0x1 +#define TD_CE_EDC_LO__ADDRESS__SHIFT 0x2 +#define TD_CE_EDC_LO__MEM_ID__SHIFT 0x18 +#define TD_CE_EDC_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define TD_CE_EDC_LO__ADDRESS_INFO_VALID_FLAG_MASK 0x00000002L +#define TD_CE_EDC_LO__ADDRESS_MASK 0x00FFFFFCL +#define TD_CE_EDC_LO__MEM_ID_MASK 0xFF000000L +//TD_CE_EDC_HI +#define TD_CE_EDC_HI__ECC__SHIFT 0x0 +#define TD_CE_EDC_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define TD_CE_EDC_HI__ERR_INFO__SHIFT 0x3 +#define TD_CE_EDC_HI__CE_CNT__SHIFT 0x17 +#define TD_CE_EDC_HI__POISON__SHIFT 0x1a +#define TD_CE_EDC_HI__ECC_MASK 0x00000001L +#define TD_CE_EDC_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define TD_CE_EDC_HI__ERR_INFO_MASK 0x007FFFF8L +#define TD_CE_EDC_HI__CE_CNT_MASK 0x03800000L +#define TD_CE_EDC_HI__POISON_MASK 0x04000000L //TD_DSM_CNTL #define TD_DSM_CNTL__TD_SS_FIFO_LO_DSM_IRRITATOR_DATA__SHIFT 0x0 #define TD_DSM_CNTL__TD_SS_FIFO_LO_ENABLE_SINGLE_WRITE__SHIFT 0x2 @@ -4771,6 +5095,48 @@ #define TA_DSM_CNTL2__TA_FS_AFIFO_HI_ENABLE_ERROR_INJECT_MASK 0x000C0000L #define TA_DSM_CNTL2__TA_FS_AFIFO_HI_SELECT_INJECT_DELAY_MASK 0x00100000L #define TA_DSM_CNTL2__TA_INJECT_DELAY_MASK 0xFC000000L +//TA_UE_EDC_LO +#define TA_UE_EDC_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define TA_UE_EDC_LO__ADDRESS_INFO_VALID_FLAG__SHIFT 0x1 +#define TA_UE_EDC_LO__ADDRESS__SHIFT 0x2 +#define TA_UE_EDC_LO__MEM_ID__SHIFT 0x18 +#define TA_UE_EDC_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define TA_UE_EDC_LO__ADDRESS_INFO_VALID_FLAG_MASK 0x00000002L +#define TA_UE_EDC_LO__ADDRESS_MASK 0x00FFFFFCL +#define TA_UE_EDC_LO__MEM_ID_MASK 0xFF000000L +//TA_UE_EDC_HI +#define TA_UE_EDC_HI__ECC__SHIFT 0x0 +#define TA_UE_EDC_HI__PARITY__SHIFT 0x1 +#define TA_UE_EDC_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define TA_UE_EDC_HI__ERR_INFO__SHIFT 0x3 +#define TA_UE_EDC_HI__UE_CNT__SHIFT 0x17 +#define TA_UE_EDC_HI__FED_CNT__SHIFT 0x1a +#define TA_UE_EDC_HI__ECC_MASK 0x00000001L +#define TA_UE_EDC_HI__PARITY_MASK 0x00000002L +#define TA_UE_EDC_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define TA_UE_EDC_HI__ERR_INFO_MASK 0x007FFFF8L +#define TA_UE_EDC_HI__UE_CNT_MASK 0x03800000L +#define TA_UE_EDC_HI__FED_CNT_MASK 0x1C000000L +//TA_CE_EDC_LO +#define TA_CE_EDC_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define TA_CE_EDC_LO__ADDRESS_INFO_VALID_FLAG__SHIFT 0x1 +#define TA_CE_EDC_LO__ADDRESS__SHIFT 0x2 +#define TA_CE_EDC_LO__MEM_ID__SHIFT 0x18 +#define TA_CE_EDC_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define TA_CE_EDC_LO__ADDRESS_INFO_VALID_FLAG_MASK 0x00000002L +#define TA_CE_EDC_LO__ADDRESS_MASK 0x00FFFFFCL +#define TA_CE_EDC_LO__MEM_ID_MASK 0xFF000000L +//TA_CE_EDC_HI +#define TA_CE_EDC_HI__ECC__SHIFT 0x0 +#define TA_CE_EDC_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define TA_CE_EDC_HI__ERR_INFO__SHIFT 0x3 +#define TA_CE_EDC_HI__CE_CNT__SHIFT 0x17 +#define TA_CE_EDC_HI__POISON__SHIFT 0x1a +#define TA_CE_EDC_HI__ECC_MASK 0x00000001L +#define TA_CE_EDC_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define TA_CE_EDC_HI__ERR_INFO_MASK 0x007FFFF8L +#define TA_CE_EDC_HI__CE_CNT_MASK 0x03800000L +#define TA_CE_EDC_HI__POISON_MASK 0x04000000L // addressBlock: xcd0_gc_gdsdec @@ -5015,6 +5381,54 @@ #define GDS_WD_GDS_CSB__UNUSED__SHIFT 0xd #define GDS_WD_GDS_CSB__COUNTER_MASK 0x00001FFFL #define GDS_WD_GDS_CSB__UNUSED_MASK 0xFFFFE000L +//GDS_UE_ERR_STATUS_LO +#define GDS_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define GDS_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define GDS_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define GDS_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define GDS_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define GDS_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define GDS_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define GDS_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//GDS_UE_ERR_STATUS_HI +#define GDS_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define GDS_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define GDS_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define GDS_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define GDS_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define GDS_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define GDS_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define GDS_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define GDS_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define GDS_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define GDS_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define GDS_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define GDS_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define GDS_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L +//GDS_CE_ERR_STATUS_LO +#define GDS_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define GDS_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define GDS_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define GDS_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define GDS_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define GDS_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define GDS_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define GDS_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//GDS_CE_ERR_STATUS_HI +#define GDS_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define GDS_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define GDS_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define GDS_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define GDS_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define GDS_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define GDS_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define GDS_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define GDS_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define GDS_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define GDS_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define GDS_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define GDS_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define GDS_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L // addressBlock: xcd0_gc_rbdec @@ -7370,6 +7784,30 @@ #define GCEA_MAM_CTRL2__ARAM_FLUSH_NOALLOC_MASK 0x00000040L #define GCEA_MAM_CTRL2__RESERVED_FIELD_MASK 0x00FFFF80L #define GCEA_MAM_CTRL2__ADDR_HI_MASK 0xFF000000L +//GCEA_UE_ERR_STATUS_LO +#define GCEA_UE_ERR_STATUS_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define GCEA_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define GCEA_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define GCEA_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define GCEA_UE_ERR_STATUS_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define GCEA_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define GCEA_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define GCEA_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//GCEA_UE_ERR_STATUS_HI +#define GCEA_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define GCEA_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define GCEA_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define GCEA_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define GCEA_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define GCEA_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define GCEA_UE_ERR_STATUS_HI__RESERVED_FIELD__SHIFT 0x1d +#define GCEA_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define GCEA_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define GCEA_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define GCEA_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define GCEA_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define GCEA_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define GCEA_UE_ERR_STATUS_HI__RESERVED_FIELD_MASK 0xE0000000L //GCEA_DSM_CNTL #define GCEA_DSM_CNTL__DRAMRD_CMDMEM_DSM_IRRITATOR_DATA__SHIFT 0x0 #define GCEA_DSM_CNTL__DRAMRD_CMDMEM_ENABLE_SINGLE_WRITE__SHIFT 0x2 @@ -7745,6 +8183,30 @@ #define GCEA_SDP_BACKDOOR_MISCCREDITS__WRRSP_CREDITS_RELEASED_MASK 0x0000FF00L #define GCEA_SDP_BACKDOOR_MISCCREDITS__PRB_REQ_CREDITS_RELEASED_MASK 0x007F0000L #define GCEA_SDP_BACKDOOR_MISCCREDITS__PRB_RSP_CREDITS_RECEIVED_MASK 0x3F800000L +//GCEA_CE_ERR_STATUS_LO +#define GCEA_CE_ERR_STATUS_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define GCEA_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define GCEA_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define GCEA_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define GCEA_CE_ERR_STATUS_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define GCEA_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define GCEA_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define GCEA_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//GCEA_CE_ERR_STATUS_HI +#define GCEA_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define GCEA_CE_ERR_STATUS_HI__RESERVED_FIELD0__SHIFT 0x1 +#define GCEA_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define GCEA_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define GCEA_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define GCEA_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define GCEA_CE_ERR_STATUS_HI__RESERVED_FIELD1__SHIFT 0x1b +#define GCEA_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define GCEA_CE_ERR_STATUS_HI__RESERVED_FIELD0_MASK 0x00000002L +#define GCEA_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define GCEA_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define GCEA_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define GCEA_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define GCEA_CE_ERR_STATUS_HI__RESERVED_FIELD1_MASK 0xF8000000L //GCEA_SDP_ENABLE #define GCEA_SDP_ENABLE__ENABLE__SHIFT 0x0 #define GCEA_SDP_ENABLE__ENABLE_MASK 0x00000001L @@ -8440,6 +8902,54 @@ //ATC_L2_MM_GROUP_RT_CLASSES #define ATC_L2_MM_GROUP_RT_CLASSES__GROUP_RT_CLASS__SHIFT 0x0 #define ATC_L2_MM_GROUP_RT_CLASSES__GROUP_RT_CLASS_MASK 0xFFFFFFFFL +//ATC_L2_UE_ERR_STATUS_LO +#define ATC_L2_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define ATC_L2_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define ATC_L2_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define ATC_L2_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define ATC_L2_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define ATC_L2_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define ATC_L2_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define ATC_L2_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//ATC_L2_UE_ERR_STATUS_HI +#define ATC_L2_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define ATC_L2_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define ATC_L2_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define ATC_L2_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define ATC_L2_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define ATC_L2_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define ATC_L2_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define ATC_L2_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define ATC_L2_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define ATC_L2_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define ATC_L2_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define ATC_L2_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define ATC_L2_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define ATC_L2_UE_ERR_STATUS_HI__RESERVED_MASK 0x60000000L +//ATC_L2_CE_ERR_STATUS_LO +#define ATC_L2_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define ATC_L2_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define ATC_L2_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define ATC_L2_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define ATC_L2_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define ATC_L2_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define ATC_L2_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define ATC_L2_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//ATC_L2_CE_ERR_STATUS_HI +#define ATC_L2_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define ATC_L2_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define ATC_L2_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define ATC_L2_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define ATC_L2_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define ATC_L2_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define ATC_L2_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define ATC_L2_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define ATC_L2_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define ATC_L2_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define ATC_L2_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define ATC_L2_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define ATC_L2_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define ATC_L2_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L // addressBlock: xcd0_gc_utcl2_vml2pfdec @@ -8888,6 +9398,150 @@ #define UTCL2_EDC_CONFIG__DIS_EDC__SHIFT 0x1 #define UTCL2_EDC_CONFIG__WRITE_DIS_MASK 0x00000001L #define UTCL2_EDC_CONFIG__DIS_EDC_MASK 0x00000002L +//VML2_UE_ERR_STATUS_LO +#define VML2_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define VML2_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define VML2_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define VML2_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define VML2_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define VML2_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define VML2_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define VML2_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//VML2_WALKER_UE_ERR_STATUS_LO +#define VML2_WALKER_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define VML2_WALKER_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define VML2_WALKER_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define VML2_WALKER_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define VML2_WALKER_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define VML2_WALKER_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define VML2_WALKER_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define VML2_WALKER_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//UTCL2_UE_ERR_STATUS_LO +#define UTCL2_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define UTCL2_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define UTCL2_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define UTCL2_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define UTCL2_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define UTCL2_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define UTCL2_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define UTCL2_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//VML2_UE_ERR_STATUS_HI +#define VML2_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define VML2_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define VML2_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define VML2_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define VML2_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define VML2_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define VML2_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define VML2_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define VML2_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define VML2_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define VML2_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define VML2_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define VML2_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define VML2_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L +//VML2_WALKER_UE_ERR_STATUS_HI +#define VML2_WALKER_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define VML2_WALKER_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define VML2_WALKER_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define VML2_WALKER_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define VML2_WALKER_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define VML2_WALKER_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define VML2_WALKER_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define VML2_WALKER_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define VML2_WALKER_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define VML2_WALKER_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define VML2_WALKER_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define VML2_WALKER_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define VML2_WALKER_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define VML2_WALKER_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L +//UTCL2_UE_ERR_STATUS_HI +#define UTCL2_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define UTCL2_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define UTCL2_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define UTCL2_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define UTCL2_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define UTCL2_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define UTCL2_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define UTCL2_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define UTCL2_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define UTCL2_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define UTCL2_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define UTCL2_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define UTCL2_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define UTCL2_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L +//VML2_CE_ERR_STATUS_LO +#define VML2_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define VML2_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define VML2_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define VML2_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define VML2_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define VML2_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define VML2_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define VML2_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//VML2_WALKER_CE_ERR_STATUS_LO +#define VML2_WALKER_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define VML2_WALKER_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define VML2_WALKER_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define VML2_WALKER_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define VML2_WALKER_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define VML2_WALKER_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define VML2_WALKER_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define VML2_WALKER_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//UTCL2_CE_ERR_STATUS_LO +#define UTCL2_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define UTCL2_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define UTCL2_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define UTCL2_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define UTCL2_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define UTCL2_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define UTCL2_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define UTCL2_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//VML2_CE_ERR_STATUS_HI +#define VML2_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define VML2_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define VML2_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define VML2_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define VML2_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define VML2_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define VML2_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define VML2_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define VML2_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define VML2_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define VML2_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define VML2_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define VML2_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define VML2_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L +//VML2_WALKER_CE_ERR_STATUS_HI +#define VML2_WALKER_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define VML2_WALKER_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define VML2_WALKER_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define VML2_WALKER_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define VML2_WALKER_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define VML2_WALKER_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define VML2_WALKER_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define VML2_WALKER_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define VML2_WALKER_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define VML2_WALKER_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define VML2_WALKER_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define VML2_WALKER_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define VML2_WALKER_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define VML2_WALKER_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L +//UTCL2_CE_ERR_STATUS_HI +#define UTCL2_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define UTCL2_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define UTCL2_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define UTCL2_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define UTCL2_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define UTCL2_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define UTCL2_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define UTCL2_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define UTCL2_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define UTCL2_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define UTCL2_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define UTCL2_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define UTCL2_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define UTCL2_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L // addressBlock: xcd0_gc_utcl2_vml2vcdec @@ -11139,6 +11793,98 @@ //TC_CFG_L2_VOLATILE #define TC_CFG_L2_VOLATILE__VOL__SHIFT 0x0 #define TC_CFG_L2_VOLATILE__VOL_MASK 0x0000000FL +//TCP_UE_EDC_HI_REG +#define TCP_UE_EDC_HI_REG__ECC__SHIFT 0x0 +#define TCP_UE_EDC_HI_REG__PARITY__SHIFT 0x1 +#define TCP_UE_EDC_HI_REG__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define TCP_UE_EDC_HI_REG__ERR_INFO__SHIFT 0x3 +#define TCP_UE_EDC_HI_REG__UE_CNT__SHIFT 0x17 +#define TCP_UE_EDC_HI_REG__FED_CNT__SHIFT 0x1a +#define TCP_UE_EDC_HI_REG__RESERVED__SHIFT 0x1d +#define TCP_UE_EDC_HI_REG__ECC_MASK 0x00000001L +#define TCP_UE_EDC_HI_REG__PARITY_MASK 0x00000002L +#define TCP_UE_EDC_HI_REG__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define TCP_UE_EDC_HI_REG__ERR_INFO_MASK 0x007FFFF8L +#define TCP_UE_EDC_HI_REG__UE_CNT_MASK 0x03800000L +#define TCP_UE_EDC_HI_REG__FED_CNT_MASK 0x1C000000L +#define TCP_UE_EDC_HI_REG__RESERVED_MASK 0xE0000000L +//TCP_UE_EDC_LO_REG +#define TCP_UE_EDC_LO_REG__STATUS_VALID_FLAG__SHIFT 0x0 +#define TCP_UE_EDC_LO_REG__ADDRESS_INFO_VALID_FLAG__SHIFT 0x1 +#define TCP_UE_EDC_LO_REG__ADDRESS__SHIFT 0x2 +#define TCP_UE_EDC_LO_REG__MEM_ID__SHIFT 0x18 +#define TCP_UE_EDC_LO_REG__STATUS_VALID_FLAG_MASK 0x00000001L +#define TCP_UE_EDC_LO_REG__ADDRESS_INFO_VALID_FLAG_MASK 0x00000002L +#define TCP_UE_EDC_LO_REG__ADDRESS_MASK 0x00FFFFFCL +#define TCP_UE_EDC_LO_REG__MEM_ID_MASK 0xFF000000L +//TCP_CE_EDC_HI_REG +#define TCP_CE_EDC_HI_REG__ECC__SHIFT 0x0 +#define TCP_CE_EDC_HI_REG__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define TCP_CE_EDC_HI_REG__ERR_INFO__SHIFT 0x3 +#define TCP_CE_EDC_HI_REG__CE_CNT__SHIFT 0x17 +#define TCP_CE_EDC_HI_REG__POISON__SHIFT 0x1a +#define TCP_CE_EDC_HI_REG__RESERVED__SHIFT 0x1b +#define TCP_CE_EDC_HI_REG__ECC_MASK 0x00000001L +#define TCP_CE_EDC_HI_REG__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define TCP_CE_EDC_HI_REG__ERR_INFO_MASK 0x007FFFF8L +#define TCP_CE_EDC_HI_REG__CE_CNT_MASK 0x03800000L +#define TCP_CE_EDC_HI_REG__POISON_MASK 0x04000000L +#define TCP_CE_EDC_HI_REG__RESERVED_MASK 0xF8000000L +//TCP_CE_EDC_LO_REG +#define TCP_CE_EDC_LO_REG__STATUS_VALID_FLAG__SHIFT 0x0 +#define TCP_CE_EDC_LO_REG__ADDRESS_INFO_VALID_FLAG__SHIFT 0x1 +#define TCP_CE_EDC_LO_REG__ADDRESS__SHIFT 0x2 +#define TCP_CE_EDC_LO_REG__MEM_ID__SHIFT 0x18 +#define TCP_CE_EDC_LO_REG__STATUS_VALID_FLAG_MASK 0x00000001L +#define TCP_CE_EDC_LO_REG__ADDRESS_INFO_VALID_FLAG_MASK 0x00000002L +#define TCP_CE_EDC_LO_REG__ADDRESS_MASK 0x00FFFFFCL +#define TCP_CE_EDC_LO_REG__MEM_ID_MASK 0xFF000000L +//TCI_UE_EDC_HI_REG +#define TCI_UE_EDC_HI_REG__ECC__SHIFT 0x0 +#define TCI_UE_EDC_HI_REG__PARITY__SHIFT 0x1 +#define TCI_UE_EDC_HI_REG__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define TCI_UE_EDC_HI_REG__ERR_INFO__SHIFT 0x3 +#define TCI_UE_EDC_HI_REG__UE_CNT__SHIFT 0x17 +#define TCI_UE_EDC_HI_REG__FED_CNT__SHIFT 0x1a +#define TCI_UE_EDC_HI_REG__RESERVED__SHIFT 0x1d +#define TCI_UE_EDC_HI_REG__ECC_MASK 0x00000001L +#define TCI_UE_EDC_HI_REG__PARITY_MASK 0x00000002L +#define TCI_UE_EDC_HI_REG__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define TCI_UE_EDC_HI_REG__ERR_INFO_MASK 0x007FFFF8L +#define TCI_UE_EDC_HI_REG__UE_CNT_MASK 0x03800000L +#define TCI_UE_EDC_HI_REG__FED_CNT_MASK 0x1C000000L +#define TCI_UE_EDC_HI_REG__RESERVED_MASK 0xE0000000L +//TCI_UE_EDC_LO_REG +#define TCI_UE_EDC_LO_REG__STATUS_VALID_FLAG__SHIFT 0x0 +#define TCI_UE_EDC_LO_REG__ADDRESS_INFO_VALID_FLAG__SHIFT 0x1 +#define TCI_UE_EDC_LO_REG__ADDRESS__SHIFT 0x2 +#define TCI_UE_EDC_LO_REG__MEM_ID__SHIFT 0x18 +#define TCI_UE_EDC_LO_REG__STATUS_VALID_FLAG_MASK 0x00000001L +#define TCI_UE_EDC_LO_REG__ADDRESS_INFO_VALID_FLAG_MASK 0x00000002L +#define TCI_UE_EDC_LO_REG__ADDRESS_MASK 0x00FFFFFCL +#define TCI_UE_EDC_LO_REG__MEM_ID_MASK 0xFF000000L +//TCI_CE_EDC_HI_REG +#define TCI_CE_EDC_HI_REG__ECC__SHIFT 0x0 +#define TCI_CE_EDC_HI_REG__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define TCI_CE_EDC_HI_REG__ERR_INFO__SHIFT 0x3 +#define TCI_CE_EDC_HI_REG__CE_CNT__SHIFT 0x17 +#define TCI_CE_EDC_HI_REG__POISON__SHIFT 0x1a +#define TCI_CE_EDC_HI_REG__RESERVED__SHIFT 0x1b +#define TCI_CE_EDC_HI_REG__ECC_MASK 0x00000001L +#define TCI_CE_EDC_HI_REG__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define TCI_CE_EDC_HI_REG__ERR_INFO_MASK 0x007FFFF8L +#define TCI_CE_EDC_HI_REG__CE_CNT_MASK 0x03800000L +#define TCI_CE_EDC_HI_REG__POISON_MASK 0x04000000L +#define TCI_CE_EDC_HI_REG__RESERVED_MASK 0xF8000000L +//TCI_CE_EDC_LO_REG +#define TCI_CE_EDC_LO_REG__STATUS_VALID_FLAG__SHIFT 0x0 +#define TCI_CE_EDC_LO_REG__ADDRESS_INFO_VALID_FLAG__SHIFT 0x1 +#define TCI_CE_EDC_LO_REG__ADDRESS__SHIFT 0x2 +#define TCI_CE_EDC_LO_REG__MEM_ID__SHIFT 0x18 +#define TCI_CE_EDC_LO_REG__STATUS_VALID_FLAG_MASK 0x00000001L +#define TCI_CE_EDC_LO_REG__ADDRESS_INFO_VALID_FLAG_MASK 0x00000002L +#define TCI_CE_EDC_LO_REG__ADDRESS_MASK 0x00FFFFFCL +#define TCI_CE_EDC_LO_REG__MEM_ID_MASK 0xFF000000L //TCI_MISC #define TCI_MISC__FGCG_REPEATER_DISABLE__SHIFT 0x0 #define TCI_MISC__LEGACY_MGCG_DISABLE__SHIFT 0x1 @@ -11560,6 +12306,112 @@ #define TCX_DSM_CNTL2__SED_ENABLE_ERROR_INJECT_MASK 0x00000003L #define TCX_DSM_CNTL2__SED_SELECT_INJECT_DELAY_MASK 0x00000004L #define TCX_DSM_CNTL2__INJECT_DELAY_MASK 0xFC000000L +//TCA_UE_ERR_STATUS_LO +#define TCA_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define TCA_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define TCA_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define TCA_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define TCA_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define TCA_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define TCA_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define TCA_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//TCA_UE_ERR_STATUS_HI +#define TCA_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define TCA_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define TCA_UE_ERR_STATUS_HI__ERROR_INFO_VALID_FLAG__SHIFT 0x2 +#define TCA_UE_ERR_STATUS_HI__ERROR_INFO__SHIFT 0x3 +#define TCA_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define TCA_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define TCA_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define TCA_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define TCA_UE_ERR_STATUS_HI__ERROR_INFO_VALID_FLAG_MASK 0x00000004L +#define TCA_UE_ERR_STATUS_HI__ERROR_INFO_MASK 0x007FFFF8L +#define TCA_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define TCA_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +//TCX_UE_ERR_STATUS_LO +#define TCX_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define TCX_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define TCX_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define TCX_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define TCX_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define TCX_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define TCX_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define TCX_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//TCX_UE_ERR_STATUS_HI +#define TCX_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define TCX_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define TCX_UE_ERR_STATUS_HI__ERROR_INFO_VALID_FLAG__SHIFT 0x2 +#define TCX_UE_ERR_STATUS_HI__ERROR_INFO__SHIFT 0x3 +#define TCX_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define TCX_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define TCX_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define TCX_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define TCX_UE_ERR_STATUS_HI__ERROR_INFO_VALID_FLAG_MASK 0x00000004L +#define TCX_UE_ERR_STATUS_HI__ERROR_INFO_MASK 0x007FFFF8L +#define TCX_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define TCX_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +//TCX_CE_ERR_STATUS_LO +#define TCX_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define TCX_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define TCX_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define TCX_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define TCX_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define TCX_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define TCX_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define TCX_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//TCX_CE_ERR_STATUS_HI +#define TCX_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define TCX_CE_ERR_STATUS_HI__ERROR_INFO_VALID_FLAG__SHIFT 0x2 +#define TCX_CE_ERR_STATUS_HI__ERROR_INFO__SHIFT 0x3 +#define TCX_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define TCX_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define TCX_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define TCX_CE_ERR_STATUS_HI__ERROR_INFO_VALID_FLAG_MASK 0x00000004L +#define TCX_CE_ERR_STATUS_HI__ERROR_INFO_MASK 0x007FFFF8L +#define TCX_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define TCX_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +//TCC_UE_ERR_STATUS_LO +#define TCC_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define TCC_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define TCC_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define TCC_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define TCC_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define TCC_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define TCC_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define TCC_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//TCC_UE_ERR_STATUS_HI +#define TCC_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define TCC_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define TCC_UE_ERR_STATUS_HI__ERROR_INFO_VALID_FLAG__SHIFT 0x2 +#define TCC_UE_ERR_STATUS_HI__ERROR_INFO__SHIFT 0x3 +#define TCC_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define TCC_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define TCC_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define TCC_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define TCC_UE_ERR_STATUS_HI__ERROR_INFO_VALID_FLAG_MASK 0x00000004L +#define TCC_UE_ERR_STATUS_HI__ERROR_INFO_MASK 0x007FFFF8L +#define TCC_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define TCC_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +//TCC_CE_ERR_STATUS_LO +#define TCC_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define TCC_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define TCC_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define TCC_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define TCC_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define TCC_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define TCC_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define TCC_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//TCC_CE_ERR_STATUS_HI +#define TCC_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define TCC_CE_ERR_STATUS_HI__ERROR_INFO_VALID_FLAG__SHIFT 0x2 +#define TCC_CE_ERR_STATUS_HI__ERROR_INFO__SHIFT 0x3 +#define TCC_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define TCC_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define TCC_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define TCC_CE_ERR_STATUS_HI__ERROR_INFO_VALID_FLAG_MASK 0x00000004L +#define TCC_CE_ERR_STATUS_HI__ERROR_INFO_MASK 0x007FFFF8L +#define TCC_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define TCC_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L // addressBlock: xcd0_gc_shdec @@ -14384,6 +15236,150 @@ #define CP_VMID_STATUS__PREEMPT_CE_STATUS__SHIFT 0x10 #define CP_VMID_STATUS__PREEMPT_DE_STATUS_MASK 0x0000FFFFL #define CP_VMID_STATUS__PREEMPT_CE_STATUS_MASK 0xFFFF0000L +//CPC_UE_ERR_STATUS_LO +#define CPC_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define CPC_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define CPC_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define CPC_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define CPC_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define CPC_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define CPC_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define CPC_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//CPC_UE_ERR_STATUS_HI +#define CPC_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define CPC_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define CPC_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define CPC_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define CPC_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define CPC_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define CPC_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define CPC_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define CPC_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define CPC_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define CPC_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define CPC_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define CPC_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define CPC_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L +//CPC_CE_ERR_STATUS_LO +#define CPC_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define CPC_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define CPC_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define CPC_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define CPC_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define CPC_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define CPC_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define CPC_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//CPC_CE_ERR_STATUS_HI +#define CPC_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define CPC_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define CPC_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define CPC_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define CPC_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define CPC_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define CPC_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define CPC_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define CPC_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define CPC_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define CPC_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define CPC_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define CPC_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define CPC_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L +//CPF_UE_ERR_STATUS_LO +#define CPF_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define CPF_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define CPF_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define CPF_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define CPF_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define CPF_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define CPF_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define CPF_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//CPF_UE_ERR_STATUS_HI +#define CPF_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define CPF_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define CPF_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define CPF_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define CPF_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define CPF_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define CPF_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define CPF_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define CPF_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define CPF_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define CPF_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define CPF_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define CPF_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define CPF_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L +//CPF_CE_ERR_STATUS_LO +#define CPF_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define CPF_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define CPF_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define CPF_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define CPF_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define CPF_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define CPF_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define CPF_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//CPF_CE_ERR_STATUS_HI +#define CPF_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define CPF_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define CPF_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define CPF_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define CPF_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define CPF_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define CPF_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define CPF_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define CPF_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define CPF_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define CPF_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define CPF_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define CPF_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define CPF_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L +//CPG_UE_ERR_STATUS_LO +#define CPG_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define CPG_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define CPG_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define CPG_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define CPG_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define CPG_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define CPG_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define CPG_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//CPG_UE_ERR_STATUS_HI +#define CPG_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define CPG_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define CPG_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define CPG_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define CPG_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define CPG_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define CPG_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define CPG_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define CPG_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define CPG_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define CPG_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define CPG_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define CPG_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define CPG_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L +//CPG_CE_ERR_STATUS_LO +#define CPG_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define CPG_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define CPG_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define CPG_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define CPG_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define CPG_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define CPG_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define CPG_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//CPG_CE_ERR_STATUS_HI +#define CPG_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define CPG_CE_ERR_STATUS_HI__OTHER__SHIFT 0x1 +#define CPG_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define CPG_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define CPG_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define CPG_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define CPG_CE_ERR_STATUS_HI__RESERVED__SHIFT 0x1b +#define CPG_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define CPG_CE_ERR_STATUS_HI__OTHER_MASK 0x00000002L +#define CPG_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define CPG_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define CPG_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define CPG_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define CPG_CE_ERR_STATUS_HI__RESERVED_MASK 0xF8000000L // addressBlock: xcd0_gc_cppdec2 @@ -22764,6 +23760,74 @@ #define SPI_WAVE_LIMIT_CNTL__GS_WAVE_GRAN_MASK 0x00000030L #define SPI_WAVE_LIMIT_CNTL__HS_WAVE_GRAN_MASK 0x000000C0L +// addressBlock: xcd0_gc_gccanedec +//GC_CANE_ERR_STATUS +#define GC_CANE_ERR_STATUS__SDPM_RDRSP_STATUS__SHIFT 0x0 +#define GC_CANE_ERR_STATUS__SDPM_WRRSP_STATUS__SHIFT 0x4 +#define GC_CANE_ERR_STATUS__SDPM_RDRSP_DATASTATUS__SHIFT 0x8 +#define GC_CANE_ERR_STATUS__SDPM_RDRSP_DATAPARITY_ERROR__SHIFT 0xa +#define GC_CANE_ERR_STATUS__SDPS_DAT_ERROR__SHIFT 0xb +#define GC_CANE_ERR_STATUS__SDPS_DAT_PARITY_ERROR__SHIFT 0xc +#define GC_CANE_ERR_STATUS__CLEAR_ERROR_STATUS__SHIFT 0xd +#define GC_CANE_ERR_STATUS__BUSY_ON_ERROR__SHIFT 0xe +#define GC_CANE_ERR_STATUS__BUSY_ON_UER_ERROR__SHIFT 0xf +#define GC_CANE_ERR_STATUS__FUE_FLAG__SHIFT 0x10 +#define GC_CANE_ERR_STATUS__INTERRUPT_ON_FATAL__SHIFT 0x11 +#define GC_CANE_ERR_STATUS__LEVEL_INTERRUPT__SHIFT 0x12 +#define GC_CANE_ERR_STATUS__SDPM_RDRSP_STATUS_MASK 0x0000000FL +#define GC_CANE_ERR_STATUS__SDPM_WRRSP_STATUS_MASK 0x000000F0L +#define GC_CANE_ERR_STATUS__SDPM_RDRSP_DATASTATUS_MASK 0x00000300L +#define GC_CANE_ERR_STATUS__SDPM_RDRSP_DATAPARITY_ERROR_MASK 0x00000400L +#define GC_CANE_ERR_STATUS__SDPS_DAT_ERROR_MASK 0x00000800L +#define GC_CANE_ERR_STATUS__SDPS_DAT_PARITY_ERROR_MASK 0x00001000L +#define GC_CANE_ERR_STATUS__CLEAR_ERROR_STATUS_MASK 0x00002000L +#define GC_CANE_ERR_STATUS__BUSY_ON_ERROR_MASK 0x00004000L +#define GC_CANE_ERR_STATUS__BUSY_ON_UER_ERROR_MASK 0x00008000L +#define GC_CANE_ERR_STATUS__FUE_FLAG_MASK 0x00010000L +#define GC_CANE_ERR_STATUS__INTERRUPT_ON_FATAL_MASK 0x00020000L +#define GC_CANE_ERR_STATUS__LEVEL_INTERRUPT_MASK 0x00040000L +//GC_CANE_UE_ERR_STATUS_LO +#define GC_CANE_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define GC_CANE_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define GC_CANE_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define GC_CANE_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define GC_CANE_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define GC_CANE_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define GC_CANE_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define GC_CANE_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//GC_CANE_UE_ERR_STATUS_HI +#define GC_CANE_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define GC_CANE_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define GC_CANE_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define GC_CANE_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define GC_CANE_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define GC_CANE_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define GC_CANE_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define GC_CANE_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define GC_CANE_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define GC_CANE_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define GC_CANE_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define GC_CANE_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +//GC_CANE_CE_ERR_STATUS_LO +#define GC_CANE_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define GC_CANE_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define GC_CANE_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define GC_CANE_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define GC_CANE_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define GC_CANE_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define GC_CANE_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define GC_CANE_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//GC_CANE_CE_ERR_STATUS_HI +#define GC_CANE_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define GC_CANE_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define GC_CANE_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define GC_CANE_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define GC_CANE_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define GC_CANE_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define GC_CANE_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define GC_CANE_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define GC_CANE_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define GC_CANE_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L // addressBlock: xcd0_gc_perfddec //CPG_PERFCOUNTER1_LO @@ -26471,6 +27535,30 @@ //RLC_CPG_STAT_INVAL #define RLC_CPG_STAT_INVAL__CPG_stat_inval__SHIFT 0x0 #define RLC_CPG_STAT_INVAL__CPG_stat_inval_MASK 0x00000001L +//RLC_UE_ERR_STATUS_LOW +#define RLC_UE_ERR_STATUS_LOW__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define RLC_UE_ERR_STATUS_LOW__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define RLC_UE_ERR_STATUS_LOW__ADDRESS__SHIFT 0x2 +#define RLC_UE_ERR_STATUS_LOW__MEMORY_ID__SHIFT 0x18 +#define RLC_UE_ERR_STATUS_LOW__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define RLC_UE_ERR_STATUS_LOW__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define RLC_UE_ERR_STATUS_LOW__ADDRESS_MASK 0x00FFFFFCL +#define RLC_UE_ERR_STATUS_LOW__MEMORY_ID_MASK 0xFF000000L +//RLC_UE_ERR_STATUS_HIGH +#define RLC_UE_ERR_STATUS_HIGH__ECC__SHIFT 0x0 +#define RLC_UE_ERR_STATUS_HIGH__PARITY__SHIFT 0x1 +#define RLC_UE_ERR_STATUS_HIGH__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define RLC_UE_ERR_STATUS_HIGH__ERR_INFO__SHIFT 0x3 +#define RLC_UE_ERR_STATUS_HIGH__UE_CNT__SHIFT 0x17 +#define RLC_UE_ERR_STATUS_HIGH__FED_CNT__SHIFT 0x1a +#define RLC_UE_ERR_STATUS_HIGH__RESERVED__SHIFT 0x1d +#define RLC_UE_ERR_STATUS_HIGH__ECC_MASK 0x00000001L +#define RLC_UE_ERR_STATUS_HIGH__PARITY_MASK 0x00000002L +#define RLC_UE_ERR_STATUS_HIGH__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define RLC_UE_ERR_STATUS_HIGH__ERR_INFO_MASK 0x007FFFF8L +#define RLC_UE_ERR_STATUS_HIGH__UE_CNT_MASK 0x03800000L +#define RLC_UE_ERR_STATUS_HIGH__FED_CNT_MASK 0x1C000000L +#define RLC_UE_ERR_STATUS_HIGH__RESERVED_MASK 0xE0000000L //RLC_DSM_CNTL #define RLC_DSM_CNTL__RLCG_INSTR_RAM_IRRITATOR_DATA_SEL__SHIFT 0x0 #define RLC_DSM_CNTL__RLCG_INSTR_RAM_IRRITATOR_SINGLE_WRITE__SHIFT 0x2 @@ -26573,6 +27661,30 @@ #define RLC_DSM_CNTL2A__RLC_SPM_SE2_SCRATCH_RAM_SELECT_INJECT_DELAY_MASK 0x00000100L #define RLC_DSM_CNTL2A__RLC_SPM_SE3_SCRATCH_RAM_ENABLE_ERROR_INJECT_MASK 0x00000600L #define RLC_DSM_CNTL2A__RLC_SPM_SE3_SCRATCH_RAM_SELECT_INJECT_DELAY_MASK 0x00000800L +//RLC_CE_ERR_STATUS_LOW +#define RLC_CE_ERR_STATUS_LOW__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define RLC_CE_ERR_STATUS_LOW__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define RLC_CE_ERR_STATUS_LOW__ADDRESS__SHIFT 0x2 +#define RLC_CE_ERR_STATUS_LOW__MEMORY_ID__SHIFT 0x18 +#define RLC_CE_ERR_STATUS_LOW__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define RLC_CE_ERR_STATUS_LOW__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define RLC_CE_ERR_STATUS_LOW__ADDRESS_MASK 0x00FFFFFCL +#define RLC_CE_ERR_STATUS_LOW__MEMORY_ID_MASK 0xFF000000L +//RLC_CE_ERR_STATUS_HIGH +#define RLC_CE_ERR_STATUS_HIGH__ECC__SHIFT 0x0 +#define RLC_CE_ERR_STATUS_HIGH__OTHER__SHIFT 0x1 +#define RLC_CE_ERR_STATUS_HIGH__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define RLC_CE_ERR_STATUS_HIGH__ERR_INFO__SHIFT 0x3 +#define RLC_CE_ERR_STATUS_HIGH__CE_CNT__SHIFT 0x17 +#define RLC_CE_ERR_STATUS_HIGH__POISON__SHIFT 0x1a +#define RLC_CE_ERR_STATUS_HIGH__RESERVED__SHIFT 0x1b +#define RLC_CE_ERR_STATUS_HIGH__ECC_MASK 0x00000001L +#define RLC_CE_ERR_STATUS_HIGH__OTHER_MASK 0x00000002L +#define RLC_CE_ERR_STATUS_HIGH__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define RLC_CE_ERR_STATUS_HIGH__ERR_INFO_MASK 0x007FFFF8L +#define RLC_CE_ERR_STATUS_HIGH__CE_CNT_MASK 0x03800000L +#define RLC_CE_ERR_STATUS_HIGH__POISON_MASK 0x04000000L +#define RLC_CE_ERR_STATUS_HIGH__RESERVED_MASK 0xF8000000L //RLC_RLCV_SPARE_INT #define RLC_RLCV_SPARE_INT__INTERRUPT__SHIFT 0x0 #define RLC_RLCV_SPARE_INT__RESERVED__SHIFT 0x1 diff --git a/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_1_8_0_offset.h b/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_1_8_0_offset.h index 8bcc81f2dfc0..879ee9de3ff3 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_1_8_0_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_1_8_0_offset.h @@ -1491,6 +1491,10 @@ #define regMMEA0_PERFCOUNTER1_CFG_BASE_IDX 0 #define regMMEA0_PERFCOUNTER_RSLT_CNTL 0x0400 #define regMMEA0_PERFCOUNTER_RSLT_CNTL_BASE_IDX 0 +#define regMMEA0_UE_ERR_STATUS_LO 0x0406 +#define regMMEA0_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regMMEA0_UE_ERR_STATUS_HI 0x0407 +#define regMMEA0_UE_ERR_STATUS_HI_BASE_IDX 0 #define regMMEA0_DSM_CNTL 0x0408 #define regMMEA0_DSM_CNTL_BASE_IDX 0 #define regMMEA0_DSM_CNTLA 0x0409 @@ -1511,8 +1515,12 @@ #define regMMEA0_ERR_STATUS_BASE_IDX 0 #define regMMEA0_MISC2 0x0412 #define regMMEA0_MISC2_BASE_IDX 0 +#define regMMEA0_CE_ERR_STATUS_LO 0x0414 +#define regMMEA0_CE_ERR_STATUS_LO_BASE_IDX 0 #define regMMEA0_MISC_AON 0x0415 #define regMMEA0_MISC_AON_BASE_IDX 0 +#define regMMEA0_CE_ERR_STATUS_HI 0x0416 +#define regMMEA0_CE_ERR_STATUS_HI_BASE_IDX 0 // addressBlock: aid_mmhub_ea_mmeadec1 @@ -1709,6 +1717,10 @@ #define regMMEA1_PERFCOUNTER1_CFG_BASE_IDX 0 #define regMMEA1_PERFCOUNTER_RSLT_CNTL 0x0540 #define regMMEA1_PERFCOUNTER_RSLT_CNTL_BASE_IDX 0 +#define regMMEA1_UE_ERR_STATUS_LO 0x0546 +#define regMMEA1_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regMMEA1_UE_ERR_STATUS_HI 0x0547 +#define regMMEA1_UE_ERR_STATUS_HI_BASE_IDX 0 #define regMMEA1_DSM_CNTL 0x0548 #define regMMEA1_DSM_CNTL_BASE_IDX 0 #define regMMEA1_DSM_CNTLA 0x0549 @@ -1729,8 +1741,12 @@ #define regMMEA1_ERR_STATUS_BASE_IDX 0 #define regMMEA1_MISC2 0x0552 #define regMMEA1_MISC2_BASE_IDX 0 +#define regMMEA1_CE_ERR_STATUS_LO 0x0554 +#define regMMEA1_CE_ERR_STATUS_LO_BASE_IDX 0 #define regMMEA1_MISC_AON 0x0555 #define regMMEA1_MISC_AON_BASE_IDX 0 +#define regMMEA1_CE_ERR_STATUS_HI 0x0556 +#define regMMEA1_CE_ERR_STATUS_HI_BASE_IDX 0 // addressBlock: aid_mmhub_ea_mmeadec2 @@ -1927,6 +1943,10 @@ #define regMMEA2_PERFCOUNTER1_CFG_BASE_IDX 0 #define regMMEA2_PERFCOUNTER_RSLT_CNTL 0x0680 #define regMMEA2_PERFCOUNTER_RSLT_CNTL_BASE_IDX 0 +#define regMMEA2_UE_ERR_STATUS_LO 0x0686 +#define regMMEA2_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regMMEA2_UE_ERR_STATUS_HI 0x0687 +#define regMMEA2_UE_ERR_STATUS_HI_BASE_IDX 0 #define regMMEA2_DSM_CNTL 0x0688 #define regMMEA2_DSM_CNTL_BASE_IDX 0 #define regMMEA2_DSM_CNTLA 0x0689 @@ -1947,8 +1967,12 @@ #define regMMEA2_ERR_STATUS_BASE_IDX 0 #define regMMEA2_MISC2 0x0692 #define regMMEA2_MISC2_BASE_IDX 0 +#define regMMEA2_CE_ERR_STATUS_LO 0x0694 +#define regMMEA2_CE_ERR_STATUS_LO_BASE_IDX 0 #define regMMEA2_MISC_AON 0x0695 #define regMMEA2_MISC_AON_BASE_IDX 0 +#define regMMEA2_CE_ERR_STATUS_HI 0x0696 +#define regMMEA2_CE_ERR_STATUS_HI_BASE_IDX 0 // addressBlock: aid_mmhub_ea_mmeadec3 @@ -2145,6 +2169,10 @@ #define regMMEA3_PERFCOUNTER1_CFG_BASE_IDX 0 #define regMMEA3_PERFCOUNTER_RSLT_CNTL 0x07c0 #define regMMEA3_PERFCOUNTER_RSLT_CNTL_BASE_IDX 0 +#define regMMEA3_UE_ERR_STATUS_LO 0x07c6 +#define regMMEA3_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regMMEA3_UE_ERR_STATUS_HI 0x07c7 +#define regMMEA3_UE_ERR_STATUS_HI_BASE_IDX 0 #define regMMEA3_DSM_CNTL 0x07c8 #define regMMEA3_DSM_CNTL_BASE_IDX 0 #define regMMEA3_DSM_CNTLA 0x07c9 @@ -2165,9 +2193,12 @@ #define regMMEA3_ERR_STATUS_BASE_IDX 0 #define regMMEA3_MISC2 0x07d2 #define regMMEA3_MISC2_BASE_IDX 0 +#define regMMEA3_CE_ERR_STATUS_LO 0x07d4 +#define regMMEA3_CE_ERR_STATUS_LO_BASE_IDX 0 #define regMMEA3_MISC_AON 0x07d5 #define regMMEA3_MISC_AON_BASE_IDX 0 - +#define regMMEA3_CE_ERR_STATUS_HI 0x07d6 +#define regMMEA3_CE_ERR_STATUS_HI_BASE_IDX 0 // addressBlock: aid_mmhub_ea_mmeadec4 // base address: 0x62000 @@ -2363,6 +2394,10 @@ #define regMMEA4_PERFCOUNTER1_CFG_BASE_IDX 0 #define regMMEA4_PERFCOUNTER_RSLT_CNTL 0x0900 #define regMMEA4_PERFCOUNTER_RSLT_CNTL_BASE_IDX 0 +#define regMMEA4_UE_ERR_STATUS_LO 0x0906 +#define regMMEA4_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regMMEA4_UE_ERR_STATUS_HI 0x0907 +#define regMMEA4_UE_ERR_STATUS_HI_BASE_IDX 0 #define regMMEA4_DSM_CNTL 0x0908 #define regMMEA4_DSM_CNTL_BASE_IDX 0 #define regMMEA4_DSM_CNTLA 0x0909 @@ -2383,9 +2418,12 @@ #define regMMEA4_ERR_STATUS_BASE_IDX 0 #define regMMEA4_MISC2 0x0912 #define regMMEA4_MISC2_BASE_IDX 0 +#define regMMEA4_CE_ERR_STATUS_LO 0x0914 +#define regMMEA4_CE_ERR_STATUS_LO_BASE_IDX 0 #define regMMEA4_MISC_AON 0x0915 #define regMMEA4_MISC_AON_BASE_IDX 0 - +#define regMMEA4_CE_ERR_STATUS_HI 0x0916 +#define regMMEA4_CE_ERR_STATUS_HI_BASE_IDX 0 // addressBlock: aid_mmhub_pctldec0 // base address: 0x62a00 @@ -3310,5 +3348,19 @@ #define regL2TLB_PERFCOUNTER_HI 0x0d2d #define regL2TLB_PERFCOUNTER_HI_BASE_IDX 0 +// addressBlock: aid_mmhub_mm_cane_mmcanedec +// base address: 0x635f0 +#define regMM_CANE_ICG_CTRL 0x0d8a +#define regMM_CANE_ICG_CTRL_BASE_IDX 0 +#define regMM_CANE_ERR_STATUS 0x0d8c +#define regMM_CANE_ERR_STATUS_BASE_IDX 0 +#define regMM_CANE_UE_ERR_STATUS_LO 0x0d8d +#define regMM_CANE_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regMM_CANE_UE_ERR_STATUS_HI 0x0d8e +#define regMM_CANE_UE_ERR_STATUS_HI_BASE_IDX 0 +#define regMM_CANE_CE_ERR_STATUS_LO 0x0d8f +#define regMM_CANE_CE_ERR_STATUS_LO_BASE_IDX 0 +#define regMM_CANE_CE_ERR_STATUS_HI 0x0d90 +#define regMM_CANE_CE_ERR_STATUS_HI_BASE_IDX 0 #endif diff --git a/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_1_8_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_1_8_0_sh_mask.h index af41468ce69f..088c1f02aa43 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_1_8_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_1_8_0_sh_mask.h @@ -10470,6 +10470,30 @@ #define MMEA0_PERFCOUNTER_RSLT_CNTL__ENABLE_ANY_MASK 0x01000000L #define MMEA0_PERFCOUNTER_RSLT_CNTL__CLEAR_ALL_MASK 0x02000000L #define MMEA0_PERFCOUNTER_RSLT_CNTL__STOP_ALL_ON_SATURATE_MASK 0x04000000L +//MMEA0_UE_ERR_STATUS_LO +#define MMEA0_UE_ERR_STATUS_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define MMEA0_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define MMEA0_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define MMEA0_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define MMEA0_UE_ERR_STATUS_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define MMEA0_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define MMEA0_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define MMEA0_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//MMEA0_UE_ERR_STATUS_HI +#define MMEA0_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define MMEA0_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define MMEA0_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define MMEA0_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define MMEA0_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define MMEA0_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define MMEA0_UE_ERR_STATUS_HI__RESERVED_FIELD__SHIFT 0x1d +#define MMEA0_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define MMEA0_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define MMEA0_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define MMEA0_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define MMEA0_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define MMEA0_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define MMEA0_UE_ERR_STATUS_HI__RESERVED_FIELD_MASK 0xE0000000L //MMEA0_DSM_CNTL #define MMEA0_DSM_CNTL__DRAMRD_CMDMEM_DSM_IRRITATOR_DATA__SHIFT 0x0 #define MMEA0_DSM_CNTL__DRAMRD_CMDMEM_ENABLE_SINGLE_WRITE__SHIFT 0x2 @@ -10718,12 +10742,35 @@ #define MMEA0_MISC2__DRAM_WR_THROTTLE_MASK 0x00020000L #define MMEA0_MISC2__GMI_RD_THROTTLE_MASK 0x00040000L #define MMEA0_MISC2__GMI_WR_THROTTLE_MASK 0x00080000L +//MMEA0_CE_ERR_STATUS_LO +#define MMEA0_CE_ERR_STATUS_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define MMEA0_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define MMEA0_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define MMEA0_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define MMEA0_CE_ERR_STATUS_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define MMEA0_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define MMEA0_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define MMEA0_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L //MMEA0_MISC_AON #define MMEA0_MISC_AON__LINKMGR_PARTACK_HYSTERESIS__SHIFT 0x0 #define MMEA0_MISC_AON__LINKMGR_PARTACK_DEASSERT_MODE__SHIFT 0x2 #define MMEA0_MISC_AON__LINKMGR_PARTACK_HYSTERESIS_MASK 0x00000003L #define MMEA0_MISC_AON__LINKMGR_PARTACK_DEASSERT_MODE_MASK 0x00000004L - +//MMEA0_CE_ERR_STATUS_HI +#define MMEA0_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define MMEA0_CE_ERR_STATUS_HI__RESERVED_FIELD0__SHIFT 0x1 +#define MMEA0_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define MMEA0_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define MMEA0_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define MMEA0_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define MMEA0_CE_ERR_STATUS_HI__RESERVED_FIELD1__SHIFT 0x1b +#define MMEA0_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define MMEA0_CE_ERR_STATUS_HI__RESERVED_FIELD0_MASK 0x00000002L +#define MMEA0_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define MMEA0_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define MMEA0_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define MMEA0_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define MMEA0_CE_ERR_STATUS_HI__RESERVED_FIELD1_MASK 0xF8000000L // addressBlock: aid_mmhub_ea_mmeadec1 //MMEA1_DRAM_RD_CLI2GRP_MAP0 @@ -12418,6 +12465,30 @@ #define MMEA1_PERFCOUNTER_RSLT_CNTL__ENABLE_ANY_MASK 0x01000000L #define MMEA1_PERFCOUNTER_RSLT_CNTL__CLEAR_ALL_MASK 0x02000000L #define MMEA1_PERFCOUNTER_RSLT_CNTL__STOP_ALL_ON_SATURATE_MASK 0x04000000L +//MMEA1_UE_ERR_STATUS_LO +#define MMEA1_UE_ERR_STATUS_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define MMEA1_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define MMEA1_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define MMEA1_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define MMEA1_UE_ERR_STATUS_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define MMEA1_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define MMEA1_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define MMEA1_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//MMEA1_UE_ERR_STATUS_HI +#define MMEA1_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define MMEA1_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define MMEA1_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define MMEA1_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define MMEA1_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define MMEA1_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define MMEA1_UE_ERR_STATUS_HI__RESERVED_FIELD__SHIFT 0x1d +#define MMEA1_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define MMEA1_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define MMEA1_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define MMEA1_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define MMEA1_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define MMEA1_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define MMEA1_UE_ERR_STATUS_HI__RESERVED_FIELD_MASK 0xE0000000L //MMEA1_DSM_CNTL #define MMEA1_DSM_CNTL__DRAMRD_CMDMEM_DSM_IRRITATOR_DATA__SHIFT 0x0 #define MMEA1_DSM_CNTL__DRAMRD_CMDMEM_ENABLE_SINGLE_WRITE__SHIFT 0x2 @@ -12666,12 +12737,35 @@ #define MMEA1_MISC2__DRAM_WR_THROTTLE_MASK 0x00020000L #define MMEA1_MISC2__GMI_RD_THROTTLE_MASK 0x00040000L #define MMEA1_MISC2__GMI_WR_THROTTLE_MASK 0x00080000L +//MMEA1_CE_ERR_STATUS_LO +#define MMEA1_CE_ERR_STATUS_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define MMEA1_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define MMEA1_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define MMEA1_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define MMEA1_CE_ERR_STATUS_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define MMEA1_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define MMEA1_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define MMEA1_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L //MMEA1_MISC_AON #define MMEA1_MISC_AON__LINKMGR_PARTACK_HYSTERESIS__SHIFT 0x0 #define MMEA1_MISC_AON__LINKMGR_PARTACK_DEASSERT_MODE__SHIFT 0x2 #define MMEA1_MISC_AON__LINKMGR_PARTACK_HYSTERESIS_MASK 0x00000003L #define MMEA1_MISC_AON__LINKMGR_PARTACK_DEASSERT_MODE_MASK 0x00000004L - +//MMEA1_CE_ERR_STATUS_HI +#define MMEA1_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define MMEA1_CE_ERR_STATUS_HI__RESERVED_FIELD0__SHIFT 0x1 +#define MMEA1_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define MMEA1_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define MMEA1_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define MMEA1_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define MMEA1_CE_ERR_STATUS_HI__RESERVED_FIELD1__SHIFT 0x1b +#define MMEA1_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define MMEA1_CE_ERR_STATUS_HI__RESERVED_FIELD0_MASK 0x00000002L +#define MMEA1_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define MMEA1_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define MMEA1_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define MMEA1_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define MMEA1_CE_ERR_STATUS_HI__RESERVED_FIELD1_MASK 0xF8000000L // addressBlock: aid_mmhub_ea_mmeadec2 //MMEA2_DRAM_RD_CLI2GRP_MAP0 @@ -14366,6 +14460,30 @@ #define MMEA2_PERFCOUNTER_RSLT_CNTL__ENABLE_ANY_MASK 0x01000000L #define MMEA2_PERFCOUNTER_RSLT_CNTL__CLEAR_ALL_MASK 0x02000000L #define MMEA2_PERFCOUNTER_RSLT_CNTL__STOP_ALL_ON_SATURATE_MASK 0x04000000L +//MMEA2_UE_ERR_STATUS_LO +#define MMEA2_UE_ERR_STATUS_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define MMEA2_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define MMEA2_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define MMEA2_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define MMEA2_UE_ERR_STATUS_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define MMEA2_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define MMEA2_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define MMEA2_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//MMEA2_UE_ERR_STATUS_HI +#define MMEA2_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define MMEA2_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define MMEA2_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define MMEA2_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define MMEA2_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define MMEA2_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define MMEA2_UE_ERR_STATUS_HI__RESERVED_FIELD__SHIFT 0x1d +#define MMEA2_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define MMEA2_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define MMEA2_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define MMEA2_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define MMEA2_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define MMEA2_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define MMEA2_UE_ERR_STATUS_HI__RESERVED_FIELD_MASK 0xE0000000L //MMEA2_DSM_CNTL #define MMEA2_DSM_CNTL__DRAMRD_CMDMEM_DSM_IRRITATOR_DATA__SHIFT 0x0 #define MMEA2_DSM_CNTL__DRAMRD_CMDMEM_ENABLE_SINGLE_WRITE__SHIFT 0x2 @@ -14614,12 +14732,35 @@ #define MMEA2_MISC2__DRAM_WR_THROTTLE_MASK 0x00020000L #define MMEA2_MISC2__GMI_RD_THROTTLE_MASK 0x00040000L #define MMEA2_MISC2__GMI_WR_THROTTLE_MASK 0x00080000L +//MMEA2_CE_ERR_STATUS_LO +#define MMEA2_CE_ERR_STATUS_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define MMEA2_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define MMEA2_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define MMEA2_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define MMEA2_CE_ERR_STATUS_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define MMEA2_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define MMEA2_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define MMEA2_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L //MMEA2_MISC_AON #define MMEA2_MISC_AON__LINKMGR_PARTACK_HYSTERESIS__SHIFT 0x0 #define MMEA2_MISC_AON__LINKMGR_PARTACK_DEASSERT_MODE__SHIFT 0x2 #define MMEA2_MISC_AON__LINKMGR_PARTACK_HYSTERESIS_MASK 0x00000003L #define MMEA2_MISC_AON__LINKMGR_PARTACK_DEASSERT_MODE_MASK 0x00000004L - +//MMEA2_CE_ERR_STATUS_HI +#define MMEA2_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define MMEA2_CE_ERR_STATUS_HI__RESERVED_FIELD0__SHIFT 0x1 +#define MMEA2_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define MMEA2_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define MMEA2_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define MMEA2_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define MMEA2_CE_ERR_STATUS_HI__RESERVED_FIELD1__SHIFT 0x1b +#define MMEA2_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define MMEA2_CE_ERR_STATUS_HI__RESERVED_FIELD0_MASK 0x00000002L +#define MMEA2_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define MMEA2_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define MMEA2_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define MMEA2_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define MMEA2_CE_ERR_STATUS_HI__RESERVED_FIELD1_MASK 0xF8000000L // addressBlock: aid_mmhub_ea_mmeadec3 //MMEA3_DRAM_RD_CLI2GRP_MAP0 @@ -16314,6 +16455,30 @@ #define MMEA3_PERFCOUNTER_RSLT_CNTL__ENABLE_ANY_MASK 0x01000000L #define MMEA3_PERFCOUNTER_RSLT_CNTL__CLEAR_ALL_MASK 0x02000000L #define MMEA3_PERFCOUNTER_RSLT_CNTL__STOP_ALL_ON_SATURATE_MASK 0x04000000L +//MMEA3_UE_ERR_STATUS_LO +#define MMEA3_UE_ERR_STATUS_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define MMEA3_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define MMEA3_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define MMEA3_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define MMEA3_UE_ERR_STATUS_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define MMEA3_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define MMEA3_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define MMEA3_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//MMEA3_UE_ERR_STATUS_HI +#define MMEA3_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define MMEA3_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define MMEA3_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define MMEA3_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define MMEA3_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define MMEA3_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define MMEA3_UE_ERR_STATUS_HI__RESERVED_FIELD__SHIFT 0x1d +#define MMEA3_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define MMEA3_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define MMEA3_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define MMEA3_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define MMEA3_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define MMEA3_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define MMEA3_UE_ERR_STATUS_HI__RESERVED_FIELD_MASK 0xE0000000L //MMEA3_DSM_CNTL #define MMEA3_DSM_CNTL__DRAMRD_CMDMEM_DSM_IRRITATOR_DATA__SHIFT 0x0 #define MMEA3_DSM_CNTL__DRAMRD_CMDMEM_ENABLE_SINGLE_WRITE__SHIFT 0x2 @@ -16562,12 +16727,35 @@ #define MMEA3_MISC2__DRAM_WR_THROTTLE_MASK 0x00020000L #define MMEA3_MISC2__GMI_RD_THROTTLE_MASK 0x00040000L #define MMEA3_MISC2__GMI_WR_THROTTLE_MASK 0x00080000L +//MMEA3_CE_ERR_STATUS_LO +#define MMEA3_CE_ERR_STATUS_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define MMEA3_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define MMEA3_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define MMEA3_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define MMEA3_CE_ERR_STATUS_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define MMEA3_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define MMEA3_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define MMEA3_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L //MMEA3_MISC_AON #define MMEA3_MISC_AON__LINKMGR_PARTACK_HYSTERESIS__SHIFT 0x0 #define MMEA3_MISC_AON__LINKMGR_PARTACK_DEASSERT_MODE__SHIFT 0x2 #define MMEA3_MISC_AON__LINKMGR_PARTACK_HYSTERESIS_MASK 0x00000003L #define MMEA3_MISC_AON__LINKMGR_PARTACK_DEASSERT_MODE_MASK 0x00000004L - +//MMEA3_CE_ERR_STATUS_HI +#define MMEA3_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define MMEA3_CE_ERR_STATUS_HI__RESERVED_FIELD0__SHIFT 0x1 +#define MMEA3_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define MMEA3_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define MMEA3_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define MMEA3_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define MMEA3_CE_ERR_STATUS_HI__RESERVED_FIELD1__SHIFT 0x1b +#define MMEA3_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define MMEA3_CE_ERR_STATUS_HI__RESERVED_FIELD0_MASK 0x00000002L +#define MMEA3_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define MMEA3_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define MMEA3_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define MMEA3_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define MMEA3_CE_ERR_STATUS_HI__RESERVED_FIELD1_MASK 0xF8000000L // addressBlock: aid_mmhub_ea_mmeadec4 //MMEA4_DRAM_RD_CLI2GRP_MAP0 @@ -18262,6 +18450,30 @@ #define MMEA4_PERFCOUNTER_RSLT_CNTL__ENABLE_ANY_MASK 0x01000000L #define MMEA4_PERFCOUNTER_RSLT_CNTL__CLEAR_ALL_MASK 0x02000000L #define MMEA4_PERFCOUNTER_RSLT_CNTL__STOP_ALL_ON_SATURATE_MASK 0x04000000L +//MMEA4_UE_ERR_STATUS_LO +#define MMEA4_UE_ERR_STATUS_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define MMEA4_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define MMEA4_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define MMEA4_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define MMEA4_UE_ERR_STATUS_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define MMEA4_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define MMEA4_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define MMEA4_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//MMEA4_UE_ERR_STATUS_HI +#define MMEA4_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define MMEA4_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define MMEA4_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define MMEA4_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define MMEA4_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define MMEA4_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define MMEA4_UE_ERR_STATUS_HI__RESERVED_FIELD__SHIFT 0x1d +#define MMEA4_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define MMEA4_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define MMEA4_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define MMEA4_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define MMEA4_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define MMEA4_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define MMEA4_UE_ERR_STATUS_HI__RESERVED_FIELD_MASK 0xE0000000L //MMEA4_DSM_CNTL #define MMEA4_DSM_CNTL__DRAMRD_CMDMEM_DSM_IRRITATOR_DATA__SHIFT 0x0 #define MMEA4_DSM_CNTL__DRAMRD_CMDMEM_ENABLE_SINGLE_WRITE__SHIFT 0x2 @@ -18510,12 +18722,35 @@ #define MMEA4_MISC2__DRAM_WR_THROTTLE_MASK 0x00020000L #define MMEA4_MISC2__GMI_RD_THROTTLE_MASK 0x00040000L #define MMEA4_MISC2__GMI_WR_THROTTLE_MASK 0x00080000L +//MMEA4_CE_ERR_STATUS_LO +#define MMEA4_CE_ERR_STATUS_LO__STATUS_VALID_FLAG__SHIFT 0x0 +#define MMEA4_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define MMEA4_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define MMEA4_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define MMEA4_CE_ERR_STATUS_LO__STATUS_VALID_FLAG_MASK 0x00000001L +#define MMEA4_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define MMEA4_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define MMEA4_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L //MMEA4_MISC_AON #define MMEA4_MISC_AON__LINKMGR_PARTACK_HYSTERESIS__SHIFT 0x0 #define MMEA4_MISC_AON__LINKMGR_PARTACK_DEASSERT_MODE__SHIFT 0x2 #define MMEA4_MISC_AON__LINKMGR_PARTACK_HYSTERESIS_MASK 0x00000003L #define MMEA4_MISC_AON__LINKMGR_PARTACK_DEASSERT_MODE_MASK 0x00000004L - +//MMEA4_CE_ERR_STATUS_HI +#define MMEA4_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define MMEA4_CE_ERR_STATUS_HI__RESERVED_FIELD0__SHIFT 0x1 +#define MMEA4_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define MMEA4_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define MMEA4_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define MMEA4_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define MMEA4_CE_ERR_STATUS_HI__RESERVED_FIELD1__SHIFT 0x1b +#define MMEA4_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define MMEA4_CE_ERR_STATUS_HI__RESERVED_FIELD0_MASK 0x00000002L +#define MMEA4_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define MMEA4_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define MMEA4_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define MMEA4_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L +#define MMEA4_CE_ERR_STATUS_HI__RESERVED_FIELD1_MASK 0xF8000000L // addressBlock: aid_mmhub_pctldec0 //PCTL0_CTRL @@ -22311,5 +22546,83 @@ #define L2TLB_PERFCOUNTER_HI__COUNTER_HI_MASK 0x0000FFFFL #define L2TLB_PERFCOUNTER_HI__COMPARE_VALUE_MASK 0xFFFF0000L - +// addressBlock: aid_mmhub_mm_cane_mmcanedec +//MM_CANE_ICG_CTRL +#define MM_CANE_ICG_CTRL__SOFT_OVERRIDE_IREQ0__SHIFT 0x0 +#define MM_CANE_ICG_CTRL__SOFT_OVERRIDE_ATRET__SHIFT 0x1 +#define MM_CANE_ICG_CTRL__SOFT_OVERRIDE_OREQ__SHIFT 0x2 +#define MM_CANE_ICG_CTRL__SOFT_OVERRIDE_REGISTER__SHIFT 0x3 +#define MM_CANE_ICG_CTRL__SOFT_OVERRIDE_SDPM_RETURN__SHIFT 0x4 +#define MM_CANE_ICG_CTRL__SOFT_OVERRIDE_IREQ0_MASK 0x00000001L +#define MM_CANE_ICG_CTRL__SOFT_OVERRIDE_ATRET_MASK 0x00000002L +#define MM_CANE_ICG_CTRL__SOFT_OVERRIDE_OREQ_MASK 0x00000004L +#define MM_CANE_ICG_CTRL__SOFT_OVERRIDE_REGISTER_MASK 0x00000008L +#define MM_CANE_ICG_CTRL__SOFT_OVERRIDE_SDPM_RETURN_MASK 0x00000010L +//MM_CANE_ERR_STATUS +#define MM_CANE_ERR_STATUS__SDPM_RDRSP_STATUS__SHIFT 0x0 +#define MM_CANE_ERR_STATUS__SDPM_WRRSP_STATUS__SHIFT 0x4 +#define MM_CANE_ERR_STATUS__SDPM_RDRSP_DATASTATUS__SHIFT 0x8 +#define MM_CANE_ERR_STATUS__SDPM_RDRSP_DATAPARITY_ERROR__SHIFT 0xa +#define MM_CANE_ERR_STATUS__SDPS_DAT_ERROR__SHIFT 0xb +#define MM_CANE_ERR_STATUS__SDPS_DAT_PARITY_ERROR__SHIFT 0xc +#define MM_CANE_ERR_STATUS__CLEAR_ERROR_STATUS__SHIFT 0xd +#define MM_CANE_ERR_STATUS__BUSY_ON_ERROR__SHIFT 0xe +#define MM_CANE_ERR_STATUS__BUSY_ON_UER_ERROR__SHIFT 0xf +#define MM_CANE_ERR_STATUS__FUE_FLAG__SHIFT 0x10 +#define MM_CANE_ERR_STATUS__INTERRUPT_ON_FATAL__SHIFT 0x11 +#define MM_CANE_ERR_STATUS__LEVEL_INTERRUPT__SHIFT 0x12 +#define MM_CANE_ERR_STATUS__SDPM_RDRSP_STATUS_MASK 0x0000000FL +#define MM_CANE_ERR_STATUS__SDPM_WRRSP_STATUS_MASK 0x000000F0L +#define MM_CANE_ERR_STATUS__SDPM_RDRSP_DATASTATUS_MASK 0x00000300L +#define MM_CANE_ERR_STATUS__SDPM_RDRSP_DATAPARITY_ERROR_MASK 0x00000400L +#define MM_CANE_ERR_STATUS__SDPS_DAT_ERROR_MASK 0x00000800L +#define MM_CANE_ERR_STATUS__SDPS_DAT_PARITY_ERROR_MASK 0x00001000L +#define MM_CANE_ERR_STATUS__CLEAR_ERROR_STATUS_MASK 0x00002000L +#define MM_CANE_ERR_STATUS__BUSY_ON_ERROR_MASK 0x00004000L +#define MM_CANE_ERR_STATUS__BUSY_ON_UER_ERROR_MASK 0x00008000L +#define MM_CANE_ERR_STATUS__FUE_FLAG_MASK 0x00010000L +#define MM_CANE_ERR_STATUS__INTERRUPT_ON_FATAL_MASK 0x00020000L +#define MM_CANE_ERR_STATUS__LEVEL_INTERRUPT_MASK 0x00040000L +//MM_CANE_UE_ERR_STATUS_LO +#define MM_CANE_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define MM_CANE_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define MM_CANE_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define MM_CANE_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define MM_CANE_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define MM_CANE_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define MM_CANE_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define MM_CANE_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//MM_CANE_UE_ERR_STATUS_HI +#define MM_CANE_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define MM_CANE_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define MM_CANE_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define MM_CANE_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define MM_CANE_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define MM_CANE_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define MM_CANE_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define MM_CANE_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define MM_CANE_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define MM_CANE_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define MM_CANE_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define MM_CANE_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +//MM_CANE_CE_ERR_STATUS_LO +#define MM_CANE_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define MM_CANE_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define MM_CANE_CE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define MM_CANE_CE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define MM_CANE_CE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define MM_CANE_CE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define MM_CANE_CE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define MM_CANE_CE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//MM_CANE_CE_ERR_STATUS_HI +#define MM_CANE_CE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define MM_CANE_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define MM_CANE_CE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define MM_CANE_CE_ERR_STATUS_HI__CE_CNT__SHIFT 0x17 +#define MM_CANE_CE_ERR_STATUS_HI__POISON__SHIFT 0x1a +#define MM_CANE_CE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define MM_CANE_CE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define MM_CANE_CE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define MM_CANE_CE_ERR_STATUS_HI__CE_CNT_MASK 0x03800000L +#define MM_CANE_CE_ERR_STATUS_HI__POISON_MASK 0x04000000L #endif diff --git a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_9_0_offset.h b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_9_0_offset.h index 033f2796c1e3..c8a15c8f4822 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_9_0_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_9_0_offset.h @@ -6201,6 +6201,8 @@ #define regNBIF_SHUB_TODET_SYNCFLOOD_CTRL2_BASE_IDX 8 #define regBIFC_BME_ERR_LOG_HB 0xe8ab #define regBIFC_BME_ERR_LOG_HB_BASE_IDX 8 +#define regBIFC_GFX_INT_MONITOR_MASK 0xe8ad +#define regBIFC_GFX_INT_MONITOR_MASK_BASE_IDX 8 #define regBIFC_HRP_SDP_WRRSP_POOLCRED_ALLOC 0xe8c0 #define regBIFC_HRP_SDP_WRRSP_POOLCRED_ALLOC_BASE_IDX 8 #define regBIFC_HRP_SDP_RDRSP_POOLCRED_ALLOC 0xe8c1 diff --git a/drivers/gpu/drm/amd/include/asic_reg/sdma/sdma_4_4_2_offset.h b/drivers/gpu/drm/amd/include/asic_reg/sdma/sdma_4_4_2_offset.h index 31bef0776ded..ead81aeffd67 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/sdma/sdma_4_4_2_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/sdma/sdma_4_4_2_offset.h @@ -211,6 +211,10 @@ #define regSDMA_RAS_STATUS_BASE_IDX 0 #define regSDMA_CLK_STATUS 0x0068 #define regSDMA_CLK_STATUS_BASE_IDX 0 +#define regSDMA_UE_ERR_STATUS_LO 0x0069 +#define regSDMA_UE_ERR_STATUS_LO_BASE_IDX 0 +#define regSDMA_UE_ERR_STATUS_HI 0x006a +#define regSDMA_UE_ERR_STATUS_HI_BASE_IDX 0 #define regSDMA_POWER_CNTL 0x006b #define regSDMA_POWER_CNTL_BASE_IDX 0 #define regSDMA_CLK_CTRL 0x006c diff --git a/drivers/gpu/drm/amd/include/asic_reg/sdma/sdma_4_4_2_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/sdma/sdma_4_4_2_sh_mask.h index e46cb3339355..290953bdf1d6 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/sdma/sdma_4_4_2_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/sdma/sdma_4_4_2_sh_mask.h @@ -1171,6 +1171,30 @@ #define SDMA_CLK_STATUS__F32_CLK_MASK 0x00000008L #define SDMA_CLK_STATUS__CE_CLK_MASK 0x00000010L #define SDMA_CLK_STATUS__PERF_CLK_MASK 0x00000020L +//SDMA_UE_ERR_STATUS_LO +#define SDMA_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define SDMA_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG__SHIFT 0x1 +#define SDMA_UE_ERR_STATUS_LO__ADDRESS__SHIFT 0x2 +#define SDMA_UE_ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define SDMA_UE_ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define SDMA_UE_ERR_STATUS_LO__ADDRESS_VALID_FLAG_MASK 0x00000002L +#define SDMA_UE_ERR_STATUS_LO__ADDRESS_MASK 0x00FFFFFCL +#define SDMA_UE_ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +//SDMA_UE_ERR_STATUS_HI +#define SDMA_UE_ERR_STATUS_HI__ECC__SHIFT 0x0 +#define SDMA_UE_ERR_STATUS_HI__PARITY__SHIFT 0x1 +#define SDMA_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define SDMA_UE_ERR_STATUS_HI__ERR_INFO__SHIFT 0x3 +#define SDMA_UE_ERR_STATUS_HI__UE_CNT__SHIFT 0x17 +#define SDMA_UE_ERR_STATUS_HI__FED_CNT__SHIFT 0x1a +#define SDMA_UE_ERR_STATUS_HI__RESERVED__SHIFT 0x1d +#define SDMA_UE_ERR_STATUS_HI__ECC_MASK 0x00000001L +#define SDMA_UE_ERR_STATUS_HI__PARITY_MASK 0x00000002L +#define SDMA_UE_ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define SDMA_UE_ERR_STATUS_HI__ERR_INFO_MASK 0x007FFFF8L +#define SDMA_UE_ERR_STATUS_HI__UE_CNT_MASK 0x03800000L +#define SDMA_UE_ERR_STATUS_HI__FED_CNT_MASK 0x1C000000L +#define SDMA_UE_ERR_STATUS_HI__RESERVED_MASK 0xE0000000L //SDMA_POWER_CNTL #define SDMA_POWER_CNTL__PG_CNTL_ENABLE__SHIFT 0x0 #define SDMA_POWER_CNTL__EXT_PG_POWER_ON_REQ__SHIFT 0x1 diff --git a/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_13_0_3_offset.h b/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_13_0_3_offset.h new file mode 100644 index 000000000000..b62b489402c5 --- /dev/null +++ b/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_13_0_3_offset.h @@ -0,0 +1,177 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#ifndef _smuio_13_0_3_OFFSET_HEADER +#define _smuio_13_0_3_OFFSET_HEADER + + + +// addressBlock: aid_smuio_smuio_reset_SmuSmuioDec +// base address: 0x5a300 +#define regSMUIO_MP_RESET_INTR 0x00c1 +#define regSMUIO_MP_RESET_INTR_BASE_IDX 1 +#define regSMUIO_SOC_HALT 0x00c2 +#define regSMUIO_SOC_HALT_BASE_IDX 1 + + +// addressBlock: aid_smuio_smuio_tsc_SmuSmuioDec +// base address: 0x5a8a0 +#define regPWROK_REFCLK_GAP_CYCLES 0x0028 +#define regPWROK_REFCLK_GAP_CYCLES_BASE_IDX 2 +#define regGOLDEN_TSC_INCREMENT_UPPER 0x002b +#define regGOLDEN_TSC_INCREMENT_UPPER_BASE_IDX 2 +#define regGOLDEN_TSC_INCREMENT_LOWER 0x002c +#define regGOLDEN_TSC_INCREMENT_LOWER_BASE_IDX 2 +#define regGOLDEN_TSC_COUNT_UPPER 0x002d +#define regGOLDEN_TSC_COUNT_UPPER_BASE_IDX 2 +#define regGOLDEN_TSC_COUNT_LOWER 0x002e +#define regGOLDEN_TSC_COUNT_LOWER_BASE_IDX 2 +#define regSOC_GOLDEN_TSC_SHADOW_UPPER 0x002f +#define regSOC_GOLDEN_TSC_SHADOW_UPPER_BASE_IDX 2 +#define regSOC_GOLDEN_TSC_SHADOW_LOWER 0x0030 +#define regSOC_GOLDEN_TSC_SHADOW_LOWER_BASE_IDX 2 +#define regSOC_GAP_PWROK 0x0031 +#define regSOC_GAP_PWROK_BASE_IDX 2 + + +// addressBlock: aid_smuio_smuio_swtimer_SmuSmuioDec +// base address: 0x5ac70 +#define regPWR_VIRT_RESET_REQ 0x011c +#define regPWR_VIRT_RESET_REQ_BASE_IDX 2 +#define regPWR_DISP_TIMER_CONTROL 0x011d +#define regPWR_DISP_TIMER_CONTROL_BASE_IDX 2 +#define regPWR_DISP_TIMER_DEBUG 0x011e +#define regPWR_DISP_TIMER_DEBUG_BASE_IDX 2 +#define regPWR_DISP_TIMER2_CONTROL 0x011f +#define regPWR_DISP_TIMER2_CONTROL_BASE_IDX 2 +#define regPWR_DISP_TIMER2_DEBUG 0x0120 +#define regPWR_DISP_TIMER2_DEBUG_BASE_IDX 2 +#define regPWR_DISP_TIMER_GLOBAL_CONTROL 0x0121 +#define regPWR_DISP_TIMER_GLOBAL_CONTROL_BASE_IDX 2 +#define regPWR_IH_CONTROL 0x0122 +#define regPWR_IH_CONTROL_BASE_IDX 2 + + +// addressBlock: aid_smuio_smuio_misc_SmuSmuioDec +// base address: 0x5a000 +#define regSMUIO_MCM_CONFIG 0x0023 +#define regSMUIO_MCM_CONFIG_BASE_IDX 1 +#define regIP_DISCOVERY_VERSION 0x0000 +#define regIP_DISCOVERY_VERSION_BASE_IDX 2 +#define regSCRATCH_REGISTER0 0x01bd +#define regSCRATCH_REGISTER0_BASE_IDX 2 +#define regSCRATCH_REGISTER1 0x01be +#define regSCRATCH_REGISTER1_BASE_IDX 2 +#define regSCRATCH_REGISTER2 0x01bf +#define regSCRATCH_REGISTER2_BASE_IDX 2 +#define regSCRATCH_REGISTER3 0x01c0 +#define regSCRATCH_REGISTER3_BASE_IDX 2 +#define regSCRATCH_REGISTER4 0x01c1 +#define regSCRATCH_REGISTER4_BASE_IDX 2 +#define regSCRATCH_REGISTER5 0x01c2 +#define regSCRATCH_REGISTER5_BASE_IDX 2 +#define regSCRATCH_REGISTER6 0x01c3 +#define regSCRATCH_REGISTER6_BASE_IDX 2 +#define regSCRATCH_REGISTER7 0x01c4 +#define regSCRATCH_REGISTER7_BASE_IDX 2 + + +// addressBlock: aid_smuio_smuio_gpio_SmuSmuioDec +// base address: 0x5a500 +#define regSMU_GPIOPAD_SW_INT_STAT 0x0140 +#define regSMU_GPIOPAD_SW_INT_STAT_BASE_IDX 1 +#define regSMU_GPIOPAD_MASK 0x0141 +#define regSMU_GPIOPAD_MASK_BASE_IDX 1 +#define regSMU_GPIOPAD_A 0x0142 +#define regSMU_GPIOPAD_A_BASE_IDX 1 +#define regSMU_GPIOPAD_TXIMPSEL 0x0143 +#define regSMU_GPIOPAD_TXIMPSEL_BASE_IDX 1 +#define regSMU_GPIOPAD_EN 0x0144 +#define regSMU_GPIOPAD_EN_BASE_IDX 1 +#define regSMU_GPIOPAD_Y 0x0145 +#define regSMU_GPIOPAD_Y_BASE_IDX 1 +#define regSMU_GPIOPAD_RXEN 0x0146 +#define regSMU_GPIOPAD_RXEN_BASE_IDX 1 +#define regSMU_GPIOPAD_RCVR_SEL0 0x0147 +#define regSMU_GPIOPAD_RCVR_SEL0_BASE_IDX 1 +#define regSMU_GPIOPAD_RCVR_SEL1 0x0148 +#define regSMU_GPIOPAD_RCVR_SEL1_BASE_IDX 1 +#define regSMU_GPIOPAD_PU_EN 0x0149 +#define regSMU_GPIOPAD_PU_EN_BASE_IDX 1 +#define regSMU_GPIOPAD_PD_EN 0x014a +#define regSMU_GPIOPAD_PD_EN_BASE_IDX 1 +#define regSMU_GPIOPAD_PINSTRAPS 0x014b +#define regSMU_GPIOPAD_PINSTRAPS_BASE_IDX 1 +#define regDFT_PINSTRAPS 0x014c +#define regDFT_PINSTRAPS_BASE_IDX 1 +#define regSMU_GPIOPAD_INT_STAT_EN 0x014d +#define regSMU_GPIOPAD_INT_STAT_EN_BASE_IDX 1 +#define regSMU_GPIOPAD_INT_STAT 0x014e +#define regSMU_GPIOPAD_INT_STAT_BASE_IDX 1 +#define regSMU_GPIOPAD_INT_STAT_AK 0x014f +#define regSMU_GPIOPAD_INT_STAT_AK_BASE_IDX 1 +#define regSMU_GPIOPAD_INT_EN 0x0150 +#define regSMU_GPIOPAD_INT_EN_BASE_IDX 1 +#define regSMU_GPIOPAD_INT_TYPE 0x0151 +#define regSMU_GPIOPAD_INT_TYPE_BASE_IDX 1 +#define regSMU_GPIOPAD_INT_POLARITY 0x0152 +#define regSMU_GPIOPAD_INT_POLARITY_BASE_IDX 1 +#define regSMUIO_PCC_GPIO_SELECT 0x0155 +#define regSMUIO_PCC_GPIO_SELECT_BASE_IDX 1 +#define regSMU_GPIOPAD_S0 0x0156 +#define regSMU_GPIOPAD_S0_BASE_IDX 1 +#define regSMU_GPIOPAD_S1 0x0157 +#define regSMU_GPIOPAD_S1_BASE_IDX 1 +#define regSMU_GPIOPAD_SCHMEN 0x0158 +#define regSMU_GPIOPAD_SCHMEN_BASE_IDX 1 +#define regSMU_GPIOPAD_SCL_EN 0x0159 +#define regSMU_GPIOPAD_SCL_EN_BASE_IDX 1 +#define regSMU_GPIOPAD_SDA_EN 0x015a +#define regSMU_GPIOPAD_SDA_EN_BASE_IDX 1 +#define regSMUIO_GPIO_INT0_SELECT 0x015b +#define regSMUIO_GPIO_INT0_SELECT_BASE_IDX 1 +#define regSMUIO_GPIO_INT1_SELECT 0x015c +#define regSMUIO_GPIO_INT1_SELECT_BASE_IDX 1 +#define regSMUIO_GPIO_INT2_SELECT 0x015d +#define regSMUIO_GPIO_INT2_SELECT_BASE_IDX 1 +#define regSMUIO_GPIO_INT3_SELECT 0x015e +#define regSMUIO_GPIO_INT3_SELECT_BASE_IDX 1 +#define regSMU_GPIOPAD_MP_INT0_STAT 0x015f +#define regSMU_GPIOPAD_MP_INT0_STAT_BASE_IDX 1 +#define regSMU_GPIOPAD_MP_INT1_STAT 0x0160 +#define regSMU_GPIOPAD_MP_INT1_STAT_BASE_IDX 1 +#define regSMU_GPIOPAD_MP_INT2_STAT 0x0161 +#define regSMU_GPIOPAD_MP_INT2_STAT_BASE_IDX 1 +#define regSMU_GPIOPAD_MP_INT3_STAT 0x0162 +#define regSMU_GPIOPAD_MP_INT3_STAT_BASE_IDX 1 +#define regSMIO_INDEX 0x0163 +#define regSMIO_INDEX_BASE_IDX 1 +#define regS0_VID_SMIO_CNTL 0x0164 +#define regS0_VID_SMIO_CNTL_BASE_IDX 1 +#define regS1_VID_SMIO_CNTL 0x0165 +#define regS1_VID_SMIO_CNTL_BASE_IDX 1 +#define regOPEN_DRAIN_SELECT 0x0166 +#define regOPEN_DRAIN_SELECT_BASE_IDX 1 +#define regSMIO_ENABLE 0x0167 +#define regSMIO_ENABLE_BASE_IDX 1 + +#endif diff --git a/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_13_0_3_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_13_0_3_sh_mask.h new file mode 100644 index 000000000000..be896f3089fe --- /dev/null +++ b/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_13_0_3_sh_mask.h @@ -0,0 +1,428 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#ifndef _smuio_13_0_3_SH_MASK_HEADER +#define _smuio_13_0_3_SH_MASK_HEADER + + +// addressBlock: aid_smuio_smuio_reset_SmuSmuioDec +//SMUIO_MP_RESET_INTR +#define SMUIO_MP_RESET_INTR__SMUIO_MP_RESET_INTR__SHIFT 0x0 +#define SMUIO_MP_RESET_INTR__SMUIO_MP_RESET_INTR_MASK 0x00000001L +//SMUIO_SOC_HALT +#define SMUIO_SOC_HALT__WDT_FORCE_PWROK_EN__SHIFT 0x2 +#define SMUIO_SOC_HALT__WDT_FORCE_RESETn_EN__SHIFT 0x3 +#define SMUIO_SOC_HALT__WDT_FORCE_PWROK_EN_MASK 0x00000004L +#define SMUIO_SOC_HALT__WDT_FORCE_RESETn_EN_MASK 0x00000008L + + +// addressBlock: aid_smuio_smuio_tsc_SmuSmuioDec +//PWROK_REFCLK_GAP_CYCLES +#define PWROK_REFCLK_GAP_CYCLES__Pwrok_PreAssertion_clkgap_cycles__SHIFT 0x0 +#define PWROK_REFCLK_GAP_CYCLES__Pwrok_PostAssertion_clkgap_cycles__SHIFT 0x8 +#define PWROK_REFCLK_GAP_CYCLES__Pwrok_PreAssertion_clkgap_cycles_MASK 0x000000FFL +#define PWROK_REFCLK_GAP_CYCLES__Pwrok_PostAssertion_clkgap_cycles_MASK 0x0000FF00L +//GOLDEN_TSC_INCREMENT_UPPER +#define GOLDEN_TSC_INCREMENT_UPPER__GoldenTscIncrementUpper__SHIFT 0x0 +#define GOLDEN_TSC_INCREMENT_UPPER__GoldenTscIncrementUpper_MASK 0x00FFFFFFL +//GOLDEN_TSC_INCREMENT_LOWER +#define GOLDEN_TSC_INCREMENT_LOWER__GoldenTscIncrementLower__SHIFT 0x0 +#define GOLDEN_TSC_INCREMENT_LOWER__GoldenTscIncrementLower_MASK 0xFFFFFFFFL +//GOLDEN_TSC_COUNT_UPPER +#define GOLDEN_TSC_COUNT_UPPER__GoldenTscCountUpper__SHIFT 0x0 +#define GOLDEN_TSC_COUNT_UPPER__GoldenTscCountUpper_MASK 0x00FFFFFFL +//GOLDEN_TSC_COUNT_LOWER +#define GOLDEN_TSC_COUNT_LOWER__GoldenTscCountLower__SHIFT 0x0 +#define GOLDEN_TSC_COUNT_LOWER__GoldenTscCountLower_MASK 0xFFFFFFFFL +//SOC_GOLDEN_TSC_SHADOW_UPPER +#define SOC_GOLDEN_TSC_SHADOW_UPPER__SocGoldenTscShadowUpper__SHIFT 0x0 +#define SOC_GOLDEN_TSC_SHADOW_UPPER__SocGoldenTscShadowUpper_MASK 0x00FFFFFFL +//SOC_GOLDEN_TSC_SHADOW_LOWER +#define SOC_GOLDEN_TSC_SHADOW_LOWER__SocGoldenTscShadowLower__SHIFT 0x0 +#define SOC_GOLDEN_TSC_SHADOW_LOWER__SocGoldenTscShadowLower_MASK 0xFFFFFFFFL +//SOC_GAP_PWROK +#define SOC_GAP_PWROK__soc_gap_pwrok__SHIFT 0x0 +#define SOC_GAP_PWROK__soc_gap_pwrok_MASK 0x00000001L + + +// addressBlock: aid_smuio_smuio_swtimer_SmuSmuioDec +//PWR_VIRT_RESET_REQ +#define PWR_VIRT_RESET_REQ__VF_FLR__SHIFT 0x0 +#define PWR_VIRT_RESET_REQ__PF_FLR__SHIFT 0x1f +#define PWR_VIRT_RESET_REQ__VF_FLR_MASK 0x7FFFFFFFL +#define PWR_VIRT_RESET_REQ__PF_FLR_MASK 0x80000000L +//PWR_DISP_TIMER_CONTROL +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_COUNT__SHIFT 0x0 +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_ENABLE__SHIFT 0x19 +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_DISABLE__SHIFT 0x1a +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_MASK__SHIFT 0x1b +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_STAT_AK__SHIFT 0x1c +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_TYPE__SHIFT 0x1d +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_MODE__SHIFT 0x1e +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_COUNT_MASK 0x01FFFFFFL +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_ENABLE_MASK 0x02000000L +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_DISABLE_MASK 0x04000000L +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_MASK_MASK 0x08000000L +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_STAT_AK_MASK 0x10000000L +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_TYPE_MASK 0x20000000L +#define PWR_DISP_TIMER_CONTROL__DISP_TIMER_INT_MODE_MASK 0x40000000L +//PWR_DISP_TIMER_DEBUG +#define PWR_DISP_TIMER_DEBUG__DISP_TIMER_INT_RUNNING__SHIFT 0x0 +#define PWR_DISP_TIMER_DEBUG__DISP_TIMER_INT_STAT__SHIFT 0x1 +#define PWR_DISP_TIMER_DEBUG__DISP_TIMER_INT__SHIFT 0x2 +#define PWR_DISP_TIMER_DEBUG__DISP_TIMER_RUN_VAL__SHIFT 0x7 +#define PWR_DISP_TIMER_DEBUG__DISP_TIMER_INT_RUNNING_MASK 0x00000001L +#define PWR_DISP_TIMER_DEBUG__DISP_TIMER_INT_STAT_MASK 0x00000002L +#define PWR_DISP_TIMER_DEBUG__DISP_TIMER_INT_MASK 0x00000004L +#define PWR_DISP_TIMER_DEBUG__DISP_TIMER_RUN_VAL_MASK 0xFFFFFF80L +//PWR_DISP_TIMER2_CONTROL +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_COUNT__SHIFT 0x0 +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_ENABLE__SHIFT 0x19 +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_DISABLE__SHIFT 0x1a +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_MASK__SHIFT 0x1b +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_STAT_AK__SHIFT 0x1c +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_TYPE__SHIFT 0x1d +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_MODE__SHIFT 0x1e +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_COUNT_MASK 0x01FFFFFFL +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_ENABLE_MASK 0x02000000L +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_DISABLE_MASK 0x04000000L +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_MASK_MASK 0x08000000L +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_STAT_AK_MASK 0x10000000L +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_TYPE_MASK 0x20000000L +#define PWR_DISP_TIMER2_CONTROL__DISP_TIMER_INT_MODE_MASK 0x40000000L +//PWR_DISP_TIMER2_DEBUG +#define PWR_DISP_TIMER2_DEBUG__DISP_TIMER_INT_RUNNING__SHIFT 0x0 +#define PWR_DISP_TIMER2_DEBUG__DISP_TIMER_INT_STAT__SHIFT 0x1 +#define PWR_DISP_TIMER2_DEBUG__DISP_TIMER_INT__SHIFT 0x2 +#define PWR_DISP_TIMER2_DEBUG__DISP_TIMER_RUN_VAL__SHIFT 0x7 +#define PWR_DISP_TIMER2_DEBUG__DISP_TIMER_INT_RUNNING_MASK 0x00000001L +#define PWR_DISP_TIMER2_DEBUG__DISP_TIMER_INT_STAT_MASK 0x00000002L +#define PWR_DISP_TIMER2_DEBUG__DISP_TIMER_INT_MASK 0x00000004L +#define PWR_DISP_TIMER2_DEBUG__DISP_TIMER_RUN_VAL_MASK 0xFFFFFF80L +//PWR_DISP_TIMER_GLOBAL_CONTROL +#define PWR_DISP_TIMER_GLOBAL_CONTROL__DISP_TIMER_PULSE_WIDTH__SHIFT 0x0 +#define PWR_DISP_TIMER_GLOBAL_CONTROL__DISP_TIMER_PULSE_EN__SHIFT 0xa +#define PWR_DISP_TIMER_GLOBAL_CONTROL__DISP_TIMER_PULSE_WIDTH_MASK 0x000003FFL +#define PWR_DISP_TIMER_GLOBAL_CONTROL__DISP_TIMER_PULSE_EN_MASK 0x00000400L +//PWR_IH_CONTROL +#define PWR_IH_CONTROL__MAX_CREDIT__SHIFT 0x0 +#define PWR_IH_CONTROL__DISP_TIMER_TRIGGER_MASK__SHIFT 0x5 +#define PWR_IH_CONTROL__DISP_TIMER2_TRIGGER_MASK__SHIFT 0x6 +#define PWR_IH_CONTROL__PWR_IH_CLK_GATE_EN__SHIFT 0x1f +#define PWR_IH_CONTROL__MAX_CREDIT_MASK 0x0000001FL +#define PWR_IH_CONTROL__DISP_TIMER_TRIGGER_MASK_MASK 0x00000020L +#define PWR_IH_CONTROL__DISP_TIMER2_TRIGGER_MASK_MASK 0x00000040L +#define PWR_IH_CONTROL__PWR_IH_CLK_GATE_EN_MASK 0x80000000L + + +// addressBlock: aid_smuio_smuio_misc_SmuSmuioDec +//SMUIO_MCM_CONFIG +#define SMUIO_MCM_CONFIG__DIE_ID__SHIFT 0x0 +#define SMUIO_MCM_CONFIG__PKG_TYPE__SHIFT 0x2 +#define SMUIO_MCM_CONFIG__SOCKET_ID__SHIFT 0x8 +#define SMUIO_MCM_CONFIG__PKG_SUBTYPE__SHIFT 0xc +#define SMUIO_MCM_CONFIG__CONSOLE_K__SHIFT 0x10 +#define SMUIO_MCM_CONFIG__CONSOLE_A__SHIFT 0x11 +#define SMUIO_MCM_CONFIG__TOPOLOGY_ID__SHIFT 0x12 +#define SMUIO_MCM_CONFIG__DIE_ID_MASK 0x00000003L +#define SMUIO_MCM_CONFIG__PKG_TYPE_MASK 0x0000003CL +#define SMUIO_MCM_CONFIG__SOCKET_ID_MASK 0x00000F00L +#define SMUIO_MCM_CONFIG__PKG_SUBTYPE_MASK 0x00001000L +#define SMUIO_MCM_CONFIG__CONSOLE_K_MASK 0x00010000L +#define SMUIO_MCM_CONFIG__CONSOLE_A_MASK 0x00020000L +#define SMUIO_MCM_CONFIG__TOPOLOGY_ID_MASK 0x007C0000L +//IP_DISCOVERY_VERSION +#define IP_DISCOVERY_VERSION__IP_DISCOVERY_VERSION__SHIFT 0x0 +#define IP_DISCOVERY_VERSION__IP_DISCOVERY_VERSION_MASK 0xFFFFFFFFL +//SCRATCH_REGISTER0 +#define SCRATCH_REGISTER0__ScratchPad0__SHIFT 0x0 +#define SCRATCH_REGISTER0__ScratchPad0_MASK 0xFFFFFFFFL +//SCRATCH_REGISTER1 +#define SCRATCH_REGISTER1__ScratchPad1__SHIFT 0x0 +#define SCRATCH_REGISTER1__ScratchPad1_MASK 0xFFFFFFFFL +//SCRATCH_REGISTER2 +#define SCRATCH_REGISTER2__ScratchPad2__SHIFT 0x0 +#define SCRATCH_REGISTER2__ScratchPad2_MASK 0xFFFFFFFFL +//SCRATCH_REGISTER3 +#define SCRATCH_REGISTER3__ScratchPad3__SHIFT 0x0 +#define SCRATCH_REGISTER3__ScratchPad3_MASK 0xFFFFFFFFL +//SCRATCH_REGISTER4 +#define SCRATCH_REGISTER4__ScratchPad4__SHIFT 0x0 +#define SCRATCH_REGISTER4__ScratchPad4_MASK 0xFFFFFFFFL +//SCRATCH_REGISTER5 +#define SCRATCH_REGISTER5__ScratchPad5__SHIFT 0x0 +#define SCRATCH_REGISTER5__ScratchPad5_MASK 0xFFFFFFFFL +//SCRATCH_REGISTER6 +#define SCRATCH_REGISTER6__ScratchPad6__SHIFT 0x0 +#define SCRATCH_REGISTER6__ScratchPad6_MASK 0xFFFFFFFFL +//SCRATCH_REGISTER7 +#define SCRATCH_REGISTER7__ScratchPad7__SHIFT 0x0 +#define SCRATCH_REGISTER7__ScratchPad7_MASK 0xFFFFFFFFL + + +// addressBlock: aid_smuio_smuio_gpio_SmuSmuioDec +//SMU_GPIOPAD_SW_INT_STAT +#define SMU_GPIOPAD_SW_INT_STAT__SW_INT_STAT__SHIFT 0x0 +#define SMU_GPIOPAD_SW_INT_STAT__SW_INT_STAT_MASK 0x00000001L +//SMU_GPIOPAD_MASK +#define SMU_GPIOPAD_MASK__GPIO_MASK__SHIFT 0x0 +#define SMU_GPIOPAD_MASK__GPIO_MASK_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_A +#define SMU_GPIOPAD_A__GPIO_A__SHIFT 0x0 +#define SMU_GPIOPAD_A__GPIO_A_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_TXIMPSEL +#define SMU_GPIOPAD_TXIMPSEL__GPIO_TXIMPSEL__SHIFT 0x0 +#define SMU_GPIOPAD_TXIMPSEL__GPIO_TXIMPSEL_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_EN +#define SMU_GPIOPAD_EN__GPIO_EN__SHIFT 0x0 +#define SMU_GPIOPAD_EN__GPIO_EN_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_Y +#define SMU_GPIOPAD_Y__GPIO_Y__SHIFT 0x0 +#define SMU_GPIOPAD_Y__GPIO_Y_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_RXEN +#define SMU_GPIOPAD_RXEN__GPIO_RXEN__SHIFT 0x0 +#define SMU_GPIOPAD_RXEN__GPIO_RXEN_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_RCVR_SEL0 +#define SMU_GPIOPAD_RCVR_SEL0__GPIO_RCVR_SEL0__SHIFT 0x0 +#define SMU_GPIOPAD_RCVR_SEL0__GPIO_RCVR_SEL0_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_RCVR_SEL1 +#define SMU_GPIOPAD_RCVR_SEL1__GPIO_RCVR_SEL1__SHIFT 0x0 +#define SMU_GPIOPAD_RCVR_SEL1__GPIO_RCVR_SEL1_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_PU_EN +#define SMU_GPIOPAD_PU_EN__GPIO_PU_EN__SHIFT 0x0 +#define SMU_GPIOPAD_PU_EN__GPIO_PU_EN_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_PD_EN +#define SMU_GPIOPAD_PD_EN__GPIO_PD_EN__SHIFT 0x0 +#define SMU_GPIOPAD_PD_EN__GPIO_PD_EN_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_PINSTRAPS +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_0__SHIFT 0x0 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_1__SHIFT 0x1 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_2__SHIFT 0x2 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_3__SHIFT 0x3 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_4__SHIFT 0x4 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_5__SHIFT 0x5 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_6__SHIFT 0x6 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_7__SHIFT 0x7 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_8__SHIFT 0x8 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_9__SHIFT 0x9 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_10__SHIFT 0xa +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_11__SHIFT 0xb +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_12__SHIFT 0xc +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_13__SHIFT 0xd +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_14__SHIFT 0xe +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_15__SHIFT 0xf +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_16__SHIFT 0x10 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_17__SHIFT 0x11 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_18__SHIFT 0x12 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_19__SHIFT 0x13 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_20__SHIFT 0x14 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_21__SHIFT 0x15 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_22__SHIFT 0x16 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_23__SHIFT 0x17 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_24__SHIFT 0x18 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_25__SHIFT 0x19 +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_26__SHIFT 0x1a +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_27__SHIFT 0x1b +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_28__SHIFT 0x1c +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_29__SHIFT 0x1d +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_30__SHIFT 0x1e +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_0_MASK 0x00000001L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_1_MASK 0x00000002L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_2_MASK 0x00000004L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_3_MASK 0x00000008L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_4_MASK 0x00000010L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_5_MASK 0x00000020L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_6_MASK 0x00000040L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_7_MASK 0x00000080L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_8_MASK 0x00000100L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_9_MASK 0x00000200L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_10_MASK 0x00000400L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_11_MASK 0x00000800L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_12_MASK 0x00001000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_13_MASK 0x00002000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_14_MASK 0x00004000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_15_MASK 0x00008000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_16_MASK 0x00010000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_17_MASK 0x00020000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_18_MASK 0x00040000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_19_MASK 0x00080000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_20_MASK 0x00100000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_21_MASK 0x00200000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_22_MASK 0x00400000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_23_MASK 0x00800000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_24_MASK 0x01000000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_25_MASK 0x02000000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_26_MASK 0x04000000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_27_MASK 0x08000000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_28_MASK 0x10000000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_29_MASK 0x20000000L +#define SMU_GPIOPAD_PINSTRAPS__GPIO_PINSTRAP_30_MASK 0x40000000L +//DFT_PINSTRAPS +#define DFT_PINSTRAPS__DFT_PINSTRAPS__SHIFT 0x0 +#define DFT_PINSTRAPS__DFT_PINSTRAPS_MASK 0x000003FFL +//SMU_GPIOPAD_INT_STAT_EN +#define SMU_GPIOPAD_INT_STAT_EN__GPIO_INT_STAT_EN__SHIFT 0x0 +#define SMU_GPIOPAD_INT_STAT_EN__SW_INITIATED_INT_STAT_EN__SHIFT 0x1f +#define SMU_GPIOPAD_INT_STAT_EN__GPIO_INT_STAT_EN_MASK 0x1FFFFFFFL +#define SMU_GPIOPAD_INT_STAT_EN__SW_INITIATED_INT_STAT_EN_MASK 0x80000000L +//SMU_GPIOPAD_INT_STAT +#define SMU_GPIOPAD_INT_STAT__GPIO_INT_STAT__SHIFT 0x0 +#define SMU_GPIOPAD_INT_STAT__SW_INITIATED_INT_STAT__SHIFT 0x1f +#define SMU_GPIOPAD_INT_STAT__GPIO_INT_STAT_MASK 0x1FFFFFFFL +#define SMU_GPIOPAD_INT_STAT__SW_INITIATED_INT_STAT_MASK 0x80000000L +//SMU_GPIOPAD_INT_STAT_AK +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_0__SHIFT 0x0 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_1__SHIFT 0x1 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_2__SHIFT 0x2 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_3__SHIFT 0x3 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_4__SHIFT 0x4 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_5__SHIFT 0x5 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_6__SHIFT 0x6 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_7__SHIFT 0x7 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_8__SHIFT 0x8 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_9__SHIFT 0x9 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_10__SHIFT 0xa +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_11__SHIFT 0xb +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_12__SHIFT 0xc +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_13__SHIFT 0xd +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_14__SHIFT 0xe +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_15__SHIFT 0xf +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_16__SHIFT 0x10 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_17__SHIFT 0x11 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_18__SHIFT 0x12 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_19__SHIFT 0x13 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_20__SHIFT 0x14 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_21__SHIFT 0x15 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_22__SHIFT 0x16 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_23__SHIFT 0x17 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_24__SHIFT 0x18 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_25__SHIFT 0x19 +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_26__SHIFT 0x1a +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_27__SHIFT 0x1b +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_28__SHIFT 0x1c +#define SMU_GPIOPAD_INT_STAT_AK__SW_INITIATED_INT_STAT_AK__SHIFT 0x1f +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_0_MASK 0x00000001L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_1_MASK 0x00000002L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_2_MASK 0x00000004L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_3_MASK 0x00000008L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_4_MASK 0x00000010L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_5_MASK 0x00000020L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_6_MASK 0x00000040L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_7_MASK 0x00000080L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_8_MASK 0x00000100L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_9_MASK 0x00000200L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_10_MASK 0x00000400L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_11_MASK 0x00000800L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_12_MASK 0x00001000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_13_MASK 0x00002000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_14_MASK 0x00004000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_15_MASK 0x00008000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_16_MASK 0x00010000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_17_MASK 0x00020000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_18_MASK 0x00040000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_19_MASK 0x00080000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_20_MASK 0x00100000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_21_MASK 0x00200000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_22_MASK 0x00400000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_23_MASK 0x00800000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_24_MASK 0x01000000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_25_MASK 0x02000000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_26_MASK 0x04000000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_27_MASK 0x08000000L +#define SMU_GPIOPAD_INT_STAT_AK__GPIO_INT_STAT_AK_28_MASK 0x10000000L +#define SMU_GPIOPAD_INT_STAT_AK__SW_INITIATED_INT_STAT_AK_MASK 0x80000000L +//SMU_GPIOPAD_INT_EN +#define SMU_GPIOPAD_INT_EN__GPIO_INT_EN__SHIFT 0x0 +#define SMU_GPIOPAD_INT_EN__SW_INITIATED_INT_EN__SHIFT 0x1f +#define SMU_GPIOPAD_INT_EN__GPIO_INT_EN_MASK 0x1FFFFFFFL +#define SMU_GPIOPAD_INT_EN__SW_INITIATED_INT_EN_MASK 0x80000000L +//SMU_GPIOPAD_INT_TYPE +#define SMU_GPIOPAD_INT_TYPE__GPIO_INT_TYPE__SHIFT 0x0 +#define SMU_GPIOPAD_INT_TYPE__SW_INITIATED_INT_TYPE__SHIFT 0x1f +#define SMU_GPIOPAD_INT_TYPE__GPIO_INT_TYPE_MASK 0x1FFFFFFFL +#define SMU_GPIOPAD_INT_TYPE__SW_INITIATED_INT_TYPE_MASK 0x80000000L +//SMU_GPIOPAD_INT_POLARITY +#define SMU_GPIOPAD_INT_POLARITY__GPIO_INT_POLARITY__SHIFT 0x0 +#define SMU_GPIOPAD_INT_POLARITY__SW_INITIATED_INT_POLARITY__SHIFT 0x1f +#define SMU_GPIOPAD_INT_POLARITY__GPIO_INT_POLARITY_MASK 0x1FFFFFFFL +#define SMU_GPIOPAD_INT_POLARITY__SW_INITIATED_INT_POLARITY_MASK 0x80000000L +//SMUIO_PCC_GPIO_SELECT +#define SMUIO_PCC_GPIO_SELECT__GPIO__SHIFT 0x0 +#define SMUIO_PCC_GPIO_SELECT__GPIO_MASK 0xFFFFFFFFL +//SMU_GPIOPAD_S0 +#define SMU_GPIOPAD_S0__GPIO_S0__SHIFT 0x0 +#define SMU_GPIOPAD_S0__GPIO_S0_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_S1 +#define SMU_GPIOPAD_S1__GPIO_S1__SHIFT 0x0 +#define SMU_GPIOPAD_S1__GPIO_S1_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_SCHMEN +#define SMU_GPIOPAD_SCHMEN__GPIO_SCHMEN__SHIFT 0x0 +#define SMU_GPIOPAD_SCHMEN__GPIO_SCHMEN_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_SCL_EN +#define SMU_GPIOPAD_SCL_EN__GPIO_SCL_EN__SHIFT 0x0 +#define SMU_GPIOPAD_SCL_EN__GPIO_SCL_EN_MASK 0x7FFFFFFFL +//SMU_GPIOPAD_SDA_EN +#define SMU_GPIOPAD_SDA_EN__GPIO_SDA_EN__SHIFT 0x0 +#define SMU_GPIOPAD_SDA_EN__GPIO_SDA_EN_MASK 0x7FFFFFFFL +//SMUIO_GPIO_INT0_SELECT +#define SMUIO_GPIO_INT0_SELECT__GPIO_INT0_SELECT__SHIFT 0x0 +#define SMUIO_GPIO_INT0_SELECT__GPIO_INT0_SELECT_MASK 0xFFFFFFFFL +//SMUIO_GPIO_INT1_SELECT +#define SMUIO_GPIO_INT1_SELECT__GPIO_INT1_SELECT__SHIFT 0x0 +#define SMUIO_GPIO_INT1_SELECT__GPIO_INT1_SELECT_MASK 0xFFFFFFFFL +//SMUIO_GPIO_INT2_SELECT +#define SMUIO_GPIO_INT2_SELECT__GPIO_INT2_SELECT__SHIFT 0x0 +#define SMUIO_GPIO_INT2_SELECT__GPIO_INT2_SELECT_MASK 0xFFFFFFFFL +//SMUIO_GPIO_INT3_SELECT +#define SMUIO_GPIO_INT3_SELECT__GPIO_INT3_SELECT__SHIFT 0x0 +#define SMUIO_GPIO_INT3_SELECT__GPIO_INT3_SELECT_MASK 0xFFFFFFFFL +//SMU_GPIOPAD_MP_INT0_STAT +#define SMU_GPIOPAD_MP_INT0_STAT__GPIO_MP_INT0_STAT__SHIFT 0x0 +#define SMU_GPIOPAD_MP_INT0_STAT__GPIO_MP_INT0_STAT_MASK 0x1FFFFFFFL +//SMU_GPIOPAD_MP_INT1_STAT +#define SMU_GPIOPAD_MP_INT1_STAT__GPIO_MP_INT1_STAT__SHIFT 0x0 +#define SMU_GPIOPAD_MP_INT1_STAT__GPIO_MP_INT1_STAT_MASK 0x1FFFFFFFL +//SMU_GPIOPAD_MP_INT2_STAT +#define SMU_GPIOPAD_MP_INT2_STAT__GPIO_MP_INT2_STAT__SHIFT 0x0 +#define SMU_GPIOPAD_MP_INT2_STAT__GPIO_MP_INT2_STAT_MASK 0x1FFFFFFFL +//SMU_GPIOPAD_MP_INT3_STAT +#define SMU_GPIOPAD_MP_INT3_STAT__GPIO_MP_INT3_STAT__SHIFT 0x0 +#define SMU_GPIOPAD_MP_INT3_STAT__GPIO_MP_INT3_STAT_MASK 0x1FFFFFFFL +//SMIO_INDEX +#define SMIO_INDEX__SW_SMIO_INDEX__SHIFT 0x0 +#define SMIO_INDEX__SW_SMIO_INDEX_MASK 0x00000001L +//S0_VID_SMIO_CNTL +#define S0_VID_SMIO_CNTL__S0_SMIO_VALUES__SHIFT 0x0 +#define S0_VID_SMIO_CNTL__S0_SMIO_VALUES_MASK 0xFFFFFFFFL +//S1_VID_SMIO_CNTL +#define S1_VID_SMIO_CNTL__S1_SMIO_VALUES__SHIFT 0x0 +#define S1_VID_SMIO_CNTL__S1_SMIO_VALUES_MASK 0xFFFFFFFFL +//OPEN_DRAIN_SELECT +#define OPEN_DRAIN_SELECT__OPEN_DRAIN_SELECT__SHIFT 0x0 +#define OPEN_DRAIN_SELECT__RESERVED__SHIFT 0x1f +#define OPEN_DRAIN_SELECT__OPEN_DRAIN_SELECT_MASK 0x7FFFFFFFL +#define OPEN_DRAIN_SELECT__RESERVED_MASK 0x80000000L +//SMIO_ENABLE +#define SMIO_ENABLE__SMIO_ENABLE__SHIFT 0x0 +#define SMIO_ENABLE__SMIO_ENABLE_MASK 0xFFFFFFFFL + +#endif diff --git a/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_4_0_3_offset.h b/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_4_0_3_offset.h new file mode 100644 index 000000000000..e9742d10de1c --- /dev/null +++ b/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_4_0_3_offset.h @@ -0,0 +1,2332 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#ifndef _vcn_4_0_3_OFFSET_HEADER +#define _vcn_4_0_3_OFFSET_HEADER + + + +// addressBlock: aid_uvd0_uvddec +// base address: 0x1fb00 +#define regUVD_TOP_CTRL 0x00c0 +#define regUVD_TOP_CTRL_BASE_IDX 1 +#define regUVD_CGC_GATE 0x00c1 +#define regUVD_CGC_GATE_BASE_IDX 1 +#define regUVD_CGC_CTRL 0x00c2 +#define regUVD_CGC_CTRL_BASE_IDX 1 +#define regAVM_SUVD_CGC_GATE 0x00c4 +#define regAVM_SUVD_CGC_GATE_BASE_IDX 1 +#define regCDEFE_SUVD_CGC_GATE 0x00c4 +#define regCDEFE_SUVD_CGC_GATE_BASE_IDX 1 +#define regEFC_SUVD_CGC_GATE 0x00c4 +#define regEFC_SUVD_CGC_GATE_BASE_IDX 1 +#define regENT_SUVD_CGC_GATE 0x00c4 +#define regENT_SUVD_CGC_GATE_BASE_IDX 1 +#define regIME_SUVD_CGC_GATE 0x00c4 +#define regIME_SUVD_CGC_GATE_BASE_IDX 1 +#define regPPU_SUVD_CGC_GATE 0x00c4 +#define regPPU_SUVD_CGC_GATE_BASE_IDX 1 +#define regSAOE_SUVD_CGC_GATE 0x00c4 +#define regSAOE_SUVD_CGC_GATE_BASE_IDX 1 +#define regSCM_SUVD_CGC_GATE 0x00c4 +#define regSCM_SUVD_CGC_GATE_BASE_IDX 1 +#define regSDB_SUVD_CGC_GATE 0x00c4 +#define regSDB_SUVD_CGC_GATE_BASE_IDX 1 +#define regSIT0_NXT_SUVD_CGC_GATE 0x00c4 +#define regSIT0_NXT_SUVD_CGC_GATE_BASE_IDX 1 +#define regSIT1_NXT_SUVD_CGC_GATE 0x00c4 +#define regSIT1_NXT_SUVD_CGC_GATE_BASE_IDX 1 +#define regSIT2_NXT_SUVD_CGC_GATE 0x00c4 +#define regSIT2_NXT_SUVD_CGC_GATE_BASE_IDX 1 +#define regSIT_SUVD_CGC_GATE 0x00c4 +#define regSIT_SUVD_CGC_GATE_BASE_IDX 1 +#define regSMPA_SUVD_CGC_GATE 0x00c4 +#define regSMPA_SUVD_CGC_GATE_BASE_IDX 1 +#define regSMP_SUVD_CGC_GATE 0x00c4 +#define regSMP_SUVD_CGC_GATE_BASE_IDX 1 +#define regSRE_SUVD_CGC_GATE 0x00c4 +#define regSRE_SUVD_CGC_GATE_BASE_IDX 1 +#define regUVD_MPBE0_SUVD_CGC_GATE 0x00c4 +#define regUVD_MPBE0_SUVD_CGC_GATE_BASE_IDX 1 +#define regUVD_MPBE1_SUVD_CGC_GATE 0x00c4 +#define regUVD_MPBE1_SUVD_CGC_GATE_BASE_IDX 1 +#define regUVD_SUVD_CGC_GATE 0x00c4 +#define regUVD_SUVD_CGC_GATE_BASE_IDX 1 +#define regAVM_SUVD_CGC_GATE2 0x00c5 +#define regAVM_SUVD_CGC_GATE2_BASE_IDX 1 +#define regCDEFE_SUVD_CGC_GATE2 0x00c5 +#define regCDEFE_SUVD_CGC_GATE2_BASE_IDX 1 +#define regDBR_SUVD_CGC_GATE2 0x00c5 +#define regDBR_SUVD_CGC_GATE2_BASE_IDX 1 +#define regENT_SUVD_CGC_GATE2 0x00c5 +#define regENT_SUVD_CGC_GATE2_BASE_IDX 1 +#define regIME_SUVD_CGC_GATE2 0x00c5 +#define regIME_SUVD_CGC_GATE2_BASE_IDX 1 +#define regMPC1_SUVD_CGC_GATE2 0x00c5 +#define regMPC1_SUVD_CGC_GATE2_BASE_IDX 1 +#define regSAOE_SUVD_CGC_GATE2 0x00c5 +#define regSAOE_SUVD_CGC_GATE2_BASE_IDX 1 +#define regSDB_SUVD_CGC_GATE2 0x00c5 +#define regSDB_SUVD_CGC_GATE2_BASE_IDX 1 +#define regSIT0_NXT_SUVD_CGC_GATE2 0x00c5 +#define regSIT0_NXT_SUVD_CGC_GATE2_BASE_IDX 1 +#define regSIT1_NXT_SUVD_CGC_GATE2 0x00c5 +#define regSIT1_NXT_SUVD_CGC_GATE2_BASE_IDX 1 +#define regSIT2_NXT_SUVD_CGC_GATE2 0x00c5 +#define regSIT2_NXT_SUVD_CGC_GATE2_BASE_IDX 1 +#define regSIT_SUVD_CGC_GATE2 0x00c5 +#define regSIT_SUVD_CGC_GATE2_BASE_IDX 1 +#define regSMPA_SUVD_CGC_GATE2 0x00c5 +#define regSMPA_SUVD_CGC_GATE2_BASE_IDX 1 +#define regSMP_SUVD_CGC_GATE2 0x00c5 +#define regSMP_SUVD_CGC_GATE2_BASE_IDX 1 +#define regSRE_SUVD_CGC_GATE2 0x00c5 +#define regSRE_SUVD_CGC_GATE2_BASE_IDX 1 +#define regUVD_MPBE0_SUVD_CGC_GATE2 0x00c5 +#define regUVD_MPBE0_SUVD_CGC_GATE2_BASE_IDX 1 +#define regUVD_MPBE1_SUVD_CGC_GATE2 0x00c5 +#define regUVD_MPBE1_SUVD_CGC_GATE2_BASE_IDX 1 +#define regUVD_SUVD_CGC_GATE2 0x00c5 +#define regUVD_SUVD_CGC_GATE2_BASE_IDX 1 +#define regAVM_SUVD_CGC_CTRL 0x00c6 +#define regAVM_SUVD_CGC_CTRL_BASE_IDX 1 +#define regCDEFE_SUVD_CGC_CTRL 0x00c6 +#define regCDEFE_SUVD_CGC_CTRL_BASE_IDX 1 +#define regDBR_SUVD_CGC_CTRL 0x00c6 +#define regDBR_SUVD_CGC_CTRL_BASE_IDX 1 +#define regEFC_SUVD_CGC_CTRL 0x00c6 +#define regEFC_SUVD_CGC_CTRL_BASE_IDX 1 +#define regENT_SUVD_CGC_CTRL 0x00c6 +#define regENT_SUVD_CGC_CTRL_BASE_IDX 1 +#define regIME_SUVD_CGC_CTRL 0x00c6 +#define regIME_SUVD_CGC_CTRL_BASE_IDX 1 +#define regMPC1_SUVD_CGC_CTRL 0x00c6 +#define regMPC1_SUVD_CGC_CTRL_BASE_IDX 1 +#define regPPU_SUVD_CGC_CTRL 0x00c6 +#define regPPU_SUVD_CGC_CTRL_BASE_IDX 1 +#define regSAOE_SUVD_CGC_CTRL 0x00c6 +#define regSAOE_SUVD_CGC_CTRL_BASE_IDX 1 +#define regSCM_SUVD_CGC_CTRL 0x00c6 +#define regSCM_SUVD_CGC_CTRL_BASE_IDX 1 +#define regSDB_SUVD_CGC_CTRL 0x00c6 +#define regSDB_SUVD_CGC_CTRL_BASE_IDX 1 +#define regSIT0_NXT_SUVD_CGC_CTRL 0x00c6 +#define regSIT0_NXT_SUVD_CGC_CTRL_BASE_IDX 1 +#define regSIT1_NXT_SUVD_CGC_CTRL 0x00c6 +#define regSIT1_NXT_SUVD_CGC_CTRL_BASE_IDX 1 +#define regSIT2_NXT_SUVD_CGC_CTRL 0x00c6 +#define regSIT2_NXT_SUVD_CGC_CTRL_BASE_IDX 1 +#define regSIT_SUVD_CGC_CTRL 0x00c6 +#define regSIT_SUVD_CGC_CTRL_BASE_IDX 1 +#define regSMPA_SUVD_CGC_CTRL 0x00c6 +#define regSMPA_SUVD_CGC_CTRL_BASE_IDX 1 +#define regSMP_SUVD_CGC_CTRL 0x00c6 +#define regSMP_SUVD_CGC_CTRL_BASE_IDX 1 +#define regSRE_SUVD_CGC_CTRL 0x00c6 +#define regSRE_SUVD_CGC_CTRL_BASE_IDX 1 +#define regUVD_MPBE0_SUVD_CGC_CTRL 0x00c6 +#define regUVD_MPBE0_SUVD_CGC_CTRL_BASE_IDX 1 +#define regUVD_MPBE1_SUVD_CGC_CTRL 0x00c6 +#define regUVD_MPBE1_SUVD_CGC_CTRL_BASE_IDX 1 +#define regUVD_SUVD_CGC_CTRL 0x00c6 +#define regUVD_SUVD_CGC_CTRL_BASE_IDX 1 +#define regUVD_CGC_CTRL3 0x00ca +#define regUVD_CGC_CTRL3_BASE_IDX 1 +#define regUVD_GPCOM_VCPU_DATA0 0x00d0 +#define regUVD_GPCOM_VCPU_DATA0_BASE_IDX 1 +#define regUVD_GPCOM_VCPU_DATA1 0x00d1 +#define regUVD_GPCOM_VCPU_DATA1_BASE_IDX 1 +#define regUVD_GPCOM_SYS_CMD 0x00d2 +#define regUVD_GPCOM_SYS_CMD_BASE_IDX 1 +#define regUVD_GPCOM_SYS_DATA0 0x00d3 +#define regUVD_GPCOM_SYS_DATA0_BASE_IDX 1 +#define regUVD_GPCOM_SYS_DATA1 0x00d4 +#define regUVD_GPCOM_SYS_DATA1_BASE_IDX 1 +#define regUVD_VCPU_INT_EN 0x00d5 +#define regUVD_VCPU_INT_EN_BASE_IDX 1 +#define regUVD_VCPU_INT_STATUS 0x00d6 +#define regUVD_VCPU_INT_STATUS_BASE_IDX 1 +#define regUVD_VCPU_INT_ACK 0x00d7 +#define regUVD_VCPU_INT_ACK_BASE_IDX 1 +#define regUVD_VCPU_INT_ROUTE 0x00d8 +#define regUVD_VCPU_INT_ROUTE_BASE_IDX 1 +#define regUVD_DRV_FW_MSG 0x00d9 +#define regUVD_DRV_FW_MSG_BASE_IDX 1 +#define regUVD_FW_DRV_MSG_ACK 0x00da +#define regUVD_FW_DRV_MSG_ACK_BASE_IDX 1 +#define regUVD_SUVD_INT_EN 0x00db +#define regUVD_SUVD_INT_EN_BASE_IDX 1 +#define regUVD_SUVD_INT_STATUS 0x00dc +#define regUVD_SUVD_INT_STATUS_BASE_IDX 1 +#define regUVD_SUVD_INT_ACK 0x00dd +#define regUVD_SUVD_INT_ACK_BASE_IDX 1 +#define regUVD_ENC_VCPU_INT_EN 0x00de +#define regUVD_ENC_VCPU_INT_EN_BASE_IDX 1 +#define regUVD_ENC_VCPU_INT_STATUS 0x00df +#define regUVD_ENC_VCPU_INT_STATUS_BASE_IDX 1 +#define regUVD_ENC_VCPU_INT_ACK 0x00e0 +#define regUVD_ENC_VCPU_INT_ACK_BASE_IDX 1 +#define regUVD_MASTINT_EN 0x00e1 +#define regUVD_MASTINT_EN_BASE_IDX 1 +#define regUVD_SYS_INT_EN 0x00e2 +#define regUVD_SYS_INT_EN_BASE_IDX 1 +#define regUVD_SYS_INT_STATUS 0x00e3 +#define regUVD_SYS_INT_STATUS_BASE_IDX 1 +#define regUVD_SYS_INT_ACK 0x00e4 +#define regUVD_SYS_INT_ACK_BASE_IDX 1 +#define regUVD_JOB_DONE 0x00e5 +#define regUVD_JOB_DONE_BASE_IDX 1 +#define regUVD_CBUF_ID 0x00e6 +#define regUVD_CBUF_ID_BASE_IDX 1 +#define regUVD_CONTEXT_ID 0x00e7 +#define regUVD_CONTEXT_ID_BASE_IDX 1 +#define regUVD_CONTEXT_ID2 0x00e8 +#define regUVD_CONTEXT_ID2_BASE_IDX 1 +#define regUVD_NO_OP 0x00e9 +#define regUVD_NO_OP_BASE_IDX 1 +#define regUVD_RB_BASE_LO 0x00ea +#define regUVD_RB_BASE_LO_BASE_IDX 1 +#define regUVD_RB_BASE_HI 0x00eb +#define regUVD_RB_BASE_HI_BASE_IDX 1 +#define regUVD_RB_SIZE 0x00ec +#define regUVD_RB_SIZE_BASE_IDX 1 +#define regUVD_RB_BASE_LO2 0x00ef +#define regUVD_RB_BASE_LO2_BASE_IDX 1 +#define regUVD_RB_BASE_HI2 0x00f0 +#define regUVD_RB_BASE_HI2_BASE_IDX 1 +#define regUVD_RB_SIZE2 0x00f1 +#define regUVD_RB_SIZE2_BASE_IDX 1 +#define regUVD_RB_BASE_LO3 0x00f4 +#define regUVD_RB_BASE_LO3_BASE_IDX 1 +#define regUVD_RB_BASE_HI3 0x00f5 +#define regUVD_RB_BASE_HI3_BASE_IDX 1 +#define regUVD_RB_SIZE3 0x00f6 +#define regUVD_RB_SIZE3_BASE_IDX 1 +#define regUVD_RB_BASE_LO4 0x00f9 +#define regUVD_RB_BASE_LO4_BASE_IDX 1 +#define regUVD_RB_BASE_HI4 0x00fa +#define regUVD_RB_BASE_HI4_BASE_IDX 1 +#define regUVD_RB_SIZE4 0x00fb +#define regUVD_RB_SIZE4_BASE_IDX 1 +#define regUVD_OUT_RB_BASE_LO 0x00fe +#define regUVD_OUT_RB_BASE_LO_BASE_IDX 1 +#define regUVD_OUT_RB_BASE_HI 0x00ff +#define regUVD_OUT_RB_BASE_HI_BASE_IDX 1 +#define regUVD_OUT_RB_SIZE 0x0100 +#define regUVD_OUT_RB_SIZE_BASE_IDX 1 +#define regUVD_IOV_ACTIVE_FCN_ID 0x0103 +#define regUVD_IOV_ACTIVE_FCN_ID_BASE_IDX 1 +#define regUVD_IOV_MAILBOX 0x0104 +#define regUVD_IOV_MAILBOX_BASE_IDX 1 +#define regUVD_IOV_MAILBOX_RESP 0x0105 +#define regUVD_IOV_MAILBOX_RESP_BASE_IDX 1 +#define regUVD_RB_ARB_CTRL 0x0106 +#define regUVD_RB_ARB_CTRL_BASE_IDX 1 +#define regUVD_CTX_INDEX 0x0107 +#define regUVD_CTX_INDEX_BASE_IDX 1 +#define regUVD_CTX_DATA 0x0108 +#define regUVD_CTX_DATA_BASE_IDX 1 +#define regUVD_CXW_WR 0x0109 +#define regUVD_CXW_WR_BASE_IDX 1 +#define regUVD_CXW_WR_INT_ID 0x010a +#define regUVD_CXW_WR_INT_ID_BASE_IDX 1 +#define regUVD_CXW_WR_INT_CTX_ID 0x010b +#define regUVD_CXW_WR_INT_CTX_ID_BASE_IDX 1 +#define regUVD_CXW_INT_ID 0x010c +#define regUVD_CXW_INT_ID_BASE_IDX 1 +#define regUVD_MPEG2_ERROR 0x010d +#define regUVD_MPEG2_ERROR_BASE_IDX 1 +#define regUVD_YBASE 0x0110 +#define regUVD_YBASE_BASE_IDX 1 +#define regUVD_UVBASE 0x0111 +#define regUVD_UVBASE_BASE_IDX 1 +#define regUVD_PITCH 0x0112 +#define regUVD_PITCH_BASE_IDX 1 +#define regUVD_WIDTH 0x0113 +#define regUVD_WIDTH_BASE_IDX 1 +#define regUVD_HEIGHT 0x0114 +#define regUVD_HEIGHT_BASE_IDX 1 +#define regUVD_PICCOUNT 0x0115 +#define regUVD_PICCOUNT_BASE_IDX 1 +#define regUVD_MPRD_INITIAL_XY 0x0116 +#define regUVD_MPRD_INITIAL_XY_BASE_IDX 1 +#define regUVD_MPEG2_CTRL 0x0117 +#define regUVD_MPEG2_CTRL_BASE_IDX 1 +#define regUVD_MB_CTL_BUF_BASE 0x0118 +#define regUVD_MB_CTL_BUF_BASE_BASE_IDX 1 +#define regUVD_PIC_CTL_BUF_BASE 0x0119 +#define regUVD_PIC_CTL_BUF_BASE_BASE_IDX 1 +#define regUVD_DXVA_BUF_SIZE 0x011a +#define regUVD_DXVA_BUF_SIZE_BASE_IDX 1 +#define regUVD_SCRATCH_NP 0x011b +#define regUVD_SCRATCH_NP_BASE_IDX 1 +#define regUVD_CLK_SWT_HANDSHAKE 0x011c +#define regUVD_CLK_SWT_HANDSHAKE_BASE_IDX 1 +#define regUVD_GP_SCRATCH0 0x011e +#define regUVD_GP_SCRATCH0_BASE_IDX 1 +#define regUVD_GP_SCRATCH1 0x011f +#define regUVD_GP_SCRATCH1_BASE_IDX 1 +#define regUVD_GP_SCRATCH2 0x0120 +#define regUVD_GP_SCRATCH2_BASE_IDX 1 +#define regUVD_GP_SCRATCH3 0x0121 +#define regUVD_GP_SCRATCH3_BASE_IDX 1 +#define regUVD_GP_SCRATCH4 0x0122 +#define regUVD_GP_SCRATCH4_BASE_IDX 1 +#define regUVD_GP_SCRATCH5 0x0123 +#define regUVD_GP_SCRATCH5_BASE_IDX 1 +#define regUVD_GP_SCRATCH6 0x0124 +#define regUVD_GP_SCRATCH6_BASE_IDX 1 +#define regUVD_GP_SCRATCH7 0x0125 +#define regUVD_GP_SCRATCH7_BASE_IDX 1 +#define regUVD_GP_SCRATCH8 0x0126 +#define regUVD_GP_SCRATCH8_BASE_IDX 1 +#define regUVD_GP_SCRATCH9 0x0127 +#define regUVD_GP_SCRATCH9_BASE_IDX 1 +#define regUVD_GP_SCRATCH10 0x0128 +#define regUVD_GP_SCRATCH10_BASE_IDX 1 +#define regUVD_GP_SCRATCH11 0x0129 +#define regUVD_GP_SCRATCH11_BASE_IDX 1 +#define regUVD_GP_SCRATCH12 0x012a +#define regUVD_GP_SCRATCH12_BASE_IDX 1 +#define regUVD_GP_SCRATCH13 0x012b +#define regUVD_GP_SCRATCH13_BASE_IDX 1 +#define regUVD_GP_SCRATCH14 0x012c +#define regUVD_GP_SCRATCH14_BASE_IDX 1 +#define regUVD_GP_SCRATCH15 0x012d +#define regUVD_GP_SCRATCH15_BASE_IDX 1 +#define regUVD_GP_SCRATCH16 0x012e +#define regUVD_GP_SCRATCH16_BASE_IDX 1 +#define regUVD_GP_SCRATCH17 0x012f +#define regUVD_GP_SCRATCH17_BASE_IDX 1 +#define regUVD_GP_SCRATCH18 0x0130 +#define regUVD_GP_SCRATCH18_BASE_IDX 1 +#define regUVD_GP_SCRATCH19 0x0131 +#define regUVD_GP_SCRATCH19_BASE_IDX 1 +#define regUVD_GP_SCRATCH20 0x0132 +#define regUVD_GP_SCRATCH20_BASE_IDX 1 +#define regUVD_GP_SCRATCH21 0x0133 +#define regUVD_GP_SCRATCH21_BASE_IDX 1 +#define regUVD_GP_SCRATCH22 0x0134 +#define regUVD_GP_SCRATCH22_BASE_IDX 1 +#define regUVD_GP_SCRATCH23 0x0135 +#define regUVD_GP_SCRATCH23_BASE_IDX 1 +#define regUVD_AUDIO_RB_BASE_LO 0x0136 +#define regUVD_AUDIO_RB_BASE_LO_BASE_IDX 1 +#define regUVD_AUDIO_RB_BASE_HI 0x0137 +#define regUVD_AUDIO_RB_BASE_HI_BASE_IDX 1 +#define regUVD_AUDIO_RB_SIZE 0x0138 +#define regUVD_AUDIO_RB_SIZE_BASE_IDX 1 +#define regUVD_VCPU_INT_STATUS2 0x013b +#define regUVD_VCPU_INT_STATUS2_BASE_IDX 1 +#define regUVD_VCPU_INT_ACK2 0x013c +#define regUVD_VCPU_INT_ACK2_BASE_IDX 1 +#define regUVD_VCPU_INT_EN2 0x013d +#define regUVD_VCPU_INT_EN2_BASE_IDX 1 +#define regUVD_SUVD_CGC_STATUS2 0x013e +#define regUVD_SUVD_CGC_STATUS2_BASE_IDX 1 +#define regUVD_SUVD_INT_STATUS2 0x0140 +#define regUVD_SUVD_INT_STATUS2_BASE_IDX 1 +#define regUVD_SUVD_INT_EN2 0x0141 +#define regUVD_SUVD_INT_EN2_BASE_IDX 1 +#define regUVD_SUVD_INT_ACK2 0x0142 +#define regUVD_SUVD_INT_ACK2_BASE_IDX 1 +#define regUVD_STATUS 0x0143 +#define regUVD_STATUS_BASE_IDX 1 +#define regUVD_ENC_PIPE_BUSY 0x0144 +#define regUVD_ENC_PIPE_BUSY_BASE_IDX 1 +#define regUVD_FW_POWER_STATUS 0x0145 +#define regUVD_FW_POWER_STATUS_BASE_IDX 1 +#define regUVD_CNTL 0x0146 +#define regUVD_CNTL_BASE_IDX 1 +#define regUVD_SOFT_RESET 0x0147 +#define regUVD_SOFT_RESET_BASE_IDX 1 +#define regUVD_SOFT_RESET2 0x0148 +#define regUVD_SOFT_RESET2_BASE_IDX 1 +#define regUVD_MMSCH_SOFT_RESET 0x0149 +#define regUVD_MMSCH_SOFT_RESET_BASE_IDX 1 +#define regUVD_WIG_CTRL 0x014a +#define regUVD_WIG_CTRL_BASE_IDX 1 +#define regUVD_CGC_STATUS 0x014c +#define regUVD_CGC_STATUS_BASE_IDX 1 +#define regUVD_CGC_UDEC_STATUS 0x014e +#define regUVD_CGC_UDEC_STATUS_BASE_IDX 1 +#define regUVD_SUVD_CGC_STATUS 0x0150 +#define regUVD_SUVD_CGC_STATUS_BASE_IDX 1 +#define regUVD_GPCOM_VCPU_CMD 0x0152 +#define regUVD_GPCOM_VCPU_CMD_BASE_IDX 1 + + +// addressBlock: aid_uvd0_ecpudec +// base address: 0x1fe00 +#define regUVD_VCPU_CACHE_OFFSET0 0x0180 +#define regUVD_VCPU_CACHE_OFFSET0_BASE_IDX 1 +#define regUVD_VCPU_CACHE_SIZE0 0x0181 +#define regUVD_VCPU_CACHE_SIZE0_BASE_IDX 1 +#define regUVD_VCPU_CACHE_OFFSET1 0x0182 +#define regUVD_VCPU_CACHE_OFFSET1_BASE_IDX 1 +#define regUVD_VCPU_CACHE_SIZE1 0x0183 +#define regUVD_VCPU_CACHE_SIZE1_BASE_IDX 1 +#define regUVD_VCPU_CACHE_OFFSET2 0x0184 +#define regUVD_VCPU_CACHE_OFFSET2_BASE_IDX 1 +#define regUVD_VCPU_CACHE_SIZE2 0x0185 +#define regUVD_VCPU_CACHE_SIZE2_BASE_IDX 1 +#define regUVD_VCPU_CACHE_OFFSET3 0x0186 +#define regUVD_VCPU_CACHE_OFFSET3_BASE_IDX 1 +#define regUVD_VCPU_CACHE_SIZE3 0x0187 +#define regUVD_VCPU_CACHE_SIZE3_BASE_IDX 1 +#define regUVD_VCPU_CACHE_OFFSET4 0x0188 +#define regUVD_VCPU_CACHE_OFFSET4_BASE_IDX 1 +#define regUVD_VCPU_CACHE_SIZE4 0x0189 +#define regUVD_VCPU_CACHE_SIZE4_BASE_IDX 1 +#define regUVD_VCPU_CACHE_OFFSET5 0x018a +#define regUVD_VCPU_CACHE_OFFSET5_BASE_IDX 1 +#define regUVD_VCPU_CACHE_SIZE5 0x018b +#define regUVD_VCPU_CACHE_SIZE5_BASE_IDX 1 +#define regUVD_VCPU_CACHE_OFFSET6 0x018c +#define regUVD_VCPU_CACHE_OFFSET6_BASE_IDX 1 +#define regUVD_VCPU_CACHE_SIZE6 0x018d +#define regUVD_VCPU_CACHE_SIZE6_BASE_IDX 1 +#define regUVD_VCPU_CACHE_OFFSET7 0x018e +#define regUVD_VCPU_CACHE_OFFSET7_BASE_IDX 1 +#define regUVD_VCPU_CACHE_SIZE7 0x018f +#define regUVD_VCPU_CACHE_SIZE7_BASE_IDX 1 +#define regUVD_VCPU_CACHE_OFFSET8 0x0190 +#define regUVD_VCPU_CACHE_OFFSET8_BASE_IDX 1 +#define regUVD_VCPU_CACHE_SIZE8 0x0191 +#define regUVD_VCPU_CACHE_SIZE8_BASE_IDX 1 +#define regUVD_VCPU_NONCACHE_OFFSET0 0x0192 +#define regUVD_VCPU_NONCACHE_OFFSET0_BASE_IDX 1 +#define regUVD_VCPU_NONCACHE_SIZE0 0x0193 +#define regUVD_VCPU_NONCACHE_SIZE0_BASE_IDX 1 +#define regUVD_VCPU_NONCACHE_OFFSET1 0x0194 +#define regUVD_VCPU_NONCACHE_OFFSET1_BASE_IDX 1 +#define regUVD_VCPU_NONCACHE_SIZE1 0x0195 +#define regUVD_VCPU_NONCACHE_SIZE1_BASE_IDX 1 +#define regUVD_VCPU_CNTL 0x0196 +#define regUVD_VCPU_CNTL_BASE_IDX 1 +#define regUVD_VCPU_PRID 0x0197 +#define regUVD_VCPU_PRID_BASE_IDX 1 +#define regUVD_VCPU_TRCE 0x0198 +#define regUVD_VCPU_TRCE_BASE_IDX 1 +#define regUVD_VCPU_TRCE_RD 0x0199 +#define regUVD_VCPU_TRCE_RD_BASE_IDX 1 +#define regUVD_VCPU_IND_INDEX 0x019b +#define regUVD_VCPU_IND_INDEX_BASE_IDX 1 +#define regUVD_VCPU_IND_DATA 0x019c +#define regUVD_VCPU_IND_DATA_BASE_IDX 1 + + +// addressBlock: aid_uvd0_uvd_mpcdec +// base address: 0x1ff30 +#define regUVD_MP_SWAP_CNTL 0x01cc +#define regUVD_MP_SWAP_CNTL_BASE_IDX 1 +#define regUVD_MP_SWAP_CNTL2 0x01cd +#define regUVD_MP_SWAP_CNTL2_BASE_IDX 1 +#define regUVD_MPC_LUMA_SRCH 0x01ce +#define regUVD_MPC_LUMA_SRCH_BASE_IDX 1 +#define regUVD_MPC_LUMA_HIT 0x01cf +#define regUVD_MPC_LUMA_HIT_BASE_IDX 1 +#define regUVD_MPC_LUMA_HITPEND 0x01d0 +#define regUVD_MPC_LUMA_HITPEND_BASE_IDX 1 +#define regUVD_MPC_CHROMA_SRCH 0x01d1 +#define regUVD_MPC_CHROMA_SRCH_BASE_IDX 1 +#define regUVD_MPC_CHROMA_HIT 0x01d2 +#define regUVD_MPC_CHROMA_HIT_BASE_IDX 1 +#define regUVD_MPC_CHROMA_HITPEND 0x01d3 +#define regUVD_MPC_CHROMA_HITPEND_BASE_IDX 1 +#define regUVD_MPC_CNTL 0x01d4 +#define regUVD_MPC_CNTL_BASE_IDX 1 +#define regUVD_MPC_PITCH 0x01d5 +#define regUVD_MPC_PITCH_BASE_IDX 1 +#define regUVD_MPC_SET_MUXA0 0x01d6 +#define regUVD_MPC_SET_MUXA0_BASE_IDX 1 +#define regUVD_MPC_SET_MUXA1 0x01d7 +#define regUVD_MPC_SET_MUXA1_BASE_IDX 1 +#define regUVD_MPC_SET_MUXB0 0x01d8 +#define regUVD_MPC_SET_MUXB0_BASE_IDX 1 +#define regUVD_MPC_SET_MUXB1 0x01d9 +#define regUVD_MPC_SET_MUXB1_BASE_IDX 1 +#define regUVD_MPC_SET_MUX 0x01da +#define regUVD_MPC_SET_MUX_BASE_IDX 1 +#define regUVD_MPC_SET_ALU 0x01db +#define regUVD_MPC_SET_ALU_BASE_IDX 1 +#define regUVD_MPC_PERF0 0x01dc +#define regUVD_MPC_PERF0_BASE_IDX 1 +#define regUVD_MPC_PERF1 0x01dd +#define regUVD_MPC_PERF1_BASE_IDX 1 +#define regUVD_MPC_IND_INDEX 0x01de +#define regUVD_MPC_IND_INDEX_BASE_IDX 1 +#define regUVD_MPC_IND_DATA 0x01df +#define regUVD_MPC_IND_DATA_BASE_IDX 1 + + +// addressBlock: aid_uvd0_uvd_rbcdec +// base address: 0x1ff90 +#define regUVD_RBC_IB_SIZE 0x01e4 +#define regUVD_RBC_IB_SIZE_BASE_IDX 1 +#define regUVD_RBC_IB_SIZE_UPDATE 0x01e5 +#define regUVD_RBC_IB_SIZE_UPDATE_BASE_IDX 1 +#define regUVD_RBC_RB_CNTL 0x01e6 +#define regUVD_RBC_RB_CNTL_BASE_IDX 1 +#define regUVD_RBC_RB_RPTR_ADDR 0x01e7 +#define regUVD_RBC_RB_RPTR_ADDR_BASE_IDX 1 +#define regUVD_RBC_VCPU_ACCESS 0x01ea +#define regUVD_RBC_VCPU_ACCESS_BASE_IDX 1 +#define regUVD_FW_SEMAPHORE_CNTL 0x01eb +#define regUVD_FW_SEMAPHORE_CNTL_BASE_IDX 1 +#define regUVD_RBC_READ_REQ_URGENT_CNTL 0x01ed +#define regUVD_RBC_READ_REQ_URGENT_CNTL_BASE_IDX 1 +#define regUVD_RBC_RB_WPTR_CNTL 0x01ee +#define regUVD_RBC_RB_WPTR_CNTL_BASE_IDX 1 +#define regUVD_RBC_WPTR_STATUS 0x01ef +#define regUVD_RBC_WPTR_STATUS_BASE_IDX 1 +#define regUVD_RBC_WPTR_POLL_CNTL 0x01f0 +#define regUVD_RBC_WPTR_POLL_CNTL_BASE_IDX 1 +#define regUVD_RBC_WPTR_POLL_ADDR 0x01f1 +#define regUVD_RBC_WPTR_POLL_ADDR_BASE_IDX 1 +#define regUVD_SEMA_CMD 0x01f2 +#define regUVD_SEMA_CMD_BASE_IDX 1 +#define regUVD_SEMA_ADDR_LOW 0x01f3 +#define regUVD_SEMA_ADDR_LOW_BASE_IDX 1 +#define regUVD_SEMA_ADDR_HIGH 0x01f4 +#define regUVD_SEMA_ADDR_HIGH_BASE_IDX 1 +#define regUVD_ENGINE_CNTL 0x01f5 +#define regUVD_ENGINE_CNTL_BASE_IDX 1 +#define regUVD_SEMA_TIMEOUT_STATUS 0x01f6 +#define regUVD_SEMA_TIMEOUT_STATUS_BASE_IDX 1 +#define regUVD_SEMA_CNTL 0x01f7 +#define regUVD_SEMA_CNTL_BASE_IDX 1 +#define regUVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL 0x01f8 +#define regUVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL_BASE_IDX 1 +#define regUVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL 0x01f9 +#define regUVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL_BASE_IDX 1 +#define regUVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL 0x01fa +#define regUVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL_BASE_IDX 1 +#define regUVD_JOB_START 0x01fb +#define regUVD_JOB_START_BASE_IDX 1 +#define regUVD_RBC_BUF_STATUS 0x01fc +#define regUVD_RBC_BUF_STATUS_BASE_IDX 1 +#define regUVD_RBC_SWAP_CNTL 0x01fd +#define regUVD_RBC_SWAP_CNTL_BASE_IDX 1 + + +// addressBlock: aid_uvd0_lmi_adpdec +// base address: 0x20090 +#define regUVD_LMI_RE_64BIT_BAR_LOW 0x0224 +#define regUVD_LMI_RE_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_RE_64BIT_BAR_HIGH 0x0225 +#define regUVD_LMI_RE_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_IT_64BIT_BAR_LOW 0x0226 +#define regUVD_LMI_IT_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_IT_64BIT_BAR_HIGH 0x0227 +#define regUVD_LMI_IT_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MP_64BIT_BAR_LOW 0x0228 +#define regUVD_LMI_MP_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MP_64BIT_BAR_HIGH 0x0229 +#define regUVD_LMI_MP_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_CM_64BIT_BAR_LOW 0x022a +#define regUVD_LMI_CM_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_CM_64BIT_BAR_HIGH 0x022b +#define regUVD_LMI_CM_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_DB_64BIT_BAR_LOW 0x022c +#define regUVD_LMI_DB_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_DB_64BIT_BAR_HIGH 0x022d +#define regUVD_LMI_DB_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_DBW_64BIT_BAR_LOW 0x022e +#define regUVD_LMI_DBW_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_DBW_64BIT_BAR_HIGH 0x022f +#define regUVD_LMI_DBW_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_IDCT_64BIT_BAR_LOW 0x0230 +#define regUVD_LMI_IDCT_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_IDCT_64BIT_BAR_HIGH 0x0231 +#define regUVD_LMI_IDCT_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MPRD_S0_64BIT_BAR_LOW 0x0232 +#define regUVD_LMI_MPRD_S0_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MPRD_S0_64BIT_BAR_HIGH 0x0233 +#define regUVD_LMI_MPRD_S0_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MPRD_S1_64BIT_BAR_LOW 0x0234 +#define regUVD_LMI_MPRD_S1_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MPRD_S1_64BIT_BAR_HIGH 0x0235 +#define regUVD_LMI_MPRD_S1_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MPRD_DBW_64BIT_BAR_LOW 0x0236 +#define regUVD_LMI_MPRD_DBW_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MPRD_DBW_64BIT_BAR_HIGH 0x0237 +#define regUVD_LMI_MPRD_DBW_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MPC_64BIT_BAR_LOW 0x0238 +#define regUVD_LMI_MPC_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MPC_64BIT_BAR_HIGH 0x0239 +#define regUVD_LMI_MPC_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_RBC_RB_64BIT_BAR_LOW 0x023a +#define regUVD_LMI_RBC_RB_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_RBC_RB_64BIT_BAR_HIGH 0x023b +#define regUVD_LMI_RBC_RB_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_RBC_IB_64BIT_BAR_LOW 0x023c +#define regUVD_LMI_RBC_IB_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_RBC_IB_64BIT_BAR_HIGH 0x023d +#define regUVD_LMI_RBC_IB_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_LBSI_64BIT_BAR_LOW 0x023e +#define regUVD_LMI_LBSI_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_LBSI_64BIT_BAR_HIGH 0x023f +#define regUVD_LMI_LBSI_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW 0x0240 +#define regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH 0x0241 +#define regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_VCPU_NC1_64BIT_BAR_LOW 0x0242 +#define regUVD_LMI_VCPU_NC1_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_VCPU_NC1_64BIT_BAR_HIGH 0x0243 +#define regUVD_LMI_VCPU_NC1_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW 0x0244 +#define regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH 0x0245 +#define regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_CENC_64BIT_BAR_LOW 0x0246 +#define regUVD_LMI_CENC_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_CENC_64BIT_BAR_HIGH 0x0247 +#define regUVD_LMI_CENC_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_SRE_64BIT_BAR_LOW 0x0248 +#define regUVD_LMI_SRE_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_SRE_64BIT_BAR_HIGH 0x0249 +#define regUVD_LMI_SRE_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_GPGPU_64BIT_BAR_LOW 0x024a +#define regUVD_LMI_MIF_GPGPU_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_GPGPU_64BIT_BAR_HIGH 0x024b +#define regUVD_LMI_MIF_GPGPU_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_CURR_LUMA_64BIT_BAR_LOW 0x024c +#define regUVD_LMI_MIF_CURR_LUMA_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_CURR_LUMA_64BIT_BAR_HIGH 0x024d +#define regUVD_LMI_MIF_CURR_LUMA_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_CURR_CHROMA_64BIT_BAR_LOW 0x024e +#define regUVD_LMI_MIF_CURR_CHROMA_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_CURR_CHROMA_64BIT_BAR_HIGH 0x024f +#define regUVD_LMI_MIF_CURR_CHROMA_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_REF_64BIT_BAR_LOW 0x0250 +#define regUVD_LMI_MIF_REF_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_REF_64BIT_BAR_HIGH 0x0251 +#define regUVD_LMI_MIF_REF_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_DBW_64BIT_BAR_LOW 0x0252 +#define regUVD_LMI_MIF_DBW_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_DBW_64BIT_BAR_HIGH 0x0253 +#define regUVD_LMI_MIF_DBW_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_CM_COLOC_64BIT_BAR_LOW 0x0254 +#define regUVD_LMI_MIF_CM_COLOC_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_CM_COLOC_64BIT_BAR_HIGH 0x0255 +#define regUVD_LMI_MIF_CM_COLOC_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_BSP0_64BIT_BAR_LOW 0x0256 +#define regUVD_LMI_MIF_BSP0_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_BSP0_64BIT_BAR_HIGH 0x0257 +#define regUVD_LMI_MIF_BSP0_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_BSP1_64BIT_BAR_LOW 0x0258 +#define regUVD_LMI_MIF_BSP1_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_BSP1_64BIT_BAR_HIGH 0x0259 +#define regUVD_LMI_MIF_BSP1_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_BSP2_64BIT_BAR_LOW 0x025a +#define regUVD_LMI_MIF_BSP2_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_BSP2_64BIT_BAR_HIGH 0x025b +#define regUVD_LMI_MIF_BSP2_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_BSP3_64BIT_BAR_LOW 0x025c +#define regUVD_LMI_MIF_BSP3_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_BSP3_64BIT_BAR_HIGH 0x025d +#define regUVD_LMI_MIF_BSP3_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_BSD0_64BIT_BAR_LOW 0x025e +#define regUVD_LMI_MIF_BSD0_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_BSD0_64BIT_BAR_HIGH 0x025f +#define regUVD_LMI_MIF_BSD0_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_BSD1_64BIT_BAR_LOW 0x0260 +#define regUVD_LMI_MIF_BSD1_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_BSD1_64BIT_BAR_HIGH 0x0261 +#define regUVD_LMI_MIF_BSD1_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_BSD2_64BIT_BAR_LOW 0x0262 +#define regUVD_LMI_MIF_BSD2_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_BSD2_64BIT_BAR_HIGH 0x0263 +#define regUVD_LMI_MIF_BSD2_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_BSD3_64BIT_BAR_LOW 0x0264 +#define regUVD_LMI_MIF_BSD3_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_BSD3_64BIT_BAR_HIGH 0x0265 +#define regUVD_LMI_MIF_BSD3_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_BSD4_64BIT_BAR_LOW 0x0266 +#define regUVD_LMI_MIF_BSD4_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_BSD4_64BIT_BAR_HIGH 0x0267 +#define regUVD_LMI_MIF_BSD4_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW 0x0270 +#define regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH 0x0271 +#define regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE8_64BIT_BAR_LOW 0x0272 +#define regUVD_LMI_VCPU_CACHE8_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE8_64BIT_BAR_HIGH 0x0273 +#define regUVD_LMI_VCPU_CACHE8_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW 0x0274 +#define regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH 0x0275 +#define regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE3_64BIT_BAR_LOW 0x0276 +#define regUVD_LMI_VCPU_CACHE3_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE3_64BIT_BAR_HIGH 0x0277 +#define regUVD_LMI_VCPU_CACHE3_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE4_64BIT_BAR_LOW 0x0278 +#define regUVD_LMI_VCPU_CACHE4_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE4_64BIT_BAR_HIGH 0x0279 +#define regUVD_LMI_VCPU_CACHE4_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE5_64BIT_BAR_LOW 0x027a +#define regUVD_LMI_VCPU_CACHE5_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE5_64BIT_BAR_HIGH 0x027b +#define regUVD_LMI_VCPU_CACHE5_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE6_64BIT_BAR_LOW 0x027c +#define regUVD_LMI_VCPU_CACHE6_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE6_64BIT_BAR_HIGH 0x027d +#define regUVD_LMI_VCPU_CACHE6_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE7_64BIT_BAR_LOW 0x027e +#define regUVD_LMI_VCPU_CACHE7_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE7_64BIT_BAR_HIGH 0x027f +#define regUVD_LMI_VCPU_CACHE7_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_SCLR_64BIT_BAR_LOW 0x0280 +#define regUVD_LMI_MIF_SCLR_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_SCLR_64BIT_BAR_HIGH 0x0281 +#define regUVD_LMI_MIF_SCLR_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_SCLR2_64BIT_BAR_LOW 0x0282 +#define regUVD_LMI_MIF_SCLR2_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_SCLR2_64BIT_BAR_HIGH 0x0283 +#define regUVD_LMI_MIF_SCLR2_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_SPH_64BIT_BAR_HIGH 0x0284 +#define regUVD_LMI_SPH_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_IMAGEPASTE_LUMA_64BIT_BAR_LOW 0x0298 +#define regUVD_LMI_MIF_IMAGEPASTE_LUMA_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_IMAGEPASTE_LUMA_64BIT_BAR_HIGH 0x0299 +#define regUVD_LMI_MIF_IMAGEPASTE_LUMA_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_IMAGEPASTE_CHROMA_64BIT_BAR_LOW 0x029a +#define regUVD_LMI_MIF_IMAGEPASTE_CHROMA_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_IMAGEPASTE_CHROMA_64BIT_BAR_HIGH 0x029b +#define regUVD_LMI_MIF_IMAGEPASTE_CHROMA_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_PRIVACY_LUMA_64BIT_BAR_LOW 0x029c +#define regUVD_LMI_MIF_PRIVACY_LUMA_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_PRIVACY_LUMA_64BIT_BAR_HIGH 0x029d +#define regUVD_LMI_MIF_PRIVACY_LUMA_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MIF_PRIVACY_CHROMA_64BIT_BAR_LOW 0x029e +#define regUVD_LMI_MIF_PRIVACY_CHROMA_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_PRIVACY_CHROMA_64BIT_BAR_HIGH 0x029f +#define regUVD_LMI_MIF_PRIVACY_CHROMA_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_ADP_ATOMIC_CONFIG 0x02a1 +#define regUVD_ADP_ATOMIC_CONFIG_BASE_IDX 1 +#define regUVD_LMI_ARB_CTRL2 0x02a2 +#define regUVD_LMI_ARB_CTRL2_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE_VMIDS_MULTI 0x02a7 +#define regUVD_LMI_VCPU_CACHE_VMIDS_MULTI_BASE_IDX 1 +#define regUVD_LMI_VCPU_NC_VMIDS_MULTI 0x02a8 +#define regUVD_LMI_VCPU_NC_VMIDS_MULTI_BASE_IDX 1 +#define regUVD_LMI_LAT_CTRL 0x02a9 +#define regUVD_LMI_LAT_CTRL_BASE_IDX 1 +#define regUVD_LMI_LAT_CNTR 0x02aa +#define regUVD_LMI_LAT_CNTR_BASE_IDX 1 +#define regUVD_LMI_AVG_LAT_CNTR 0x02ab +#define regUVD_LMI_AVG_LAT_CNTR_BASE_IDX 1 +#define regUVD_LMI_SPH 0x02ac +#define regUVD_LMI_SPH_BASE_IDX 1 +#define regUVD_LMI_VCPU_CACHE_VMID 0x02ad +#define regUVD_LMI_VCPU_CACHE_VMID_BASE_IDX 1 +#define regUVD_LMI_CTRL2 0x02ae +#define regUVD_LMI_CTRL2_BASE_IDX 1 +#define regUVD_LMI_URGENT_CTRL 0x02af +#define regUVD_LMI_URGENT_CTRL_BASE_IDX 1 +#define regUVD_LMI_CTRL 0x02b0 +#define regUVD_LMI_CTRL_BASE_IDX 1 +#define regUVD_LMI_STATUS 0x02b1 +#define regUVD_LMI_STATUS_BASE_IDX 1 +#define regUVD_LMI_PERFMON_CTRL 0x02b4 +#define regUVD_LMI_PERFMON_CTRL_BASE_IDX 1 +#define regUVD_LMI_PERFMON_COUNT_LO 0x02b5 +#define regUVD_LMI_PERFMON_COUNT_LO_BASE_IDX 1 +#define regUVD_LMI_PERFMON_COUNT_HI 0x02b6 +#define regUVD_LMI_PERFMON_COUNT_HI_BASE_IDX 1 +#define regUVD_LMI_ADP_SWAP_CNTL 0x02b7 +#define regUVD_LMI_ADP_SWAP_CNTL_BASE_IDX 1 +#define regUVD_LMI_RBC_RB_VMID 0x02b8 +#define regUVD_LMI_RBC_RB_VMID_BASE_IDX 1 +#define regUVD_LMI_RBC_IB_VMID 0x02b9 +#define regUVD_LMI_RBC_IB_VMID_BASE_IDX 1 +#define regUVD_LMI_MC_CREDITS 0x02ba +#define regUVD_LMI_MC_CREDITS_BASE_IDX 1 +#define regUVD_LMI_ADP_IND_INDEX 0x02be +#define regUVD_LMI_ADP_IND_INDEX_BASE_IDX 1 +#define regUVD_LMI_ADP_IND_DATA 0x02bf +#define regUVD_LMI_ADP_IND_DATA_BASE_IDX 1 +#define regUVD_LMI_ADP_PF_EN 0x02c0 +#define regUVD_LMI_ADP_PF_EN_BASE_IDX 1 +#define regUVD_LMI_PREF_CTRL 0x02c2 +#define regUVD_LMI_PREF_CTRL_BASE_IDX 1 +#define regUVD_LMI_MIF_REF_LUMA_64BIT_BAR_LOW 0x02dd +#define regUVD_LMI_MIF_REF_LUMA_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MIF_REF_LUMA_64BIT_BAR_HIGH 0x02de +#define regUVD_LMI_MIF_REF_LUMA_64BIT_BAR_HIGH_BASE_IDX 1 +#define regVCN_RAS_CNTL 0x02df +#define regVCN_RAS_CNTL_BASE_IDX 1 + + +// addressBlock: aid_uvd0_uvd_jpeg0_jpegnpdec +// base address: 0x20f00 +#define regUVD_JPEG_CNTL 0x05c0 +#define regUVD_JPEG_CNTL_BASE_IDX 1 +#define regUVD_JPEG_RB_BASE 0x05c1 +#define regUVD_JPEG_RB_BASE_BASE_IDX 1 +#define regUVD_JPEG_RB_WPTR 0x05c2 +#define regUVD_JPEG_RB_WPTR_BASE_IDX 1 +#define regUVD_JPEG_RB_RPTR 0x05c3 +#define regUVD_JPEG_RB_RPTR_BASE_IDX 1 +#define regUVD_JPEG_RB_SIZE 0x05c4 +#define regUVD_JPEG_RB_SIZE_BASE_IDX 1 +#define regUVD_JPEG_DEC_CNT 0x05c5 +#define regUVD_JPEG_DEC_CNT_BASE_IDX 1 +#define regUVD_JPEG_SPS_INFO 0x05c6 +#define regUVD_JPEG_SPS_INFO_BASE_IDX 1 +#define regUVD_JPEG_SPS1_INFO 0x05c7 +#define regUVD_JPEG_SPS1_INFO_BASE_IDX 1 +#define regUVD_JPEG_RE_TIMER 0x05c8 +#define regUVD_JPEG_RE_TIMER_BASE_IDX 1 +#define regUVD_JPEG_DEC_SCRATCH0 0x05c9 +#define regUVD_JPEG_DEC_SCRATCH0_BASE_IDX 1 +#define regUVD_JPEG_INT_EN 0x05ca +#define regUVD_JPEG_INT_EN_BASE_IDX 1 +#define regUVD_JPEG_INT_STAT 0x05cb +#define regUVD_JPEG_INT_STAT_BASE_IDX 1 +#define regUVD_JPEG_TIER_CNTL0 0x05cc +#define regUVD_JPEG_TIER_CNTL0_BASE_IDX 1 +#define regUVD_JPEG_TIER_CNTL1 0x05cd +#define regUVD_JPEG_TIER_CNTL1_BASE_IDX 1 +#define regUVD_JPEG_TIER_CNTL2 0x05ce +#define regUVD_JPEG_TIER_CNTL2_BASE_IDX 1 +#define regUVD_JPEG_TIER_STATUS 0x05cf +#define regUVD_JPEG_TIER_STATUS_BASE_IDX 1 + + +// addressBlock: aid_uvd0_uvd_jpeg_sclk0_jpegnpsclkdec +// base address: 0x21000 +#define regUVD_JPEG_OUTBUF_CNTL 0x0600 +#define regUVD_JPEG_OUTBUF_CNTL_BASE_IDX 1 +#define regUVD_JPEG_OUTBUF_WPTR 0x0601 +#define regUVD_JPEG_OUTBUF_WPTR_BASE_IDX 1 +#define regUVD_JPEG_OUTBUF_RPTR 0x0602 +#define regUVD_JPEG_OUTBUF_RPTR_BASE_IDX 1 +#define regUVD_JPEG_PITCH 0x0603 +#define regUVD_JPEG_PITCH_BASE_IDX 1 +#define regUVD_JPEG_UV_PITCH 0x0604 +#define regUVD_JPEG_UV_PITCH_BASE_IDX 1 +#define regJPEG_DEC_Y_GFX8_TILING_SURFACE 0x0605 +#define regJPEG_DEC_Y_GFX8_TILING_SURFACE_BASE_IDX 1 +#define regJPEG_DEC_UV_GFX8_TILING_SURFACE 0x0606 +#define regJPEG_DEC_UV_GFX8_TILING_SURFACE_BASE_IDX 1 +#define regJPEG_DEC_GFX8_ADDR_CONFIG 0x0607 +#define regJPEG_DEC_GFX8_ADDR_CONFIG_BASE_IDX 1 +#define regJPEG_DEC_Y_GFX10_TILING_SURFACE 0x0608 +#define regJPEG_DEC_Y_GFX10_TILING_SURFACE_BASE_IDX 1 +#define regJPEG_DEC_UV_GFX10_TILING_SURFACE 0x0609 +#define regJPEG_DEC_UV_GFX10_TILING_SURFACE_BASE_IDX 1 +#define regJPEG_DEC_GFX10_ADDR_CONFIG 0x060a +#define regJPEG_DEC_GFX10_ADDR_CONFIG_BASE_IDX 1 +#define regJPEG_DEC_ADDR_MODE 0x060b +#define regJPEG_DEC_ADDR_MODE_BASE_IDX 1 +#define regUVD_JPEG_OUTPUT_XY 0x060c +#define regUVD_JPEG_OUTPUT_XY_BASE_IDX 1 +#define regUVD_JPEG_GPCOM_CMD 0x060d +#define regUVD_JPEG_GPCOM_CMD_BASE_IDX 1 +#define regUVD_JPEG_GPCOM_DATA0 0x060e +#define regUVD_JPEG_GPCOM_DATA0_BASE_IDX 1 +#define regUVD_JPEG_GPCOM_DATA1 0x060f +#define regUVD_JPEG_GPCOM_DATA1_BASE_IDX 1 +#define regUVD_JPEG_SCRATCH1 0x0610 +#define regUVD_JPEG_SCRATCH1_BASE_IDX 1 +#define regUVD_JPEG_DEC_SOFT_RST 0x0611 +#define regUVD_JPEG_DEC_SOFT_RST_BASE_IDX 1 + + +// addressBlock: aid_uvd0_uvd_jrbc0_uvd_jrbc_dec +// base address: 0x21100 +#define regUVD_JRBC0_UVD_JRBC_RB_WPTR 0x0640 +#define regUVD_JRBC0_UVD_JRBC_RB_WPTR_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_RB_CNTL 0x0641 +#define regUVD_JRBC0_UVD_JRBC_RB_CNTL_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_IB_SIZE 0x0642 +#define regUVD_JRBC0_UVD_JRBC_IB_SIZE_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_URGENT_CNTL 0x0643 +#define regUVD_JRBC0_UVD_JRBC_URGENT_CNTL_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_RB_REF_DATA 0x0644 +#define regUVD_JRBC0_UVD_JRBC_RB_REF_DATA_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_RB_COND_RD_TIMER 0x0645 +#define regUVD_JRBC0_UVD_JRBC_RB_COND_RD_TIMER_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_SOFT_RESET 0x0648 +#define regUVD_JRBC0_UVD_JRBC_SOFT_RESET_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_STATUS 0x0649 +#define regUVD_JRBC0_UVD_JRBC_STATUS_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_RB_RPTR 0x064a +#define regUVD_JRBC0_UVD_JRBC_RB_RPTR_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_RB_BUF_STATUS 0x064b +#define regUVD_JRBC0_UVD_JRBC_RB_BUF_STATUS_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_IB_BUF_STATUS 0x064c +#define regUVD_JRBC0_UVD_JRBC_IB_BUF_STATUS_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_IB_SIZE_UPDATE 0x064d +#define regUVD_JRBC0_UVD_JRBC_IB_SIZE_UPDATE_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_IB_COND_RD_TIMER 0x064e +#define regUVD_JRBC0_UVD_JRBC_IB_COND_RD_TIMER_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_IB_REF_DATA 0x064f +#define regUVD_JRBC0_UVD_JRBC_IB_REF_DATA_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JPEG_PREEMPT_CMD 0x0650 +#define regUVD_JRBC0_UVD_JPEG_PREEMPT_CMD_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JPEG_PREEMPT_FENCE_DATA0 0x0651 +#define regUVD_JRBC0_UVD_JPEG_PREEMPT_FENCE_DATA0_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JPEG_PREEMPT_FENCE_DATA1 0x0652 +#define regUVD_JRBC0_UVD_JPEG_PREEMPT_FENCE_DATA1_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_RB_SIZE 0x0653 +#define regUVD_JRBC0_UVD_JRBC_RB_SIZE_BASE_IDX 1 +#define regUVD_JRBC0_UVD_JRBC_SCRATCH0 0x0654 +#define regUVD_JRBC0_UVD_JRBC_SCRATCH0_BASE_IDX 1 + + +// addressBlock: aid_uvd0_uvd_jmi0_uvd_jmi_dec +// base address: 0x21180 +#define regUVD_JMI0_UVD_JPEG_DEC_PF_CTRL 0x0660 +#define regUVD_JMI0_UVD_JPEG_DEC_PF_CTRL_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JRBC_CTRL 0x0661 +#define regUVD_JMI0_UVD_LMI_JRBC_CTRL_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JPEG_CTRL 0x0662 +#define regUVD_JMI0_UVD_LMI_JPEG_CTRL_BASE_IDX 1 +#define regUVD_JMI0_JPEG_LMI_DROP 0x0663 +#define regUVD_JMI0_JPEG_LMI_DROP_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JRBC_IB_VMID 0x0664 +#define regUVD_JMI0_UVD_LMI_JRBC_IB_VMID_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JRBC_RB_VMID 0x0665 +#define regUVD_JMI0_UVD_LMI_JRBC_RB_VMID_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JPEG_VMID 0x0666 +#define regUVD_JMI0_UVD_LMI_JPEG_VMID_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW 0x0667 +#define regUVD_JMI0_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH 0x0668 +#define regUVD_JMI0_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_LOW 0x0669 +#define regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH 0x066a +#define regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW 0x066b +#define regUVD_JMI0_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH 0x066c +#define regUVD_JMI0_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JPEG_PREEMPT_VMID 0x066d +#define regUVD_JMI0_UVD_LMI_JPEG_PREEMPT_VMID_BASE_IDX 1 +#define regUVD_JMI0_UVD_JMI_DEC_SWAP_CNTL 0x066e +#define regUVD_JMI0_UVD_JMI_DEC_SWAP_CNTL_BASE_IDX 1 +#define regUVD_JMI0_UVD_JMI_ATOMIC_CNTL 0x066f +#define regUVD_JMI0_UVD_JMI_ATOMIC_CNTL_BASE_IDX 1 +#define regUVD_JMI0_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW 0x0670 +#define regUVD_JMI0_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_JMI0_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH 0x0671 +#define regUVD_JMI0_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JPEG_READ_64BIT_BAR_LOW 0x0672 +#define regUVD_JMI0_UVD_LMI_JPEG_READ_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH 0x0673 +#define regUVD_JMI0_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW 0x0674 +#define regUVD_JMI0_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH 0x0675 +#define regUVD_JMI0_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JRBC_IB_64BIT_BAR_LOW 0x0676 +#define regUVD_JMI0_UVD_LMI_JRBC_IB_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH 0x0677 +#define regUVD_JMI0_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW 0x0678 +#define regUVD_JMI0_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_JMI0_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH 0x0679 +#define regUVD_JMI0_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_JMI0_UVD_JMI_ATOMIC_CNTL2 0x067d +#define regUVD_JMI0_UVD_JMI_ATOMIC_CNTL2_BASE_IDX 1 + + +// addressBlock: aid_uvd0_uvd_jmi_common_dec +// base address: 0x21300 +#define regUVD_JADP_MCIF_URGENT_CTRL 0x06c1 +#define regUVD_JADP_MCIF_URGENT_CTRL_BASE_IDX 1 +#define regUVD_JMI_URGENT_CTRL 0x06c2 +#define regUVD_JMI_URGENT_CTRL_BASE_IDX 1 +#define regUVD_JMI_CTRL 0x06c3 +#define regUVD_JMI_CTRL_BASE_IDX 1 +#define regJPEG_MEMCHECK_CLAMPING_CNTL 0x06c4 +#define regJPEG_MEMCHECK_CLAMPING_CNTL_BASE_IDX 1 +#define regJPEG_MEMCHECK_SAFE_ADDR 0x06c5 +#define regJPEG_MEMCHECK_SAFE_ADDR_BASE_IDX 1 +#define regJPEG_MEMCHECK_SAFE_ADDR_64BIT 0x06c6 +#define regJPEG_MEMCHECK_SAFE_ADDR_64BIT_BASE_IDX 1 +#define regUVD_JMI_LAT_CTRL 0x06c7 +#define regUVD_JMI_LAT_CTRL_BASE_IDX 1 +#define regUVD_JMI_LAT_CNTR 0x06c8 +#define regUVD_JMI_LAT_CNTR_BASE_IDX 1 +#define regUVD_JMI_AVG_LAT_CNTR 0x06c9 +#define regUVD_JMI_AVG_LAT_CNTR_BASE_IDX 1 +#define regUVD_JMI_PERFMON_CTRL 0x06ca +#define regUVD_JMI_PERFMON_CTRL_BASE_IDX 1 +#define regUVD_JMI_PERFMON_COUNT_LO 0x06cb +#define regUVD_JMI_PERFMON_COUNT_LO_BASE_IDX 1 +#define regUVD_JMI_PERFMON_COUNT_HI 0x06cc +#define regUVD_JMI_PERFMON_COUNT_HI_BASE_IDX 1 +#define regUVD_JMI_CLEAN_STATUS 0x06cd +#define regUVD_JMI_CLEAN_STATUS_BASE_IDX 1 +#define regUVD_JMI_CNTL 0x06ce +#define regUVD_JMI_CNTL_BASE_IDX 1 + + +// addressBlock: aid_uvd0_uvd_jpeg_common_dec +// base address: 0x21400 +#define regJPEG_SOFT_RESET_STATUS 0x0700 +#define regJPEG_SOFT_RESET_STATUS_BASE_IDX 1 +#define regJPEG_SYS_INT_EN 0x0701 +#define regJPEG_SYS_INT_EN_BASE_IDX 1 +#define regJPEG_SYS_INT_EN1 0x0702 +#define regJPEG_SYS_INT_EN1_BASE_IDX 1 +#define regJPEG_SYS_INT_STATUS 0x0703 +#define regJPEG_SYS_INT_STATUS_BASE_IDX 1 +#define regJPEG_SYS_INT_STATUS1 0x0704 +#define regJPEG_SYS_INT_STATUS1_BASE_IDX 1 +#define regJPEG_SYS_INT_ACK 0x0705 +#define regJPEG_SYS_INT_ACK_BASE_IDX 1 +#define regJPEG_SYS_INT_ACK1 0x0706 +#define regJPEG_SYS_INT_ACK1_BASE_IDX 1 +#define regJPEG_MEMCHECK_SYS_INT_EN 0x0707 +#define regJPEG_MEMCHECK_SYS_INT_EN_BASE_IDX 1 +#define regJPEG_MEMCHECK_SYS_INT_EN1 0x0708 +#define regJPEG_MEMCHECK_SYS_INT_EN1_BASE_IDX 1 +#define regJPEG_MEMCHECK_SYS_INT_STAT 0x0709 +#define regJPEG_MEMCHECK_SYS_INT_STAT_BASE_IDX 1 +#define regJPEG_MEMCHECK_SYS_INT_STAT1 0x070a +#define regJPEG_MEMCHECK_SYS_INT_STAT1_BASE_IDX 1 +#define regJPEG_MEMCHECK_SYS_INT_STAT2 0x070b +#define regJPEG_MEMCHECK_SYS_INT_STAT2_BASE_IDX 1 +#define regJPEG_MEMCHECK_SYS_INT_ACK 0x070c +#define regJPEG_MEMCHECK_SYS_INT_ACK_BASE_IDX 1 +#define regJPEG_MEMCHECK_SYS_INT_ACK1 0x070d +#define regJPEG_MEMCHECK_SYS_INT_ACK1_BASE_IDX 1 +#define regJPEG_MEMCHECK_SYS_INT_ACK2 0x070e +#define regJPEG_MEMCHECK_SYS_INT_ACK2_BASE_IDX 1 +#define regJPEG_MASTINT_EN 0x0710 +#define regJPEG_MASTINT_EN_BASE_IDX 1 +#define regJPEG_IH_CTRL 0x0711 +#define regJPEG_IH_CTRL_BASE_IDX 1 +#define regJRBBM_ARB_CTRL 0x0713 +#define regJRBBM_ARB_CTRL_BASE_IDX 1 + + +// addressBlock: aid_uvd0_uvd_jpeg_common_sclk_dec +// base address: 0x21480 +#define regJPEG_CGC_GATE 0x0720 +#define regJPEG_CGC_GATE_BASE_IDX 1 +#define regJPEG_CGC_CTRL 0x0721 +#define regJPEG_CGC_CTRL_BASE_IDX 1 +#define regJPEG_CGC_STATUS 0x0722 +#define regJPEG_CGC_STATUS_BASE_IDX 1 +#define regJPEG_COMN_CGC_MEM_CTRL 0x0723 +#define regJPEG_COMN_CGC_MEM_CTRL_BASE_IDX 1 +#define regJPEG_DEC_CGC_MEM_CTRL 0x0724 +#define regJPEG_DEC_CGC_MEM_CTRL_BASE_IDX 1 +#define regJPEG_ENC_CGC_MEM_CTRL 0x0726 +#define regJPEG_ENC_CGC_MEM_CTRL_BASE_IDX 1 +#define regJPEG_PERF_BANK_CONF 0x0727 +#define regJPEG_PERF_BANK_CONF_BASE_IDX 1 +#define regJPEG_PERF_BANK_EVENT_SEL 0x0728 +#define regJPEG_PERF_BANK_EVENT_SEL_BASE_IDX 1 +#define regJPEG_PERF_BANK_COUNT0 0x0729 +#define regJPEG_PERF_BANK_COUNT0_BASE_IDX 1 +#define regJPEG_PERF_BANK_COUNT1 0x072a +#define regJPEG_PERF_BANK_COUNT1_BASE_IDX 1 +#define regJPEG_PERF_BANK_COUNT2 0x072b +#define regJPEG_PERF_BANK_COUNT2_BASE_IDX 1 +#define regJPEG_PERF_BANK_COUNT3 0x072c +#define regJPEG_PERF_BANK_COUNT3_BASE_IDX 1 + + +// addressBlock: aid_uvd0_uvd_pg_dec +// base address: 0x1f800 +#define regUVD_PGFSM_CONFIG 0x0000 +#define regUVD_PGFSM_CONFIG_BASE_IDX 1 +#define regUVD_PGFSM_STATUS 0x0001 +#define regUVD_PGFSM_STATUS_BASE_IDX 1 +#define regUVD_POWER_STATUS 0x0002 +#define regUVD_POWER_STATUS_BASE_IDX 1 +#define regUVD_JPEG_POWER_STATUS 0x0003 +#define regUVD_JPEG_POWER_STATUS_BASE_IDX 1 +#define regUVD_MC_DJPEG_RD_SPACE 0x0006 +#define regUVD_MC_DJPEG_RD_SPACE_BASE_IDX 1 +#define regUVD_MC_DJPEG_WR_SPACE 0x0007 +#define regUVD_MC_DJPEG_WR_SPACE_BASE_IDX 1 +#define regUVD_MC_EJPEG_RD_SPACE 0x0008 +#define regUVD_MC_EJPEG_RD_SPACE_BASE_IDX 1 +#define regUVD_MC_EJPEG_WR_SPACE 0x0009 +#define regUVD_MC_EJPEG_WR_SPACE_BASE_IDX 1 +#define regUVD_PG_IND_INDEX 0x000c +#define regUVD_PG_IND_INDEX_BASE_IDX 1 +#define regUVD_PG_IND_DATA 0x000e +#define regUVD_PG_IND_DATA_BASE_IDX 1 +#define regCC_UVD_HARVESTING 0x000f +#define regCC_UVD_HARVESTING_BASE_IDX 1 +#define regUVD_DPG_LMA_CTL 0x0011 +#define regUVD_DPG_LMA_CTL_BASE_IDX 1 +#define regUVD_DPG_LMA_DATA 0x0012 +#define regUVD_DPG_LMA_DATA_BASE_IDX 1 +#define regUVD_DPG_LMA_MASK 0x0013 +#define regUVD_DPG_LMA_MASK_BASE_IDX 1 +#define regUVD_DPG_PAUSE 0x0014 +#define regUVD_DPG_PAUSE_BASE_IDX 1 +#define regUVD_SCRATCH1 0x0015 +#define regUVD_SCRATCH1_BASE_IDX 1 +#define regUVD_SCRATCH2 0x0016 +#define regUVD_SCRATCH2_BASE_IDX 1 +#define regUVD_SCRATCH3 0x0017 +#define regUVD_SCRATCH3_BASE_IDX 1 +#define regUVD_SCRATCH4 0x0018 +#define regUVD_SCRATCH4_BASE_IDX 1 +#define regUVD_SCRATCH5 0x0019 +#define regUVD_SCRATCH5_BASE_IDX 1 +#define regUVD_SCRATCH6 0x001a +#define regUVD_SCRATCH6_BASE_IDX 1 +#define regUVD_SCRATCH7 0x001b +#define regUVD_SCRATCH7_BASE_IDX 1 +#define regUVD_SCRATCH8 0x001c +#define regUVD_SCRATCH8_BASE_IDX 1 +#define regUVD_SCRATCH9 0x001d +#define regUVD_SCRATCH9_BASE_IDX 1 +#define regUVD_SCRATCH10 0x001e +#define regUVD_SCRATCH10_BASE_IDX 1 +#define regUVD_SCRATCH11 0x001f +#define regUVD_SCRATCH11_BASE_IDX 1 +#define regUVD_SCRATCH12 0x0020 +#define regUVD_SCRATCH12_BASE_IDX 1 +#define regUVD_SCRATCH13 0x0021 +#define regUVD_SCRATCH13_BASE_IDX 1 +#define regUVD_SCRATCH14 0x0022 +#define regUVD_SCRATCH14_BASE_IDX 1 +#define regUVD_FREE_COUNTER_REG 0x0023 +#define regUVD_FREE_COUNTER_REG_BASE_IDX 1 +#define regUVD_DPG_LMI_VCPU_CACHE_64BIT_BAR_LOW 0x0024 +#define regUVD_DPG_LMI_VCPU_CACHE_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_DPG_LMI_VCPU_CACHE_64BIT_BAR_HIGH 0x0025 +#define regUVD_DPG_LMI_VCPU_CACHE_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_DPG_VCPU_CACHE_OFFSET0 0x0026 +#define regUVD_DPG_VCPU_CACHE_OFFSET0_BASE_IDX 1 +#define regUVD_DPG_LMI_VCPU_CACHE_VMID 0x0027 +#define regUVD_DPG_LMI_VCPU_CACHE_VMID_BASE_IDX 1 +#define regUVD_REG_FILTER_EN 0x0028 +#define regUVD_REG_FILTER_EN_BASE_IDX 1 +#define regUVD_SECURITY_REG_VIO_REPORT 0x0029 +#define regUVD_SECURITY_REG_VIO_REPORT_BASE_IDX 1 +#define regUVD_FW_VERSION 0x002a +#define regUVD_FW_VERSION_BASE_IDX 1 +#define regUVD_PF_STATUS 0x002c +#define regUVD_PF_STATUS_BASE_IDX 1 +#define regUVD_DPG_CLK_EN_VCPU_REPORT 0x002e +#define regUVD_DPG_CLK_EN_VCPU_REPORT_BASE_IDX 1 +#define regCC_UVD_VCPU_ERR_DETECT_BOT_LO 0x002f +#define regCC_UVD_VCPU_ERR_DETECT_BOT_LO_BASE_IDX 1 +#define regCC_UVD_VCPU_ERR_DETECT_BOT_HI 0x0030 +#define regCC_UVD_VCPU_ERR_DETECT_BOT_HI_BASE_IDX 1 +#define regCC_UVD_VCPU_ERR_DETECT_TOP_LO 0x0031 +#define regCC_UVD_VCPU_ERR_DETECT_TOP_LO_BASE_IDX 1 +#define regCC_UVD_VCPU_ERR_DETECT_TOP_HI 0x0032 +#define regCC_UVD_VCPU_ERR_DETECT_TOP_HI_BASE_IDX 1 +#define regCC_UVD_VCPU_ERR 0x0033 +#define regCC_UVD_VCPU_ERR_BASE_IDX 1 +#define regCC_UVD_VCPU_ERR_INST_ADDR_LO 0x0034 +#define regCC_UVD_VCPU_ERR_INST_ADDR_LO_BASE_IDX 1 +#define regCC_UVD_VCPU_ERR_INST_ADDR_HI 0x0035 +#define regCC_UVD_VCPU_ERR_INST_ADDR_HI_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC_SPACE 0x003d +#define regUVD_LMI_MMSCH_NC_SPACE_BASE_IDX 1 +#define regUVD_LMI_ATOMIC_SPACE 0x003e +#define regUVD_LMI_ATOMIC_SPACE_BASE_IDX 1 +#define regUVD_GFX8_ADDR_CONFIG 0x0041 +#define regUVD_GFX8_ADDR_CONFIG_BASE_IDX 1 +#define regUVD_GFX10_ADDR_CONFIG 0x0042 +#define regUVD_GFX10_ADDR_CONFIG_BASE_IDX 1 +#define regUVD_GPCNT2_CNTL 0x0043 +#define regUVD_GPCNT2_CNTL_BASE_IDX 1 +#define regUVD_GPCNT2_TARGET_LOWER 0x0044 +#define regUVD_GPCNT2_TARGET_LOWER_BASE_IDX 1 +#define regUVD_GPCNT2_STATUS_LOWER 0x0045 +#define regUVD_GPCNT2_STATUS_LOWER_BASE_IDX 1 +#define regUVD_GPCNT2_TARGET_UPPER 0x0046 +#define regUVD_GPCNT2_TARGET_UPPER_BASE_IDX 1 +#define regUVD_GPCNT2_STATUS_UPPER 0x0047 +#define regUVD_GPCNT2_STATUS_UPPER_BASE_IDX 1 +#define regUVD_GPCNT3_CNTL 0x0048 +#define regUVD_GPCNT3_CNTL_BASE_IDX 1 +#define regUVD_GPCNT3_TARGET_LOWER 0x0049 +#define regUVD_GPCNT3_TARGET_LOWER_BASE_IDX 1 +#define regUVD_GPCNT3_STATUS_LOWER 0x004a +#define regUVD_GPCNT3_STATUS_LOWER_BASE_IDX 1 +#define regUVD_GPCNT3_TARGET_UPPER 0x004b +#define regUVD_GPCNT3_TARGET_UPPER_BASE_IDX 1 +#define regUVD_GPCNT3_STATUS_UPPER 0x004c +#define regUVD_GPCNT3_STATUS_UPPER_BASE_IDX 1 +#define regUVD_VCLK_DS_CNTL 0x004d +#define regUVD_VCLK_DS_CNTL_BASE_IDX 1 +#define regUVD_DCLK_DS_CNTL 0x004e +#define regUVD_DCLK_DS_CNTL_BASE_IDX 1 +#define regUVD_TSC_LOWER 0x004f +#define regUVD_TSC_LOWER_BASE_IDX 1 +#define regUVD_TSC_UPPER 0x0050 +#define regUVD_TSC_UPPER_BASE_IDX 1 +#define regVCN_FEATURES 0x0051 +#define regVCN_FEATURES_BASE_IDX 1 +#define regUVD_GPUIOV_STATUS 0x0055 +#define regUVD_GPUIOV_STATUS_BASE_IDX 1 +#define regUVD_RAS_VCPU_VCODEC_STATUS 0x0057 +#define regUVD_RAS_VCPU_VCODEC_STATUS_BASE_IDX 1 +#define regUVD_RAS_MMSCH_FATAL_ERROR 0x0058 +#define regUVD_RAS_MMSCH_FATAL_ERROR_BASE_IDX 1 +#define regUVD_RAS_JPEG0_STATUS 0x0059 +#define regUVD_RAS_JPEG0_STATUS_BASE_IDX 1 +#define regUVD_RAS_JPEG1_STATUS 0x005a +#define regUVD_RAS_JPEG1_STATUS_BASE_IDX 1 +#define regUVD_RAS_CNTL_PMI_ARB 0x005b +#define regUVD_RAS_CNTL_PMI_ARB_BASE_IDX 1 +#define regUVD_SCRATCH15 0x005c +#define regUVD_SCRATCH15_BASE_IDX 1 +#define regVCN_JPEG_DB_CTRL1 0x005d +#define regVCN_JPEG_DB_CTRL1_BASE_IDX 1 +#define regVCN_JPEG_DB_CTRL2 0x005e +#define regVCN_JPEG_DB_CTRL2_BASE_IDX 1 +#define regVCN_JPEG_DB_CTRL3 0x005f +#define regVCN_JPEG_DB_CTRL3_BASE_IDX 1 +#define regVCN_JPEG_DB_CTRL4 0x0060 +#define regVCN_JPEG_DB_CTRL4_BASE_IDX 1 +#define regVCN_JPEG_DB_CTRL5 0x0061 +#define regVCN_JPEG_DB_CTRL5_BASE_IDX 1 +#define regVCN_JPEG_DB_CTRL6 0x0062 +#define regVCN_JPEG_DB_CTRL6_BASE_IDX 1 +#define regVCN_JPEG_DB_CTRL7 0x0063 +#define regVCN_JPEG_DB_CTRL7_BASE_IDX 1 +#define regUVD_SCRATCH32 0x006d +#define regUVD_SCRATCH32_BASE_IDX 1 +#define regUVD_VERSION 0x006e +#define regUVD_VERSION_BASE_IDX 1 +#define regVCN_RB_DB_CTRL 0x0070 +#define regVCN_RB_DB_CTRL_BASE_IDX 1 +#define regVCN_JPEG_DB_CTRL 0x0071 +#define regVCN_JPEG_DB_CTRL_BASE_IDX 1 +#define regVCN_RB1_DB_CTRL 0x0072 +#define regVCN_RB1_DB_CTRL_BASE_IDX 1 +#define regVCN_RB2_DB_CTRL 0x0073 +#define regVCN_RB2_DB_CTRL_BASE_IDX 1 +#define regVCN_RB3_DB_CTRL 0x0074 +#define regVCN_RB3_DB_CTRL_BASE_IDX 1 +#define regVCN_RB4_DB_CTRL 0x0075 +#define regVCN_RB4_DB_CTRL_BASE_IDX 1 +#define regVCN_RB_ENABLE 0x0085 +#define regVCN_RB_ENABLE_BASE_IDX 1 +#define regVCN_RB_WPTR_CTRL 0x0086 +#define regVCN_RB_WPTR_CTRL_BASE_IDX 1 +#define regUVD_RB_RPTR 0x00ac +#define regUVD_RB_RPTR_BASE_IDX 1 +#define regUVD_RB_WPTR 0x00ad +#define regUVD_RB_WPTR_BASE_IDX 1 +#define regUVD_RB_RPTR2 0x00ae +#define regUVD_RB_RPTR2_BASE_IDX 1 +#define regUVD_RB_WPTR2 0x00af +#define regUVD_RB_WPTR2_BASE_IDX 1 +#define regUVD_RB_RPTR3 0x00b0 +#define regUVD_RB_RPTR3_BASE_IDX 1 +#define regUVD_RB_WPTR3 0x00b1 +#define regUVD_RB_WPTR3_BASE_IDX 1 +#define regUVD_RB_RPTR4 0x00b2 +#define regUVD_RB_RPTR4_BASE_IDX 1 +#define regUVD_RB_WPTR4 0x00b3 +#define regUVD_RB_WPTR4_BASE_IDX 1 +#define regUVD_OUT_RB_RPTR 0x00b4 +#define regUVD_OUT_RB_RPTR_BASE_IDX 1 +#define regUVD_OUT_RB_WPTR 0x00b5 +#define regUVD_OUT_RB_WPTR_BASE_IDX 1 +#define regUVD_AUDIO_RB_RPTR 0x00b6 +#define regUVD_AUDIO_RB_RPTR_BASE_IDX 1 +#define regUVD_AUDIO_RB_WPTR 0x00b7 +#define regUVD_AUDIO_RB_WPTR_BASE_IDX 1 +#define regUVD_RBC_RB_RPTR 0x00b8 +#define regUVD_RBC_RB_RPTR_BASE_IDX 1 +#define regUVD_RBC_RB_WPTR 0x00b9 +#define regUVD_RBC_RB_WPTR_BASE_IDX 1 +#define regUVD_DPG_LMA_CTL2 0x00bb +#define regUVD_DPG_LMA_CTL2_BASE_IDX 1 + + +// addressBlock: aid_uvd0_mmsch_dec +// base address: 0x20d00 +#define regMMSCH_UCODE_ADDR 0x0540 +#define regMMSCH_UCODE_ADDR_BASE_IDX 1 +#define regMMSCH_UCODE_DATA 0x0541 +#define regMMSCH_UCODE_DATA_BASE_IDX 1 +#define regMMSCH_SRAM_ADDR 0x0542 +#define regMMSCH_SRAM_ADDR_BASE_IDX 1 +#define regMMSCH_SRAM_DATA 0x0543 +#define regMMSCH_SRAM_DATA_BASE_IDX 1 +#define regMMSCH_VF_SRAM_OFFSET 0x0544 +#define regMMSCH_VF_SRAM_OFFSET_BASE_IDX 1 +#define regMMSCH_DB_SRAM_OFFSET 0x0545 +#define regMMSCH_DB_SRAM_OFFSET_BASE_IDX 1 +#define regMMSCH_CTX_SRAM_OFFSET 0x0546 +#define regMMSCH_CTX_SRAM_OFFSET_BASE_IDX 1 +#define regMMSCH_CTL 0x0547 +#define regMMSCH_CTL_BASE_IDX 1 +#define regMMSCH_INTR 0x0548 +#define regMMSCH_INTR_BASE_IDX 1 +#define regMMSCH_INTR_ACK 0x0549 +#define regMMSCH_INTR_ACK_BASE_IDX 1 +#define regMMSCH_INTR_STATUS 0x054a +#define regMMSCH_INTR_STATUS_BASE_IDX 1 +#define regMMSCH_VF_VMID 0x054b +#define regMMSCH_VF_VMID_BASE_IDX 1 +#define regMMSCH_VF_CTX_ADDR_LO 0x054c +#define regMMSCH_VF_CTX_ADDR_LO_BASE_IDX 1 +#define regMMSCH_VF_CTX_ADDR_HI 0x054d +#define regMMSCH_VF_CTX_ADDR_HI_BASE_IDX 1 +#define regMMSCH_VF_CTX_SIZE 0x054e +#define regMMSCH_VF_CTX_SIZE_BASE_IDX 1 +#define regMMSCH_VF_GPCOM_ADDR_LO 0x054f +#define regMMSCH_VF_GPCOM_ADDR_LO_BASE_IDX 1 +#define regMMSCH_VF_GPCOM_ADDR_HI 0x0550 +#define regMMSCH_VF_GPCOM_ADDR_HI_BASE_IDX 1 +#define regMMSCH_VF_GPCOM_SIZE 0x0551 +#define regMMSCH_VF_GPCOM_SIZE_BASE_IDX 1 +#define regMMSCH_VF_MAILBOX_HOST 0x0552 +#define regMMSCH_VF_MAILBOX_HOST_BASE_IDX 1 +#define regMMSCH_VF_MAILBOX_RESP 0x0553 +#define regMMSCH_VF_MAILBOX_RESP_BASE_IDX 1 +#define regMMSCH_VF_MAILBOX_0 0x0554 +#define regMMSCH_VF_MAILBOX_0_BASE_IDX 1 +#define regMMSCH_VF_MAILBOX_0_RESP 0x0555 +#define regMMSCH_VF_MAILBOX_0_RESP_BASE_IDX 1 +#define regMMSCH_VF_MAILBOX_1 0x0556 +#define regMMSCH_VF_MAILBOX_1_BASE_IDX 1 +#define regMMSCH_VF_MAILBOX_1_RESP 0x0557 +#define regMMSCH_VF_MAILBOX_1_RESP_BASE_IDX 1 +#define regMMSCH_CNTL 0x055c +#define regMMSCH_CNTL_BASE_IDX 1 +#define regMMSCH_NONCACHE_OFFSET0 0x055d +#define regMMSCH_NONCACHE_OFFSET0_BASE_IDX 1 +#define regMMSCH_NONCACHE_SIZE0 0x055e +#define regMMSCH_NONCACHE_SIZE0_BASE_IDX 1 +#define regMMSCH_NONCACHE_OFFSET1 0x055f +#define regMMSCH_NONCACHE_OFFSET1_BASE_IDX 1 +#define regMMSCH_NONCACHE_SIZE1 0x0560 +#define regMMSCH_NONCACHE_SIZE1_BASE_IDX 1 +#define regMMSCH_PROC_STATE1 0x0566 +#define regMMSCH_PROC_STATE1_BASE_IDX 1 +#define regMMSCH_LAST_MC_ADDR 0x0567 +#define regMMSCH_LAST_MC_ADDR_BASE_IDX 1 +#define regMMSCH_LAST_MEM_ACCESS_HI 0x0568 +#define regMMSCH_LAST_MEM_ACCESS_HI_BASE_IDX 1 +#define regMMSCH_LAST_MEM_ACCESS_LO 0x0569 +#define regMMSCH_LAST_MEM_ACCESS_LO_BASE_IDX 1 +#define regMMSCH_IOV_ACTIVE_FCN_ID 0x056a +#define regMMSCH_IOV_ACTIVE_FCN_ID_BASE_IDX 1 +#define regMMSCH_SCRATCH_0 0x056b +#define regMMSCH_SCRATCH_0_BASE_IDX 1 +#define regMMSCH_SCRATCH_1 0x056c +#define regMMSCH_SCRATCH_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_SCH_BLOCK_0 0x056d +#define regMMSCH_GPUIOV_SCH_BLOCK_0_BASE_IDX 1 +#define regMMSCH_GPUIOV_CMD_CONTROL_0 0x056e +#define regMMSCH_GPUIOV_CMD_CONTROL_0_BASE_IDX 1 +#define regMMSCH_GPUIOV_CMD_STATUS_0 0x056f +#define regMMSCH_GPUIOV_CMD_STATUS_0_BASE_IDX 1 +#define regMMSCH_GPUIOV_VM_BUSY_STATUS_0 0x0570 +#define regMMSCH_GPUIOV_VM_BUSY_STATUS_0_BASE_IDX 1 +#define regMMSCH_GPUIOV_ACTIVE_FCNS_0 0x0571 +#define regMMSCH_GPUIOV_ACTIVE_FCNS_0_BASE_IDX 1 +#define regMMSCH_GPUIOV_ACTIVE_FCN_ID_0 0x0572 +#define regMMSCH_GPUIOV_ACTIVE_FCN_ID_0_BASE_IDX 1 +#define regMMSCH_GPUIOV_DW6_0 0x0573 +#define regMMSCH_GPUIOV_DW6_0_BASE_IDX 1 +#define regMMSCH_GPUIOV_DW7_0 0x0574 +#define regMMSCH_GPUIOV_DW7_0_BASE_IDX 1 +#define regMMSCH_GPUIOV_DW8_0 0x0575 +#define regMMSCH_GPUIOV_DW8_0_BASE_IDX 1 +#define regMMSCH_GPUIOV_SCH_BLOCK_1 0x0576 +#define regMMSCH_GPUIOV_SCH_BLOCK_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_CMD_CONTROL_1 0x0577 +#define regMMSCH_GPUIOV_CMD_CONTROL_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_CMD_STATUS_1 0x0578 +#define regMMSCH_GPUIOV_CMD_STATUS_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_VM_BUSY_STATUS_1 0x0579 +#define regMMSCH_GPUIOV_VM_BUSY_STATUS_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_ACTIVE_FCNS_1 0x057a +#define regMMSCH_GPUIOV_ACTIVE_FCNS_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_ACTIVE_FCN_ID_1 0x057b +#define regMMSCH_GPUIOV_ACTIVE_FCN_ID_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_DW6_1 0x057c +#define regMMSCH_GPUIOV_DW6_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_DW7_1 0x057d +#define regMMSCH_GPUIOV_DW7_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_DW8_1 0x057e +#define regMMSCH_GPUIOV_DW8_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_CNTXT 0x057f +#define regMMSCH_GPUIOV_CNTXT_BASE_IDX 1 +#define regMMSCH_SCRATCH_2 0x0580 +#define regMMSCH_SCRATCH_2_BASE_IDX 1 +#define regMMSCH_SCRATCH_3 0x0581 +#define regMMSCH_SCRATCH_3_BASE_IDX 1 +#define regMMSCH_SCRATCH_4 0x0582 +#define regMMSCH_SCRATCH_4_BASE_IDX 1 +#define regMMSCH_SCRATCH_5 0x0583 +#define regMMSCH_SCRATCH_5_BASE_IDX 1 +#define regMMSCH_SCRATCH_6 0x0584 +#define regMMSCH_SCRATCH_6_BASE_IDX 1 +#define regMMSCH_SCRATCH_7 0x0585 +#define regMMSCH_SCRATCH_7_BASE_IDX 1 +#define regMMSCH_VFID_FIFO_HEAD_0 0x0586 +#define regMMSCH_VFID_FIFO_HEAD_0_BASE_IDX 1 +#define regMMSCH_VFID_FIFO_TAIL_0 0x0587 +#define regMMSCH_VFID_FIFO_TAIL_0_BASE_IDX 1 +#define regMMSCH_VFID_FIFO_HEAD_1 0x0588 +#define regMMSCH_VFID_FIFO_HEAD_1_BASE_IDX 1 +#define regMMSCH_VFID_FIFO_TAIL_1 0x0589 +#define regMMSCH_VFID_FIFO_TAIL_1_BASE_IDX 1 +#define regMMSCH_NACK_STATUS 0x058a +#define regMMSCH_NACK_STATUS_BASE_IDX 1 +#define regMMSCH_VF_MAILBOX0_DATA 0x058b +#define regMMSCH_VF_MAILBOX0_DATA_BASE_IDX 1 +#define regMMSCH_VF_MAILBOX1_DATA 0x058c +#define regMMSCH_VF_MAILBOX1_DATA_BASE_IDX 1 +#define regMMSCH_GPUIOV_SCH_BLOCK_IP_0 0x058d +#define regMMSCH_GPUIOV_SCH_BLOCK_IP_0_BASE_IDX 1 +#define regMMSCH_GPUIOV_CMD_STATUS_IP_0 0x058e +#define regMMSCH_GPUIOV_CMD_STATUS_IP_0_BASE_IDX 1 +#define regMMSCH_GPUIOV_ACTIVE_FCN_ID_IP_0 0x058f +#define regMMSCH_GPUIOV_ACTIVE_FCN_ID_IP_0_BASE_IDX 1 +#define regMMSCH_GPUIOV_SCH_BLOCK_IP_1 0x0590 +#define regMMSCH_GPUIOV_SCH_BLOCK_IP_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_CMD_STATUS_IP_1 0x0591 +#define regMMSCH_GPUIOV_CMD_STATUS_IP_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_ACTIVE_FCN_ID_IP_1 0x0592 +#define regMMSCH_GPUIOV_ACTIVE_FCN_ID_IP_1_BASE_IDX 1 +#define regMMSCH_GPUIOV_CNTXT_IP 0x0593 +#define regMMSCH_GPUIOV_CNTXT_IP_BASE_IDX 1 +#define regMMSCH_GPUIOV_SCH_BLOCK_2 0x0594 +#define regMMSCH_GPUIOV_SCH_BLOCK_2_BASE_IDX 1 +#define regMMSCH_GPUIOV_CMD_CONTROL_2 0x0595 +#define regMMSCH_GPUIOV_CMD_CONTROL_2_BASE_IDX 1 +#define regMMSCH_GPUIOV_CMD_STATUS_2 0x0596 +#define regMMSCH_GPUIOV_CMD_STATUS_2_BASE_IDX 1 +#define regMMSCH_GPUIOV_VM_BUSY_STATUS_2 0x0597 +#define regMMSCH_GPUIOV_VM_BUSY_STATUS_2_BASE_IDX 1 +#define regMMSCH_GPUIOV_ACTIVE_FCNS_2 0x0598 +#define regMMSCH_GPUIOV_ACTIVE_FCNS_2_BASE_IDX 1 +#define regMMSCH_GPUIOV_ACTIVE_FCN_ID_2 0x0599 +#define regMMSCH_GPUIOV_ACTIVE_FCN_ID_2_BASE_IDX 1 +#define regMMSCH_GPUIOV_DW6_2 0x059a +#define regMMSCH_GPUIOV_DW6_2_BASE_IDX 1 +#define regMMSCH_GPUIOV_DW7_2 0x059b +#define regMMSCH_GPUIOV_DW7_2_BASE_IDX 1 +#define regMMSCH_GPUIOV_DW8_2 0x059c +#define regMMSCH_GPUIOV_DW8_2_BASE_IDX 1 +#define regMMSCH_GPUIOV_SCH_BLOCK_IP_2 0x059d +#define regMMSCH_GPUIOV_SCH_BLOCK_IP_2_BASE_IDX 1 +#define regMMSCH_GPUIOV_CMD_STATUS_IP_2 0x059e +#define regMMSCH_GPUIOV_CMD_STATUS_IP_2_BASE_IDX 1 +#define regMMSCH_GPUIOV_ACTIVE_FCN_ID_IP_2 0x059f +#define regMMSCH_GPUIOV_ACTIVE_FCN_ID_IP_2_BASE_IDX 1 +#define regMMSCH_VFID_FIFO_HEAD_2 0x05a0 +#define regMMSCH_VFID_FIFO_HEAD_2_BASE_IDX 1 +#define regMMSCH_VFID_FIFO_TAIL_2 0x05a1 +#define regMMSCH_VFID_FIFO_TAIL_2_BASE_IDX 1 +#define regMMSCH_VM_BUSY_STATUS_0 0x05a2 +#define regMMSCH_VM_BUSY_STATUS_0_BASE_IDX 1 +#define regMMSCH_VM_BUSY_STATUS_1 0x05a3 +#define regMMSCH_VM_BUSY_STATUS_1_BASE_IDX 1 +#define regMMSCH_VM_BUSY_STATUS_2 0x05a4 +#define regMMSCH_VM_BUSY_STATUS_2_BASE_IDX 1 + + +// addressBlock: aid_uvd0_slmi_adpdec +// base address: 0x21c00 +#define regUVD_LMI_MMSCH_NC0_64BIT_BAR_LOW 0x0900 +#define regUVD_LMI_MMSCH_NC0_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC0_64BIT_BAR_HIGH 0x0901 +#define regUVD_LMI_MMSCH_NC0_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC1_64BIT_BAR_LOW 0x0902 +#define regUVD_LMI_MMSCH_NC1_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC1_64BIT_BAR_HIGH 0x0903 +#define regUVD_LMI_MMSCH_NC1_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC2_64BIT_BAR_LOW 0x0904 +#define regUVD_LMI_MMSCH_NC2_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC2_64BIT_BAR_HIGH 0x0905 +#define regUVD_LMI_MMSCH_NC2_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC3_64BIT_BAR_LOW 0x0906 +#define regUVD_LMI_MMSCH_NC3_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC3_64BIT_BAR_HIGH 0x0907 +#define regUVD_LMI_MMSCH_NC3_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC4_64BIT_BAR_LOW 0x0908 +#define regUVD_LMI_MMSCH_NC4_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC4_64BIT_BAR_HIGH 0x0909 +#define regUVD_LMI_MMSCH_NC4_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC5_64BIT_BAR_LOW 0x090a +#define regUVD_LMI_MMSCH_NC5_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC5_64BIT_BAR_HIGH 0x090b +#define regUVD_LMI_MMSCH_NC5_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC6_64BIT_BAR_LOW 0x090c +#define regUVD_LMI_MMSCH_NC6_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC6_64BIT_BAR_HIGH 0x090d +#define regUVD_LMI_MMSCH_NC6_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC7_64BIT_BAR_LOW 0x090e +#define regUVD_LMI_MMSCH_NC7_64BIT_BAR_LOW_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC7_64BIT_BAR_HIGH 0x090f +#define regUVD_LMI_MMSCH_NC7_64BIT_BAR_HIGH_BASE_IDX 1 +#define regUVD_LMI_MMSCH_NC_VMID 0x0910 +#define regUVD_LMI_MMSCH_NC_VMID_BASE_IDX 1 +#define regUVD_LMI_MMSCH_CTRL 0x0911 +#define regUVD_LMI_MMSCH_CTRL_BASE_IDX 1 +#define regUVD_MMSCH_LMI_STATUS 0x0912 +#define regUVD_MMSCH_LMI_STATUS_BASE_IDX 1 +#define regVCN_RAS_CNTL_MMSCH 0x0914 +#define regVCN_RAS_CNTL_MMSCH_BASE_IDX 1 + +// addressBlock: aid_uvd0_vcn_edcc_dec +// base address: 0x21d20 +#define regVCN_UE_ERR_STATUS_LO_VIDD 0x094c +#define regVCN_UE_ERR_STATUS_LO_VIDD_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_VIDD 0x094d +#define regVCN_UE_ERR_STATUS_HI_VIDD_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_VIDV 0x094e +#define regVCN_UE_ERR_STATUS_LO_VIDV_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_VIDV 0x094f +#define regVCN_UE_ERR_STATUS_HI_VIDV_BASE_IDX 1 +#define regVCN_CE_ERR_STATUS_LO_MMSCHD 0x0950 +#define regVCN_CE_ERR_STATUS_LO_MMSCHD_BASE_IDX 1 +#define regVCN_CE_ERR_STATUS_HI_MMSCHD 0x0951 +#define regVCN_CE_ERR_STATUS_HI_MMSCHD_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG0S 0x0952 +#define regVCN_UE_ERR_STATUS_LO_JPEG0S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG0S 0x0953 +#define regVCN_UE_ERR_STATUS_HI_JPEG0S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG0D 0x0954 +#define regVCN_UE_ERR_STATUS_LO_JPEG0D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG0D 0x0955 +#define regVCN_UE_ERR_STATUS_HI_JPEG0D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG1S 0x0956 +#define regVCN_UE_ERR_STATUS_LO_JPEG1S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG1S 0x0957 +#define regVCN_UE_ERR_STATUS_HI_JPEG1S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG1D 0x0958 +#define regVCN_UE_ERR_STATUS_LO_JPEG1D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG1D 0x0959 +#define regVCN_UE_ERR_STATUS_HI_JPEG1D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG2S 0x095a +#define regVCN_UE_ERR_STATUS_LO_JPEG2S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG2S 0x095b +#define regVCN_UE_ERR_STATUS_HI_JPEG2S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG2D 0x095c +#define regVCN_UE_ERR_STATUS_LO_JPEG2D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG2D 0x095d +#define regVCN_UE_ERR_STATUS_HI_JPEG2D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG3S 0x095e +#define regVCN_UE_ERR_STATUS_LO_JPEG3S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG3S 0x095f +#define regVCN_UE_ERR_STATUS_HI_JPEG3S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG3D 0x0960 +#define regVCN_UE_ERR_STATUS_LO_JPEG3D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG3D 0x0961 +#define regVCN_UE_ERR_STATUS_HI_JPEG3D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG4S 0x0962 +#define regVCN_UE_ERR_STATUS_LO_JPEG4S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG4S 0x0963 +#define regVCN_UE_ERR_STATUS_HI_JPEG4S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG4D 0x0964 +#define regVCN_UE_ERR_STATUS_LO_JPEG4D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG4D 0x0965 +#define regVCN_UE_ERR_STATUS_HI_JPEG4D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG5S 0x0966 +#define regVCN_UE_ERR_STATUS_LO_JPEG5S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG5S 0x0967 +#define regVCN_UE_ERR_STATUS_HI_JPEG5S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG5D 0x0968 +#define regVCN_UE_ERR_STATUS_LO_JPEG5D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG5D 0x0969 +#define regVCN_UE_ERR_STATUS_HI_JPEG5D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG6S 0x096a +#define regVCN_UE_ERR_STATUS_LO_JPEG6S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG6S 0x096b +#define regVCN_UE_ERR_STATUS_HI_JPEG6S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG6D 0x096c +#define regVCN_UE_ERR_STATUS_LO_JPEG6D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG6D 0x096d +#define regVCN_UE_ERR_STATUS_HI_JPEG6D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG7S 0x096e +#define regVCN_UE_ERR_STATUS_LO_JPEG7S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG7S 0x096f +#define regVCN_UE_ERR_STATUS_HI_JPEG7S_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_LO_JPEG7D 0x0970 +#define regVCN_UE_ERR_STATUS_LO_JPEG7D_BASE_IDX 1 +#define regVCN_UE_ERR_STATUS_HI_JPEG7D 0x0971 +#define regVCN_UE_ERR_STATUS_HI_JPEG7D_BASE_IDX 1 + +// addressBlock: aid_uvd0_uvd_jrbc1_uvd_jrbc_dec +// base address: 0x1e000 +#define regUVD_JRBC1_UVD_JRBC_RB_WPTR 0x0000 +#define regUVD_JRBC1_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_RB_CNTL 0x0001 +#define regUVD_JRBC1_UVD_JRBC_RB_CNTL_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_IB_SIZE 0x0002 +#define regUVD_JRBC1_UVD_JRBC_IB_SIZE_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_URGENT_CNTL 0x0003 +#define regUVD_JRBC1_UVD_JRBC_URGENT_CNTL_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_RB_REF_DATA 0x0004 +#define regUVD_JRBC1_UVD_JRBC_RB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_RB_COND_RD_TIMER 0x0005 +#define regUVD_JRBC1_UVD_JRBC_RB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_SOFT_RESET 0x0008 +#define regUVD_JRBC1_UVD_JRBC_SOFT_RESET_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_STATUS 0x0009 +#define regUVD_JRBC1_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_RB_RPTR 0x000a +#define regUVD_JRBC1_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_RB_BUF_STATUS 0x000b +#define regUVD_JRBC1_UVD_JRBC_RB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_IB_BUF_STATUS 0x000c +#define regUVD_JRBC1_UVD_JRBC_IB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_IB_SIZE_UPDATE 0x000d +#define regUVD_JRBC1_UVD_JRBC_IB_SIZE_UPDATE_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_IB_COND_RD_TIMER 0x000e +#define regUVD_JRBC1_UVD_JRBC_IB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_IB_REF_DATA 0x000f +#define regUVD_JRBC1_UVD_JRBC_IB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JPEG_PREEMPT_CMD 0x0010 +#define regUVD_JRBC1_UVD_JPEG_PREEMPT_CMD_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JPEG_PREEMPT_FENCE_DATA0 0x0011 +#define regUVD_JRBC1_UVD_JPEG_PREEMPT_FENCE_DATA0_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JPEG_PREEMPT_FENCE_DATA1 0x0012 +#define regUVD_JRBC1_UVD_JPEG_PREEMPT_FENCE_DATA1_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_RB_SIZE 0x0013 +#define regUVD_JRBC1_UVD_JRBC_RB_SIZE_BASE_IDX 0 +#define regUVD_JRBC1_UVD_JRBC_SCRATCH0 0x0014 +#define regUVD_JRBC1_UVD_JRBC_SCRATCH0_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jrbc2_uvd_jrbc_dec +// base address: 0x1e100 +#define regUVD_JRBC2_UVD_JRBC_RB_WPTR 0x0040 +#define regUVD_JRBC2_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_RB_CNTL 0x0041 +#define regUVD_JRBC2_UVD_JRBC_RB_CNTL_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_IB_SIZE 0x0042 +#define regUVD_JRBC2_UVD_JRBC_IB_SIZE_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_URGENT_CNTL 0x0043 +#define regUVD_JRBC2_UVD_JRBC_URGENT_CNTL_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_RB_REF_DATA 0x0044 +#define regUVD_JRBC2_UVD_JRBC_RB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_RB_COND_RD_TIMER 0x0045 +#define regUVD_JRBC2_UVD_JRBC_RB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_SOFT_RESET 0x0048 +#define regUVD_JRBC2_UVD_JRBC_SOFT_RESET_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_STATUS 0x0049 +#define regUVD_JRBC2_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_RB_RPTR 0x004a +#define regUVD_JRBC2_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_RB_BUF_STATUS 0x004b +#define regUVD_JRBC2_UVD_JRBC_RB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_IB_BUF_STATUS 0x004c +#define regUVD_JRBC2_UVD_JRBC_IB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_IB_SIZE_UPDATE 0x004d +#define regUVD_JRBC2_UVD_JRBC_IB_SIZE_UPDATE_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_IB_COND_RD_TIMER 0x004e +#define regUVD_JRBC2_UVD_JRBC_IB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_IB_REF_DATA 0x004f +#define regUVD_JRBC2_UVD_JRBC_IB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JPEG_PREEMPT_CMD 0x0050 +#define regUVD_JRBC2_UVD_JPEG_PREEMPT_CMD_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JPEG_PREEMPT_FENCE_DATA0 0x0051 +#define regUVD_JRBC2_UVD_JPEG_PREEMPT_FENCE_DATA0_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JPEG_PREEMPT_FENCE_DATA1 0x0052 +#define regUVD_JRBC2_UVD_JPEG_PREEMPT_FENCE_DATA1_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_RB_SIZE 0x0053 +#define regUVD_JRBC2_UVD_JRBC_RB_SIZE_BASE_IDX 0 +#define regUVD_JRBC2_UVD_JRBC_SCRATCH0 0x0054 +#define regUVD_JRBC2_UVD_JRBC_SCRATCH0_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jrbc3_uvd_jrbc_dec +// base address: 0x1e200 +#define regUVD_JRBC3_UVD_JRBC_RB_WPTR 0x0080 +#define regUVD_JRBC3_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_RB_CNTL 0x0081 +#define regUVD_JRBC3_UVD_JRBC_RB_CNTL_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_IB_SIZE 0x0082 +#define regUVD_JRBC3_UVD_JRBC_IB_SIZE_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_URGENT_CNTL 0x0083 +#define regUVD_JRBC3_UVD_JRBC_URGENT_CNTL_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_RB_REF_DATA 0x0084 +#define regUVD_JRBC3_UVD_JRBC_RB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_RB_COND_RD_TIMER 0x0085 +#define regUVD_JRBC3_UVD_JRBC_RB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_SOFT_RESET 0x0088 +#define regUVD_JRBC3_UVD_JRBC_SOFT_RESET_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_STATUS 0x0089 +#define regUVD_JRBC3_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_RB_RPTR 0x008a +#define regUVD_JRBC3_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_RB_BUF_STATUS 0x008b +#define regUVD_JRBC3_UVD_JRBC_RB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_IB_BUF_STATUS 0x008c +#define regUVD_JRBC3_UVD_JRBC_IB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_IB_SIZE_UPDATE 0x008d +#define regUVD_JRBC3_UVD_JRBC_IB_SIZE_UPDATE_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_IB_COND_RD_TIMER 0x008e +#define regUVD_JRBC3_UVD_JRBC_IB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_IB_REF_DATA 0x008f +#define regUVD_JRBC3_UVD_JRBC_IB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JPEG_PREEMPT_CMD 0x0090 +#define regUVD_JRBC3_UVD_JPEG_PREEMPT_CMD_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JPEG_PREEMPT_FENCE_DATA0 0x0091 +#define regUVD_JRBC3_UVD_JPEG_PREEMPT_FENCE_DATA0_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JPEG_PREEMPT_FENCE_DATA1 0x0092 +#define regUVD_JRBC3_UVD_JPEG_PREEMPT_FENCE_DATA1_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_RB_SIZE 0x0093 +#define regUVD_JRBC3_UVD_JRBC_RB_SIZE_BASE_IDX 0 +#define regUVD_JRBC3_UVD_JRBC_SCRATCH0 0x0094 +#define regUVD_JRBC3_UVD_JRBC_SCRATCH0_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jrbc4_uvd_jrbc_dec +// base address: 0x1e300 +#define regUVD_JRBC4_UVD_JRBC_RB_WPTR 0x00c0 +#define regUVD_JRBC4_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_RB_CNTL 0x00c1 +#define regUVD_JRBC4_UVD_JRBC_RB_CNTL_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_IB_SIZE 0x00c2 +#define regUVD_JRBC4_UVD_JRBC_IB_SIZE_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_URGENT_CNTL 0x00c3 +#define regUVD_JRBC4_UVD_JRBC_URGENT_CNTL_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_RB_REF_DATA 0x00c4 +#define regUVD_JRBC4_UVD_JRBC_RB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_RB_COND_RD_TIMER 0x00c5 +#define regUVD_JRBC4_UVD_JRBC_RB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_SOFT_RESET 0x00c8 +#define regUVD_JRBC4_UVD_JRBC_SOFT_RESET_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_STATUS 0x00c9 +#define regUVD_JRBC4_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_RB_RPTR 0x00ca +#define regUVD_JRBC4_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_RB_BUF_STATUS 0x00cb +#define regUVD_JRBC4_UVD_JRBC_RB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_IB_BUF_STATUS 0x00cc +#define regUVD_JRBC4_UVD_JRBC_IB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_IB_SIZE_UPDATE 0x00cd +#define regUVD_JRBC4_UVD_JRBC_IB_SIZE_UPDATE_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_IB_COND_RD_TIMER 0x00ce +#define regUVD_JRBC4_UVD_JRBC_IB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_IB_REF_DATA 0x00cf +#define regUVD_JRBC4_UVD_JRBC_IB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JPEG_PREEMPT_CMD 0x00d0 +#define regUVD_JRBC4_UVD_JPEG_PREEMPT_CMD_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JPEG_PREEMPT_FENCE_DATA0 0x00d1 +#define regUVD_JRBC4_UVD_JPEG_PREEMPT_FENCE_DATA0_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JPEG_PREEMPT_FENCE_DATA1 0x00d2 +#define regUVD_JRBC4_UVD_JPEG_PREEMPT_FENCE_DATA1_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_RB_SIZE 0x00d3 +#define regUVD_JRBC4_UVD_JRBC_RB_SIZE_BASE_IDX 0 +#define regUVD_JRBC4_UVD_JRBC_SCRATCH0 0x00d4 +#define regUVD_JRBC4_UVD_JRBC_SCRATCH0_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jrbc5_uvd_jrbc_dec +// base address: 0x1e400 +#define regUVD_JRBC5_UVD_JRBC_RB_WPTR 0x0100 +#define regUVD_JRBC5_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_RB_CNTL 0x0101 +#define regUVD_JRBC5_UVD_JRBC_RB_CNTL_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_IB_SIZE 0x0102 +#define regUVD_JRBC5_UVD_JRBC_IB_SIZE_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_URGENT_CNTL 0x0103 +#define regUVD_JRBC5_UVD_JRBC_URGENT_CNTL_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_RB_REF_DATA 0x0104 +#define regUVD_JRBC5_UVD_JRBC_RB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_RB_COND_RD_TIMER 0x0105 +#define regUVD_JRBC5_UVD_JRBC_RB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_SOFT_RESET 0x0108 +#define regUVD_JRBC5_UVD_JRBC_SOFT_RESET_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_STATUS 0x0109 +#define regUVD_JRBC5_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_RB_RPTR 0x010a +#define regUVD_JRBC5_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_RB_BUF_STATUS 0x010b +#define regUVD_JRBC5_UVD_JRBC_RB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_IB_BUF_STATUS 0x010c +#define regUVD_JRBC5_UVD_JRBC_IB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_IB_SIZE_UPDATE 0x010d +#define regUVD_JRBC5_UVD_JRBC_IB_SIZE_UPDATE_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_IB_COND_RD_TIMER 0x010e +#define regUVD_JRBC5_UVD_JRBC_IB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_IB_REF_DATA 0x010f +#define regUVD_JRBC5_UVD_JRBC_IB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JPEG_PREEMPT_CMD 0x0110 +#define regUVD_JRBC5_UVD_JPEG_PREEMPT_CMD_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JPEG_PREEMPT_FENCE_DATA0 0x0111 +#define regUVD_JRBC5_UVD_JPEG_PREEMPT_FENCE_DATA0_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JPEG_PREEMPT_FENCE_DATA1 0x0112 +#define regUVD_JRBC5_UVD_JPEG_PREEMPT_FENCE_DATA1_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_RB_SIZE 0x0113 +#define regUVD_JRBC5_UVD_JRBC_RB_SIZE_BASE_IDX 0 +#define regUVD_JRBC5_UVD_JRBC_SCRATCH0 0x0114 +#define regUVD_JRBC5_UVD_JRBC_SCRATCH0_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jrbc6_uvd_jrbc_dec +// base address: 0x1e500 +#define regUVD_JRBC6_UVD_JRBC_RB_WPTR 0x0140 +#define regUVD_JRBC6_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_RB_CNTL 0x0141 +#define regUVD_JRBC6_UVD_JRBC_RB_CNTL_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_IB_SIZE 0x0142 +#define regUVD_JRBC6_UVD_JRBC_IB_SIZE_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_URGENT_CNTL 0x0143 +#define regUVD_JRBC6_UVD_JRBC_URGENT_CNTL_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_RB_REF_DATA 0x0144 +#define regUVD_JRBC6_UVD_JRBC_RB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_RB_COND_RD_TIMER 0x0145 +#define regUVD_JRBC6_UVD_JRBC_RB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_SOFT_RESET 0x0148 +#define regUVD_JRBC6_UVD_JRBC_SOFT_RESET_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_STATUS 0x0149 +#define regUVD_JRBC6_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_RB_RPTR 0x014a +#define regUVD_JRBC6_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_RB_BUF_STATUS 0x014b +#define regUVD_JRBC6_UVD_JRBC_RB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_IB_BUF_STATUS 0x014c +#define regUVD_JRBC6_UVD_JRBC_IB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_IB_SIZE_UPDATE 0x014d +#define regUVD_JRBC6_UVD_JRBC_IB_SIZE_UPDATE_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_IB_COND_RD_TIMER 0x014e +#define regUVD_JRBC6_UVD_JRBC_IB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_IB_REF_DATA 0x014f +#define regUVD_JRBC6_UVD_JRBC_IB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JPEG_PREEMPT_CMD 0x0150 +#define regUVD_JRBC6_UVD_JPEG_PREEMPT_CMD_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JPEG_PREEMPT_FENCE_DATA0 0x0151 +#define regUVD_JRBC6_UVD_JPEG_PREEMPT_FENCE_DATA0_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JPEG_PREEMPT_FENCE_DATA1 0x0152 +#define regUVD_JRBC6_UVD_JPEG_PREEMPT_FENCE_DATA1_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_RB_SIZE 0x0153 +#define regUVD_JRBC6_UVD_JRBC_RB_SIZE_BASE_IDX 0 +#define regUVD_JRBC6_UVD_JRBC_SCRATCH0 0x0154 +#define regUVD_JRBC6_UVD_JRBC_SCRATCH0_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jrbc7_uvd_jrbc_dec +// base address: 0x1e600 +#define regUVD_JRBC7_UVD_JRBC_RB_WPTR 0x0180 +#define regUVD_JRBC7_UVD_JRBC_RB_WPTR_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_RB_CNTL 0x0181 +#define regUVD_JRBC7_UVD_JRBC_RB_CNTL_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_IB_SIZE 0x0182 +#define regUVD_JRBC7_UVD_JRBC_IB_SIZE_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_URGENT_CNTL 0x0183 +#define regUVD_JRBC7_UVD_JRBC_URGENT_CNTL_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_RB_REF_DATA 0x0184 +#define regUVD_JRBC7_UVD_JRBC_RB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_RB_COND_RD_TIMER 0x0185 +#define regUVD_JRBC7_UVD_JRBC_RB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_SOFT_RESET 0x0188 +#define regUVD_JRBC7_UVD_JRBC_SOFT_RESET_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_STATUS 0x0189 +#define regUVD_JRBC7_UVD_JRBC_STATUS_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_RB_RPTR 0x018a +#define regUVD_JRBC7_UVD_JRBC_RB_RPTR_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_RB_BUF_STATUS 0x018b +#define regUVD_JRBC7_UVD_JRBC_RB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_IB_BUF_STATUS 0x018c +#define regUVD_JRBC7_UVD_JRBC_IB_BUF_STATUS_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_IB_SIZE_UPDATE 0x018d +#define regUVD_JRBC7_UVD_JRBC_IB_SIZE_UPDATE_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_IB_COND_RD_TIMER 0x018e +#define regUVD_JRBC7_UVD_JRBC_IB_COND_RD_TIMER_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_IB_REF_DATA 0x018f +#define regUVD_JRBC7_UVD_JRBC_IB_REF_DATA_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JPEG_PREEMPT_CMD 0x0190 +#define regUVD_JRBC7_UVD_JPEG_PREEMPT_CMD_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JPEG_PREEMPT_FENCE_DATA0 0x0191 +#define regUVD_JRBC7_UVD_JPEG_PREEMPT_FENCE_DATA0_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JPEG_PREEMPT_FENCE_DATA1 0x0192 +#define regUVD_JRBC7_UVD_JPEG_PREEMPT_FENCE_DATA1_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_RB_SIZE 0x0193 +#define regUVD_JRBC7_UVD_JRBC_RB_SIZE_BASE_IDX 0 +#define regUVD_JRBC7_UVD_JRBC_SCRATCH0 0x0194 +#define regUVD_JRBC7_UVD_JRBC_SCRATCH0_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jmi1_uvd_jmi_dec +// base address: 0x1e080 +#define regUVD_JMI1_UVD_JPEG_DEC_PF_CTRL 0x0020 +#define regUVD_JMI1_UVD_JPEG_DEC_PF_CTRL_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JRBC_CTRL 0x0021 +#define regUVD_JMI1_UVD_LMI_JRBC_CTRL_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JPEG_CTRL 0x0022 +#define regUVD_JMI1_UVD_LMI_JPEG_CTRL_BASE_IDX 0 +#define regUVD_JMI1_JPEG_LMI_DROP 0x0023 +#define regUVD_JMI1_JPEG_LMI_DROP_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JRBC_IB_VMID 0x0024 +#define regUVD_JMI1_UVD_LMI_JRBC_IB_VMID_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JRBC_RB_VMID 0x0025 +#define regUVD_JMI1_UVD_LMI_JRBC_RB_VMID_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JPEG_VMID 0x0026 +#define regUVD_JMI1_UVD_LMI_JPEG_VMID_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW 0x0027 +#define regUVD_JMI1_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH 0x0028 +#define regUVD_JMI1_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JRBC_RB_64BIT_BAR_LOW 0x0029 +#define regUVD_JMI1_UVD_LMI_JRBC_RB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH 0x002a +#define regUVD_JMI1_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW 0x002b +#define regUVD_JMI1_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH 0x002c +#define regUVD_JMI1_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JPEG_PREEMPT_VMID 0x002d +#define regUVD_JMI1_UVD_LMI_JPEG_PREEMPT_VMID_BASE_IDX 0 +#define regUVD_JMI1_UVD_JMI_DEC_SWAP_CNTL 0x002e +#define regUVD_JMI1_UVD_JMI_DEC_SWAP_CNTL_BASE_IDX 0 +#define regUVD_JMI1_UVD_JMI_ATOMIC_CNTL 0x002f +#define regUVD_JMI1_UVD_JMI_ATOMIC_CNTL_BASE_IDX 0 +#define regUVD_JMI1_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW 0x0030 +#define regUVD_JMI1_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI1_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH 0x0031 +#define regUVD_JMI1_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JPEG_READ_64BIT_BAR_LOW 0x0032 +#define regUVD_JMI1_UVD_LMI_JPEG_READ_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH 0x0033 +#define regUVD_JMI1_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW 0x0034 +#define regUVD_JMI1_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH 0x0035 +#define regUVD_JMI1_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JRBC_IB_64BIT_BAR_LOW 0x0036 +#define regUVD_JMI1_UVD_LMI_JRBC_IB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH 0x0037 +#define regUVD_JMI1_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW 0x0038 +#define regUVD_JMI1_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI1_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH 0x0039 +#define regUVD_JMI1_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI1_UVD_JMI_ATOMIC_CNTL2 0x003d +#define regUVD_JMI1_UVD_JMI_ATOMIC_CNTL2_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jmi2_uvd_jmi_dec +// base address: 0x1e180 +#define regUVD_JMI2_UVD_JPEG_DEC_PF_CTRL 0x0060 +#define regUVD_JMI2_UVD_JPEG_DEC_PF_CTRL_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JRBC_CTRL 0x0061 +#define regUVD_JMI2_UVD_LMI_JRBC_CTRL_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JPEG_CTRL 0x0062 +#define regUVD_JMI2_UVD_LMI_JPEG_CTRL_BASE_IDX 0 +#define regUVD_JMI2_JPEG_LMI_DROP 0x0063 +#define regUVD_JMI2_JPEG_LMI_DROP_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JRBC_IB_VMID 0x0064 +#define regUVD_JMI2_UVD_LMI_JRBC_IB_VMID_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JRBC_RB_VMID 0x0065 +#define regUVD_JMI2_UVD_LMI_JRBC_RB_VMID_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JPEG_VMID 0x0066 +#define regUVD_JMI2_UVD_LMI_JPEG_VMID_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW 0x0067 +#define regUVD_JMI2_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH 0x0068 +#define regUVD_JMI2_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JRBC_RB_64BIT_BAR_LOW 0x0069 +#define regUVD_JMI2_UVD_LMI_JRBC_RB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH 0x006a +#define regUVD_JMI2_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW 0x006b +#define regUVD_JMI2_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH 0x006c +#define regUVD_JMI2_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JPEG_PREEMPT_VMID 0x006d +#define regUVD_JMI2_UVD_LMI_JPEG_PREEMPT_VMID_BASE_IDX 0 +#define regUVD_JMI2_UVD_JMI_DEC_SWAP_CNTL 0x006e +#define regUVD_JMI2_UVD_JMI_DEC_SWAP_CNTL_BASE_IDX 0 +#define regUVD_JMI2_UVD_JMI_ATOMIC_CNTL 0x006f +#define regUVD_JMI2_UVD_JMI_ATOMIC_CNTL_BASE_IDX 0 +#define regUVD_JMI2_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW 0x0070 +#define regUVD_JMI2_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI2_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH 0x0071 +#define regUVD_JMI2_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JPEG_READ_64BIT_BAR_LOW 0x0072 +#define regUVD_JMI2_UVD_LMI_JPEG_READ_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH 0x0073 +#define regUVD_JMI2_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW 0x0074 +#define regUVD_JMI2_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH 0x0075 +#define regUVD_JMI2_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JRBC_IB_64BIT_BAR_LOW 0x0076 +#define regUVD_JMI2_UVD_LMI_JRBC_IB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH 0x0077 +#define regUVD_JMI2_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW 0x0078 +#define regUVD_JMI2_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI2_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH 0x0079 +#define regUVD_JMI2_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI2_UVD_JMI_ATOMIC_CNTL2 0x007d +#define regUVD_JMI2_UVD_JMI_ATOMIC_CNTL2_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jmi3_uvd_jmi_dec +// base address: 0x1e280 +#define regUVD_JMI3_UVD_JPEG_DEC_PF_CTRL 0x00a0 +#define regUVD_JMI3_UVD_JPEG_DEC_PF_CTRL_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JRBC_CTRL 0x00a1 +#define regUVD_JMI3_UVD_LMI_JRBC_CTRL_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JPEG_CTRL 0x00a2 +#define regUVD_JMI3_UVD_LMI_JPEG_CTRL_BASE_IDX 0 +#define regUVD_JMI3_JPEG_LMI_DROP 0x00a3 +#define regUVD_JMI3_JPEG_LMI_DROP_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JRBC_IB_VMID 0x00a4 +#define regUVD_JMI3_UVD_LMI_JRBC_IB_VMID_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JRBC_RB_VMID 0x00a5 +#define regUVD_JMI3_UVD_LMI_JRBC_RB_VMID_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JPEG_VMID 0x00a6 +#define regUVD_JMI3_UVD_LMI_JPEG_VMID_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW 0x00a7 +#define regUVD_JMI3_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH 0x00a8 +#define regUVD_JMI3_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JRBC_RB_64BIT_BAR_LOW 0x00a9 +#define regUVD_JMI3_UVD_LMI_JRBC_RB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH 0x00aa +#define regUVD_JMI3_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW 0x00ab +#define regUVD_JMI3_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH 0x00ac +#define regUVD_JMI3_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JPEG_PREEMPT_VMID 0x00ad +#define regUVD_JMI3_UVD_LMI_JPEG_PREEMPT_VMID_BASE_IDX 0 +#define regUVD_JMI3_UVD_JMI_DEC_SWAP_CNTL 0x00ae +#define regUVD_JMI3_UVD_JMI_DEC_SWAP_CNTL_BASE_IDX 0 +#define regUVD_JMI3_UVD_JMI_ATOMIC_CNTL 0x00af +#define regUVD_JMI3_UVD_JMI_ATOMIC_CNTL_BASE_IDX 0 +#define regUVD_JMI3_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW 0x00b0 +#define regUVD_JMI3_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI3_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH 0x00b1 +#define regUVD_JMI3_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JPEG_READ_64BIT_BAR_LOW 0x00b2 +#define regUVD_JMI3_UVD_LMI_JPEG_READ_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH 0x00b3 +#define regUVD_JMI3_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW 0x00b4 +#define regUVD_JMI3_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH 0x00b5 +#define regUVD_JMI3_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JRBC_IB_64BIT_BAR_LOW 0x00b6 +#define regUVD_JMI3_UVD_LMI_JRBC_IB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH 0x00b7 +#define regUVD_JMI3_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW 0x00b8 +#define regUVD_JMI3_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI3_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH 0x00b9 +#define regUVD_JMI3_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI3_UVD_JMI_ATOMIC_CNTL2 0x00bd +#define regUVD_JMI3_UVD_JMI_ATOMIC_CNTL2_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jmi4_uvd_jmi_dec +// base address: 0x1e380 +#define regUVD_JMI4_UVD_JPEG_DEC_PF_CTRL 0x00e0 +#define regUVD_JMI4_UVD_JPEG_DEC_PF_CTRL_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JRBC_CTRL 0x00e1 +#define regUVD_JMI4_UVD_LMI_JRBC_CTRL_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JPEG_CTRL 0x00e2 +#define regUVD_JMI4_UVD_LMI_JPEG_CTRL_BASE_IDX 0 +#define regUVD_JMI4_JPEG_LMI_DROP 0x00e3 +#define regUVD_JMI4_JPEG_LMI_DROP_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JRBC_IB_VMID 0x00e4 +#define regUVD_JMI4_UVD_LMI_JRBC_IB_VMID_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JRBC_RB_VMID 0x00e5 +#define regUVD_JMI4_UVD_LMI_JRBC_RB_VMID_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JPEG_VMID 0x00e6 +#define regUVD_JMI4_UVD_LMI_JPEG_VMID_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW 0x00e7 +#define regUVD_JMI4_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH 0x00e8 +#define regUVD_JMI4_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JRBC_RB_64BIT_BAR_LOW 0x00e9 +#define regUVD_JMI4_UVD_LMI_JRBC_RB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH 0x00ea +#define regUVD_JMI4_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW 0x00eb +#define regUVD_JMI4_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH 0x00ec +#define regUVD_JMI4_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JPEG_PREEMPT_VMID 0x00ed +#define regUVD_JMI4_UVD_LMI_JPEG_PREEMPT_VMID_BASE_IDX 0 +#define regUVD_JMI4_UVD_JMI_DEC_SWAP_CNTL 0x00ee +#define regUVD_JMI4_UVD_JMI_DEC_SWAP_CNTL_BASE_IDX 0 +#define regUVD_JMI4_UVD_JMI_ATOMIC_CNTL 0x00ef +#define regUVD_JMI4_UVD_JMI_ATOMIC_CNTL_BASE_IDX 0 +#define regUVD_JMI4_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW 0x00f0 +#define regUVD_JMI4_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI4_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH 0x00f1 +#define regUVD_JMI4_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JPEG_READ_64BIT_BAR_LOW 0x00f2 +#define regUVD_JMI4_UVD_LMI_JPEG_READ_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH 0x00f3 +#define regUVD_JMI4_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW 0x00f4 +#define regUVD_JMI4_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH 0x00f5 +#define regUVD_JMI4_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JRBC_IB_64BIT_BAR_LOW 0x00f6 +#define regUVD_JMI4_UVD_LMI_JRBC_IB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH 0x00f7 +#define regUVD_JMI4_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW 0x00f8 +#define regUVD_JMI4_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI4_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH 0x00f9 +#define regUVD_JMI4_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI4_UVD_JMI_ATOMIC_CNTL2 0x00fd +#define regUVD_JMI4_UVD_JMI_ATOMIC_CNTL2_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jmi5_uvd_jmi_dec +// base address: 0x1e480 +#define regUVD_JMI5_UVD_JPEG_DEC_PF_CTRL 0x0120 +#define regUVD_JMI5_UVD_JPEG_DEC_PF_CTRL_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JRBC_CTRL 0x0121 +#define regUVD_JMI5_UVD_LMI_JRBC_CTRL_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JPEG_CTRL 0x0122 +#define regUVD_JMI5_UVD_LMI_JPEG_CTRL_BASE_IDX 0 +#define regUVD_JMI5_JPEG_LMI_DROP 0x0123 +#define regUVD_JMI5_JPEG_LMI_DROP_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JRBC_IB_VMID 0x0124 +#define regUVD_JMI5_UVD_LMI_JRBC_IB_VMID_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JRBC_RB_VMID 0x0125 +#define regUVD_JMI5_UVD_LMI_JRBC_RB_VMID_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JPEG_VMID 0x0126 +#define regUVD_JMI5_UVD_LMI_JPEG_VMID_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW 0x0127 +#define regUVD_JMI5_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH 0x0128 +#define regUVD_JMI5_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JRBC_RB_64BIT_BAR_LOW 0x0129 +#define regUVD_JMI5_UVD_LMI_JRBC_RB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH 0x012a +#define regUVD_JMI5_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW 0x012b +#define regUVD_JMI5_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH 0x012c +#define regUVD_JMI5_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JPEG_PREEMPT_VMID 0x012d +#define regUVD_JMI5_UVD_LMI_JPEG_PREEMPT_VMID_BASE_IDX 0 +#define regUVD_JMI5_UVD_JMI_DEC_SWAP_CNTL 0x012e +#define regUVD_JMI5_UVD_JMI_DEC_SWAP_CNTL_BASE_IDX 0 +#define regUVD_JMI5_UVD_JMI_ATOMIC_CNTL 0x012f +#define regUVD_JMI5_UVD_JMI_ATOMIC_CNTL_BASE_IDX 0 +#define regUVD_JMI5_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW 0x0130 +#define regUVD_JMI5_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI5_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH 0x0131 +#define regUVD_JMI5_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JPEG_READ_64BIT_BAR_LOW 0x0132 +#define regUVD_JMI5_UVD_LMI_JPEG_READ_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH 0x0133 +#define regUVD_JMI5_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW 0x0134 +#define regUVD_JMI5_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH 0x0135 +#define regUVD_JMI5_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JRBC_IB_64BIT_BAR_LOW 0x0136 +#define regUVD_JMI5_UVD_LMI_JRBC_IB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH 0x0137 +#define regUVD_JMI5_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW 0x0138 +#define regUVD_JMI5_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI5_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH 0x0139 +#define regUVD_JMI5_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI5_UVD_JMI_ATOMIC_CNTL2 0x013d +#define regUVD_JMI5_UVD_JMI_ATOMIC_CNTL2_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jmi6_uvd_jmi_dec +// base address: 0x1e580 +#define regUVD_JMI6_UVD_JPEG_DEC_PF_CTRL 0x0160 +#define regUVD_JMI6_UVD_JPEG_DEC_PF_CTRL_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JRBC_CTRL 0x0161 +#define regUVD_JMI6_UVD_LMI_JRBC_CTRL_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JPEG_CTRL 0x0162 +#define regUVD_JMI6_UVD_LMI_JPEG_CTRL_BASE_IDX 0 +#define regUVD_JMI6_JPEG_LMI_DROP 0x0163 +#define regUVD_JMI6_JPEG_LMI_DROP_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JRBC_IB_VMID 0x0164 +#define regUVD_JMI6_UVD_LMI_JRBC_IB_VMID_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JRBC_RB_VMID 0x0165 +#define regUVD_JMI6_UVD_LMI_JRBC_RB_VMID_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JPEG_VMID 0x0166 +#define regUVD_JMI6_UVD_LMI_JPEG_VMID_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW 0x0167 +#define regUVD_JMI6_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH 0x0168 +#define regUVD_JMI6_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JRBC_RB_64BIT_BAR_LOW 0x0169 +#define regUVD_JMI6_UVD_LMI_JRBC_RB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH 0x016a +#define regUVD_JMI6_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW 0x016b +#define regUVD_JMI6_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH 0x016c +#define regUVD_JMI6_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JPEG_PREEMPT_VMID 0x016d +#define regUVD_JMI6_UVD_LMI_JPEG_PREEMPT_VMID_BASE_IDX 0 +#define regUVD_JMI6_UVD_JMI_DEC_SWAP_CNTL 0x016e +#define regUVD_JMI6_UVD_JMI_DEC_SWAP_CNTL_BASE_IDX 0 +#define regUVD_JMI6_UVD_JMI_ATOMIC_CNTL 0x016f +#define regUVD_JMI6_UVD_JMI_ATOMIC_CNTL_BASE_IDX 0 +#define regUVD_JMI6_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW 0x0170 +#define regUVD_JMI6_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI6_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH 0x0171 +#define regUVD_JMI6_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JPEG_READ_64BIT_BAR_LOW 0x0172 +#define regUVD_JMI6_UVD_LMI_JPEG_READ_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH 0x0173 +#define regUVD_JMI6_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW 0x0174 +#define regUVD_JMI6_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH 0x0175 +#define regUVD_JMI6_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JRBC_IB_64BIT_BAR_LOW 0x0176 +#define regUVD_JMI6_UVD_LMI_JRBC_IB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH 0x0177 +#define regUVD_JMI6_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW 0x0178 +#define regUVD_JMI6_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI6_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH 0x0179 +#define regUVD_JMI6_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI6_UVD_JMI_ATOMIC_CNTL2 0x017d +#define regUVD_JMI6_UVD_JMI_ATOMIC_CNTL2_BASE_IDX 0 + + +// addressBlock: aid_uvd0_uvd_jmi7_uvd_jmi_dec +// base address: 0x1e680 +#define regUVD_JMI7_UVD_JPEG_DEC_PF_CTRL 0x01a0 +#define regUVD_JMI7_UVD_JPEG_DEC_PF_CTRL_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JRBC_CTRL 0x01a1 +#define regUVD_JMI7_UVD_LMI_JRBC_CTRL_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JPEG_CTRL 0x01a2 +#define regUVD_JMI7_UVD_LMI_JPEG_CTRL_BASE_IDX 0 +#define regUVD_JMI7_JPEG_LMI_DROP 0x01a3 +#define regUVD_JMI7_JPEG_LMI_DROP_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JRBC_IB_VMID 0x01a4 +#define regUVD_JMI7_UVD_LMI_JRBC_IB_VMID_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JRBC_RB_VMID 0x01a5 +#define regUVD_JMI7_UVD_LMI_JRBC_RB_VMID_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JPEG_VMID 0x01a6 +#define regUVD_JMI7_UVD_LMI_JPEG_VMID_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW 0x01a7 +#define regUVD_JMI7_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH 0x01a8 +#define regUVD_JMI7_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JRBC_RB_64BIT_BAR_LOW 0x01a9 +#define regUVD_JMI7_UVD_LMI_JRBC_RB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH 0x01aa +#define regUVD_JMI7_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW 0x01ab +#define regUVD_JMI7_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH 0x01ac +#define regUVD_JMI7_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JPEG_PREEMPT_VMID 0x01ad +#define regUVD_JMI7_UVD_LMI_JPEG_PREEMPT_VMID_BASE_IDX 0 +#define regUVD_JMI7_UVD_JMI_DEC_SWAP_CNTL 0x01ae +#define regUVD_JMI7_UVD_JMI_DEC_SWAP_CNTL_BASE_IDX 0 +#define regUVD_JMI7_UVD_JMI_ATOMIC_CNTL 0x01af +#define regUVD_JMI7_UVD_JMI_ATOMIC_CNTL_BASE_IDX 0 +#define regUVD_JMI7_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW 0x01b0 +#define regUVD_JMI7_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI7_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH 0x01b1 +#define regUVD_JMI7_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JPEG_READ_64BIT_BAR_LOW 0x01b2 +#define regUVD_JMI7_UVD_LMI_JPEG_READ_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH 0x01b3 +#define regUVD_JMI7_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW 0x01b4 +#define regUVD_JMI7_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH 0x01b5 +#define regUVD_JMI7_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JRBC_IB_64BIT_BAR_LOW 0x01b6 +#define regUVD_JMI7_UVD_LMI_JRBC_IB_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH 0x01b7 +#define regUVD_JMI7_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW 0x01b8 +#define regUVD_JMI7_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW_BASE_IDX 0 +#define regUVD_JMI7_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH 0x01b9 +#define regUVD_JMI7_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH_BASE_IDX 0 +#define regUVD_JMI7_UVD_JMI_ATOMIC_CNTL2 0x01bd +#define regUVD_JMI7_UVD_JMI_ATOMIC_CNTL2_BASE_IDX 0 + + +// addressBlock: uvdctxind +// base address: 0x0 +#define ixUVD_CGC_MEM_CTRL 0x0000 +#define ixUVD_CGC_CTRL2 0x0001 +#define ixUVD_CGC_MEM_DS_CTRL 0x0002 +#define ixUVD_CGC_MEM_SD_CTRL 0x0003 +#define ixUVD_SW_SCRATCH_00 0x0004 +#define ixUVD_SW_SCRATCH_01 0x0005 +#define ixUVD_SW_SCRATCH_02 0x0006 +#define ixUVD_SW_SCRATCH_03 0x0007 +#define ixUVD_SW_SCRATCH_04 0x0008 +#define ixUVD_SW_SCRATCH_05 0x0009 +#define ixUVD_SW_SCRATCH_06 0x000a +#define ixUVD_SW_SCRATCH_07 0x000b +#define ixUVD_SW_SCRATCH_08 0x000c +#define ixUVD_SW_SCRATCH_09 0x000d +#define ixUVD_SW_SCRATCH_10 0x000e +#define ixUVD_SW_SCRATCH_11 0x000f +#define ixUVD_SW_SCRATCH_12 0x0010 +#define ixUVD_SW_SCRATCH_13 0x0011 +#define ixUVD_SW_SCRATCH_14 0x0012 +#define ixUVD_SW_SCRATCH_15 0x0013 +#define ixUVD_IH_SEM_CTRL 0x001e + + +// addressBlock: lmi_adp_indirect +// base address: 0x0 +#define ixUVD_LMI_CRC0 0x0000 +#define ixUVD_LMI_CRC1 0x0001 +#define ixUVD_LMI_CRC2 0x0002 +#define ixUVD_LMI_CRC3 0x0003 +#define ixUVD_LMI_CRC10 0x000a +#define ixUVD_LMI_CRC11 0x000b +#define ixUVD_LMI_CRC12 0x000c +#define ixUVD_LMI_CRC13 0x000d +#define ixUVD_LMI_CRC14 0x000e +#define ixUVD_LMI_CRC15 0x000f +#define ixUVD_LMI_SWAP_CNTL2 0x0029 +#define ixUVD_MEMCHECK_SYS_INT_EN 0x0134 +#define ixUVD_MEMCHECK_SYS_INT_STAT 0x0135 +#define ixUVD_MEMCHECK_SYS_INT_ACK 0x0136 +#define ixUVD_MEMCHECK_VCPU_INT_EN 0x0137 +#define ixUVD_MEMCHECK_VCPU_INT_STAT 0x0138 +#define ixUVD_MEMCHECK_VCPU_INT_ACK 0x0139 +#define ixUVD_MEMCHECK2_SYS_INT_STAT 0x0140 +#define ixUVD_MEMCHECK2_SYS_INT_ACK 0x0141 +#define ixUVD_MEMCHECK2_VCPU_INT_STAT 0x0142 +#define ixUVD_MEMCHECK2_VCPU_INT_ACK 0x0143 + + +#endif diff --git a/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_4_0_3_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_4_0_3_sh_mask.h new file mode 100644 index 000000000000..5bd8111bf04a --- /dev/null +++ b/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_4_0_3_sh_mask.h @@ -0,0 +1,10919 @@ +/* + * Copyright 2022 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#ifndef _vcn_4_0_3_SH_MASK_HEADER +#define _vcn_4_0_3_SH_MASK_HEADER + + +// addressBlock: aid_uvd0_uvddec +//UVD_TOP_CTRL +#define UVD_TOP_CTRL__STANDARD__SHIFT 0x0 +#define UVD_TOP_CTRL__STD_VERSION__SHIFT 0x4 +#define UVD_TOP_CTRL__STANDARD_MASK 0x0000000FL +#define UVD_TOP_CTRL__STD_VERSION_MASK 0x00000010L +//UVD_CGC_GATE +#define UVD_CGC_GATE__SYS__SHIFT 0x0 +#define UVD_CGC_GATE__UDEC__SHIFT 0x1 +#define UVD_CGC_GATE__MPEG2__SHIFT 0x2 +#define UVD_CGC_GATE__REGS__SHIFT 0x3 +#define UVD_CGC_GATE__RBC__SHIFT 0x4 +#define UVD_CGC_GATE__LMI_MC__SHIFT 0x5 +#define UVD_CGC_GATE__LMI_UMC__SHIFT 0x6 +#define UVD_CGC_GATE__IDCT__SHIFT 0x7 +#define UVD_CGC_GATE__MPRD__SHIFT 0x8 +#define UVD_CGC_GATE__MPC__SHIFT 0x9 +#define UVD_CGC_GATE__LBSI__SHIFT 0xa +#define UVD_CGC_GATE__LRBBM__SHIFT 0xb +#define UVD_CGC_GATE__UDEC_RE__SHIFT 0xc +#define UVD_CGC_GATE__UDEC_CM__SHIFT 0xd +#define UVD_CGC_GATE__UDEC_IT__SHIFT 0xe +#define UVD_CGC_GATE__UDEC_DB__SHIFT 0xf +#define UVD_CGC_GATE__UDEC_MP__SHIFT 0x10 +#define UVD_CGC_GATE__WCB__SHIFT 0x11 +#define UVD_CGC_GATE__VCPU__SHIFT 0x12 +#define UVD_CGC_GATE__MMSCH__SHIFT 0x14 +#define UVD_CGC_GATE__LCM0__SHIFT 0x15 +#define UVD_CGC_GATE__LCM1__SHIFT 0x16 +#define UVD_CGC_GATE__MIF__SHIFT 0x17 +#define UVD_CGC_GATE__VREG__SHIFT 0x18 +#define UVD_CGC_GATE__PE__SHIFT 0x19 +#define UVD_CGC_GATE__PPU__SHIFT 0x1a +#define UVD_CGC_GATE__SYS_MASK 0x00000001L +#define UVD_CGC_GATE__UDEC_MASK 0x00000002L +#define UVD_CGC_GATE__MPEG2_MASK 0x00000004L +#define UVD_CGC_GATE__REGS_MASK 0x00000008L +#define UVD_CGC_GATE__RBC_MASK 0x00000010L +#define UVD_CGC_GATE__LMI_MC_MASK 0x00000020L +#define UVD_CGC_GATE__LMI_UMC_MASK 0x00000040L +#define UVD_CGC_GATE__IDCT_MASK 0x00000080L +#define UVD_CGC_GATE__MPRD_MASK 0x00000100L +#define UVD_CGC_GATE__MPC_MASK 0x00000200L +#define UVD_CGC_GATE__LBSI_MASK 0x00000400L +#define UVD_CGC_GATE__LRBBM_MASK 0x00000800L +#define UVD_CGC_GATE__UDEC_RE_MASK 0x00001000L +#define UVD_CGC_GATE__UDEC_CM_MASK 0x00002000L +#define UVD_CGC_GATE__UDEC_IT_MASK 0x00004000L +#define UVD_CGC_GATE__UDEC_DB_MASK 0x00008000L +#define UVD_CGC_GATE__UDEC_MP_MASK 0x00010000L +#define UVD_CGC_GATE__WCB_MASK 0x00020000L +#define UVD_CGC_GATE__VCPU_MASK 0x00040000L +#define UVD_CGC_GATE__MMSCH_MASK 0x00100000L +#define UVD_CGC_GATE__LCM0_MASK 0x00200000L +#define UVD_CGC_GATE__LCM1_MASK 0x00400000L +#define UVD_CGC_GATE__MIF_MASK 0x00800000L +#define UVD_CGC_GATE__VREG_MASK 0x01000000L +#define UVD_CGC_GATE__PE_MASK 0x02000000L +#define UVD_CGC_GATE__PPU_MASK 0x04000000L +//UVD_CGC_CTRL +#define UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT 0x0 +#define UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT 0x2 +#define UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT 0x6 +#define UVD_CGC_CTRL__UDEC_RE_MODE__SHIFT 0xb +#define UVD_CGC_CTRL__UDEC_CM_MODE__SHIFT 0xc +#define UVD_CGC_CTRL__UDEC_IT_MODE__SHIFT 0xd +#define UVD_CGC_CTRL__UDEC_DB_MODE__SHIFT 0xe +#define UVD_CGC_CTRL__UDEC_MP_MODE__SHIFT 0xf +#define UVD_CGC_CTRL__SYS_MODE__SHIFT 0x10 +#define UVD_CGC_CTRL__UDEC_MODE__SHIFT 0x11 +#define UVD_CGC_CTRL__MPEG2_MODE__SHIFT 0x12 +#define UVD_CGC_CTRL__REGS_MODE__SHIFT 0x13 +#define UVD_CGC_CTRL__RBC_MODE__SHIFT 0x14 +#define UVD_CGC_CTRL__LMI_MC_MODE__SHIFT 0x15 +#define UVD_CGC_CTRL__LMI_UMC_MODE__SHIFT 0x16 +#define UVD_CGC_CTRL__IDCT_MODE__SHIFT 0x17 +#define UVD_CGC_CTRL__MPRD_MODE__SHIFT 0x18 +#define UVD_CGC_CTRL__MPC_MODE__SHIFT 0x19 +#define UVD_CGC_CTRL__LBSI_MODE__SHIFT 0x1a +#define UVD_CGC_CTRL__LRBBM_MODE__SHIFT 0x1b +#define UVD_CGC_CTRL__WCB_MODE__SHIFT 0x1c +#define UVD_CGC_CTRL__VCPU_MODE__SHIFT 0x1d +#define UVD_CGC_CTRL__MMSCH_MODE__SHIFT 0x1f +#define UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK 0x00000001L +#define UVD_CGC_CTRL__CLK_GATE_DLY_TIMER_MASK 0x0000003CL +#define UVD_CGC_CTRL__CLK_OFF_DELAY_MASK 0x000007C0L +#define UVD_CGC_CTRL__UDEC_RE_MODE_MASK 0x00000800L +#define UVD_CGC_CTRL__UDEC_CM_MODE_MASK 0x00001000L +#define UVD_CGC_CTRL__UDEC_IT_MODE_MASK 0x00002000L +#define UVD_CGC_CTRL__UDEC_DB_MODE_MASK 0x00004000L +#define UVD_CGC_CTRL__UDEC_MP_MODE_MASK 0x00008000L +#define UVD_CGC_CTRL__SYS_MODE_MASK 0x00010000L +#define UVD_CGC_CTRL__UDEC_MODE_MASK 0x00020000L +#define UVD_CGC_CTRL__MPEG2_MODE_MASK 0x00040000L +#define UVD_CGC_CTRL__REGS_MODE_MASK 0x00080000L +#define UVD_CGC_CTRL__RBC_MODE_MASK 0x00100000L +#define UVD_CGC_CTRL__LMI_MC_MODE_MASK 0x00200000L +#define UVD_CGC_CTRL__LMI_UMC_MODE_MASK 0x00400000L +#define UVD_CGC_CTRL__IDCT_MODE_MASK 0x00800000L +#define UVD_CGC_CTRL__MPRD_MODE_MASK 0x01000000L +#define UVD_CGC_CTRL__MPC_MODE_MASK 0x02000000L +#define UVD_CGC_CTRL__LBSI_MODE_MASK 0x04000000L +#define UVD_CGC_CTRL__LRBBM_MODE_MASK 0x08000000L +#define UVD_CGC_CTRL__WCB_MODE_MASK 0x10000000L +#define UVD_CGC_CTRL__VCPU_MODE_MASK 0x20000000L +#define UVD_CGC_CTRL__MMSCH_MODE_MASK 0x80000000L +//AVM_SUVD_CGC_GATE +#define AVM_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define AVM_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define AVM_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define AVM_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define AVM_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define AVM_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define AVM_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define AVM_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define AVM_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define AVM_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define AVM_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define AVM_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define AVM_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define AVM_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define AVM_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define AVM_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define AVM_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define AVM_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define AVM_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define AVM_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define AVM_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define AVM_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define AVM_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define AVM_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define AVM_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define AVM_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define AVM_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define AVM_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define AVM_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define AVM_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define AVM_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define AVM_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define AVM_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define AVM_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define AVM_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define AVM_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define AVM_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define AVM_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define AVM_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define AVM_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define AVM_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define AVM_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define AVM_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define AVM_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define AVM_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define AVM_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define AVM_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define AVM_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define AVM_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define AVM_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define AVM_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define AVM_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define AVM_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define AVM_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define AVM_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define AVM_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define AVM_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define AVM_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define AVM_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define AVM_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define AVM_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define AVM_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define AVM_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define AVM_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//CDEFE_SUVD_CGC_GATE +#define CDEFE_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define CDEFE_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define CDEFE_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define CDEFE_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define CDEFE_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define CDEFE_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define CDEFE_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define CDEFE_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define CDEFE_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define CDEFE_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define CDEFE_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define CDEFE_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define CDEFE_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define CDEFE_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define CDEFE_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define CDEFE_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define CDEFE_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define CDEFE_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define CDEFE_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define CDEFE_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define CDEFE_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define CDEFE_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define CDEFE_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define CDEFE_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define CDEFE_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define CDEFE_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define CDEFE_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define CDEFE_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define CDEFE_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define CDEFE_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define CDEFE_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define CDEFE_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define CDEFE_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define CDEFE_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define CDEFE_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define CDEFE_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define CDEFE_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define CDEFE_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define CDEFE_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define CDEFE_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define CDEFE_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define CDEFE_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define CDEFE_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define CDEFE_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define CDEFE_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define CDEFE_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define CDEFE_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define CDEFE_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define CDEFE_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define CDEFE_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define CDEFE_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define CDEFE_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define CDEFE_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define CDEFE_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define CDEFE_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define CDEFE_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define CDEFE_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define CDEFE_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define CDEFE_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define CDEFE_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define CDEFE_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define CDEFE_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define CDEFE_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define CDEFE_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//EFC_SUVD_CGC_GATE +#define EFC_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define EFC_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define EFC_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define EFC_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define EFC_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define EFC_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define EFC_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define EFC_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define EFC_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define EFC_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define EFC_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define EFC_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define EFC_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define EFC_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define EFC_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define EFC_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define EFC_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define EFC_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define EFC_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define EFC_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define EFC_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define EFC_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define EFC_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define EFC_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define EFC_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define EFC_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define EFC_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define EFC_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define EFC_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define EFC_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define EFC_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define EFC_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define EFC_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define EFC_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define EFC_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define EFC_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define EFC_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define EFC_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define EFC_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define EFC_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define EFC_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define EFC_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define EFC_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define EFC_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define EFC_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define EFC_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define EFC_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define EFC_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define EFC_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define EFC_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define EFC_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define EFC_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define EFC_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define EFC_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define EFC_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define EFC_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define EFC_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define EFC_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define EFC_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define EFC_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define EFC_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define EFC_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define EFC_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define EFC_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//ENT_SUVD_CGC_GATE +#define ENT_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define ENT_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define ENT_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define ENT_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define ENT_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define ENT_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define ENT_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define ENT_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define ENT_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define ENT_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define ENT_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define ENT_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define ENT_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define ENT_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define ENT_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define ENT_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define ENT_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define ENT_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define ENT_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define ENT_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define ENT_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define ENT_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define ENT_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define ENT_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define ENT_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define ENT_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define ENT_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define ENT_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define ENT_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define ENT_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define ENT_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define ENT_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define ENT_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define ENT_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define ENT_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define ENT_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define ENT_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define ENT_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define ENT_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define ENT_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define ENT_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define ENT_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define ENT_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define ENT_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define ENT_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define ENT_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define ENT_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define ENT_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define ENT_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define ENT_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define ENT_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define ENT_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define ENT_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define ENT_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define ENT_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define ENT_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define ENT_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define ENT_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define ENT_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define ENT_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define ENT_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define ENT_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define ENT_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define ENT_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//IME_SUVD_CGC_GATE +#define IME_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define IME_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define IME_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define IME_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define IME_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define IME_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define IME_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define IME_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define IME_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define IME_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define IME_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define IME_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define IME_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define IME_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define IME_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define IME_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define IME_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define IME_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define IME_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define IME_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define IME_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define IME_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define IME_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define IME_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define IME_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define IME_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define IME_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define IME_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define IME_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define IME_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define IME_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define IME_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define IME_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define IME_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define IME_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define IME_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define IME_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define IME_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define IME_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define IME_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define IME_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define IME_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define IME_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define IME_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define IME_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define IME_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define IME_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define IME_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define IME_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define IME_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define IME_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define IME_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define IME_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define IME_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define IME_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define IME_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define IME_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define IME_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define IME_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define IME_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define IME_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define IME_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define IME_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define IME_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//PPU_SUVD_CGC_GATE +#define PPU_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define PPU_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define PPU_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define PPU_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define PPU_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define PPU_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define PPU_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define PPU_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define PPU_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define PPU_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define PPU_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define PPU_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define PPU_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define PPU_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define PPU_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define PPU_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define PPU_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define PPU_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define PPU_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define PPU_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define PPU_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define PPU_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define PPU_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define PPU_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define PPU_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define PPU_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define PPU_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define PPU_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define PPU_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define PPU_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define PPU_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define PPU_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define PPU_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define PPU_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define PPU_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define PPU_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define PPU_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define PPU_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define PPU_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define PPU_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define PPU_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define PPU_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define PPU_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define PPU_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define PPU_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define PPU_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define PPU_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define PPU_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define PPU_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define PPU_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define PPU_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define PPU_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define PPU_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define PPU_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define PPU_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define PPU_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define PPU_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define PPU_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define PPU_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define PPU_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define PPU_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define PPU_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define PPU_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define PPU_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//SAOE_SUVD_CGC_GATE +#define SAOE_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define SAOE_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define SAOE_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define SAOE_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define SAOE_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define SAOE_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define SAOE_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define SAOE_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define SAOE_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define SAOE_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define SAOE_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define SAOE_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define SAOE_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define SAOE_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define SAOE_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define SAOE_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define SAOE_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define SAOE_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define SAOE_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define SAOE_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define SAOE_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define SAOE_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define SAOE_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define SAOE_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define SAOE_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define SAOE_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define SAOE_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define SAOE_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define SAOE_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define SAOE_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define SAOE_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define SAOE_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define SAOE_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define SAOE_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define SAOE_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define SAOE_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define SAOE_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define SAOE_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define SAOE_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define SAOE_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define SAOE_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define SAOE_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define SAOE_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define SAOE_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define SAOE_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define SAOE_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define SAOE_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define SAOE_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define SAOE_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define SAOE_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define SAOE_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define SAOE_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define SAOE_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define SAOE_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define SAOE_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define SAOE_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define SAOE_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define SAOE_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define SAOE_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define SAOE_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define SAOE_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define SAOE_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define SAOE_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define SAOE_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//SCM_SUVD_CGC_GATE +#define SCM_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define SCM_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define SCM_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define SCM_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define SCM_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define SCM_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define SCM_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define SCM_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define SCM_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define SCM_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define SCM_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define SCM_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define SCM_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define SCM_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define SCM_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define SCM_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define SCM_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define SCM_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define SCM_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define SCM_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define SCM_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define SCM_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define SCM_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define SCM_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define SCM_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define SCM_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define SCM_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define SCM_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define SCM_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define SCM_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define SCM_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define SCM_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define SCM_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define SCM_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define SCM_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define SCM_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define SCM_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define SCM_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define SCM_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define SCM_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define SCM_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define SCM_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define SCM_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define SCM_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define SCM_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define SCM_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define SCM_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define SCM_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define SCM_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define SCM_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define SCM_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define SCM_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define SCM_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define SCM_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define SCM_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define SCM_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define SCM_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define SCM_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define SCM_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define SCM_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define SCM_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define SCM_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define SCM_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define SCM_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//SDB_SUVD_CGC_GATE +#define SDB_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define SDB_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define SDB_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define SDB_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define SDB_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define SDB_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define SDB_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define SDB_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define SDB_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define SDB_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define SDB_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define SDB_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define SDB_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define SDB_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define SDB_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define SDB_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define SDB_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define SDB_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define SDB_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define SDB_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define SDB_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define SDB_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define SDB_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define SDB_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define SDB_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define SDB_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define SDB_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define SDB_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define SDB_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define SDB_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define SDB_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define SDB_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define SDB_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define SDB_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define SDB_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define SDB_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define SDB_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define SDB_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define SDB_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define SDB_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define SDB_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define SDB_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define SDB_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define SDB_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define SDB_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define SDB_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define SDB_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define SDB_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define SDB_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define SDB_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define SDB_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define SDB_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define SDB_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define SDB_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define SDB_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define SDB_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define SDB_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define SDB_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define SDB_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define SDB_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define SDB_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define SDB_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define SDB_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define SDB_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//SIT0_NXT_SUVD_CGC_GATE +#define SIT0_NXT_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define SIT0_NXT_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define SIT0_NXT_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define SIT0_NXT_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define SIT0_NXT_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define SIT0_NXT_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define SIT0_NXT_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define SIT0_NXT_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define SIT0_NXT_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define SIT0_NXT_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define SIT0_NXT_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define SIT0_NXT_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define SIT0_NXT_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define SIT0_NXT_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define SIT0_NXT_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define SIT0_NXT_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define SIT0_NXT_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define SIT0_NXT_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define SIT0_NXT_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define SIT0_NXT_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define SIT0_NXT_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define SIT0_NXT_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define SIT0_NXT_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define SIT0_NXT_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define SIT0_NXT_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define SIT0_NXT_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define SIT0_NXT_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define SIT0_NXT_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define SIT0_NXT_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define SIT0_NXT_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define SIT0_NXT_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define SIT0_NXT_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define SIT0_NXT_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define SIT0_NXT_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define SIT0_NXT_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define SIT0_NXT_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define SIT0_NXT_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define SIT0_NXT_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define SIT0_NXT_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define SIT0_NXT_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define SIT0_NXT_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define SIT0_NXT_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define SIT0_NXT_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define SIT0_NXT_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define SIT0_NXT_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define SIT0_NXT_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define SIT0_NXT_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define SIT0_NXT_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define SIT0_NXT_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define SIT0_NXT_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define SIT0_NXT_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define SIT0_NXT_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define SIT0_NXT_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define SIT0_NXT_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define SIT0_NXT_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define SIT0_NXT_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define SIT0_NXT_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define SIT0_NXT_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define SIT0_NXT_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define SIT0_NXT_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define SIT0_NXT_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define SIT0_NXT_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define SIT0_NXT_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define SIT0_NXT_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//SIT1_NXT_SUVD_CGC_GATE +#define SIT1_NXT_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define SIT1_NXT_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define SIT1_NXT_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define SIT1_NXT_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define SIT1_NXT_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define SIT1_NXT_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define SIT1_NXT_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define SIT1_NXT_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define SIT1_NXT_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define SIT1_NXT_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define SIT1_NXT_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define SIT1_NXT_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define SIT1_NXT_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define SIT1_NXT_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define SIT1_NXT_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define SIT1_NXT_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define SIT1_NXT_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define SIT1_NXT_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define SIT1_NXT_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define SIT1_NXT_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define SIT1_NXT_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define SIT1_NXT_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define SIT1_NXT_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define SIT1_NXT_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define SIT1_NXT_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define SIT1_NXT_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define SIT1_NXT_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define SIT1_NXT_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define SIT1_NXT_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define SIT1_NXT_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define SIT1_NXT_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define SIT1_NXT_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define SIT1_NXT_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define SIT1_NXT_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define SIT1_NXT_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define SIT1_NXT_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define SIT1_NXT_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define SIT1_NXT_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define SIT1_NXT_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define SIT1_NXT_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define SIT1_NXT_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define SIT1_NXT_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define SIT1_NXT_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define SIT1_NXT_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define SIT1_NXT_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define SIT1_NXT_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define SIT1_NXT_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define SIT1_NXT_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define SIT1_NXT_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define SIT1_NXT_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define SIT1_NXT_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define SIT1_NXT_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define SIT1_NXT_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define SIT1_NXT_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define SIT1_NXT_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define SIT1_NXT_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define SIT1_NXT_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define SIT1_NXT_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define SIT1_NXT_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define SIT1_NXT_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define SIT1_NXT_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define SIT1_NXT_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define SIT1_NXT_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define SIT1_NXT_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//SIT2_NXT_SUVD_CGC_GATE +#define SIT2_NXT_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define SIT2_NXT_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define SIT2_NXT_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define SIT2_NXT_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define SIT2_NXT_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define SIT2_NXT_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define SIT2_NXT_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define SIT2_NXT_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define SIT2_NXT_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define SIT2_NXT_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define SIT2_NXT_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define SIT2_NXT_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define SIT2_NXT_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define SIT2_NXT_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define SIT2_NXT_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define SIT2_NXT_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define SIT2_NXT_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define SIT2_NXT_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define SIT2_NXT_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define SIT2_NXT_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define SIT2_NXT_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define SIT2_NXT_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define SIT2_NXT_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define SIT2_NXT_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define SIT2_NXT_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define SIT2_NXT_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define SIT2_NXT_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define SIT2_NXT_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define SIT2_NXT_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define SIT2_NXT_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define SIT2_NXT_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define SIT2_NXT_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define SIT2_NXT_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define SIT2_NXT_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define SIT2_NXT_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define SIT2_NXT_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define SIT2_NXT_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define SIT2_NXT_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define SIT2_NXT_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define SIT2_NXT_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define SIT2_NXT_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define SIT2_NXT_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define SIT2_NXT_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define SIT2_NXT_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define SIT2_NXT_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define SIT2_NXT_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define SIT2_NXT_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define SIT2_NXT_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define SIT2_NXT_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define SIT2_NXT_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define SIT2_NXT_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define SIT2_NXT_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define SIT2_NXT_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define SIT2_NXT_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define SIT2_NXT_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define SIT2_NXT_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define SIT2_NXT_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define SIT2_NXT_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define SIT2_NXT_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define SIT2_NXT_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define SIT2_NXT_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define SIT2_NXT_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define SIT2_NXT_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define SIT2_NXT_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//SIT_SUVD_CGC_GATE +#define SIT_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define SIT_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define SIT_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define SIT_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define SIT_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define SIT_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define SIT_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define SIT_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define SIT_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define SIT_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define SIT_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define SIT_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define SIT_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define SIT_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define SIT_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define SIT_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define SIT_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define SIT_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define SIT_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define SIT_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define SIT_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define SIT_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define SIT_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define SIT_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define SIT_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define SIT_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define SIT_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define SIT_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define SIT_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define SIT_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define SIT_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define SIT_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define SIT_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define SIT_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define SIT_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define SIT_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define SIT_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define SIT_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define SIT_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define SIT_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define SIT_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define SIT_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define SIT_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define SIT_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define SIT_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define SIT_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define SIT_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define SIT_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define SIT_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define SIT_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define SIT_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define SIT_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define SIT_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define SIT_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define SIT_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define SIT_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define SIT_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define SIT_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define SIT_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define SIT_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define SIT_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define SIT_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define SIT_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define SIT_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//SMPA_SUVD_CGC_GATE +#define SMPA_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define SMPA_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define SMPA_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define SMPA_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define SMPA_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define SMPA_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define SMPA_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define SMPA_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define SMPA_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define SMPA_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define SMPA_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define SMPA_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define SMPA_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define SMPA_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define SMPA_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define SMPA_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define SMPA_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define SMPA_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define SMPA_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define SMPA_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define SMPA_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define SMPA_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define SMPA_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define SMPA_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define SMPA_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define SMPA_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define SMPA_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define SMPA_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define SMPA_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define SMPA_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define SMPA_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define SMPA_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define SMPA_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define SMPA_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define SMPA_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define SMPA_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define SMPA_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define SMPA_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define SMPA_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define SMPA_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define SMPA_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define SMPA_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define SMPA_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define SMPA_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define SMPA_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define SMPA_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define SMPA_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define SMPA_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define SMPA_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define SMPA_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define SMPA_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define SMPA_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define SMPA_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define SMPA_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define SMPA_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define SMPA_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define SMPA_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define SMPA_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define SMPA_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define SMPA_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define SMPA_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define SMPA_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define SMPA_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define SMPA_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//SMP_SUVD_CGC_GATE +#define SMP_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define SMP_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define SMP_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define SMP_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define SMP_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define SMP_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define SMP_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define SMP_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define SMP_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define SMP_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define SMP_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define SMP_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define SMP_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define SMP_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define SMP_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define SMP_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define SMP_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define SMP_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define SMP_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define SMP_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define SMP_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define SMP_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define SMP_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define SMP_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define SMP_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define SMP_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define SMP_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define SMP_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define SMP_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define SMP_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define SMP_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define SMP_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define SMP_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define SMP_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define SMP_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define SMP_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define SMP_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define SMP_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define SMP_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define SMP_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define SMP_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define SMP_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define SMP_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define SMP_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define SMP_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define SMP_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define SMP_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define SMP_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define SMP_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define SMP_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define SMP_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define SMP_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define SMP_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define SMP_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define SMP_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define SMP_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define SMP_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define SMP_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define SMP_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define SMP_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define SMP_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define SMP_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define SMP_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define SMP_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//SRE_SUVD_CGC_GATE +#define SRE_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define SRE_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define SRE_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define SRE_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define SRE_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define SRE_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define SRE_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define SRE_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define SRE_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define SRE_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define SRE_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define SRE_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define SRE_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define SRE_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define SRE_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define SRE_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define SRE_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define SRE_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define SRE_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define SRE_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define SRE_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define SRE_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define SRE_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define SRE_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define SRE_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define SRE_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define SRE_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define SRE_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define SRE_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define SRE_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define SRE_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define SRE_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define SRE_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define SRE_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define SRE_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define SRE_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define SRE_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define SRE_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define SRE_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define SRE_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define SRE_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define SRE_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define SRE_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define SRE_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define SRE_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define SRE_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define SRE_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define SRE_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define SRE_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define SRE_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define SRE_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define SRE_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define SRE_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define SRE_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define SRE_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define SRE_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define SRE_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define SRE_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define SRE_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define SRE_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define SRE_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define SRE_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define SRE_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define SRE_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//UVD_MPBE0_SUVD_CGC_GATE +#define UVD_MPBE0_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define UVD_MPBE0_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define UVD_MPBE0_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define UVD_MPBE0_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define UVD_MPBE0_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define UVD_MPBE0_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define UVD_MPBE0_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define UVD_MPBE0_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define UVD_MPBE0_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define UVD_MPBE0_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define UVD_MPBE0_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define UVD_MPBE0_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define UVD_MPBE0_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define UVD_MPBE0_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define UVD_MPBE0_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define UVD_MPBE0_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define UVD_MPBE0_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define UVD_MPBE0_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define UVD_MPBE0_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define UVD_MPBE0_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define UVD_MPBE0_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define UVD_MPBE0_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define UVD_MPBE0_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define UVD_MPBE0_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define UVD_MPBE0_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define UVD_MPBE0_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define UVD_MPBE0_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define UVD_MPBE0_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define UVD_MPBE0_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define UVD_MPBE0_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define UVD_MPBE0_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define UVD_MPBE0_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define UVD_MPBE0_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define UVD_MPBE0_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define UVD_MPBE0_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define UVD_MPBE0_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define UVD_MPBE0_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define UVD_MPBE0_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define UVD_MPBE0_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define UVD_MPBE0_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define UVD_MPBE0_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define UVD_MPBE0_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define UVD_MPBE0_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define UVD_MPBE0_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define UVD_MPBE0_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define UVD_MPBE0_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define UVD_MPBE0_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define UVD_MPBE0_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define UVD_MPBE0_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define UVD_MPBE0_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define UVD_MPBE0_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define UVD_MPBE0_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define UVD_MPBE0_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define UVD_MPBE0_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define UVD_MPBE0_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define UVD_MPBE0_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define UVD_MPBE0_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define UVD_MPBE0_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define UVD_MPBE0_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define UVD_MPBE0_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define UVD_MPBE0_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define UVD_MPBE0_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define UVD_MPBE0_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define UVD_MPBE0_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//UVD_MPBE1_SUVD_CGC_GATE +#define UVD_MPBE1_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define UVD_MPBE1_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define UVD_MPBE1_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define UVD_MPBE1_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define UVD_MPBE1_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define UVD_MPBE1_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define UVD_MPBE1_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define UVD_MPBE1_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define UVD_MPBE1_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define UVD_MPBE1_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define UVD_MPBE1_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define UVD_MPBE1_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define UVD_MPBE1_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define UVD_MPBE1_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define UVD_MPBE1_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define UVD_MPBE1_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define UVD_MPBE1_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define UVD_MPBE1_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define UVD_MPBE1_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define UVD_MPBE1_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define UVD_MPBE1_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define UVD_MPBE1_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define UVD_MPBE1_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define UVD_MPBE1_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define UVD_MPBE1_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define UVD_MPBE1_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define UVD_MPBE1_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define UVD_MPBE1_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define UVD_MPBE1_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define UVD_MPBE1_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define UVD_MPBE1_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define UVD_MPBE1_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define UVD_MPBE1_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define UVD_MPBE1_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define UVD_MPBE1_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define UVD_MPBE1_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define UVD_MPBE1_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define UVD_MPBE1_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define UVD_MPBE1_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define UVD_MPBE1_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define UVD_MPBE1_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define UVD_MPBE1_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define UVD_MPBE1_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define UVD_MPBE1_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define UVD_MPBE1_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define UVD_MPBE1_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define UVD_MPBE1_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define UVD_MPBE1_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define UVD_MPBE1_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define UVD_MPBE1_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define UVD_MPBE1_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define UVD_MPBE1_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define UVD_MPBE1_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define UVD_MPBE1_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define UVD_MPBE1_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define UVD_MPBE1_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define UVD_MPBE1_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define UVD_MPBE1_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define UVD_MPBE1_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define UVD_MPBE1_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define UVD_MPBE1_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define UVD_MPBE1_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define UVD_MPBE1_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define UVD_MPBE1_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//UVD_SUVD_CGC_GATE +#define UVD_SUVD_CGC_GATE__SRE__SHIFT 0x0 +#define UVD_SUVD_CGC_GATE__SIT__SHIFT 0x1 +#define UVD_SUVD_CGC_GATE__SMP__SHIFT 0x2 +#define UVD_SUVD_CGC_GATE__SCM__SHIFT 0x3 +#define UVD_SUVD_CGC_GATE__SDB__SHIFT 0x4 +#define UVD_SUVD_CGC_GATE__SRE_H264__SHIFT 0x5 +#define UVD_SUVD_CGC_GATE__SRE_HEVC__SHIFT 0x6 +#define UVD_SUVD_CGC_GATE__SIT_H264__SHIFT 0x7 +#define UVD_SUVD_CGC_GATE__SIT_HEVC__SHIFT 0x8 +#define UVD_SUVD_CGC_GATE__SCM_H264__SHIFT 0x9 +#define UVD_SUVD_CGC_GATE__SCM_HEVC__SHIFT 0xa +#define UVD_SUVD_CGC_GATE__SDB_H264__SHIFT 0xb +#define UVD_SUVD_CGC_GATE__SDB_HEVC__SHIFT 0xc +#define UVD_SUVD_CGC_GATE__SCLR__SHIFT 0xd +#define UVD_SUVD_CGC_GATE__UVD_SC__SHIFT 0xe +#define UVD_SUVD_CGC_GATE__ENT__SHIFT 0xf +#define UVD_SUVD_CGC_GATE__IME__SHIFT 0x10 +#define UVD_SUVD_CGC_GATE__SIT_HEVC_DEC__SHIFT 0x11 +#define UVD_SUVD_CGC_GATE__SIT_HEVC_ENC__SHIFT 0x12 +#define UVD_SUVD_CGC_GATE__SITE__SHIFT 0x13 +#define UVD_SUVD_CGC_GATE__SRE_VP9__SHIFT 0x14 +#define UVD_SUVD_CGC_GATE__SCM_VP9__SHIFT 0x15 +#define UVD_SUVD_CGC_GATE__SIT_VP9_DEC__SHIFT 0x16 +#define UVD_SUVD_CGC_GATE__SDB_VP9__SHIFT 0x17 +#define UVD_SUVD_CGC_GATE__IME_HEVC__SHIFT 0x18 +#define UVD_SUVD_CGC_GATE__EFC__SHIFT 0x19 +#define UVD_SUVD_CGC_GATE__SAOE__SHIFT 0x1a +#define UVD_SUVD_CGC_GATE__SRE_AV1__SHIFT 0x1b +#define UVD_SUVD_CGC_GATE__FBC_PCLK__SHIFT 0x1c +#define UVD_SUVD_CGC_GATE__FBC_CCLK__SHIFT 0x1d +#define UVD_SUVD_CGC_GATE__SCM_AV1__SHIFT 0x1e +#define UVD_SUVD_CGC_GATE__SMPA__SHIFT 0x1f +#define UVD_SUVD_CGC_GATE__SRE_MASK 0x00000001L +#define UVD_SUVD_CGC_GATE__SIT_MASK 0x00000002L +#define UVD_SUVD_CGC_GATE__SMP_MASK 0x00000004L +#define UVD_SUVD_CGC_GATE__SCM_MASK 0x00000008L +#define UVD_SUVD_CGC_GATE__SDB_MASK 0x00000010L +#define UVD_SUVD_CGC_GATE__SRE_H264_MASK 0x00000020L +#define UVD_SUVD_CGC_GATE__SRE_HEVC_MASK 0x00000040L +#define UVD_SUVD_CGC_GATE__SIT_H264_MASK 0x00000080L +#define UVD_SUVD_CGC_GATE__SIT_HEVC_MASK 0x00000100L +#define UVD_SUVD_CGC_GATE__SCM_H264_MASK 0x00000200L +#define UVD_SUVD_CGC_GATE__SCM_HEVC_MASK 0x00000400L +#define UVD_SUVD_CGC_GATE__SDB_H264_MASK 0x00000800L +#define UVD_SUVD_CGC_GATE__SDB_HEVC_MASK 0x00001000L +#define UVD_SUVD_CGC_GATE__SCLR_MASK 0x00002000L +#define UVD_SUVD_CGC_GATE__UVD_SC_MASK 0x00004000L +#define UVD_SUVD_CGC_GATE__ENT_MASK 0x00008000L +#define UVD_SUVD_CGC_GATE__IME_MASK 0x00010000L +#define UVD_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK 0x00020000L +#define UVD_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK 0x00040000L +#define UVD_SUVD_CGC_GATE__SITE_MASK 0x00080000L +#define UVD_SUVD_CGC_GATE__SRE_VP9_MASK 0x00100000L +#define UVD_SUVD_CGC_GATE__SCM_VP9_MASK 0x00200000L +#define UVD_SUVD_CGC_GATE__SIT_VP9_DEC_MASK 0x00400000L +#define UVD_SUVD_CGC_GATE__SDB_VP9_MASK 0x00800000L +#define UVD_SUVD_CGC_GATE__IME_HEVC_MASK 0x01000000L +#define UVD_SUVD_CGC_GATE__EFC_MASK 0x02000000L +#define UVD_SUVD_CGC_GATE__SAOE_MASK 0x04000000L +#define UVD_SUVD_CGC_GATE__SRE_AV1_MASK 0x08000000L +#define UVD_SUVD_CGC_GATE__FBC_PCLK_MASK 0x10000000L +#define UVD_SUVD_CGC_GATE__FBC_CCLK_MASK 0x20000000L +#define UVD_SUVD_CGC_GATE__SCM_AV1_MASK 0x40000000L +#define UVD_SUVD_CGC_GATE__SMPA_MASK 0x80000000L +//AVM_SUVD_CGC_GATE2 +#define AVM_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define AVM_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define AVM_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define AVM_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define AVM_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define AVM_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define AVM_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define AVM_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define AVM_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define AVM_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define AVM_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define AVM_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define AVM_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define AVM_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define AVM_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define AVM_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define AVM_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define AVM_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define AVM_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define AVM_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define AVM_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define AVM_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define AVM_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define AVM_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//CDEFE_SUVD_CGC_GATE2 +#define CDEFE_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define CDEFE_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define CDEFE_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define CDEFE_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define CDEFE_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define CDEFE_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define CDEFE_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define CDEFE_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define CDEFE_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define CDEFE_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define CDEFE_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define CDEFE_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define CDEFE_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define CDEFE_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define CDEFE_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define CDEFE_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define CDEFE_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define CDEFE_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define CDEFE_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define CDEFE_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define CDEFE_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define CDEFE_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define CDEFE_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define CDEFE_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//DBR_SUVD_CGC_GATE2 +#define DBR_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define DBR_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define DBR_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define DBR_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define DBR_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define DBR_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define DBR_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define DBR_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define DBR_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define DBR_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define DBR_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define DBR_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define DBR_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define DBR_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define DBR_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define DBR_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define DBR_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define DBR_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define DBR_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define DBR_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define DBR_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define DBR_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define DBR_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define DBR_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//ENT_SUVD_CGC_GATE2 +#define ENT_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define ENT_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define ENT_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define ENT_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define ENT_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define ENT_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define ENT_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define ENT_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define ENT_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define ENT_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define ENT_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define ENT_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define ENT_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define ENT_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define ENT_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define ENT_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define ENT_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define ENT_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define ENT_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define ENT_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define ENT_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define ENT_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define ENT_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define ENT_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//IME_SUVD_CGC_GATE2 +#define IME_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define IME_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define IME_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define IME_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define IME_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define IME_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define IME_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define IME_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define IME_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define IME_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define IME_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define IME_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define IME_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define IME_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define IME_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define IME_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define IME_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define IME_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define IME_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define IME_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define IME_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define IME_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define IME_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define IME_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//MPC1_SUVD_CGC_GATE2 +#define MPC1_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define MPC1_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define MPC1_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define MPC1_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define MPC1_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define MPC1_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define MPC1_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define MPC1_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define MPC1_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define MPC1_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define MPC1_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define MPC1_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define MPC1_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define MPC1_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define MPC1_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define MPC1_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define MPC1_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define MPC1_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define MPC1_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define MPC1_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define MPC1_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define MPC1_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define MPC1_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define MPC1_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//SAOE_SUVD_CGC_GATE2 +#define SAOE_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define SAOE_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define SAOE_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define SAOE_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define SAOE_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define SAOE_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define SAOE_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define SAOE_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define SAOE_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define SAOE_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define SAOE_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define SAOE_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define SAOE_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define SAOE_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define SAOE_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define SAOE_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define SAOE_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define SAOE_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define SAOE_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define SAOE_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define SAOE_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define SAOE_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define SAOE_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define SAOE_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//SDB_SUVD_CGC_GATE2 +#define SDB_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define SDB_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define SDB_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define SDB_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define SDB_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define SDB_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define SDB_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define SDB_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define SDB_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define SDB_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define SDB_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define SDB_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define SDB_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define SDB_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define SDB_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define SDB_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define SDB_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define SDB_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define SDB_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define SDB_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define SDB_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define SDB_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define SDB_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define SDB_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//SIT0_NXT_SUVD_CGC_GATE2 +#define SIT0_NXT_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define SIT0_NXT_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define SIT0_NXT_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define SIT0_NXT_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define SIT0_NXT_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define SIT0_NXT_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define SIT0_NXT_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define SIT0_NXT_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define SIT0_NXT_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define SIT0_NXT_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define SIT0_NXT_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define SIT0_NXT_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define SIT0_NXT_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define SIT0_NXT_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define SIT0_NXT_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define SIT0_NXT_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define SIT0_NXT_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define SIT0_NXT_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define SIT0_NXT_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define SIT0_NXT_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define SIT0_NXT_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define SIT0_NXT_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define SIT0_NXT_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define SIT0_NXT_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//SIT1_NXT_SUVD_CGC_GATE2 +#define SIT1_NXT_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define SIT1_NXT_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define SIT1_NXT_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define SIT1_NXT_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define SIT1_NXT_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define SIT1_NXT_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define SIT1_NXT_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define SIT1_NXT_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define SIT1_NXT_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define SIT1_NXT_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define SIT1_NXT_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define SIT1_NXT_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define SIT1_NXT_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define SIT1_NXT_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define SIT1_NXT_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define SIT1_NXT_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define SIT1_NXT_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define SIT1_NXT_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define SIT1_NXT_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define SIT1_NXT_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define SIT1_NXT_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define SIT1_NXT_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define SIT1_NXT_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define SIT1_NXT_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//SIT2_NXT_SUVD_CGC_GATE2 +#define SIT2_NXT_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define SIT2_NXT_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define SIT2_NXT_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define SIT2_NXT_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define SIT2_NXT_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define SIT2_NXT_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define SIT2_NXT_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define SIT2_NXT_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define SIT2_NXT_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define SIT2_NXT_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define SIT2_NXT_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define SIT2_NXT_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define SIT2_NXT_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define SIT2_NXT_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define SIT2_NXT_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define SIT2_NXT_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define SIT2_NXT_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define SIT2_NXT_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define SIT2_NXT_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define SIT2_NXT_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define SIT2_NXT_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define SIT2_NXT_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define SIT2_NXT_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define SIT2_NXT_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//SIT_SUVD_CGC_GATE2 +#define SIT_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define SIT_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define SIT_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define SIT_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define SIT_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define SIT_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define SIT_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define SIT_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define SIT_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define SIT_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define SIT_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define SIT_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define SIT_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define SIT_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define SIT_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define SIT_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define SIT_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define SIT_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define SIT_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define SIT_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define SIT_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define SIT_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define SIT_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define SIT_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//SMPA_SUVD_CGC_GATE2 +#define SMPA_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define SMPA_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define SMPA_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define SMPA_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define SMPA_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define SMPA_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define SMPA_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define SMPA_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define SMPA_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define SMPA_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define SMPA_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define SMPA_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define SMPA_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define SMPA_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define SMPA_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define SMPA_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define SMPA_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define SMPA_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define SMPA_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define SMPA_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define SMPA_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define SMPA_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define SMPA_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define SMPA_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//SMP_SUVD_CGC_GATE2 +#define SMP_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define SMP_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define SMP_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define SMP_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define SMP_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define SMP_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define SMP_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define SMP_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define SMP_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define SMP_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define SMP_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define SMP_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define SMP_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define SMP_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define SMP_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define SMP_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define SMP_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define SMP_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define SMP_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define SMP_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define SMP_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define SMP_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define SMP_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define SMP_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//SRE_SUVD_CGC_GATE2 +#define SRE_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define SRE_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define SRE_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define SRE_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define SRE_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define SRE_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define SRE_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define SRE_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define SRE_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define SRE_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define SRE_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define SRE_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define SRE_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define SRE_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define SRE_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define SRE_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define SRE_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define SRE_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define SRE_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define SRE_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define SRE_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define SRE_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define SRE_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define SRE_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//UVD_MPBE0_SUVD_CGC_GATE2 +#define UVD_MPBE0_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define UVD_MPBE0_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define UVD_MPBE0_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define UVD_MPBE0_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define UVD_MPBE0_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define UVD_MPBE0_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define UVD_MPBE0_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define UVD_MPBE0_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define UVD_MPBE0_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define UVD_MPBE0_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define UVD_MPBE0_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define UVD_MPBE0_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define UVD_MPBE0_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define UVD_MPBE0_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +//UVD_MPBE1_SUVD_CGC_GATE2 +#define UVD_MPBE1_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define UVD_MPBE1_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define UVD_MPBE1_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define UVD_MPBE1_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define UVD_MPBE1_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define UVD_MPBE1_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define UVD_MPBE1_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define UVD_MPBE1_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define UVD_MPBE1_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define UVD_MPBE1_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define UVD_MPBE1_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define UVD_MPBE1_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define UVD_MPBE1_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define UVD_MPBE1_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +//UVD_SUVD_CGC_GATE2 +#define UVD_SUVD_CGC_GATE2__MPBE0__SHIFT 0x0 +#define UVD_SUVD_CGC_GATE2__MPBE1__SHIFT 0x1 +#define UVD_SUVD_CGC_GATE2__SIT_AV1__SHIFT 0x2 +#define UVD_SUVD_CGC_GATE2__SDB_AV1__SHIFT 0x3 +#define UVD_SUVD_CGC_GATE2__MPC1__SHIFT 0x4 +#define UVD_SUVD_CGC_GATE2__SRE_AV1_ENC__SHIFT 0x5 +#define UVD_SUVD_CGC_GATE2__CDEFE__SHIFT 0x6 +#define UVD_SUVD_CGC_GATE2__AVM_0__SHIFT 0x7 +#define UVD_SUVD_CGC_GATE2__AVM_1__SHIFT 0x8 +#define UVD_SUVD_CGC_GATE2__SIT_NXT_CMN__SHIFT 0x9 +#define UVD_SUVD_CGC_GATE2__SIT_NXT_DEC__SHIFT 0xa +#define UVD_SUVD_CGC_GATE2__SIT_NXT_ENC__SHIFT 0xb +#define UVD_SUVD_CGC_GATE2__MPBE0_MASK 0x00000001L +#define UVD_SUVD_CGC_GATE2__MPBE1_MASK 0x00000002L +#define UVD_SUVD_CGC_GATE2__SIT_AV1_MASK 0x00000004L +#define UVD_SUVD_CGC_GATE2__SDB_AV1_MASK 0x00000008L +#define UVD_SUVD_CGC_GATE2__MPC1_MASK 0x00000010L +#define UVD_SUVD_CGC_GATE2__SRE_AV1_ENC_MASK 0x00000020L +#define UVD_SUVD_CGC_GATE2__CDEFE_MASK 0x00000040L +#define UVD_SUVD_CGC_GATE2__AVM_0_MASK 0x00000080L +#define UVD_SUVD_CGC_GATE2__AVM_1_MASK 0x00000100L +#define UVD_SUVD_CGC_GATE2__SIT_NXT_CMN_MASK 0x00000200L +#define UVD_SUVD_CGC_GATE2__SIT_NXT_DEC_MASK 0x00000400L +#define UVD_SUVD_CGC_GATE2__SIT_NXT_ENC_MASK 0x00000800L +//AVM_SUVD_CGC_CTRL +#define AVM_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define AVM_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define AVM_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define AVM_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define AVM_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define AVM_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define AVM_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define AVM_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define AVM_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define AVM_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define AVM_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define AVM_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define AVM_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define AVM_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define AVM_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define AVM_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define AVM_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define AVM_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define AVM_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define AVM_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define AVM_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define AVM_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define AVM_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define AVM_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define AVM_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define AVM_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define AVM_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define AVM_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define AVM_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define AVM_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define AVM_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define AVM_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define AVM_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define AVM_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define AVM_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define AVM_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define AVM_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define AVM_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define AVM_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define AVM_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define AVM_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define AVM_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define AVM_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define AVM_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define AVM_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define AVM_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define AVM_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define AVM_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define AVM_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define AVM_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define AVM_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define AVM_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//CDEFE_SUVD_CGC_CTRL +#define CDEFE_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define CDEFE_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define CDEFE_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define CDEFE_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define CDEFE_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define CDEFE_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define CDEFE_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define CDEFE_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define CDEFE_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define CDEFE_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define CDEFE_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define CDEFE_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define CDEFE_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define CDEFE_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define CDEFE_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define CDEFE_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define CDEFE_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define CDEFE_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define CDEFE_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define CDEFE_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define CDEFE_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define CDEFE_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define CDEFE_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define CDEFE_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define CDEFE_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define CDEFE_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define CDEFE_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define CDEFE_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define CDEFE_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define CDEFE_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define CDEFE_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define CDEFE_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define CDEFE_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define CDEFE_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define CDEFE_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define CDEFE_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define CDEFE_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define CDEFE_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define CDEFE_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define CDEFE_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define CDEFE_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define CDEFE_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define CDEFE_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define CDEFE_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define CDEFE_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define CDEFE_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define CDEFE_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define CDEFE_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define CDEFE_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define CDEFE_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define CDEFE_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define CDEFE_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//DBR_SUVD_CGC_CTRL +#define DBR_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define DBR_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define DBR_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define DBR_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define DBR_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define DBR_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define DBR_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define DBR_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define DBR_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define DBR_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define DBR_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define DBR_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define DBR_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define DBR_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define DBR_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define DBR_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define DBR_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define DBR_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define DBR_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define DBR_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define DBR_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define DBR_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define DBR_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define DBR_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define DBR_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define DBR_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define DBR_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define DBR_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define DBR_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define DBR_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define DBR_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define DBR_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define DBR_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define DBR_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define DBR_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define DBR_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define DBR_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define DBR_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define DBR_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define DBR_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define DBR_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define DBR_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define DBR_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define DBR_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define DBR_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define DBR_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define DBR_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define DBR_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define DBR_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define DBR_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define DBR_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define DBR_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//EFC_SUVD_CGC_CTRL +#define EFC_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define EFC_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define EFC_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define EFC_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define EFC_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define EFC_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define EFC_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define EFC_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define EFC_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define EFC_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define EFC_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define EFC_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define EFC_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define EFC_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define EFC_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define EFC_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define EFC_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define EFC_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define EFC_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define EFC_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define EFC_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define EFC_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define EFC_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define EFC_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define EFC_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define EFC_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define EFC_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define EFC_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define EFC_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define EFC_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define EFC_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define EFC_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define EFC_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define EFC_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define EFC_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define EFC_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define EFC_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define EFC_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define EFC_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define EFC_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define EFC_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define EFC_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define EFC_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define EFC_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define EFC_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define EFC_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define EFC_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define EFC_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define EFC_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define EFC_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define EFC_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define EFC_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//ENT_SUVD_CGC_CTRL +#define ENT_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define ENT_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define ENT_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define ENT_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define ENT_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define ENT_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define ENT_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define ENT_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define ENT_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define ENT_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define ENT_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define ENT_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define ENT_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define ENT_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define ENT_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define ENT_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define ENT_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define ENT_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define ENT_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define ENT_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define ENT_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define ENT_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define ENT_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define ENT_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define ENT_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define ENT_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define ENT_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define ENT_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define ENT_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define ENT_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define ENT_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define ENT_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define ENT_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define ENT_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define ENT_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define ENT_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define ENT_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define ENT_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define ENT_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define ENT_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define ENT_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define ENT_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define ENT_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define ENT_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define ENT_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define ENT_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define ENT_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define ENT_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define ENT_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define ENT_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define ENT_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define ENT_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//IME_SUVD_CGC_CTRL +#define IME_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define IME_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define IME_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define IME_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define IME_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define IME_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define IME_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define IME_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define IME_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define IME_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define IME_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define IME_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define IME_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define IME_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define IME_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define IME_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define IME_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define IME_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define IME_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define IME_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define IME_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define IME_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define IME_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define IME_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define IME_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define IME_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define IME_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define IME_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define IME_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define IME_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define IME_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define IME_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define IME_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define IME_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define IME_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define IME_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define IME_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define IME_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define IME_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define IME_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define IME_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define IME_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define IME_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define IME_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define IME_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define IME_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define IME_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define IME_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define IME_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define IME_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define IME_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define IME_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//MPC1_SUVD_CGC_CTRL +#define MPC1_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define MPC1_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define MPC1_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define MPC1_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define MPC1_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define MPC1_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define MPC1_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define MPC1_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define MPC1_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define MPC1_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define MPC1_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define MPC1_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define MPC1_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define MPC1_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define MPC1_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define MPC1_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define MPC1_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define MPC1_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define MPC1_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define MPC1_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define MPC1_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define MPC1_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define MPC1_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define MPC1_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define MPC1_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define MPC1_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define MPC1_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define MPC1_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define MPC1_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define MPC1_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define MPC1_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define MPC1_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define MPC1_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define MPC1_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define MPC1_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define MPC1_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define MPC1_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define MPC1_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define MPC1_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define MPC1_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define MPC1_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define MPC1_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define MPC1_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define MPC1_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define MPC1_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define MPC1_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define MPC1_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define MPC1_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define MPC1_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define MPC1_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define MPC1_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define MPC1_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//PPU_SUVD_CGC_CTRL +#define PPU_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define PPU_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define PPU_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define PPU_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define PPU_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define PPU_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define PPU_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define PPU_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define PPU_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define PPU_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define PPU_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define PPU_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define PPU_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define PPU_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define PPU_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define PPU_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define PPU_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define PPU_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define PPU_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define PPU_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define PPU_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define PPU_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define PPU_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define PPU_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define PPU_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define PPU_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define PPU_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define PPU_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define PPU_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define PPU_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define PPU_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define PPU_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define PPU_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define PPU_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define PPU_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define PPU_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define PPU_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define PPU_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define PPU_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define PPU_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define PPU_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define PPU_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define PPU_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define PPU_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define PPU_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define PPU_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define PPU_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define PPU_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define PPU_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define PPU_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define PPU_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define PPU_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//SAOE_SUVD_CGC_CTRL +#define SAOE_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define SAOE_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define SAOE_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define SAOE_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define SAOE_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define SAOE_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define SAOE_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define SAOE_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define SAOE_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define SAOE_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define SAOE_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define SAOE_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define SAOE_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define SAOE_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define SAOE_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define SAOE_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define SAOE_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define SAOE_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define SAOE_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define SAOE_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define SAOE_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define SAOE_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define SAOE_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define SAOE_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define SAOE_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define SAOE_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define SAOE_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define SAOE_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define SAOE_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define SAOE_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define SAOE_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define SAOE_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define SAOE_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define SAOE_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define SAOE_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define SAOE_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define SAOE_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define SAOE_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define SAOE_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define SAOE_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define SAOE_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define SAOE_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define SAOE_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define SAOE_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define SAOE_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define SAOE_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define SAOE_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define SAOE_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define SAOE_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define SAOE_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define SAOE_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define SAOE_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//SCM_SUVD_CGC_CTRL +#define SCM_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define SCM_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define SCM_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define SCM_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define SCM_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define SCM_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define SCM_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define SCM_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define SCM_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define SCM_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define SCM_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define SCM_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define SCM_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define SCM_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define SCM_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define SCM_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define SCM_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define SCM_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define SCM_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define SCM_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define SCM_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define SCM_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define SCM_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define SCM_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define SCM_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define SCM_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define SCM_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define SCM_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define SCM_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define SCM_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define SCM_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define SCM_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define SCM_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define SCM_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define SCM_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define SCM_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define SCM_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define SCM_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define SCM_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define SCM_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define SCM_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define SCM_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define SCM_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define SCM_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define SCM_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define SCM_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define SCM_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define SCM_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define SCM_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define SCM_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define SCM_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define SCM_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//SDB_SUVD_CGC_CTRL +#define SDB_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define SDB_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define SDB_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define SDB_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define SDB_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define SDB_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define SDB_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define SDB_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define SDB_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define SDB_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define SDB_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define SDB_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define SDB_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define SDB_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define SDB_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define SDB_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define SDB_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define SDB_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define SDB_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define SDB_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define SDB_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define SDB_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define SDB_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define SDB_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define SDB_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define SDB_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define SDB_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define SDB_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define SDB_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define SDB_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define SDB_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define SDB_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define SDB_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define SDB_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define SDB_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define SDB_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define SDB_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define SDB_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define SDB_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define SDB_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define SDB_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define SDB_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define SDB_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define SDB_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define SDB_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define SDB_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define SDB_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define SDB_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define SDB_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define SDB_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define SDB_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define SDB_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//SIT0_NXT_SUVD_CGC_CTRL +#define SIT0_NXT_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define SIT0_NXT_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define SIT0_NXT_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define SIT0_NXT_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define SIT0_NXT_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define SIT0_NXT_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define SIT0_NXT_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define SIT0_NXT_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define SIT0_NXT_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define SIT0_NXT_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define SIT0_NXT_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define SIT0_NXT_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define SIT0_NXT_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define SIT0_NXT_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define SIT0_NXT_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define SIT0_NXT_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define SIT0_NXT_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define SIT0_NXT_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define SIT0_NXT_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define SIT0_NXT_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define SIT0_NXT_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define SIT0_NXT_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define SIT0_NXT_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define SIT0_NXT_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define SIT0_NXT_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define SIT0_NXT_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define SIT0_NXT_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define SIT0_NXT_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define SIT0_NXT_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define SIT0_NXT_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define SIT0_NXT_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define SIT0_NXT_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define SIT0_NXT_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define SIT0_NXT_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define SIT0_NXT_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define SIT0_NXT_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define SIT0_NXT_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define SIT0_NXT_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define SIT0_NXT_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define SIT0_NXT_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define SIT0_NXT_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define SIT0_NXT_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define SIT0_NXT_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define SIT0_NXT_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define SIT0_NXT_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define SIT0_NXT_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define SIT0_NXT_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define SIT0_NXT_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define SIT0_NXT_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define SIT0_NXT_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define SIT0_NXT_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define SIT0_NXT_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//SIT1_NXT_SUVD_CGC_CTRL +#define SIT1_NXT_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define SIT1_NXT_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define SIT1_NXT_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define SIT1_NXT_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define SIT1_NXT_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define SIT1_NXT_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define SIT1_NXT_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define SIT1_NXT_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define SIT1_NXT_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define SIT1_NXT_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define SIT1_NXT_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define SIT1_NXT_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define SIT1_NXT_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define SIT1_NXT_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define SIT1_NXT_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define SIT1_NXT_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define SIT1_NXT_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define SIT1_NXT_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define SIT1_NXT_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define SIT1_NXT_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define SIT1_NXT_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define SIT1_NXT_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define SIT1_NXT_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define SIT1_NXT_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define SIT1_NXT_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define SIT1_NXT_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define SIT1_NXT_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define SIT1_NXT_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define SIT1_NXT_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define SIT1_NXT_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define SIT1_NXT_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define SIT1_NXT_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define SIT1_NXT_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define SIT1_NXT_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define SIT1_NXT_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define SIT1_NXT_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define SIT1_NXT_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define SIT1_NXT_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define SIT1_NXT_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define SIT1_NXT_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define SIT1_NXT_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define SIT1_NXT_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define SIT1_NXT_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define SIT1_NXT_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define SIT1_NXT_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define SIT1_NXT_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define SIT1_NXT_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define SIT1_NXT_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define SIT1_NXT_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define SIT1_NXT_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define SIT1_NXT_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define SIT1_NXT_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//SIT2_NXT_SUVD_CGC_CTRL +#define SIT2_NXT_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define SIT2_NXT_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define SIT2_NXT_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define SIT2_NXT_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define SIT2_NXT_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define SIT2_NXT_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define SIT2_NXT_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define SIT2_NXT_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define SIT2_NXT_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define SIT2_NXT_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define SIT2_NXT_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define SIT2_NXT_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define SIT2_NXT_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define SIT2_NXT_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define SIT2_NXT_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define SIT2_NXT_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define SIT2_NXT_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define SIT2_NXT_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define SIT2_NXT_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define SIT2_NXT_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define SIT2_NXT_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define SIT2_NXT_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define SIT2_NXT_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define SIT2_NXT_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define SIT2_NXT_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define SIT2_NXT_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define SIT2_NXT_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define SIT2_NXT_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define SIT2_NXT_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define SIT2_NXT_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define SIT2_NXT_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define SIT2_NXT_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define SIT2_NXT_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define SIT2_NXT_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define SIT2_NXT_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define SIT2_NXT_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define SIT2_NXT_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define SIT2_NXT_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define SIT2_NXT_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define SIT2_NXT_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define SIT2_NXT_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define SIT2_NXT_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define SIT2_NXT_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define SIT2_NXT_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define SIT2_NXT_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define SIT2_NXT_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define SIT2_NXT_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define SIT2_NXT_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define SIT2_NXT_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define SIT2_NXT_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define SIT2_NXT_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define SIT2_NXT_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//SIT_SUVD_CGC_CTRL +#define SIT_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define SIT_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define SIT_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define SIT_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define SIT_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define SIT_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define SIT_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define SIT_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define SIT_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define SIT_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define SIT_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define SIT_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define SIT_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define SIT_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define SIT_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define SIT_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define SIT_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define SIT_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define SIT_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define SIT_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define SIT_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define SIT_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define SIT_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define SIT_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define SIT_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define SIT_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define SIT_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define SIT_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define SIT_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define SIT_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define SIT_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define SIT_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define SIT_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define SIT_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define SIT_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define SIT_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define SIT_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define SIT_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define SIT_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define SIT_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define SIT_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define SIT_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define SIT_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define SIT_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define SIT_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define SIT_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define SIT_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define SIT_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define SIT_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define SIT_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define SIT_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define SIT_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//SMPA_SUVD_CGC_CTRL +#define SMPA_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define SMPA_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define SMPA_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define SMPA_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define SMPA_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define SMPA_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define SMPA_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define SMPA_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define SMPA_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define SMPA_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define SMPA_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define SMPA_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define SMPA_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define SMPA_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define SMPA_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define SMPA_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define SMPA_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define SMPA_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define SMPA_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define SMPA_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define SMPA_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define SMPA_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define SMPA_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define SMPA_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define SMPA_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define SMPA_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define SMPA_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define SMPA_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define SMPA_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define SMPA_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define SMPA_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define SMPA_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define SMPA_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define SMPA_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define SMPA_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define SMPA_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define SMPA_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define SMPA_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define SMPA_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define SMPA_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define SMPA_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define SMPA_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define SMPA_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define SMPA_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define SMPA_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define SMPA_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define SMPA_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define SMPA_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define SMPA_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define SMPA_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define SMPA_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define SMPA_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//SMP_SUVD_CGC_CTRL +#define SMP_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define SMP_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define SMP_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define SMP_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define SMP_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define SMP_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define SMP_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define SMP_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define SMP_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define SMP_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define SMP_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define SMP_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define SMP_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define SMP_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define SMP_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define SMP_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define SMP_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define SMP_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define SMP_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define SMP_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define SMP_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define SMP_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define SMP_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define SMP_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define SMP_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define SMP_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define SMP_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define SMP_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define SMP_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define SMP_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define SMP_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define SMP_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define SMP_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define SMP_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define SMP_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define SMP_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define SMP_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define SMP_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define SMP_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define SMP_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define SMP_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define SMP_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define SMP_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define SMP_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define SMP_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define SMP_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define SMP_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define SMP_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define SMP_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define SMP_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define SMP_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define SMP_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//SRE_SUVD_CGC_CTRL +#define SRE_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define SRE_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define SRE_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define SRE_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define SRE_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define SRE_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define SRE_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define SRE_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define SRE_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define SRE_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define SRE_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define SRE_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define SRE_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define SRE_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define SRE_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define SRE_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define SRE_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define SRE_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define SRE_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define SRE_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define SRE_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define SRE_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define SRE_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define SRE_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define SRE_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define SRE_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define SRE_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define SRE_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define SRE_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define SRE_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define SRE_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define SRE_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define SRE_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define SRE_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define SRE_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define SRE_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define SRE_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define SRE_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define SRE_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define SRE_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define SRE_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define SRE_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define SRE_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define SRE_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define SRE_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define SRE_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define SRE_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define SRE_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define SRE_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define SRE_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define SRE_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define SRE_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//UVD_MPBE0_SUVD_CGC_CTRL +#define UVD_MPBE0_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define UVD_MPBE0_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define UVD_MPBE0_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define UVD_MPBE0_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define UVD_MPBE0_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define UVD_MPBE0_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define UVD_MPBE0_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define UVD_MPBE0_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define UVD_MPBE0_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define UVD_MPBE0_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define UVD_MPBE0_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define UVD_MPBE0_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define UVD_MPBE0_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define UVD_MPBE0_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define UVD_MPBE0_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define UVD_MPBE0_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define UVD_MPBE0_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define UVD_MPBE0_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define UVD_MPBE0_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define UVD_MPBE0_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define UVD_MPBE0_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define UVD_MPBE0_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define UVD_MPBE0_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define UVD_MPBE0_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define UVD_MPBE0_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define UVD_MPBE0_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define UVD_MPBE0_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define UVD_MPBE0_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define UVD_MPBE0_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define UVD_MPBE0_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define UVD_MPBE0_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define UVD_MPBE0_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define UVD_MPBE0_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define UVD_MPBE0_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define UVD_MPBE0_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define UVD_MPBE0_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define UVD_MPBE0_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define UVD_MPBE0_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define UVD_MPBE0_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define UVD_MPBE0_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define UVD_MPBE0_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define UVD_MPBE0_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//UVD_MPBE1_SUVD_CGC_CTRL +#define UVD_MPBE1_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define UVD_MPBE1_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define UVD_MPBE1_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define UVD_MPBE1_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define UVD_MPBE1_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define UVD_MPBE1_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define UVD_MPBE1_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define UVD_MPBE1_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define UVD_MPBE1_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define UVD_MPBE1_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define UVD_MPBE1_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define UVD_MPBE1_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define UVD_MPBE1_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define UVD_MPBE1_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define UVD_MPBE1_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define UVD_MPBE1_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define UVD_MPBE1_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define UVD_MPBE1_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define UVD_MPBE1_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define UVD_MPBE1_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define UVD_MPBE1_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define UVD_MPBE1_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define UVD_MPBE1_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define UVD_MPBE1_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define UVD_MPBE1_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define UVD_MPBE1_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define UVD_MPBE1_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define UVD_MPBE1_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define UVD_MPBE1_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define UVD_MPBE1_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define UVD_MPBE1_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define UVD_MPBE1_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define UVD_MPBE1_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define UVD_MPBE1_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define UVD_MPBE1_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define UVD_MPBE1_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define UVD_MPBE1_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define UVD_MPBE1_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define UVD_MPBE1_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define UVD_MPBE1_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define UVD_MPBE1_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define UVD_MPBE1_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//UVD_SUVD_CGC_CTRL +#define UVD_SUVD_CGC_CTRL__SRE_MODE__SHIFT 0x0 +#define UVD_SUVD_CGC_CTRL__SIT_MODE__SHIFT 0x1 +#define UVD_SUVD_CGC_CTRL__SMP_MODE__SHIFT 0x2 +#define UVD_SUVD_CGC_CTRL__SCM_MODE__SHIFT 0x3 +#define UVD_SUVD_CGC_CTRL__SDB_MODE__SHIFT 0x4 +#define UVD_SUVD_CGC_CTRL__SCLR_MODE__SHIFT 0x5 +#define UVD_SUVD_CGC_CTRL__UVD_SC_MODE__SHIFT 0x6 +#define UVD_SUVD_CGC_CTRL__ENT_MODE__SHIFT 0x7 +#define UVD_SUVD_CGC_CTRL__IME_MODE__SHIFT 0x8 +#define UVD_SUVD_CGC_CTRL__SITE_MODE__SHIFT 0x9 +#define UVD_SUVD_CGC_CTRL__EFC_MODE__SHIFT 0xa +#define UVD_SUVD_CGC_CTRL__SAOE_MODE__SHIFT 0xb +#define UVD_SUVD_CGC_CTRL__SMPA_MODE__SHIFT 0xc +#define UVD_SUVD_CGC_CTRL__MPBE0_MODE__SHIFT 0xd +#define UVD_SUVD_CGC_CTRL__MPBE1_MODE__SHIFT 0xe +#define UVD_SUVD_CGC_CTRL__SIT_AV1_MODE__SHIFT 0xf +#define UVD_SUVD_CGC_CTRL__SDB_AV1_MODE__SHIFT 0x10 +#define UVD_SUVD_CGC_CTRL__MPC1_MODE__SHIFT 0x11 +#define UVD_SUVD_CGC_CTRL__AVM_0_MODE__SHIFT 0x12 +#define UVD_SUVD_CGC_CTRL__AVM_1_MODE__SHIFT 0x13 +#define UVD_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE__SHIFT 0x14 +#define UVD_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE__SHIFT 0x15 +#define UVD_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE__SHIFT 0x16 +#define UVD_SUVD_CGC_CTRL__FBC_PCLK__SHIFT 0x1c +#define UVD_SUVD_CGC_CTRL__FBC_CCLK__SHIFT 0x1d +#define UVD_SUVD_CGC_CTRL__CDEFE_MODE__SHIFT 0x1e +#define UVD_SUVD_CGC_CTRL__SRE_MODE_MASK 0x00000001L +#define UVD_SUVD_CGC_CTRL__SIT_MODE_MASK 0x00000002L +#define UVD_SUVD_CGC_CTRL__SMP_MODE_MASK 0x00000004L +#define UVD_SUVD_CGC_CTRL__SCM_MODE_MASK 0x00000008L +#define UVD_SUVD_CGC_CTRL__SDB_MODE_MASK 0x00000010L +#define UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK 0x00000020L +#define UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK 0x00000040L +#define UVD_SUVD_CGC_CTRL__ENT_MODE_MASK 0x00000080L +#define UVD_SUVD_CGC_CTRL__IME_MODE_MASK 0x00000100L +#define UVD_SUVD_CGC_CTRL__SITE_MODE_MASK 0x00000200L +#define UVD_SUVD_CGC_CTRL__EFC_MODE_MASK 0x00000400L +#define UVD_SUVD_CGC_CTRL__SAOE_MODE_MASK 0x00000800L +#define UVD_SUVD_CGC_CTRL__SMPA_MODE_MASK 0x00001000L +#define UVD_SUVD_CGC_CTRL__MPBE0_MODE_MASK 0x00002000L +#define UVD_SUVD_CGC_CTRL__MPBE1_MODE_MASK 0x00004000L +#define UVD_SUVD_CGC_CTRL__SIT_AV1_MODE_MASK 0x00008000L +#define UVD_SUVD_CGC_CTRL__SDB_AV1_MODE_MASK 0x00010000L +#define UVD_SUVD_CGC_CTRL__MPC1_MODE_MASK 0x00020000L +#define UVD_SUVD_CGC_CTRL__AVM_0_MODE_MASK 0x00040000L +#define UVD_SUVD_CGC_CTRL__AVM_1_MODE_MASK 0x00080000L +#define UVD_SUVD_CGC_CTRL__SIT_NXT_CMN_MODE_MASK 0x00100000L +#define UVD_SUVD_CGC_CTRL__SIT_NXT_DEC_MODE_MASK 0x00200000L +#define UVD_SUVD_CGC_CTRL__SIT_NXT_ENC_MODE_MASK 0x00400000L +#define UVD_SUVD_CGC_CTRL__FBC_PCLK_MASK 0x10000000L +#define UVD_SUVD_CGC_CTRL__FBC_CCLK_MASK 0x20000000L +#define UVD_SUVD_CGC_CTRL__CDEFE_MODE_MASK 0x40000000L +//UVD_CGC_CTRL3 +#define UVD_CGC_CTRL3__CGC_CLK_OFF_DELAY__SHIFT 0x0 +#define UVD_CGC_CTRL3__LCM0_MODE__SHIFT 0xb +#define UVD_CGC_CTRL3__LCM1_MODE__SHIFT 0xc +#define UVD_CGC_CTRL3__MIF_MODE__SHIFT 0xd +#define UVD_CGC_CTRL3__VREG_MODE__SHIFT 0xe +#define UVD_CGC_CTRL3__PE_MODE__SHIFT 0xf +#define UVD_CGC_CTRL3__PPU_MODE__SHIFT 0x10 +#define UVD_CGC_CTRL3__CGC_CLK_OFF_DELAY_MASK 0x000000FFL +#define UVD_CGC_CTRL3__LCM0_MODE_MASK 0x00000800L +#define UVD_CGC_CTRL3__LCM1_MODE_MASK 0x00001000L +#define UVD_CGC_CTRL3__MIF_MODE_MASK 0x00002000L +#define UVD_CGC_CTRL3__VREG_MODE_MASK 0x00004000L +#define UVD_CGC_CTRL3__PE_MODE_MASK 0x00008000L +#define UVD_CGC_CTRL3__PPU_MODE_MASK 0x00010000L +//UVD_GPCOM_VCPU_DATA0 +#define UVD_GPCOM_VCPU_DATA0__DATA0__SHIFT 0x0 +#define UVD_GPCOM_VCPU_DATA0__DATA0_MASK 0xFFFFFFFFL +//UVD_GPCOM_VCPU_DATA1 +#define UVD_GPCOM_VCPU_DATA1__DATA1__SHIFT 0x0 +#define UVD_GPCOM_VCPU_DATA1__DATA1_MASK 0xFFFFFFFFL +//UVD_GPCOM_SYS_CMD +#define UVD_GPCOM_SYS_CMD__CMD_SEND__SHIFT 0x0 +#define UVD_GPCOM_SYS_CMD__CMD__SHIFT 0x1 +#define UVD_GPCOM_SYS_CMD__CMD_SOURCE__SHIFT 0x1f +#define UVD_GPCOM_SYS_CMD__CMD_SEND_MASK 0x00000001L +#define UVD_GPCOM_SYS_CMD__CMD_MASK 0x7FFFFFFEL +#define UVD_GPCOM_SYS_CMD__CMD_SOURCE_MASK 0x80000000L +//UVD_GPCOM_SYS_DATA0 +#define UVD_GPCOM_SYS_DATA0__DATA0__SHIFT 0x0 +#define UVD_GPCOM_SYS_DATA0__DATA0_MASK 0xFFFFFFFFL +//UVD_GPCOM_SYS_DATA1 +#define UVD_GPCOM_SYS_DATA1__DATA1__SHIFT 0x0 +#define UVD_GPCOM_SYS_DATA1__DATA1_MASK 0xFFFFFFFFL +//UVD_VCPU_INT_EN +#define UVD_VCPU_INT_EN__PIF_ADDR_ERR_EN__SHIFT 0x0 +#define UVD_VCPU_INT_EN__SEMA_WAIT_FAULT_TIMEOUT_EN__SHIFT 0x1 +#define UVD_VCPU_INT_EN__SEMA_SIGNAL_INCOMPLETE_TIMEOUT_EN__SHIFT 0x2 +#define UVD_VCPU_INT_EN__NJ_PF_RPT_EN__SHIFT 0x3 +#define UVD_VCPU_INT_EN__SW_RB1_INT_EN__SHIFT 0x4 +#define UVD_VCPU_INT_EN__SW_RB2_INT_EN__SHIFT 0x5 +#define UVD_VCPU_INT_EN__RBC_REG_PRIV_FAULT_EN__SHIFT 0x6 +#define UVD_VCPU_INT_EN__SW_RB3_INT_EN__SHIFT 0x7 +#define UVD_VCPU_INT_EN__SW_RB4_INT_EN__SHIFT 0x9 +#define UVD_VCPU_INT_EN__SW_RB5_INT_EN__SHIFT 0xa +#define UVD_VCPU_INT_EN__LBSI_EN__SHIFT 0xb +#define UVD_VCPU_INT_EN__UDEC_EN__SHIFT 0xc +#define UVD_VCPU_INT_EN__LMI_AXI_UNSUPPORTED_LEN_EN__SHIFT 0xd +#define UVD_VCPU_INT_EN__LMI_AXI_UNSUPPORTED_ADR_ALIGN_EN__SHIFT 0xe +#define UVD_VCPU_INT_EN__SUVD_EN__SHIFT 0xf +#define UVD_VCPU_INT_EN__RPTR_WR_EN__SHIFT 0x10 +#define UVD_VCPU_INT_EN__JOB_START_EN__SHIFT 0x11 +#define UVD_VCPU_INT_EN__NJ_PF_EN__SHIFT 0x12 +#define UVD_VCPU_INT_EN__SEMA_WAIT_FAIL_SIG_EN__SHIFT 0x17 +#define UVD_VCPU_INT_EN__IDCT_EN__SHIFT 0x18 +#define UVD_VCPU_INT_EN__MPRD_EN__SHIFT 0x19 +#define UVD_VCPU_INT_EN__AVM_INT_EN__SHIFT 0x1a +#define UVD_VCPU_INT_EN__CLK_SWT_EN__SHIFT 0x1b +#define UVD_VCPU_INT_EN__MIF_HWINT_EN__SHIFT 0x1c +#define UVD_VCPU_INT_EN__MPRD_ERR_EN__SHIFT 0x1d +#define UVD_VCPU_INT_EN__DRV_FW_REQ_EN__SHIFT 0x1e +#define UVD_VCPU_INT_EN__DRV_FW_ACK_EN__SHIFT 0x1f +#define UVD_VCPU_INT_EN__PIF_ADDR_ERR_EN_MASK 0x00000001L +#define UVD_VCPU_INT_EN__SEMA_WAIT_FAULT_TIMEOUT_EN_MASK 0x00000002L +#define UVD_VCPU_INT_EN__SEMA_SIGNAL_INCOMPLETE_TIMEOUT_EN_MASK 0x00000004L +#define UVD_VCPU_INT_EN__NJ_PF_RPT_EN_MASK 0x00000008L +#define UVD_VCPU_INT_EN__SW_RB1_INT_EN_MASK 0x00000010L +#define UVD_VCPU_INT_EN__SW_RB2_INT_EN_MASK 0x00000020L +#define UVD_VCPU_INT_EN__RBC_REG_PRIV_FAULT_EN_MASK 0x00000040L +#define UVD_VCPU_INT_EN__SW_RB3_INT_EN_MASK 0x00000080L +#define UVD_VCPU_INT_EN__SW_RB4_INT_EN_MASK 0x00000200L +#define UVD_VCPU_INT_EN__SW_RB5_INT_EN_MASK 0x00000400L +#define UVD_VCPU_INT_EN__LBSI_EN_MASK 0x00000800L +#define UVD_VCPU_INT_EN__UDEC_EN_MASK 0x00001000L +#define UVD_VCPU_INT_EN__LMI_AXI_UNSUPPORTED_LEN_EN_MASK 0x00002000L +#define UVD_VCPU_INT_EN__LMI_AXI_UNSUPPORTED_ADR_ALIGN_EN_MASK 0x00004000L +#define UVD_VCPU_INT_EN__SUVD_EN_MASK 0x00008000L +#define UVD_VCPU_INT_EN__RPTR_WR_EN_MASK 0x00010000L +#define UVD_VCPU_INT_EN__JOB_START_EN_MASK 0x00020000L +#define UVD_VCPU_INT_EN__NJ_PF_EN_MASK 0x00040000L +#define UVD_VCPU_INT_EN__SEMA_WAIT_FAIL_SIG_EN_MASK 0x00800000L +#define UVD_VCPU_INT_EN__IDCT_EN_MASK 0x01000000L +#define UVD_VCPU_INT_EN__MPRD_EN_MASK 0x02000000L +#define UVD_VCPU_INT_EN__AVM_INT_EN_MASK 0x04000000L +#define UVD_VCPU_INT_EN__CLK_SWT_EN_MASK 0x08000000L +#define UVD_VCPU_INT_EN__MIF_HWINT_EN_MASK 0x10000000L +#define UVD_VCPU_INT_EN__MPRD_ERR_EN_MASK 0x20000000L +#define UVD_VCPU_INT_EN__DRV_FW_REQ_EN_MASK 0x40000000L +#define UVD_VCPU_INT_EN__DRV_FW_ACK_EN_MASK 0x80000000L +//UVD_VCPU_INT_STATUS +#define UVD_VCPU_INT_STATUS__PIF_ADDR_ERR_INT__SHIFT 0x0 +#define UVD_VCPU_INT_STATUS__SEMA_WAIT_FAULT_TIMEOUT_INT__SHIFT 0x1 +#define UVD_VCPU_INT_STATUS__SEMA_SIGNAL_INCOMPLETE_TIMEOUT_INT__SHIFT 0x2 +#define UVD_VCPU_INT_STATUS__NJ_PF_RPT_INT__SHIFT 0x3 +#define UVD_VCPU_INT_STATUS__SW_RB1_INT__SHIFT 0x4 +#define UVD_VCPU_INT_STATUS__SW_RB2_INT__SHIFT 0x5 +#define UVD_VCPU_INT_STATUS__RBC_REG_PRIV_FAULT_INT__SHIFT 0x6 +#define UVD_VCPU_INT_STATUS__SW_RB3_INT__SHIFT 0x7 +#define UVD_VCPU_INT_STATUS__SW_RB4_INT__SHIFT 0x9 +#define UVD_VCPU_INT_STATUS__SW_RB5_INT__SHIFT 0xa +#define UVD_VCPU_INT_STATUS__LBSI_INT__SHIFT 0xb +#define UVD_VCPU_INT_STATUS__UDEC_INT__SHIFT 0xc +#define UVD_VCPU_INT_STATUS__LMI_AXI_UNSUPPORTED_LEN_INT__SHIFT 0xd +#define UVD_VCPU_INT_STATUS__LMI_AXI_UNSUPPORTED_ADR_ALIGN_INT__SHIFT 0xe +#define UVD_VCPU_INT_STATUS__SUVD_INT__SHIFT 0xf +#define UVD_VCPU_INT_STATUS__RPTR_WR_INT__SHIFT 0x10 +#define UVD_VCPU_INT_STATUS__JOB_START_INT__SHIFT 0x11 +#define UVD_VCPU_INT_STATUS__NJ_PF_INT__SHIFT 0x12 +#define UVD_VCPU_INT_STATUS__GPCOM_INT__SHIFT 0x14 +#define UVD_VCPU_INT_STATUS__SEMA_WAIT_FAIL_SIG_INT__SHIFT 0x17 +#define UVD_VCPU_INT_STATUS__IDCT_INT__SHIFT 0x18 +#define UVD_VCPU_INT_STATUS__MPRD_INT__SHIFT 0x19 +#define UVD_VCPU_INT_STATUS__AVM_INT__SHIFT 0x1a +#define UVD_VCPU_INT_STATUS__CLK_SWT_INT__SHIFT 0x1b +#define UVD_VCPU_INT_STATUS__MIF_HWINT__SHIFT 0x1c +#define UVD_VCPU_INT_STATUS__MPRD_ERR_INT__SHIFT 0x1d +#define UVD_VCPU_INT_STATUS__DRV_FW_REQ_INT__SHIFT 0x1e +#define UVD_VCPU_INT_STATUS__DRV_FW_ACK_INT__SHIFT 0x1f +#define UVD_VCPU_INT_STATUS__PIF_ADDR_ERR_INT_MASK 0x00000001L +#define UVD_VCPU_INT_STATUS__SEMA_WAIT_FAULT_TIMEOUT_INT_MASK 0x00000002L +#define UVD_VCPU_INT_STATUS__SEMA_SIGNAL_INCOMPLETE_TIMEOUT_INT_MASK 0x00000004L +#define UVD_VCPU_INT_STATUS__NJ_PF_RPT_INT_MASK 0x00000008L +#define UVD_VCPU_INT_STATUS__SW_RB1_INT_MASK 0x00000010L +#define UVD_VCPU_INT_STATUS__SW_RB2_INT_MASK 0x00000020L +#define UVD_VCPU_INT_STATUS__RBC_REG_PRIV_FAULT_INT_MASK 0x00000040L +#define UVD_VCPU_INT_STATUS__SW_RB3_INT_MASK 0x00000080L +#define UVD_VCPU_INT_STATUS__SW_RB4_INT_MASK 0x00000200L +#define UVD_VCPU_INT_STATUS__SW_RB5_INT_MASK 0x00000400L +#define UVD_VCPU_INT_STATUS__LBSI_INT_MASK 0x00000800L +#define UVD_VCPU_INT_STATUS__UDEC_INT_MASK 0x00001000L +#define UVD_VCPU_INT_STATUS__LMI_AXI_UNSUPPORTED_LEN_INT_MASK 0x00002000L +#define UVD_VCPU_INT_STATUS__LMI_AXI_UNSUPPORTED_ADR_ALIGN_INT_MASK 0x00004000L +#define UVD_VCPU_INT_STATUS__SUVD_INT_MASK 0x00008000L +#define UVD_VCPU_INT_STATUS__RPTR_WR_INT_MASK 0x00010000L +#define UVD_VCPU_INT_STATUS__JOB_START_INT_MASK 0x00020000L +#define UVD_VCPU_INT_STATUS__NJ_PF_INT_MASK 0x00040000L +#define UVD_VCPU_INT_STATUS__GPCOM_INT_MASK 0x00100000L +#define UVD_VCPU_INT_STATUS__SEMA_WAIT_FAIL_SIG_INT_MASK 0x00800000L +#define UVD_VCPU_INT_STATUS__IDCT_INT_MASK 0x01000000L +#define UVD_VCPU_INT_STATUS__MPRD_INT_MASK 0x02000000L +#define UVD_VCPU_INT_STATUS__AVM_INT_MASK 0x04000000L +#define UVD_VCPU_INT_STATUS__CLK_SWT_INT_MASK 0x08000000L +#define UVD_VCPU_INT_STATUS__MIF_HWINT_MASK 0x10000000L +#define UVD_VCPU_INT_STATUS__MPRD_ERR_INT_MASK 0x20000000L +#define UVD_VCPU_INT_STATUS__DRV_FW_REQ_INT_MASK 0x40000000L +#define UVD_VCPU_INT_STATUS__DRV_FW_ACK_INT_MASK 0x80000000L +//UVD_VCPU_INT_ACK +#define UVD_VCPU_INT_ACK__PIF_ADDR_ERR_ACK__SHIFT 0x0 +#define UVD_VCPU_INT_ACK__SEMA_WAIT_FAULT_TIMEOUT_ACK__SHIFT 0x1 +#define UVD_VCPU_INT_ACK__SEMA_SIGNAL_INCOMPLETE_TIMEOUT_ACK__SHIFT 0x2 +#define UVD_VCPU_INT_ACK__NJ_PF_RPT_ACK__SHIFT 0x3 +#define UVD_VCPU_INT_ACK__SW_RB1_INT_ACK__SHIFT 0x4 +#define UVD_VCPU_INT_ACK__SW_RB2_INT_ACK__SHIFT 0x5 +#define UVD_VCPU_INT_ACK__RBC_REG_PRIV_FAULT_ACK__SHIFT 0x6 +#define UVD_VCPU_INT_ACK__SW_RB3_INT_ACK__SHIFT 0x7 +#define UVD_VCPU_INT_ACK__SW_RB4_INT_ACK__SHIFT 0x9 +#define UVD_VCPU_INT_ACK__SW_RB5_INT_ACK__SHIFT 0xa +#define UVD_VCPU_INT_ACK__LBSI_ACK__SHIFT 0xb +#define UVD_VCPU_INT_ACK__UDEC_ACK__SHIFT 0xc +#define UVD_VCPU_INT_ACK__LMI_AXI_UNSUPPORTED_LEN_ACK__SHIFT 0xd +#define UVD_VCPU_INT_ACK__LMI_AXI_UNSUPPORTED_ADR_ALIGN_ACK__SHIFT 0xe +#define UVD_VCPU_INT_ACK__SUVD_ACK__SHIFT 0xf +#define UVD_VCPU_INT_ACK__RPTR_WR_ACK__SHIFT 0x10 +#define UVD_VCPU_INT_ACK__JOB_START_ACK__SHIFT 0x11 +#define UVD_VCPU_INT_ACK__NJ_PF_ACK__SHIFT 0x12 +#define UVD_VCPU_INT_ACK__SEMA_WAIT_FAIL_SIG_ACK__SHIFT 0x17 +#define UVD_VCPU_INT_ACK__IDCT_ACK__SHIFT 0x18 +#define UVD_VCPU_INT_ACK__MPRD_ACK__SHIFT 0x19 +#define UVD_VCPU_INT_ACK__AVM_INT_ACK__SHIFT 0x1a +#define UVD_VCPU_INT_ACK__CLK_SWT_ACK__SHIFT 0x1b +#define UVD_VCPU_INT_ACK__MIF_HWINT_ACK__SHIFT 0x1c +#define UVD_VCPU_INT_ACK__MPRD_ERR_ACK__SHIFT 0x1d +#define UVD_VCPU_INT_ACK__DRV_FW_REQ_ACK__SHIFT 0x1e +#define UVD_VCPU_INT_ACK__DRV_FW_ACK_ACK__SHIFT 0x1f +#define UVD_VCPU_INT_ACK__PIF_ADDR_ERR_ACK_MASK 0x00000001L +#define UVD_VCPU_INT_ACK__SEMA_WAIT_FAULT_TIMEOUT_ACK_MASK 0x00000002L +#define UVD_VCPU_INT_ACK__SEMA_SIGNAL_INCOMPLETE_TIMEOUT_ACK_MASK 0x00000004L +#define UVD_VCPU_INT_ACK__NJ_PF_RPT_ACK_MASK 0x00000008L +#define UVD_VCPU_INT_ACK__SW_RB1_INT_ACK_MASK 0x00000010L +#define UVD_VCPU_INT_ACK__SW_RB2_INT_ACK_MASK 0x00000020L +#define UVD_VCPU_INT_ACK__RBC_REG_PRIV_FAULT_ACK_MASK 0x00000040L +#define UVD_VCPU_INT_ACK__SW_RB3_INT_ACK_MASK 0x00000080L +#define UVD_VCPU_INT_ACK__SW_RB4_INT_ACK_MASK 0x00000200L +#define UVD_VCPU_INT_ACK__SW_RB5_INT_ACK_MASK 0x00000400L +#define UVD_VCPU_INT_ACK__LBSI_ACK_MASK 0x00000800L +#define UVD_VCPU_INT_ACK__UDEC_ACK_MASK 0x00001000L +#define UVD_VCPU_INT_ACK__LMI_AXI_UNSUPPORTED_LEN_ACK_MASK 0x00002000L +#define UVD_VCPU_INT_ACK__LMI_AXI_UNSUPPORTED_ADR_ALIGN_ACK_MASK 0x00004000L +#define UVD_VCPU_INT_ACK__SUVD_ACK_MASK 0x00008000L +#define UVD_VCPU_INT_ACK__RPTR_WR_ACK_MASK 0x00010000L +#define UVD_VCPU_INT_ACK__JOB_START_ACK_MASK 0x00020000L +#define UVD_VCPU_INT_ACK__NJ_PF_ACK_MASK 0x00040000L +#define UVD_VCPU_INT_ACK__SEMA_WAIT_FAIL_SIG_ACK_MASK 0x00800000L +#define UVD_VCPU_INT_ACK__IDCT_ACK_MASK 0x01000000L +#define UVD_VCPU_INT_ACK__MPRD_ACK_MASK 0x02000000L +#define UVD_VCPU_INT_ACK__AVM_INT_ACK_MASK 0x04000000L +#define UVD_VCPU_INT_ACK__CLK_SWT_ACK_MASK 0x08000000L +#define UVD_VCPU_INT_ACK__MIF_HWINT_ACK_MASK 0x10000000L +#define UVD_VCPU_INT_ACK__MPRD_ERR_ACK_MASK 0x20000000L +#define UVD_VCPU_INT_ACK__DRV_FW_REQ_ACK_MASK 0x40000000L +#define UVD_VCPU_INT_ACK__DRV_FW_ACK_ACK_MASK 0x80000000L +//UVD_VCPU_INT_ROUTE +#define UVD_VCPU_INT_ROUTE__DRV_FW_MSG__SHIFT 0x0 +#define UVD_VCPU_INT_ROUTE__FW_DRV_MSG_ACK__SHIFT 0x1 +#define UVD_VCPU_INT_ROUTE__VCPU_GPCOM__SHIFT 0x2 +#define UVD_VCPU_INT_ROUTE__DRV_FW_MSG_MASK 0x00000001L +#define UVD_VCPU_INT_ROUTE__FW_DRV_MSG_ACK_MASK 0x00000002L +#define UVD_VCPU_INT_ROUTE__VCPU_GPCOM_MASK 0x00000004L +//UVD_DRV_FW_MSG +#define UVD_DRV_FW_MSG__MSG__SHIFT 0x0 +#define UVD_DRV_FW_MSG__MSG_MASK 0xFFFFFFFFL +//UVD_FW_DRV_MSG_ACK +#define UVD_FW_DRV_MSG_ACK__ACK__SHIFT 0x0 +#define UVD_FW_DRV_MSG_ACK__ACK_MASK 0x00000001L +//UVD_SUVD_INT_EN +#define UVD_SUVD_INT_EN__SRE_FUNC_INT_EN__SHIFT 0x0 +#define UVD_SUVD_INT_EN__SRE_ERR_INT_EN__SHIFT 0x5 +#define UVD_SUVD_INT_EN__SIT_FUNC_INT_EN__SHIFT 0x6 +#define UVD_SUVD_INT_EN__SIT_ERR_INT_EN__SHIFT 0xb +#define UVD_SUVD_INT_EN__SMP_FUNC_INT_EN__SHIFT 0xc +#define UVD_SUVD_INT_EN__SMP_ERR_INT_EN__SHIFT 0x11 +#define UVD_SUVD_INT_EN__SCM_FUNC_INT_EN__SHIFT 0x12 +#define UVD_SUVD_INT_EN__SCM_ERR_INT_EN__SHIFT 0x17 +#define UVD_SUVD_INT_EN__SDB_FUNC_INT_EN__SHIFT 0x18 +#define UVD_SUVD_INT_EN__SDB_ERR_INT_EN__SHIFT 0x1d +#define UVD_SUVD_INT_EN__FBC_ERR_INT_EN__SHIFT 0x1e +#define UVD_SUVD_INT_EN__SRE_FUNC_INT_EN_MASK 0x0000001FL +#define UVD_SUVD_INT_EN__SRE_ERR_INT_EN_MASK 0x00000020L +#define UVD_SUVD_INT_EN__SIT_FUNC_INT_EN_MASK 0x000007C0L +#define UVD_SUVD_INT_EN__SIT_ERR_INT_EN_MASK 0x00000800L +#define UVD_SUVD_INT_EN__SMP_FUNC_INT_EN_MASK 0x0001F000L +#define UVD_SUVD_INT_EN__SMP_ERR_INT_EN_MASK 0x00020000L +#define UVD_SUVD_INT_EN__SCM_FUNC_INT_EN_MASK 0x007C0000L +#define UVD_SUVD_INT_EN__SCM_ERR_INT_EN_MASK 0x00800000L +#define UVD_SUVD_INT_EN__SDB_FUNC_INT_EN_MASK 0x1F000000L +#define UVD_SUVD_INT_EN__SDB_ERR_INT_EN_MASK 0x20000000L +#define UVD_SUVD_INT_EN__FBC_ERR_INT_EN_MASK 0x40000000L +//UVD_SUVD_INT_STATUS +#define UVD_SUVD_INT_STATUS__SRE_FUNC_INT__SHIFT 0x0 +#define UVD_SUVD_INT_STATUS__SRE_ERR_INT__SHIFT 0x5 +#define UVD_SUVD_INT_STATUS__SIT_FUNC_INT__SHIFT 0x6 +#define UVD_SUVD_INT_STATUS__SIT_ERR_INT__SHIFT 0xb +#define UVD_SUVD_INT_STATUS__SMP_FUNC_INT__SHIFT 0xc +#define UVD_SUVD_INT_STATUS__SMP_ERR_INT__SHIFT 0x11 +#define UVD_SUVD_INT_STATUS__SCM_FUNC_INT__SHIFT 0x12 +#define UVD_SUVD_INT_STATUS__SCM_ERR_INT__SHIFT 0x17 +#define UVD_SUVD_INT_STATUS__SDB_FUNC_INT__SHIFT 0x18 +#define UVD_SUVD_INT_STATUS__SDB_ERR_INT__SHIFT 0x1d +#define UVD_SUVD_INT_STATUS__FBC_ERR_INT__SHIFT 0x1e +#define UVD_SUVD_INT_STATUS__SRE_FUNC_INT_MASK 0x0000001FL +#define UVD_SUVD_INT_STATUS__SRE_ERR_INT_MASK 0x00000020L +#define UVD_SUVD_INT_STATUS__SIT_FUNC_INT_MASK 0x000007C0L +#define UVD_SUVD_INT_STATUS__SIT_ERR_INT_MASK 0x00000800L +#define UVD_SUVD_INT_STATUS__SMP_FUNC_INT_MASK 0x0001F000L +#define UVD_SUVD_INT_STATUS__SMP_ERR_INT_MASK 0x00020000L +#define UVD_SUVD_INT_STATUS__SCM_FUNC_INT_MASK 0x007C0000L +#define UVD_SUVD_INT_STATUS__SCM_ERR_INT_MASK 0x00800000L +#define UVD_SUVD_INT_STATUS__SDB_FUNC_INT_MASK 0x1F000000L +#define UVD_SUVD_INT_STATUS__SDB_ERR_INT_MASK 0x20000000L +#define UVD_SUVD_INT_STATUS__FBC_ERR_INT_MASK 0x40000000L +//UVD_SUVD_INT_ACK +#define UVD_SUVD_INT_ACK__SRE_FUNC_INT_ACK__SHIFT 0x0 +#define UVD_SUVD_INT_ACK__SRE_ERR_INT_ACK__SHIFT 0x5 +#define UVD_SUVD_INT_ACK__SIT_FUNC_INT_ACK__SHIFT 0x6 +#define UVD_SUVD_INT_ACK__SIT_ERR_INT_ACK__SHIFT 0xb +#define UVD_SUVD_INT_ACK__SMP_FUNC_INT_ACK__SHIFT 0xc +#define UVD_SUVD_INT_ACK__SMP_ERR_INT_ACK__SHIFT 0x11 +#define UVD_SUVD_INT_ACK__SCM_FUNC_INT_ACK__SHIFT 0x12 +#define UVD_SUVD_INT_ACK__SCM_ERR_INT_ACK__SHIFT 0x17 +#define UVD_SUVD_INT_ACK__SDB_FUNC_INT_ACK__SHIFT 0x18 +#define UVD_SUVD_INT_ACK__SDB_ERR_INT_ACK__SHIFT 0x1d +#define UVD_SUVD_INT_ACK__FBC_ERR_INT_ACK__SHIFT 0x1e +#define UVD_SUVD_INT_ACK__SRE_FUNC_INT_ACK_MASK 0x0000001FL +#define UVD_SUVD_INT_ACK__SRE_ERR_INT_ACK_MASK 0x00000020L +#define UVD_SUVD_INT_ACK__SIT_FUNC_INT_ACK_MASK 0x000007C0L +#define UVD_SUVD_INT_ACK__SIT_ERR_INT_ACK_MASK 0x00000800L +#define UVD_SUVD_INT_ACK__SMP_FUNC_INT_ACK_MASK 0x0001F000L +#define UVD_SUVD_INT_ACK__SMP_ERR_INT_ACK_MASK 0x00020000L +#define UVD_SUVD_INT_ACK__SCM_FUNC_INT_ACK_MASK 0x007C0000L +#define UVD_SUVD_INT_ACK__SCM_ERR_INT_ACK_MASK 0x00800000L +#define UVD_SUVD_INT_ACK__SDB_FUNC_INT_ACK_MASK 0x1F000000L +#define UVD_SUVD_INT_ACK__SDB_ERR_INT_ACK_MASK 0x20000000L +#define UVD_SUVD_INT_ACK__FBC_ERR_INT_ACK_MASK 0x40000000L +//UVD_ENC_VCPU_INT_EN +#define UVD_ENC_VCPU_INT_EN__DCE_UVD_SCAN_IN_BUFMGR_EN__SHIFT 0x0 +#define UVD_ENC_VCPU_INT_EN__DCE_UVD_SCAN_IN_BUFMGR2_EN__SHIFT 0x1 +#define UVD_ENC_VCPU_INT_EN__DCE_UVD_SCAN_IN_BUFMGR3_EN__SHIFT 0x2 +#define UVD_ENC_VCPU_INT_EN__DCE_UVD_SCAN_IN_BUFMGR_EN_MASK 0x00000001L +#define UVD_ENC_VCPU_INT_EN__DCE_UVD_SCAN_IN_BUFMGR2_EN_MASK 0x00000002L +#define UVD_ENC_VCPU_INT_EN__DCE_UVD_SCAN_IN_BUFMGR3_EN_MASK 0x00000004L +//UVD_ENC_VCPU_INT_STATUS +#define UVD_ENC_VCPU_INT_STATUS__DCE_UVD_SCAN_IN_BUFMGR_INT__SHIFT 0x0 +#define UVD_ENC_VCPU_INT_STATUS__DCE_UVD_SCAN_IN_BUFMGR2_INT__SHIFT 0x1 +#define UVD_ENC_VCPU_INT_STATUS__DCE_UVD_SCAN_IN_BUFMGR3_INT__SHIFT 0x2 +#define UVD_ENC_VCPU_INT_STATUS__DCE_UVD_SCAN_IN_BUFMGR_INT_MASK 0x00000001L +#define UVD_ENC_VCPU_INT_STATUS__DCE_UVD_SCAN_IN_BUFMGR2_INT_MASK 0x00000002L +#define UVD_ENC_VCPU_INT_STATUS__DCE_UVD_SCAN_IN_BUFMGR3_INT_MASK 0x00000004L +//UVD_ENC_VCPU_INT_ACK +#define UVD_ENC_VCPU_INT_ACK__DCE_UVD_SCAN_IN_BUFMGR_ACK__SHIFT 0x0 +#define UVD_ENC_VCPU_INT_ACK__DCE_UVD_SCAN_IN_BUFMGR2_ACK__SHIFT 0x1 +#define UVD_ENC_VCPU_INT_ACK__DCE_UVD_SCAN_IN_BUFMGR3_ACK__SHIFT 0x2 +#define UVD_ENC_VCPU_INT_ACK__DCE_UVD_SCAN_IN_BUFMGR_ACK_MASK 0x00000001L +#define UVD_ENC_VCPU_INT_ACK__DCE_UVD_SCAN_IN_BUFMGR2_ACK_MASK 0x00000002L +#define UVD_ENC_VCPU_INT_ACK__DCE_UVD_SCAN_IN_BUFMGR3_ACK_MASK 0x00000004L +//UVD_MASTINT_EN +#define UVD_MASTINT_EN__OVERRUN_RST__SHIFT 0x0 +#define UVD_MASTINT_EN__VCPU_EN__SHIFT 0x1 +#define UVD_MASTINT_EN__SYS_EN__SHIFT 0x2 +#define UVD_MASTINT_EN__INT_OVERRUN__SHIFT 0x4 +#define UVD_MASTINT_EN__OVERRUN_RST_MASK 0x00000001L +#define UVD_MASTINT_EN__VCPU_EN_MASK 0x00000002L +#define UVD_MASTINT_EN__SYS_EN_MASK 0x00000004L +#define UVD_MASTINT_EN__INT_OVERRUN_MASK 0x00FFFFF0L +//UVD_SYS_INT_EN +#define UVD_SYS_INT_EN__PIF_ADDR_ERR_EN__SHIFT 0x0 +#define UVD_SYS_INT_EN__SEMA_WAIT_FAULT_TIMEOUT_EN__SHIFT 0x1 +#define UVD_SYS_INT_EN__SEMA_SIGNAL_INCOMPLETE_TIMEOUT_EN__SHIFT 0x2 +#define UVD_SYS_INT_EN__CXW_WR_EN__SHIFT 0x3 +#define UVD_SYS_INT_EN__RBC_REG_PRIV_FAULT_EN__SHIFT 0x6 +#define UVD_SYS_INT_EN__LBSI_EN__SHIFT 0xb +#define UVD_SYS_INT_EN__UDEC_EN__SHIFT 0xc +#define UVD_SYS_INT_EN__LMI_AXI_UNSUPPORTED_LEN_EN__SHIFT 0xd +#define UVD_SYS_INT_EN__LMI_AXI_UNSUPPORTED_ADR_ALIGN_EN__SHIFT 0xe +#define UVD_SYS_INT_EN__SUVD_EN__SHIFT 0xf +#define UVD_SYS_INT_EN__JOB_DONE_EN__SHIFT 0x10 +#define UVD_SYS_INT_EN__SEMA_WAIT_FAIL_SIG_EN__SHIFT 0x17 +#define UVD_SYS_INT_EN__IDCT_EN__SHIFT 0x18 +#define UVD_SYS_INT_EN__MPRD_EN__SHIFT 0x19 +#define UVD_SYS_INT_EN__RASCNTL_VCPU_VCODEC_EN__SHIFT 0x1a +#define UVD_SYS_INT_EN__CLK_SWT_EN__SHIFT 0x1b +#define UVD_SYS_INT_EN__MIF_HWINT_EN__SHIFT 0x1c +#define UVD_SYS_INT_EN__MPRD_ERR_EN__SHIFT 0x1d +#define UVD_SYS_INT_EN__AVM_INT_EN__SHIFT 0x1f +#define UVD_SYS_INT_EN__PIF_ADDR_ERR_EN_MASK 0x00000001L +#define UVD_SYS_INT_EN__SEMA_WAIT_FAULT_TIMEOUT_EN_MASK 0x00000002L +#define UVD_SYS_INT_EN__SEMA_SIGNAL_INCOMPLETE_TIMEOUT_EN_MASK 0x00000004L +#define UVD_SYS_INT_EN__CXW_WR_EN_MASK 0x00000008L +#define UVD_SYS_INT_EN__RBC_REG_PRIV_FAULT_EN_MASK 0x00000040L +#define UVD_SYS_INT_EN__LBSI_EN_MASK 0x00000800L +#define UVD_SYS_INT_EN__UDEC_EN_MASK 0x00001000L +#define UVD_SYS_INT_EN__LMI_AXI_UNSUPPORTED_LEN_EN_MASK 0x00002000L +#define UVD_SYS_INT_EN__LMI_AXI_UNSUPPORTED_ADR_ALIGN_EN_MASK 0x00004000L +#define UVD_SYS_INT_EN__SUVD_EN_MASK 0x00008000L +#define UVD_SYS_INT_EN__JOB_DONE_EN_MASK 0x00010000L +#define UVD_SYS_INT_EN__SEMA_WAIT_FAIL_SIG_EN_MASK 0x00800000L +#define UVD_SYS_INT_EN__IDCT_EN_MASK 0x01000000L +#define UVD_SYS_INT_EN__MPRD_EN_MASK 0x02000000L +#define UVD_SYS_INT_EN__RASCNTL_VCPU_VCODEC_EN_MASK 0x04000000L +#define UVD_SYS_INT_EN__CLK_SWT_EN_MASK 0x08000000L +#define UVD_SYS_INT_EN__MIF_HWINT_EN_MASK 0x10000000L +#define UVD_SYS_INT_EN__MPRD_ERR_EN_MASK 0x20000000L +#define UVD_SYS_INT_EN__AVM_INT_EN_MASK 0x80000000L +//UVD_SYS_INT_STATUS +#define UVD_SYS_INT_STATUS__PIF_ADDR_ERR_INT__SHIFT 0x0 +#define UVD_SYS_INT_STATUS__SEMA_WAIT_FAULT_TIMEOUT_INT__SHIFT 0x1 +#define UVD_SYS_INT_STATUS__SEMA_SIGNAL_INCOMPLETE_TIMEOUT_INT__SHIFT 0x2 +#define UVD_SYS_INT_STATUS__CXW_WR_INT__SHIFT 0x3 +#define UVD_SYS_INT_STATUS__RBC_REG_PRIV_FAULT_INT__SHIFT 0x6 +#define UVD_SYS_INT_STATUS__LBSI_INT__SHIFT 0xb +#define UVD_SYS_INT_STATUS__UDEC_INT__SHIFT 0xc +#define UVD_SYS_INT_STATUS__LMI_AXI_UNSUPPORTED_LEN_INT__SHIFT 0xd +#define UVD_SYS_INT_STATUS__LMI_AXI_UNSUPPORTED_ADR_ALIGN_INT__SHIFT 0xe +#define UVD_SYS_INT_STATUS__SUVD_INT__SHIFT 0xf +#define UVD_SYS_INT_STATUS__JOB_DONE_INT__SHIFT 0x10 +#define UVD_SYS_INT_STATUS__GPCOM_INT__SHIFT 0x12 +#define UVD_SYS_INT_STATUS__SEMA_WAIT_FAIL_SIG_INT__SHIFT 0x17 +#define UVD_SYS_INT_STATUS__IDCT_INT__SHIFT 0x18 +#define UVD_SYS_INT_STATUS__MPRD_INT__SHIFT 0x19 +#define UVD_SYS_INT_STATUS__CLK_SWT_INT__SHIFT 0x1b +#define UVD_SYS_INT_STATUS__MIF_HWINT__SHIFT 0x1c +#define UVD_SYS_INT_STATUS__MPRD_ERR_INT__SHIFT 0x1d +#define UVD_SYS_INT_STATUS__RASCNTL_VCPU_VCODEC_INT__SHIFT 0x1e +#define UVD_SYS_INT_STATUS__AVM_INT__SHIFT 0x1f +#define UVD_SYS_INT_STATUS__PIF_ADDR_ERR_INT_MASK 0x00000001L +#define UVD_SYS_INT_STATUS__SEMA_WAIT_FAULT_TIMEOUT_INT_MASK 0x00000002L +#define UVD_SYS_INT_STATUS__SEMA_SIGNAL_INCOMPLETE_TIMEOUT_INT_MASK 0x00000004L +#define UVD_SYS_INT_STATUS__CXW_WR_INT_MASK 0x00000008L +#define UVD_SYS_INT_STATUS__RBC_REG_PRIV_FAULT_INT_MASK 0x00000040L +#define UVD_SYS_INT_STATUS__LBSI_INT_MASK 0x00000800L +#define UVD_SYS_INT_STATUS__UDEC_INT_MASK 0x00001000L +#define UVD_SYS_INT_STATUS__LMI_AXI_UNSUPPORTED_LEN_INT_MASK 0x00002000L +#define UVD_SYS_INT_STATUS__LMI_AXI_UNSUPPORTED_ADR_ALIGN_INT_MASK 0x00004000L +#define UVD_SYS_INT_STATUS__SUVD_INT_MASK 0x00008000L +#define UVD_SYS_INT_STATUS__JOB_DONE_INT_MASK 0x00010000L +#define UVD_SYS_INT_STATUS__GPCOM_INT_MASK 0x00040000L +#define UVD_SYS_INT_STATUS__SEMA_WAIT_FAIL_SIG_INT_MASK 0x00800000L +#define UVD_SYS_INT_STATUS__IDCT_INT_MASK 0x01000000L +#define UVD_SYS_INT_STATUS__MPRD_INT_MASK 0x02000000L +#define UVD_SYS_INT_STATUS__CLK_SWT_INT_MASK 0x08000000L +#define UVD_SYS_INT_STATUS__MIF_HWINT_MASK 0x10000000L +#define UVD_SYS_INT_STATUS__MPRD_ERR_INT_MASK 0x20000000L +#define UVD_SYS_INT_STATUS__RASCNTL_VCPU_VCODEC_INT_MASK 0x40000000L +#define UVD_SYS_INT_STATUS__AVM_INT_MASK 0x80000000L +//UVD_SYS_INT_ACK +#define UVD_SYS_INT_ACK__PIF_ADDR_ERR_ACK__SHIFT 0x0 +#define UVD_SYS_INT_ACK__SEMA_WAIT_FAULT_TIMEOUT_ACK__SHIFT 0x1 +#define UVD_SYS_INT_ACK__SEMA_SIGNAL_INCOMPLETE_TIMEOUT_ACK__SHIFT 0x2 +#define UVD_SYS_INT_ACK__CXW_WR_ACK__SHIFT 0x3 +#define UVD_SYS_INT_ACK__RBC_REG_PRIV_FAULT_ACK__SHIFT 0x6 +#define UVD_SYS_INT_ACK__LBSI_ACK__SHIFT 0xb +#define UVD_SYS_INT_ACK__UDEC_ACK__SHIFT 0xc +#define UVD_SYS_INT_ACK__LMI_AXI_UNSUPPORTED_LEN_ACK__SHIFT 0xd +#define UVD_SYS_INT_ACK__LMI_AXI_UNSUPPORTED_ADR_ALIGN_ACK__SHIFT 0xe +#define UVD_SYS_INT_ACK__SUVD_ACK__SHIFT 0xf +#define UVD_SYS_INT_ACK__JOB_DONE_ACK__SHIFT 0x10 +#define UVD_SYS_INT_ACK__SEMA_WAIT_FAIL_SIG_ACK__SHIFT 0x17 +#define UVD_SYS_INT_ACK__IDCT_ACK__SHIFT 0x18 +#define UVD_SYS_INT_ACK__MPRD_ACK__SHIFT 0x19 +#define UVD_SYS_INT_ACK__CLK_SWT_ACK__SHIFT 0x1b +#define UVD_SYS_INT_ACK__MIF_HWINT_ACK__SHIFT 0x1c +#define UVD_SYS_INT_ACK__MPRD_ERR_ACK__SHIFT 0x1d +#define UVD_SYS_INT_ACK__RASCNTL_VCPU_VCODEC_ACK__SHIFT 0x1e +#define UVD_SYS_INT_ACK__AVM_INT_ACK__SHIFT 0x1f +#define UVD_SYS_INT_ACK__PIF_ADDR_ERR_ACK_MASK 0x00000001L +#define UVD_SYS_INT_ACK__SEMA_WAIT_FAULT_TIMEOUT_ACK_MASK 0x00000002L +#define UVD_SYS_INT_ACK__SEMA_SIGNAL_INCOMPLETE_TIMEOUT_ACK_MASK 0x00000004L +#define UVD_SYS_INT_ACK__CXW_WR_ACK_MASK 0x00000008L +#define UVD_SYS_INT_ACK__RBC_REG_PRIV_FAULT_ACK_MASK 0x00000040L +#define UVD_SYS_INT_ACK__LBSI_ACK_MASK 0x00000800L +#define UVD_SYS_INT_ACK__UDEC_ACK_MASK 0x00001000L +#define UVD_SYS_INT_ACK__LMI_AXI_UNSUPPORTED_LEN_ACK_MASK 0x00002000L +#define UVD_SYS_INT_ACK__LMI_AXI_UNSUPPORTED_ADR_ALIGN_ACK_MASK 0x00004000L +#define UVD_SYS_INT_ACK__SUVD_ACK_MASK 0x00008000L +#define UVD_SYS_INT_ACK__JOB_DONE_ACK_MASK 0x00010000L +#define UVD_SYS_INT_ACK__SEMA_WAIT_FAIL_SIG_ACK_MASK 0x00800000L +#define UVD_SYS_INT_ACK__IDCT_ACK_MASK 0x01000000L +#define UVD_SYS_INT_ACK__MPRD_ACK_MASK 0x02000000L +#define UVD_SYS_INT_ACK__CLK_SWT_ACK_MASK 0x08000000L +#define UVD_SYS_INT_ACK__MIF_HWINT_ACK_MASK 0x10000000L +#define UVD_SYS_INT_ACK__MPRD_ERR_ACK_MASK 0x20000000L +#define UVD_SYS_INT_ACK__RASCNTL_VCPU_VCODEC_ACK_MASK 0x40000000L +#define UVD_SYS_INT_ACK__AVM_INT_ACK_MASK 0x80000000L +//UVD_JOB_DONE +#define UVD_JOB_DONE__JOB_DONE__SHIFT 0x0 +#define UVD_JOB_DONE__JOB_DONE_MASK 0x00000003L +//UVD_CBUF_ID +#define UVD_CBUF_ID__CBUF_ID__SHIFT 0x0 +#define UVD_CBUF_ID__CBUF_ID_MASK 0xFFFFFFFFL +//UVD_CONTEXT_ID +#define UVD_CONTEXT_ID__CONTEXT_ID__SHIFT 0x0 +#define UVD_CONTEXT_ID__CONTEXT_ID_MASK 0xFFFFFFFFL +//UVD_CONTEXT_ID2 +#define UVD_CONTEXT_ID2__CONTEXT_ID2__SHIFT 0x0 +#define UVD_CONTEXT_ID2__CONTEXT_ID2_MASK 0xFFFFFFFFL +//UVD_NO_OP +#define UVD_NO_OP__NO_OP__SHIFT 0x0 +#define UVD_NO_OP__NO_OP_MASK 0xFFFFFFFFL +//UVD_RB_BASE_LO +#define UVD_RB_BASE_LO__RB_BASE_LO__SHIFT 0x6 +#define UVD_RB_BASE_LO__RB_BASE_LO_MASK 0xFFFFFFC0L +//UVD_RB_BASE_HI +#define UVD_RB_BASE_HI__RB_BASE_HI__SHIFT 0x0 +#define UVD_RB_BASE_HI__RB_BASE_HI_MASK 0xFFFFFFFFL +//UVD_RB_SIZE +#define UVD_RB_SIZE__RB_SIZE__SHIFT 0x4 +#define UVD_RB_SIZE__RB_SIZE_MASK 0x007FFFF0L +//UVD_RB_BASE_LO2 +#define UVD_RB_BASE_LO2__RB_BASE_LO__SHIFT 0x6 +#define UVD_RB_BASE_LO2__RB_BASE_LO_MASK 0xFFFFFFC0L +//UVD_RB_BASE_HI2 +#define UVD_RB_BASE_HI2__RB_BASE_HI__SHIFT 0x0 +#define UVD_RB_BASE_HI2__RB_BASE_HI_MASK 0xFFFFFFFFL +//UVD_RB_SIZE2 +#define UVD_RB_SIZE2__RB_SIZE__SHIFT 0x4 +#define UVD_RB_SIZE2__RB_SIZE_MASK 0x007FFFF0L +//UVD_RB_BASE_LO3 +#define UVD_RB_BASE_LO3__RB_BASE_LO__SHIFT 0x6 +#define UVD_RB_BASE_LO3__RB_BASE_LO_MASK 0xFFFFFFC0L +//UVD_RB_BASE_HI3 +#define UVD_RB_BASE_HI3__RB_BASE_HI__SHIFT 0x0 +#define UVD_RB_BASE_HI3__RB_BASE_HI_MASK 0xFFFFFFFFL +//UVD_RB_SIZE3 +#define UVD_RB_SIZE3__RB_SIZE__SHIFT 0x4 +#define UVD_RB_SIZE3__RB_SIZE_MASK 0x007FFFF0L +//UVD_RB_BASE_LO4 +#define UVD_RB_BASE_LO4__RB_BASE_LO__SHIFT 0x6 +#define UVD_RB_BASE_LO4__RB_BASE_LO_MASK 0xFFFFFFC0L +//UVD_RB_BASE_HI4 +#define UVD_RB_BASE_HI4__RB_BASE_HI__SHIFT 0x0 +#define UVD_RB_BASE_HI4__RB_BASE_HI_MASK 0xFFFFFFFFL +//UVD_RB_SIZE4 +#define UVD_RB_SIZE4__RB_SIZE__SHIFT 0x4 +#define UVD_RB_SIZE4__RB_SIZE_MASK 0x007FFFF0L +//UVD_OUT_RB_BASE_LO +#define UVD_OUT_RB_BASE_LO__RB_BASE_LO__SHIFT 0x6 +#define UVD_OUT_RB_BASE_LO__RB_BASE_LO_MASK 0xFFFFFFC0L +//UVD_OUT_RB_BASE_HI +#define UVD_OUT_RB_BASE_HI__RB_BASE_HI__SHIFT 0x0 +#define UVD_OUT_RB_BASE_HI__RB_BASE_HI_MASK 0xFFFFFFFFL +//UVD_OUT_RB_SIZE +#define UVD_OUT_RB_SIZE__RB_SIZE__SHIFT 0x4 +#define UVD_OUT_RB_SIZE__RB_SIZE_MASK 0x007FFFF0L +//UVD_IOV_ACTIVE_FCN_ID +#define UVD_IOV_ACTIVE_FCN_ID__VF_ID__SHIFT 0x0 +#define UVD_IOV_ACTIVE_FCN_ID__PF_VF__SHIFT 0x1f +#define UVD_IOV_ACTIVE_FCN_ID__VF_ID_MASK 0x0000003FL +#define UVD_IOV_ACTIVE_FCN_ID__PF_VF_MASK 0x80000000L +//UVD_IOV_MAILBOX +#define UVD_IOV_MAILBOX__MAILBOX__SHIFT 0x0 +#define UVD_IOV_MAILBOX__MAILBOX_MASK 0xFFFFFFFFL +//UVD_IOV_MAILBOX_RESP +#define UVD_IOV_MAILBOX_RESP__RESP__SHIFT 0x0 +#define UVD_IOV_MAILBOX_RESP__RESP_MASK 0xFFFFFFFFL +//UVD_RB_ARB_CTRL +#define UVD_RB_ARB_CTRL__SRBM_DROP__SHIFT 0x0 +#define UVD_RB_ARB_CTRL__SRBM_DIS__SHIFT 0x1 +#define UVD_RB_ARB_CTRL__VCPU_DROP__SHIFT 0x2 +#define UVD_RB_ARB_CTRL__VCPU_DIS__SHIFT 0x3 +#define UVD_RB_ARB_CTRL__RBC_DROP__SHIFT 0x4 +#define UVD_RB_ARB_CTRL__RBC_DIS__SHIFT 0x5 +#define UVD_RB_ARB_CTRL__FWOFLD_DROP__SHIFT 0x6 +#define UVD_RB_ARB_CTRL__FWOFLD_DIS__SHIFT 0x7 +#define UVD_RB_ARB_CTRL__FAST_PATH_EN__SHIFT 0x8 +#define UVD_RB_ARB_CTRL__UVD_RB_DBG_EN__SHIFT 0x9 +#define UVD_RB_ARB_CTRL__SRBM_DROP_MASK 0x00000001L +#define UVD_RB_ARB_CTRL__SRBM_DIS_MASK 0x00000002L +#define UVD_RB_ARB_CTRL__VCPU_DROP_MASK 0x00000004L +#define UVD_RB_ARB_CTRL__VCPU_DIS_MASK 0x00000008L +#define UVD_RB_ARB_CTRL__RBC_DROP_MASK 0x00000010L +#define UVD_RB_ARB_CTRL__RBC_DIS_MASK 0x00000020L +#define UVD_RB_ARB_CTRL__FWOFLD_DROP_MASK 0x00000040L +#define UVD_RB_ARB_CTRL__FWOFLD_DIS_MASK 0x00000080L +#define UVD_RB_ARB_CTRL__FAST_PATH_EN_MASK 0x00000100L +#define UVD_RB_ARB_CTRL__UVD_RB_DBG_EN_MASK 0x00000200L +//UVD_CTX_INDEX +#define UVD_CTX_INDEX__INDEX__SHIFT 0x0 +#define UVD_CTX_INDEX__INDEX_MASK 0x000001FFL +//UVD_CTX_DATA +#define UVD_CTX_DATA__DATA__SHIFT 0x0 +#define UVD_CTX_DATA__DATA_MASK 0xFFFFFFFFL +//UVD_CXW_WR +#define UVD_CXW_WR__DAT__SHIFT 0x0 +#define UVD_CXW_WR__STAT__SHIFT 0x1f +#define UVD_CXW_WR__DAT_MASK 0x0FFFFFFFL +#define UVD_CXW_WR__STAT_MASK 0x80000000L +//UVD_CXW_WR_INT_ID +#define UVD_CXW_WR_INT_ID__ID__SHIFT 0x0 +#define UVD_CXW_WR_INT_ID__ID_MASK 0x000000FFL +//UVD_CXW_WR_INT_CTX_ID +#define UVD_CXW_WR_INT_CTX_ID__ID__SHIFT 0x0 +#define UVD_CXW_WR_INT_CTX_ID__ID_MASK 0x0FFFFFFFL +//UVD_CXW_INT_ID +#define UVD_CXW_INT_ID__ID__SHIFT 0x0 +#define UVD_CXW_INT_ID__ID_MASK 0x000000FFL +//UVD_MPEG2_ERROR +#define UVD_MPEG2_ERROR__STATUS__SHIFT 0x0 +#define UVD_MPEG2_ERROR__STATUS_MASK 0xFFFFFFFFL +//UVD_YBASE +#define UVD_YBASE__DUM__SHIFT 0x0 +#define UVD_YBASE__DUM_MASK 0xFFFFFFFFL +//UVD_UVBASE +#define UVD_UVBASE__DUM__SHIFT 0x0 +#define UVD_UVBASE__DUM_MASK 0xFFFFFFFFL +//UVD_PITCH +#define UVD_PITCH__DUM__SHIFT 0x0 +#define UVD_PITCH__DUM_MASK 0xFFFFFFFFL +//UVD_WIDTH +#define UVD_WIDTH__DUM__SHIFT 0x0 +#define UVD_WIDTH__DUM_MASK 0xFFFFFFFFL +//UVD_HEIGHT +#define UVD_HEIGHT__DUM__SHIFT 0x0 +#define UVD_HEIGHT__DUM_MASK 0xFFFFFFFFL +//UVD_PICCOUNT +#define UVD_PICCOUNT__DUM__SHIFT 0x0 +#define UVD_PICCOUNT__DUM_MASK 0xFFFFFFFFL +//UVD_MPRD_INITIAL_XY +#define UVD_MPRD_INITIAL_XY__MPRD_SCREEN_X__SHIFT 0x0 +#define UVD_MPRD_INITIAL_XY__MPRD_SCREEN_Y__SHIFT 0x10 +#define UVD_MPRD_INITIAL_XY__MPRD_SCREEN_X_MASK 0x00000FFFL +#define UVD_MPRD_INITIAL_XY__MPRD_SCREEN_Y_MASK 0x0FFF0000L +//UVD_MPEG2_CTRL +#define UVD_MPEG2_CTRL__EN__SHIFT 0x0 +#define UVD_MPEG2_CTRL__TRICK_MODE__SHIFT 0x1 +#define UVD_MPEG2_CTRL__NUM_MB_PER_JOB__SHIFT 0x10 +#define UVD_MPEG2_CTRL__EN_MASK 0x00000001L +#define UVD_MPEG2_CTRL__TRICK_MODE_MASK 0x00000002L +#define UVD_MPEG2_CTRL__NUM_MB_PER_JOB_MASK 0xFFFF0000L +//UVD_MB_CTL_BUF_BASE +#define UVD_MB_CTL_BUF_BASE__BASE__SHIFT 0x0 +#define UVD_MB_CTL_BUF_BASE__BASE_MASK 0xFFFFFFFFL +//UVD_PIC_CTL_BUF_BASE +#define UVD_PIC_CTL_BUF_BASE__BASE__SHIFT 0x0 +#define UVD_PIC_CTL_BUF_BASE__BASE_MASK 0xFFFFFFFFL +//UVD_DXVA_BUF_SIZE +#define UVD_DXVA_BUF_SIZE__PIC_SIZE__SHIFT 0x0 +#define UVD_DXVA_BUF_SIZE__MB_SIZE__SHIFT 0x10 +#define UVD_DXVA_BUF_SIZE__PIC_SIZE_MASK 0x0000FFFFL +#define UVD_DXVA_BUF_SIZE__MB_SIZE_MASK 0xFFFF0000L +//UVD_SCRATCH_NP +#define UVD_SCRATCH_NP__DATA__SHIFT 0x0 +#define UVD_SCRATCH_NP__DATA_MASK 0xFFFFFFFFL +//UVD_CLK_SWT_HANDSHAKE +#define UVD_CLK_SWT_HANDSHAKE__CLK_SWT_TYPE__SHIFT 0x0 +#define UVD_CLK_SWT_HANDSHAKE__CLK_DOMAIN_SWT__SHIFT 0x8 +#define UVD_CLK_SWT_HANDSHAKE__CLK_SWT_TYPE_MASK 0x00000003L +#define UVD_CLK_SWT_HANDSHAKE__CLK_DOMAIN_SWT_MASK 0x00000300L +//UVD_GP_SCRATCH0 +#define UVD_GP_SCRATCH0__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH0__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH1 +#define UVD_GP_SCRATCH1__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH1__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH2 +#define UVD_GP_SCRATCH2__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH2__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH3 +#define UVD_GP_SCRATCH3__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH3__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH4 +#define UVD_GP_SCRATCH4__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH4__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH5 +#define UVD_GP_SCRATCH5__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH5__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH6 +#define UVD_GP_SCRATCH6__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH6__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH7 +#define UVD_GP_SCRATCH7__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH7__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH8 +#define UVD_GP_SCRATCH8__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH8__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH9 +#define UVD_GP_SCRATCH9__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH9__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH10 +#define UVD_GP_SCRATCH10__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH10__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH11 +#define UVD_GP_SCRATCH11__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH11__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH12 +#define UVD_GP_SCRATCH12__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH12__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH13 +#define UVD_GP_SCRATCH13__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH13__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH14 +#define UVD_GP_SCRATCH14__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH14__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH15 +#define UVD_GP_SCRATCH15__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH15__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH16 +#define UVD_GP_SCRATCH16__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH16__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH17 +#define UVD_GP_SCRATCH17__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH17__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH18 +#define UVD_GP_SCRATCH18__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH18__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH19 +#define UVD_GP_SCRATCH19__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH19__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH20 +#define UVD_GP_SCRATCH20__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH20__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH21 +#define UVD_GP_SCRATCH21__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH21__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH22 +#define UVD_GP_SCRATCH22__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH22__DATA_MASK 0xFFFFFFFFL +//UVD_GP_SCRATCH23 +#define UVD_GP_SCRATCH23__DATA__SHIFT 0x0 +#define UVD_GP_SCRATCH23__DATA_MASK 0xFFFFFFFFL +//UVD_AUDIO_RB_BASE_LO +#define UVD_AUDIO_RB_BASE_LO__RB_BASE_LO__SHIFT 0x6 +#define UVD_AUDIO_RB_BASE_LO__RB_BASE_LO_MASK 0xFFFFFFC0L +//UVD_AUDIO_RB_BASE_HI +#define UVD_AUDIO_RB_BASE_HI__RB_BASE_HI__SHIFT 0x0 +#define UVD_AUDIO_RB_BASE_HI__RB_BASE_HI_MASK 0xFFFFFFFFL +//UVD_AUDIO_RB_SIZE +#define UVD_AUDIO_RB_SIZE__RB_SIZE__SHIFT 0x4 +#define UVD_AUDIO_RB_SIZE__RB_SIZE_MASK 0x007FFFF0L +//UVD_VCPU_INT_STATUS2 +#define UVD_VCPU_INT_STATUS2__SW_RB6_INT__SHIFT 0x0 +#define UVD_VCPU_INT_STATUS2__RASCNTL_VCPU_VCODEC_INT__SHIFT 0x15 +#define UVD_VCPU_INT_STATUS2__SW_RB6_INT_MASK 0x00000001L +#define UVD_VCPU_INT_STATUS2__RASCNTL_VCPU_VCODEC_INT_MASK 0x00200000L +//UVD_VCPU_INT_ACK2 +#define UVD_VCPU_INT_ACK2__SW_RB6_INT_ACK__SHIFT 0x0 +#define UVD_VCPU_INT_ACK2__RASCNTL_VCPU_VCODEC_ACK__SHIFT 0x16 +#define UVD_VCPU_INT_ACK2__SW_RB6_INT_ACK_MASK 0x00000001L +#define UVD_VCPU_INT_ACK2__RASCNTL_VCPU_VCODEC_ACK_MASK 0x00400000L +//UVD_VCPU_INT_EN2 +#define UVD_VCPU_INT_EN2__SW_RB6_INT_EN__SHIFT 0x0 +#define UVD_VCPU_INT_EN2__RASCNTL_VCPU_VCODEC_EN__SHIFT 0x1 +#define UVD_VCPU_INT_EN2__SW_RB6_INT_EN_MASK 0x00000001L +#define UVD_VCPU_INT_EN2__RASCNTL_VCPU_VCODEC_EN_MASK 0x00000002L +//UVD_SUVD_CGC_STATUS2 +#define UVD_SUVD_CGC_STATUS2__SMPA_VCLK__SHIFT 0x0 +#define UVD_SUVD_CGC_STATUS2__SMPA_DCLK__SHIFT 0x1 +#define UVD_SUVD_CGC_STATUS2__MPBE1_DCLK__SHIFT 0x3 +#define UVD_SUVD_CGC_STATUS2__SIT_AV1_DCLK__SHIFT 0x4 +#define UVD_SUVD_CGC_STATUS2__SDB_AV1_DCLK__SHIFT 0x5 +#define UVD_SUVD_CGC_STATUS2__MPC1_DCLK__SHIFT 0x6 +#define UVD_SUVD_CGC_STATUS2__MPC1_SCLK__SHIFT 0x7 +#define UVD_SUVD_CGC_STATUS2__MPC1_VCLK__SHIFT 0x8 +#define UVD_SUVD_CGC_STATUS2__SRE_AV1_ENC_DCLK__SHIFT 0x9 +#define UVD_SUVD_CGC_STATUS2__CDEFE_DCLK__SHIFT 0xa +#define UVD_SUVD_CGC_STATUS2__SIT0_DCLK__SHIFT 0xb +#define UVD_SUVD_CGC_STATUS2__SIT1_DCLK__SHIFT 0xc +#define UVD_SUVD_CGC_STATUS2__SIT2_DCLK__SHIFT 0xd +#define UVD_SUVD_CGC_STATUS2__FBC_PCLK__SHIFT 0x1c +#define UVD_SUVD_CGC_STATUS2__FBC_CCLK__SHIFT 0x1d +#define UVD_SUVD_CGC_STATUS2__SMPA_VCLK_MASK 0x00000001L +#define UVD_SUVD_CGC_STATUS2__SMPA_DCLK_MASK 0x00000002L +#define UVD_SUVD_CGC_STATUS2__MPBE1_DCLK_MASK 0x00000008L +#define UVD_SUVD_CGC_STATUS2__SIT_AV1_DCLK_MASK 0x00000010L +#define UVD_SUVD_CGC_STATUS2__SDB_AV1_DCLK_MASK 0x00000020L +#define UVD_SUVD_CGC_STATUS2__MPC1_DCLK_MASK 0x00000040L +#define UVD_SUVD_CGC_STATUS2__MPC1_SCLK_MASK 0x00000080L +#define UVD_SUVD_CGC_STATUS2__MPC1_VCLK_MASK 0x00000100L +#define UVD_SUVD_CGC_STATUS2__SRE_AV1_ENC_DCLK_MASK 0x00000200L +#define UVD_SUVD_CGC_STATUS2__CDEFE_DCLK_MASK 0x00000400L +#define UVD_SUVD_CGC_STATUS2__SIT0_DCLK_MASK 0x00000800L +#define UVD_SUVD_CGC_STATUS2__SIT1_DCLK_MASK 0x00001000L +#define UVD_SUVD_CGC_STATUS2__SIT2_DCLK_MASK 0x00002000L +#define UVD_SUVD_CGC_STATUS2__FBC_PCLK_MASK 0x10000000L +#define UVD_SUVD_CGC_STATUS2__FBC_CCLK_MASK 0x20000000L +//UVD_SUVD_INT_STATUS2 +#define UVD_SUVD_INT_STATUS2__SMPA_FUNC_INT__SHIFT 0x0 +#define UVD_SUVD_INT_STATUS2__SMPA_ERR_INT__SHIFT 0x5 +#define UVD_SUVD_INT_STATUS2__SDB_AV1_FUNC_INT__SHIFT 0x6 +#define UVD_SUVD_INT_STATUS2__SDB_AV1_ERR_INT__SHIFT 0xb +#define UVD_SUVD_INT_STATUS2__SMPA_FUNC_INT_MASK 0x0000001FL +#define UVD_SUVD_INT_STATUS2__SMPA_ERR_INT_MASK 0x00000020L +#define UVD_SUVD_INT_STATUS2__SDB_AV1_FUNC_INT_MASK 0x000007C0L +#define UVD_SUVD_INT_STATUS2__SDB_AV1_ERR_INT_MASK 0x00000800L +//UVD_SUVD_INT_EN2 +#define UVD_SUVD_INT_EN2__SMPA_FUNC_INT_EN__SHIFT 0x0 +#define UVD_SUVD_INT_EN2__SMPA_ERR_INT_EN__SHIFT 0x5 +#define UVD_SUVD_INT_EN2__SDB_AV1_FUNC_INT_EN__SHIFT 0x6 +#define UVD_SUVD_INT_EN2__SDB_AV1_ERR_INT_EN__SHIFT 0xb +#define UVD_SUVD_INT_EN2__SMPA_FUNC_INT_EN_MASK 0x0000001FL +#define UVD_SUVD_INT_EN2__SMPA_ERR_INT_EN_MASK 0x00000020L +#define UVD_SUVD_INT_EN2__SDB_AV1_FUNC_INT_EN_MASK 0x000007C0L +#define UVD_SUVD_INT_EN2__SDB_AV1_ERR_INT_EN_MASK 0x00000800L +//UVD_SUVD_INT_ACK2 +#define UVD_SUVD_INT_ACK2__SMPA_FUNC_INT_ACK__SHIFT 0x0 +#define UVD_SUVD_INT_ACK2__SMPA_ERR_INT_ACK__SHIFT 0x5 +#define UVD_SUVD_INT_ACK2__SDB_AV1_FUNC_INT_ACK__SHIFT 0x6 +#define UVD_SUVD_INT_ACK2__SDB_AV1_ERR_INT_ACK__SHIFT 0xb +#define UVD_SUVD_INT_ACK2__SMPA_FUNC_INT_ACK_MASK 0x0000001FL +#define UVD_SUVD_INT_ACK2__SMPA_ERR_INT_ACK_MASK 0x00000020L +#define UVD_SUVD_INT_ACK2__SDB_AV1_FUNC_INT_ACK_MASK 0x000007C0L +#define UVD_SUVD_INT_ACK2__SDB_AV1_ERR_INT_ACK_MASK 0x00000800L +//UVD_STATUS +#define UVD_STATUS__RBC_BUSY__SHIFT 0x0 +#define UVD_STATUS__VCPU_REPORT__SHIFT 0x1 +#define UVD_STATUS__FILL_0__SHIFT 0x8 +#define UVD_STATUS__RBC_ACCESS_GPCOM__SHIFT 0x10 +#define UVD_STATUS__DRM_BUSY__SHIFT 0x11 +#define UVD_STATUS__FILL_1__SHIFT 0x12 +#define UVD_STATUS__SYS_GPCOM_REQ__SHIFT 0x1f +#define UVD_STATUS__RBC_BUSY_MASK 0x00000001L +#define UVD_STATUS__VCPU_REPORT_MASK 0x000000FEL +#define UVD_STATUS__FILL_0_MASK 0x0000FF00L +#define UVD_STATUS__RBC_ACCESS_GPCOM_MASK 0x00010000L +#define UVD_STATUS__DRM_BUSY_MASK 0x00020000L +#define UVD_STATUS__FILL_1_MASK 0x7FFC0000L +#define UVD_STATUS__SYS_GPCOM_REQ_MASK 0x80000000L +//UVD_ENC_PIPE_BUSY +#define UVD_ENC_PIPE_BUSY__IME_BUSY__SHIFT 0x0 +#define UVD_ENC_PIPE_BUSY__SMP_BUSY__SHIFT 0x1 +#define UVD_ENC_PIPE_BUSY__SIT_BUSY__SHIFT 0x2 +#define UVD_ENC_PIPE_BUSY__SDB_BUSY__SHIFT 0x3 +#define UVD_ENC_PIPE_BUSY__ENT_BUSY__SHIFT 0x4 +#define UVD_ENC_PIPE_BUSY__ENT_HEADER_BUSY__SHIFT 0x5 +#define UVD_ENC_PIPE_BUSY__LCM_BUSY__SHIFT 0x6 +#define UVD_ENC_PIPE_BUSY__MDM_RD_CUR_BUSY__SHIFT 0x7 +#define UVD_ENC_PIPE_BUSY__MDM_RD_REF_BUSY__SHIFT 0x8 +#define UVD_ENC_PIPE_BUSY__MDM_RD_GEN_BUSY__SHIFT 0x9 +#define UVD_ENC_PIPE_BUSY__MDM_WR_RECON_BUSY__SHIFT 0xa +#define UVD_ENC_PIPE_BUSY__MDM_WR_GEN_BUSY__SHIFT 0xb +#define UVD_ENC_PIPE_BUSY__EFC_BUSY__SHIFT 0xc +#define UVD_ENC_PIPE_BUSY__MDM_PPU_BUSY__SHIFT 0xd +#define UVD_ENC_PIPE_BUSY__MIF_AUTODMA_BUSY__SHIFT 0xe +#define UVD_ENC_PIPE_BUSY__CDEFE_BUSY__SHIFT 0xf +#define UVD_ENC_PIPE_BUSY__MIF_RD_CUR_BUSY__SHIFT 0x10 +#define UVD_ENC_PIPE_BUSY__MIF_RD_REF0_BUSY__SHIFT 0x11 +#define UVD_ENC_PIPE_BUSY__MIF_WR_GEN0_BUSY__SHIFT 0x12 +#define UVD_ENC_PIPE_BUSY__MIF_RD_GEN0_BUSY__SHIFT 0x13 +#define UVD_ENC_PIPE_BUSY__MIF_WR_GEN1_BUSY__SHIFT 0x14 +#define UVD_ENC_PIPE_BUSY__MIF_RD_GEN1_BUSY__SHIFT 0x15 +#define UVD_ENC_PIPE_BUSY__MIF_WR_BSP0_BUSY__SHIFT 0x16 +#define UVD_ENC_PIPE_BUSY__MIF_WR_BSP1_BUSY__SHIFT 0x17 +#define UVD_ENC_PIPE_BUSY__MIF_RD_BSD0_BUSY__SHIFT 0x18 +#define UVD_ENC_PIPE_BUSY__MIF_RD_BSD1_BUSY__SHIFT 0x19 +#define UVD_ENC_PIPE_BUSY__MIF_RD_BSD2_BUSY__SHIFT 0x1a +#define UVD_ENC_PIPE_BUSY__MIF_RD_BSD3_BUSY__SHIFT 0x1b +#define UVD_ENC_PIPE_BUSY__MIF_RD_BSD4_BUSY__SHIFT 0x1c +#define UVD_ENC_PIPE_BUSY__MIF_WR_BSP2_BUSY__SHIFT 0x1d +#define UVD_ENC_PIPE_BUSY__MIF_WR_BSP3_BUSY__SHIFT 0x1e +#define UVD_ENC_PIPE_BUSY__SAOE_BUSY__SHIFT 0x1f +#define UVD_ENC_PIPE_BUSY__IME_BUSY_MASK 0x00000001L +#define UVD_ENC_PIPE_BUSY__SMP_BUSY_MASK 0x00000002L +#define UVD_ENC_PIPE_BUSY__SIT_BUSY_MASK 0x00000004L +#define UVD_ENC_PIPE_BUSY__SDB_BUSY_MASK 0x00000008L +#define UVD_ENC_PIPE_BUSY__ENT_BUSY_MASK 0x00000010L +#define UVD_ENC_PIPE_BUSY__ENT_HEADER_BUSY_MASK 0x00000020L +#define UVD_ENC_PIPE_BUSY__LCM_BUSY_MASK 0x00000040L +#define UVD_ENC_PIPE_BUSY__MDM_RD_CUR_BUSY_MASK 0x00000080L +#define UVD_ENC_PIPE_BUSY__MDM_RD_REF_BUSY_MASK 0x00000100L +#define UVD_ENC_PIPE_BUSY__MDM_RD_GEN_BUSY_MASK 0x00000200L +#define UVD_ENC_PIPE_BUSY__MDM_WR_RECON_BUSY_MASK 0x00000400L +#define UVD_ENC_PIPE_BUSY__MDM_WR_GEN_BUSY_MASK 0x00000800L +#define UVD_ENC_PIPE_BUSY__EFC_BUSY_MASK 0x00001000L +#define UVD_ENC_PIPE_BUSY__MDM_PPU_BUSY_MASK 0x00002000L +#define UVD_ENC_PIPE_BUSY__MIF_AUTODMA_BUSY_MASK 0x00004000L +#define UVD_ENC_PIPE_BUSY__CDEFE_BUSY_MASK 0x00008000L +#define UVD_ENC_PIPE_BUSY__MIF_RD_CUR_BUSY_MASK 0x00010000L +#define UVD_ENC_PIPE_BUSY__MIF_RD_REF0_BUSY_MASK 0x00020000L +#define UVD_ENC_PIPE_BUSY__MIF_WR_GEN0_BUSY_MASK 0x00040000L +#define UVD_ENC_PIPE_BUSY__MIF_RD_GEN0_BUSY_MASK 0x00080000L +#define UVD_ENC_PIPE_BUSY__MIF_WR_GEN1_BUSY_MASK 0x00100000L +#define UVD_ENC_PIPE_BUSY__MIF_RD_GEN1_BUSY_MASK 0x00200000L +#define UVD_ENC_PIPE_BUSY__MIF_WR_BSP0_BUSY_MASK 0x00400000L +#define UVD_ENC_PIPE_BUSY__MIF_WR_BSP1_BUSY_MASK 0x00800000L +#define UVD_ENC_PIPE_BUSY__MIF_RD_BSD0_BUSY_MASK 0x01000000L +#define UVD_ENC_PIPE_BUSY__MIF_RD_BSD1_BUSY_MASK 0x02000000L +#define UVD_ENC_PIPE_BUSY__MIF_RD_BSD2_BUSY_MASK 0x04000000L +#define UVD_ENC_PIPE_BUSY__MIF_RD_BSD3_BUSY_MASK 0x08000000L +#define UVD_ENC_PIPE_BUSY__MIF_RD_BSD4_BUSY_MASK 0x10000000L +#define UVD_ENC_PIPE_BUSY__MIF_WR_BSP2_BUSY_MASK 0x20000000L +#define UVD_ENC_PIPE_BUSY__MIF_WR_BSP3_BUSY_MASK 0x40000000L +#define UVD_ENC_PIPE_BUSY__SAOE_BUSY_MASK 0x80000000L +//UVD_FW_POWER_STATUS +#define UVD_FW_POWER_STATUS__UVDF_PWR_OFF__SHIFT 0x0 +#define UVD_FW_POWER_STATUS__UVDTC_PWR_OFF__SHIFT 0x1 +#define UVD_FW_POWER_STATUS__UVDB_PWR_OFF__SHIFT 0x2 +#define UVD_FW_POWER_STATUS__UVDTA_PWR_OFF__SHIFT 0x3 +#define UVD_FW_POWER_STATUS__UVDTD_PWR_OFF__SHIFT 0x4 +#define UVD_FW_POWER_STATUS__UVDTE_PWR_OFF__SHIFT 0x5 +#define UVD_FW_POWER_STATUS__UVDE_PWR_OFF__SHIFT 0x6 +#define UVD_FW_POWER_STATUS__UVDAB_PWR_OFF__SHIFT 0x7 +#define UVD_FW_POWER_STATUS__UVDTB_PWR_OFF__SHIFT 0x8 +#define UVD_FW_POWER_STATUS__UVDNA_PWR_OFF__SHIFT 0x9 +#define UVD_FW_POWER_STATUS__UVDNB_PWR_OFF__SHIFT 0xa +#define UVD_FW_POWER_STATUS__UVDF_PWR_OFF_MASK 0x00000001L +#define UVD_FW_POWER_STATUS__UVDTC_PWR_OFF_MASK 0x00000002L +#define UVD_FW_POWER_STATUS__UVDB_PWR_OFF_MASK 0x00000004L +#define UVD_FW_POWER_STATUS__UVDTA_PWR_OFF_MASK 0x00000008L +#define UVD_FW_POWER_STATUS__UVDTD_PWR_OFF_MASK 0x00000010L +#define UVD_FW_POWER_STATUS__UVDTE_PWR_OFF_MASK 0x00000020L +#define UVD_FW_POWER_STATUS__UVDE_PWR_OFF_MASK 0x00000040L +#define UVD_FW_POWER_STATUS__UVDAB_PWR_OFF_MASK 0x00000080L +#define UVD_FW_POWER_STATUS__UVDTB_PWR_OFF_MASK 0x00000100L +#define UVD_FW_POWER_STATUS__UVDNA_PWR_OFF_MASK 0x00000200L +#define UVD_FW_POWER_STATUS__UVDNB_PWR_OFF_MASK 0x00000400L +//UVD_CNTL +#define UVD_CNTL__MIF_WR_LOW_THRESHOLD_BP__SHIFT 0x11 +#define UVD_CNTL__SUVD_EN__SHIFT 0x13 +#define UVD_CNTL__CABAC_MB_ACC__SHIFT 0x1c +#define UVD_CNTL__LRBBM_SAFE_SYNC_DIS__SHIFT 0x1f +#define UVD_CNTL__MIF_WR_LOW_THRESHOLD_BP_MASK 0x00020000L +#define UVD_CNTL__SUVD_EN_MASK 0x00080000L +#define UVD_CNTL__CABAC_MB_ACC_MASK 0x10000000L +#define UVD_CNTL__LRBBM_SAFE_SYNC_DIS_MASK 0x80000000L +//UVD_SOFT_RESET +#define UVD_SOFT_RESET__RBC_SOFT_RESET__SHIFT 0x0 +#define UVD_SOFT_RESET__LBSI_SOFT_RESET__SHIFT 0x1 +#define UVD_SOFT_RESET__LMI_SOFT_RESET__SHIFT 0x2 +#define UVD_SOFT_RESET__VCPU_SOFT_RESET__SHIFT 0x3 +#define UVD_SOFT_RESET__UDEC_SOFT_RESET__SHIFT 0x4 +#define UVD_SOFT_RESET__CXW_SOFT_RESET__SHIFT 0x6 +#define UVD_SOFT_RESET__TAP_SOFT_RESET__SHIFT 0x7 +#define UVD_SOFT_RESET__MPC_SOFT_RESET__SHIFT 0x8 +#define UVD_SOFT_RESET__EFC_SOFT_RESET__SHIFT 0x9 +#define UVD_SOFT_RESET__IH_SOFT_RESET__SHIFT 0xa +#define UVD_SOFT_RESET__MPRD_SOFT_RESET__SHIFT 0xb +#define UVD_SOFT_RESET__IDCT_SOFT_RESET__SHIFT 0xc +#define UVD_SOFT_RESET__LMI_UMC_SOFT_RESET__SHIFT 0xd +#define UVD_SOFT_RESET__SPH_SOFT_RESET__SHIFT 0xe +#define UVD_SOFT_RESET__MIF_SOFT_RESET__SHIFT 0xf +#define UVD_SOFT_RESET__LCM_SOFT_RESET__SHIFT 0x10 +#define UVD_SOFT_RESET__SUVD_SOFT_RESET__SHIFT 0x11 +#define UVD_SOFT_RESET__LBSI_VCLK_RESET_STATUS__SHIFT 0x12 +#define UVD_SOFT_RESET__VCPU_VCLK_RESET_STATUS__SHIFT 0x13 +#define UVD_SOFT_RESET__UDEC_VCLK_RESET_STATUS__SHIFT 0x14 +#define UVD_SOFT_RESET__UDEC_DCLK_RESET_STATUS__SHIFT 0x15 +#define UVD_SOFT_RESET__MPC_DCLK_RESET_STATUS__SHIFT 0x16 +#define UVD_SOFT_RESET__MPRD_VCLK_RESET_STATUS__SHIFT 0x17 +#define UVD_SOFT_RESET__MPRD_DCLK_RESET_STATUS__SHIFT 0x18 +#define UVD_SOFT_RESET__IDCT_VCLK_RESET_STATUS__SHIFT 0x19 +#define UVD_SOFT_RESET__MIF_DCLK_RESET_STATUS__SHIFT 0x1a +#define UVD_SOFT_RESET__LCM_DCLK_RESET_STATUS__SHIFT 0x1b +#define UVD_SOFT_RESET__SUVD_VCLK_RESET_STATUS__SHIFT 0x1c +#define UVD_SOFT_RESET__SUVD_DCLK_RESET_STATUS__SHIFT 0x1d +#define UVD_SOFT_RESET__RE_DCLK_RESET_STATUS__SHIFT 0x1e +#define UVD_SOFT_RESET__SRE_DCLK_RESET_STATUS__SHIFT 0x1f +#define UVD_SOFT_RESET__RBC_SOFT_RESET_MASK 0x00000001L +#define UVD_SOFT_RESET__LBSI_SOFT_RESET_MASK 0x00000002L +#define UVD_SOFT_RESET__LMI_SOFT_RESET_MASK 0x00000004L +#define UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK 0x00000008L +#define UVD_SOFT_RESET__UDEC_SOFT_RESET_MASK 0x00000010L +#define UVD_SOFT_RESET__CXW_SOFT_RESET_MASK 0x00000040L +#define UVD_SOFT_RESET__TAP_SOFT_RESET_MASK 0x00000080L +#define UVD_SOFT_RESET__MPC_SOFT_RESET_MASK 0x00000100L +#define UVD_SOFT_RESET__EFC_SOFT_RESET_MASK 0x00000200L +#define UVD_SOFT_RESET__IH_SOFT_RESET_MASK 0x00000400L +#define UVD_SOFT_RESET__MPRD_SOFT_RESET_MASK 0x00000800L +#define UVD_SOFT_RESET__IDCT_SOFT_RESET_MASK 0x00001000L +#define UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK 0x00002000L +#define UVD_SOFT_RESET__SPH_SOFT_RESET_MASK 0x00004000L +#define UVD_SOFT_RESET__MIF_SOFT_RESET_MASK 0x00008000L +#define UVD_SOFT_RESET__LCM_SOFT_RESET_MASK 0x00010000L +#define UVD_SOFT_RESET__SUVD_SOFT_RESET_MASK 0x00020000L +#define UVD_SOFT_RESET__LBSI_VCLK_RESET_STATUS_MASK 0x00040000L +#define UVD_SOFT_RESET__VCPU_VCLK_RESET_STATUS_MASK 0x00080000L +#define UVD_SOFT_RESET__UDEC_VCLK_RESET_STATUS_MASK 0x00100000L +#define UVD_SOFT_RESET__UDEC_DCLK_RESET_STATUS_MASK 0x00200000L +#define UVD_SOFT_RESET__MPC_DCLK_RESET_STATUS_MASK 0x00400000L +#define UVD_SOFT_RESET__MPRD_VCLK_RESET_STATUS_MASK 0x00800000L +#define UVD_SOFT_RESET__MPRD_DCLK_RESET_STATUS_MASK 0x01000000L +#define UVD_SOFT_RESET__IDCT_VCLK_RESET_STATUS_MASK 0x02000000L +#define UVD_SOFT_RESET__MIF_DCLK_RESET_STATUS_MASK 0x04000000L +#define UVD_SOFT_RESET__LCM_DCLK_RESET_STATUS_MASK 0x08000000L +#define UVD_SOFT_RESET__SUVD_VCLK_RESET_STATUS_MASK 0x10000000L +#define UVD_SOFT_RESET__SUVD_DCLK_RESET_STATUS_MASK 0x20000000L +#define UVD_SOFT_RESET__RE_DCLK_RESET_STATUS_MASK 0x40000000L +#define UVD_SOFT_RESET__SRE_DCLK_RESET_STATUS_MASK 0x80000000L +//UVD_SOFT_RESET2 +#define UVD_SOFT_RESET2__ATOMIC_SOFT_RESET__SHIFT 0x0 +#define UVD_SOFT_RESET2__PPU_SOFT_RESET__SHIFT 0x1 +#define UVD_SOFT_RESET2__MMSCH_VCLK_RESET_STATUS__SHIFT 0x10 +#define UVD_SOFT_RESET2__MMSCH_SCLK_RESET_STATUS__SHIFT 0x11 +#define UVD_SOFT_RESET2__ATOMIC_SOFT_RESET_MASK 0x00000001L +#define UVD_SOFT_RESET2__PPU_SOFT_RESET_MASK 0x00000002L +#define UVD_SOFT_RESET2__MMSCH_VCLK_RESET_STATUS_MASK 0x00010000L +#define UVD_SOFT_RESET2__MMSCH_SCLK_RESET_STATUS_MASK 0x00020000L +//UVD_MMSCH_SOFT_RESET +#define UVD_MMSCH_SOFT_RESET__MMSCH_RESET__SHIFT 0x0 +#define UVD_MMSCH_SOFT_RESET__TAP_SOFT_RESET__SHIFT 0x1 +#define UVD_MMSCH_SOFT_RESET__MMSCH_LOCK__SHIFT 0x1f +#define UVD_MMSCH_SOFT_RESET__MMSCH_RESET_MASK 0x00000001L +#define UVD_MMSCH_SOFT_RESET__TAP_SOFT_RESET_MASK 0x00000002L +#define UVD_MMSCH_SOFT_RESET__MMSCH_LOCK_MASK 0x80000000L +//UVD_WIG_CTRL +#define UVD_WIG_CTRL__AVM_SOFT_RESET__SHIFT 0x0 +#define UVD_WIG_CTRL__ACAP_SOFT_RESET__SHIFT 0x1 +#define UVD_WIG_CTRL__WIG_SOFT_RESET__SHIFT 0x2 +#define UVD_WIG_CTRL__WIG_REGCLK_FORCE_ON__SHIFT 0x3 +#define UVD_WIG_CTRL__AVM_REGCLK_FORCE_ON__SHIFT 0x4 +#define UVD_WIG_CTRL__AVM_SOFT_RESET_MASK 0x00000001L +#define UVD_WIG_CTRL__ACAP_SOFT_RESET_MASK 0x00000002L +#define UVD_WIG_CTRL__WIG_SOFT_RESET_MASK 0x00000004L +#define UVD_WIG_CTRL__WIG_REGCLK_FORCE_ON_MASK 0x00000008L +#define UVD_WIG_CTRL__AVM_REGCLK_FORCE_ON_MASK 0x00000010L +//UVD_CGC_STATUS +#define UVD_CGC_STATUS__SYS_SCLK__SHIFT 0x0 +#define UVD_CGC_STATUS__SYS_DCLK__SHIFT 0x1 +#define UVD_CGC_STATUS__SYS_VCLK__SHIFT 0x2 +#define UVD_CGC_STATUS__UDEC_SCLK__SHIFT 0x3 +#define UVD_CGC_STATUS__UDEC_DCLK__SHIFT 0x4 +#define UVD_CGC_STATUS__UDEC_VCLK__SHIFT 0x5 +#define UVD_CGC_STATUS__MPEG2_SCLK__SHIFT 0x6 +#define UVD_CGC_STATUS__MPEG2_DCLK__SHIFT 0x7 +#define UVD_CGC_STATUS__MPEG2_VCLK__SHIFT 0x8 +#define UVD_CGC_STATUS__REGS_SCLK__SHIFT 0x9 +#define UVD_CGC_STATUS__REGS_VCLK__SHIFT 0xa +#define UVD_CGC_STATUS__RBC_SCLK__SHIFT 0xb +#define UVD_CGC_STATUS__LMI_MC_SCLK__SHIFT 0xc +#define UVD_CGC_STATUS__LMI_UMC_SCLK__SHIFT 0xd +#define UVD_CGC_STATUS__IDCT_SCLK__SHIFT 0xe +#define UVD_CGC_STATUS__IDCT_VCLK__SHIFT 0xf +#define UVD_CGC_STATUS__MPRD_SCLK__SHIFT 0x10 +#define UVD_CGC_STATUS__MPRD_DCLK__SHIFT 0x11 +#define UVD_CGC_STATUS__MPRD_VCLK__SHIFT 0x12 +#define UVD_CGC_STATUS__MPC_SCLK__SHIFT 0x13 +#define UVD_CGC_STATUS__MPC_DCLK__SHIFT 0x14 +#define UVD_CGC_STATUS__LBSI_SCLK__SHIFT 0x15 +#define UVD_CGC_STATUS__LBSI_VCLK__SHIFT 0x16 +#define UVD_CGC_STATUS__LRBBM_SCLK__SHIFT 0x17 +#define UVD_CGC_STATUS__WCB_SCLK__SHIFT 0x18 +#define UVD_CGC_STATUS__VCPU_SCLK__SHIFT 0x19 +#define UVD_CGC_STATUS__VCPU_VCLK__SHIFT 0x1a +#define UVD_CGC_STATUS__MMSCH_SCLK__SHIFT 0x1b +#define UVD_CGC_STATUS__MMSCH_VCLK__SHIFT 0x1c +#define UVD_CGC_STATUS__ALL_ENC_ACTIVE__SHIFT 0x1d +#define UVD_CGC_STATUS__LRBBM_DCLK__SHIFT 0x1e +#define UVD_CGC_STATUS__ALL_DEC_ACTIVE__SHIFT 0x1f +#define UVD_CGC_STATUS__SYS_SCLK_MASK 0x00000001L +#define UVD_CGC_STATUS__SYS_DCLK_MASK 0x00000002L +#define UVD_CGC_STATUS__SYS_VCLK_MASK 0x00000004L +#define UVD_CGC_STATUS__UDEC_SCLK_MASK 0x00000008L +#define UVD_CGC_STATUS__UDEC_DCLK_MASK 0x00000010L +#define UVD_CGC_STATUS__UDEC_VCLK_MASK 0x00000020L +#define UVD_CGC_STATUS__MPEG2_SCLK_MASK 0x00000040L +#define UVD_CGC_STATUS__MPEG2_DCLK_MASK 0x00000080L +#define UVD_CGC_STATUS__MPEG2_VCLK_MASK 0x00000100L +#define UVD_CGC_STATUS__REGS_SCLK_MASK 0x00000200L +#define UVD_CGC_STATUS__REGS_VCLK_MASK 0x00000400L +#define UVD_CGC_STATUS__RBC_SCLK_MASK 0x00000800L +#define UVD_CGC_STATUS__LMI_MC_SCLK_MASK 0x00001000L +#define UVD_CGC_STATUS__LMI_UMC_SCLK_MASK 0x00002000L +#define UVD_CGC_STATUS__IDCT_SCLK_MASK 0x00004000L +#define UVD_CGC_STATUS__IDCT_VCLK_MASK 0x00008000L +#define UVD_CGC_STATUS__MPRD_SCLK_MASK 0x00010000L +#define UVD_CGC_STATUS__MPRD_DCLK_MASK 0x00020000L +#define UVD_CGC_STATUS__MPRD_VCLK_MASK 0x00040000L +#define UVD_CGC_STATUS__MPC_SCLK_MASK 0x00080000L +#define UVD_CGC_STATUS__MPC_DCLK_MASK 0x00100000L +#define UVD_CGC_STATUS__LBSI_SCLK_MASK 0x00200000L +#define UVD_CGC_STATUS__LBSI_VCLK_MASK 0x00400000L +#define UVD_CGC_STATUS__LRBBM_SCLK_MASK 0x00800000L +#define UVD_CGC_STATUS__WCB_SCLK_MASK 0x01000000L +#define UVD_CGC_STATUS__VCPU_SCLK_MASK 0x02000000L +#define UVD_CGC_STATUS__VCPU_VCLK_MASK 0x04000000L +#define UVD_CGC_STATUS__MMSCH_SCLK_MASK 0x08000000L +#define UVD_CGC_STATUS__MMSCH_VCLK_MASK 0x10000000L +#define UVD_CGC_STATUS__ALL_ENC_ACTIVE_MASK 0x20000000L +#define UVD_CGC_STATUS__LRBBM_DCLK_MASK 0x40000000L +#define UVD_CGC_STATUS__ALL_DEC_ACTIVE_MASK 0x80000000L +//UVD_CGC_UDEC_STATUS +#define UVD_CGC_UDEC_STATUS__RE_SCLK__SHIFT 0x0 +#define UVD_CGC_UDEC_STATUS__RE_DCLK__SHIFT 0x1 +#define UVD_CGC_UDEC_STATUS__RE_VCLK__SHIFT 0x2 +#define UVD_CGC_UDEC_STATUS__CM_SCLK__SHIFT 0x3 +#define UVD_CGC_UDEC_STATUS__CM_DCLK__SHIFT 0x4 +#define UVD_CGC_UDEC_STATUS__CM_VCLK__SHIFT 0x5 +#define UVD_CGC_UDEC_STATUS__IT_SCLK__SHIFT 0x6 +#define UVD_CGC_UDEC_STATUS__IT_DCLK__SHIFT 0x7 +#define UVD_CGC_UDEC_STATUS__IT_VCLK__SHIFT 0x8 +#define UVD_CGC_UDEC_STATUS__DB_SCLK__SHIFT 0x9 +#define UVD_CGC_UDEC_STATUS__DB_DCLK__SHIFT 0xa +#define UVD_CGC_UDEC_STATUS__DB_VCLK__SHIFT 0xb +#define UVD_CGC_UDEC_STATUS__MP_SCLK__SHIFT 0xc +#define UVD_CGC_UDEC_STATUS__MP_DCLK__SHIFT 0xd +#define UVD_CGC_UDEC_STATUS__MP_VCLK__SHIFT 0xe +#define UVD_CGC_UDEC_STATUS__RE_SCLK_MASK 0x00000001L +#define UVD_CGC_UDEC_STATUS__RE_DCLK_MASK 0x00000002L +#define UVD_CGC_UDEC_STATUS__RE_VCLK_MASK 0x00000004L +#define UVD_CGC_UDEC_STATUS__CM_SCLK_MASK 0x00000008L +#define UVD_CGC_UDEC_STATUS__CM_DCLK_MASK 0x00000010L +#define UVD_CGC_UDEC_STATUS__CM_VCLK_MASK 0x00000020L +#define UVD_CGC_UDEC_STATUS__IT_SCLK_MASK 0x00000040L +#define UVD_CGC_UDEC_STATUS__IT_DCLK_MASK 0x00000080L +#define UVD_CGC_UDEC_STATUS__IT_VCLK_MASK 0x00000100L +#define UVD_CGC_UDEC_STATUS__DB_SCLK_MASK 0x00000200L +#define UVD_CGC_UDEC_STATUS__DB_DCLK_MASK 0x00000400L +#define UVD_CGC_UDEC_STATUS__DB_VCLK_MASK 0x00000800L +#define UVD_CGC_UDEC_STATUS__MP_SCLK_MASK 0x00001000L +#define UVD_CGC_UDEC_STATUS__MP_DCLK_MASK 0x00002000L +#define UVD_CGC_UDEC_STATUS__MP_VCLK_MASK 0x00004000L +//UVD_SUVD_CGC_STATUS +#define UVD_SUVD_CGC_STATUS__SRE_VCLK__SHIFT 0x0 +#define UVD_SUVD_CGC_STATUS__SRE_DCLK__SHIFT 0x1 +#define UVD_SUVD_CGC_STATUS__SIT_DCLK__SHIFT 0x2 +#define UVD_SUVD_CGC_STATUS__SMP_DCLK__SHIFT 0x3 +#define UVD_SUVD_CGC_STATUS__SCM_DCLK__SHIFT 0x4 +#define UVD_SUVD_CGC_STATUS__SDB_DCLK__SHIFT 0x5 +#define UVD_SUVD_CGC_STATUS__SRE_H264_VCLK__SHIFT 0x6 +#define UVD_SUVD_CGC_STATUS__SRE_HEVC_VCLK__SHIFT 0x7 +#define UVD_SUVD_CGC_STATUS__SIT_H264_DCLK__SHIFT 0x8 +#define UVD_SUVD_CGC_STATUS__SIT_HEVC_DCLK__SHIFT 0x9 +#define UVD_SUVD_CGC_STATUS__SCM_H264_DCLK__SHIFT 0xa +#define UVD_SUVD_CGC_STATUS__SCM_HEVC_DCLK__SHIFT 0xb +#define UVD_SUVD_CGC_STATUS__SDB_H264_DCLK__SHIFT 0xc +#define UVD_SUVD_CGC_STATUS__SDB_HEVC_DCLK__SHIFT 0xd +#define UVD_SUVD_CGC_STATUS__SCLR_DCLK__SHIFT 0xe +#define UVD_SUVD_CGC_STATUS__UVD_SC__SHIFT 0xf +#define UVD_SUVD_CGC_STATUS__ENT_DCLK__SHIFT 0x10 +#define UVD_SUVD_CGC_STATUS__IME_DCLK__SHIFT 0x11 +#define UVD_SUVD_CGC_STATUS__SIT_HEVC_DEC_DCLK__SHIFT 0x12 +#define UVD_SUVD_CGC_STATUS__SIT_HEVC_ENC_DCLK__SHIFT 0x13 +#define UVD_SUVD_CGC_STATUS__SITE_DCLK__SHIFT 0x14 +#define UVD_SUVD_CGC_STATUS__SITE_HEVC_DCLK__SHIFT 0x15 +#define UVD_SUVD_CGC_STATUS__SITE_HEVC_ENC_DCLK__SHIFT 0x16 +#define UVD_SUVD_CGC_STATUS__SRE_VP9_VCLK__SHIFT 0x17 +#define UVD_SUVD_CGC_STATUS__SCM_VP9_VCLK__SHIFT 0x18 +#define UVD_SUVD_CGC_STATUS__SIT_VP9_DEC_DCLK__SHIFT 0x19 +#define UVD_SUVD_CGC_STATUS__SDB_VP9_DCLK__SHIFT 0x1a +#define UVD_SUVD_CGC_STATUS__IME_HEVC_DCLK__SHIFT 0x1b +#define UVD_SUVD_CGC_STATUS__EFC_DCLK__SHIFT 0x1c +#define UVD_SUVD_CGC_STATUS__SAOE_DCLK__SHIFT 0x1d +#define UVD_SUVD_CGC_STATUS__SRE_AV1_VCLK__SHIFT 0x1e +#define UVD_SUVD_CGC_STATUS__SCM_AV1_DCLK__SHIFT 0x1f +#define UVD_SUVD_CGC_STATUS__SRE_VCLK_MASK 0x00000001L +#define UVD_SUVD_CGC_STATUS__SRE_DCLK_MASK 0x00000002L +#define UVD_SUVD_CGC_STATUS__SIT_DCLK_MASK 0x00000004L +#define UVD_SUVD_CGC_STATUS__SMP_DCLK_MASK 0x00000008L +#define UVD_SUVD_CGC_STATUS__SCM_DCLK_MASK 0x00000010L +#define UVD_SUVD_CGC_STATUS__SDB_DCLK_MASK 0x00000020L +#define UVD_SUVD_CGC_STATUS__SRE_H264_VCLK_MASK 0x00000040L +#define UVD_SUVD_CGC_STATUS__SRE_HEVC_VCLK_MASK 0x00000080L +#define UVD_SUVD_CGC_STATUS__SIT_H264_DCLK_MASK 0x00000100L +#define UVD_SUVD_CGC_STATUS__SIT_HEVC_DCLK_MASK 0x00000200L +#define UVD_SUVD_CGC_STATUS__SCM_H264_DCLK_MASK 0x00000400L +#define UVD_SUVD_CGC_STATUS__SCM_HEVC_DCLK_MASK 0x00000800L +#define UVD_SUVD_CGC_STATUS__SDB_H264_DCLK_MASK 0x00001000L +#define UVD_SUVD_CGC_STATUS__SDB_HEVC_DCLK_MASK 0x00002000L +#define UVD_SUVD_CGC_STATUS__SCLR_DCLK_MASK 0x00004000L +#define UVD_SUVD_CGC_STATUS__UVD_SC_MASK 0x00008000L +#define UVD_SUVD_CGC_STATUS__ENT_DCLK_MASK 0x00010000L +#define UVD_SUVD_CGC_STATUS__IME_DCLK_MASK 0x00020000L +#define UVD_SUVD_CGC_STATUS__SIT_HEVC_DEC_DCLK_MASK 0x00040000L +#define UVD_SUVD_CGC_STATUS__SIT_HEVC_ENC_DCLK_MASK 0x00080000L +#define UVD_SUVD_CGC_STATUS__SITE_DCLK_MASK 0x00100000L +#define UVD_SUVD_CGC_STATUS__SITE_HEVC_DCLK_MASK 0x00200000L +#define UVD_SUVD_CGC_STATUS__SITE_HEVC_ENC_DCLK_MASK 0x00400000L +#define UVD_SUVD_CGC_STATUS__SRE_VP9_VCLK_MASK 0x00800000L +#define UVD_SUVD_CGC_STATUS__SCM_VP9_VCLK_MASK 0x01000000L +#define UVD_SUVD_CGC_STATUS__SIT_VP9_DEC_DCLK_MASK 0x02000000L +#define UVD_SUVD_CGC_STATUS__SDB_VP9_DCLK_MASK 0x04000000L +#define UVD_SUVD_CGC_STATUS__IME_HEVC_DCLK_MASK 0x08000000L +#define UVD_SUVD_CGC_STATUS__EFC_DCLK_MASK 0x10000000L +#define UVD_SUVD_CGC_STATUS__SAOE_DCLK_MASK 0x20000000L +#define UVD_SUVD_CGC_STATUS__SRE_AV1_VCLK_MASK 0x40000000L +#define UVD_SUVD_CGC_STATUS__SCM_AV1_DCLK_MASK 0x80000000L +//UVD_GPCOM_VCPU_CMD +#define UVD_GPCOM_VCPU_CMD__CMD_SEND__SHIFT 0x0 +#define UVD_GPCOM_VCPU_CMD__CMD__SHIFT 0x1 +#define UVD_GPCOM_VCPU_CMD__CMD_SOURCE__SHIFT 0x1f +#define UVD_GPCOM_VCPU_CMD__CMD_SEND_MASK 0x00000001L +#define UVD_GPCOM_VCPU_CMD__CMD_MASK 0x7FFFFFFEL +#define UVD_GPCOM_VCPU_CMD__CMD_SOURCE_MASK 0x80000000L + + +// addressBlock: aid_uvd0_ecpudec +//UVD_VCPU_CACHE_OFFSET0 +#define UVD_VCPU_CACHE_OFFSET0__CACHE_OFFSET0__SHIFT 0x0 +#define UVD_VCPU_CACHE_OFFSET0__CACHE_OFFSET0_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_SIZE0 +#define UVD_VCPU_CACHE_SIZE0__CACHE_SIZE0__SHIFT 0x0 +#define UVD_VCPU_CACHE_SIZE0__CACHE_SIZE0_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_OFFSET1 +#define UVD_VCPU_CACHE_OFFSET1__CACHE_OFFSET1__SHIFT 0x0 +#define UVD_VCPU_CACHE_OFFSET1__CACHE_OFFSET1_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_SIZE1 +#define UVD_VCPU_CACHE_SIZE1__CACHE_SIZE1__SHIFT 0x0 +#define UVD_VCPU_CACHE_SIZE1__CACHE_SIZE1_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_OFFSET2 +#define UVD_VCPU_CACHE_OFFSET2__CACHE_OFFSET2__SHIFT 0x0 +#define UVD_VCPU_CACHE_OFFSET2__CACHE_OFFSET2_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_SIZE2 +#define UVD_VCPU_CACHE_SIZE2__CACHE_SIZE2__SHIFT 0x0 +#define UVD_VCPU_CACHE_SIZE2__CACHE_SIZE2_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_OFFSET3 +#define UVD_VCPU_CACHE_OFFSET3__CACHE_OFFSET3__SHIFT 0x0 +#define UVD_VCPU_CACHE_OFFSET3__CACHE_OFFSET3_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_SIZE3 +#define UVD_VCPU_CACHE_SIZE3__CACHE_SIZE3__SHIFT 0x0 +#define UVD_VCPU_CACHE_SIZE3__CACHE_SIZE3_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_OFFSET4 +#define UVD_VCPU_CACHE_OFFSET4__CACHE_OFFSET4__SHIFT 0x0 +#define UVD_VCPU_CACHE_OFFSET4__CACHE_OFFSET4_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_SIZE4 +#define UVD_VCPU_CACHE_SIZE4__CACHE_SIZE4__SHIFT 0x0 +#define UVD_VCPU_CACHE_SIZE4__CACHE_SIZE4_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_OFFSET5 +#define UVD_VCPU_CACHE_OFFSET5__CACHE_OFFSET5__SHIFT 0x0 +#define UVD_VCPU_CACHE_OFFSET5__CACHE_OFFSET5_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_SIZE5 +#define UVD_VCPU_CACHE_SIZE5__CACHE_SIZE5__SHIFT 0x0 +#define UVD_VCPU_CACHE_SIZE5__CACHE_SIZE5_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_OFFSET6 +#define UVD_VCPU_CACHE_OFFSET6__CACHE_OFFSET6__SHIFT 0x0 +#define UVD_VCPU_CACHE_OFFSET6__CACHE_OFFSET6_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_SIZE6 +#define UVD_VCPU_CACHE_SIZE6__CACHE_SIZE6__SHIFT 0x0 +#define UVD_VCPU_CACHE_SIZE6__CACHE_SIZE6_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_OFFSET7 +#define UVD_VCPU_CACHE_OFFSET7__CACHE_OFFSET7__SHIFT 0x0 +#define UVD_VCPU_CACHE_OFFSET7__CACHE_OFFSET7_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_SIZE7 +#define UVD_VCPU_CACHE_SIZE7__CACHE_SIZE7__SHIFT 0x0 +#define UVD_VCPU_CACHE_SIZE7__CACHE_SIZE7_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_OFFSET8 +#define UVD_VCPU_CACHE_OFFSET8__CACHE_OFFSET8__SHIFT 0x0 +#define UVD_VCPU_CACHE_OFFSET8__CACHE_OFFSET8_MASK 0x001FFFFFL +//UVD_VCPU_CACHE_SIZE8 +#define UVD_VCPU_CACHE_SIZE8__CACHE_SIZE8__SHIFT 0x0 +#define UVD_VCPU_CACHE_SIZE8__CACHE_SIZE8_MASK 0x001FFFFFL +//UVD_VCPU_NONCACHE_OFFSET0 +#define UVD_VCPU_NONCACHE_OFFSET0__NONCACHE_OFFSET0__SHIFT 0x0 +#define UVD_VCPU_NONCACHE_OFFSET0__NONCACHE_OFFSET0_MASK 0x01FFFFFFL +//UVD_VCPU_NONCACHE_SIZE0 +#define UVD_VCPU_NONCACHE_SIZE0__NONCACHE_SIZE0__SHIFT 0x0 +#define UVD_VCPU_NONCACHE_SIZE0__NONCACHE_SIZE0_MASK 0x001FFFFFL +//UVD_VCPU_NONCACHE_OFFSET1 +#define UVD_VCPU_NONCACHE_OFFSET1__NONCACHE_OFFSET1__SHIFT 0x0 +#define UVD_VCPU_NONCACHE_OFFSET1__NONCACHE_OFFSET1_MASK 0x01FFFFFFL +//UVD_VCPU_NONCACHE_SIZE1 +#define UVD_VCPU_NONCACHE_SIZE1__NONCACHE_SIZE1__SHIFT 0x0 +#define UVD_VCPU_NONCACHE_SIZE1__NONCACHE_SIZE1_MASK 0x001FFFFFL +//UVD_VCPU_CNTL +#define UVD_VCPU_CNTL__IRQ_ERR__SHIFT 0x0 +#define UVD_VCPU_CNTL__AXI_MAX_BRST_SIZE_IS_4__SHIFT 0x4 +#define UVD_VCPU_CNTL__PMB_ED_ENABLE__SHIFT 0x5 +#define UVD_VCPU_CNTL__PMB_SOFT_RESET__SHIFT 0x6 +#define UVD_VCPU_CNTL__RBBM_SOFT_RESET__SHIFT 0x7 +#define UVD_VCPU_CNTL__ABORT_REQ__SHIFT 0x8 +#define UVD_VCPU_CNTL__CLK_EN__SHIFT 0x9 +#define UVD_VCPU_CNTL__TRCE_EN__SHIFT 0xa +#define UVD_VCPU_CNTL__TRCE_MUX__SHIFT 0xb +#define UVD_VCPU_CNTL__DBG_MUX__SHIFT 0xd +#define UVD_VCPU_CNTL__JTAG_EN__SHIFT 0x10 +#define UVD_VCPU_CNTL__TIMEOUT_DIS__SHIFT 0x12 +#define UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT 0x14 +#define UVD_VCPU_CNTL__BLK_RST__SHIFT 0x1c +#define UVD_VCPU_CNTL__RUNSTALL__SHIFT 0x1d +#define UVD_VCPU_CNTL__SRE_CMDIF_DRST__SHIFT 0x1e +#define UVD_VCPU_CNTL__SRE_CMDIF_VRST__SHIFT 0x1f +#define UVD_VCPU_CNTL__IRQ_ERR_MASK 0x0000000FL +#define UVD_VCPU_CNTL__AXI_MAX_BRST_SIZE_IS_4_MASK 0x00000010L +#define UVD_VCPU_CNTL__PMB_ED_ENABLE_MASK 0x00000020L +#define UVD_VCPU_CNTL__PMB_SOFT_RESET_MASK 0x00000040L +#define UVD_VCPU_CNTL__RBBM_SOFT_RESET_MASK 0x00000080L +#define UVD_VCPU_CNTL__ABORT_REQ_MASK 0x00000100L +#define UVD_VCPU_CNTL__CLK_EN_MASK 0x00000200L +#define UVD_VCPU_CNTL__TRCE_EN_MASK 0x00000400L +#define UVD_VCPU_CNTL__TRCE_MUX_MASK 0x00001800L +#define UVD_VCPU_CNTL__DBG_MUX_MASK 0x0000E000L +#define UVD_VCPU_CNTL__JTAG_EN_MASK 0x00010000L +#define UVD_VCPU_CNTL__TIMEOUT_DIS_MASK 0x00040000L +#define UVD_VCPU_CNTL__PRB_TIMEOUT_VAL_MASK 0x0FF00000L +#define UVD_VCPU_CNTL__BLK_RST_MASK 0x10000000L +#define UVD_VCPU_CNTL__RUNSTALL_MASK 0x20000000L +#define UVD_VCPU_CNTL__SRE_CMDIF_DRST_MASK 0x40000000L +#define UVD_VCPU_CNTL__SRE_CMDIF_VRST_MASK 0x80000000L +//UVD_VCPU_PRID +#define UVD_VCPU_PRID__PRID__SHIFT 0x0 +#define UVD_VCPU_PRID__PRID_MASK 0x0000FFFFL +//UVD_VCPU_TRCE +#define UVD_VCPU_TRCE__PC__SHIFT 0x0 +#define UVD_VCPU_TRCE__PC_MASK 0x0FFFFFFFL +//UVD_VCPU_TRCE_RD +#define UVD_VCPU_TRCE_RD__DATA__SHIFT 0x0 +#define UVD_VCPU_TRCE_RD__DATA_MASK 0xFFFFFFFFL +//UVD_VCPU_IND_INDEX +#define UVD_VCPU_IND_INDEX__INDEX__SHIFT 0x0 +#define UVD_VCPU_IND_INDEX__INDEX_MASK 0x000001FFL +//UVD_VCPU_IND_DATA +#define UVD_VCPU_IND_DATA__DATA__SHIFT 0x0 +#define UVD_VCPU_IND_DATA__DATA_MASK 0xFFFFFFFFL + + +// addressBlock: aid_uvd0_uvd_mpcdec +//UVD_MP_SWAP_CNTL +#define UVD_MP_SWAP_CNTL__MP_REF0_MC_SWAP__SHIFT 0x0 +#define UVD_MP_SWAP_CNTL__MP_REF1_MC_SWAP__SHIFT 0x2 +#define UVD_MP_SWAP_CNTL__MP_REF2_MC_SWAP__SHIFT 0x4 +#define UVD_MP_SWAP_CNTL__MP_REF3_MC_SWAP__SHIFT 0x6 +#define UVD_MP_SWAP_CNTL__MP_REF4_MC_SWAP__SHIFT 0x8 +#define UVD_MP_SWAP_CNTL__MP_REF5_MC_SWAP__SHIFT 0xa +#define UVD_MP_SWAP_CNTL__MP_REF6_MC_SWAP__SHIFT 0xc +#define UVD_MP_SWAP_CNTL__MP_REF7_MC_SWAP__SHIFT 0xe +#define UVD_MP_SWAP_CNTL__MP_REF8_MC_SWAP__SHIFT 0x10 +#define UVD_MP_SWAP_CNTL__MP_REF9_MC_SWAP__SHIFT 0x12 +#define UVD_MP_SWAP_CNTL__MP_REF10_MC_SWAP__SHIFT 0x14 +#define UVD_MP_SWAP_CNTL__MP_REF11_MC_SWAP__SHIFT 0x16 +#define UVD_MP_SWAP_CNTL__MP_REF12_MC_SWAP__SHIFT 0x18 +#define UVD_MP_SWAP_CNTL__MP_REF13_MC_SWAP__SHIFT 0x1a +#define UVD_MP_SWAP_CNTL__MP_REF14_MC_SWAP__SHIFT 0x1c +#define UVD_MP_SWAP_CNTL__MP_REF15_MC_SWAP__SHIFT 0x1e +#define UVD_MP_SWAP_CNTL__MP_REF0_MC_SWAP_MASK 0x00000003L +#define UVD_MP_SWAP_CNTL__MP_REF1_MC_SWAP_MASK 0x0000000CL +#define UVD_MP_SWAP_CNTL__MP_REF2_MC_SWAP_MASK 0x00000030L +#define UVD_MP_SWAP_CNTL__MP_REF3_MC_SWAP_MASK 0x000000C0L +#define UVD_MP_SWAP_CNTL__MP_REF4_MC_SWAP_MASK 0x00000300L +#define UVD_MP_SWAP_CNTL__MP_REF5_MC_SWAP_MASK 0x00000C00L +#define UVD_MP_SWAP_CNTL__MP_REF6_MC_SWAP_MASK 0x00003000L +#define UVD_MP_SWAP_CNTL__MP_REF7_MC_SWAP_MASK 0x0000C000L +#define UVD_MP_SWAP_CNTL__MP_REF8_MC_SWAP_MASK 0x00030000L +#define UVD_MP_SWAP_CNTL__MP_REF9_MC_SWAP_MASK 0x000C0000L +#define UVD_MP_SWAP_CNTL__MP_REF10_MC_SWAP_MASK 0x00300000L +#define UVD_MP_SWAP_CNTL__MP_REF11_MC_SWAP_MASK 0x00C00000L +#define UVD_MP_SWAP_CNTL__MP_REF12_MC_SWAP_MASK 0x03000000L +#define UVD_MP_SWAP_CNTL__MP_REF13_MC_SWAP_MASK 0x0C000000L +#define UVD_MP_SWAP_CNTL__MP_REF14_MC_SWAP_MASK 0x30000000L +#define UVD_MP_SWAP_CNTL__MP_REF15_MC_SWAP_MASK 0xC0000000L +//UVD_MP_SWAP_CNTL2 +#define UVD_MP_SWAP_CNTL2__MP_REF16_MC_SWAP__SHIFT 0x0 +#define UVD_MP_SWAP_CNTL2__MP_REF16_MC_SWAP_MASK 0x00000003L +//UVD_MPC_LUMA_SRCH +#define UVD_MPC_LUMA_SRCH__CNTR__SHIFT 0x0 +#define UVD_MPC_LUMA_SRCH__CNTR_MASK 0xFFFFFFFFL +//UVD_MPC_LUMA_HIT +#define UVD_MPC_LUMA_HIT__CNTR__SHIFT 0x0 +#define UVD_MPC_LUMA_HIT__CNTR_MASK 0xFFFFFFFFL +//UVD_MPC_LUMA_HITPEND +#define UVD_MPC_LUMA_HITPEND__CNTR__SHIFT 0x0 +#define UVD_MPC_LUMA_HITPEND__CNTR_MASK 0xFFFFFFFFL +//UVD_MPC_CHROMA_SRCH +#define UVD_MPC_CHROMA_SRCH__CNTR__SHIFT 0x0 +#define UVD_MPC_CHROMA_SRCH__CNTR_MASK 0xFFFFFFFFL +//UVD_MPC_CHROMA_HIT +#define UVD_MPC_CHROMA_HIT__CNTR__SHIFT 0x0 +#define UVD_MPC_CHROMA_HIT__CNTR_MASK 0xFFFFFFFFL +//UVD_MPC_CHROMA_HITPEND +#define UVD_MPC_CHROMA_HITPEND__CNTR__SHIFT 0x0 +#define UVD_MPC_CHROMA_HITPEND__CNTR_MASK 0xFFFFFFFFL +//UVD_MPC_CNTL +#define UVD_MPC_CNTL__BLK_RST__SHIFT 0x0 +#define UVD_MPC_CNTL__REG_MPC1_PERF_SELECT__SHIFT 0x1 +#define UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT 0x3 +#define UVD_MPC_CNTL__PERF_RST__SHIFT 0x6 +#define UVD_MPC_CNTL__REG_MPC_CNTL_BACKWARD_COMPATIBILITY__SHIFT 0x7 +#define UVD_MPC_CNTL__DBG_MUX__SHIFT 0x8 +#define UVD_MPC_CNTL__AVE_WEIGHT__SHIFT 0x10 +#define UVD_MPC_CNTL__URGENT_EN__SHIFT 0x12 +#define UVD_MPC_CNTL__SMPAT_REQ_SPEED_UP__SHIFT 0x13 +#define UVD_MPC_CNTL__TEST_MODE_EN__SHIFT 0x14 +#define UVD_MPC_CNTL__BLK_RST_MASK 0x00000001L +#define UVD_MPC_CNTL__REG_MPC1_PERF_SELECT_MASK 0x00000002L +#define UVD_MPC_CNTL__REPLACEMENT_MODE_MASK 0x00000038L +#define UVD_MPC_CNTL__PERF_RST_MASK 0x00000040L +#define UVD_MPC_CNTL__REG_MPC_CNTL_BACKWARD_COMPATIBILITY_MASK 0x00000080L +#define UVD_MPC_CNTL__DBG_MUX_MASK 0x00000F00L +#define UVD_MPC_CNTL__AVE_WEIGHT_MASK 0x00030000L +#define UVD_MPC_CNTL__URGENT_EN_MASK 0x00040000L +#define UVD_MPC_CNTL__SMPAT_REQ_SPEED_UP_MASK 0x00080000L +#define UVD_MPC_CNTL__TEST_MODE_EN_MASK 0x00300000L +//UVD_MPC_PITCH +#define UVD_MPC_PITCH__LUMA_PITCH__SHIFT 0x0 +#define UVD_MPC_PITCH__LUMA_PITCH_MASK 0x000007FFL +//UVD_MPC_SET_MUXA0 +#define UVD_MPC_SET_MUXA0__VARA_0__SHIFT 0x0 +#define UVD_MPC_SET_MUXA0__VARA_1__SHIFT 0x6 +#define UVD_MPC_SET_MUXA0__VARA_2__SHIFT 0xc +#define UVD_MPC_SET_MUXA0__VARA_3__SHIFT 0x12 +#define UVD_MPC_SET_MUXA0__VARA_4__SHIFT 0x18 +#define UVD_MPC_SET_MUXA0__VARA_0_MASK 0x0000003FL +#define UVD_MPC_SET_MUXA0__VARA_1_MASK 0x00000FC0L +#define UVD_MPC_SET_MUXA0__VARA_2_MASK 0x0003F000L +#define UVD_MPC_SET_MUXA0__VARA_3_MASK 0x00FC0000L +#define UVD_MPC_SET_MUXA0__VARA_4_MASK 0x3F000000L +//UVD_MPC_SET_MUXA1 +#define UVD_MPC_SET_MUXA1__VARA_5__SHIFT 0x0 +#define UVD_MPC_SET_MUXA1__VARA_6__SHIFT 0x6 +#define UVD_MPC_SET_MUXA1__VARA_7__SHIFT 0xc +#define UVD_MPC_SET_MUXA1__VARA_5_MASK 0x0000003FL +#define UVD_MPC_SET_MUXA1__VARA_6_MASK 0x00000FC0L +#define UVD_MPC_SET_MUXA1__VARA_7_MASK 0x0003F000L +//UVD_MPC_SET_MUXB0 +#define UVD_MPC_SET_MUXB0__VARB_0__SHIFT 0x0 +#define UVD_MPC_SET_MUXB0__VARB_1__SHIFT 0x6 +#define UVD_MPC_SET_MUXB0__VARB_2__SHIFT 0xc +#define UVD_MPC_SET_MUXB0__VARB_3__SHIFT 0x12 +#define UVD_MPC_SET_MUXB0__VARB_4__SHIFT 0x18 +#define UVD_MPC_SET_MUXB0__VARB_0_MASK 0x0000003FL +#define UVD_MPC_SET_MUXB0__VARB_1_MASK 0x00000FC0L +#define UVD_MPC_SET_MUXB0__VARB_2_MASK 0x0003F000L +#define UVD_MPC_SET_MUXB0__VARB_3_MASK 0x00FC0000L +#define UVD_MPC_SET_MUXB0__VARB_4_MASK 0x3F000000L +//UVD_MPC_SET_MUXB1 +#define UVD_MPC_SET_MUXB1__VARB_5__SHIFT 0x0 +#define UVD_MPC_SET_MUXB1__VARB_6__SHIFT 0x6 +#define UVD_MPC_SET_MUXB1__VARB_7__SHIFT 0xc +#define UVD_MPC_SET_MUXB1__VARB_5_MASK 0x0000003FL +#define UVD_MPC_SET_MUXB1__VARB_6_MASK 0x00000FC0L +#define UVD_MPC_SET_MUXB1__VARB_7_MASK 0x0003F000L +//UVD_MPC_SET_MUX +#define UVD_MPC_SET_MUX__SET_0__SHIFT 0x0 +#define UVD_MPC_SET_MUX__SET_1__SHIFT 0x3 +#define UVD_MPC_SET_MUX__SET_2__SHIFT 0x6 +#define UVD_MPC_SET_MUX__SET_0_MASK 0x00000007L +#define UVD_MPC_SET_MUX__SET_1_MASK 0x00000038L +#define UVD_MPC_SET_MUX__SET_2_MASK 0x000001C0L +//UVD_MPC_SET_ALU +#define UVD_MPC_SET_ALU__FUNCT__SHIFT 0x0 +#define UVD_MPC_SET_ALU__OPERAND__SHIFT 0x4 +#define UVD_MPC_SET_ALU__FUNCT_MASK 0x00000007L +#define UVD_MPC_SET_ALU__OPERAND_MASK 0x00000FF0L +//UVD_MPC_PERF0 +#define UVD_MPC_PERF0__MAX_LAT__SHIFT 0x0 +#define UVD_MPC_PERF0__MAX_LAT_MASK 0x000003FFL +//UVD_MPC_PERF1 +#define UVD_MPC_PERF1__AVE_LAT__SHIFT 0x0 +#define UVD_MPC_PERF1__AVE_LAT_MASK 0x000003FFL +//UVD_MPC_IND_INDEX +#define UVD_MPC_IND_INDEX__INDEX__SHIFT 0x0 +#define UVD_MPC_IND_INDEX__INDEX_MASK 0x000001FFL +//UVD_MPC_IND_DATA +#define UVD_MPC_IND_DATA__DATA__SHIFT 0x0 +#define UVD_MPC_IND_DATA__DATA_MASK 0xFFFFFFFFL + + +// addressBlock: aid_uvd0_uvd_rbcdec +//UVD_RBC_IB_SIZE +#define UVD_RBC_IB_SIZE__IB_SIZE__SHIFT 0x4 +#define UVD_RBC_IB_SIZE__IB_SIZE_MASK 0x007FFFF0L +//UVD_RBC_IB_SIZE_UPDATE +#define UVD_RBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE__SHIFT 0x4 +#define UVD_RBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE_MASK 0x007FFFF0L +//UVD_RBC_RB_CNTL +#define UVD_RBC_RB_CNTL__RB_BUFSZ__SHIFT 0x0 +#define UVD_RBC_RB_CNTL__RB_BLKSZ__SHIFT 0x8 +#define UVD_RBC_RB_CNTL__RB_NO_FETCH__SHIFT 0x10 +#define UVD_RBC_RB_CNTL__RB_WPTR_POLL_EN__SHIFT 0x14 +#define UVD_RBC_RB_CNTL__RB_NO_UPDATE__SHIFT 0x18 +#define UVD_RBC_RB_CNTL__RB_RPTR_WR_EN__SHIFT 0x1c +#define UVD_RBC_RB_CNTL__BLK_RST__SHIFT 0x1d +#define UVD_RBC_RB_CNTL__RB_BUFSZ_MASK 0x0000001FL +#define UVD_RBC_RB_CNTL__RB_BLKSZ_MASK 0x00001F00L +#define UVD_RBC_RB_CNTL__RB_NO_FETCH_MASK 0x00010000L +#define UVD_RBC_RB_CNTL__RB_WPTR_POLL_EN_MASK 0x00100000L +#define UVD_RBC_RB_CNTL__RB_NO_UPDATE_MASK 0x01000000L +#define UVD_RBC_RB_CNTL__RB_RPTR_WR_EN_MASK 0x10000000L +#define UVD_RBC_RB_CNTL__BLK_RST_MASK 0x20000000L +//UVD_RBC_RB_RPTR_ADDR +#define UVD_RBC_RB_RPTR_ADDR__RB_RPTR_ADDR__SHIFT 0x0 +#define UVD_RBC_RB_RPTR_ADDR__RB_RPTR_ADDR_MASK 0xFFFFFFFFL +//UVD_RBC_VCPU_ACCESS +#define UVD_RBC_VCPU_ACCESS__ENABLE_RBC__SHIFT 0x0 +#define UVD_RBC_VCPU_ACCESS__ENABLE_RBC_MASK 0x00000001L +//UVD_FW_SEMAPHORE_CNTL +#define UVD_FW_SEMAPHORE_CNTL__START__SHIFT 0x0 +#define UVD_FW_SEMAPHORE_CNTL__BUSY__SHIFT 0x8 +#define UVD_FW_SEMAPHORE_CNTL__PASS__SHIFT 0x9 +#define UVD_FW_SEMAPHORE_CNTL__START_MASK 0x00000001L +#define UVD_FW_SEMAPHORE_CNTL__BUSY_MASK 0x00000100L +#define UVD_FW_SEMAPHORE_CNTL__PASS_MASK 0x00000200L +//UVD_RBC_READ_REQ_URGENT_CNTL +#define UVD_RBC_READ_REQ_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK__SHIFT 0x0 +#define UVD_RBC_READ_REQ_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK_MASK 0x00000003L +//UVD_RBC_RB_WPTR_CNTL +#define UVD_RBC_RB_WPTR_CNTL__RB_PRE_WRITE_TIMER__SHIFT 0x0 +#define UVD_RBC_RB_WPTR_CNTL__RB_PRE_WRITE_TIMER_MASK 0x00007FFFL +//UVD_RBC_WPTR_STATUS +#define UVD_RBC_WPTR_STATUS__RB_WPTR_IN_USE__SHIFT 0x4 +#define UVD_RBC_WPTR_STATUS__RB_WPTR_IN_USE_MASK 0x007FFFF0L +//UVD_RBC_WPTR_POLL_CNTL +#define UVD_RBC_WPTR_POLL_CNTL__POLL_FREQ__SHIFT 0x0 +#define UVD_RBC_WPTR_POLL_CNTL__IDLE_POLL_COUNT__SHIFT 0x10 +#define UVD_RBC_WPTR_POLL_CNTL__POLL_FREQ_MASK 0x0000FFFFL +#define UVD_RBC_WPTR_POLL_CNTL__IDLE_POLL_COUNT_MASK 0xFFFF0000L +//UVD_RBC_WPTR_POLL_ADDR +#define UVD_RBC_WPTR_POLL_ADDR__POLL_ADDR__SHIFT 0x2 +#define UVD_RBC_WPTR_POLL_ADDR__POLL_ADDR_MASK 0xFFFFFFFCL +//UVD_SEMA_CMD +#define UVD_SEMA_CMD__REQ_CMD__SHIFT 0x0 +#define UVD_SEMA_CMD__WR_PHASE__SHIFT 0x4 +#define UVD_SEMA_CMD__MODE__SHIFT 0x6 +#define UVD_SEMA_CMD__VMID_EN__SHIFT 0x7 +#define UVD_SEMA_CMD__VMID__SHIFT 0x8 +#define UVD_SEMA_CMD__REQ_CMD_MASK 0x0000000FL +#define UVD_SEMA_CMD__WR_PHASE_MASK 0x00000030L +#define UVD_SEMA_CMD__MODE_MASK 0x00000040L +#define UVD_SEMA_CMD__VMID_EN_MASK 0x00000080L +#define UVD_SEMA_CMD__VMID_MASK 0x00000F00L +//UVD_SEMA_ADDR_LOW +#define UVD_SEMA_ADDR_LOW__ADDR_26_3__SHIFT 0x0 +#define UVD_SEMA_ADDR_LOW__ADDR_26_3_MASK 0x00FFFFFFL +//UVD_SEMA_ADDR_HIGH +#define UVD_SEMA_ADDR_HIGH__ADDR_47_27__SHIFT 0x0 +#define UVD_SEMA_ADDR_HIGH__ADDR_47_27_MASK 0x001FFFFFL +//UVD_ENGINE_CNTL +#define UVD_ENGINE_CNTL__ENGINE_START__SHIFT 0x0 +#define UVD_ENGINE_CNTL__ENGINE_START_MODE__SHIFT 0x1 +#define UVD_ENGINE_CNTL__NJ_PF_HANDLE_DISABLE__SHIFT 0x2 +#define UVD_ENGINE_CNTL__ENGINE_START_MASK 0x00000001L +#define UVD_ENGINE_CNTL__ENGINE_START_MODE_MASK 0x00000002L +#define UVD_ENGINE_CNTL__NJ_PF_HANDLE_DISABLE_MASK 0x00000004L +//UVD_SEMA_TIMEOUT_STATUS +#define UVD_SEMA_TIMEOUT_STATUS__SEMAPHORE_WAIT_INCOMPLETE_TIMEOUT_STAT__SHIFT 0x0 +#define UVD_SEMA_TIMEOUT_STATUS__SEMAPHORE_WAIT_FAULT_TIMEOUT_STAT__SHIFT 0x1 +#define UVD_SEMA_TIMEOUT_STATUS__SEMAPHORE_SIGNAL_INCOMPLETE_TIMEOUT_STAT__SHIFT 0x2 +#define UVD_SEMA_TIMEOUT_STATUS__SEMAPHORE_TIMEOUT_CLEAR__SHIFT 0x3 +#define UVD_SEMA_TIMEOUT_STATUS__SEMAPHORE_WAIT_INCOMPLETE_TIMEOUT_STAT_MASK 0x00000001L +#define UVD_SEMA_TIMEOUT_STATUS__SEMAPHORE_WAIT_FAULT_TIMEOUT_STAT_MASK 0x00000002L +#define UVD_SEMA_TIMEOUT_STATUS__SEMAPHORE_SIGNAL_INCOMPLETE_TIMEOUT_STAT_MASK 0x00000004L +#define UVD_SEMA_TIMEOUT_STATUS__SEMAPHORE_TIMEOUT_CLEAR_MASK 0x00000008L +//UVD_SEMA_CNTL +#define UVD_SEMA_CNTL__SEMAPHORE_EN__SHIFT 0x0 +#define UVD_SEMA_CNTL__ADVANCED_MODE_DIS__SHIFT 0x1 +#define UVD_SEMA_CNTL__SEMAPHORE_EN_MASK 0x00000001L +#define UVD_SEMA_CNTL__ADVANCED_MODE_DIS_MASK 0x00000002L +//UVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL +#define UVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL__SIGNAL_INCOMPLETE_EN__SHIFT 0x0 +#define UVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL__SIGNAL_INCOMPLETE_COUNT__SHIFT 0x1 +#define UVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL__RESEND_TIMER__SHIFT 0x18 +#define UVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL__SIGNAL_INCOMPLETE_EN_MASK 0x00000001L +#define UVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL__SIGNAL_INCOMPLETE_COUNT_MASK 0x001FFFFEL +#define UVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL__RESEND_TIMER_MASK 0x07000000L +//UVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL +#define UVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL__WAIT_FAULT_EN__SHIFT 0x0 +#define UVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL__WAIT_FAULT_COUNT__SHIFT 0x1 +#define UVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL__RESEND_TIMER__SHIFT 0x18 +#define UVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL__WAIT_FAULT_EN_MASK 0x00000001L +#define UVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL__WAIT_FAULT_COUNT_MASK 0x001FFFFEL +#define UVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL__RESEND_TIMER_MASK 0x07000000L +//UVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL +#define UVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL__WAIT_INCOMPLETE_EN__SHIFT 0x0 +#define UVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL__WAIT_INCOMPLETE_COUNT__SHIFT 0x1 +#define UVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL__RESEND_TIMER__SHIFT 0x18 +#define UVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL__WAIT_INCOMPLETE_EN_MASK 0x00000001L +#define UVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL__WAIT_INCOMPLETE_COUNT_MASK 0x001FFFFEL +#define UVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL__RESEND_TIMER_MASK 0x07000000L +//UVD_JOB_START +#define UVD_JOB_START__JOB_START__SHIFT 0x0 +#define UVD_JOB_START__JOB_START_MASK 0x00000001L +//UVD_RBC_BUF_STATUS +#define UVD_RBC_BUF_STATUS__RB_BUF_VALID__SHIFT 0x0 +#define UVD_RBC_BUF_STATUS__IB_BUF_VALID__SHIFT 0x8 +#define UVD_RBC_BUF_STATUS__RB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_RBC_BUF_STATUS__IB_BUF_RD_ADDR__SHIFT 0x13 +#define UVD_RBC_BUF_STATUS__RB_BUF_WR_ADDR__SHIFT 0x16 +#define UVD_RBC_BUF_STATUS__IB_BUF_WR_ADDR__SHIFT 0x19 +#define UVD_RBC_BUF_STATUS__RB_BUF_VALID_MASK 0x000000FFL +#define UVD_RBC_BUF_STATUS__IB_BUF_VALID_MASK 0x0000FF00L +#define UVD_RBC_BUF_STATUS__RB_BUF_RD_ADDR_MASK 0x00070000L +#define UVD_RBC_BUF_STATUS__IB_BUF_RD_ADDR_MASK 0x00380000L +#define UVD_RBC_BUF_STATUS__RB_BUF_WR_ADDR_MASK 0x01C00000L +#define UVD_RBC_BUF_STATUS__IB_BUF_WR_ADDR_MASK 0x0E000000L +//UVD_RBC_SWAP_CNTL +#define UVD_RBC_SWAP_CNTL__RB_MC_SWAP__SHIFT 0x0 +#define UVD_RBC_SWAP_CNTL__IB_MC_SWAP__SHIFT 0x2 +#define UVD_RBC_SWAP_CNTL__RB_RPTR_MC_SWAP__SHIFT 0x4 +#define UVD_RBC_SWAP_CNTL__RB_WR_MC_SWAP__SHIFT 0x1a +#define UVD_RBC_SWAP_CNTL__RB_MC_SWAP_MASK 0x00000003L +#define UVD_RBC_SWAP_CNTL__IB_MC_SWAP_MASK 0x0000000CL +#define UVD_RBC_SWAP_CNTL__RB_RPTR_MC_SWAP_MASK 0x00000030L +#define UVD_RBC_SWAP_CNTL__RB_WR_MC_SWAP_MASK 0x0C000000L + + +// addressBlock: aid_uvd0_lmi_adpdec +//UVD_LMI_RE_64BIT_BAR_LOW +#define UVD_LMI_RE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_RE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_RE_64BIT_BAR_HIGH +#define UVD_LMI_RE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_RE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_IT_64BIT_BAR_LOW +#define UVD_LMI_IT_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_IT_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_IT_64BIT_BAR_HIGH +#define UVD_LMI_IT_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_IT_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MP_64BIT_BAR_LOW +#define UVD_LMI_MP_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MP_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MP_64BIT_BAR_HIGH +#define UVD_LMI_MP_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MP_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_CM_64BIT_BAR_LOW +#define UVD_LMI_CM_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_CM_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_CM_64BIT_BAR_HIGH +#define UVD_LMI_CM_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_CM_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_DB_64BIT_BAR_LOW +#define UVD_LMI_DB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_DB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_DB_64BIT_BAR_HIGH +#define UVD_LMI_DB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_DB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_DBW_64BIT_BAR_LOW +#define UVD_LMI_DBW_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_DBW_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_DBW_64BIT_BAR_HIGH +#define UVD_LMI_DBW_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_DBW_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_IDCT_64BIT_BAR_LOW +#define UVD_LMI_IDCT_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_IDCT_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_IDCT_64BIT_BAR_HIGH +#define UVD_LMI_IDCT_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_IDCT_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MPRD_S0_64BIT_BAR_LOW +#define UVD_LMI_MPRD_S0_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MPRD_S0_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MPRD_S0_64BIT_BAR_HIGH +#define UVD_LMI_MPRD_S0_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MPRD_S0_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MPRD_S1_64BIT_BAR_LOW +#define UVD_LMI_MPRD_S1_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MPRD_S1_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MPRD_S1_64BIT_BAR_HIGH +#define UVD_LMI_MPRD_S1_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MPRD_S1_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MPRD_DBW_64BIT_BAR_LOW +#define UVD_LMI_MPRD_DBW_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MPRD_DBW_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MPRD_DBW_64BIT_BAR_HIGH +#define UVD_LMI_MPRD_DBW_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MPRD_DBW_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MPC_64BIT_BAR_LOW +#define UVD_LMI_MPC_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MPC_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MPC_64BIT_BAR_HIGH +#define UVD_LMI_MPC_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MPC_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_RBC_RB_64BIT_BAR_LOW +#define UVD_LMI_RBC_RB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_RBC_RB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_RBC_RB_64BIT_BAR_HIGH +#define UVD_LMI_RBC_RB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_RBC_RB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_RBC_IB_64BIT_BAR_LOW +#define UVD_LMI_RBC_IB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_RBC_IB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_RBC_IB_64BIT_BAR_HIGH +#define UVD_LMI_RBC_IB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_RBC_IB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_LBSI_64BIT_BAR_LOW +#define UVD_LMI_LBSI_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_LBSI_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_LBSI_64BIT_BAR_HIGH +#define UVD_LMI_LBSI_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_LBSI_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_NC0_64BIT_BAR_LOW +#define UVD_LMI_VCPU_NC0_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_VCPU_NC0_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_NC0_64BIT_BAR_HIGH +#define UVD_LMI_VCPU_NC0_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_VCPU_NC0_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_NC1_64BIT_BAR_LOW +#define UVD_LMI_VCPU_NC1_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_VCPU_NC1_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_NC1_64BIT_BAR_HIGH +#define UVD_LMI_VCPU_NC1_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_VCPU_NC1_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE_64BIT_BAR_LOW +#define UVD_LMI_VCPU_CACHE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH +#define UVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_CENC_64BIT_BAR_LOW +#define UVD_LMI_CENC_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_CENC_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_CENC_64BIT_BAR_HIGH +#define UVD_LMI_CENC_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_CENC_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_SRE_64BIT_BAR_LOW +#define UVD_LMI_SRE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_SRE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_SRE_64BIT_BAR_HIGH +#define UVD_LMI_SRE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_SRE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_GPGPU_64BIT_BAR_LOW +#define UVD_LMI_MIF_GPGPU_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_GPGPU_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_GPGPU_64BIT_BAR_HIGH +#define UVD_LMI_MIF_GPGPU_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_GPGPU_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_CURR_LUMA_64BIT_BAR_LOW +#define UVD_LMI_MIF_CURR_LUMA_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_CURR_LUMA_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_CURR_LUMA_64BIT_BAR_HIGH +#define UVD_LMI_MIF_CURR_LUMA_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_CURR_LUMA_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_CURR_CHROMA_64BIT_BAR_LOW +#define UVD_LMI_MIF_CURR_CHROMA_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_CURR_CHROMA_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_CURR_CHROMA_64BIT_BAR_HIGH +#define UVD_LMI_MIF_CURR_CHROMA_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_CURR_CHROMA_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_REF_64BIT_BAR_LOW +#define UVD_LMI_MIF_REF_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_REF_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_REF_64BIT_BAR_HIGH +#define UVD_LMI_MIF_REF_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_REF_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_DBW_64BIT_BAR_LOW +#define UVD_LMI_MIF_DBW_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_DBW_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_DBW_64BIT_BAR_HIGH +#define UVD_LMI_MIF_DBW_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_DBW_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_CM_COLOC_64BIT_BAR_LOW +#define UVD_LMI_MIF_CM_COLOC_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_CM_COLOC_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_CM_COLOC_64BIT_BAR_HIGH +#define UVD_LMI_MIF_CM_COLOC_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_CM_COLOC_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSP0_64BIT_BAR_LOW +#define UVD_LMI_MIF_BSP0_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_BSP0_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSP0_64BIT_BAR_HIGH +#define UVD_LMI_MIF_BSP0_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_BSP0_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSP1_64BIT_BAR_LOW +#define UVD_LMI_MIF_BSP1_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_BSP1_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSP1_64BIT_BAR_HIGH +#define UVD_LMI_MIF_BSP1_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_BSP1_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSP2_64BIT_BAR_LOW +#define UVD_LMI_MIF_BSP2_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_BSP2_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSP2_64BIT_BAR_HIGH +#define UVD_LMI_MIF_BSP2_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_BSP2_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSP3_64BIT_BAR_LOW +#define UVD_LMI_MIF_BSP3_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_BSP3_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSP3_64BIT_BAR_HIGH +#define UVD_LMI_MIF_BSP3_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_BSP3_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSD0_64BIT_BAR_LOW +#define UVD_LMI_MIF_BSD0_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_BSD0_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSD0_64BIT_BAR_HIGH +#define UVD_LMI_MIF_BSD0_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_BSD0_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSD1_64BIT_BAR_LOW +#define UVD_LMI_MIF_BSD1_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_BSD1_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSD1_64BIT_BAR_HIGH +#define UVD_LMI_MIF_BSD1_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_BSD1_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSD2_64BIT_BAR_LOW +#define UVD_LMI_MIF_BSD2_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_BSD2_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSD2_64BIT_BAR_HIGH +#define UVD_LMI_MIF_BSD2_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_BSD2_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSD3_64BIT_BAR_LOW +#define UVD_LMI_MIF_BSD3_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_BSD3_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSD3_64BIT_BAR_HIGH +#define UVD_LMI_MIF_BSD3_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_BSD3_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSD4_64BIT_BAR_LOW +#define UVD_LMI_MIF_BSD4_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_BSD4_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_BSD4_64BIT_BAR_HIGH +#define UVD_LMI_MIF_BSD4_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_BSD4_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW +#define UVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH +#define UVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE8_64BIT_BAR_LOW +#define UVD_LMI_VCPU_CACHE8_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE8_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE8_64BIT_BAR_HIGH +#define UVD_LMI_VCPU_CACHE8_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE8_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW +#define UVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH +#define UVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE3_64BIT_BAR_LOW +#define UVD_LMI_VCPU_CACHE3_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE3_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE3_64BIT_BAR_HIGH +#define UVD_LMI_VCPU_CACHE3_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE3_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE4_64BIT_BAR_LOW +#define UVD_LMI_VCPU_CACHE4_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE4_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE4_64BIT_BAR_HIGH +#define UVD_LMI_VCPU_CACHE4_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE4_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE5_64BIT_BAR_LOW +#define UVD_LMI_VCPU_CACHE5_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE5_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE5_64BIT_BAR_HIGH +#define UVD_LMI_VCPU_CACHE5_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE5_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE6_64BIT_BAR_LOW +#define UVD_LMI_VCPU_CACHE6_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE6_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE6_64BIT_BAR_HIGH +#define UVD_LMI_VCPU_CACHE6_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE6_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE7_64BIT_BAR_LOW +#define UVD_LMI_VCPU_CACHE7_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE7_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_VCPU_CACHE7_64BIT_BAR_HIGH +#define UVD_LMI_VCPU_CACHE7_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE7_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_SCLR_64BIT_BAR_LOW +#define UVD_LMI_MIF_SCLR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_SCLR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_SCLR_64BIT_BAR_HIGH +#define UVD_LMI_MIF_SCLR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_SCLR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_SCLR2_64BIT_BAR_LOW +#define UVD_LMI_MIF_SCLR2_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_SCLR2_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_SCLR2_64BIT_BAR_HIGH +#define UVD_LMI_MIF_SCLR2_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_SCLR2_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_SPH_64BIT_BAR_HIGH +#define UVD_LMI_SPH_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_SPH_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_IMAGEPASTE_LUMA_64BIT_BAR_LOW +#define UVD_LMI_MIF_IMAGEPASTE_LUMA_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_IMAGEPASTE_LUMA_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_IMAGEPASTE_LUMA_64BIT_BAR_HIGH +#define UVD_LMI_MIF_IMAGEPASTE_LUMA_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_IMAGEPASTE_LUMA_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_IMAGEPASTE_CHROMA_64BIT_BAR_LOW +#define UVD_LMI_MIF_IMAGEPASTE_CHROMA_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_IMAGEPASTE_CHROMA_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_IMAGEPASTE_CHROMA_64BIT_BAR_HIGH +#define UVD_LMI_MIF_IMAGEPASTE_CHROMA_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_IMAGEPASTE_CHROMA_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_PRIVACY_LUMA_64BIT_BAR_LOW +#define UVD_LMI_MIF_PRIVACY_LUMA_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_PRIVACY_LUMA_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_PRIVACY_LUMA_64BIT_BAR_HIGH +#define UVD_LMI_MIF_PRIVACY_LUMA_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_PRIVACY_LUMA_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_PRIVACY_CHROMA_64BIT_BAR_LOW +#define UVD_LMI_MIF_PRIVACY_CHROMA_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_PRIVACY_CHROMA_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_PRIVACY_CHROMA_64BIT_BAR_HIGH +#define UVD_LMI_MIF_PRIVACY_CHROMA_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_PRIVACY_CHROMA_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_ADP_ATOMIC_CONFIG +#define UVD_ADP_ATOMIC_CONFIG__ATOMIC_USER0_WR_CACHE__SHIFT 0x0 +#define UVD_ADP_ATOMIC_CONFIG__ATOMIC_USER1_WR_CACHE__SHIFT 0x4 +#define UVD_ADP_ATOMIC_CONFIG__ATOMIC_USER2_WR_CACHE__SHIFT 0x8 +#define UVD_ADP_ATOMIC_CONFIG__ATOMIC_USER3_WR_CACHE__SHIFT 0xc +#define UVD_ADP_ATOMIC_CONFIG__ATOMIC_RD_URG__SHIFT 0x10 +#define UVD_ADP_ATOMIC_CONFIG__ATOMIC_USER0_WR_CACHE_MASK 0x0000000FL +#define UVD_ADP_ATOMIC_CONFIG__ATOMIC_USER1_WR_CACHE_MASK 0x000000F0L +#define UVD_ADP_ATOMIC_CONFIG__ATOMIC_USER2_WR_CACHE_MASK 0x00000F00L +#define UVD_ADP_ATOMIC_CONFIG__ATOMIC_USER3_WR_CACHE_MASK 0x0000F000L +#define UVD_ADP_ATOMIC_CONFIG__ATOMIC_RD_URG_MASK 0x000F0000L +//UVD_LMI_ARB_CTRL2 +#define UVD_LMI_ARB_CTRL2__CENC_RD_WAIT_EN__SHIFT 0x0 +#define UVD_LMI_ARB_CTRL2__ATOMIC_WR_WAIT_EN__SHIFT 0x1 +#define UVD_LMI_ARB_CTRL2__CENC_RD_MAX_BURST__SHIFT 0x2 +#define UVD_LMI_ARB_CTRL2__ATOMIC_WR_MAX_BURST__SHIFT 0x6 +#define UVD_LMI_ARB_CTRL2__MIF_RD_REQ_RET_MAX__SHIFT 0xa +#define UVD_LMI_ARB_CTRL2__MIF_WR_REQ_RET_MAX__SHIFT 0x14 +#define UVD_LMI_ARB_CTRL2__CENC_RD_WAIT_EN_MASK 0x00000001L +#define UVD_LMI_ARB_CTRL2__ATOMIC_WR_WAIT_EN_MASK 0x00000002L +#define UVD_LMI_ARB_CTRL2__CENC_RD_MAX_BURST_MASK 0x0000003CL +#define UVD_LMI_ARB_CTRL2__ATOMIC_WR_MAX_BURST_MASK 0x000003C0L +#define UVD_LMI_ARB_CTRL2__MIF_RD_REQ_RET_MAX_MASK 0x000FFC00L +#define UVD_LMI_ARB_CTRL2__MIF_WR_REQ_RET_MAX_MASK 0xFFF00000L +//UVD_LMI_VCPU_CACHE_VMIDS_MULTI +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE1_VMID__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE2_VMID__SHIFT 0x4 +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE3_VMID__SHIFT 0x8 +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE4_VMID__SHIFT 0xc +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE5_VMID__SHIFT 0x10 +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE6_VMID__SHIFT 0x14 +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE7_VMID__SHIFT 0x18 +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE8_VMID__SHIFT 0x1c +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE1_VMID_MASK 0x0000000FL +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE2_VMID_MASK 0x000000F0L +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE3_VMID_MASK 0x00000F00L +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE4_VMID_MASK 0x0000F000L +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE5_VMID_MASK 0x000F0000L +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE6_VMID_MASK 0x00F00000L +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE7_VMID_MASK 0x0F000000L +#define UVD_LMI_VCPU_CACHE_VMIDS_MULTI__VCPU_CACHE8_VMID_MASK 0xF0000000L +//UVD_LMI_VCPU_NC_VMIDS_MULTI +#define UVD_LMI_VCPU_NC_VMIDS_MULTI__VCPU_NC2_VMID__SHIFT 0x4 +#define UVD_LMI_VCPU_NC_VMIDS_MULTI__VCPU_NC3_VMID__SHIFT 0x8 +#define UVD_LMI_VCPU_NC_VMIDS_MULTI__VCPU_NC4_VMID__SHIFT 0xc +#define UVD_LMI_VCPU_NC_VMIDS_MULTI__VCPU_NC5_VMID__SHIFT 0x10 +#define UVD_LMI_VCPU_NC_VMIDS_MULTI__VCPU_NC6_VMID__SHIFT 0x14 +#define UVD_LMI_VCPU_NC_VMIDS_MULTI__VCPU_NC7_VMID__SHIFT 0x18 +#define UVD_LMI_VCPU_NC_VMIDS_MULTI__VCPU_NC2_VMID_MASK 0x000000F0L +#define UVD_LMI_VCPU_NC_VMIDS_MULTI__VCPU_NC3_VMID_MASK 0x00000F00L +#define UVD_LMI_VCPU_NC_VMIDS_MULTI__VCPU_NC4_VMID_MASK 0x0000F000L +#define UVD_LMI_VCPU_NC_VMIDS_MULTI__VCPU_NC5_VMID_MASK 0x000F0000L +#define UVD_LMI_VCPU_NC_VMIDS_MULTI__VCPU_NC6_VMID_MASK 0x00F00000L +#define UVD_LMI_VCPU_NC_VMIDS_MULTI__VCPU_NC7_VMID_MASK 0x0F000000L +//UVD_LMI_LAT_CTRL +#define UVD_LMI_LAT_CTRL__SCALE__SHIFT 0x0 +#define UVD_LMI_LAT_CTRL__MAX_START__SHIFT 0x8 +#define UVD_LMI_LAT_CTRL__MIN_START__SHIFT 0x9 +#define UVD_LMI_LAT_CTRL__AVG_START__SHIFT 0xa +#define UVD_LMI_LAT_CTRL__PERFMON_SYNC__SHIFT 0xb +#define UVD_LMI_LAT_CTRL__SKIP__SHIFT 0x10 +#define UVD_LMI_LAT_CTRL__SCALE_MASK 0x000000FFL +#define UVD_LMI_LAT_CTRL__MAX_START_MASK 0x00000100L +#define UVD_LMI_LAT_CTRL__MIN_START_MASK 0x00000200L +#define UVD_LMI_LAT_CTRL__AVG_START_MASK 0x00000400L +#define UVD_LMI_LAT_CTRL__PERFMON_SYNC_MASK 0x00000800L +#define UVD_LMI_LAT_CTRL__SKIP_MASK 0x000F0000L +//UVD_LMI_LAT_CNTR +#define UVD_LMI_LAT_CNTR__MAX_LAT__SHIFT 0x0 +#define UVD_LMI_LAT_CNTR__MIN_LAT__SHIFT 0x8 +#define UVD_LMI_LAT_CNTR__MAX_LAT_MASK 0x000000FFL +#define UVD_LMI_LAT_CNTR__MIN_LAT_MASK 0x0000FF00L +//UVD_LMI_AVG_LAT_CNTR +#define UVD_LMI_AVG_LAT_CNTR__ENV_LOW__SHIFT 0x0 +#define UVD_LMI_AVG_LAT_CNTR__ENV_HIGH__SHIFT 0x8 +#define UVD_LMI_AVG_LAT_CNTR__ENV_HIT__SHIFT 0x10 +#define UVD_LMI_AVG_LAT_CNTR__ENV_LOW_MASK 0x000000FFL +#define UVD_LMI_AVG_LAT_CNTR__ENV_HIGH_MASK 0x0000FF00L +#define UVD_LMI_AVG_LAT_CNTR__ENV_HIT_MASK 0xFFFF0000L +//UVD_LMI_SPH +#define UVD_LMI_SPH__ADDR__SHIFT 0x0 +#define UVD_LMI_SPH__STS__SHIFT 0x1c +#define UVD_LMI_SPH__STS_VALID__SHIFT 0x1e +#define UVD_LMI_SPH__STS_OVERFLOW__SHIFT 0x1f +#define UVD_LMI_SPH__ADDR_MASK 0x0FFFFFFFL +#define UVD_LMI_SPH__STS_MASK 0x30000000L +#define UVD_LMI_SPH__STS_VALID_MASK 0x40000000L +#define UVD_LMI_SPH__STS_OVERFLOW_MASK 0x80000000L +//UVD_LMI_VCPU_CACHE_VMID +#define UVD_LMI_VCPU_CACHE_VMID__VCPU_CACHE_VMID__SHIFT 0x0 +#define UVD_LMI_VCPU_CACHE_VMID__VCPU_CACHE_VMID_MASK 0x0000000FL +//UVD_LMI_CTRL2 +#define UVD_LMI_CTRL2__SPH_DIS__SHIFT 0x0 +#define UVD_LMI_CTRL2__STALL_ARB__SHIFT 0x1 +#define UVD_LMI_CTRL2__ASSERT_UMC_URGENT__SHIFT 0x2 +#define UVD_LMI_CTRL2__MASK_UMC_URGENT__SHIFT 0x3 +#define UVD_LMI_CTRL2__CRC1_RESET__SHIFT 0x4 +#define UVD_LMI_CTRL2__DRCITF_BUBBLE_FIX_DIS__SHIFT 0x7 +#define UVD_LMI_CTRL2__STALL_ARB_UMC__SHIFT 0x8 +#define UVD_LMI_CTRL2__MC_READ_ID_SEL__SHIFT 0x9 +#define UVD_LMI_CTRL2__MC_WRITE_ID_SEL__SHIFT 0xb +#define UVD_LMI_CTRL2__VCPU_NC0_EXT_EN__SHIFT 0xd +#define UVD_LMI_CTRL2__VCPU_NC1_EXT_EN__SHIFT 0xe +#define UVD_LMI_CTRL2__SPU_EXTRA_CID_EN__SHIFT 0xf +#define UVD_LMI_CTRL2__RE_OFFLOAD_EN__SHIFT 0x10 +#define UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM__SHIFT 0x11 +#define UVD_LMI_CTRL2__CLEAR_NJ_PF_BP__SHIFT 0x19 +#define UVD_LMI_CTRL2__NJ_MIF_GATING__SHIFT 0x1a +#define UVD_LMI_CTRL2__CRC1_SEL__SHIFT 0x1b +#define UVD_LMI_CTRL2__SPH_DIS_MASK 0x00000001L +#define UVD_LMI_CTRL2__STALL_ARB_MASK 0x00000002L +#define UVD_LMI_CTRL2__ASSERT_UMC_URGENT_MASK 0x00000004L +#define UVD_LMI_CTRL2__MASK_UMC_URGENT_MASK 0x00000008L +#define UVD_LMI_CTRL2__CRC1_RESET_MASK 0x00000010L +#define UVD_LMI_CTRL2__DRCITF_BUBBLE_FIX_DIS_MASK 0x00000080L +#define UVD_LMI_CTRL2__STALL_ARB_UMC_MASK 0x00000100L +#define UVD_LMI_CTRL2__MC_READ_ID_SEL_MASK 0x00000600L +#define UVD_LMI_CTRL2__MC_WRITE_ID_SEL_MASK 0x00001800L +#define UVD_LMI_CTRL2__VCPU_NC0_EXT_EN_MASK 0x00002000L +#define UVD_LMI_CTRL2__VCPU_NC1_EXT_EN_MASK 0x00004000L +#define UVD_LMI_CTRL2__SPU_EXTRA_CID_EN_MASK 0x00008000L +#define UVD_LMI_CTRL2__RE_OFFLOAD_EN_MASK 0x00010000L +#define UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM_MASK 0x01FE0000L +#define UVD_LMI_CTRL2__CLEAR_NJ_PF_BP_MASK 0x02000000L +#define UVD_LMI_CTRL2__NJ_MIF_GATING_MASK 0x04000000L +#define UVD_LMI_CTRL2__CRC1_SEL_MASK 0xF8000000L +//UVD_LMI_URGENT_CTRL +#define UVD_LMI_URGENT_CTRL__ENABLE_MC_RD_URGENT_STALL__SHIFT 0x0 +#define UVD_LMI_URGENT_CTRL__ASSERT_MC_RD_STALL__SHIFT 0x1 +#define UVD_LMI_URGENT_CTRL__ASSERT_MC_RD_URGENT__SHIFT 0x2 +#define UVD_LMI_URGENT_CTRL__ENABLE_MC_WR_URGENT_STALL__SHIFT 0x8 +#define UVD_LMI_URGENT_CTRL__ASSERT_MC_WR_STALL__SHIFT 0x9 +#define UVD_LMI_URGENT_CTRL__ASSERT_MC_WR_URGENT__SHIFT 0xa +#define UVD_LMI_URGENT_CTRL__ENABLE_UMC_RD_URGENT_STALL__SHIFT 0x10 +#define UVD_LMI_URGENT_CTRL__ASSERT_UMC_RD_STALL__SHIFT 0x11 +#define UVD_LMI_URGENT_CTRL__ASSERT_UMC_RD_URGENT__SHIFT 0x12 +#define UVD_LMI_URGENT_CTRL__ENABLE_UMC_WR_URGENT_STALL__SHIFT 0x18 +#define UVD_LMI_URGENT_CTRL__ASSERT_UMC_WR_STALL__SHIFT 0x19 +#define UVD_LMI_URGENT_CTRL__ASSERT_UMC_WR_URGENT__SHIFT 0x1a +#define UVD_LMI_URGENT_CTRL__ENABLE_MC_RD_URGENT_STALL_MASK 0x00000001L +#define UVD_LMI_URGENT_CTRL__ASSERT_MC_RD_STALL_MASK 0x00000002L +#define UVD_LMI_URGENT_CTRL__ASSERT_MC_RD_URGENT_MASK 0x0000003CL +#define UVD_LMI_URGENT_CTRL__ENABLE_MC_WR_URGENT_STALL_MASK 0x00000100L +#define UVD_LMI_URGENT_CTRL__ASSERT_MC_WR_STALL_MASK 0x00000200L +#define UVD_LMI_URGENT_CTRL__ASSERT_MC_WR_URGENT_MASK 0x00003C00L +#define UVD_LMI_URGENT_CTRL__ENABLE_UMC_RD_URGENT_STALL_MASK 0x00010000L +#define UVD_LMI_URGENT_CTRL__ASSERT_UMC_RD_STALL_MASK 0x00020000L +#define UVD_LMI_URGENT_CTRL__ASSERT_UMC_RD_URGENT_MASK 0x003C0000L +#define UVD_LMI_URGENT_CTRL__ENABLE_UMC_WR_URGENT_STALL_MASK 0x01000000L +#define UVD_LMI_URGENT_CTRL__ASSERT_UMC_WR_STALL_MASK 0x02000000L +#define UVD_LMI_URGENT_CTRL__ASSERT_UMC_WR_URGENT_MASK 0x3C000000L +//UVD_LMI_CTRL +#define UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT 0x0 +#define UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN__SHIFT 0x8 +#define UVD_LMI_CTRL__REQ_MODE__SHIFT 0x9 +#define UVD_LMI_CTRL__ASSERT_MC_URGENT__SHIFT 0xb +#define UVD_LMI_CTRL__MASK_MC_URGENT__SHIFT 0xc +#define UVD_LMI_CTRL__DATA_COHERENCY_EN__SHIFT 0xd +#define UVD_LMI_CTRL__CRC_RESET__SHIFT 0xe +#define UVD_LMI_CTRL__CRC_SEL__SHIFT 0xf +#define UVD_LMI_CTRL__DISABLE_ON_FWV_FAIL__SHIFT 0x14 +#define UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN__SHIFT 0x15 +#define UVD_LMI_CTRL__CM_DATA_COHERENCY_EN__SHIFT 0x16 +#define UVD_LMI_CTRL__DB_DB_DATA_COHERENCY_EN__SHIFT 0x17 +#define UVD_LMI_CTRL__DB_IT_DATA_COHERENCY_EN__SHIFT 0x18 +#define UVD_LMI_CTRL__IT_IT_DATA_COHERENCY_EN__SHIFT 0x19 +#define UVD_LMI_CTRL__MIF_MIF_DATA_COHERENCY_EN__SHIFT 0x1a +#define UVD_LMI_CTRL__MIF_LESS_OUTSTANDING_RD_REQ__SHIFT 0x1b +#define UVD_LMI_CTRL__MC_BLK_RST__SHIFT 0x1c +#define UVD_LMI_CTRL__UMC_BLK_RST__SHIFT 0x1d +#define UVD_LMI_CTRL__RFU__SHIFT 0x1e +#define UVD_LMI_CTRL__WRITE_CLEAN_TIMER_MASK 0x000000FFL +#define UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK 0x00000100L +#define UVD_LMI_CTRL__REQ_MODE_MASK 0x00000200L +#define UVD_LMI_CTRL__ASSERT_MC_URGENT_MASK 0x00000800L +#define UVD_LMI_CTRL__MASK_MC_URGENT_MASK 0x00001000L +#define UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK 0x00002000L +#define UVD_LMI_CTRL__CRC_RESET_MASK 0x00004000L +#define UVD_LMI_CTRL__CRC_SEL_MASK 0x000F8000L +#define UVD_LMI_CTRL__DISABLE_ON_FWV_FAIL_MASK 0x00100000L +#define UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK 0x00200000L +#define UVD_LMI_CTRL__CM_DATA_COHERENCY_EN_MASK 0x00400000L +#define UVD_LMI_CTRL__DB_DB_DATA_COHERENCY_EN_MASK 0x00800000L +#define UVD_LMI_CTRL__DB_IT_DATA_COHERENCY_EN_MASK 0x01000000L +#define UVD_LMI_CTRL__IT_IT_DATA_COHERENCY_EN_MASK 0x02000000L +#define UVD_LMI_CTRL__MIF_MIF_DATA_COHERENCY_EN_MASK 0x04000000L +#define UVD_LMI_CTRL__MIF_LESS_OUTSTANDING_RD_REQ_MASK 0x08000000L +#define UVD_LMI_CTRL__MC_BLK_RST_MASK 0x10000000L +#define UVD_LMI_CTRL__UMC_BLK_RST_MASK 0x20000000L +#define UVD_LMI_CTRL__RFU_MASK 0xC0000000L +//UVD_LMI_STATUS +#define UVD_LMI_STATUS__READ_CLEAN__SHIFT 0x0 +#define UVD_LMI_STATUS__WRITE_CLEAN__SHIFT 0x1 +#define UVD_LMI_STATUS__WRITE_CLEAN_RAW__SHIFT 0x2 +#define UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN__SHIFT 0x3 +#define UVD_LMI_STATUS__UMC_READ_CLEAN__SHIFT 0x4 +#define UVD_LMI_STATUS__UMC_WRITE_CLEAN__SHIFT 0x5 +#define UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW__SHIFT 0x6 +#define UVD_LMI_STATUS__PENDING_UVD_MC_WRITE__SHIFT 0x7 +#define UVD_LMI_STATUS__READ_CLEAN_RAW__SHIFT 0x8 +#define UVD_LMI_STATUS__UMC_READ_CLEAN_RAW__SHIFT 0x9 +#define UVD_LMI_STATUS__UMC_UVD_IDLE__SHIFT 0xa +#define UVD_LMI_STATUS__UMC_AVP_IDLE__SHIFT 0xb +#define UVD_LMI_STATUS__ADP_MC_READ_CLEAN__SHIFT 0xc +#define UVD_LMI_STATUS__ADP_UMC_READ_CLEAN__SHIFT 0xd +#define UVD_LMI_STATUS__BSP0_WRITE_CLEAN__SHIFT 0x12 +#define UVD_LMI_STATUS__BSP1_WRITE_CLEAN__SHIFT 0x13 +#define UVD_LMI_STATUS__BSP2_WRITE_CLEAN__SHIFT 0x14 +#define UVD_LMI_STATUS__BSP3_WRITE_CLEAN__SHIFT 0x15 +#define UVD_LMI_STATUS__CENC_READ_CLEAN__SHIFT 0x16 +#define UVD_LMI_STATUS__READ_CLEAN_MASK 0x00000001L +#define UVD_LMI_STATUS__WRITE_CLEAN_MASK 0x00000002L +#define UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK 0x00000004L +#define UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK 0x00000008L +#define UVD_LMI_STATUS__UMC_READ_CLEAN_MASK 0x00000010L +#define UVD_LMI_STATUS__UMC_WRITE_CLEAN_MASK 0x00000020L +#define UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK 0x00000040L +#define UVD_LMI_STATUS__PENDING_UVD_MC_WRITE_MASK 0x00000080L +#define UVD_LMI_STATUS__READ_CLEAN_RAW_MASK 0x00000100L +#define UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK 0x00000200L +#define UVD_LMI_STATUS__UMC_UVD_IDLE_MASK 0x00000400L +#define UVD_LMI_STATUS__UMC_AVP_IDLE_MASK 0x00000800L +#define UVD_LMI_STATUS__ADP_MC_READ_CLEAN_MASK 0x00001000L +#define UVD_LMI_STATUS__ADP_UMC_READ_CLEAN_MASK 0x00002000L +#define UVD_LMI_STATUS__BSP0_WRITE_CLEAN_MASK 0x00040000L +#define UVD_LMI_STATUS__BSP1_WRITE_CLEAN_MASK 0x00080000L +#define UVD_LMI_STATUS__BSP2_WRITE_CLEAN_MASK 0x00100000L +#define UVD_LMI_STATUS__BSP3_WRITE_CLEAN_MASK 0x00200000L +#define UVD_LMI_STATUS__CENC_READ_CLEAN_MASK 0x00400000L +//UVD_LMI_PERFMON_CTRL +#define UVD_LMI_PERFMON_CTRL__PERFMON_STATE__SHIFT 0x0 +#define UVD_LMI_PERFMON_CTRL__PERFMON_SEL__SHIFT 0x8 +#define UVD_LMI_PERFMON_CTRL__PERFMON_STATE_MASK 0x00000003L +#define UVD_LMI_PERFMON_CTRL__PERFMON_SEL_MASK 0x00001F00L +//UVD_LMI_PERFMON_COUNT_LO +#define UVD_LMI_PERFMON_COUNT_LO__PERFMON_COUNT__SHIFT 0x0 +#define UVD_LMI_PERFMON_COUNT_LO__PERFMON_COUNT_MASK 0xFFFFFFFFL +//UVD_LMI_PERFMON_COUNT_HI +#define UVD_LMI_PERFMON_COUNT_HI__PERFMON_COUNT__SHIFT 0x0 +#define UVD_LMI_PERFMON_COUNT_HI__PERFMON_COUNT_MASK 0x0000FFFFL +//UVD_LMI_ADP_SWAP_CNTL +#define UVD_LMI_ADP_SWAP_CNTL__VCPU_R_MC_SWAP__SHIFT 0x6 +#define UVD_LMI_ADP_SWAP_CNTL__VCPU_W_MC_SWAP__SHIFT 0x8 +#define UVD_LMI_ADP_SWAP_CNTL__CM_MC_SWAP__SHIFT 0xa +#define UVD_LMI_ADP_SWAP_CNTL__IT_MC_SWAP__SHIFT 0xc +#define UVD_LMI_ADP_SWAP_CNTL__DB_R_MC_SWAP__SHIFT 0xe +#define UVD_LMI_ADP_SWAP_CNTL__DB_W_MC_SWAP__SHIFT 0x10 +#define UVD_LMI_ADP_SWAP_CNTL__CSM_MC_SWAP__SHIFT 0x12 +#define UVD_LMI_ADP_SWAP_CNTL__PREF_MC_SWAP__SHIFT 0x14 +#define UVD_LMI_ADP_SWAP_CNTL__DBW_MC_SWAP__SHIFT 0x18 +#define UVD_LMI_ADP_SWAP_CNTL__RE_MC_SWAP__SHIFT 0x1c +#define UVD_LMI_ADP_SWAP_CNTL__MP_MC_SWAP__SHIFT 0x1e +#define UVD_LMI_ADP_SWAP_CNTL__VCPU_R_MC_SWAP_MASK 0x000000C0L +#define UVD_LMI_ADP_SWAP_CNTL__VCPU_W_MC_SWAP_MASK 0x00000300L +#define UVD_LMI_ADP_SWAP_CNTL__CM_MC_SWAP_MASK 0x00000C00L +#define UVD_LMI_ADP_SWAP_CNTL__IT_MC_SWAP_MASK 0x00003000L +#define UVD_LMI_ADP_SWAP_CNTL__DB_R_MC_SWAP_MASK 0x0000C000L +#define UVD_LMI_ADP_SWAP_CNTL__DB_W_MC_SWAP_MASK 0x00030000L +#define UVD_LMI_ADP_SWAP_CNTL__CSM_MC_SWAP_MASK 0x000C0000L +#define UVD_LMI_ADP_SWAP_CNTL__PREF_MC_SWAP_MASK 0x00300000L +#define UVD_LMI_ADP_SWAP_CNTL__DBW_MC_SWAP_MASK 0x03000000L +#define UVD_LMI_ADP_SWAP_CNTL__RE_MC_SWAP_MASK 0x30000000L +#define UVD_LMI_ADP_SWAP_CNTL__MP_MC_SWAP_MASK 0xC0000000L +//UVD_LMI_RBC_RB_VMID +#define UVD_LMI_RBC_RB_VMID__RB_VMID__SHIFT 0x0 +#define UVD_LMI_RBC_RB_VMID__RB_VMID_MASK 0x0000000FL +//UVD_LMI_RBC_IB_VMID +#define UVD_LMI_RBC_IB_VMID__IB_VMID__SHIFT 0x0 +#define UVD_LMI_RBC_IB_VMID__IB_VMID_MASK 0x0000000FL +//UVD_LMI_MC_CREDITS +#define UVD_LMI_MC_CREDITS__UVD_RD_CREDITS__SHIFT 0x0 +#define UVD_LMI_MC_CREDITS__UVD_WR_CREDITS__SHIFT 0x8 +#define UVD_LMI_MC_CREDITS__UMC_RD_CREDITS__SHIFT 0x10 +#define UVD_LMI_MC_CREDITS__UMC_WR_CREDITS__SHIFT 0x18 +#define UVD_LMI_MC_CREDITS__UVD_RD_CREDITS_MASK 0x0000003FL +#define UVD_LMI_MC_CREDITS__UVD_WR_CREDITS_MASK 0x00003F00L +#define UVD_LMI_MC_CREDITS__UMC_RD_CREDITS_MASK 0x003F0000L +#define UVD_LMI_MC_CREDITS__UMC_WR_CREDITS_MASK 0x3F000000L +//UVD_LMI_ADP_IND_INDEX +#define UVD_LMI_ADP_IND_INDEX__INDEX__SHIFT 0x0 +#define UVD_LMI_ADP_IND_INDEX__INDEX_MASK 0x00001FFFL +//UVD_LMI_ADP_IND_DATA +#define UVD_LMI_ADP_IND_DATA__DATA__SHIFT 0x0 +#define UVD_LMI_ADP_IND_DATA__DATA_MASK 0xFFFFFFFFL +//UVD_LMI_ADP_PF_EN +#define UVD_LMI_ADP_PF_EN__VCPU_CACHE0_PF_EN__SHIFT 0x0 +#define UVD_LMI_ADP_PF_EN__VCPU_CACHE1_PF_EN__SHIFT 0x1 +#define UVD_LMI_ADP_PF_EN__VCPU_CACHE2_PF_EN__SHIFT 0x2 +#define UVD_LMI_ADP_PF_EN__VCPU_CACHE0_PF_EN_MASK 0x00000001L +#define UVD_LMI_ADP_PF_EN__VCPU_CACHE1_PF_EN_MASK 0x00000002L +#define UVD_LMI_ADP_PF_EN__VCPU_CACHE2_PF_EN_MASK 0x00000004L +//UVD_LMI_PREF_CTRL +#define UVD_LMI_PREF_CTRL__PREF_RST__SHIFT 0x0 +#define UVD_LMI_PREF_CTRL__PREF_BUSY_STATUS__SHIFT 0x1 +#define UVD_LMI_PREF_CTRL__PREF_WSTRB__SHIFT 0x2 +#define UVD_LMI_PREF_CTRL__PREF_WRITE_SIZE__SHIFT 0x3 +#define UVD_LMI_PREF_CTRL__PREF_STEP_SIZE__SHIFT 0x4 +#define UVD_LMI_PREF_CTRL__PREF_SIZE__SHIFT 0x13 +#define UVD_LMI_PREF_CTRL__PREF_RST_MASK 0x00000001L +#define UVD_LMI_PREF_CTRL__PREF_BUSY_STATUS_MASK 0x00000002L +#define UVD_LMI_PREF_CTRL__PREF_WSTRB_MASK 0x00000004L +#define UVD_LMI_PREF_CTRL__PREF_WRITE_SIZE_MASK 0x00000008L +#define UVD_LMI_PREF_CTRL__PREF_STEP_SIZE_MASK 0x00000070L +#define UVD_LMI_PREF_CTRL__PREF_SIZE_MASK 0xFFF80000L +//UVD_LMI_MIF_REF_LUMA_64BIT_BAR_LOW +#define UVD_LMI_MIF_REF_LUMA_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MIF_REF_LUMA_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MIF_REF_LUMA_64BIT_BAR_HIGH +#define UVD_LMI_MIF_REF_LUMA_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MIF_REF_LUMA_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//VCN_RAS_CNTL +#define VCN_RAS_CNTL__VCPU_VCODEC_IH_EN__SHIFT 0x0 +#define VCN_RAS_CNTL__VCPU_VCODEC_PMI_EN__SHIFT 0x4 +#define VCN_RAS_CNTL__VCPU_VCODEC_REARM__SHIFT 0x8 +#define VCN_RAS_CNTL__VCPU_VCODEC_STALL_EN__SHIFT 0xc +#define VCN_RAS_CNTL__VCPU_VCODEC_READY__SHIFT 0x10 +#define VCN_RAS_CNTL__VCPU_VCODEC_IH_EN_MASK 0x00000001L +#define VCN_RAS_CNTL__VCPU_VCODEC_PMI_EN_MASK 0x00000010L +#define VCN_RAS_CNTL__VCPU_VCODEC_REARM_MASK 0x00000100L +#define VCN_RAS_CNTL__VCPU_VCODEC_STALL_EN_MASK 0x00001000L +#define VCN_RAS_CNTL__VCPU_VCODEC_READY_MASK 0x00010000L + + +// addressBlock: aid_uvd0_uvd_jpeg0_jpegnpdec +//UVD_JPEG_CNTL +#define UVD_JPEG_CNTL__REQUEST_EN__SHIFT 0x1 +#define UVD_JPEG_CNTL__ERR_RST_EN__SHIFT 0x2 +#define UVD_JPEG_CNTL__DBG_MUX_SEL__SHIFT 0x8 +#define UVD_JPEG_CNTL__FORMAT_CONV_EN__SHIFT 0x10 +#define UVD_JPEG_CNTL__VUP_MODE__SHIFT 0x11 +#define UVD_JPEG_CNTL__FC_TIMEOUT_EN__SHIFT 0x12 +#define UVD_JPEG_CNTL__ROI_CROP_EN__SHIFT 0x18 +#define UVD_JPEG_CNTL__ROI_CROP_EARLY_DECODE_STOP_DIS__SHIFT 0x19 +#define UVD_JPEG_CNTL__REQUEST_EN_MASK 0x00000002L +#define UVD_JPEG_CNTL__ERR_RST_EN_MASK 0x00000004L +#define UVD_JPEG_CNTL__DBG_MUX_SEL_MASK 0x00007F00L +#define UVD_JPEG_CNTL__FORMAT_CONV_EN_MASK 0x00010000L +#define UVD_JPEG_CNTL__VUP_MODE_MASK 0x00020000L +#define UVD_JPEG_CNTL__FC_TIMEOUT_EN_MASK 0x00040000L +#define UVD_JPEG_CNTL__ROI_CROP_EN_MASK 0x01000000L +#define UVD_JPEG_CNTL__ROI_CROP_EARLY_DECODE_STOP_DIS_MASK 0x02000000L +//UVD_JPEG_RB_BASE +#define UVD_JPEG_RB_BASE__RB_BYTE_OFF__SHIFT 0x0 +#define UVD_JPEG_RB_BASE__RB_BASE__SHIFT 0x6 +#define UVD_JPEG_RB_BASE__RB_BYTE_OFF_MASK 0x0000003FL +#define UVD_JPEG_RB_BASE__RB_BASE_MASK 0xFFFFFFC0L +//UVD_JPEG_RB_WPTR +#define UVD_JPEG_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_JPEG_RB_WPTR__RB_WPTR_MASK 0x3FFFFFF0L +//UVD_JPEG_RB_RPTR +#define UVD_JPEG_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_JPEG_RB_RPTR__RB_RPTR_MASK 0x3FFFFFF0L +//UVD_JPEG_RB_SIZE +#define UVD_JPEG_RB_SIZE__RB_SIZE__SHIFT 0x4 +#define UVD_JPEG_RB_SIZE__RB_SIZE_MASK 0x3FFFFFF0L +//UVD_JPEG_DEC_CNT +#define UVD_JPEG_DEC_CNT__DECODE_COUNT__SHIFT 0x0 +#define UVD_JPEG_DEC_CNT__DECODE_COUNT_MASK 0xFFFFFFFFL +//UVD_JPEG_SPS_INFO +#define UVD_JPEG_SPS_INFO__PIC_WIDTH__SHIFT 0x0 +#define UVD_JPEG_SPS_INFO__PIC_HEIGHT__SHIFT 0x10 +#define UVD_JPEG_SPS_INFO__PIC_WIDTH_MASK 0x0000FFFFL +#define UVD_JPEG_SPS_INFO__PIC_HEIGHT_MASK 0xFFFF0000L +//UVD_JPEG_SPS1_INFO +#define UVD_JPEG_SPS1_INFO__CHROMA_FORMAT_IDC__SHIFT 0x0 +#define UVD_JPEG_SPS1_INFO__YUV422_SUBFORMAT__SHIFT 0x3 +#define UVD_JPEG_SPS1_INFO__OUT_FMT_422__SHIFT 0x4 +#define UVD_JPEG_SPS1_INFO__CHROMA_FORMAT_IDC_MASK 0x00000007L +#define UVD_JPEG_SPS1_INFO__YUV422_SUBFORMAT_MASK 0x00000008L +#define UVD_JPEG_SPS1_INFO__OUT_FMT_422_MASK 0x00000010L +//UVD_JPEG_RE_TIMER +#define UVD_JPEG_RE_TIMER__TIMER_OUT__SHIFT 0x0 +#define UVD_JPEG_RE_TIMER__TIMER_OUT_EN__SHIFT 0x10 +#define UVD_JPEG_RE_TIMER__TIMER_OUT_MASK 0x000000FFL +#define UVD_JPEG_RE_TIMER__TIMER_OUT_EN_MASK 0x00010000L +//UVD_JPEG_DEC_SCRATCH0 +#define UVD_JPEG_DEC_SCRATCH0__SCRATCH0__SHIFT 0x0 +#define UVD_JPEG_DEC_SCRATCH0__SCRATCH0_MASK 0xFFFFFFFFL +//UVD_JPEG_INT_EN +#define UVD_JPEG_INT_EN__OUTBUF_WPTR_INC_EN__SHIFT 0x0 +#define UVD_JPEG_INT_EN__JOB_AVAIL_EN__SHIFT 0x1 +#define UVD_JPEG_INT_EN__FENCE_VAL_EN__SHIFT 0x2 +#define UVD_JPEG_INT_EN__FIFO_OVERFLOW_ERR_EN__SHIFT 0x6 +#define UVD_JPEG_INT_EN__BLK_CNT_OUT_OF_SYNC_ERR_EN__SHIFT 0x7 +#define UVD_JPEG_INT_EN__EOI_ERR_EN__SHIFT 0x8 +#define UVD_JPEG_INT_EN__HFM_ERR_EN__SHIFT 0x9 +#define UVD_JPEG_INT_EN__RST_ERR_EN__SHIFT 0xa +#define UVD_JPEG_INT_EN__ECS_MK_ERR_EN__SHIFT 0xb +#define UVD_JPEG_INT_EN__TIMEOUT_ERR_EN__SHIFT 0xc +#define UVD_JPEG_INT_EN__MARKER_ERR_EN__SHIFT 0xd +#define UVD_JPEG_INT_EN__FMT_ERR_EN__SHIFT 0xe +#define UVD_JPEG_INT_EN__PROFILE_ERR_EN__SHIFT 0xf +#define UVD_JPEG_INT_EN__FC_TIMEOUT_ERR_EN__SHIFT 0x10 +#define UVD_JPEG_INT_EN__FC_FMT_ERR_EN__SHIFT 0x11 +#define UVD_JPEG_INT_EN__FC_SRC_ERR_EN__SHIFT 0x12 +#define UVD_JPEG_INT_EN__CROP_SIZE_ERR_EN__SHIFT 0x13 +#define UVD_JPEG_INT_EN__OUTBUF_WPTR_INC_EN_MASK 0x00000001L +#define UVD_JPEG_INT_EN__JOB_AVAIL_EN_MASK 0x00000002L +#define UVD_JPEG_INT_EN__FENCE_VAL_EN_MASK 0x00000004L +#define UVD_JPEG_INT_EN__FIFO_OVERFLOW_ERR_EN_MASK 0x00000040L +#define UVD_JPEG_INT_EN__BLK_CNT_OUT_OF_SYNC_ERR_EN_MASK 0x00000080L +#define UVD_JPEG_INT_EN__EOI_ERR_EN_MASK 0x00000100L +#define UVD_JPEG_INT_EN__HFM_ERR_EN_MASK 0x00000200L +#define UVD_JPEG_INT_EN__RST_ERR_EN_MASK 0x00000400L +#define UVD_JPEG_INT_EN__ECS_MK_ERR_EN_MASK 0x00000800L +#define UVD_JPEG_INT_EN__TIMEOUT_ERR_EN_MASK 0x00001000L +#define UVD_JPEG_INT_EN__MARKER_ERR_EN_MASK 0x00002000L +#define UVD_JPEG_INT_EN__FMT_ERR_EN_MASK 0x00004000L +#define UVD_JPEG_INT_EN__PROFILE_ERR_EN_MASK 0x00008000L +#define UVD_JPEG_INT_EN__FC_TIMEOUT_ERR_EN_MASK 0x00010000L +#define UVD_JPEG_INT_EN__FC_FMT_ERR_EN_MASK 0x00020000L +#define UVD_JPEG_INT_EN__FC_SRC_ERR_EN_MASK 0x00040000L +#define UVD_JPEG_INT_EN__CROP_SIZE_ERR_EN_MASK 0x00080000L +//UVD_JPEG_INT_STAT +#define UVD_JPEG_INT_STAT__OUTBUF_WPTR_INC_INT__SHIFT 0x0 +#define UVD_JPEG_INT_STAT__JOB_AVAIL_INT__SHIFT 0x1 +#define UVD_JPEG_INT_STAT__FENCE_VAL_INT__SHIFT 0x2 +#define UVD_JPEG_INT_STAT__FIFO_OVERFLOW_ERR_INT__SHIFT 0x6 +#define UVD_JPEG_INT_STAT__BLK_CNT_OUT_OF_SYNC_ERR_INT__SHIFT 0x7 +#define UVD_JPEG_INT_STAT__EOI_ERR_INT__SHIFT 0x8 +#define UVD_JPEG_INT_STAT__HFM_ERR_INT__SHIFT 0x9 +#define UVD_JPEG_INT_STAT__RST_ERR_INT__SHIFT 0xa +#define UVD_JPEG_INT_STAT__ECS_MK_ERR_INT__SHIFT 0xb +#define UVD_JPEG_INT_STAT__TIMEOUT_ERR_INT__SHIFT 0xc +#define UVD_JPEG_INT_STAT__MARKER_ERR_INT__SHIFT 0xd +#define UVD_JPEG_INT_STAT__FMT_ERR_INT__SHIFT 0xe +#define UVD_JPEG_INT_STAT__PROFILE_ERR_INT__SHIFT 0xf +#define UVD_JPEG_INT_STAT__FC_TIMEOUT_ERR_INT__SHIFT 0x10 +#define UVD_JPEG_INT_STAT__FC_FMT_ERR_INT__SHIFT 0x11 +#define UVD_JPEG_INT_STAT__FC_SRC_ERR_INT__SHIFT 0x12 +#define UVD_JPEG_INT_STAT__CROP_SIZE_ERR_INT__SHIFT 0x13 +#define UVD_JPEG_INT_STAT__OUTBUF_WPTR_INC_INT_MASK 0x00000001L +#define UVD_JPEG_INT_STAT__JOB_AVAIL_INT_MASK 0x00000002L +#define UVD_JPEG_INT_STAT__FENCE_VAL_INT_MASK 0x00000004L +#define UVD_JPEG_INT_STAT__FIFO_OVERFLOW_ERR_INT_MASK 0x00000040L +#define UVD_JPEG_INT_STAT__BLK_CNT_OUT_OF_SYNC_ERR_INT_MASK 0x00000080L +#define UVD_JPEG_INT_STAT__EOI_ERR_INT_MASK 0x00000100L +#define UVD_JPEG_INT_STAT__HFM_ERR_INT_MASK 0x00000200L +#define UVD_JPEG_INT_STAT__RST_ERR_INT_MASK 0x00000400L +#define UVD_JPEG_INT_STAT__ECS_MK_ERR_INT_MASK 0x00000800L +#define UVD_JPEG_INT_STAT__TIMEOUT_ERR_INT_MASK 0x00001000L +#define UVD_JPEG_INT_STAT__MARKER_ERR_INT_MASK 0x00002000L +#define UVD_JPEG_INT_STAT__FMT_ERR_INT_MASK 0x00004000L +#define UVD_JPEG_INT_STAT__PROFILE_ERR_INT_MASK 0x00008000L +#define UVD_JPEG_INT_STAT__FC_TIMEOUT_ERR_INT_MASK 0x00010000L +#define UVD_JPEG_INT_STAT__FC_FMT_ERR_INT_MASK 0x00020000L +#define UVD_JPEG_INT_STAT__FC_SRC_ERR_INT_MASK 0x00040000L +#define UVD_JPEG_INT_STAT__CROP_SIZE_ERR_INT_MASK 0x00080000L +//UVD_JPEG_TIER_CNTL0 +#define UVD_JPEG_TIER_CNTL0__TIER_SEL__SHIFT 0x0 +#define UVD_JPEG_TIER_CNTL0__Y_COMP_ID__SHIFT 0x2 +#define UVD_JPEG_TIER_CNTL0__U_COMP_ID__SHIFT 0x4 +#define UVD_JPEG_TIER_CNTL0__V_COMP_ID__SHIFT 0x6 +#define UVD_JPEG_TIER_CNTL0__Y_H_SAMP_FAC__SHIFT 0x8 +#define UVD_JPEG_TIER_CNTL0__Y_V_SAMP_FAC__SHIFT 0xb +#define UVD_JPEG_TIER_CNTL0__U_H_SAMP_FAC__SHIFT 0xe +#define UVD_JPEG_TIER_CNTL0__U_V_SAMP_FAC__SHIFT 0x11 +#define UVD_JPEG_TIER_CNTL0__V_H_SAMP_FAC__SHIFT 0x14 +#define UVD_JPEG_TIER_CNTL0__V_V_SAMP_FAC__SHIFT 0x17 +#define UVD_JPEG_TIER_CNTL0__Y_TQ__SHIFT 0x1a +#define UVD_JPEG_TIER_CNTL0__U_TQ__SHIFT 0x1c +#define UVD_JPEG_TIER_CNTL0__V_TQ__SHIFT 0x1e +#define UVD_JPEG_TIER_CNTL0__TIER_SEL_MASK 0x00000003L +#define UVD_JPEG_TIER_CNTL0__Y_COMP_ID_MASK 0x0000000CL +#define UVD_JPEG_TIER_CNTL0__U_COMP_ID_MASK 0x00000030L +#define UVD_JPEG_TIER_CNTL0__V_COMP_ID_MASK 0x000000C0L +#define UVD_JPEG_TIER_CNTL0__Y_H_SAMP_FAC_MASK 0x00000700L +#define UVD_JPEG_TIER_CNTL0__Y_V_SAMP_FAC_MASK 0x00003800L +#define UVD_JPEG_TIER_CNTL0__U_H_SAMP_FAC_MASK 0x0001C000L +#define UVD_JPEG_TIER_CNTL0__U_V_SAMP_FAC_MASK 0x000E0000L +#define UVD_JPEG_TIER_CNTL0__V_H_SAMP_FAC_MASK 0x00700000L +#define UVD_JPEG_TIER_CNTL0__V_V_SAMP_FAC_MASK 0x03800000L +#define UVD_JPEG_TIER_CNTL0__Y_TQ_MASK 0x0C000000L +#define UVD_JPEG_TIER_CNTL0__U_TQ_MASK 0x30000000L +#define UVD_JPEG_TIER_CNTL0__V_TQ_MASK 0xC0000000L +//UVD_JPEG_TIER_CNTL1 +#define UVD_JPEG_TIER_CNTL1__SRC_WIDTH__SHIFT 0x0 +#define UVD_JPEG_TIER_CNTL1__SRC_HEIGHT__SHIFT 0x10 +#define UVD_JPEG_TIER_CNTL1__SRC_WIDTH_MASK 0x0000FFFFL +#define UVD_JPEG_TIER_CNTL1__SRC_HEIGHT_MASK 0xFFFF0000L +//UVD_JPEG_TIER_CNTL2 +#define UVD_JPEG_TIER_CNTL2__TBL_ECS_SEL__SHIFT 0x0 +#define UVD_JPEG_TIER_CNTL2__TBL_TYPE__SHIFT 0x1 +#define UVD_JPEG_TIER_CNTL2__TQ__SHIFT 0x2 +#define UVD_JPEG_TIER_CNTL2__TH__SHIFT 0x4 +#define UVD_JPEG_TIER_CNTL2__TC__SHIFT 0x6 +#define UVD_JPEG_TIER_CNTL2__TD__SHIFT 0x7 +#define UVD_JPEG_TIER_CNTL2__TA__SHIFT 0xa +#define UVD_JPEG_TIER_CNTL2__TIER2_HTBL_CNTLEN__SHIFT 0xe +#define UVD_JPEG_TIER_CNTL2__DRI_VAL__SHIFT 0x10 +#define UVD_JPEG_TIER_CNTL2__TBL_ECS_SEL_MASK 0x00000001L +#define UVD_JPEG_TIER_CNTL2__TBL_TYPE_MASK 0x00000002L +#define UVD_JPEG_TIER_CNTL2__TQ_MASK 0x0000000CL +#define UVD_JPEG_TIER_CNTL2__TH_MASK 0x00000030L +#define UVD_JPEG_TIER_CNTL2__TC_MASK 0x00000040L +#define UVD_JPEG_TIER_CNTL2__TD_MASK 0x00000380L +#define UVD_JPEG_TIER_CNTL2__TA_MASK 0x00001C00L +#define UVD_JPEG_TIER_CNTL2__TIER2_HTBL_CNTLEN_MASK 0x00004000L +#define UVD_JPEG_TIER_CNTL2__DRI_VAL_MASK 0xFFFF0000L +//UVD_JPEG_TIER_STATUS +#define UVD_JPEG_TIER_STATUS__BSI_FETCH_DONE__SHIFT 0x0 +#define UVD_JPEG_TIER_STATUS__DECODE_DONE__SHIFT 0x1 +#define UVD_JPEG_TIER_STATUS__BSI_FETCH_DONE_MASK 0x00000001L +#define UVD_JPEG_TIER_STATUS__DECODE_DONE_MASK 0x00000002L + + +// addressBlock: aid_uvd0_uvd_jpeg_sclk0_jpegnpsclkdec +//UVD_JPEG_OUTBUF_CNTL +#define UVD_JPEG_OUTBUF_CNTL__OUTBUF_CNT__SHIFT 0x0 +#define UVD_JPEG_OUTBUF_CNTL__HGT_ALIGN__SHIFT 0x2 +#define UVD_JPEG_OUTBUF_CNTL__JPEG0_DECODE_DONE_FIX__SHIFT 0x6 +#define UVD_JPEG_OUTBUF_CNTL__JPEG0_WR_COMB_MAX_CNT__SHIFT 0x7 +#define UVD_JPEG_OUTBUF_CNTL__JPEG0_WR_COMB_TIMER__SHIFT 0x9 +#define UVD_JPEG_OUTBUF_CNTL__DIS_OBUF_AVAIL_CHECK__SHIFT 0x10 +#define UVD_JPEG_OUTBUF_CNTL__OUTBUF_CNT_MASK 0x00000003L +#define UVD_JPEG_OUTBUF_CNTL__HGT_ALIGN_MASK 0x00000004L +#define UVD_JPEG_OUTBUF_CNTL__JPEG0_DECODE_DONE_FIX_MASK 0x00000040L +#define UVD_JPEG_OUTBUF_CNTL__JPEG0_WR_COMB_MAX_CNT_MASK 0x00000180L +#define UVD_JPEG_OUTBUF_CNTL__JPEG0_WR_COMB_TIMER_MASK 0x00001E00L +#define UVD_JPEG_OUTBUF_CNTL__DIS_OBUF_AVAIL_CHECK_MASK 0x00010000L +//UVD_JPEG_OUTBUF_WPTR +#define UVD_JPEG_OUTBUF_WPTR__OUTBUF_WPTR__SHIFT 0x0 +#define UVD_JPEG_OUTBUF_WPTR__OUTBUF_WPTR_MASK 0xFFFFFFFFL +//UVD_JPEG_OUTBUF_RPTR +#define UVD_JPEG_OUTBUF_RPTR__OUTBUF_RPTR__SHIFT 0x0 +#define UVD_JPEG_OUTBUF_RPTR__OUTBUF_RPTR_MASK 0xFFFFFFFFL +//UVD_JPEG_PITCH +#define UVD_JPEG_PITCH__PITCH__SHIFT 0x0 +#define UVD_JPEG_PITCH__PITCH_MASK 0xFFFFFFFFL +//UVD_JPEG_UV_PITCH +#define UVD_JPEG_UV_PITCH__UV_PITCH__SHIFT 0x0 +#define UVD_JPEG_UV_PITCH__UV_PITCH_MASK 0xFFFFFFFFL +//JPEG_DEC_Y_GFX8_TILING_SURFACE +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__BANK_WIDTH__SHIFT 0x0 +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__BANK_HEIGHT__SHIFT 0x2 +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__MACRO_TILE_ASPECT__SHIFT 0x4 +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__NUM_BANKS__SHIFT 0x6 +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__PIPE_CONFIG__SHIFT 0x8 +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__TILE_SPLIT__SHIFT 0xd +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__ARRAY_MODE__SHIFT 0x10 +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__BANK_WIDTH_MASK 0x00000003L +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__BANK_HEIGHT_MASK 0x0000000CL +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__MACRO_TILE_ASPECT_MASK 0x00000030L +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__NUM_BANKS_MASK 0x000000C0L +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__PIPE_CONFIG_MASK 0x00001F00L +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__TILE_SPLIT_MASK 0x0000E000L +#define JPEG_DEC_Y_GFX8_TILING_SURFACE__ARRAY_MODE_MASK 0x000F0000L +//JPEG_DEC_UV_GFX8_TILING_SURFACE +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__BANK_WIDTH__SHIFT 0x0 +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__BANK_HEIGHT__SHIFT 0x2 +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__MACRO_TILE_ASPECT__SHIFT 0x4 +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__NUM_BANKS__SHIFT 0x6 +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__PIPE_CONFIG__SHIFT 0x8 +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__TILE_SPLIT__SHIFT 0xd +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__ARRAY_MODE__SHIFT 0x10 +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__BANK_WIDTH_MASK 0x00000003L +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__BANK_HEIGHT_MASK 0x0000000CL +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__MACRO_TILE_ASPECT_MASK 0x00000030L +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__NUM_BANKS_MASK 0x000000C0L +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__PIPE_CONFIG_MASK 0x00001F00L +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__TILE_SPLIT_MASK 0x0000E000L +#define JPEG_DEC_UV_GFX8_TILING_SURFACE__ARRAY_MODE_MASK 0x000F0000L +//JPEG_DEC_GFX8_ADDR_CONFIG +#define JPEG_DEC_GFX8_ADDR_CONFIG__PIPE_INTERLEAVE_SIZE__SHIFT 0x4 +#define JPEG_DEC_GFX8_ADDR_CONFIG__PIPE_INTERLEAVE_SIZE_MASK 0x00000070L +//JPEG_DEC_Y_GFX10_TILING_SURFACE +#define JPEG_DEC_Y_GFX10_TILING_SURFACE__SWIZZLE_MODE__SHIFT 0x0 +#define JPEG_DEC_Y_GFX10_TILING_SURFACE__SWIZZLE_MODE_MASK 0x0000001FL +//JPEG_DEC_UV_GFX10_TILING_SURFACE +#define JPEG_DEC_UV_GFX10_TILING_SURFACE__SWIZZLE_MODE__SHIFT 0x0 +#define JPEG_DEC_UV_GFX10_TILING_SURFACE__SWIZZLE_MODE_MASK 0x0000001FL +//JPEG_DEC_GFX10_ADDR_CONFIG +#define JPEG_DEC_GFX10_ADDR_CONFIG__NUM_PIPES__SHIFT 0x0 +#define JPEG_DEC_GFX10_ADDR_CONFIG__PIPE_INTERLEAVE_SIZE__SHIFT 0x3 +#define JPEG_DEC_GFX10_ADDR_CONFIG__NUM_PKRS__SHIFT 0x8 +#define JPEG_DEC_GFX10_ADDR_CONFIG__NUM_BANKS__SHIFT 0xc +#define JPEG_DEC_GFX10_ADDR_CONFIG__NUM_SHADER_ENGINES__SHIFT 0x13 +#define JPEG_DEC_GFX10_ADDR_CONFIG__NUM_PIPES_MASK 0x00000007L +#define JPEG_DEC_GFX10_ADDR_CONFIG__PIPE_INTERLEAVE_SIZE_MASK 0x00000038L +#define JPEG_DEC_GFX10_ADDR_CONFIG__NUM_PKRS_MASK 0x00000700L +#define JPEG_DEC_GFX10_ADDR_CONFIG__NUM_BANKS_MASK 0x00007000L +#define JPEG_DEC_GFX10_ADDR_CONFIG__NUM_SHADER_ENGINES_MASK 0x00180000L +//JPEG_DEC_ADDR_MODE +#define JPEG_DEC_ADDR_MODE__ADDR_MODE_Y__SHIFT 0x0 +#define JPEG_DEC_ADDR_MODE__ADDR_MODE_UV__SHIFT 0x2 +#define JPEG_DEC_ADDR_MODE__ADDR_LIB_SEL__SHIFT 0xc +#define JPEG_DEC_ADDR_MODE__ADDR_MODE_Y_MASK 0x00000003L +#define JPEG_DEC_ADDR_MODE__ADDR_MODE_UV_MASK 0x0000000CL +#define JPEG_DEC_ADDR_MODE__ADDR_LIB_SEL_MASK 0x00007000L +//UVD_JPEG_OUTPUT_XY +#define UVD_JPEG_OUTPUT_XY__OUTPUT_X__SHIFT 0x0 +#define UVD_JPEG_OUTPUT_XY__OUTPUT_Y__SHIFT 0x10 +#define UVD_JPEG_OUTPUT_XY__OUTPUT_X_MASK 0x00003FFFL +#define UVD_JPEG_OUTPUT_XY__OUTPUT_Y_MASK 0x3FFF0000L +//UVD_JPEG_GPCOM_CMD +#define UVD_JPEG_GPCOM_CMD__CMD__SHIFT 0x1 +#define UVD_JPEG_GPCOM_CMD__CMD_MASK 0x0000000EL +//UVD_JPEG_GPCOM_DATA0 +#define UVD_JPEG_GPCOM_DATA0__DATA0__SHIFT 0x0 +#define UVD_JPEG_GPCOM_DATA0__DATA0_MASK 0xFFFFFFFFL +//UVD_JPEG_GPCOM_DATA1 +#define UVD_JPEG_GPCOM_DATA1__DATA1__SHIFT 0x0 +#define UVD_JPEG_GPCOM_DATA1__DATA1_MASK 0xFFFFFFFFL +//UVD_JPEG_SCRATCH1 +#define UVD_JPEG_SCRATCH1__SCRATCH1__SHIFT 0x0 +#define UVD_JPEG_SCRATCH1__SCRATCH1_MASK 0xFFFFFFFFL +//UVD_JPEG_DEC_SOFT_RST +#define UVD_JPEG_DEC_SOFT_RST__SOFT_RESET__SHIFT 0x0 +#define UVD_JPEG_DEC_SOFT_RST__RESET_STATUS__SHIFT 0x10 +#define UVD_JPEG_DEC_SOFT_RST__SOFT_RESET_MASK 0x00000001L +#define UVD_JPEG_DEC_SOFT_RST__RESET_STATUS_MASK 0x00010000L + +// addressBlock: aid_uvd0_vcn_edcc_dec +//VCN_UE_ERR_STATUS_LO_VIDD +#define VCN_UE_ERR_STATUS_LO_VIDD__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_VIDD__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_VIDD__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_VIDD__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_VIDD__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_VIDD__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_VIDD__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_VIDD__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_VIDD +#define VCN_UE_ERR_STATUS_HI_VIDD__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_VIDD__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_VIDD__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_VIDD__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_VIDD__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_VIDD__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_VIDD__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_VIDD__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_VIDD__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_VIDD__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_VIDD__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_VIDD__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_VIDD__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_VIDD__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_VIDD__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_VIDD__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_VIDV +#define VCN_UE_ERR_STATUS_LO_VIDV__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_VIDV__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_VIDV__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_VIDV__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_VIDV__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_VIDV__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_VIDV__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_VIDV__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_VIDV +#define VCN_UE_ERR_STATUS_HI_VIDV__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_VIDV__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_VIDV__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_VIDV__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_VIDV__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_VIDV__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_VIDV__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_VIDV__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_VIDV__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_VIDV__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_VIDV__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_VIDV__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_VIDV__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_VIDV__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_VIDV__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_VIDV__Err_clr_MASK 0x80000000L +//VCN_CE_ERR_STATUS_LO_MMSCHD +#define VCN_CE_ERR_STATUS_LO_MMSCHD__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_CE_ERR_STATUS_LO_MMSCHD__Address_Valid_Flag__SHIFT 0x1 +#define VCN_CE_ERR_STATUS_LO_MMSCHD__Address__SHIFT 0x2 +#define VCN_CE_ERR_STATUS_LO_MMSCHD__Memory_id__SHIFT 0x18 +#define VCN_CE_ERR_STATUS_LO_MMSCHD__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_CE_ERR_STATUS_LO_MMSCHD__Address_Valid_Flag_MASK 0x00000002L +#define VCN_CE_ERR_STATUS_LO_MMSCHD__Address_MASK 0x00FFFFFCL +#define VCN_CE_ERR_STATUS_LO_MMSCHD__Memory_id_MASK 0xFF000000L +//VCN_CE_ERR_STATUS_HI_MMSCHD +#define VCN_CE_ERR_STATUS_HI_MMSCHD__ECC__SHIFT 0x0 +#define VCN_CE_ERR_STATUS_HI_MMSCHD__Other__SHIFT 0x1 +#define VCN_CE_ERR_STATUS_HI_MMSCHD__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_CE_ERR_STATUS_HI_MMSCHD__Err_Info__SHIFT 0x3 +#define VCN_CE_ERR_STATUS_HI_MMSCHD__CE_Cnt__SHIFT 0x17 +#define VCN_CE_ERR_STATUS_HI_MMSCHD__Poison__SHIFT 0x1c +#define VCN_CE_ERR_STATUS_HI_MMSCHD__RESERVED__SHIFT 0x1d +#define VCN_CE_ERR_STATUS_HI_MMSCHD__Err_clr__SHIFT 0x1f +#define VCN_CE_ERR_STATUS_HI_MMSCHD__ECC_MASK 0x00000001L +#define VCN_CE_ERR_STATUS_HI_MMSCHD__Other_MASK 0x00000002L +#define VCN_CE_ERR_STATUS_HI_MMSCHD__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_CE_ERR_STATUS_HI_MMSCHD__Err_Info_MASK 0x007FFFF8L +#define VCN_CE_ERR_STATUS_HI_MMSCHD__CE_Cnt_MASK 0x03800000L +#define VCN_CE_ERR_STATUS_HI_MMSCHD__Poison_MASK 0x10000000L +#define VCN_CE_ERR_STATUS_HI_MMSCHD__RESERVED_MASK 0x60000000L +#define VCN_CE_ERR_STATUS_HI_MMSCHD__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG0S +#define VCN_UE_ERR_STATUS_LO_JPEG0S__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG0S__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG0S__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG0S__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG0S__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG0S__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG0S__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG0S__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG0S +#define VCN_UE_ERR_STATUS_HI_JPEG0S__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG0S__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG0S__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG0S__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG0S__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG0S__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG0S__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG0S__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG0S__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG0S__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG0S__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG0S__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG0S__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG0S__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG0S__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG0S__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG0D +#define VCN_UE_ERR_STATUS_LO_JPEG0D__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG0D__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG0D__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG0D__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG0D__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG0D__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG0D__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG0D__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG0D +#define VCN_UE_ERR_STATUS_HI_JPEG0D__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG0D__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG0D__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG0D__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG0D__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG0D__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG0D__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG0D__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG0D__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG0D__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG0D__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG0D__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG0D__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG0D__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG0D__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG0D__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG1S +#define VCN_UE_ERR_STATUS_LO_JPEG1S__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG1S__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG1S__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG1S__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG1S__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG1S__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG1S__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG1S__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG1S +#define VCN_UE_ERR_STATUS_HI_JPEG1S__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG1S__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG1S__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG1S__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG1S__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG1S__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG1S__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG1S__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG1S__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG1S__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG1S__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG1S__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG1S__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG1S__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG1S__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG1S__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG1D +#define VCN_UE_ERR_STATUS_LO_JPEG1D__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG1D__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG1D__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG1D__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG1D__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG1D__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG1D__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG1D__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG1D +#define VCN_UE_ERR_STATUS_HI_JPEG1D__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG1D__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG1D__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG1D__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG1D__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG1D__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG1D__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG1D__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG1D__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG1D__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG1D__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG1D__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG1D__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG1D__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG1D__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG1D__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG2S +#define VCN_UE_ERR_STATUS_LO_JPEG2S__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG2S__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG2S__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG2S__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG2S__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG2S__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG2S__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG2S__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG2S +#define VCN_UE_ERR_STATUS_HI_JPEG2S__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG2S__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG2S__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG2S__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG2S__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG2S__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG2S__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG2S__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG2S__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG2S__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG2S__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG2S__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG2S__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG2S__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG2S__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG2S__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG2D +#define VCN_UE_ERR_STATUS_LO_JPEG2D__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG2D__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG2D__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG2D__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG2D__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG2D__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG2D__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG2D__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG2D +#define VCN_UE_ERR_STATUS_HI_JPEG2D__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG2D__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG2D__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG2D__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG2D__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG2D__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG2D__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG2D__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG2D__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG2D__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG2D__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG2D__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG2D__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG2D__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG2D__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG2D__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG3S +#define VCN_UE_ERR_STATUS_LO_JPEG3S__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG3S__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG3S__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG3S__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG3S__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG3S__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG3S__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG3S__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG3S +#define VCN_UE_ERR_STATUS_HI_JPEG3S__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG3S__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG3S__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG3S__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG3S__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG3S__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG3S__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG3S__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG3S__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG3S__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG3S__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG3S__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG3S__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG3S__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG3S__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG3S__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG3D +#define VCN_UE_ERR_STATUS_LO_JPEG3D__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG3D__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG3D__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG3D__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG3D__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG3D__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG3D__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG3D__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG3D +#define VCN_UE_ERR_STATUS_HI_JPEG3D__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG3D__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG3D__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG3D__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG3D__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG3D__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG3D__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG3D__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG3D__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG3D__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG3D__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG3D__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG3D__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG3D__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG3D__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG3D__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG4S +#define VCN_UE_ERR_STATUS_LO_JPEG4S__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG4S__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG4S__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG4S__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG4S__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG4S__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG4S__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG4S__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG4S +#define VCN_UE_ERR_STATUS_HI_JPEG4S__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG4S__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG4S__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG4S__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG4S__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG4S__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG4S__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG4S__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG4S__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG4S__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG4S__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG4S__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG4S__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG4S__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG4S__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG4S__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG4D +#define VCN_UE_ERR_STATUS_LO_JPEG4D__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG4D__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG4D__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG4D__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG4D__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG4D__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG4D__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG4D__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG4D +#define VCN_UE_ERR_STATUS_HI_JPEG4D__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG4D__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG4D__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG4D__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG4D__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG4D__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG4D__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG4D__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG4D__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG4D__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG4D__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG4D__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG4D__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG4D__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG4D__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG4D__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG5S +#define VCN_UE_ERR_STATUS_LO_JPEG5S__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG5S__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG5S__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG5S__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG5S__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG5S__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG5S__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG5S__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG5S +#define VCN_UE_ERR_STATUS_HI_JPEG5S__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG5S__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG5S__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG5S__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG5S__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG5S__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG5S__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG5S__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG5S__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG5S__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG5S__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG5S__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG5S__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG5S__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG5S__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG5S__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG5D +#define VCN_UE_ERR_STATUS_LO_JPEG5D__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG5D__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG5D__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG5D__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG5D__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG5D__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG5D__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG5D__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG5D +#define VCN_UE_ERR_STATUS_HI_JPEG5D__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG5D__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG5D__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG5D__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG5D__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG5D__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG5D__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG5D__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG5D__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG5D__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG5D__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG5D__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG5D__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG5D__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG5D__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG5D__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG6S +#define VCN_UE_ERR_STATUS_LO_JPEG6S__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG6S__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG6S__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG6S__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG6S__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG6S__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG6S__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG6S__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG6S +#define VCN_UE_ERR_STATUS_HI_JPEG6S__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG6S__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG6S__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG6S__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG6S__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG6S__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG6S__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG6S__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG6S__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG6S__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG6S__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG6S__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG6S__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG6S__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG6S__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG6S__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG6D +#define VCN_UE_ERR_STATUS_LO_JPEG6D__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG6D__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG6D__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG6D__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG6D__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG6D__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG6D__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG6D__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG6D +#define VCN_UE_ERR_STATUS_HI_JPEG6D__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG6D__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG6D__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG6D__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG6D__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG6D__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG6D__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG6D__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG6D__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG6D__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG6D__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG6D__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG6D__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG6D__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG6D__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG6D__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG7S +#define VCN_UE_ERR_STATUS_LO_JPEG7S__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG7S__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG7S__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG7S__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG7S__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG7S__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG7S__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG7S__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG7S +#define VCN_UE_ERR_STATUS_HI_JPEG7S__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG7S__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG7S__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG7S__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG7S__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG7S__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG7S__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG7S__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG7S__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG7S__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG7S__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG7S__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG7S__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG7S__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG7S__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG7S__Err_clr_MASK 0x80000000L +//VCN_UE_ERR_STATUS_LO_JPEG7D +#define VCN_UE_ERR_STATUS_LO_JPEG7D__Err_Status_Valid_Flag__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_LO_JPEG7D__Address_Valid_Flag__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_LO_JPEG7D__Address__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_LO_JPEG7D__Memory_id__SHIFT 0x18 +#define VCN_UE_ERR_STATUS_LO_JPEG7D__Err_Status_Valid_Flag_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_LO_JPEG7D__Address_Valid_Flag_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_LO_JPEG7D__Address_MASK 0x00FFFFFCL +#define VCN_UE_ERR_STATUS_LO_JPEG7D__Memory_id_MASK 0xFF000000L +//VCN_UE_ERR_STATUS_HI_JPEG7D +#define VCN_UE_ERR_STATUS_HI_JPEG7D__ECC__SHIFT 0x0 +#define VCN_UE_ERR_STATUS_HI_JPEG7D__Parity__SHIFT 0x1 +#define VCN_UE_ERR_STATUS_HI_JPEG7D__Err_Info_Valid_Flag__SHIFT 0x2 +#define VCN_UE_ERR_STATUS_HI_JPEG7D__Err_Info__SHIFT 0x3 +#define VCN_UE_ERR_STATUS_HI_JPEG7D__UE_Cnt__SHIFT 0x17 +#define VCN_UE_ERR_STATUS_HI_JPEG7D__FED_Cnt__SHIFT 0x1a +#define VCN_UE_ERR_STATUS_HI_JPEG7D__RESERVED__SHIFT 0x1d +#define VCN_UE_ERR_STATUS_HI_JPEG7D__Err_clr__SHIFT 0x1f +#define VCN_UE_ERR_STATUS_HI_JPEG7D__ECC_MASK 0x00000001L +#define VCN_UE_ERR_STATUS_HI_JPEG7D__Parity_MASK 0x00000002L +#define VCN_UE_ERR_STATUS_HI_JPEG7D__Err_Info_Valid_Flag_MASK 0x00000004L +#define VCN_UE_ERR_STATUS_HI_JPEG7D__Err_Info_MASK 0x007FFFF8L +#define VCN_UE_ERR_STATUS_HI_JPEG7D__UE_Cnt_MASK 0x03800000L +#define VCN_UE_ERR_STATUS_HI_JPEG7D__FED_Cnt_MASK 0x1C000000L +#define VCN_UE_ERR_STATUS_HI_JPEG7D__RESERVED_MASK 0x60000000L +#define VCN_UE_ERR_STATUS_HI_JPEG7D__Err_clr_MASK 0x80000000L + +// addressBlock: aid_uvd0_uvd_jrbc0_uvd_jrbc_dec +//UVD_JRBC0_UVD_JRBC_RB_WPTR +#define UVD_JRBC0_UVD_JRBC_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_JRBC0_UVD_JRBC_RB_WPTR__RB_WPTR_MASK 0x007FFFF0L +//UVD_JRBC0_UVD_JRBC_RB_CNTL +#define UVD_JRBC0_UVD_JRBC_RB_CNTL__RB_NO_FETCH__SHIFT 0x0 +#define UVD_JRBC0_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN__SHIFT 0x1 +#define UVD_JRBC0_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER__SHIFT 0x4 +#define UVD_JRBC0_UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK 0x00000001L +#define UVD_JRBC0_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK 0x00000002L +#define UVD_JRBC0_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER_MASK 0x0007FFF0L +//UVD_JRBC0_UVD_JRBC_IB_SIZE +#define UVD_JRBC0_UVD_JRBC_IB_SIZE__IB_SIZE__SHIFT 0x4 +#define UVD_JRBC0_UVD_JRBC_IB_SIZE__IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC0_UVD_JRBC_URGENT_CNTL +#define UVD_JRBC0_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK__SHIFT 0x0 +#define UVD_JRBC0_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK_MASK 0x00000003L +//UVD_JRBC0_UVD_JRBC_RB_REF_DATA +#define UVD_JRBC0_UVD_JRBC_RB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC0_UVD_JRBC_RB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC0_UVD_JRBC_RB_COND_RD_TIMER +#define UVD_JRBC0_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC0_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC0_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC0_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC0_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC0_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC0_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC0_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC0_UVD_JRBC_SOFT_RESET +#define UVD_JRBC0_UVD_JRBC_SOFT_RESET__RESET__SHIFT 0x0 +#define UVD_JRBC0_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS__SHIFT 0x11 +#define UVD_JRBC0_UVD_JRBC_SOFT_RESET__RESET_MASK 0x00000001L +#define UVD_JRBC0_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS_MASK 0x00020000L +//UVD_JRBC0_UVD_JRBC_STATUS +#define UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE__SHIFT 0x0 +#define UVD_JRBC0_UVD_JRBC_STATUS__IB_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC0_UVD_JRBC_STATUS__RB_ILLEGAL_CMD__SHIFT 0x2 +#define UVD_JRBC0_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT__SHIFT 0x3 +#define UVD_JRBC0_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT__SHIFT 0x4 +#define UVD_JRBC0_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT__SHIFT 0x5 +#define UVD_JRBC0_UVD_JRBC_STATUS__IB_ILLEGAL_CMD__SHIFT 0x6 +#define UVD_JRBC0_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT__SHIFT 0x7 +#define UVD_JRBC0_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT__SHIFT 0x8 +#define UVD_JRBC0_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT__SHIFT 0x9 +#define UVD_JRBC0_UVD_JRBC_STATUS__RB_TRAP_STATUS__SHIFT 0xa +#define UVD_JRBC0_UVD_JRBC_STATUS__PREEMPT_STATUS__SHIFT 0xb +#define UVD_JRBC0_UVD_JRBC_STATUS__IB_TRAP_STATUS__SHIFT 0xc +#define UVD_JRBC0_UVD_JRBC_STATUS__INT_EN__SHIFT 0x10 +#define UVD_JRBC0_UVD_JRBC_STATUS__INT_ACK__SHIFT 0x11 +#define UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK 0x00000001L +#define UVD_JRBC0_UVD_JRBC_STATUS__IB_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC0_UVD_JRBC_STATUS__RB_ILLEGAL_CMD_MASK 0x00000004L +#define UVD_JRBC0_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT_MASK 0x00000008L +#define UVD_JRBC0_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT_MASK 0x00000010L +#define UVD_JRBC0_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT_MASK 0x00000020L +#define UVD_JRBC0_UVD_JRBC_STATUS__IB_ILLEGAL_CMD_MASK 0x00000040L +#define UVD_JRBC0_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT_MASK 0x00000080L +#define UVD_JRBC0_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT_MASK 0x00000100L +#define UVD_JRBC0_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT_MASK 0x00000200L +#define UVD_JRBC0_UVD_JRBC_STATUS__RB_TRAP_STATUS_MASK 0x00000400L +#define UVD_JRBC0_UVD_JRBC_STATUS__PREEMPT_STATUS_MASK 0x00000800L +#define UVD_JRBC0_UVD_JRBC_STATUS__IB_TRAP_STATUS_MASK 0x00001000L +#define UVD_JRBC0_UVD_JRBC_STATUS__INT_EN_MASK 0x00010000L +#define UVD_JRBC0_UVD_JRBC_STATUS__INT_ACK_MASK 0x00020000L +//UVD_JRBC0_UVD_JRBC_RB_RPTR +#define UVD_JRBC0_UVD_JRBC_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_JRBC0_UVD_JRBC_RB_RPTR__RB_RPTR_MASK 0x007FFFF0L +//UVD_JRBC0_UVD_JRBC_RB_BUF_STATUS +#define UVD_JRBC0_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC0_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC0_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC0_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC0_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC0_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC0_UVD_JRBC_IB_BUF_STATUS +#define UVD_JRBC0_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC0_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC0_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC0_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC0_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC0_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC0_UVD_JRBC_IB_SIZE_UPDATE +#define UVD_JRBC0_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE__SHIFT 0x4 +#define UVD_JRBC0_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC0_UVD_JRBC_IB_COND_RD_TIMER +#define UVD_JRBC0_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC0_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC0_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC0_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC0_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC0_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC0_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC0_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC0_UVD_JRBC_IB_REF_DATA +#define UVD_JRBC0_UVD_JRBC_IB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC0_UVD_JRBC_IB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC0_UVD_JPEG_PREEMPT_CMD +#define UVD_JRBC0_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN__SHIFT 0x0 +#define UVD_JRBC0_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC0_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD__SHIFT 0x2 +#define UVD_JRBC0_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN_MASK 0x00000001L +#define UVD_JRBC0_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC0_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD_MASK 0x00000004L +//UVD_JRBC0_UVD_JPEG_PREEMPT_FENCE_DATA0 +#define UVD_JRBC0_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0__SHIFT 0x0 +#define UVD_JRBC0_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0_MASK 0xFFFFFFFFL +//UVD_JRBC0_UVD_JPEG_PREEMPT_FENCE_DATA1 +#define UVD_JRBC0_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1__SHIFT 0x0 +#define UVD_JRBC0_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1_MASK 0xFFFFFFFFL +//UVD_JRBC0_UVD_JRBC_RB_SIZE +#define UVD_JRBC0_UVD_JRBC_RB_SIZE__RB_SIZE__SHIFT 0x4 +#define UVD_JRBC0_UVD_JRBC_RB_SIZE__RB_SIZE_MASK 0x00FFFFF0L +//UVD_JRBC0_UVD_JRBC_SCRATCH0 +#define UVD_JRBC0_UVD_JRBC_SCRATCH0__SCRATCH0__SHIFT 0x0 +#define UVD_JRBC0_UVD_JRBC_SCRATCH0__SCRATCH0_MASK 0xFFFFFFFFL + + +// addressBlock: aid_uvd0_uvd_jmi0_uvd_jmi_dec +//UVD_JMI0_UVD_JPEG_DEC_PF_CTRL +#define UVD_JMI0_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS__SHIFT 0x0 +#define UVD_JMI0_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING__SHIFT 0x1 +#define UVD_JMI0_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS_MASK 0x00000001L +#define UVD_JMI0_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING_MASK 0x00000002L +//UVD_JMI0_UVD_LMI_JRBC_CTRL +#define UVD_JMI0_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI0_UVD_LMI_JRBC_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI0_UVD_LMI_JRBC_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI0_UVD_LMI_JRBC_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI0_UVD_LMI_JRBC_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI0_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI0_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI0_UVD_LMI_JRBC_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI0_UVD_LMI_JRBC_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI0_UVD_LMI_JRBC_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI0_UVD_LMI_JRBC_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI0_UVD_LMI_JPEG_CTRL +#define UVD_JMI0_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI0_UVD_LMI_JPEG_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI0_UVD_LMI_JPEG_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI0_UVD_LMI_JPEG_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI0_UVD_LMI_JPEG_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI0_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI0_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI0_UVD_LMI_JPEG_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI0_UVD_LMI_JPEG_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI0_UVD_LMI_JPEG_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI0_UVD_LMI_JPEG_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI0_JPEG_LMI_DROP +#define UVD_JMI0_JPEG_LMI_DROP__JPEG_WR_DROP__SHIFT 0x0 +#define UVD_JMI0_JPEG_LMI_DROP__JRBC_WR_DROP__SHIFT 0x1 +#define UVD_JMI0_JPEG_LMI_DROP__JPEG_RD_DROP__SHIFT 0x2 +#define UVD_JMI0_JPEG_LMI_DROP__JRBC_RD_DROP__SHIFT 0x3 +#define UVD_JMI0_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP__SHIFT 0x4 +#define UVD_JMI0_JPEG_LMI_DROP__JPEG_WR_DROP_MASK 0x00000001L +#define UVD_JMI0_JPEG_LMI_DROP__JRBC_WR_DROP_MASK 0x00000002L +#define UVD_JMI0_JPEG_LMI_DROP__JPEG_RD_DROP_MASK 0x00000004L +#define UVD_JMI0_JPEG_LMI_DROP__JRBC_RD_DROP_MASK 0x00000008L +#define UVD_JMI0_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP_MASK 0x00000010L +//UVD_JMI0_UVD_LMI_JRBC_IB_VMID +#define UVD_JMI0_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID__SHIFT 0x4 +#define UVD_JMI0_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI0_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI0_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI0_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI0_UVD_LMI_JRBC_RB_VMID +#define UVD_JMI0_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID__SHIFT 0x4 +#define UVD_JMI0_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI0_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI0_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI0_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI0_UVD_LMI_JPEG_VMID +#define UVD_JMI0_UVD_LMI_JPEG_VMID__JPEG_RD_VMID__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JPEG_VMID__JPEG_WR_VMID__SHIFT 0x4 +#define UVD_JMI0_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID__SHIFT 0x8 +#define UVD_JMI0_UVD_LMI_JPEG_VMID__JPEG_RD_VMID_MASK 0x0000000FL +#define UVD_JMI0_UVD_LMI_JPEG_VMID__JPEG_WR_VMID_MASK 0x000000F0L +#define UVD_JMI0_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID_MASK 0x00000F00L +//UVD_JMI0_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW +#define UVD_JMI0_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH +#define UVD_JMI0_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_LOW +#define UVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH +#define UVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI0_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI0_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JPEG_PREEMPT_VMID +#define UVD_JMI0_UVD_LMI_JPEG_PREEMPT_VMID__VMID__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JPEG_PREEMPT_VMID__VMID_MASK 0x0000000FL +//UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP__SHIFT 0x0 +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP__SHIFT 0x2 +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP__SHIFT 0x4 +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP__SHIFT 0x6 +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP__SHIFT 0x8 +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP__SHIFT 0xa +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP__SHIFT 0xc +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP__SHIFT 0xe +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP__SHIFT 0x10 +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP_MASK 0x00000003L +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP_MASK 0x0000000CL +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP_MASK 0x00000030L +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP_MASK 0x000000C0L +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP_MASK 0x00000300L +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP_MASK 0x00000C00L +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP_MASK 0x00003000L +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP_MASK 0x0000C000L +#define UVD_JMI0_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP_MASK 0x00030000L +//UVD_JMI0_UVD_JMI_ATOMIC_CNTL +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en__SHIFT 0x0 +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL__atomic_max_burst__SHIFT 0x1 +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop__SHIFT 0x5 +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en__SHIFT 0x6 +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG__SHIFT 0x7 +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE__SHIFT 0xb +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en_MASK 0x00000001L +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL__atomic_max_burst_MASK 0x0000001EL +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop_MASK 0x00000020L +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en_MASK 0x00000040L +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG_MASK 0x00000780L +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE_MASK 0x00000800L +//UVD_JMI0_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW +#define UVD_JMI0_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI0_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH +#define UVD_JMI0_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI0_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JPEG_READ_64BIT_BAR_LOW +#define UVD_JMI0_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH +#define UVD_JMI0_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW +#define UVD_JMI0_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH +#define UVD_JMI0_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JRBC_IB_64BIT_BAR_LOW +#define UVD_JMI0_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH +#define UVD_JMI0_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI0_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI0_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI0_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI0_UVD_JMI_ATOMIC_CNTL2 +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap__SHIFT 0x10 +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP__SHIFT 0x18 +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap_MASK 0x00FF0000L +#define UVD_JMI0_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP_MASK 0xFF000000L + + +// addressBlock: aid_uvd0_uvd_jmi_common_dec +//UVD_JADP_MCIF_URGENT_CTRL +#define UVD_JADP_MCIF_URGENT_CTRL__WR_WATERMARK__SHIFT 0x0 +#define UVD_JADP_MCIF_URGENT_CTRL__RD_WATERMARK__SHIFT 0x6 +#define UVD_JADP_MCIF_URGENT_CTRL__WR_RD_URGENT_TIMER__SHIFT 0xb +#define UVD_JADP_MCIF_URGENT_CTRL__WR_URGENT_PROG_STEP__SHIFT 0x11 +#define UVD_JADP_MCIF_URGENT_CTRL__RD_URGENT_PROG_STEP__SHIFT 0x15 +#define UVD_JADP_MCIF_URGENT_CTRL__WR_QOS_EN__SHIFT 0x19 +#define UVD_JADP_MCIF_URGENT_CTRL__RD_QOS_EN__SHIFT 0x1a +#define UVD_JADP_MCIF_URGENT_CTRL__WR_WATERMARK_MASK 0x0000003FL +#define UVD_JADP_MCIF_URGENT_CTRL__RD_WATERMARK_MASK 0x000007C0L +#define UVD_JADP_MCIF_URGENT_CTRL__WR_RD_URGENT_TIMER_MASK 0x0001F800L +#define UVD_JADP_MCIF_URGENT_CTRL__WR_URGENT_PROG_STEP_MASK 0x001E0000L +#define UVD_JADP_MCIF_URGENT_CTRL__RD_URGENT_PROG_STEP_MASK 0x01E00000L +#define UVD_JADP_MCIF_URGENT_CTRL__WR_QOS_EN_MASK 0x02000000L +#define UVD_JADP_MCIF_URGENT_CTRL__RD_QOS_EN_MASK 0x04000000L +//UVD_JMI_URGENT_CTRL +#define UVD_JMI_URGENT_CTRL__ENABLE_MC_RD_URGENT_STALL__SHIFT 0x0 +#define UVD_JMI_URGENT_CTRL__ASSERT_MC_RD_URGENT__SHIFT 0x4 +#define UVD_JMI_URGENT_CTRL__ENABLE_MC_WR_URGENT_STALL__SHIFT 0x10 +#define UVD_JMI_URGENT_CTRL__ASSERT_MC_WR_URGENT__SHIFT 0x14 +#define UVD_JMI_URGENT_CTRL__ENABLE_MC_RD_URGENT_STALL_MASK 0x00000001L +#define UVD_JMI_URGENT_CTRL__ASSERT_MC_RD_URGENT_MASK 0x000000F0L +#define UVD_JMI_URGENT_CTRL__ENABLE_MC_WR_URGENT_STALL_MASK 0x00010000L +#define UVD_JMI_URGENT_CTRL__ASSERT_MC_WR_URGENT_MASK 0x00F00000L +//UVD_JMI_CTRL +#define UVD_JMI_CTRL__STALL_MC_ARB__SHIFT 0x0 +#define UVD_JMI_CTRL__MASK_MC_URGENT__SHIFT 0x1 +#define UVD_JMI_CTRL__ASSERT_MC_URGENT__SHIFT 0x2 +#define UVD_JMI_CTRL__MC_RD_ARB_WAIT_TIMER__SHIFT 0x8 +#define UVD_JMI_CTRL__MC_WR_ARB_WAIT_TIMER__SHIFT 0x10 +#define UVD_JMI_CTRL__STALL_MC_ARB_MASK 0x00000001L +#define UVD_JMI_CTRL__MASK_MC_URGENT_MASK 0x00000002L +#define UVD_JMI_CTRL__ASSERT_MC_URGENT_MASK 0x00000004L +#define UVD_JMI_CTRL__MC_RD_ARB_WAIT_TIMER_MASK 0x0000FF00L +#define UVD_JMI_CTRL__MC_WR_ARB_WAIT_TIMER_MASK 0x00FF0000L +//JPEG_MEMCHECK_CLAMPING_CNTL +#define JPEG_MEMCHECK_CLAMPING_CNTL__CLAMP_TO_SAFE_ADDR_EN__SHIFT 0x0 +#define JPEG_MEMCHECK_CLAMPING_CNTL__CLAMP_TO_SAFE_ADDR_EN_MASK 0x00000001L +//JPEG_MEMCHECK_SAFE_ADDR +#define JPEG_MEMCHECK_SAFE_ADDR__MEMCHECK_SAFE_ADDR__SHIFT 0x0 +#define JPEG_MEMCHECK_SAFE_ADDR__MEMCHECK_SAFE_ADDR_MASK 0xFFFFFFFFL +//JPEG_MEMCHECK_SAFE_ADDR_64BIT +#define JPEG_MEMCHECK_SAFE_ADDR_64BIT__MEMCHECK_SAFE_ADDR_64BIT__SHIFT 0x0 +#define JPEG_MEMCHECK_SAFE_ADDR_64BIT__MEMCHECK_SAFE_ADDR_64BIT_MASK 0xFFFFFFFFL +//UVD_JMI_LAT_CTRL +#define UVD_JMI_LAT_CTRL__SCALE__SHIFT 0x0 +#define UVD_JMI_LAT_CTRL__MAX_START__SHIFT 0x8 +#define UVD_JMI_LAT_CTRL__MIN_START__SHIFT 0x9 +#define UVD_JMI_LAT_CTRL__AVG_START__SHIFT 0xa +#define UVD_JMI_LAT_CTRL__PERFMON_SYNC__SHIFT 0xb +#define UVD_JMI_LAT_CTRL__SKIP__SHIFT 0x10 +#define UVD_JMI_LAT_CTRL__SCALE_MASK 0x000000FFL +#define UVD_JMI_LAT_CTRL__MAX_START_MASK 0x00000100L +#define UVD_JMI_LAT_CTRL__MIN_START_MASK 0x00000200L +#define UVD_JMI_LAT_CTRL__AVG_START_MASK 0x00000400L +#define UVD_JMI_LAT_CTRL__PERFMON_SYNC_MASK 0x00000800L +#define UVD_JMI_LAT_CTRL__SKIP_MASK 0x000F0000L +//UVD_JMI_LAT_CNTR +#define UVD_JMI_LAT_CNTR__MAX_LAT__SHIFT 0x0 +#define UVD_JMI_LAT_CNTR__MIN_LAT__SHIFT 0x8 +#define UVD_JMI_LAT_CNTR__MAX_LAT_MASK 0x000000FFL +#define UVD_JMI_LAT_CNTR__MIN_LAT_MASK 0x0000FF00L +//UVD_JMI_AVG_LAT_CNTR +#define UVD_JMI_AVG_LAT_CNTR__ENV_LOW__SHIFT 0x0 +#define UVD_JMI_AVG_LAT_CNTR__ENV_HIGH__SHIFT 0x8 +#define UVD_JMI_AVG_LAT_CNTR__ENV_HIT__SHIFT 0x10 +#define UVD_JMI_AVG_LAT_CNTR__ENV_LOW_MASK 0x000000FFL +#define UVD_JMI_AVG_LAT_CNTR__ENV_HIGH_MASK 0x0000FF00L +#define UVD_JMI_AVG_LAT_CNTR__ENV_HIT_MASK 0xFFFF0000L +//UVD_JMI_PERFMON_CTRL +#define UVD_JMI_PERFMON_CTRL__PERFMON_STATE__SHIFT 0x0 +#define UVD_JMI_PERFMON_CTRL__PERFMON_SEL__SHIFT 0x8 +#define UVD_JMI_PERFMON_CTRL__PERFMON_STATE_MASK 0x00000003L +#define UVD_JMI_PERFMON_CTRL__PERFMON_SEL_MASK 0x00001F00L +//UVD_JMI_PERFMON_COUNT_LO +#define UVD_JMI_PERFMON_COUNT_LO__PERFMON_COUNT__SHIFT 0x0 +#define UVD_JMI_PERFMON_COUNT_LO__PERFMON_COUNT_MASK 0xFFFFFFFFL +//UVD_JMI_PERFMON_COUNT_HI +#define UVD_JMI_PERFMON_COUNT_HI__PERFMON_COUNT__SHIFT 0x0 +#define UVD_JMI_PERFMON_COUNT_HI__PERFMON_COUNT_MASK 0x0000FFFFL +//UVD_JMI_CLEAN_STATUS +#define UVD_JMI_CLEAN_STATUS__LMI_READ_CLEAN__SHIFT 0x0 +#define UVD_JMI_CLEAN_STATUS__LMI_READ_CLEAN_RAW__SHIFT 0x1 +#define UVD_JMI_CLEAN_STATUS__LMI_WRITE_CLEAN__SHIFT 0x2 +#define UVD_JMI_CLEAN_STATUS__LMI_WRITE_CLEAN_RAW__SHIFT 0x3 +#define UVD_JMI_CLEAN_STATUS__MC_WRITE_PENDING__SHIFT 0x4 +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE0_READ_CLEAN__SHIFT 0x8 +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE1_READ_CLEAN__SHIFT 0x9 +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE2_READ_CLEAN__SHIFT 0xa +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE3_READ_CLEAN__SHIFT 0xb +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE4_READ_CLEAN__SHIFT 0xc +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE5_READ_CLEAN__SHIFT 0xd +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE6_READ_CLEAN__SHIFT 0xe +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE7_READ_CLEAN__SHIFT 0xf +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE0_WRITE_CLEAN__SHIFT 0x10 +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE1_WRITE_CLEAN__SHIFT 0x11 +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE2_WRITE_CLEAN__SHIFT 0x12 +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE3_WRITE_CLEAN__SHIFT 0x13 +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE4_WRITE_CLEAN__SHIFT 0x14 +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE5_WRITE_CLEAN__SHIFT 0x15 +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE6_WRITE_CLEAN__SHIFT 0x16 +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE7_WRITE_CLEAN__SHIFT 0x17 +#define UVD_JMI_CLEAN_STATUS__LMI_READ_CLEAN_MASK 0x00000001L +#define UVD_JMI_CLEAN_STATUS__LMI_READ_CLEAN_RAW_MASK 0x00000002L +#define UVD_JMI_CLEAN_STATUS__LMI_WRITE_CLEAN_MASK 0x00000004L +#define UVD_JMI_CLEAN_STATUS__LMI_WRITE_CLEAN_RAW_MASK 0x00000008L +#define UVD_JMI_CLEAN_STATUS__MC_WRITE_PENDING_MASK 0x00000010L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE0_READ_CLEAN_MASK 0x00000100L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE1_READ_CLEAN_MASK 0x00000200L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE2_READ_CLEAN_MASK 0x00000400L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE3_READ_CLEAN_MASK 0x00000800L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE4_READ_CLEAN_MASK 0x00001000L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE5_READ_CLEAN_MASK 0x00002000L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE6_READ_CLEAN_MASK 0x00004000L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE7_READ_CLEAN_MASK 0x00008000L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE0_WRITE_CLEAN_MASK 0x00010000L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE1_WRITE_CLEAN_MASK 0x00020000L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE2_WRITE_CLEAN_MASK 0x00040000L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE3_WRITE_CLEAN_MASK 0x00080000L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE4_WRITE_CLEAN_MASK 0x00100000L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE5_WRITE_CLEAN_MASK 0x00200000L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE6_WRITE_CLEAN_MASK 0x00400000L +#define UVD_JMI_CLEAN_STATUS__DJPEG_CORE7_WRITE_CLEAN_MASK 0x00800000L +//UVD_JMI_CNTL +#define UVD_JMI_CNTL__SOFT_RESET__SHIFT 0x0 +#define UVD_JMI_CNTL__MC_RD_REQ_RET_MAX__SHIFT 0x8 +#define UVD_JMI_CNTL__SOFT_RESET_MASK 0x00000001L +#define UVD_JMI_CNTL__MC_RD_REQ_RET_MAX_MASK 0x0003FF00L + + +// addressBlock: aid_uvd0_uvd_jpeg_common_dec +//JPEG_SOFT_RESET_STATUS +#define JPEG_SOFT_RESET_STATUS__JPEG0_DEC_RESET_STATUS__SHIFT 0x0 +#define JPEG_SOFT_RESET_STATUS__JPEG1_DEC_RESET_STATUS__SHIFT 0x1 +#define JPEG_SOFT_RESET_STATUS__JPEG2_DEC_RESET_STATUS__SHIFT 0x2 +#define JPEG_SOFT_RESET_STATUS__JPEG3_DEC_RESET_STATUS__SHIFT 0x3 +#define JPEG_SOFT_RESET_STATUS__JPEG4_DEC_RESET_STATUS__SHIFT 0x4 +#define JPEG_SOFT_RESET_STATUS__JPEG5_DEC_RESET_STATUS__SHIFT 0x5 +#define JPEG_SOFT_RESET_STATUS__JPEG6_DEC_RESET_STATUS__SHIFT 0x6 +#define JPEG_SOFT_RESET_STATUS__JPEG7_DEC_RESET_STATUS__SHIFT 0x7 +#define JPEG_SOFT_RESET_STATUS__DJRBC0_RESET_STATUS__SHIFT 0x8 +#define JPEG_SOFT_RESET_STATUS__DJRBC1_RESET_STATUS__SHIFT 0x9 +#define JPEG_SOFT_RESET_STATUS__DJRBC2_RESET_STATUS__SHIFT 0xa +#define JPEG_SOFT_RESET_STATUS__DJRBC3_RESET_STATUS__SHIFT 0xb +#define JPEG_SOFT_RESET_STATUS__DJRBC4_RESET_STATUS__SHIFT 0xc +#define JPEG_SOFT_RESET_STATUS__DJRBC5_RESET_STATUS__SHIFT 0xd +#define JPEG_SOFT_RESET_STATUS__DJRBC6_RESET_STATUS__SHIFT 0xe +#define JPEG_SOFT_RESET_STATUS__DJRBC7_RESET_STATUS__SHIFT 0xf +#define JPEG_SOFT_RESET_STATUS__JPEG_ENC_RESET_STATUS__SHIFT 0x11 +#define JPEG_SOFT_RESET_STATUS__EJRBC_RESET_STATUS__SHIFT 0x12 +#define JPEG_SOFT_RESET_STATUS__JMCIF_RESET_STATUS__SHIFT 0x18 +#define JPEG_SOFT_RESET_STATUS__JPEG0_DEC_RESET_STATUS_MASK 0x00000001L +#define JPEG_SOFT_RESET_STATUS__JPEG1_DEC_RESET_STATUS_MASK 0x00000002L +#define JPEG_SOFT_RESET_STATUS__JPEG2_DEC_RESET_STATUS_MASK 0x00000004L +#define JPEG_SOFT_RESET_STATUS__JPEG3_DEC_RESET_STATUS_MASK 0x00000008L +#define JPEG_SOFT_RESET_STATUS__JPEG4_DEC_RESET_STATUS_MASK 0x00000010L +#define JPEG_SOFT_RESET_STATUS__JPEG5_DEC_RESET_STATUS_MASK 0x00000020L +#define JPEG_SOFT_RESET_STATUS__JPEG6_DEC_RESET_STATUS_MASK 0x00000040L +#define JPEG_SOFT_RESET_STATUS__JPEG7_DEC_RESET_STATUS_MASK 0x00000080L +#define JPEG_SOFT_RESET_STATUS__DJRBC0_RESET_STATUS_MASK 0x00000100L +#define JPEG_SOFT_RESET_STATUS__DJRBC1_RESET_STATUS_MASK 0x00000200L +#define JPEG_SOFT_RESET_STATUS__DJRBC2_RESET_STATUS_MASK 0x00000400L +#define JPEG_SOFT_RESET_STATUS__DJRBC3_RESET_STATUS_MASK 0x00000800L +#define JPEG_SOFT_RESET_STATUS__DJRBC4_RESET_STATUS_MASK 0x00001000L +#define JPEG_SOFT_RESET_STATUS__DJRBC5_RESET_STATUS_MASK 0x00002000L +#define JPEG_SOFT_RESET_STATUS__DJRBC6_RESET_STATUS_MASK 0x00004000L +#define JPEG_SOFT_RESET_STATUS__DJRBC7_RESET_STATUS_MASK 0x00008000L +#define JPEG_SOFT_RESET_STATUS__JPEG_ENC_RESET_STATUS_MASK 0x00020000L +#define JPEG_SOFT_RESET_STATUS__EJRBC_RESET_STATUS_MASK 0x00040000L +#define JPEG_SOFT_RESET_STATUS__JMCIF_RESET_STATUS_MASK 0x01000000L +//JPEG_SYS_INT_EN +#define JPEG_SYS_INT_EN__DJPEG0_CORE__SHIFT 0x0 +#define JPEG_SYS_INT_EN__DJPEG1_CORE__SHIFT 0x1 +#define JPEG_SYS_INT_EN__DJPEG2_CORE__SHIFT 0x2 +#define JPEG_SYS_INT_EN__DJPEG3_CORE__SHIFT 0x3 +#define JPEG_SYS_INT_EN__DJPEG4_CORE__SHIFT 0x4 +#define JPEG_SYS_INT_EN__DJPEG5_CORE__SHIFT 0x5 +#define JPEG_SYS_INT_EN__DJPEG6_CORE__SHIFT 0x6 +#define JPEG_SYS_INT_EN__DJPEG7_CORE__SHIFT 0x7 +#define JPEG_SYS_INT_EN__DJRBC0__SHIFT 0x8 +#define JPEG_SYS_INT_EN__DJRBC1__SHIFT 0x9 +#define JPEG_SYS_INT_EN__DJRBC2__SHIFT 0xa +#define JPEG_SYS_INT_EN__DJRBC3__SHIFT 0xb +#define JPEG_SYS_INT_EN__DJRBC4__SHIFT 0xc +#define JPEG_SYS_INT_EN__DJRBC5__SHIFT 0xd +#define JPEG_SYS_INT_EN__DJRBC6__SHIFT 0xe +#define JPEG_SYS_INT_EN__DJRBC7__SHIFT 0xf +#define JPEG_SYS_INT_EN__DJPEG0_PF_RPT__SHIFT 0x10 +#define JPEG_SYS_INT_EN__DJPEG1_PF_RPT__SHIFT 0x11 +#define JPEG_SYS_INT_EN__DJPEG2_PF_RPT__SHIFT 0x12 +#define JPEG_SYS_INT_EN__DJPEG3_PF_RPT__SHIFT 0x13 +#define JPEG_SYS_INT_EN__DJPEG4_PF_RPT__SHIFT 0x14 +#define JPEG_SYS_INT_EN__DJPEG5_PF_RPT__SHIFT 0x15 +#define JPEG_SYS_INT_EN__DJPEG6_PF_RPT__SHIFT 0x16 +#define JPEG_SYS_INT_EN__DJPEG7_PF_RPT__SHIFT 0x17 +#define JPEG_SYS_INT_EN__DJPEG0_RAS_CNTL__SHIFT 0x18 +#define JPEG_SYS_INT_EN__DJPEG1_RAS_CNTL__SHIFT 0x19 +#define JPEG_SYS_INT_EN__DJPEG0_CORE_MASK 0x00000001L +#define JPEG_SYS_INT_EN__DJPEG1_CORE_MASK 0x00000002L +#define JPEG_SYS_INT_EN__DJPEG2_CORE_MASK 0x00000004L +#define JPEG_SYS_INT_EN__DJPEG3_CORE_MASK 0x00000008L +#define JPEG_SYS_INT_EN__DJPEG4_CORE_MASK 0x00000010L +#define JPEG_SYS_INT_EN__DJPEG5_CORE_MASK 0x00000020L +#define JPEG_SYS_INT_EN__DJPEG6_CORE_MASK 0x00000040L +#define JPEG_SYS_INT_EN__DJPEG7_CORE_MASK 0x00000080L +#define JPEG_SYS_INT_EN__DJRBC0_MASK 0x00000100L +#define JPEG_SYS_INT_EN__DJRBC1_MASK 0x00000200L +#define JPEG_SYS_INT_EN__DJRBC2_MASK 0x00000400L +#define JPEG_SYS_INT_EN__DJRBC3_MASK 0x00000800L +#define JPEG_SYS_INT_EN__DJRBC4_MASK 0x00001000L +#define JPEG_SYS_INT_EN__DJRBC5_MASK 0x00002000L +#define JPEG_SYS_INT_EN__DJRBC6_MASK 0x00004000L +#define JPEG_SYS_INT_EN__DJRBC7_MASK 0x00008000L +#define JPEG_SYS_INT_EN__DJPEG0_PF_RPT_MASK 0x00010000L +#define JPEG_SYS_INT_EN__DJPEG1_PF_RPT_MASK 0x00020000L +#define JPEG_SYS_INT_EN__DJPEG2_PF_RPT_MASK 0x00040000L +#define JPEG_SYS_INT_EN__DJPEG3_PF_RPT_MASK 0x00080000L +#define JPEG_SYS_INT_EN__DJPEG4_PF_RPT_MASK 0x00100000L +#define JPEG_SYS_INT_EN__DJPEG5_PF_RPT_MASK 0x00200000L +#define JPEG_SYS_INT_EN__DJPEG6_PF_RPT_MASK 0x00400000L +#define JPEG_SYS_INT_EN__DJPEG7_PF_RPT_MASK 0x00800000L +#define JPEG_SYS_INT_EN__DJPEG0_RAS_CNTL_MASK 0x01000000L +#define JPEG_SYS_INT_EN__DJPEG1_RAS_CNTL_MASK 0x02000000L +//JPEG_SYS_INT_EN1 +#define JPEG_SYS_INT_EN1__EJPEG_PF_RPT__SHIFT 0x0 +#define JPEG_SYS_INT_EN1__EJPEG_CORE__SHIFT 0x1 +#define JPEG_SYS_INT_EN1__EJRBC__SHIFT 0x2 +#define JPEG_SYS_INT_EN1__EJPEG_RAS_CNTL__SHIFT 0x3 +#define JPEG_SYS_INT_EN1__EJPEG_PF_RPT_MASK 0x00000001L +#define JPEG_SYS_INT_EN1__EJPEG_CORE_MASK 0x00000002L +#define JPEG_SYS_INT_EN1__EJRBC_MASK 0x00000004L +#define JPEG_SYS_INT_EN1__EJPEG_RAS_CNTL_MASK 0x00000008L +//JPEG_SYS_INT_STATUS +#define JPEG_SYS_INT_STATUS__DJPEG0_CORE__SHIFT 0x0 +#define JPEG_SYS_INT_STATUS__DJPEG1_CORE__SHIFT 0x1 +#define JPEG_SYS_INT_STATUS__DJPEG2_CORE__SHIFT 0x2 +#define JPEG_SYS_INT_STATUS__DJPEG3_CORE__SHIFT 0x3 +#define JPEG_SYS_INT_STATUS__DJPEG4_CORE__SHIFT 0x4 +#define JPEG_SYS_INT_STATUS__DJPEG5_CORE__SHIFT 0x5 +#define JPEG_SYS_INT_STATUS__DJPEG6_CORE__SHIFT 0x6 +#define JPEG_SYS_INT_STATUS__DJPEG7_CORE__SHIFT 0x7 +#define JPEG_SYS_INT_STATUS__DJRBC0__SHIFT 0x8 +#define JPEG_SYS_INT_STATUS__DJRBC1__SHIFT 0x9 +#define JPEG_SYS_INT_STATUS__DJRBC2__SHIFT 0xa +#define JPEG_SYS_INT_STATUS__DJRBC3__SHIFT 0xb +#define JPEG_SYS_INT_STATUS__DJRBC4__SHIFT 0xc +#define JPEG_SYS_INT_STATUS__DJRBC5__SHIFT 0xd +#define JPEG_SYS_INT_STATUS__DJRBC6__SHIFT 0xe +#define JPEG_SYS_INT_STATUS__DJRBC7__SHIFT 0xf +#define JPEG_SYS_INT_STATUS__DJPEG0_PF_RPT__SHIFT 0x10 +#define JPEG_SYS_INT_STATUS__DJPEG1_PF_RPT__SHIFT 0x11 +#define JPEG_SYS_INT_STATUS__DJPEG2_PF_RPT__SHIFT 0x12 +#define JPEG_SYS_INT_STATUS__DJPEG3_PF_RPT__SHIFT 0x13 +#define JPEG_SYS_INT_STATUS__DJPEG4_PF_RPT__SHIFT 0x14 +#define JPEG_SYS_INT_STATUS__DJPEG5_PF_RPT__SHIFT 0x15 +#define JPEG_SYS_INT_STATUS__DJPEG6_PF_RPT__SHIFT 0x16 +#define JPEG_SYS_INT_STATUS__DJPEG7_PF_RPT__SHIFT 0x17 +#define JPEG_SYS_INT_STATUS__DJPEG0_RAS_CNTL__SHIFT 0x18 +#define JPEG_SYS_INT_STATUS__DJPEG1_RAS_CNTL__SHIFT 0x19 +#define JPEG_SYS_INT_STATUS__DJPEG0_CORE_MASK 0x00000001L +#define JPEG_SYS_INT_STATUS__DJPEG1_CORE_MASK 0x00000002L +#define JPEG_SYS_INT_STATUS__DJPEG2_CORE_MASK 0x00000004L +#define JPEG_SYS_INT_STATUS__DJPEG3_CORE_MASK 0x00000008L +#define JPEG_SYS_INT_STATUS__DJPEG4_CORE_MASK 0x00000010L +#define JPEG_SYS_INT_STATUS__DJPEG5_CORE_MASK 0x00000020L +#define JPEG_SYS_INT_STATUS__DJPEG6_CORE_MASK 0x00000040L +#define JPEG_SYS_INT_STATUS__DJPEG7_CORE_MASK 0x00000080L +#define JPEG_SYS_INT_STATUS__DJRBC0_MASK 0x00000100L +#define JPEG_SYS_INT_STATUS__DJRBC1_MASK 0x00000200L +#define JPEG_SYS_INT_STATUS__DJRBC2_MASK 0x00000400L +#define JPEG_SYS_INT_STATUS__DJRBC3_MASK 0x00000800L +#define JPEG_SYS_INT_STATUS__DJRBC4_MASK 0x00001000L +#define JPEG_SYS_INT_STATUS__DJRBC5_MASK 0x00002000L +#define JPEG_SYS_INT_STATUS__DJRBC6_MASK 0x00004000L +#define JPEG_SYS_INT_STATUS__DJRBC7_MASK 0x00008000L +#define JPEG_SYS_INT_STATUS__DJPEG0_PF_RPT_MASK 0x00010000L +#define JPEG_SYS_INT_STATUS__DJPEG1_PF_RPT_MASK 0x00020000L +#define JPEG_SYS_INT_STATUS__DJPEG2_PF_RPT_MASK 0x00040000L +#define JPEG_SYS_INT_STATUS__DJPEG3_PF_RPT_MASK 0x00080000L +#define JPEG_SYS_INT_STATUS__DJPEG4_PF_RPT_MASK 0x00100000L +#define JPEG_SYS_INT_STATUS__DJPEG5_PF_RPT_MASK 0x00200000L +#define JPEG_SYS_INT_STATUS__DJPEG6_PF_RPT_MASK 0x00400000L +#define JPEG_SYS_INT_STATUS__DJPEG7_PF_RPT_MASK 0x00800000L +#define JPEG_SYS_INT_STATUS__DJPEG0_RAS_CNTL_MASK 0x01000000L +#define JPEG_SYS_INT_STATUS__DJPEG1_RAS_CNTL_MASK 0x02000000L +//JPEG_SYS_INT_STATUS1 +#define JPEG_SYS_INT_STATUS1__EJPEG_PF_RPT__SHIFT 0x0 +#define JPEG_SYS_INT_STATUS1__EJPEG_CORE__SHIFT 0x1 +#define JPEG_SYS_INT_STATUS1__EJRBC__SHIFT 0x2 +#define JPEG_SYS_INT_STATUS1__EJPEG_RAS_CNTL__SHIFT 0x3 +#define JPEG_SYS_INT_STATUS1__EJPEG_PF_RPT_MASK 0x00000001L +#define JPEG_SYS_INT_STATUS1__EJPEG_CORE_MASK 0x00000002L +#define JPEG_SYS_INT_STATUS1__EJRBC_MASK 0x00000004L +#define JPEG_SYS_INT_STATUS1__EJPEG_RAS_CNTL_MASK 0x00000008L +//JPEG_SYS_INT_ACK +#define JPEG_SYS_INT_ACK__DJPEG0_CORE__SHIFT 0x0 +#define JPEG_SYS_INT_ACK__DJPEG1_CORE__SHIFT 0x1 +#define JPEG_SYS_INT_ACK__DJPEG2_CORE__SHIFT 0x2 +#define JPEG_SYS_INT_ACK__DJPEG3_CORE__SHIFT 0x3 +#define JPEG_SYS_INT_ACK__DJPEG4_CORE__SHIFT 0x4 +#define JPEG_SYS_INT_ACK__DJPEG5_CORE__SHIFT 0x5 +#define JPEG_SYS_INT_ACK__DJPEG6_CORE__SHIFT 0x6 +#define JPEG_SYS_INT_ACK__DJPEG7_CORE__SHIFT 0x7 +#define JPEG_SYS_INT_ACK__DJRBC0__SHIFT 0x8 +#define JPEG_SYS_INT_ACK__DJRBC1__SHIFT 0x9 +#define JPEG_SYS_INT_ACK__DJRBC2__SHIFT 0xa +#define JPEG_SYS_INT_ACK__DJRBC3__SHIFT 0xb +#define JPEG_SYS_INT_ACK__DJRBC4__SHIFT 0xc +#define JPEG_SYS_INT_ACK__DJRBC5__SHIFT 0xd +#define JPEG_SYS_INT_ACK__DJRBC6__SHIFT 0xe +#define JPEG_SYS_INT_ACK__DJRBC7__SHIFT 0xf +#define JPEG_SYS_INT_ACK__DJPEG0_PF_RPT__SHIFT 0x10 +#define JPEG_SYS_INT_ACK__DJPEG1_PF_RPT__SHIFT 0x11 +#define JPEG_SYS_INT_ACK__DJPEG2_PF_RPT__SHIFT 0x12 +#define JPEG_SYS_INT_ACK__DJPEG3_PF_RPT__SHIFT 0x13 +#define JPEG_SYS_INT_ACK__DJPEG4_PF_RPT__SHIFT 0x14 +#define JPEG_SYS_INT_ACK__DJPEG5_PF_RPT__SHIFT 0x15 +#define JPEG_SYS_INT_ACK__DJPEG6_PF_RPT__SHIFT 0x16 +#define JPEG_SYS_INT_ACK__DJPEG7_PF_RPT__SHIFT 0x17 +#define JPEG_SYS_INT_ACK__DJPEG0_RAS_CNTL__SHIFT 0x18 +#define JPEG_SYS_INT_ACK__DJPEG1_RAS_CNTL__SHIFT 0x19 +#define JPEG_SYS_INT_ACK__DJPEG0_CORE_MASK 0x00000001L +#define JPEG_SYS_INT_ACK__DJPEG1_CORE_MASK 0x00000002L +#define JPEG_SYS_INT_ACK__DJPEG2_CORE_MASK 0x00000004L +#define JPEG_SYS_INT_ACK__DJPEG3_CORE_MASK 0x00000008L +#define JPEG_SYS_INT_ACK__DJPEG4_CORE_MASK 0x00000010L +#define JPEG_SYS_INT_ACK__DJPEG5_CORE_MASK 0x00000020L +#define JPEG_SYS_INT_ACK__DJPEG6_CORE_MASK 0x00000040L +#define JPEG_SYS_INT_ACK__DJPEG7_CORE_MASK 0x00000080L +#define JPEG_SYS_INT_ACK__DJRBC0_MASK 0x00000100L +#define JPEG_SYS_INT_ACK__DJRBC1_MASK 0x00000200L +#define JPEG_SYS_INT_ACK__DJRBC2_MASK 0x00000400L +#define JPEG_SYS_INT_ACK__DJRBC3_MASK 0x00000800L +#define JPEG_SYS_INT_ACK__DJRBC4_MASK 0x00001000L +#define JPEG_SYS_INT_ACK__DJRBC5_MASK 0x00002000L +#define JPEG_SYS_INT_ACK__DJRBC6_MASK 0x00004000L +#define JPEG_SYS_INT_ACK__DJRBC7_MASK 0x00008000L +#define JPEG_SYS_INT_ACK__DJPEG0_PF_RPT_MASK 0x00010000L +#define JPEG_SYS_INT_ACK__DJPEG1_PF_RPT_MASK 0x00020000L +#define JPEG_SYS_INT_ACK__DJPEG2_PF_RPT_MASK 0x00040000L +#define JPEG_SYS_INT_ACK__DJPEG3_PF_RPT_MASK 0x00080000L +#define JPEG_SYS_INT_ACK__DJPEG4_PF_RPT_MASK 0x00100000L +#define JPEG_SYS_INT_ACK__DJPEG5_PF_RPT_MASK 0x00200000L +#define JPEG_SYS_INT_ACK__DJPEG6_PF_RPT_MASK 0x00400000L +#define JPEG_SYS_INT_ACK__DJPEG7_PF_RPT_MASK 0x00800000L +#define JPEG_SYS_INT_ACK__DJPEG0_RAS_CNTL_MASK 0x01000000L +#define JPEG_SYS_INT_ACK__DJPEG1_RAS_CNTL_MASK 0x02000000L +//JPEG_SYS_INT_ACK1 +#define JPEG_SYS_INT_ACK1__EJPEG_PF_RPT__SHIFT 0x0 +#define JPEG_SYS_INT_ACK1__EJPEG_CORE__SHIFT 0x1 +#define JPEG_SYS_INT_ACK1__EJRBC__SHIFT 0x2 +#define JPEG_SYS_INT_ACK1__EJPEG_RAS_CNTL__SHIFT 0x3 +#define JPEG_SYS_INT_ACK1__EJPEG_PF_RPT_MASK 0x00000001L +#define JPEG_SYS_INT_ACK1__EJPEG_CORE_MASK 0x00000002L +#define JPEG_SYS_INT_ACK1__EJRBC_MASK 0x00000004L +#define JPEG_SYS_INT_ACK1__EJPEG_RAS_CNTL_MASK 0x00000008L +//JPEG_MEMCHECK_SYS_INT_EN +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC0_RD_ERR_EN__SHIFT 0x0 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC1_RD_ERR_EN__SHIFT 0x1 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC2_RD_ERR_EN__SHIFT 0x2 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC3_RD_ERR_EN__SHIFT 0x3 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC4_RD_ERR_EN__SHIFT 0x4 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC5_RD_ERR_EN__SHIFT 0x5 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC6_RD_ERR_EN__SHIFT 0x6 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC7_RD_ERR_EN__SHIFT 0x7 +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH0_RD_ERR_EN__SHIFT 0x8 +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH1_RD_ERR_EN__SHIFT 0x9 +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH2_RD_ERR_EN__SHIFT 0xa +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH3_RD_ERR_EN__SHIFT 0xb +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH4_RD_ERR_EN__SHIFT 0xc +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH5_RD_ERR_EN__SHIFT 0xd +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH6_RD_ERR_EN__SHIFT 0xe +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH7_RD_ERR_EN__SHIFT 0xf +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC0_WR_ERR_EN__SHIFT 0x10 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC1_WR_ERR_EN__SHIFT 0x11 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC2_WR_ERR_EN__SHIFT 0x12 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC3_WR_ERR_EN__SHIFT 0x13 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC4_WR_ERR_EN__SHIFT 0x14 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC5_WR_ERR_EN__SHIFT 0x15 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC6_WR_ERR_EN__SHIFT 0x16 +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC7_WR_ERR_EN__SHIFT 0x17 +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF0_WR_ERR_EN__SHIFT 0x18 +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF1_WR_ERR_EN__SHIFT 0x19 +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF2_WR_ERR_EN__SHIFT 0x1a +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF3_WR_ERR_EN__SHIFT 0x1b +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF4_WR_ERR_EN__SHIFT 0x1c +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF5_WR_ERR_EN__SHIFT 0x1d +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF6_WR_ERR_EN__SHIFT 0x1e +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF7_WR_ERR_EN__SHIFT 0x1f +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC0_RD_ERR_EN_MASK 0x00000001L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC1_RD_ERR_EN_MASK 0x00000002L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC2_RD_ERR_EN_MASK 0x00000004L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC3_RD_ERR_EN_MASK 0x00000008L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC4_RD_ERR_EN_MASK 0x00000010L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC5_RD_ERR_EN_MASK 0x00000020L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC6_RD_ERR_EN_MASK 0x00000040L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC7_RD_ERR_EN_MASK 0x00000080L +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH0_RD_ERR_EN_MASK 0x00000100L +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH1_RD_ERR_EN_MASK 0x00000200L +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH2_RD_ERR_EN_MASK 0x00000400L +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH3_RD_ERR_EN_MASK 0x00000800L +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH4_RD_ERR_EN_MASK 0x00001000L +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH5_RD_ERR_EN_MASK 0x00002000L +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH6_RD_ERR_EN_MASK 0x00004000L +#define JPEG_MEMCHECK_SYS_INT_EN__BSFETCH7_RD_ERR_EN_MASK 0x00008000L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC0_WR_ERR_EN_MASK 0x00010000L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC1_WR_ERR_EN_MASK 0x00020000L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC2_WR_ERR_EN_MASK 0x00040000L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC3_WR_ERR_EN_MASK 0x00080000L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC4_WR_ERR_EN_MASK 0x00100000L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC5_WR_ERR_EN_MASK 0x00200000L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC6_WR_ERR_EN_MASK 0x00400000L +#define JPEG_MEMCHECK_SYS_INT_EN__DJRBC7_WR_ERR_EN_MASK 0x00800000L +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF0_WR_ERR_EN_MASK 0x01000000L +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF1_WR_ERR_EN_MASK 0x02000000L +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF2_WR_ERR_EN_MASK 0x04000000L +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF3_WR_ERR_EN_MASK 0x08000000L +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF4_WR_ERR_EN_MASK 0x10000000L +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF5_WR_ERR_EN_MASK 0x20000000L +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF6_WR_ERR_EN_MASK 0x40000000L +#define JPEG_MEMCHECK_SYS_INT_EN__OBUF7_WR_ERR_EN_MASK 0x80000000L +//JPEG_MEMCHECK_SYS_INT_EN1 +#define JPEG_MEMCHECK_SYS_INT_EN1__EJRBC_RD_ERR_EN__SHIFT 0x0 +#define JPEG_MEMCHECK_SYS_INT_EN1__PELFETCH_RD_ERR_EN__SHIFT 0x1 +#define JPEG_MEMCHECK_SYS_INT_EN1__SCALAR_RD_ERR_EN__SHIFT 0x2 +#define JPEG_MEMCHECK_SYS_INT_EN1__EJRBC_WR_ERR_EN__SHIFT 0x3 +#define JPEG_MEMCHECK_SYS_INT_EN1__BS_WR_ERR_EN__SHIFT 0x4 +#define JPEG_MEMCHECK_SYS_INT_EN1__SCALAR_WR_ERR_EN__SHIFT 0x5 +#define JPEG_MEMCHECK_SYS_INT_EN1__EJRBC_RD_ERR_EN_MASK 0x00000001L +#define JPEG_MEMCHECK_SYS_INT_EN1__PELFETCH_RD_ERR_EN_MASK 0x00000002L +#define JPEG_MEMCHECK_SYS_INT_EN1__SCALAR_RD_ERR_EN_MASK 0x00000004L +#define JPEG_MEMCHECK_SYS_INT_EN1__EJRBC_WR_ERR_EN_MASK 0x00000008L +#define JPEG_MEMCHECK_SYS_INT_EN1__BS_WR_ERR_EN_MASK 0x00000010L +#define JPEG_MEMCHECK_SYS_INT_EN1__SCALAR_WR_ERR_EN_MASK 0x00000020L +//JPEG_MEMCHECK_SYS_INT_STAT +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH0_RD_HI_ERR__SHIFT 0x0 +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH1_RD_HI_ERR__SHIFT 0x1 +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH2_RD_HI_ERR__SHIFT 0x2 +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH3_RD_HI_ERR__SHIFT 0x3 +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH4_RD_HI_ERR__SHIFT 0x4 +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH5_RD_HI_ERR__SHIFT 0x5 +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH6_RD_HI_ERR__SHIFT 0x6 +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH7_RD_HI_ERR__SHIFT 0x7 +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH0_RD_LO_ERR__SHIFT 0x8 +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH1_RD_LO_ERR__SHIFT 0x9 +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH2_RD_LO_ERR__SHIFT 0xa +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH3_RD_LO_ERR__SHIFT 0xb +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH4_RD_LO_ERR__SHIFT 0xc +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH5_RD_LO_ERR__SHIFT 0xd +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH6_RD_LO_ERR__SHIFT 0xe +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH7_RD_LO_ERR__SHIFT 0xf +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF0_WR_HI_ERR__SHIFT 0x10 +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF1_WR_HI_ERR__SHIFT 0x11 +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF2_WR_HI_ERR__SHIFT 0x12 +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF3_WR_HI_ERR__SHIFT 0x13 +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF4_WR_HI_ERR__SHIFT 0x14 +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF5_WR_HI_ERR__SHIFT 0x15 +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF6_WR_HI_ERR__SHIFT 0x16 +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF7_WR_HI_ERR__SHIFT 0x17 +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF0_WR_LO_ERR__SHIFT 0x18 +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF1_WR_LO_ERR__SHIFT 0x19 +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF2_WR_LO_ERR__SHIFT 0x1a +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF3_WR_LO_ERR__SHIFT 0x1b +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF4_WR_LO_ERR__SHIFT 0x1c +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF5_WR_LO_ERR__SHIFT 0x1d +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF6_WR_LO_ERR__SHIFT 0x1e +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF7_WR_LO_ERR__SHIFT 0x1f +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH0_RD_HI_ERR_MASK 0x00000001L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH1_RD_HI_ERR_MASK 0x00000002L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH2_RD_HI_ERR_MASK 0x00000004L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH3_RD_HI_ERR_MASK 0x00000008L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH4_RD_HI_ERR_MASK 0x00000010L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH5_RD_HI_ERR_MASK 0x00000020L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH6_RD_HI_ERR_MASK 0x00000040L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH7_RD_HI_ERR_MASK 0x00000080L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH0_RD_LO_ERR_MASK 0x00000100L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH1_RD_LO_ERR_MASK 0x00000200L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH2_RD_LO_ERR_MASK 0x00000400L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH3_RD_LO_ERR_MASK 0x00000800L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH4_RD_LO_ERR_MASK 0x00001000L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH5_RD_LO_ERR_MASK 0x00002000L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH6_RD_LO_ERR_MASK 0x00004000L +#define JPEG_MEMCHECK_SYS_INT_STAT__BSFETCH7_RD_LO_ERR_MASK 0x00008000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF0_WR_HI_ERR_MASK 0x00010000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF1_WR_HI_ERR_MASK 0x00020000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF2_WR_HI_ERR_MASK 0x00040000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF3_WR_HI_ERR_MASK 0x00080000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF4_WR_HI_ERR_MASK 0x00100000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF5_WR_HI_ERR_MASK 0x00200000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF6_WR_HI_ERR_MASK 0x00400000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF7_WR_HI_ERR_MASK 0x00800000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF0_WR_LO_ERR_MASK 0x01000000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF1_WR_LO_ERR_MASK 0x02000000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF2_WR_LO_ERR_MASK 0x04000000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF3_WR_LO_ERR_MASK 0x08000000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF4_WR_LO_ERR_MASK 0x10000000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF5_WR_LO_ERR_MASK 0x20000000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF6_WR_LO_ERR_MASK 0x40000000L +#define JPEG_MEMCHECK_SYS_INT_STAT__OBUF7_WR_LO_ERR_MASK 0x80000000L +//JPEG_MEMCHECK_SYS_INT_STAT1 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC0_RD_HI_ERR__SHIFT 0x0 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC1_RD_HI_ERR__SHIFT 0x1 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC2_RD_HI_ERR__SHIFT 0x2 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC3_RD_HI_ERR__SHIFT 0x3 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC4_RD_HI_ERR__SHIFT 0x4 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC5_RD_HI_ERR__SHIFT 0x5 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC6_RD_HI_ERR__SHIFT 0x6 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC7_RD_HI_ERR__SHIFT 0x7 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC0_RD_LO_ERR__SHIFT 0x8 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC1_RD_LO_ERR__SHIFT 0x9 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC2_RD_LO_ERR__SHIFT 0xa +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC3_RD_LO_ERR__SHIFT 0xb +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC4_RD_LO_ERR__SHIFT 0xc +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC5_RD_LO_ERR__SHIFT 0xd +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC6_RD_LO_ERR__SHIFT 0xe +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC7_RD_LO_ERR__SHIFT 0xf +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC0_WR_HI_ERR__SHIFT 0x10 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC1_WR_HI_ERR__SHIFT 0x11 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC2_WR_HI_ERR__SHIFT 0x12 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC3_WR_HI_ERR__SHIFT 0x13 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC4_WR_HI_ERR__SHIFT 0x14 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC5_WR_HI_ERR__SHIFT 0x15 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC6_WR_HI_ERR__SHIFT 0x16 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC7_WR_HI_ERR__SHIFT 0x17 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC0_WR_LO_ERR__SHIFT 0x18 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC1_WR_LO_ERR__SHIFT 0x19 +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC2_WR_LO_ERR__SHIFT 0x1a +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC3_WR_LO_ERR__SHIFT 0x1b +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC4_WR_LO_ERR__SHIFT 0x1c +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC5_WR_LO_ERR__SHIFT 0x1d +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC6_WR_LO_ERR__SHIFT 0x1e +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC7_WR_LO_ERR__SHIFT 0x1f +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC0_RD_HI_ERR_MASK 0x00000001L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC1_RD_HI_ERR_MASK 0x00000002L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC2_RD_HI_ERR_MASK 0x00000004L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC3_RD_HI_ERR_MASK 0x00000008L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC4_RD_HI_ERR_MASK 0x00000010L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC5_RD_HI_ERR_MASK 0x00000020L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC6_RD_HI_ERR_MASK 0x00000040L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC7_RD_HI_ERR_MASK 0x00000080L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC0_RD_LO_ERR_MASK 0x00000100L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC1_RD_LO_ERR_MASK 0x00000200L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC2_RD_LO_ERR_MASK 0x00000400L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC3_RD_LO_ERR_MASK 0x00000800L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC4_RD_LO_ERR_MASK 0x00001000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC5_RD_LO_ERR_MASK 0x00002000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC6_RD_LO_ERR_MASK 0x00004000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC7_RD_LO_ERR_MASK 0x00008000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC0_WR_HI_ERR_MASK 0x00010000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC1_WR_HI_ERR_MASK 0x00020000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC2_WR_HI_ERR_MASK 0x00040000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC3_WR_HI_ERR_MASK 0x00080000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC4_WR_HI_ERR_MASK 0x00100000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC5_WR_HI_ERR_MASK 0x00200000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC6_WR_HI_ERR_MASK 0x00400000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC7_WR_HI_ERR_MASK 0x00800000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC0_WR_LO_ERR_MASK 0x01000000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC1_WR_LO_ERR_MASK 0x02000000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC2_WR_LO_ERR_MASK 0x04000000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC3_WR_LO_ERR_MASK 0x08000000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC4_WR_LO_ERR_MASK 0x10000000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC5_WR_LO_ERR_MASK 0x20000000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC6_WR_LO_ERR_MASK 0x40000000L +#define JPEG_MEMCHECK_SYS_INT_STAT1__DJRBC7_WR_LO_ERR_MASK 0x80000000L +//JPEG_MEMCHECK_SYS_INT_STAT2 +#define JPEG_MEMCHECK_SYS_INT_STAT2__EJRBC_RD_HI_ERR__SHIFT 0x0 +#define JPEG_MEMCHECK_SYS_INT_STAT2__EJRBC_RD_LO_ERR__SHIFT 0x1 +#define JPEG_MEMCHECK_SYS_INT_STAT2__PELFETCH_RD_HI_ERR__SHIFT 0x2 +#define JPEG_MEMCHECK_SYS_INT_STAT2__PELFETCH_RD_LO_ERR__SHIFT 0x3 +#define JPEG_MEMCHECK_SYS_INT_STAT2__SCALAR_RD_HI_ERR__SHIFT 0x4 +#define JPEG_MEMCHECK_SYS_INT_STAT2__SCALAR_RD_LO_ERR__SHIFT 0x5 +#define JPEG_MEMCHECK_SYS_INT_STAT2__EJRBC_WR_HI_ERR__SHIFT 0x6 +#define JPEG_MEMCHECK_SYS_INT_STAT2__EJRBC_WR_LO_ERR__SHIFT 0x7 +#define JPEG_MEMCHECK_SYS_INT_STAT2__BS_WR_HI_ERR__SHIFT 0x8 +#define JPEG_MEMCHECK_SYS_INT_STAT2__BS_WR_LO_ERR__SHIFT 0x9 +#define JPEG_MEMCHECK_SYS_INT_STAT2__SCALAR_WR_HI_ERR__SHIFT 0xa +#define JPEG_MEMCHECK_SYS_INT_STAT2__SCALAR_WR_LO_ERR__SHIFT 0xb +#define JPEG_MEMCHECK_SYS_INT_STAT2__EJRBC_RD_HI_ERR_MASK 0x00000001L +#define JPEG_MEMCHECK_SYS_INT_STAT2__EJRBC_RD_LO_ERR_MASK 0x00000002L +#define JPEG_MEMCHECK_SYS_INT_STAT2__PELFETCH_RD_HI_ERR_MASK 0x00000004L +#define JPEG_MEMCHECK_SYS_INT_STAT2__PELFETCH_RD_LO_ERR_MASK 0x00000008L +#define JPEG_MEMCHECK_SYS_INT_STAT2__SCALAR_RD_HI_ERR_MASK 0x00000010L +#define JPEG_MEMCHECK_SYS_INT_STAT2__SCALAR_RD_LO_ERR_MASK 0x00000020L +#define JPEG_MEMCHECK_SYS_INT_STAT2__EJRBC_WR_HI_ERR_MASK 0x00000040L +#define JPEG_MEMCHECK_SYS_INT_STAT2__EJRBC_WR_LO_ERR_MASK 0x00000080L +#define JPEG_MEMCHECK_SYS_INT_STAT2__BS_WR_HI_ERR_MASK 0x00000100L +#define JPEG_MEMCHECK_SYS_INT_STAT2__BS_WR_LO_ERR_MASK 0x00000200L +#define JPEG_MEMCHECK_SYS_INT_STAT2__SCALAR_WR_HI_ERR_MASK 0x00000400L +#define JPEG_MEMCHECK_SYS_INT_STAT2__SCALAR_WR_LO_ERR_MASK 0x00000800L +//JPEG_MEMCHECK_SYS_INT_ACK +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH0_RD_HI_ERR__SHIFT 0x0 +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH1_RD_HI_ERR__SHIFT 0x1 +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH2_RD_HI_ERR__SHIFT 0x2 +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH3_RD_HI_ERR__SHIFT 0x3 +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH4_RD_HI_ERR__SHIFT 0x4 +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH5_RD_HI_ERR__SHIFT 0x5 +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH6_RD_HI_ERR__SHIFT 0x6 +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH7_RD_HI_ERR__SHIFT 0x7 +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH0_RD_LO_ERR__SHIFT 0x8 +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH1_RD_LO_ERR__SHIFT 0x9 +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH2_RD_LO_ERR__SHIFT 0xa +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH3_RD_LO_ERR__SHIFT 0xb +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH4_RD_LO_ERR__SHIFT 0xc +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH5_RD_LO_ERR__SHIFT 0xd +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH6_RD_LO_ERR__SHIFT 0xe +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH7_RD_LO_ERR__SHIFT 0xf +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF0_WR_HI_ERR__SHIFT 0x10 +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF1_WR_HI_ERR__SHIFT 0x11 +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF2_WR_HI_ERR__SHIFT 0x12 +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF3_WR_HI_ERR__SHIFT 0x13 +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF4_WR_HI_ERR__SHIFT 0x14 +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF5_WR_HI_ERR__SHIFT 0x15 +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF6_WR_HI_ERR__SHIFT 0x16 +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF7_WR_HI_ERR__SHIFT 0x17 +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF0_WR_LO_ERR__SHIFT 0x18 +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF1_WR_LO_ERR__SHIFT 0x19 +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF2_WR_LO_ERR__SHIFT 0x1a +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF3_WR_LO_ERR__SHIFT 0x1b +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF4_WR_LO_ERR__SHIFT 0x1c +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF5_WR_LO_ERR__SHIFT 0x1d +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF6_WR_LO_ERR__SHIFT 0x1e +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF7_WR_LO_ERR__SHIFT 0x1f +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH0_RD_HI_ERR_MASK 0x00000001L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH1_RD_HI_ERR_MASK 0x00000002L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH2_RD_HI_ERR_MASK 0x00000004L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH3_RD_HI_ERR_MASK 0x00000008L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH4_RD_HI_ERR_MASK 0x00000010L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH5_RD_HI_ERR_MASK 0x00000020L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH6_RD_HI_ERR_MASK 0x00000040L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH7_RD_HI_ERR_MASK 0x00000080L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH0_RD_LO_ERR_MASK 0x00000100L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH1_RD_LO_ERR_MASK 0x00000200L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH2_RD_LO_ERR_MASK 0x00000400L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH3_RD_LO_ERR_MASK 0x00000800L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH4_RD_LO_ERR_MASK 0x00001000L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH5_RD_LO_ERR_MASK 0x00002000L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH6_RD_LO_ERR_MASK 0x00004000L +#define JPEG_MEMCHECK_SYS_INT_ACK__BSFETCH7_RD_LO_ERR_MASK 0x00008000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF0_WR_HI_ERR_MASK 0x00010000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF1_WR_HI_ERR_MASK 0x00020000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF2_WR_HI_ERR_MASK 0x00040000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF3_WR_HI_ERR_MASK 0x00080000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF4_WR_HI_ERR_MASK 0x00100000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF5_WR_HI_ERR_MASK 0x00200000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF6_WR_HI_ERR_MASK 0x00400000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF7_WR_HI_ERR_MASK 0x00800000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF0_WR_LO_ERR_MASK 0x01000000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF1_WR_LO_ERR_MASK 0x02000000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF2_WR_LO_ERR_MASK 0x04000000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF3_WR_LO_ERR_MASK 0x08000000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF4_WR_LO_ERR_MASK 0x10000000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF5_WR_LO_ERR_MASK 0x20000000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF6_WR_LO_ERR_MASK 0x40000000L +#define JPEG_MEMCHECK_SYS_INT_ACK__OBUF7_WR_LO_ERR_MASK 0x80000000L +//JPEG_MEMCHECK_SYS_INT_ACK1 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC0_RD_HI_ERR__SHIFT 0x0 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC1_RD_HI_ERR__SHIFT 0x1 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC2_RD_HI_ERR__SHIFT 0x2 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC3_RD_HI_ERR__SHIFT 0x3 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC4_RD_HI_ERR__SHIFT 0x4 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC5_RD_HI_ERR__SHIFT 0x5 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC6_RD_HI_ERR__SHIFT 0x6 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC7_RD_HI_ERR__SHIFT 0x7 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC0_RD_LO_ERR__SHIFT 0x8 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC1_RD_LO_ERR__SHIFT 0x9 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC2_RD_LO_ERR__SHIFT 0xa +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC3_RD_LO_ERR__SHIFT 0xb +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC4_RD_LO_ERR__SHIFT 0xc +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC5_RD_LO_ERR__SHIFT 0xd +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC6_RD_LO_ERR__SHIFT 0xe +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC7_RD_LO_ERR__SHIFT 0xf +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC0_WR_HI_ERR__SHIFT 0x10 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC1_WR_HI_ERR__SHIFT 0x11 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC2_WR_HI_ERR__SHIFT 0x12 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC3_WR_HI_ERR__SHIFT 0x13 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC4_WR_HI_ERR__SHIFT 0x14 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC5_WR_HI_ERR__SHIFT 0x15 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC6_WR_HI_ERR__SHIFT 0x16 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC7_WR_HI_ERR__SHIFT 0x17 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC0_WR_LO_ERR__SHIFT 0x18 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC1_WR_LO_ERR__SHIFT 0x19 +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC2_WR_LO_ERR__SHIFT 0x1a +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC3_WR_LO_ERR__SHIFT 0x1b +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC4_WR_LO_ERR__SHIFT 0x1c +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC5_WR_LO_ERR__SHIFT 0x1d +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC6_WR_LO_ERR__SHIFT 0x1e +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC7_WR_LO_ERR__SHIFT 0x1f +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC0_RD_HI_ERR_MASK 0x00000001L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC1_RD_HI_ERR_MASK 0x00000002L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC2_RD_HI_ERR_MASK 0x00000004L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC3_RD_HI_ERR_MASK 0x00000008L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC4_RD_HI_ERR_MASK 0x00000010L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC5_RD_HI_ERR_MASK 0x00000020L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC6_RD_HI_ERR_MASK 0x00000040L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC7_RD_HI_ERR_MASK 0x00000080L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC0_RD_LO_ERR_MASK 0x00000100L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC1_RD_LO_ERR_MASK 0x00000200L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC2_RD_LO_ERR_MASK 0x00000400L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC3_RD_LO_ERR_MASK 0x00000800L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC4_RD_LO_ERR_MASK 0x00001000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC5_RD_LO_ERR_MASK 0x00002000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC6_RD_LO_ERR_MASK 0x00004000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC7_RD_LO_ERR_MASK 0x00008000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC0_WR_HI_ERR_MASK 0x00010000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC1_WR_HI_ERR_MASK 0x00020000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC2_WR_HI_ERR_MASK 0x00040000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC3_WR_HI_ERR_MASK 0x00080000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC4_WR_HI_ERR_MASK 0x00100000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC5_WR_HI_ERR_MASK 0x00200000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC6_WR_HI_ERR_MASK 0x00400000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC7_WR_HI_ERR_MASK 0x00800000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC0_WR_LO_ERR_MASK 0x01000000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC1_WR_LO_ERR_MASK 0x02000000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC2_WR_LO_ERR_MASK 0x04000000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC3_WR_LO_ERR_MASK 0x08000000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC4_WR_LO_ERR_MASK 0x10000000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC5_WR_LO_ERR_MASK 0x20000000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC6_WR_LO_ERR_MASK 0x40000000L +#define JPEG_MEMCHECK_SYS_INT_ACK1__DJRBC7_WR_LO_ERR_MASK 0x80000000L +//JPEG_MEMCHECK_SYS_INT_ACK2 +#define JPEG_MEMCHECK_SYS_INT_ACK2__EJRBC_RD_HI_ERR__SHIFT 0x0 +#define JPEG_MEMCHECK_SYS_INT_ACK2__EJRBC_RD_LO_ERR__SHIFT 0x1 +#define JPEG_MEMCHECK_SYS_INT_ACK2__PELFETCH_RD_HI_ERR__SHIFT 0x2 +#define JPEG_MEMCHECK_SYS_INT_ACK2__PELFETCH_RD_LO_ERR__SHIFT 0x3 +#define JPEG_MEMCHECK_SYS_INT_ACK2__SCALAR_RD_HI_ERR__SHIFT 0x4 +#define JPEG_MEMCHECK_SYS_INT_ACK2__SCALAR_RD_LO_ERR__SHIFT 0x5 +#define JPEG_MEMCHECK_SYS_INT_ACK2__EJRBC_WR_HI_ERR__SHIFT 0x6 +#define JPEG_MEMCHECK_SYS_INT_ACK2__EJRBC_WR_LO_ERR__SHIFT 0x7 +#define JPEG_MEMCHECK_SYS_INT_ACK2__BS_WR_HI_ERR__SHIFT 0x8 +#define JPEG_MEMCHECK_SYS_INT_ACK2__BS_WR_LO_ERR__SHIFT 0x9 +#define JPEG_MEMCHECK_SYS_INT_ACK2__SCALAR_WR_HI_ERR__SHIFT 0xa +#define JPEG_MEMCHECK_SYS_INT_ACK2__SCALAR_WR_LO_ERR__SHIFT 0xb +#define JPEG_MEMCHECK_SYS_INT_ACK2__EJRBC_RD_HI_ERR_MASK 0x00000001L +#define JPEG_MEMCHECK_SYS_INT_ACK2__EJRBC_RD_LO_ERR_MASK 0x00000002L +#define JPEG_MEMCHECK_SYS_INT_ACK2__PELFETCH_RD_HI_ERR_MASK 0x00000004L +#define JPEG_MEMCHECK_SYS_INT_ACK2__PELFETCH_RD_LO_ERR_MASK 0x00000008L +#define JPEG_MEMCHECK_SYS_INT_ACK2__SCALAR_RD_HI_ERR_MASK 0x00000010L +#define JPEG_MEMCHECK_SYS_INT_ACK2__SCALAR_RD_LO_ERR_MASK 0x00000020L +#define JPEG_MEMCHECK_SYS_INT_ACK2__EJRBC_WR_HI_ERR_MASK 0x00000040L +#define JPEG_MEMCHECK_SYS_INT_ACK2__EJRBC_WR_LO_ERR_MASK 0x00000080L +#define JPEG_MEMCHECK_SYS_INT_ACK2__BS_WR_HI_ERR_MASK 0x00000100L +#define JPEG_MEMCHECK_SYS_INT_ACK2__BS_WR_LO_ERR_MASK 0x00000200L +#define JPEG_MEMCHECK_SYS_INT_ACK2__SCALAR_WR_HI_ERR_MASK 0x00000400L +#define JPEG_MEMCHECK_SYS_INT_ACK2__SCALAR_WR_LO_ERR_MASK 0x00000800L +//JPEG_MASTINT_EN +#define JPEG_MASTINT_EN__OVERRUN_RST__SHIFT 0x0 +#define JPEG_MASTINT_EN__INT_OVERRUN__SHIFT 0x4 +#define JPEG_MASTINT_EN__OVERRUN_RST_MASK 0x00000001L +#define JPEG_MASTINT_EN__INT_OVERRUN_MASK 0x007FFFF0L +//JPEG_IH_CTRL +#define JPEG_IH_CTRL__IH_SOFT_RESET__SHIFT 0x0 +#define JPEG_IH_CTRL__IH_STALL_EN__SHIFT 0x1 +#define JPEG_IH_CTRL__IH_STATUS_CLEAN__SHIFT 0x2 +#define JPEG_IH_CTRL__IH_VMID__SHIFT 0x3 +#define JPEG_IH_CTRL__IH_USER_DATA__SHIFT 0x7 +#define JPEG_IH_CTRL__IH_RINGID__SHIFT 0x13 +#define JPEG_IH_CTRL__IH_SOFT_RESET_MASK 0x00000001L +#define JPEG_IH_CTRL__IH_STALL_EN_MASK 0x00000002L +#define JPEG_IH_CTRL__IH_STATUS_CLEAN_MASK 0x00000004L +#define JPEG_IH_CTRL__IH_VMID_MASK 0x00000078L +#define JPEG_IH_CTRL__IH_USER_DATA_MASK 0x0007FF80L +#define JPEG_IH_CTRL__IH_RINGID_MASK 0x07F80000L +//JRBBM_ARB_CTRL +#define JRBBM_ARB_CTRL__DJRBC0_DROP__SHIFT 0x0 +#define JRBBM_ARB_CTRL__DJRBC1_DROP__SHIFT 0x1 +#define JRBBM_ARB_CTRL__DJRBC2_DROP__SHIFT 0x2 +#define JRBBM_ARB_CTRL__DJRBC3_DROP__SHIFT 0x3 +#define JRBBM_ARB_CTRL__DJRBC4_DROP__SHIFT 0x4 +#define JRBBM_ARB_CTRL__DJRBC5_DROP__SHIFT 0x5 +#define JRBBM_ARB_CTRL__DJRBC6_DROP__SHIFT 0x6 +#define JRBBM_ARB_CTRL__DJRBC7_DROP__SHIFT 0x7 +#define JRBBM_ARB_CTRL__EJRBC_DROP__SHIFT 0x8 +#define JRBBM_ARB_CTRL__SRBM_DROP__SHIFT 0x9 +#define JRBBM_ARB_CTRL__DJRBC0_DROP_MASK 0x00000001L +#define JRBBM_ARB_CTRL__DJRBC1_DROP_MASK 0x00000002L +#define JRBBM_ARB_CTRL__DJRBC2_DROP_MASK 0x00000004L +#define JRBBM_ARB_CTRL__DJRBC3_DROP_MASK 0x00000008L +#define JRBBM_ARB_CTRL__DJRBC4_DROP_MASK 0x00000010L +#define JRBBM_ARB_CTRL__DJRBC5_DROP_MASK 0x00000020L +#define JRBBM_ARB_CTRL__DJRBC6_DROP_MASK 0x00000040L +#define JRBBM_ARB_CTRL__DJRBC7_DROP_MASK 0x00000080L +#define JRBBM_ARB_CTRL__EJRBC_DROP_MASK 0x00000100L +#define JRBBM_ARB_CTRL__SRBM_DROP_MASK 0x00000200L + + +// addressBlock: aid_uvd0_uvd_jpeg_common_sclk_dec +//JPEG_CGC_GATE +#define JPEG_CGC_GATE__JPEG0_DEC__SHIFT 0x0 +#define JPEG_CGC_GATE__JPEG1_DEC__SHIFT 0x1 +#define JPEG_CGC_GATE__JPEG2_DEC__SHIFT 0x2 +#define JPEG_CGC_GATE__JPEG3_DEC__SHIFT 0x3 +#define JPEG_CGC_GATE__JPEG4_DEC__SHIFT 0x4 +#define JPEG_CGC_GATE__JPEG5_DEC__SHIFT 0x5 +#define JPEG_CGC_GATE__JPEG6_DEC__SHIFT 0x6 +#define JPEG_CGC_GATE__JPEG7_DEC__SHIFT 0x7 +#define JPEG_CGC_GATE__JPEG_ENC__SHIFT 0x8 +#define JPEG_CGC_GATE__JMCIF__SHIFT 0x9 +#define JPEG_CGC_GATE__JRBBM__SHIFT 0xa +#define JPEG_CGC_GATE__JPEG0_DEC_MASK 0x00000001L +#define JPEG_CGC_GATE__JPEG1_DEC_MASK 0x00000002L +#define JPEG_CGC_GATE__JPEG2_DEC_MASK 0x00000004L +#define JPEG_CGC_GATE__JPEG3_DEC_MASK 0x00000008L +#define JPEG_CGC_GATE__JPEG4_DEC_MASK 0x00000010L +#define JPEG_CGC_GATE__JPEG5_DEC_MASK 0x00000020L +#define JPEG_CGC_GATE__JPEG6_DEC_MASK 0x00000040L +#define JPEG_CGC_GATE__JPEG7_DEC_MASK 0x00000080L +#define JPEG_CGC_GATE__JPEG_ENC_MASK 0x00000100L +#define JPEG_CGC_GATE__JMCIF_MASK 0x00000200L +#define JPEG_CGC_GATE__JRBBM_MASK 0x00000400L +//JPEG_CGC_CTRL +#define JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT 0x0 +#define JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT 0x1 +#define JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT 0x5 +#define JPEG_CGC_CTRL__JPEG0_DEC_MODE__SHIFT 0x10 +#define JPEG_CGC_CTRL__JPEG1_DEC_MODE__SHIFT 0x11 +#define JPEG_CGC_CTRL__JPEG2_DEC_MODE__SHIFT 0x12 +#define JPEG_CGC_CTRL__JPEG3_DEC_MODE__SHIFT 0x13 +#define JPEG_CGC_CTRL__JPEG4_DEC_MODE__SHIFT 0x14 +#define JPEG_CGC_CTRL__JPEG5_DEC_MODE__SHIFT 0x15 +#define JPEG_CGC_CTRL__JPEG6_DEC_MODE__SHIFT 0x16 +#define JPEG_CGC_CTRL__JPEG7_DEC_MODE__SHIFT 0x17 +#define JPEG_CGC_CTRL__JPEG_ENC_MODE__SHIFT 0x18 +#define JPEG_CGC_CTRL__JMCIF_MODE__SHIFT 0x19 +#define JPEG_CGC_CTRL__JRBBM_MODE__SHIFT 0x1a +#define JPEG_CGC_CTRL__DYN_CLOCK_MODE_MASK 0x00000001L +#define JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER_MASK 0x0000001EL +#define JPEG_CGC_CTRL__CLK_OFF_DELAY_MASK 0x00001FE0L +#define JPEG_CGC_CTRL__JPEG0_DEC_MODE_MASK 0x00010000L +#define JPEG_CGC_CTRL__JPEG1_DEC_MODE_MASK 0x00020000L +#define JPEG_CGC_CTRL__JPEG2_DEC_MODE_MASK 0x00040000L +#define JPEG_CGC_CTRL__JPEG3_DEC_MODE_MASK 0x00080000L +#define JPEG_CGC_CTRL__JPEG4_DEC_MODE_MASK 0x00100000L +#define JPEG_CGC_CTRL__JPEG5_DEC_MODE_MASK 0x00200000L +#define JPEG_CGC_CTRL__JPEG6_DEC_MODE_MASK 0x00400000L +#define JPEG_CGC_CTRL__JPEG7_DEC_MODE_MASK 0x00800000L +#define JPEG_CGC_CTRL__JPEG_ENC_MODE_MASK 0x01000000L +#define JPEG_CGC_CTRL__JMCIF_MODE_MASK 0x02000000L +#define JPEG_CGC_CTRL__JRBBM_MODE_MASK 0x04000000L +//JPEG_CGC_STATUS +#define JPEG_CGC_STATUS__JPEG0_DEC_VCLK_ACTIVE__SHIFT 0x0 +#define JPEG_CGC_STATUS__JPEG0_DEC_SCLK_ACTIVE__SHIFT 0x1 +#define JPEG_CGC_STATUS__JPEG1_DEC_VCLK_ACTIVE__SHIFT 0x2 +#define JPEG_CGC_STATUS__JPEG1_DEC_SCLK_ACTIVE__SHIFT 0x3 +#define JPEG_CGC_STATUS__JPEG2_DEC_VCLK_ACTIVE__SHIFT 0x4 +#define JPEG_CGC_STATUS__JPEG2_DEC_SCLK_ACTIVE__SHIFT 0x5 +#define JPEG_CGC_STATUS__JPEG3_DEC_VCLK_ACTIVE__SHIFT 0x6 +#define JPEG_CGC_STATUS__JPEG3_DEC_SCLK_ACTIVE__SHIFT 0x7 +#define JPEG_CGC_STATUS__JPEG4_DEC_VCLK_ACTIVE__SHIFT 0x8 +#define JPEG_CGC_STATUS__JPEG4_DEC_SCLK_ACTIVE__SHIFT 0x9 +#define JPEG_CGC_STATUS__JPEG5_DEC_VCLK_ACTIVE__SHIFT 0xa +#define JPEG_CGC_STATUS__JPEG5_DEC_SCLK_ACTIVE__SHIFT 0xb +#define JPEG_CGC_STATUS__JPEG6_DEC_VCLK_ACTIVE__SHIFT 0xc +#define JPEG_CGC_STATUS__JPEG6_DEC_SCLK_ACTIVE__SHIFT 0xd +#define JPEG_CGC_STATUS__JPEG7_DEC_VCLK_ACTIVE__SHIFT 0xe +#define JPEG_CGC_STATUS__JPEG7_DEC_SCLK_ACTIVE__SHIFT 0xf +#define JPEG_CGC_STATUS__JPEG_ENC_VCLK_ACTIVE__SHIFT 0x10 +#define JPEG_CGC_STATUS__JPEG_ENC_SCLK_ACTIVE__SHIFT 0x11 +#define JPEG_CGC_STATUS__JMCIF_SCLK_ACTIVE__SHIFT 0x12 +#define JPEG_CGC_STATUS__JRBBM_VCLK_ACTIVE__SHIFT 0x13 +#define JPEG_CGC_STATUS__JRBBM_SCLK_ACTIVE__SHIFT 0x14 +#define JPEG_CGC_STATUS__JPEG0_DEC_VCLK_ACTIVE_MASK 0x00000001L +#define JPEG_CGC_STATUS__JPEG0_DEC_SCLK_ACTIVE_MASK 0x00000002L +#define JPEG_CGC_STATUS__JPEG1_DEC_VCLK_ACTIVE_MASK 0x00000004L +#define JPEG_CGC_STATUS__JPEG1_DEC_SCLK_ACTIVE_MASK 0x00000008L +#define JPEG_CGC_STATUS__JPEG2_DEC_VCLK_ACTIVE_MASK 0x00000010L +#define JPEG_CGC_STATUS__JPEG2_DEC_SCLK_ACTIVE_MASK 0x00000020L +#define JPEG_CGC_STATUS__JPEG3_DEC_VCLK_ACTIVE_MASK 0x00000040L +#define JPEG_CGC_STATUS__JPEG3_DEC_SCLK_ACTIVE_MASK 0x00000080L +#define JPEG_CGC_STATUS__JPEG4_DEC_VCLK_ACTIVE_MASK 0x00000100L +#define JPEG_CGC_STATUS__JPEG4_DEC_SCLK_ACTIVE_MASK 0x00000200L +#define JPEG_CGC_STATUS__JPEG5_DEC_VCLK_ACTIVE_MASK 0x00000400L +#define JPEG_CGC_STATUS__JPEG5_DEC_SCLK_ACTIVE_MASK 0x00000800L +#define JPEG_CGC_STATUS__JPEG6_DEC_VCLK_ACTIVE_MASK 0x00001000L +#define JPEG_CGC_STATUS__JPEG6_DEC_SCLK_ACTIVE_MASK 0x00002000L +#define JPEG_CGC_STATUS__JPEG7_DEC_VCLK_ACTIVE_MASK 0x00004000L +#define JPEG_CGC_STATUS__JPEG7_DEC_SCLK_ACTIVE_MASK 0x00008000L +#define JPEG_CGC_STATUS__JPEG_ENC_VCLK_ACTIVE_MASK 0x00010000L +#define JPEG_CGC_STATUS__JPEG_ENC_SCLK_ACTIVE_MASK 0x00020000L +#define JPEG_CGC_STATUS__JMCIF_SCLK_ACTIVE_MASK 0x00040000L +#define JPEG_CGC_STATUS__JRBBM_VCLK_ACTIVE_MASK 0x00080000L +#define JPEG_CGC_STATUS__JRBBM_SCLK_ACTIVE_MASK 0x00100000L +//JPEG_COMN_CGC_MEM_CTRL +#define JPEG_COMN_CGC_MEM_CTRL__JMCIF_LS_EN__SHIFT 0x0 +#define JPEG_COMN_CGC_MEM_CTRL__JMCIF_DS_EN__SHIFT 0x1 +#define JPEG_COMN_CGC_MEM_CTRL__JMCIF_SD_EN__SHIFT 0x2 +#define JPEG_COMN_CGC_MEM_CTRL__JMCIF_LS_SW_EN__SHIFT 0x3 +#define JPEG_COMN_CGC_MEM_CTRL__JMCIF_LS_EN_MASK 0x00000001L +#define JPEG_COMN_CGC_MEM_CTRL__JMCIF_DS_EN_MASK 0x00000002L +#define JPEG_COMN_CGC_MEM_CTRL__JMCIF_SD_EN_MASK 0x00000004L +#define JPEG_COMN_CGC_MEM_CTRL__JMCIF_LS_SW_EN_MASK 0x00000008L +//JPEG_DEC_CGC_MEM_CTRL +#define JPEG_DEC_CGC_MEM_CTRL__JPEG0_DEC_LS_EN__SHIFT 0x0 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG0_DEC_DS_EN__SHIFT 0x1 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG0_DEC_SD_EN__SHIFT 0x2 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG0_DEC_LS_SW_EN__SHIFT 0x3 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG1_DEC_LS_EN__SHIFT 0x4 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG1_DEC_DS_EN__SHIFT 0x5 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG1_DEC_SD_EN__SHIFT 0x6 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG1_DEC_LS_SW_EN__SHIFT 0x7 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG2_DEC_LS_EN__SHIFT 0x8 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG2_DEC_DS_EN__SHIFT 0x9 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG2_DEC_SD_EN__SHIFT 0xa +#define JPEG_DEC_CGC_MEM_CTRL__JPEG2_DEC_LS_SW_EN__SHIFT 0xb +#define JPEG_DEC_CGC_MEM_CTRL__JPEG3_DEC_LS_EN__SHIFT 0xc +#define JPEG_DEC_CGC_MEM_CTRL__JPEG3_DEC_DS_EN__SHIFT 0xd +#define JPEG_DEC_CGC_MEM_CTRL__JPEG3_DEC_SD_EN__SHIFT 0xe +#define JPEG_DEC_CGC_MEM_CTRL__JPEG3_DEC_LS_SW_EN__SHIFT 0xf +#define JPEG_DEC_CGC_MEM_CTRL__JPEG4_DEC_LS_EN__SHIFT 0x10 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG4_DEC_DS_EN__SHIFT 0x11 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG4_DEC_SD_EN__SHIFT 0x12 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG4_DEC_LS_SW_EN__SHIFT 0x13 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG5_DEC_LS_EN__SHIFT 0x14 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG5_DEC_DS_EN__SHIFT 0x15 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG5_DEC_SD_EN__SHIFT 0x16 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG5_DEC_LS_SW_EN__SHIFT 0x17 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG6_DEC_LS_EN__SHIFT 0x18 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG6_DEC_DS_EN__SHIFT 0x19 +#define JPEG_DEC_CGC_MEM_CTRL__JPEG6_DEC_SD_EN__SHIFT 0x1a +#define JPEG_DEC_CGC_MEM_CTRL__JPEG6_DEC_LS_SW_EN__SHIFT 0x1b +#define JPEG_DEC_CGC_MEM_CTRL__JPEG7_DEC_LS_EN__SHIFT 0x1c +#define JPEG_DEC_CGC_MEM_CTRL__JPEG7_DEC_DS_EN__SHIFT 0x1d +#define JPEG_DEC_CGC_MEM_CTRL__JPEG7_DEC_SD_EN__SHIFT 0x1e +#define JPEG_DEC_CGC_MEM_CTRL__JPEG7_DEC_LS_SW_EN__SHIFT 0x1f +#define JPEG_DEC_CGC_MEM_CTRL__JPEG0_DEC_LS_EN_MASK 0x00000001L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG0_DEC_DS_EN_MASK 0x00000002L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG0_DEC_SD_EN_MASK 0x00000004L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG0_DEC_LS_SW_EN_MASK 0x00000008L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG1_DEC_LS_EN_MASK 0x00000010L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG1_DEC_DS_EN_MASK 0x00000020L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG1_DEC_SD_EN_MASK 0x00000040L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG1_DEC_LS_SW_EN_MASK 0x00000080L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG2_DEC_LS_EN_MASK 0x00000100L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG2_DEC_DS_EN_MASK 0x00000200L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG2_DEC_SD_EN_MASK 0x00000400L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG2_DEC_LS_SW_EN_MASK 0x00000800L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG3_DEC_LS_EN_MASK 0x00001000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG3_DEC_DS_EN_MASK 0x00002000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG3_DEC_SD_EN_MASK 0x00004000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG3_DEC_LS_SW_EN_MASK 0x00008000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG4_DEC_LS_EN_MASK 0x00010000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG4_DEC_DS_EN_MASK 0x00020000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG4_DEC_SD_EN_MASK 0x00040000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG4_DEC_LS_SW_EN_MASK 0x00080000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG5_DEC_LS_EN_MASK 0x00100000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG5_DEC_DS_EN_MASK 0x00200000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG5_DEC_SD_EN_MASK 0x00400000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG5_DEC_LS_SW_EN_MASK 0x00800000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG6_DEC_LS_EN_MASK 0x01000000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG6_DEC_DS_EN_MASK 0x02000000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG6_DEC_SD_EN_MASK 0x04000000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG6_DEC_LS_SW_EN_MASK 0x08000000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG7_DEC_LS_EN_MASK 0x10000000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG7_DEC_DS_EN_MASK 0x20000000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG7_DEC_SD_EN_MASK 0x40000000L +#define JPEG_DEC_CGC_MEM_CTRL__JPEG7_DEC_LS_SW_EN_MASK 0x80000000L +//JPEG_ENC_CGC_MEM_CTRL +#define JPEG_ENC_CGC_MEM_CTRL__JPEG_ENC_LS_EN__SHIFT 0x0 +#define JPEG_ENC_CGC_MEM_CTRL__JPEG_ENC_DS_EN__SHIFT 0x1 +#define JPEG_ENC_CGC_MEM_CTRL__JPEG_ENC_SD_EN__SHIFT 0x2 +#define JPEG_ENC_CGC_MEM_CTRL__JPEG_ENC_LS_SW_EN__SHIFT 0x3 +#define JPEG_ENC_CGC_MEM_CTRL__JPEG_ENC_LS_EN_MASK 0x00000001L +#define JPEG_ENC_CGC_MEM_CTRL__JPEG_ENC_DS_EN_MASK 0x00000002L +#define JPEG_ENC_CGC_MEM_CTRL__JPEG_ENC_SD_EN_MASK 0x00000004L +#define JPEG_ENC_CGC_MEM_CTRL__JPEG_ENC_LS_SW_EN_MASK 0x00000008L +//JPEG_PERF_BANK_CONF +#define JPEG_PERF_BANK_CONF__RESET__SHIFT 0x0 +#define JPEG_PERF_BANK_CONF__PEEK__SHIFT 0x8 +#define JPEG_PERF_BANK_CONF__CONCATENATE__SHIFT 0x10 +#define JPEG_PERF_BANK_CONF__CORE_SEL__SHIFT 0x15 +#define JPEG_PERF_BANK_CONF__RESET_MASK 0x0000000FL +#define JPEG_PERF_BANK_CONF__PEEK_MASK 0x00000F00L +#define JPEG_PERF_BANK_CONF__CONCATENATE_MASK 0x00030000L +#define JPEG_PERF_BANK_CONF__CORE_SEL_MASK 0x00E00000L +//JPEG_PERF_BANK_EVENT_SEL +#define JPEG_PERF_BANK_EVENT_SEL__SEL0__SHIFT 0x0 +#define JPEG_PERF_BANK_EVENT_SEL__SEL1__SHIFT 0x8 +#define JPEG_PERF_BANK_EVENT_SEL__SEL2__SHIFT 0x10 +#define JPEG_PERF_BANK_EVENT_SEL__SEL3__SHIFT 0x18 +#define JPEG_PERF_BANK_EVENT_SEL__SEL0_MASK 0x000000FFL +#define JPEG_PERF_BANK_EVENT_SEL__SEL1_MASK 0x0000FF00L +#define JPEG_PERF_BANK_EVENT_SEL__SEL2_MASK 0x00FF0000L +#define JPEG_PERF_BANK_EVENT_SEL__SEL3_MASK 0xFF000000L +//JPEG_PERF_BANK_COUNT0 +#define JPEG_PERF_BANK_COUNT0__COUNT__SHIFT 0x0 +#define JPEG_PERF_BANK_COUNT0__COUNT_MASK 0xFFFFFFFFL +//JPEG_PERF_BANK_COUNT1 +#define JPEG_PERF_BANK_COUNT1__COUNT__SHIFT 0x0 +#define JPEG_PERF_BANK_COUNT1__COUNT_MASK 0xFFFFFFFFL +//JPEG_PERF_BANK_COUNT2 +#define JPEG_PERF_BANK_COUNT2__COUNT__SHIFT 0x0 +#define JPEG_PERF_BANK_COUNT2__COUNT_MASK 0xFFFFFFFFL +//JPEG_PERF_BANK_COUNT3 +#define JPEG_PERF_BANK_COUNT3__COUNT__SHIFT 0x0 +#define JPEG_PERF_BANK_COUNT3__COUNT_MASK 0xFFFFFFFFL + + +// addressBlock: aid_uvd0_uvd_pg_dec +//UVD_PGFSM_CONFIG +#define UVD_PGFSM_CONFIG__UVDM_PWR_CONFIG__SHIFT 0x0 +#define UVD_PGFSM_CONFIG__UVDS_PWR_CONFIG__SHIFT 0x2 +#define UVD_PGFSM_CONFIG__UVDF_PWR_CONFIG__SHIFT 0x4 +#define UVD_PGFSM_CONFIG__UVDTC_PWR_CONFIG__SHIFT 0x6 +#define UVD_PGFSM_CONFIG__UVDB_PWR_CONFIG__SHIFT 0x8 +#define UVD_PGFSM_CONFIG__UVDTA_PWR_CONFIG__SHIFT 0xa +#define UVD_PGFSM_CONFIG__UVDLM_PWR_CONFIG__SHIFT 0xc +#define UVD_PGFSM_CONFIG__UVDTD_PWR_CONFIG__SHIFT 0xe +#define UVD_PGFSM_CONFIG__UVDTE_PWR_CONFIG__SHIFT 0x10 +#define UVD_PGFSM_CONFIG__UVDE_PWR_CONFIG__SHIFT 0x12 +#define UVD_PGFSM_CONFIG__UVDAB_PWR_CONFIG__SHIFT 0x14 +#define UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG__SHIFT 0x16 +#define UVD_PGFSM_CONFIG__UVDTB_PWR_CONFIG__SHIFT 0x18 +#define UVD_PGFSM_CONFIG__UVDNA_PWR_CONFIG__SHIFT 0x1a +#define UVD_PGFSM_CONFIG__UVDNB_PWR_CONFIG__SHIFT 0x1c +#define UVD_PGFSM_CONFIG__UVDM_PWR_CONFIG_MASK 0x00000003L +#define UVD_PGFSM_CONFIG__UVDS_PWR_CONFIG_MASK 0x0000000CL +#define UVD_PGFSM_CONFIG__UVDF_PWR_CONFIG_MASK 0x00000030L +#define UVD_PGFSM_CONFIG__UVDTC_PWR_CONFIG_MASK 0x000000C0L +#define UVD_PGFSM_CONFIG__UVDB_PWR_CONFIG_MASK 0x00000300L +#define UVD_PGFSM_CONFIG__UVDTA_PWR_CONFIG_MASK 0x00000C00L +#define UVD_PGFSM_CONFIG__UVDLM_PWR_CONFIG_MASK 0x00003000L +#define UVD_PGFSM_CONFIG__UVDTD_PWR_CONFIG_MASK 0x0000C000L +#define UVD_PGFSM_CONFIG__UVDTE_PWR_CONFIG_MASK 0x00030000L +#define UVD_PGFSM_CONFIG__UVDE_PWR_CONFIG_MASK 0x000C0000L +#define UVD_PGFSM_CONFIG__UVDAB_PWR_CONFIG_MASK 0x00300000L +#define UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG_MASK 0x00C00000L +#define UVD_PGFSM_CONFIG__UVDTB_PWR_CONFIG_MASK 0x03000000L +#define UVD_PGFSM_CONFIG__UVDNA_PWR_CONFIG_MASK 0x0C000000L +#define UVD_PGFSM_CONFIG__UVDNB_PWR_CONFIG_MASK 0x30000000L +//UVD_PGFSM_STATUS +#define UVD_PGFSM_STATUS__UVDM_PWR_STATUS__SHIFT 0x0 +#define UVD_PGFSM_STATUS__UVDS_PWR_STATUS__SHIFT 0x2 +#define UVD_PGFSM_STATUS__UVDF_PWR_STATUS__SHIFT 0x4 +#define UVD_PGFSM_STATUS__UVDTC_PWR_STATUS__SHIFT 0x6 +#define UVD_PGFSM_STATUS__UVDB_PWR_STATUS__SHIFT 0x8 +#define UVD_PGFSM_STATUS__UVDTA_PWR_STATUS__SHIFT 0xa +#define UVD_PGFSM_STATUS__UVDLM_PWR_STATUS__SHIFT 0xc +#define UVD_PGFSM_STATUS__UVDTD_PWR_STATUS__SHIFT 0xe +#define UVD_PGFSM_STATUS__UVDTE_PWR_STATUS__SHIFT 0x10 +#define UVD_PGFSM_STATUS__UVDE_PWR_STATUS__SHIFT 0x12 +#define UVD_PGFSM_STATUS__UVDAB_PWR_STATUS__SHIFT 0x14 +#define UVD_PGFSM_STATUS__UVDJ_PWR_STATUS__SHIFT 0x16 +#define UVD_PGFSM_STATUS__UVDTB_PWR_STATUS__SHIFT 0x18 +#define UVD_PGFSM_STATUS__UVDNA_PWR_STATUS__SHIFT 0x1a +#define UVD_PGFSM_STATUS__UVDNB_PWR_STATUS__SHIFT 0x1c +#define UVD_PGFSM_STATUS__UVDM_PWR_STATUS_MASK 0x00000003L +#define UVD_PGFSM_STATUS__UVDS_PWR_STATUS_MASK 0x0000000CL +#define UVD_PGFSM_STATUS__UVDF_PWR_STATUS_MASK 0x00000030L +#define UVD_PGFSM_STATUS__UVDTC_PWR_STATUS_MASK 0x000000C0L +#define UVD_PGFSM_STATUS__UVDB_PWR_STATUS_MASK 0x00000300L +#define UVD_PGFSM_STATUS__UVDTA_PWR_STATUS_MASK 0x00000C00L +#define UVD_PGFSM_STATUS__UVDLM_PWR_STATUS_MASK 0x00003000L +#define UVD_PGFSM_STATUS__UVDTD_PWR_STATUS_MASK 0x0000C000L +#define UVD_PGFSM_STATUS__UVDTE_PWR_STATUS_MASK 0x00030000L +#define UVD_PGFSM_STATUS__UVDE_PWR_STATUS_MASK 0x000C0000L +#define UVD_PGFSM_STATUS__UVDAB_PWR_STATUS_MASK 0x00300000L +#define UVD_PGFSM_STATUS__UVDJ_PWR_STATUS_MASK 0x00C00000L +#define UVD_PGFSM_STATUS__UVDTB_PWR_STATUS_MASK 0x03000000L +#define UVD_PGFSM_STATUS__UVDNA_PWR_STATUS_MASK 0x0C000000L +#define UVD_PGFSM_STATUS__UVDNB_PWR_STATUS_MASK 0x30000000L +//UVD_POWER_STATUS +#define UVD_POWER_STATUS__UVD_POWER_STATUS__SHIFT 0x0 +#define UVD_POWER_STATUS__UVD_PG_MODE__SHIFT 0x2 +#define UVD_POWER_STATUS__UVD_CG_MODE__SHIFT 0x4 +#define UVD_POWER_STATUS__UVD_PG_EN__SHIFT 0x8 +#define UVD_POWER_STATUS__RBC_SNOOP_DIS__SHIFT 0x9 +#define UVD_POWER_STATUS__SW_RB_SNOOP_DIS__SHIFT 0xb +#define UVD_POWER_STATUS__STALL_DPG_POWER_UP__SHIFT 0x1f +#define UVD_POWER_STATUS__UVD_POWER_STATUS_MASK 0x00000003L +#define UVD_POWER_STATUS__UVD_PG_MODE_MASK 0x00000004L +#define UVD_POWER_STATUS__UVD_CG_MODE_MASK 0x00000030L +#define UVD_POWER_STATUS__UVD_PG_EN_MASK 0x00000100L +#define UVD_POWER_STATUS__RBC_SNOOP_DIS_MASK 0x00000200L +#define UVD_POWER_STATUS__SW_RB_SNOOP_DIS_MASK 0x00000800L +#define UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK 0x80000000L +//UVD_JPEG_POWER_STATUS +#define UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS__SHIFT 0x0 +#define UVD_JPEG_POWER_STATUS__JPEG_PG_MODE__SHIFT 0x4 +#define UVD_JPEG_POWER_STATUS__JRBC_DEC_SNOOP_DIS__SHIFT 0x8 +#define UVD_JPEG_POWER_STATUS__JRBC_ENC_SNOOP_DIS__SHIFT 0x9 +#define UVD_JPEG_POWER_STATUS__STALL_JDPG_POWER_UP__SHIFT 0x1f +#define UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK 0x00000001L +#define UVD_JPEG_POWER_STATUS__JPEG_PG_MODE_MASK 0x00000010L +#define UVD_JPEG_POWER_STATUS__JRBC_DEC_SNOOP_DIS_MASK 0x00000100L +#define UVD_JPEG_POWER_STATUS__JRBC_ENC_SNOOP_DIS_MASK 0x00000200L +#define UVD_JPEG_POWER_STATUS__STALL_JDPG_POWER_UP_MASK 0x80000000L +//UVD_MC_DJPEG_RD_SPACE +#define UVD_MC_DJPEG_RD_SPACE__DJPEG_RD_SPACE__SHIFT 0x0 +#define UVD_MC_DJPEG_RD_SPACE__DJPEG_RD_SPACE_MASK 0x0003FFFFL +//UVD_MC_DJPEG_WR_SPACE +#define UVD_MC_DJPEG_WR_SPACE__DJPEG_WR_SPACE__SHIFT 0x0 +#define UVD_MC_DJPEG_WR_SPACE__DJPEG_WR_SPACE_MASK 0x0003FFFFL +//UVD_MC_EJPEG_RD_SPACE +#define UVD_MC_EJPEG_RD_SPACE__EJPEG_RD_SPACE__SHIFT 0x0 +#define UVD_MC_EJPEG_RD_SPACE__EJPEG_RD_SPACE_MASK 0x0003FFFFL +//UVD_MC_EJPEG_WR_SPACE +#define UVD_MC_EJPEG_WR_SPACE__EJPEG_WR_SPACE__SHIFT 0x0 +#define UVD_MC_EJPEG_WR_SPACE__EJPEG_WR_SPACE_MASK 0x0003FFFFL +//UVD_PG_IND_INDEX +#define UVD_PG_IND_INDEX__INDEX__SHIFT 0x0 +#define UVD_PG_IND_INDEX__INDEX_MASK 0x0000003FL +//UVD_PG_IND_DATA +#define UVD_PG_IND_DATA__DATA__SHIFT 0x0 +#define UVD_PG_IND_DATA__DATA_MASK 0xFFFFFFFFL +//CC_UVD_HARVESTING +#define CC_UVD_HARVESTING__MMSCH_DISABLE__SHIFT 0x0 +#define CC_UVD_HARVESTING__UVD_DISABLE__SHIFT 0x1 +#define CC_UVD_HARVESTING__MMSCH_DISABLE_MASK 0x00000001L +#define CC_UVD_HARVESTING__UVD_DISABLE_MASK 0x00000002L +//UVD_DPG_LMA_CTL +#define UVD_DPG_LMA_CTL__READ_WRITE__SHIFT 0x0 +#define UVD_DPG_LMA_CTL__MASK_EN__SHIFT 0x1 +#define UVD_DPG_LMA_CTL__ADDR_AUTO_INCREMENT__SHIFT 0x2 +#define UVD_DPG_LMA_CTL__SRAM_SEL__SHIFT 0x4 +#define UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT 0x10 +#define UVD_DPG_LMA_CTL__READ_WRITE_MASK 0x00000001L +#define UVD_DPG_LMA_CTL__MASK_EN_MASK 0x00000002L +#define UVD_DPG_LMA_CTL__ADDR_AUTO_INCREMENT_MASK 0x00000004L +#define UVD_DPG_LMA_CTL__SRAM_SEL_MASK 0x00000010L +#define UVD_DPG_LMA_CTL__READ_WRITE_ADDR_MASK 0xFFFF0000L +//UVD_DPG_LMA_DATA +#define UVD_DPG_LMA_DATA__LMA_DATA__SHIFT 0x0 +#define UVD_DPG_LMA_DATA__LMA_DATA_MASK 0xFFFFFFFFL +//UVD_DPG_LMA_MASK +#define UVD_DPG_LMA_MASK__LMA_MASK__SHIFT 0x0 +#define UVD_DPG_LMA_MASK__LMA_MASK_MASK 0xFFFFFFFFL +//UVD_DPG_PAUSE +#define UVD_DPG_PAUSE__JPEG_PAUSE_DPG_REQ__SHIFT 0x0 +#define UVD_DPG_PAUSE__JPEG_PAUSE_DPG_ACK__SHIFT 0x1 +#define UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ__SHIFT 0x2 +#define UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK__SHIFT 0x3 +#define UVD_DPG_PAUSE__JPEG_PAUSE_DPG_REQ_MASK 0x00000001L +#define UVD_DPG_PAUSE__JPEG_PAUSE_DPG_ACK_MASK 0x00000002L +#define UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK 0x00000004L +#define UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK 0x00000008L +//UVD_SCRATCH1 +#define UVD_SCRATCH1__SCRATCH1_DATA__SHIFT 0x0 +#define UVD_SCRATCH1__SCRATCH1_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH2 +#define UVD_SCRATCH2__SCRATCH2_DATA__SHIFT 0x0 +#define UVD_SCRATCH2__SCRATCH2_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH3 +#define UVD_SCRATCH3__SCRATCH3_DATA__SHIFT 0x0 +#define UVD_SCRATCH3__SCRATCH3_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH4 +#define UVD_SCRATCH4__SCRATCH4_DATA__SHIFT 0x0 +#define UVD_SCRATCH4__SCRATCH4_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH5 +#define UVD_SCRATCH5__SCRATCH5_DATA__SHIFT 0x0 +#define UVD_SCRATCH5__SCRATCH5_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH6 +#define UVD_SCRATCH6__SCRATCH6_DATA__SHIFT 0x0 +#define UVD_SCRATCH6__SCRATCH6_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH7 +#define UVD_SCRATCH7__SCRATCH7_DATA__SHIFT 0x0 +#define UVD_SCRATCH7__SCRATCH7_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH8 +#define UVD_SCRATCH8__SCRATCH8_DATA__SHIFT 0x0 +#define UVD_SCRATCH8__SCRATCH8_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH9 +#define UVD_SCRATCH9__SCRATCH9_DATA__SHIFT 0x0 +#define UVD_SCRATCH9__SCRATCH9_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH10 +#define UVD_SCRATCH10__SCRATCH10_DATA__SHIFT 0x0 +#define UVD_SCRATCH10__SCRATCH10_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH11 +#define UVD_SCRATCH11__SCRATCH11_DATA__SHIFT 0x0 +#define UVD_SCRATCH11__SCRATCH11_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH12 +#define UVD_SCRATCH12__SCRATCH12_DATA__SHIFT 0x0 +#define UVD_SCRATCH12__SCRATCH12_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH13 +#define UVD_SCRATCH13__SCRATCH13_DATA__SHIFT 0x0 +#define UVD_SCRATCH13__SCRATCH13_DATA_MASK 0xFFFFFFFFL +//UVD_SCRATCH14 +#define UVD_SCRATCH14__SCRATCH14_DATA__SHIFT 0x0 +#define UVD_SCRATCH14__SCRATCH14_DATA_MASK 0xFFFFFFFFL +//UVD_FREE_COUNTER_REG +#define UVD_FREE_COUNTER_REG__FREE_COUNTER__SHIFT 0x0 +#define UVD_FREE_COUNTER_REG__FREE_COUNTER_MASK 0xFFFFFFFFL +//UVD_DPG_LMI_VCPU_CACHE_64BIT_BAR_LOW +#define UVD_DPG_LMI_VCPU_CACHE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_DPG_LMI_VCPU_CACHE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_DPG_LMI_VCPU_CACHE_64BIT_BAR_HIGH +#define UVD_DPG_LMI_VCPU_CACHE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_DPG_LMI_VCPU_CACHE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_DPG_VCPU_CACHE_OFFSET0 +#define UVD_DPG_VCPU_CACHE_OFFSET0__CACHE_OFFSET0__SHIFT 0x0 +#define UVD_DPG_VCPU_CACHE_OFFSET0__CACHE_OFFSET0_MASK 0x01FFFFFFL +//UVD_DPG_LMI_VCPU_CACHE_VMID +#define UVD_DPG_LMI_VCPU_CACHE_VMID__VCPU_CACHE_VMID__SHIFT 0x0 +#define UVD_DPG_LMI_VCPU_CACHE_VMID__VCPU_CACHE_VMID_MASK 0x0000000FL +//UVD_REG_FILTER_EN +#define UVD_REG_FILTER_EN__UVD_REG_FILTER_EN__SHIFT 0x0 +#define UVD_REG_FILTER_EN__MMSCH_HI_PRIV__SHIFT 0x1 +#define UVD_REG_FILTER_EN__VIDEO_PRIV_EN__SHIFT 0x2 +#define UVD_REG_FILTER_EN__JPEG_PRIV_EN__SHIFT 0x3 +#define UVD_REG_FILTER_EN__UVD_REG_FILTER_EN_MASK 0x00000001L +#define UVD_REG_FILTER_EN__MMSCH_HI_PRIV_MASK 0x00000002L +#define UVD_REG_FILTER_EN__VIDEO_PRIV_EN_MASK 0x00000004L +#define UVD_REG_FILTER_EN__JPEG_PRIV_EN_MASK 0x00000008L +//UVD_SECURITY_REG_VIO_REPORT +#define UVD_SECURITY_REG_VIO_REPORT__HOST_REG_VIO__SHIFT 0x0 +#define UVD_SECURITY_REG_VIO_REPORT__VCPU_REG_VIO__SHIFT 0x1 +#define UVD_SECURITY_REG_VIO_REPORT__VIDEO_REG_VIO__SHIFT 0x2 +#define UVD_SECURITY_REG_VIO_REPORT__DPG_REG_VIO__SHIFT 0x3 +#define UVD_SECURITY_REG_VIO_REPORT__JPEG_REG_VIO__SHIFT 0x4 +#define UVD_SECURITY_REG_VIO_REPORT__JDPG_REG_VIO__SHIFT 0x5 +#define UVD_SECURITY_REG_VIO_REPORT__HOST_REG_VIO_MASK 0x00000001L +#define UVD_SECURITY_REG_VIO_REPORT__VCPU_REG_VIO_MASK 0x00000002L +#define UVD_SECURITY_REG_VIO_REPORT__VIDEO_REG_VIO_MASK 0x00000004L +#define UVD_SECURITY_REG_VIO_REPORT__DPG_REG_VIO_MASK 0x00000008L +#define UVD_SECURITY_REG_VIO_REPORT__JPEG_REG_VIO_MASK 0x00000010L +#define UVD_SECURITY_REG_VIO_REPORT__JDPG_REG_VIO_MASK 0x00000020L +//UVD_FW_VERSION +#define UVD_FW_VERSION__FW_VERSION__SHIFT 0x0 +#define UVD_FW_VERSION__FW_VERSION_MASK 0xFFFFFFFFL +//UVD_PF_STATUS +#define UVD_PF_STATUS__JPEG_PF_OCCURED__SHIFT 0x0 +#define UVD_PF_STATUS__NJ_PF_OCCURED__SHIFT 0x1 +#define UVD_PF_STATUS__ENCODER0_PF_OCCURED__SHIFT 0x2 +#define UVD_PF_STATUS__ENCODER1_PF_OCCURED__SHIFT 0x3 +#define UVD_PF_STATUS__ENCODER2_PF_OCCURED__SHIFT 0x4 +#define UVD_PF_STATUS__ENCODER3_PF_OCCURED__SHIFT 0x5 +#define UVD_PF_STATUS__ENCODER4_PF_OCCURED__SHIFT 0x6 +#define UVD_PF_STATUS__EJPEG_PF_OCCURED__SHIFT 0x7 +#define UVD_PF_STATUS__JPEG_PF_CLEAR__SHIFT 0x8 +#define UVD_PF_STATUS__NJ_PF_CLEAR__SHIFT 0x9 +#define UVD_PF_STATUS__ENCODER0_PF_CLEAR__SHIFT 0xa +#define UVD_PF_STATUS__ENCODER1_PF_CLEAR__SHIFT 0xb +#define UVD_PF_STATUS__ENCODER2_PF_CLEAR__SHIFT 0xc +#define UVD_PF_STATUS__ENCODER3_PF_CLEAR__SHIFT 0xd +#define UVD_PF_STATUS__ENCODER4_PF_CLEAR__SHIFT 0xe +#define UVD_PF_STATUS__EJPEG_PF_CLEAR__SHIFT 0xf +#define UVD_PF_STATUS__NJ_ATM_PF_OCCURED__SHIFT 0x10 +#define UVD_PF_STATUS__DJ_ATM_PF_OCCURED__SHIFT 0x11 +#define UVD_PF_STATUS__EJ_ATM_PF_OCCURED__SHIFT 0x12 +#define UVD_PF_STATUS__JPEG2_PF_OCCURED__SHIFT 0x13 +#define UVD_PF_STATUS__DJ2_ATM_PF_OCCURED__SHIFT 0x14 +#define UVD_PF_STATUS__JPEG2_PF_CLEAR__SHIFT 0x15 +#define UVD_PF_STATUS__ENCODER5_PF_OCCURED__SHIFT 0x16 +#define UVD_PF_STATUS__ENCODER5_PF_CLEAR__SHIFT 0x17 +#define UVD_PF_STATUS__JPEG_PF_OCCURED_MASK 0x00000001L +#define UVD_PF_STATUS__NJ_PF_OCCURED_MASK 0x00000002L +#define UVD_PF_STATUS__ENCODER0_PF_OCCURED_MASK 0x00000004L +#define UVD_PF_STATUS__ENCODER1_PF_OCCURED_MASK 0x00000008L +#define UVD_PF_STATUS__ENCODER2_PF_OCCURED_MASK 0x00000010L +#define UVD_PF_STATUS__ENCODER3_PF_OCCURED_MASK 0x00000020L +#define UVD_PF_STATUS__ENCODER4_PF_OCCURED_MASK 0x00000040L +#define UVD_PF_STATUS__EJPEG_PF_OCCURED_MASK 0x00000080L +#define UVD_PF_STATUS__JPEG_PF_CLEAR_MASK 0x00000100L +#define UVD_PF_STATUS__NJ_PF_CLEAR_MASK 0x00000200L +#define UVD_PF_STATUS__ENCODER0_PF_CLEAR_MASK 0x00000400L +#define UVD_PF_STATUS__ENCODER1_PF_CLEAR_MASK 0x00000800L +#define UVD_PF_STATUS__ENCODER2_PF_CLEAR_MASK 0x00001000L +#define UVD_PF_STATUS__ENCODER3_PF_CLEAR_MASK 0x00002000L +#define UVD_PF_STATUS__ENCODER4_PF_CLEAR_MASK 0x00004000L +#define UVD_PF_STATUS__EJPEG_PF_CLEAR_MASK 0x00008000L +#define UVD_PF_STATUS__NJ_ATM_PF_OCCURED_MASK 0x00010000L +#define UVD_PF_STATUS__DJ_ATM_PF_OCCURED_MASK 0x00020000L +#define UVD_PF_STATUS__EJ_ATM_PF_OCCURED_MASK 0x00040000L +#define UVD_PF_STATUS__JPEG2_PF_OCCURED_MASK 0x00080000L +#define UVD_PF_STATUS__DJ2_ATM_PF_OCCURED_MASK 0x00100000L +#define UVD_PF_STATUS__JPEG2_PF_CLEAR_MASK 0x00200000L +#define UVD_PF_STATUS__ENCODER5_PF_OCCURED_MASK 0x00400000L +#define UVD_PF_STATUS__ENCODER5_PF_CLEAR_MASK 0x00800000L +//UVD_DPG_CLK_EN_VCPU_REPORT +#define UVD_DPG_CLK_EN_VCPU_REPORT__CLK_EN__SHIFT 0x0 +#define UVD_DPG_CLK_EN_VCPU_REPORT__VCPU_REPORT__SHIFT 0x1 +#define UVD_DPG_CLK_EN_VCPU_REPORT__CLK_EN_MASK 0x00000001L +#define UVD_DPG_CLK_EN_VCPU_REPORT__VCPU_REPORT_MASK 0x000000FEL +//CC_UVD_VCPU_ERR_DETECT_BOT_LO +#define CC_UVD_VCPU_ERR_DETECT_BOT_LO__UVD_VCPU_ERR_DETECT_BOT_LO__SHIFT 0xc +#define CC_UVD_VCPU_ERR_DETECT_BOT_LO__UVD_VCPU_ERR_DETECT_BOT_LO_MASK 0xFFFFF000L +//CC_UVD_VCPU_ERR_DETECT_BOT_HI +#define CC_UVD_VCPU_ERR_DETECT_BOT_HI__UVD_VCPU_ERR_DETECT_BOT_HI__SHIFT 0x0 +#define CC_UVD_VCPU_ERR_DETECT_BOT_HI__UVD_VCPU_ERR_DETECT_BOT_HI_MASK 0x0000FFFFL +//CC_UVD_VCPU_ERR_DETECT_TOP_LO +#define CC_UVD_VCPU_ERR_DETECT_TOP_LO__UVD_VCPU_ERR_DETECT_TOP_LO__SHIFT 0xc +#define CC_UVD_VCPU_ERR_DETECT_TOP_LO__UVD_VCPU_ERR_DETECT_TOP_LO_MASK 0xFFFFF000L +//CC_UVD_VCPU_ERR_DETECT_TOP_HI +#define CC_UVD_VCPU_ERR_DETECT_TOP_HI__UVD_VCPU_ERR_DETECT_TOP_HI__SHIFT 0x0 +#define CC_UVD_VCPU_ERR_DETECT_TOP_HI__UVD_VCPU_ERR_DETECT_TOP_HI_MASK 0x0000FFFFL +//CC_UVD_VCPU_ERR +#define CC_UVD_VCPU_ERR__UVD_VCPU_ERR_STATUS__SHIFT 0x0 +#define CC_UVD_VCPU_ERR__UVD_VCPU_ERR_CLEAR__SHIFT 0x1 +#define CC_UVD_VCPU_ERR__UVD_VCPU_ERR_DETECT_EN__SHIFT 0x2 +#define CC_UVD_VCPU_ERR__UVD_TMZ_DBG_DIS__SHIFT 0x3 +#define CC_UVD_VCPU_ERR__RESET_ON_FAULT__SHIFT 0x4 +#define CC_UVD_VCPU_ERR__UVD_VCPU_ERR_STATUS_MASK 0x00000001L +#define CC_UVD_VCPU_ERR__UVD_VCPU_ERR_CLEAR_MASK 0x00000002L +#define CC_UVD_VCPU_ERR__UVD_VCPU_ERR_DETECT_EN_MASK 0x00000004L +#define CC_UVD_VCPU_ERR__UVD_TMZ_DBG_DIS_MASK 0x00000008L +#define CC_UVD_VCPU_ERR__RESET_ON_FAULT_MASK 0x00000010L +//CC_UVD_VCPU_ERR_INST_ADDR_LO +#define CC_UVD_VCPU_ERR_INST_ADDR_LO__UVD_VCPU_ERR_INST_ADDR_LO__SHIFT 0x0 +#define CC_UVD_VCPU_ERR_INST_ADDR_LO__UVD_VCPU_ERR_INST_ADDR_LO_MASK 0xFFFFFFFFL +//CC_UVD_VCPU_ERR_INST_ADDR_HI +#define CC_UVD_VCPU_ERR_INST_ADDR_HI__UVD_VCPU_ERR_INST_ADDR_HI__SHIFT 0x0 +#define CC_UVD_VCPU_ERR_INST_ADDR_HI__UVD_VCPU_ERR_INST_ADDR_HI_MASK 0x0000FFFFL +//UVD_LMI_MMSCH_NC_SPACE +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC0_SPACE__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC1_SPACE__SHIFT 0x3 +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC2_SPACE__SHIFT 0x6 +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC3_SPACE__SHIFT 0x9 +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC4_SPACE__SHIFT 0xc +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC5_SPACE__SHIFT 0xf +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC6_SPACE__SHIFT 0x12 +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC7_SPACE__SHIFT 0x15 +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC0_SPACE_MASK 0x00000007L +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC1_SPACE_MASK 0x00000038L +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC2_SPACE_MASK 0x000001C0L +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC3_SPACE_MASK 0x00000E00L +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC4_SPACE_MASK 0x00007000L +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC5_SPACE_MASK 0x00038000L +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC6_SPACE_MASK 0x001C0000L +#define UVD_LMI_MMSCH_NC_SPACE__MMSCH_NC7_SPACE_MASK 0x00E00000L +//UVD_LMI_ATOMIC_SPACE +#define UVD_LMI_ATOMIC_SPACE__ATOMIC_USER0_SPACE__SHIFT 0x0 +#define UVD_LMI_ATOMIC_SPACE__ATOMIC_USER1_SPACE__SHIFT 0x3 +#define UVD_LMI_ATOMIC_SPACE__ATOMIC_USER2_SPACE__SHIFT 0x6 +#define UVD_LMI_ATOMIC_SPACE__ATOMIC_USER3_SPACE__SHIFT 0x9 +#define UVD_LMI_ATOMIC_SPACE__ATOMIC_USER0_SPACE_MASK 0x00000007L +#define UVD_LMI_ATOMIC_SPACE__ATOMIC_USER1_SPACE_MASK 0x00000038L +#define UVD_LMI_ATOMIC_SPACE__ATOMIC_USER2_SPACE_MASK 0x000001C0L +#define UVD_LMI_ATOMIC_SPACE__ATOMIC_USER3_SPACE_MASK 0x00000E00L +//UVD_GFX8_ADDR_CONFIG +#define UVD_GFX8_ADDR_CONFIG__PIPE_INTERLEAVE_SIZE__SHIFT 0x4 +#define UVD_GFX8_ADDR_CONFIG__PIPE_INTERLEAVE_SIZE_MASK 0x00000070L +//UVD_GFX10_ADDR_CONFIG +#define UVD_GFX10_ADDR_CONFIG__NUM_PIPES__SHIFT 0x0 +#define UVD_GFX10_ADDR_CONFIG__PIPE_INTERLEAVE_SIZE__SHIFT 0x3 +#define UVD_GFX10_ADDR_CONFIG__MAX_COMPRESSED_FRAGS__SHIFT 0x6 +#define UVD_GFX10_ADDR_CONFIG__NUM_PKRS__SHIFT 0x8 +#define UVD_GFX10_ADDR_CONFIG__NUM_BANKS__SHIFT 0xc +#define UVD_GFX10_ADDR_CONFIG__NUM_SHADER_ENGINES__SHIFT 0x13 +#define UVD_GFX10_ADDR_CONFIG__NUM_PIPES_MASK 0x00000007L +#define UVD_GFX10_ADDR_CONFIG__PIPE_INTERLEAVE_SIZE_MASK 0x00000038L +#define UVD_GFX10_ADDR_CONFIG__MAX_COMPRESSED_FRAGS_MASK 0x000000C0L +#define UVD_GFX10_ADDR_CONFIG__NUM_PKRS_MASK 0x00000700L +#define UVD_GFX10_ADDR_CONFIG__NUM_BANKS_MASK 0x00007000L +#define UVD_GFX10_ADDR_CONFIG__NUM_SHADER_ENGINES_MASK 0x00180000L +//UVD_GPCNT2_CNTL +#define UVD_GPCNT2_CNTL__CLR__SHIFT 0x0 +#define UVD_GPCNT2_CNTL__START__SHIFT 0x1 +#define UVD_GPCNT2_CNTL__COUNTUP__SHIFT 0x2 +#define UVD_GPCNT2_CNTL__CLR_MASK 0x00000001L +#define UVD_GPCNT2_CNTL__START_MASK 0x00000002L +#define UVD_GPCNT2_CNTL__COUNTUP_MASK 0x00000004L +//UVD_GPCNT2_TARGET_LOWER +#define UVD_GPCNT2_TARGET_LOWER__TARGET__SHIFT 0x0 +#define UVD_GPCNT2_TARGET_LOWER__TARGET_MASK 0xFFFFFFFFL +//UVD_GPCNT2_STATUS_LOWER +#define UVD_GPCNT2_STATUS_LOWER__COUNT__SHIFT 0x0 +#define UVD_GPCNT2_STATUS_LOWER__COUNT_MASK 0xFFFFFFFFL +//UVD_GPCNT2_TARGET_UPPER +#define UVD_GPCNT2_TARGET_UPPER__TARGET__SHIFT 0x0 +#define UVD_GPCNT2_TARGET_UPPER__TARGET_MASK 0x0000FFFFL +//UVD_GPCNT2_STATUS_UPPER +#define UVD_GPCNT2_STATUS_UPPER__COUNT__SHIFT 0x0 +#define UVD_GPCNT2_STATUS_UPPER__COUNT_MASK 0x0000FFFFL +//UVD_GPCNT3_CNTL +#define UVD_GPCNT3_CNTL__CLR__SHIFT 0x0 +#define UVD_GPCNT3_CNTL__START__SHIFT 0x1 +#define UVD_GPCNT3_CNTL__COUNTUP__SHIFT 0x2 +#define UVD_GPCNT3_CNTL__FREQ__SHIFT 0x3 +#define UVD_GPCNT3_CNTL__DIV__SHIFT 0xa +#define UVD_GPCNT3_CNTL__CLR_MASK 0x00000001L +#define UVD_GPCNT3_CNTL__START_MASK 0x00000002L +#define UVD_GPCNT3_CNTL__COUNTUP_MASK 0x00000004L +#define UVD_GPCNT3_CNTL__FREQ_MASK 0x000003F8L +#define UVD_GPCNT3_CNTL__DIV_MASK 0x0001FC00L +//UVD_GPCNT3_TARGET_LOWER +#define UVD_GPCNT3_TARGET_LOWER__TARGET__SHIFT 0x0 +#define UVD_GPCNT3_TARGET_LOWER__TARGET_MASK 0xFFFFFFFFL +//UVD_GPCNT3_STATUS_LOWER +#define UVD_GPCNT3_STATUS_LOWER__COUNT__SHIFT 0x0 +#define UVD_GPCNT3_STATUS_LOWER__COUNT_MASK 0xFFFFFFFFL +//UVD_GPCNT3_TARGET_UPPER +#define UVD_GPCNT3_TARGET_UPPER__TARGET__SHIFT 0x0 +#define UVD_GPCNT3_TARGET_UPPER__TARGET_MASK 0x0000FFFFL +//UVD_GPCNT3_STATUS_UPPER +#define UVD_GPCNT3_STATUS_UPPER__COUNT__SHIFT 0x0 +#define UVD_GPCNT3_STATUS_UPPER__COUNT_MASK 0x0000FFFFL +//UVD_VCLK_DS_CNTL +#define UVD_VCLK_DS_CNTL__VCLK_DS_EN__SHIFT 0x0 +#define UVD_VCLK_DS_CNTL__VCLK_DS_STATUS__SHIFT 0x4 +#define UVD_VCLK_DS_CNTL__VCLK_DS_HYSTERESIS_CNT__SHIFT 0x10 +#define UVD_VCLK_DS_CNTL__VCLK_DS_EN_MASK 0x00000001L +#define UVD_VCLK_DS_CNTL__VCLK_DS_STATUS_MASK 0x00000010L +#define UVD_VCLK_DS_CNTL__VCLK_DS_HYSTERESIS_CNT_MASK 0xFFFF0000L +//UVD_DCLK_DS_CNTL +#define UVD_DCLK_DS_CNTL__DCLK_DS_EN__SHIFT 0x0 +#define UVD_DCLK_DS_CNTL__DCLK_DS_STATUS__SHIFT 0x4 +#define UVD_DCLK_DS_CNTL__DCLK_DS_HYSTERESIS_CNT__SHIFT 0x10 +#define UVD_DCLK_DS_CNTL__DCLK_DS_EN_MASK 0x00000001L +#define UVD_DCLK_DS_CNTL__DCLK_DS_STATUS_MASK 0x00000010L +#define UVD_DCLK_DS_CNTL__DCLK_DS_HYSTERESIS_CNT_MASK 0xFFFF0000L +//UVD_TSC_LOWER +#define UVD_TSC_LOWER__COUNT__SHIFT 0x0 +#define UVD_TSC_LOWER__COUNT_MASK 0xFFFFFFFFL +//UVD_TSC_UPPER +#define UVD_TSC_UPPER__COUNT__SHIFT 0x0 +#define UVD_TSC_UPPER__COUNT_MASK 0x00FFFFFFL +//VCN_FEATURES +#define VCN_FEATURES__HAS_VIDEO_DEC__SHIFT 0x0 +#define VCN_FEATURES__HAS_VIDEO_ENC__SHIFT 0x1 +#define VCN_FEATURES__HAS_MJPEG_DEC__SHIFT 0x2 +#define VCN_FEATURES__HAS_MJPEG_ENC__SHIFT 0x3 +#define VCN_FEATURES__HAS_VIDEO_VIRT__SHIFT 0x4 +#define VCN_FEATURES__HAS_H264_LEGACY_DEC__SHIFT 0x5 +#define VCN_FEATURES__HAS_UDEC_DEC__SHIFT 0x6 +#define VCN_FEATURES__HAS_MJPEG2_IDCT_DEC__SHIFT 0x7 +#define VCN_FEATURES__HAS_SCLR_DEC__SHIFT 0x8 +#define VCN_FEATURES__HAS_VP9_DEC__SHIFT 0x9 +#define VCN_FEATURES__HAS_AV1_DEC__SHIFT 0xa +#define VCN_FEATURES__HAS_EFC_ENC__SHIFT 0xb +#define VCN_FEATURES__HAS_EFC_HDR2SDR_ENC__SHIFT 0xc +#define VCN_FEATURES__HAS_DUAL_MJPEG_DEC__SHIFT 0xd +#define VCN_FEATURES__HAS_AV1_ENC__SHIFT 0xe +#define VCN_FEATURES__INSTANCE_ID__SHIFT 0x1c +#define VCN_FEATURES__HAS_VIDEO_DEC_MASK 0x00000001L +#define VCN_FEATURES__HAS_VIDEO_ENC_MASK 0x00000002L +#define VCN_FEATURES__HAS_MJPEG_DEC_MASK 0x00000004L +#define VCN_FEATURES__HAS_MJPEG_ENC_MASK 0x00000008L +#define VCN_FEATURES__HAS_VIDEO_VIRT_MASK 0x00000010L +#define VCN_FEATURES__HAS_H264_LEGACY_DEC_MASK 0x00000020L +#define VCN_FEATURES__HAS_UDEC_DEC_MASK 0x00000040L +#define VCN_FEATURES__HAS_MJPEG2_IDCT_DEC_MASK 0x00000080L +#define VCN_FEATURES__HAS_SCLR_DEC_MASK 0x00000100L +#define VCN_FEATURES__HAS_VP9_DEC_MASK 0x00000200L +#define VCN_FEATURES__HAS_AV1_DEC_MASK 0x00000400L +#define VCN_FEATURES__HAS_EFC_ENC_MASK 0x00000800L +#define VCN_FEATURES__HAS_EFC_HDR2SDR_ENC_MASK 0x00001000L +#define VCN_FEATURES__HAS_DUAL_MJPEG_DEC_MASK 0x00002000L +#define VCN_FEATURES__HAS_AV1_ENC_MASK 0x00004000L +#define VCN_FEATURES__INSTANCE_ID_MASK 0xF0000000L +//UVD_GPUIOV_STATUS +#define UVD_GPUIOV_STATUS__UVD_GPUIOV_STATUS_VF_ENABLE__SHIFT 0x0 +#define UVD_GPUIOV_STATUS__UVD_GPUIOV_STATUS_VF_ENABLE_MASK 0x00000001L +//UVD_RAS_VCPU_VCODEC_STATUS +#define UVD_RAS_VCPU_VCODEC_STATUS__POISONED_VF__SHIFT 0x0 +#define UVD_RAS_VCPU_VCODEC_STATUS__POISONED_PF__SHIFT 0x1f +#define UVD_RAS_VCPU_VCODEC_STATUS__POISONED_VF_MASK 0x7FFFFFFFL +#define UVD_RAS_VCPU_VCODEC_STATUS__POISONED_PF_MASK 0x80000000L +//UVD_RAS_MMSCH_FATAL_ERROR +#define UVD_RAS_MMSCH_FATAL_ERROR__POISONED_VF__SHIFT 0x0 +#define UVD_RAS_MMSCH_FATAL_ERROR__POISONED_PF__SHIFT 0x1f +#define UVD_RAS_MMSCH_FATAL_ERROR__POISONED_VF_MASK 0x7FFFFFFFL +#define UVD_RAS_MMSCH_FATAL_ERROR__POISONED_PF_MASK 0x80000000L +//UVD_RAS_JPEG0_STATUS +#define UVD_RAS_JPEG0_STATUS__POISONED_VF__SHIFT 0x0 +#define UVD_RAS_JPEG0_STATUS__POISONED_PF__SHIFT 0x1f +#define UVD_RAS_JPEG0_STATUS__POISONED_VF_MASK 0x7FFFFFFFL +#define UVD_RAS_JPEG0_STATUS__POISONED_PF_MASK 0x80000000L +//UVD_RAS_JPEG1_STATUS +#define UVD_RAS_JPEG1_STATUS__POISONED_VF__SHIFT 0x0 +#define UVD_RAS_JPEG1_STATUS__POISONED_PF__SHIFT 0x1f +#define UVD_RAS_JPEG1_STATUS__POISONED_VF_MASK 0x7FFFFFFFL +#define UVD_RAS_JPEG1_STATUS__POISONED_PF_MASK 0x80000000L +//UVD_RAS_CNTL_PMI_ARB +#define UVD_RAS_CNTL_PMI_ARB__STAT_VCPU_VCODEC__SHIFT 0x0 +#define UVD_RAS_CNTL_PMI_ARB__ACK_VCPU_VCODEC__SHIFT 0x1 +#define UVD_RAS_CNTL_PMI_ARB__STAT_MMSCH__SHIFT 0x2 +#define UVD_RAS_CNTL_PMI_ARB__ACK_MMSCH__SHIFT 0x3 +#define UVD_RAS_CNTL_PMI_ARB__STAT_JPEG0__SHIFT 0x4 +#define UVD_RAS_CNTL_PMI_ARB__ACK_JPEG0__SHIFT 0x5 +#define UVD_RAS_CNTL_PMI_ARB__STAT_JPEG1__SHIFT 0x6 +#define UVD_RAS_CNTL_PMI_ARB__ACK_JPEG1__SHIFT 0x7 +#define UVD_RAS_CNTL_PMI_ARB__STAT_VCPU_VCODEC_MASK 0x00000001L +#define UVD_RAS_CNTL_PMI_ARB__ACK_VCPU_VCODEC_MASK 0x00000002L +#define UVD_RAS_CNTL_PMI_ARB__STAT_MMSCH_MASK 0x00000004L +#define UVD_RAS_CNTL_PMI_ARB__ACK_MMSCH_MASK 0x00000008L +#define UVD_RAS_CNTL_PMI_ARB__STAT_JPEG0_MASK 0x00000010L +#define UVD_RAS_CNTL_PMI_ARB__ACK_JPEG0_MASK 0x00000020L +#define UVD_RAS_CNTL_PMI_ARB__STAT_JPEG1_MASK 0x00000040L +#define UVD_RAS_CNTL_PMI_ARB__ACK_JPEG1_MASK 0x00000080L +//UVD_SCRATCH15 +#define UVD_SCRATCH15__SCRATCH15_DATA__SHIFT 0x0 +#define UVD_SCRATCH15__SCRATCH15_DATA_MASK 0xFFFFFFFFL +//VCN_JPEG_DB_CTRL1 +#define VCN_JPEG_DB_CTRL1__OFFSET__SHIFT 0x2 +#define VCN_JPEG_DB_CTRL1__EN__SHIFT 0x1e +#define VCN_JPEG_DB_CTRL1__HIT__SHIFT 0x1f +#define VCN_JPEG_DB_CTRL1__OFFSET_MASK 0x0FFFFFFCL +#define VCN_JPEG_DB_CTRL1__EN_MASK 0x40000000L +#define VCN_JPEG_DB_CTRL1__HIT_MASK 0x80000000L +//VCN_JPEG_DB_CTRL2 +#define VCN_JPEG_DB_CTRL2__OFFSET__SHIFT 0x2 +#define VCN_JPEG_DB_CTRL2__EN__SHIFT 0x1e +#define VCN_JPEG_DB_CTRL2__HIT__SHIFT 0x1f +#define VCN_JPEG_DB_CTRL2__OFFSET_MASK 0x0FFFFFFCL +#define VCN_JPEG_DB_CTRL2__EN_MASK 0x40000000L +#define VCN_JPEG_DB_CTRL2__HIT_MASK 0x80000000L +//VCN_JPEG_DB_CTRL3 +#define VCN_JPEG_DB_CTRL3__OFFSET__SHIFT 0x2 +#define VCN_JPEG_DB_CTRL3__EN__SHIFT 0x1e +#define VCN_JPEG_DB_CTRL3__HIT__SHIFT 0x1f +#define VCN_JPEG_DB_CTRL3__OFFSET_MASK 0x0FFFFFFCL +#define VCN_JPEG_DB_CTRL3__EN_MASK 0x40000000L +#define VCN_JPEG_DB_CTRL3__HIT_MASK 0x80000000L +//VCN_JPEG_DB_CTRL4 +#define VCN_JPEG_DB_CTRL4__OFFSET__SHIFT 0x2 +#define VCN_JPEG_DB_CTRL4__EN__SHIFT 0x1e +#define VCN_JPEG_DB_CTRL4__HIT__SHIFT 0x1f +#define VCN_JPEG_DB_CTRL4__OFFSET_MASK 0x0FFFFFFCL +#define VCN_JPEG_DB_CTRL4__EN_MASK 0x40000000L +#define VCN_JPEG_DB_CTRL4__HIT_MASK 0x80000000L +//VCN_JPEG_DB_CTRL5 +#define VCN_JPEG_DB_CTRL5__OFFSET__SHIFT 0x2 +#define VCN_JPEG_DB_CTRL5__EN__SHIFT 0x1e +#define VCN_JPEG_DB_CTRL5__HIT__SHIFT 0x1f +#define VCN_JPEG_DB_CTRL5__OFFSET_MASK 0x0FFFFFFCL +#define VCN_JPEG_DB_CTRL5__EN_MASK 0x40000000L +#define VCN_JPEG_DB_CTRL5__HIT_MASK 0x80000000L +//VCN_JPEG_DB_CTRL6 +#define VCN_JPEG_DB_CTRL6__OFFSET__SHIFT 0x2 +#define VCN_JPEG_DB_CTRL6__EN__SHIFT 0x1e +#define VCN_JPEG_DB_CTRL6__HIT__SHIFT 0x1f +#define VCN_JPEG_DB_CTRL6__OFFSET_MASK 0x0FFFFFFCL +#define VCN_JPEG_DB_CTRL6__EN_MASK 0x40000000L +#define VCN_JPEG_DB_CTRL6__HIT_MASK 0x80000000L +//VCN_JPEG_DB_CTRL7 +#define VCN_JPEG_DB_CTRL7__OFFSET__SHIFT 0x2 +#define VCN_JPEG_DB_CTRL7__EN__SHIFT 0x1e +#define VCN_JPEG_DB_CTRL7__HIT__SHIFT 0x1f +#define VCN_JPEG_DB_CTRL7__OFFSET_MASK 0x0FFFFFFCL +#define VCN_JPEG_DB_CTRL7__EN_MASK 0x40000000L +#define VCN_JPEG_DB_CTRL7__HIT_MASK 0x80000000L +//UVD_SCRATCH32 +#define UVD_SCRATCH32__SCRATCH32_DATA__SHIFT 0x0 +#define UVD_SCRATCH32__SCRATCH32_DATA_MASK 0xFFFFFFFFL +//UVD_VERSION +#define UVD_VERSION__VARIANT_TYPE__SHIFT 0x0 +#define UVD_VERSION__MINOR_VERSION__SHIFT 0x8 +#define UVD_VERSION__MAJOR_VERSION__SHIFT 0x10 +#define UVD_VERSION__INSTANCE_ID__SHIFT 0x1c +#define UVD_VERSION__VARIANT_TYPE_MASK 0x000000FFL +#define UVD_VERSION__MINOR_VERSION_MASK 0x0000FF00L +#define UVD_VERSION__MAJOR_VERSION_MASK 0x0FFF0000L +#define UVD_VERSION__INSTANCE_ID_MASK 0xF0000000L +//VCN_RB_DB_CTRL +#define VCN_RB_DB_CTRL__OFFSET__SHIFT 0x2 +#define VCN_RB_DB_CTRL__EN__SHIFT 0x1e +#define VCN_RB_DB_CTRL__HIT__SHIFT 0x1f +#define VCN_RB_DB_CTRL__OFFSET_MASK 0x0FFFFFFCL +#define VCN_RB_DB_CTRL__EN_MASK 0x40000000L +#define VCN_RB_DB_CTRL__HIT_MASK 0x80000000L +//VCN_JPEG_DB_CTRL +#define VCN_JPEG_DB_CTRL__OFFSET__SHIFT 0x2 +#define VCN_JPEG_DB_CTRL__EN__SHIFT 0x1e +#define VCN_JPEG_DB_CTRL__HIT__SHIFT 0x1f +#define VCN_JPEG_DB_CTRL__OFFSET_MASK 0x0FFFFFFCL +#define VCN_JPEG_DB_CTRL__EN_MASK 0x40000000L +#define VCN_JPEG_DB_CTRL__HIT_MASK 0x80000000L +//VCN_RB1_DB_CTRL +#define VCN_RB1_DB_CTRL__OFFSET__SHIFT 0x2 +#define VCN_RB1_DB_CTRL__EN__SHIFT 0x1e +#define VCN_RB1_DB_CTRL__HIT__SHIFT 0x1f +#define VCN_RB1_DB_CTRL__OFFSET_MASK 0x0FFFFFFCL +#define VCN_RB1_DB_CTRL__EN_MASK 0x40000000L +#define VCN_RB1_DB_CTRL__HIT_MASK 0x80000000L +//VCN_RB2_DB_CTRL +#define VCN_RB2_DB_CTRL__OFFSET__SHIFT 0x2 +#define VCN_RB2_DB_CTRL__EN__SHIFT 0x1e +#define VCN_RB2_DB_CTRL__HIT__SHIFT 0x1f +#define VCN_RB2_DB_CTRL__OFFSET_MASK 0x0FFFFFFCL +#define VCN_RB2_DB_CTRL__EN_MASK 0x40000000L +#define VCN_RB2_DB_CTRL__HIT_MASK 0x80000000L +//VCN_RB3_DB_CTRL +#define VCN_RB3_DB_CTRL__OFFSET__SHIFT 0x2 +#define VCN_RB3_DB_CTRL__EN__SHIFT 0x1e +#define VCN_RB3_DB_CTRL__HIT__SHIFT 0x1f +#define VCN_RB3_DB_CTRL__OFFSET_MASK 0x0FFFFFFCL +#define VCN_RB3_DB_CTRL__EN_MASK 0x40000000L +#define VCN_RB3_DB_CTRL__HIT_MASK 0x80000000L +//VCN_RB4_DB_CTRL +#define VCN_RB4_DB_CTRL__OFFSET__SHIFT 0x2 +#define VCN_RB4_DB_CTRL__EN__SHIFT 0x1e +#define VCN_RB4_DB_CTRL__HIT__SHIFT 0x1f +#define VCN_RB4_DB_CTRL__OFFSET_MASK 0x0FFFFFFCL +#define VCN_RB4_DB_CTRL__EN_MASK 0x40000000L +#define VCN_RB4_DB_CTRL__HIT_MASK 0x80000000L +//VCN_RB_ENABLE +#define VCN_RB_ENABLE__RB_EN__SHIFT 0x0 +#define VCN_RB_ENABLE__JPEG_RB_EN__SHIFT 0x1 +#define VCN_RB_ENABLE__RB1_EN__SHIFT 0x2 +#define VCN_RB_ENABLE__RB2_EN__SHIFT 0x3 +#define VCN_RB_ENABLE__RB3_EN__SHIFT 0x4 +#define VCN_RB_ENABLE__RB4_EN__SHIFT 0x5 +#define VCN_RB_ENABLE__UMSCH_RB_EN__SHIFT 0x6 +#define VCN_RB_ENABLE__EJPEG_RB_EN__SHIFT 0x7 +#define VCN_RB_ENABLE__AUDIO_RB_EN__SHIFT 0x8 +#define VCN_RB_ENABLE__RB_EN_MASK 0x00000001L +#define VCN_RB_ENABLE__JPEG_RB_EN_MASK 0x00000002L +#define VCN_RB_ENABLE__RB1_EN_MASK 0x00000004L +#define VCN_RB_ENABLE__RB2_EN_MASK 0x00000008L +#define VCN_RB_ENABLE__RB3_EN_MASK 0x00000010L +#define VCN_RB_ENABLE__RB4_EN_MASK 0x00000020L +#define VCN_RB_ENABLE__UMSCH_RB_EN_MASK 0x00000040L +#define VCN_RB_ENABLE__EJPEG_RB_EN_MASK 0x00000080L +#define VCN_RB_ENABLE__AUDIO_RB_EN_MASK 0x00000100L +//VCN_RB_WPTR_CTRL +#define VCN_RB_WPTR_CTRL__RB_CS_EN__SHIFT 0x0 +#define VCN_RB_WPTR_CTRL__JPEG_CS_EN__SHIFT 0x1 +#define VCN_RB_WPTR_CTRL__RB1_CS_EN__SHIFT 0x2 +#define VCN_RB_WPTR_CTRL__RB2_CS_EN__SHIFT 0x3 +#define VCN_RB_WPTR_CTRL__RB3_CS_EN__SHIFT 0x4 +#define VCN_RB_WPTR_CTRL__RB4_CS_EN__SHIFT 0x5 +#define VCN_RB_WPTR_CTRL__UMSCH_RB_CS_EN__SHIFT 0x6 +#define VCN_RB_WPTR_CTRL__EJPEG_RB_CS_EN__SHIFT 0x7 +#define VCN_RB_WPTR_CTRL__AUDIO_RB_CS_EN__SHIFT 0x8 +#define VCN_RB_WPTR_CTRL__RB_CS_EN_MASK 0x00000001L +#define VCN_RB_WPTR_CTRL__JPEG_CS_EN_MASK 0x00000002L +#define VCN_RB_WPTR_CTRL__RB1_CS_EN_MASK 0x00000004L +#define VCN_RB_WPTR_CTRL__RB2_CS_EN_MASK 0x00000008L +#define VCN_RB_WPTR_CTRL__RB3_CS_EN_MASK 0x00000010L +#define VCN_RB_WPTR_CTRL__RB4_CS_EN_MASK 0x00000020L +#define VCN_RB_WPTR_CTRL__UMSCH_RB_CS_EN_MASK 0x00000040L +#define VCN_RB_WPTR_CTRL__EJPEG_RB_CS_EN_MASK 0x00000080L +#define VCN_RB_WPTR_CTRL__AUDIO_RB_CS_EN_MASK 0x00000100L +//UVD_RB_RPTR +#define UVD_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_RB_RPTR__RB_RPTR_MASK 0x007FFFF0L +//UVD_RB_WPTR +#define UVD_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_RB_WPTR__RB_WPTR_MASK 0x007FFFF0L +//UVD_RB_RPTR2 +#define UVD_RB_RPTR2__RB_RPTR__SHIFT 0x4 +#define UVD_RB_RPTR2__RB_RPTR_MASK 0x007FFFF0L +//UVD_RB_WPTR2 +#define UVD_RB_WPTR2__RB_WPTR__SHIFT 0x4 +#define UVD_RB_WPTR2__RB_WPTR_MASK 0x007FFFF0L +//UVD_RB_RPTR3 +#define UVD_RB_RPTR3__RB_RPTR__SHIFT 0x4 +#define UVD_RB_RPTR3__RB_RPTR_MASK 0x007FFFF0L +//UVD_RB_WPTR3 +#define UVD_RB_WPTR3__RB_WPTR__SHIFT 0x4 +#define UVD_RB_WPTR3__RB_WPTR_MASK 0x007FFFF0L +//UVD_RB_RPTR4 +#define UVD_RB_RPTR4__RB_RPTR__SHIFT 0x4 +#define UVD_RB_RPTR4__RB_RPTR_MASK 0x007FFFF0L +//UVD_RB_WPTR4 +#define UVD_RB_WPTR4__RB_WPTR__SHIFT 0x4 +#define UVD_RB_WPTR4__RB_WPTR_MASK 0x007FFFF0L +//UVD_OUT_RB_RPTR +#define UVD_OUT_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_OUT_RB_RPTR__RB_RPTR_MASK 0x007FFFF0L +//UVD_OUT_RB_WPTR +#define UVD_OUT_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_OUT_RB_WPTR__RB_WPTR_MASK 0x007FFFF0L +//UVD_AUDIO_RB_RPTR +#define UVD_AUDIO_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_AUDIO_RB_RPTR__RB_RPTR_MASK 0x007FFFF0L +//UVD_AUDIO_RB_WPTR +#define UVD_AUDIO_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_AUDIO_RB_WPTR__RB_WPTR_MASK 0x007FFFF0L +//UVD_RBC_RB_RPTR +#define UVD_RBC_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_RBC_RB_RPTR__RB_RPTR_MASK 0x007FFFF0L +//UVD_RBC_RB_WPTR +#define UVD_RBC_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_RBC_RB_WPTR__RB_WPTR_MASK 0x007FFFF0L +//UVD_DPG_LMA_CTL2 +#define UVD_DPG_LMA_CTL2__DIRECT_ACCESS_SRAM_SEL__SHIFT 0x0 +#define UVD_DPG_LMA_CTL2__FIFO_DIRECT_ACCESS_EN__SHIFT 0x1 +#define UVD_DPG_LMA_CTL2__VID_WRITE_PTR__SHIFT 0x2 +#define UVD_DPG_LMA_CTL2__JPEG_WRITE_PTR__SHIFT 0x9 +#define UVD_DPG_LMA_CTL2__DIRECT_ACCESS_SRAM_SEL_MASK 0x00000001L +#define UVD_DPG_LMA_CTL2__FIFO_DIRECT_ACCESS_EN_MASK 0x00000002L +#define UVD_DPG_LMA_CTL2__VID_WRITE_PTR_MASK 0x000001FCL +#define UVD_DPG_LMA_CTL2__JPEG_WRITE_PTR_MASK 0x0000FE00L + + +// addressBlock: aid_uvd0_mmsch_dec +//MMSCH_UCODE_ADDR +#define MMSCH_UCODE_ADDR__UCODE_ADDR__SHIFT 0x2 +#define MMSCH_UCODE_ADDR__UCODE_LOCK__SHIFT 0x1f +#define MMSCH_UCODE_ADDR__UCODE_ADDR_MASK 0x00003FFCL +#define MMSCH_UCODE_ADDR__UCODE_LOCK_MASK 0x80000000L +//MMSCH_UCODE_DATA +#define MMSCH_UCODE_DATA__UCODE_DATA__SHIFT 0x0 +#define MMSCH_UCODE_DATA__UCODE_DATA_MASK 0xFFFFFFFFL +//MMSCH_SRAM_ADDR +#define MMSCH_SRAM_ADDR__SRAM_ADDR__SHIFT 0x2 +#define MMSCH_SRAM_ADDR__SRAM_LOCK__SHIFT 0x1f +#define MMSCH_SRAM_ADDR__SRAM_ADDR_MASK 0x00001FFCL +#define MMSCH_SRAM_ADDR__SRAM_LOCK_MASK 0x80000000L +//MMSCH_SRAM_DATA +#define MMSCH_SRAM_DATA__SRAM_DATA__SHIFT 0x0 +#define MMSCH_SRAM_DATA__SRAM_DATA_MASK 0xFFFFFFFFL +//MMSCH_VF_SRAM_OFFSET +#define MMSCH_VF_SRAM_OFFSET__VF_SRAM_OFFSET__SHIFT 0x2 +#define MMSCH_VF_SRAM_OFFSET__VF_SRAM_NUM_DW_PER_VF__SHIFT 0x10 +#define MMSCH_VF_SRAM_OFFSET__VF_SRAM_OFFSET_MASK 0x00001FFCL +#define MMSCH_VF_SRAM_OFFSET__VF_SRAM_NUM_DW_PER_VF_MASK 0x00FF0000L +//MMSCH_DB_SRAM_OFFSET +#define MMSCH_DB_SRAM_OFFSET__DB_SRAM_OFFSET__SHIFT 0x2 +#define MMSCH_DB_SRAM_OFFSET__DB_SRAM_NUM_ENG__SHIFT 0x10 +#define MMSCH_DB_SRAM_OFFSET__DB_SRAM_NUM_RING_PER_ENG__SHIFT 0x18 +#define MMSCH_DB_SRAM_OFFSET__DB_SRAM_OFFSET_MASK 0x00001FFCL +#define MMSCH_DB_SRAM_OFFSET__DB_SRAM_NUM_ENG_MASK 0x00FF0000L +#define MMSCH_DB_SRAM_OFFSET__DB_SRAM_NUM_RING_PER_ENG_MASK 0xFF000000L +//MMSCH_CTX_SRAM_OFFSET +#define MMSCH_CTX_SRAM_OFFSET__CTX_SRAM_OFFSET__SHIFT 0x2 +#define MMSCH_CTX_SRAM_OFFSET__CTX_SRAM_SIZE__SHIFT 0x10 +#define MMSCH_CTX_SRAM_OFFSET__CTX_SRAM_OFFSET_MASK 0x00001FFCL +#define MMSCH_CTX_SRAM_OFFSET__CTX_SRAM_SIZE_MASK 0xFFFF0000L +//MMSCH_CTL +#define MMSCH_CTL__P_RUNSTALL__SHIFT 0x0 +#define MMSCH_CTL__P_RESET__SHIFT 0x1 +#define MMSCH_CTL__VFID_FIFO_EN__SHIFT 0x4 +#define MMSCH_CTL__P_LOCK__SHIFT 0x1f +#define MMSCH_CTL__P_RUNSTALL_MASK 0x00000001L +#define MMSCH_CTL__P_RESET_MASK 0x00000002L +#define MMSCH_CTL__VFID_FIFO_EN_MASK 0x00000010L +#define MMSCH_CTL__P_LOCK_MASK 0x80000000L +//MMSCH_INTR +#define MMSCH_INTR__INTR__SHIFT 0x0 +#define MMSCH_INTR__INTR_MASK 0x00001FFFL +//MMSCH_INTR_ACK +#define MMSCH_INTR_ACK__INTR__SHIFT 0x0 +#define MMSCH_INTR_ACK__INTR_MASK 0x00001FFFL +//MMSCH_INTR_STATUS +#define MMSCH_INTR_STATUS__INTR__SHIFT 0x0 +#define MMSCH_INTR_STATUS__INTR_MASK 0x00001FFFL +//MMSCH_VF_VMID +#define MMSCH_VF_VMID__VF_CTX_VMID__SHIFT 0x0 +#define MMSCH_VF_VMID__VF_GPCOM_VMID__SHIFT 0x5 +#define MMSCH_VF_VMID__VF_CTX_VMID_MASK 0x0000001FL +#define MMSCH_VF_VMID__VF_GPCOM_VMID_MASK 0x000003E0L +//MMSCH_VF_CTX_ADDR_LO +#define MMSCH_VF_CTX_ADDR_LO__VF_CTX_ADDR_LO__SHIFT 0x6 +#define MMSCH_VF_CTX_ADDR_LO__VF_CTX_ADDR_LO_MASK 0xFFFFFFC0L +//MMSCH_VF_CTX_ADDR_HI +#define MMSCH_VF_CTX_ADDR_HI__VF_CTX_ADDR_HI__SHIFT 0x0 +#define MMSCH_VF_CTX_ADDR_HI__VF_CTX_ADDR_HI_MASK 0xFFFFFFFFL +//MMSCH_VF_CTX_SIZE +#define MMSCH_VF_CTX_SIZE__VF_CTX_SIZE__SHIFT 0x0 +#define MMSCH_VF_CTX_SIZE__VF_CTX_SIZE_MASK 0xFFFFFFFFL +//MMSCH_VF_GPCOM_ADDR_LO +#define MMSCH_VF_GPCOM_ADDR_LO__VF_GPCOM_ADDR_LO__SHIFT 0x6 +#define MMSCH_VF_GPCOM_ADDR_LO__VF_GPCOM_ADDR_LO_MASK 0xFFFFFFC0L +//MMSCH_VF_GPCOM_ADDR_HI +#define MMSCH_VF_GPCOM_ADDR_HI__VF_GPCOM_ADDR_HI__SHIFT 0x0 +#define MMSCH_VF_GPCOM_ADDR_HI__VF_GPCOM_ADDR_HI_MASK 0xFFFFFFFFL +//MMSCH_VF_GPCOM_SIZE +#define MMSCH_VF_GPCOM_SIZE__VF_GPCOM_SIZE__SHIFT 0x0 +#define MMSCH_VF_GPCOM_SIZE__VF_GPCOM_SIZE_MASK 0xFFFFFFFFL +//MMSCH_VF_MAILBOX_HOST +#define MMSCH_VF_MAILBOX_HOST__DATA__SHIFT 0x0 +#define MMSCH_VF_MAILBOX_HOST__DATA_MASK 0xFFFFFFFFL +//MMSCH_VF_MAILBOX_RESP +#define MMSCH_VF_MAILBOX_RESP__RESP__SHIFT 0x0 +#define MMSCH_VF_MAILBOX_RESP__RESP_MASK 0xFFFFFFFFL +//MMSCH_VF_MAILBOX_0 +#define MMSCH_VF_MAILBOX_0__DATA__SHIFT 0x0 +#define MMSCH_VF_MAILBOX_0__DATA_MASK 0xFFFFFFFFL +//MMSCH_VF_MAILBOX_0_RESP +#define MMSCH_VF_MAILBOX_0_RESP__RESP__SHIFT 0x0 +#define MMSCH_VF_MAILBOX_0_RESP__RESP_MASK 0xFFFFFFFFL +//MMSCH_VF_MAILBOX_1 +#define MMSCH_VF_MAILBOX_1__DATA__SHIFT 0x0 +#define MMSCH_VF_MAILBOX_1__DATA_MASK 0xFFFFFFFFL +//MMSCH_VF_MAILBOX_1_RESP +#define MMSCH_VF_MAILBOX_1_RESP__RESP__SHIFT 0x0 +#define MMSCH_VF_MAILBOX_1_RESP__RESP_MASK 0xFFFFFFFFL +//MMSCH_CNTL +#define MMSCH_CNTL__CLK_EN__SHIFT 0x0 +#define MMSCH_CNTL__ED_ENABLE__SHIFT 0x1 +#define MMSCH_CNTL__AXI_MAX_BRST_SIZE_IS_4__SHIFT 0x2 +#define MMSCH_CNTL__AXI_40BIT_PIF_ADDR_FIX_EN__SHIFT 0x3 +#define MMSCH_CNTL__PDEBUG_ENABLE__SHIFT 0x4 +#define MMSCH_CNTL__MMSCH_IRQ_ERR__SHIFT 0x5 +#define MMSCH_CNTL__MMSCH_NACK_INTR_EN__SHIFT 0x9 +#define MMSCH_CNTL__MMSCH_DB_BUSY_INTR_EN__SHIFT 0xa +#define MMSCH_CNTL__PRB_TIMEOUT_VAL__SHIFT 0x14 +#define MMSCH_CNTL__TIMEOUT_DIS__SHIFT 0x1c +#define MMSCH_CNTL__MMSCH_IDLE__SHIFT 0x1d +#define MMSCH_CNTL__CLK_EN_MASK 0x00000001L +#define MMSCH_CNTL__ED_ENABLE_MASK 0x00000002L +#define MMSCH_CNTL__AXI_MAX_BRST_SIZE_IS_4_MASK 0x00000004L +#define MMSCH_CNTL__AXI_40BIT_PIF_ADDR_FIX_EN_MASK 0x00000008L +#define MMSCH_CNTL__PDEBUG_ENABLE_MASK 0x00000010L +#define MMSCH_CNTL__MMSCH_IRQ_ERR_MASK 0x000001E0L +#define MMSCH_CNTL__MMSCH_NACK_INTR_EN_MASK 0x00000200L +#define MMSCH_CNTL__MMSCH_DB_BUSY_INTR_EN_MASK 0x00000400L +#define MMSCH_CNTL__PRB_TIMEOUT_VAL_MASK 0x0FF00000L +#define MMSCH_CNTL__TIMEOUT_DIS_MASK 0x10000000L +#define MMSCH_CNTL__MMSCH_IDLE_MASK 0x20000000L +//MMSCH_NONCACHE_OFFSET0 +#define MMSCH_NONCACHE_OFFSET0__OFFSET__SHIFT 0x0 +#define MMSCH_NONCACHE_OFFSET0__OFFSET_MASK 0x0FFFFFFFL +//MMSCH_NONCACHE_SIZE0 +#define MMSCH_NONCACHE_SIZE0__SIZE__SHIFT 0x0 +#define MMSCH_NONCACHE_SIZE0__SIZE_MASK 0x00FFFFFFL +//MMSCH_NONCACHE_OFFSET1 +#define MMSCH_NONCACHE_OFFSET1__OFFSET__SHIFT 0x0 +#define MMSCH_NONCACHE_OFFSET1__OFFSET_MASK 0x0FFFFFFFL +//MMSCH_NONCACHE_SIZE1 +#define MMSCH_NONCACHE_SIZE1__SIZE__SHIFT 0x0 +#define MMSCH_NONCACHE_SIZE1__SIZE_MASK 0x00FFFFFFL +//MMSCH_PROC_STATE1 +#define MMSCH_PROC_STATE1__PC__SHIFT 0x0 +#define MMSCH_PROC_STATE1__PC_MASK 0xFFFFFFFFL +//MMSCH_LAST_MC_ADDR +#define MMSCH_LAST_MC_ADDR__MC_ADDR__SHIFT 0x0 +#define MMSCH_LAST_MC_ADDR__RW__SHIFT 0x1f +#define MMSCH_LAST_MC_ADDR__MC_ADDR_MASK 0x0FFFFFFFL +#define MMSCH_LAST_MC_ADDR__RW_MASK 0x80000000L +//MMSCH_LAST_MEM_ACCESS_HI +#define MMSCH_LAST_MEM_ACCESS_HI__PROC_CMD__SHIFT 0x0 +#define MMSCH_LAST_MEM_ACCESS_HI__FIFO_RPTR__SHIFT 0x8 +#define MMSCH_LAST_MEM_ACCESS_HI__FIFO_WPTR__SHIFT 0xc +#define MMSCH_LAST_MEM_ACCESS_HI__PROC_CMD_MASK 0x00000007L +#define MMSCH_LAST_MEM_ACCESS_HI__FIFO_RPTR_MASK 0x00000700L +#define MMSCH_LAST_MEM_ACCESS_HI__FIFO_WPTR_MASK 0x00007000L +//MMSCH_LAST_MEM_ACCESS_LO +#define MMSCH_LAST_MEM_ACCESS_LO__PROC_ADDR__SHIFT 0x0 +#define MMSCH_LAST_MEM_ACCESS_LO__PROC_ADDR_MASK 0xFFFFFFFFL +//MMSCH_IOV_ACTIVE_FCN_ID +#define MMSCH_IOV_ACTIVE_FCN_ID__ACTIVE_VF_ID__SHIFT 0x0 +#define MMSCH_IOV_ACTIVE_FCN_ID__ACTIVE_PF_VF__SHIFT 0x1f +#define MMSCH_IOV_ACTIVE_FCN_ID__ACTIVE_VF_ID_MASK 0x0000001FL +#define MMSCH_IOV_ACTIVE_FCN_ID__ACTIVE_PF_VF_MASK 0x80000000L +//MMSCH_SCRATCH_0 +#define MMSCH_SCRATCH_0__SCRATCH_0__SHIFT 0x0 +#define MMSCH_SCRATCH_0__SCRATCH_0_MASK 0xFFFFFFFFL +//MMSCH_SCRATCH_1 +#define MMSCH_SCRATCH_1__SCRATCH_1__SHIFT 0x0 +#define MMSCH_SCRATCH_1__SCRATCH_1_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_SCH_BLOCK_0 +#define MMSCH_GPUIOV_SCH_BLOCK_0__ID__SHIFT 0x0 +#define MMSCH_GPUIOV_SCH_BLOCK_0__VERSION__SHIFT 0x4 +#define MMSCH_GPUIOV_SCH_BLOCK_0__SIZE__SHIFT 0x8 +#define MMSCH_GPUIOV_SCH_BLOCK_0__ID_MASK 0x0000000FL +#define MMSCH_GPUIOV_SCH_BLOCK_0__VERSION_MASK 0x000000F0L +#define MMSCH_GPUIOV_SCH_BLOCK_0__SIZE_MASK 0x0000FF00L +//MMSCH_GPUIOV_CMD_CONTROL_0 +#define MMSCH_GPUIOV_CMD_CONTROL_0__CMD_TYPE__SHIFT 0x0 +#define MMSCH_GPUIOV_CMD_CONTROL_0__CMD_EXECUTE__SHIFT 0x4 +#define MMSCH_GPUIOV_CMD_CONTROL_0__CMD_EXECUTE_INTR_EN__SHIFT 0x5 +#define MMSCH_GPUIOV_CMD_CONTROL_0__VM_BUSY_INTR_EN__SHIFT 0x6 +#define MMSCH_GPUIOV_CMD_CONTROL_0__FUNCTINO_ID__SHIFT 0x8 +#define MMSCH_GPUIOV_CMD_CONTROL_0__NEXT_FUNCTINO_ID__SHIFT 0x10 +#define MMSCH_GPUIOV_CMD_CONTROL_0__CMD_TYPE_MASK 0x0000000FL +#define MMSCH_GPUIOV_CMD_CONTROL_0__CMD_EXECUTE_MASK 0x00000010L +#define MMSCH_GPUIOV_CMD_CONTROL_0__CMD_EXECUTE_INTR_EN_MASK 0x00000020L +#define MMSCH_GPUIOV_CMD_CONTROL_0__VM_BUSY_INTR_EN_MASK 0x00000040L +#define MMSCH_GPUIOV_CMD_CONTROL_0__FUNCTINO_ID_MASK 0x0000FF00L +#define MMSCH_GPUIOV_CMD_CONTROL_0__NEXT_FUNCTINO_ID_MASK 0x00FF0000L +//MMSCH_GPUIOV_CMD_STATUS_0 +#define MMSCH_GPUIOV_CMD_STATUS_0__CMD_STATUS__SHIFT 0x0 +#define MMSCH_GPUIOV_CMD_STATUS_0__CMD_STATUS_MASK 0x0000000FL +//MMSCH_GPUIOV_VM_BUSY_STATUS_0 +#define MMSCH_GPUIOV_VM_BUSY_STATUS_0__BUSY__SHIFT 0x0 +#define MMSCH_GPUIOV_VM_BUSY_STATUS_0__BUSY_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_ACTIVE_FCNS_0 +#define MMSCH_GPUIOV_ACTIVE_FCNS_0__ACTIVE_FCNS__SHIFT 0x0 +#define MMSCH_GPUIOV_ACTIVE_FCNS_0__ACTIVE_FCNS_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_ACTIVE_FCN_ID_0 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_0__ID__SHIFT 0x0 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_0__ID_STATUS__SHIFT 0x8 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_0__ID_MASK 0x000000FFL +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_0__ID_STATUS_MASK 0x00000F00L +//MMSCH_GPUIOV_DW6_0 +#define MMSCH_GPUIOV_DW6_0__DATA__SHIFT 0x0 +#define MMSCH_GPUIOV_DW6_0__DATA_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_DW7_0 +#define MMSCH_GPUIOV_DW7_0__DATA__SHIFT 0x0 +#define MMSCH_GPUIOV_DW7_0__DATA_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_DW8_0 +#define MMSCH_GPUIOV_DW8_0__DATA__SHIFT 0x0 +#define MMSCH_GPUIOV_DW8_0__DATA_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_SCH_BLOCK_1 +#define MMSCH_GPUIOV_SCH_BLOCK_1__ID__SHIFT 0x0 +#define MMSCH_GPUIOV_SCH_BLOCK_1__VERSION__SHIFT 0x4 +#define MMSCH_GPUIOV_SCH_BLOCK_1__SIZE__SHIFT 0x8 +#define MMSCH_GPUIOV_SCH_BLOCK_1__ID_MASK 0x0000000FL +#define MMSCH_GPUIOV_SCH_BLOCK_1__VERSION_MASK 0x000000F0L +#define MMSCH_GPUIOV_SCH_BLOCK_1__SIZE_MASK 0x0000FF00L +//MMSCH_GPUIOV_CMD_CONTROL_1 +#define MMSCH_GPUIOV_CMD_CONTROL_1__CMD_TYPE__SHIFT 0x0 +#define MMSCH_GPUIOV_CMD_CONTROL_1__CMD_EXECUTE__SHIFT 0x4 +#define MMSCH_GPUIOV_CMD_CONTROL_1__CMD_EXECUTE_INTR_EN__SHIFT 0x5 +#define MMSCH_GPUIOV_CMD_CONTROL_1__VM_BUSY_INTR_EN__SHIFT 0x6 +#define MMSCH_GPUIOV_CMD_CONTROL_1__FUNCTINO_ID__SHIFT 0x8 +#define MMSCH_GPUIOV_CMD_CONTROL_1__NEXT_FUNCTINO_ID__SHIFT 0x10 +#define MMSCH_GPUIOV_CMD_CONTROL_1__CMD_TYPE_MASK 0x0000000FL +#define MMSCH_GPUIOV_CMD_CONTROL_1__CMD_EXECUTE_MASK 0x00000010L +#define MMSCH_GPUIOV_CMD_CONTROL_1__CMD_EXECUTE_INTR_EN_MASK 0x00000020L +#define MMSCH_GPUIOV_CMD_CONTROL_1__VM_BUSY_INTR_EN_MASK 0x00000040L +#define MMSCH_GPUIOV_CMD_CONTROL_1__FUNCTINO_ID_MASK 0x0000FF00L +#define MMSCH_GPUIOV_CMD_CONTROL_1__NEXT_FUNCTINO_ID_MASK 0x00FF0000L +//MMSCH_GPUIOV_CMD_STATUS_1 +#define MMSCH_GPUIOV_CMD_STATUS_1__CMD_STATUS__SHIFT 0x0 +#define MMSCH_GPUIOV_CMD_STATUS_1__CMD_STATUS_MASK 0x0000000FL +//MMSCH_GPUIOV_VM_BUSY_STATUS_1 +#define MMSCH_GPUIOV_VM_BUSY_STATUS_1__BUSY__SHIFT 0x0 +#define MMSCH_GPUIOV_VM_BUSY_STATUS_1__BUSY_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_ACTIVE_FCNS_1 +#define MMSCH_GPUIOV_ACTIVE_FCNS_1__ACTIVE_FCNS__SHIFT 0x0 +#define MMSCH_GPUIOV_ACTIVE_FCNS_1__ACTIVE_FCNS_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_ACTIVE_FCN_ID_1 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_1__ID__SHIFT 0x0 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_1__ID_STATUS__SHIFT 0x8 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_1__ID_MASK 0x000000FFL +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_1__ID_STATUS_MASK 0x00000F00L +//MMSCH_GPUIOV_DW6_1 +#define MMSCH_GPUIOV_DW6_1__DATA__SHIFT 0x0 +#define MMSCH_GPUIOV_DW6_1__DATA_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_DW7_1 +#define MMSCH_GPUIOV_DW7_1__DATA__SHIFT 0x0 +#define MMSCH_GPUIOV_DW7_1__DATA_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_DW8_1 +#define MMSCH_GPUIOV_DW8_1__DATA__SHIFT 0x0 +#define MMSCH_GPUIOV_DW8_1__DATA_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_CNTXT +#define MMSCH_GPUIOV_CNTXT__CNTXT_SIZE__SHIFT 0x0 +#define MMSCH_GPUIOV_CNTXT__CNTXT_LOCATION__SHIFT 0x7 +#define MMSCH_GPUIOV_CNTXT__CNTXT_OFFSET__SHIFT 0xa +#define MMSCH_GPUIOV_CNTXT__CNTXT_SIZE_MASK 0x0000007FL +#define MMSCH_GPUIOV_CNTXT__CNTXT_LOCATION_MASK 0x00000080L +#define MMSCH_GPUIOV_CNTXT__CNTXT_OFFSET_MASK 0xFFFFFC00L +//MMSCH_SCRATCH_2 +#define MMSCH_SCRATCH_2__SCRATCH_2__SHIFT 0x0 +#define MMSCH_SCRATCH_2__SCRATCH_2_MASK 0xFFFFFFFFL +//MMSCH_SCRATCH_3 +#define MMSCH_SCRATCH_3__SCRATCH_3__SHIFT 0x0 +#define MMSCH_SCRATCH_3__SCRATCH_3_MASK 0xFFFFFFFFL +//MMSCH_SCRATCH_4 +#define MMSCH_SCRATCH_4__SCRATCH_4__SHIFT 0x0 +#define MMSCH_SCRATCH_4__SCRATCH_4_MASK 0xFFFFFFFFL +//MMSCH_SCRATCH_5 +#define MMSCH_SCRATCH_5__SCRATCH_5__SHIFT 0x0 +#define MMSCH_SCRATCH_5__SCRATCH_5_MASK 0xFFFFFFFFL +//MMSCH_SCRATCH_6 +#define MMSCH_SCRATCH_6__SCRATCH_6__SHIFT 0x0 +#define MMSCH_SCRATCH_6__SCRATCH_6_MASK 0xFFFFFFFFL +//MMSCH_SCRATCH_7 +#define MMSCH_SCRATCH_7__SCRATCH_7__SHIFT 0x0 +#define MMSCH_SCRATCH_7__SCRATCH_7_MASK 0xFFFFFFFFL +//MMSCH_VFID_FIFO_HEAD_0 +#define MMSCH_VFID_FIFO_HEAD_0__HEAD__SHIFT 0x0 +#define MMSCH_VFID_FIFO_HEAD_0__HEAD_MASK 0x0000003FL +//MMSCH_VFID_FIFO_TAIL_0 +#define MMSCH_VFID_FIFO_TAIL_0__TAIL__SHIFT 0x0 +#define MMSCH_VFID_FIFO_TAIL_0__TAIL_MASK 0x0000003FL +//MMSCH_VFID_FIFO_HEAD_1 +#define MMSCH_VFID_FIFO_HEAD_1__HEAD__SHIFT 0x0 +#define MMSCH_VFID_FIFO_HEAD_1__HEAD_MASK 0x0000003FL +//MMSCH_VFID_FIFO_TAIL_1 +#define MMSCH_VFID_FIFO_TAIL_1__TAIL__SHIFT 0x0 +#define MMSCH_VFID_FIFO_TAIL_1__TAIL_MASK 0x0000003FL +//MMSCH_NACK_STATUS +#define MMSCH_NACK_STATUS__WR_NACK_STATUS__SHIFT 0x0 +#define MMSCH_NACK_STATUS__RD_NACK_STATUS__SHIFT 0x2 +#define MMSCH_NACK_STATUS__WR_NACK_STATUS_MASK 0x00000003L +#define MMSCH_NACK_STATUS__RD_NACK_STATUS_MASK 0x0000000CL +//MMSCH_VF_MAILBOX0_DATA +#define MMSCH_VF_MAILBOX0_DATA__DATA__SHIFT 0x0 +#define MMSCH_VF_MAILBOX0_DATA__DATA_MASK 0xFFFFFFFFL +//MMSCH_VF_MAILBOX1_DATA +#define MMSCH_VF_MAILBOX1_DATA__DATA__SHIFT 0x0 +#define MMSCH_VF_MAILBOX1_DATA__DATA_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_SCH_BLOCK_IP_0 +#define MMSCH_GPUIOV_SCH_BLOCK_IP_0__ID__SHIFT 0x0 +#define MMSCH_GPUIOV_SCH_BLOCK_IP_0__VERSION__SHIFT 0x4 +#define MMSCH_GPUIOV_SCH_BLOCK_IP_0__SIZE__SHIFT 0x8 +#define MMSCH_GPUIOV_SCH_BLOCK_IP_0__ID_MASK 0x0000000FL +#define MMSCH_GPUIOV_SCH_BLOCK_IP_0__VERSION_MASK 0x000000F0L +#define MMSCH_GPUIOV_SCH_BLOCK_IP_0__SIZE_MASK 0x0000FF00L +//MMSCH_GPUIOV_CMD_STATUS_IP_0 +#define MMSCH_GPUIOV_CMD_STATUS_IP_0__CMD_STATUS__SHIFT 0x0 +#define MMSCH_GPUIOV_CMD_STATUS_IP_0__CMD_STATUS_MASK 0x0000000FL +//MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_0 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_0__ID__SHIFT 0x0 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_0__ID_STATUS__SHIFT 0x8 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_0__ID_MASK 0x000000FFL +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_0__ID_STATUS_MASK 0x00000F00L +//MMSCH_GPUIOV_SCH_BLOCK_IP_1 +#define MMSCH_GPUIOV_SCH_BLOCK_IP_1__ID__SHIFT 0x0 +#define MMSCH_GPUIOV_SCH_BLOCK_IP_1__VERSION__SHIFT 0x4 +#define MMSCH_GPUIOV_SCH_BLOCK_IP_1__SIZE__SHIFT 0x8 +#define MMSCH_GPUIOV_SCH_BLOCK_IP_1__ID_MASK 0x0000000FL +#define MMSCH_GPUIOV_SCH_BLOCK_IP_1__VERSION_MASK 0x000000F0L +#define MMSCH_GPUIOV_SCH_BLOCK_IP_1__SIZE_MASK 0x0000FF00L +//MMSCH_GPUIOV_CMD_STATUS_IP_1 +#define MMSCH_GPUIOV_CMD_STATUS_IP_1__CMD_STATUS__SHIFT 0x0 +#define MMSCH_GPUIOV_CMD_STATUS_IP_1__CMD_STATUS_MASK 0x0000000FL +//MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_1 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_1__ID__SHIFT 0x0 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_1__ID_STATUS__SHIFT 0x8 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_1__ID_MASK 0x000000FFL +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_1__ID_STATUS_MASK 0x00000F00L +//MMSCH_GPUIOV_CNTXT_IP +#define MMSCH_GPUIOV_CNTXT_IP__CNTXT_SIZE__SHIFT 0x0 +#define MMSCH_GPUIOV_CNTXT_IP__CNTXT_LOCATION__SHIFT 0x7 +#define MMSCH_GPUIOV_CNTXT_IP__CNTXT_SIZE_MASK 0x0000007FL +#define MMSCH_GPUIOV_CNTXT_IP__CNTXT_LOCATION_MASK 0x00000080L +//MMSCH_GPUIOV_SCH_BLOCK_2 +#define MMSCH_GPUIOV_SCH_BLOCK_2__ID__SHIFT 0x0 +#define MMSCH_GPUIOV_SCH_BLOCK_2__VERSION__SHIFT 0x4 +#define MMSCH_GPUIOV_SCH_BLOCK_2__SIZE__SHIFT 0x8 +#define MMSCH_GPUIOV_SCH_BLOCK_2__ID_MASK 0x0000000FL +#define MMSCH_GPUIOV_SCH_BLOCK_2__VERSION_MASK 0x000000F0L +#define MMSCH_GPUIOV_SCH_BLOCK_2__SIZE_MASK 0x0000FF00L +//MMSCH_GPUIOV_CMD_CONTROL_2 +#define MMSCH_GPUIOV_CMD_CONTROL_2__CMD_TYPE__SHIFT 0x0 +#define MMSCH_GPUIOV_CMD_CONTROL_2__CMD_EXECUTE__SHIFT 0x4 +#define MMSCH_GPUIOV_CMD_CONTROL_2__CMD_EXECUTE_INTR_EN__SHIFT 0x5 +#define MMSCH_GPUIOV_CMD_CONTROL_2__VM_BUSY_INTR_EN__SHIFT 0x6 +#define MMSCH_GPUIOV_CMD_CONTROL_2__FUNCTINO_ID__SHIFT 0x8 +#define MMSCH_GPUIOV_CMD_CONTROL_2__NEXT_FUNCTINO_ID__SHIFT 0x10 +#define MMSCH_GPUIOV_CMD_CONTROL_2__CMD_TYPE_MASK 0x0000000FL +#define MMSCH_GPUIOV_CMD_CONTROL_2__CMD_EXECUTE_MASK 0x00000010L +#define MMSCH_GPUIOV_CMD_CONTROL_2__CMD_EXECUTE_INTR_EN_MASK 0x00000020L +#define MMSCH_GPUIOV_CMD_CONTROL_2__VM_BUSY_INTR_EN_MASK 0x00000040L +#define MMSCH_GPUIOV_CMD_CONTROL_2__FUNCTINO_ID_MASK 0x0000FF00L +#define MMSCH_GPUIOV_CMD_CONTROL_2__NEXT_FUNCTINO_ID_MASK 0x00FF0000L +//MMSCH_GPUIOV_CMD_STATUS_2 +#define MMSCH_GPUIOV_CMD_STATUS_2__CMD_STATUS__SHIFT 0x0 +#define MMSCH_GPUIOV_CMD_STATUS_2__CMD_STATUS_MASK 0x0000000FL +//MMSCH_GPUIOV_VM_BUSY_STATUS_2 +#define MMSCH_GPUIOV_VM_BUSY_STATUS_2__BUSY__SHIFT 0x0 +#define MMSCH_GPUIOV_VM_BUSY_STATUS_2__BUSY_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_ACTIVE_FCNS_2 +#define MMSCH_GPUIOV_ACTIVE_FCNS_2__ACTIVE_FCNS__SHIFT 0x0 +#define MMSCH_GPUIOV_ACTIVE_FCNS_2__ACTIVE_FCNS_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_ACTIVE_FCN_ID_2 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_2__ID__SHIFT 0x0 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_2__ID_STATUS__SHIFT 0x8 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_2__ID_MASK 0x000000FFL +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_2__ID_STATUS_MASK 0x00000F00L +//MMSCH_GPUIOV_DW6_2 +#define MMSCH_GPUIOV_DW6_2__DATA__SHIFT 0x0 +#define MMSCH_GPUIOV_DW6_2__DATA_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_DW7_2 +#define MMSCH_GPUIOV_DW7_2__DATA__SHIFT 0x0 +#define MMSCH_GPUIOV_DW7_2__DATA_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_DW8_2 +#define MMSCH_GPUIOV_DW8_2__DATA__SHIFT 0x0 +#define MMSCH_GPUIOV_DW8_2__DATA_MASK 0xFFFFFFFFL +//MMSCH_GPUIOV_SCH_BLOCK_IP_2 +#define MMSCH_GPUIOV_SCH_BLOCK_IP_2__ID__SHIFT 0x0 +#define MMSCH_GPUIOV_SCH_BLOCK_IP_2__VERSION__SHIFT 0x4 +#define MMSCH_GPUIOV_SCH_BLOCK_IP_2__SIZE__SHIFT 0x8 +#define MMSCH_GPUIOV_SCH_BLOCK_IP_2__ID_MASK 0x0000000FL +#define MMSCH_GPUIOV_SCH_BLOCK_IP_2__VERSION_MASK 0x000000F0L +#define MMSCH_GPUIOV_SCH_BLOCK_IP_2__SIZE_MASK 0x0000FF00L +//MMSCH_GPUIOV_CMD_STATUS_IP_2 +#define MMSCH_GPUIOV_CMD_STATUS_IP_2__CMD_STATUS__SHIFT 0x0 +#define MMSCH_GPUIOV_CMD_STATUS_IP_2__CMD_STATUS_MASK 0x0000000FL +//MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_2 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_2__ID__SHIFT 0x0 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_2__ID_STATUS__SHIFT 0x8 +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_2__ID_MASK 0x000000FFL +#define MMSCH_GPUIOV_ACTIVE_FCN_ID_IP_2__ID_STATUS_MASK 0x00000F00L +//MMSCH_VFID_FIFO_HEAD_2 +#define MMSCH_VFID_FIFO_HEAD_2__HEAD__SHIFT 0x0 +#define MMSCH_VFID_FIFO_HEAD_2__HEAD_MASK 0x0000003FL +//MMSCH_VFID_FIFO_TAIL_2 +#define MMSCH_VFID_FIFO_TAIL_2__TAIL__SHIFT 0x0 +#define MMSCH_VFID_FIFO_TAIL_2__TAIL_MASK 0x0000003FL +//MMSCH_VM_BUSY_STATUS_0 +#define MMSCH_VM_BUSY_STATUS_0__BUSY__SHIFT 0x0 +#define MMSCH_VM_BUSY_STATUS_0__BUSY_MASK 0xFFFFFFFFL +//MMSCH_VM_BUSY_STATUS_1 +#define MMSCH_VM_BUSY_STATUS_1__BUSY__SHIFT 0x0 +#define MMSCH_VM_BUSY_STATUS_1__BUSY_MASK 0xFFFFFFFFL +//MMSCH_VM_BUSY_STATUS_2 +#define MMSCH_VM_BUSY_STATUS_2__BUSY__SHIFT 0x0 +#define MMSCH_VM_BUSY_STATUS_2__BUSY_MASK 0xFFFFFFFFL + + +// addressBlock: aid_uvd0_slmi_adpdec +//UVD_LMI_MMSCH_NC0_64BIT_BAR_LOW +#define UVD_LMI_MMSCH_NC0_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC0_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC0_64BIT_BAR_HIGH +#define UVD_LMI_MMSCH_NC0_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC0_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC1_64BIT_BAR_LOW +#define UVD_LMI_MMSCH_NC1_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC1_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC1_64BIT_BAR_HIGH +#define UVD_LMI_MMSCH_NC1_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC1_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC2_64BIT_BAR_LOW +#define UVD_LMI_MMSCH_NC2_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC2_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC2_64BIT_BAR_HIGH +#define UVD_LMI_MMSCH_NC2_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC2_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC3_64BIT_BAR_LOW +#define UVD_LMI_MMSCH_NC3_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC3_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC3_64BIT_BAR_HIGH +#define UVD_LMI_MMSCH_NC3_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC3_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC4_64BIT_BAR_LOW +#define UVD_LMI_MMSCH_NC4_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC4_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC4_64BIT_BAR_HIGH +#define UVD_LMI_MMSCH_NC4_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC4_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC5_64BIT_BAR_LOW +#define UVD_LMI_MMSCH_NC5_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC5_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC5_64BIT_BAR_HIGH +#define UVD_LMI_MMSCH_NC5_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC5_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC6_64BIT_BAR_LOW +#define UVD_LMI_MMSCH_NC6_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC6_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC6_64BIT_BAR_HIGH +#define UVD_LMI_MMSCH_NC6_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC6_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC7_64BIT_BAR_LOW +#define UVD_LMI_MMSCH_NC7_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC7_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC7_64BIT_BAR_HIGH +#define UVD_LMI_MMSCH_NC7_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC7_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_LMI_MMSCH_NC_VMID +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC0_VMID__SHIFT 0x0 +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC1_VMID__SHIFT 0x4 +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC2_VMID__SHIFT 0x8 +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC3_VMID__SHIFT 0xc +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC4_VMID__SHIFT 0x10 +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC5_VMID__SHIFT 0x14 +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC6_VMID__SHIFT 0x18 +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC7_VMID__SHIFT 0x1c +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC0_VMID_MASK 0x0000000FL +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC1_VMID_MASK 0x000000F0L +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC2_VMID_MASK 0x00000F00L +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC3_VMID_MASK 0x0000F000L +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC4_VMID_MASK 0x000F0000L +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC5_VMID_MASK 0x00F00000L +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC6_VMID_MASK 0x0F000000L +#define UVD_LMI_MMSCH_NC_VMID__MMSCH_NC7_VMID_MASK 0xF0000000L +//UVD_LMI_MMSCH_CTRL +#define UVD_LMI_MMSCH_CTRL__MMSCH_DATA_COHERENCY_EN__SHIFT 0x0 +#define UVD_LMI_MMSCH_CTRL__MMSCH_VM__SHIFT 0x1 +#define UVD_LMI_MMSCH_CTRL__PRIV_CLIENT_MMSCH__SHIFT 0x2 +#define UVD_LMI_MMSCH_CTRL__MMSCH_R_MC_SWAP__SHIFT 0x3 +#define UVD_LMI_MMSCH_CTRL__MMSCH_W_MC_SWAP__SHIFT 0x5 +#define UVD_LMI_MMSCH_CTRL__MMSCH_RD__SHIFT 0x7 +#define UVD_LMI_MMSCH_CTRL__MMSCH_WR__SHIFT 0x9 +#define UVD_LMI_MMSCH_CTRL__MMSCH_RD_DROP__SHIFT 0xb +#define UVD_LMI_MMSCH_CTRL__MMSCH_WR_DROP__SHIFT 0xc +#define UVD_LMI_MMSCH_CTRL__MMSCH_DATA_COHERENCY_EN_MASK 0x00000001L +#define UVD_LMI_MMSCH_CTRL__MMSCH_VM_MASK 0x00000002L +#define UVD_LMI_MMSCH_CTRL__PRIV_CLIENT_MMSCH_MASK 0x00000004L +#define UVD_LMI_MMSCH_CTRL__MMSCH_R_MC_SWAP_MASK 0x00000018L +#define UVD_LMI_MMSCH_CTRL__MMSCH_W_MC_SWAP_MASK 0x00000060L +#define UVD_LMI_MMSCH_CTRL__MMSCH_RD_MASK 0x00000180L +#define UVD_LMI_MMSCH_CTRL__MMSCH_WR_MASK 0x00000600L +#define UVD_LMI_MMSCH_CTRL__MMSCH_RD_DROP_MASK 0x00000800L +#define UVD_LMI_MMSCH_CTRL__MMSCH_WR_DROP_MASK 0x00001000L +//UVD_MMSCH_LMI_STATUS +#define UVD_MMSCH_LMI_STATUS__LMI_AXI_MMSCH_UNSUPPORTED_LEN_INT__SHIFT 0x0 +#define UVD_MMSCH_LMI_STATUS__LMI_AXI_MMSCH_UNSUPPORTED_ADR_ALIGN_INT__SHIFT 0x1 +#define UVD_MMSCH_LMI_STATUS__MMSCH_LMI_WRITE_CLEAN__SHIFT 0x2 +#define UVD_MMSCH_LMI_STATUS__AXI_MMSCH_ERR_LEN__SHIFT 0x4 +#define UVD_MMSCH_LMI_STATUS__AXI_MMSCH_ERR_ADDR_LSBS__SHIFT 0x8 +#define UVD_MMSCH_LMI_STATUS__AXI_MMSCH_ERR_AWRITE__SHIFT 0xc +#define UVD_MMSCH_LMI_STATUS__MMSCH_RD_CLEAN__SHIFT 0xd +#define UVD_MMSCH_LMI_STATUS__MMSCH_WR_CLEAN__SHIFT 0xe +#define UVD_MMSCH_LMI_STATUS__LMI_AXI_MMSCH_UNSUPPORTED_LEN_INT_MASK 0x00000001L +#define UVD_MMSCH_LMI_STATUS__LMI_AXI_MMSCH_UNSUPPORTED_ADR_ALIGN_INT_MASK 0x00000002L +#define UVD_MMSCH_LMI_STATUS__MMSCH_LMI_WRITE_CLEAN_MASK 0x00000004L +#define UVD_MMSCH_LMI_STATUS__AXI_MMSCH_ERR_LEN_MASK 0x000000F0L +#define UVD_MMSCH_LMI_STATUS__AXI_MMSCH_ERR_ADDR_LSBS_MASK 0x00000700L +#define UVD_MMSCH_LMI_STATUS__AXI_MMSCH_ERR_AWRITE_MASK 0x00001000L +#define UVD_MMSCH_LMI_STATUS__MMSCH_RD_CLEAN_MASK 0x00002000L +#define UVD_MMSCH_LMI_STATUS__MMSCH_WR_CLEAN_MASK 0x00004000L +//VCN_RAS_CNTL_MMSCH +#define VCN_RAS_CNTL_MMSCH__MMSCH_FATAL_ERROR_EN__SHIFT 0x1 +#define VCN_RAS_CNTL_MMSCH__MMSCH_PMI_EN__SHIFT 0x5 +#define VCN_RAS_CNTL_MMSCH__MMSCH_REARM__SHIFT 0x9 +#define VCN_RAS_CNTL_MMSCH__MMSCH_READY__SHIFT 0x11 +#define VCN_RAS_CNTL_MMSCH__MMSCH_FATAL_ERROR_EN_MASK 0x00000002L +#define VCN_RAS_CNTL_MMSCH__MMSCH_PMI_EN_MASK 0x00000020L +#define VCN_RAS_CNTL_MMSCH__MMSCH_REARM_MASK 0x00000200L +#define VCN_RAS_CNTL_MMSCH__MMSCH_READY_MASK 0x00020000L + + +// addressBlock: aid_uvd0_uvd_jrbc1_uvd_jrbc_dec +//UVD_JRBC1_UVD_JRBC_RB_WPTR +#define UVD_JRBC1_UVD_JRBC_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_JRBC1_UVD_JRBC_RB_WPTR__RB_WPTR_MASK 0x007FFFF0L +//UVD_JRBC1_UVD_JRBC_RB_CNTL +#define UVD_JRBC1_UVD_JRBC_RB_CNTL__RB_NO_FETCH__SHIFT 0x0 +#define UVD_JRBC1_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN__SHIFT 0x1 +#define UVD_JRBC1_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER__SHIFT 0x4 +#define UVD_JRBC1_UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK 0x00000001L +#define UVD_JRBC1_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK 0x00000002L +#define UVD_JRBC1_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER_MASK 0x0007FFF0L +//UVD_JRBC1_UVD_JRBC_IB_SIZE +#define UVD_JRBC1_UVD_JRBC_IB_SIZE__IB_SIZE__SHIFT 0x4 +#define UVD_JRBC1_UVD_JRBC_IB_SIZE__IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC1_UVD_JRBC_URGENT_CNTL +#define UVD_JRBC1_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK__SHIFT 0x0 +#define UVD_JRBC1_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK_MASK 0x00000003L +//UVD_JRBC1_UVD_JRBC_RB_REF_DATA +#define UVD_JRBC1_UVD_JRBC_RB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC1_UVD_JRBC_RB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC1_UVD_JRBC_RB_COND_RD_TIMER +#define UVD_JRBC1_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC1_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC1_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC1_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC1_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC1_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC1_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC1_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC1_UVD_JRBC_SOFT_RESET +#define UVD_JRBC1_UVD_JRBC_SOFT_RESET__RESET__SHIFT 0x0 +#define UVD_JRBC1_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS__SHIFT 0x11 +#define UVD_JRBC1_UVD_JRBC_SOFT_RESET__RESET_MASK 0x00000001L +#define UVD_JRBC1_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS_MASK 0x00020000L +//UVD_JRBC1_UVD_JRBC_STATUS +#define UVD_JRBC1_UVD_JRBC_STATUS__RB_JOB_DONE__SHIFT 0x0 +#define UVD_JRBC1_UVD_JRBC_STATUS__IB_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC1_UVD_JRBC_STATUS__RB_ILLEGAL_CMD__SHIFT 0x2 +#define UVD_JRBC1_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT__SHIFT 0x3 +#define UVD_JRBC1_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT__SHIFT 0x4 +#define UVD_JRBC1_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT__SHIFT 0x5 +#define UVD_JRBC1_UVD_JRBC_STATUS__IB_ILLEGAL_CMD__SHIFT 0x6 +#define UVD_JRBC1_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT__SHIFT 0x7 +#define UVD_JRBC1_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT__SHIFT 0x8 +#define UVD_JRBC1_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT__SHIFT 0x9 +#define UVD_JRBC1_UVD_JRBC_STATUS__RB_TRAP_STATUS__SHIFT 0xa +#define UVD_JRBC1_UVD_JRBC_STATUS__PREEMPT_STATUS__SHIFT 0xb +#define UVD_JRBC1_UVD_JRBC_STATUS__IB_TRAP_STATUS__SHIFT 0xc +#define UVD_JRBC1_UVD_JRBC_STATUS__INT_EN__SHIFT 0x10 +#define UVD_JRBC1_UVD_JRBC_STATUS__INT_ACK__SHIFT 0x11 +#define UVD_JRBC1_UVD_JRBC_STATUS__RB_JOB_DONE_MASK 0x00000001L +#define UVD_JRBC1_UVD_JRBC_STATUS__IB_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC1_UVD_JRBC_STATUS__RB_ILLEGAL_CMD_MASK 0x00000004L +#define UVD_JRBC1_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT_MASK 0x00000008L +#define UVD_JRBC1_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT_MASK 0x00000010L +#define UVD_JRBC1_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT_MASK 0x00000020L +#define UVD_JRBC1_UVD_JRBC_STATUS__IB_ILLEGAL_CMD_MASK 0x00000040L +#define UVD_JRBC1_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT_MASK 0x00000080L +#define UVD_JRBC1_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT_MASK 0x00000100L +#define UVD_JRBC1_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT_MASK 0x00000200L +#define UVD_JRBC1_UVD_JRBC_STATUS__RB_TRAP_STATUS_MASK 0x00000400L +#define UVD_JRBC1_UVD_JRBC_STATUS__PREEMPT_STATUS_MASK 0x00000800L +#define UVD_JRBC1_UVD_JRBC_STATUS__IB_TRAP_STATUS_MASK 0x00001000L +#define UVD_JRBC1_UVD_JRBC_STATUS__INT_EN_MASK 0x00010000L +#define UVD_JRBC1_UVD_JRBC_STATUS__INT_ACK_MASK 0x00020000L +//UVD_JRBC1_UVD_JRBC_RB_RPTR +#define UVD_JRBC1_UVD_JRBC_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_JRBC1_UVD_JRBC_RB_RPTR__RB_RPTR_MASK 0x007FFFF0L +//UVD_JRBC1_UVD_JRBC_RB_BUF_STATUS +#define UVD_JRBC1_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC1_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC1_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC1_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC1_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC1_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC1_UVD_JRBC_IB_BUF_STATUS +#define UVD_JRBC1_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC1_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC1_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC1_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC1_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC1_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC1_UVD_JRBC_IB_SIZE_UPDATE +#define UVD_JRBC1_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE__SHIFT 0x4 +#define UVD_JRBC1_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC1_UVD_JRBC_IB_COND_RD_TIMER +#define UVD_JRBC1_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC1_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC1_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC1_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC1_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC1_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC1_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC1_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC1_UVD_JRBC_IB_REF_DATA +#define UVD_JRBC1_UVD_JRBC_IB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC1_UVD_JRBC_IB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC1_UVD_JPEG_PREEMPT_CMD +#define UVD_JRBC1_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN__SHIFT 0x0 +#define UVD_JRBC1_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC1_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD__SHIFT 0x2 +#define UVD_JRBC1_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN_MASK 0x00000001L +#define UVD_JRBC1_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC1_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD_MASK 0x00000004L +//UVD_JRBC1_UVD_JPEG_PREEMPT_FENCE_DATA0 +#define UVD_JRBC1_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0__SHIFT 0x0 +#define UVD_JRBC1_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0_MASK 0xFFFFFFFFL +//UVD_JRBC1_UVD_JPEG_PREEMPT_FENCE_DATA1 +#define UVD_JRBC1_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1__SHIFT 0x0 +#define UVD_JRBC1_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1_MASK 0xFFFFFFFFL +//UVD_JRBC1_UVD_JRBC_RB_SIZE +#define UVD_JRBC1_UVD_JRBC_RB_SIZE__RB_SIZE__SHIFT 0x4 +#define UVD_JRBC1_UVD_JRBC_RB_SIZE__RB_SIZE_MASK 0x00FFFFF0L +//UVD_JRBC1_UVD_JRBC_SCRATCH0 +#define UVD_JRBC1_UVD_JRBC_SCRATCH0__SCRATCH0__SHIFT 0x0 +#define UVD_JRBC1_UVD_JRBC_SCRATCH0__SCRATCH0_MASK 0xFFFFFFFFL + + +// addressBlock: aid_uvd0_uvd_jrbc2_uvd_jrbc_dec +//UVD_JRBC2_UVD_JRBC_RB_WPTR +#define UVD_JRBC2_UVD_JRBC_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_JRBC2_UVD_JRBC_RB_WPTR__RB_WPTR_MASK 0x007FFFF0L +//UVD_JRBC2_UVD_JRBC_RB_CNTL +#define UVD_JRBC2_UVD_JRBC_RB_CNTL__RB_NO_FETCH__SHIFT 0x0 +#define UVD_JRBC2_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN__SHIFT 0x1 +#define UVD_JRBC2_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER__SHIFT 0x4 +#define UVD_JRBC2_UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK 0x00000001L +#define UVD_JRBC2_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK 0x00000002L +#define UVD_JRBC2_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER_MASK 0x0007FFF0L +//UVD_JRBC2_UVD_JRBC_IB_SIZE +#define UVD_JRBC2_UVD_JRBC_IB_SIZE__IB_SIZE__SHIFT 0x4 +#define UVD_JRBC2_UVD_JRBC_IB_SIZE__IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC2_UVD_JRBC_URGENT_CNTL +#define UVD_JRBC2_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK__SHIFT 0x0 +#define UVD_JRBC2_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK_MASK 0x00000003L +//UVD_JRBC2_UVD_JRBC_RB_REF_DATA +#define UVD_JRBC2_UVD_JRBC_RB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC2_UVD_JRBC_RB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC2_UVD_JRBC_RB_COND_RD_TIMER +#define UVD_JRBC2_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC2_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC2_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC2_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC2_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC2_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC2_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC2_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC2_UVD_JRBC_SOFT_RESET +#define UVD_JRBC2_UVD_JRBC_SOFT_RESET__RESET__SHIFT 0x0 +#define UVD_JRBC2_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS__SHIFT 0x11 +#define UVD_JRBC2_UVD_JRBC_SOFT_RESET__RESET_MASK 0x00000001L +#define UVD_JRBC2_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS_MASK 0x00020000L +//UVD_JRBC2_UVD_JRBC_STATUS +#define UVD_JRBC2_UVD_JRBC_STATUS__RB_JOB_DONE__SHIFT 0x0 +#define UVD_JRBC2_UVD_JRBC_STATUS__IB_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC2_UVD_JRBC_STATUS__RB_ILLEGAL_CMD__SHIFT 0x2 +#define UVD_JRBC2_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT__SHIFT 0x3 +#define UVD_JRBC2_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT__SHIFT 0x4 +#define UVD_JRBC2_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT__SHIFT 0x5 +#define UVD_JRBC2_UVD_JRBC_STATUS__IB_ILLEGAL_CMD__SHIFT 0x6 +#define UVD_JRBC2_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT__SHIFT 0x7 +#define UVD_JRBC2_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT__SHIFT 0x8 +#define UVD_JRBC2_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT__SHIFT 0x9 +#define UVD_JRBC2_UVD_JRBC_STATUS__RB_TRAP_STATUS__SHIFT 0xa +#define UVD_JRBC2_UVD_JRBC_STATUS__PREEMPT_STATUS__SHIFT 0xb +#define UVD_JRBC2_UVD_JRBC_STATUS__IB_TRAP_STATUS__SHIFT 0xc +#define UVD_JRBC2_UVD_JRBC_STATUS__INT_EN__SHIFT 0x10 +#define UVD_JRBC2_UVD_JRBC_STATUS__INT_ACK__SHIFT 0x11 +#define UVD_JRBC2_UVD_JRBC_STATUS__RB_JOB_DONE_MASK 0x00000001L +#define UVD_JRBC2_UVD_JRBC_STATUS__IB_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC2_UVD_JRBC_STATUS__RB_ILLEGAL_CMD_MASK 0x00000004L +#define UVD_JRBC2_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT_MASK 0x00000008L +#define UVD_JRBC2_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT_MASK 0x00000010L +#define UVD_JRBC2_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT_MASK 0x00000020L +#define UVD_JRBC2_UVD_JRBC_STATUS__IB_ILLEGAL_CMD_MASK 0x00000040L +#define UVD_JRBC2_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT_MASK 0x00000080L +#define UVD_JRBC2_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT_MASK 0x00000100L +#define UVD_JRBC2_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT_MASK 0x00000200L +#define UVD_JRBC2_UVD_JRBC_STATUS__RB_TRAP_STATUS_MASK 0x00000400L +#define UVD_JRBC2_UVD_JRBC_STATUS__PREEMPT_STATUS_MASK 0x00000800L +#define UVD_JRBC2_UVD_JRBC_STATUS__IB_TRAP_STATUS_MASK 0x00001000L +#define UVD_JRBC2_UVD_JRBC_STATUS__INT_EN_MASK 0x00010000L +#define UVD_JRBC2_UVD_JRBC_STATUS__INT_ACK_MASK 0x00020000L +//UVD_JRBC2_UVD_JRBC_RB_RPTR +#define UVD_JRBC2_UVD_JRBC_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_JRBC2_UVD_JRBC_RB_RPTR__RB_RPTR_MASK 0x007FFFF0L +//UVD_JRBC2_UVD_JRBC_RB_BUF_STATUS +#define UVD_JRBC2_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC2_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC2_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC2_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC2_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC2_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC2_UVD_JRBC_IB_BUF_STATUS +#define UVD_JRBC2_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC2_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC2_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC2_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC2_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC2_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC2_UVD_JRBC_IB_SIZE_UPDATE +#define UVD_JRBC2_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE__SHIFT 0x4 +#define UVD_JRBC2_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC2_UVD_JRBC_IB_COND_RD_TIMER +#define UVD_JRBC2_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC2_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC2_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC2_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC2_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC2_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC2_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC2_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC2_UVD_JRBC_IB_REF_DATA +#define UVD_JRBC2_UVD_JRBC_IB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC2_UVD_JRBC_IB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC2_UVD_JPEG_PREEMPT_CMD +#define UVD_JRBC2_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN__SHIFT 0x0 +#define UVD_JRBC2_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC2_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD__SHIFT 0x2 +#define UVD_JRBC2_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN_MASK 0x00000001L +#define UVD_JRBC2_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC2_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD_MASK 0x00000004L +//UVD_JRBC2_UVD_JPEG_PREEMPT_FENCE_DATA0 +#define UVD_JRBC2_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0__SHIFT 0x0 +#define UVD_JRBC2_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0_MASK 0xFFFFFFFFL +//UVD_JRBC2_UVD_JPEG_PREEMPT_FENCE_DATA1 +#define UVD_JRBC2_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1__SHIFT 0x0 +#define UVD_JRBC2_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1_MASK 0xFFFFFFFFL +//UVD_JRBC2_UVD_JRBC_RB_SIZE +#define UVD_JRBC2_UVD_JRBC_RB_SIZE__RB_SIZE__SHIFT 0x4 +#define UVD_JRBC2_UVD_JRBC_RB_SIZE__RB_SIZE_MASK 0x00FFFFF0L +//UVD_JRBC2_UVD_JRBC_SCRATCH0 +#define UVD_JRBC2_UVD_JRBC_SCRATCH0__SCRATCH0__SHIFT 0x0 +#define UVD_JRBC2_UVD_JRBC_SCRATCH0__SCRATCH0_MASK 0xFFFFFFFFL + + +// addressBlock: aid_uvd0_uvd_jrbc3_uvd_jrbc_dec +//UVD_JRBC3_UVD_JRBC_RB_WPTR +#define UVD_JRBC3_UVD_JRBC_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_JRBC3_UVD_JRBC_RB_WPTR__RB_WPTR_MASK 0x007FFFF0L +//UVD_JRBC3_UVD_JRBC_RB_CNTL +#define UVD_JRBC3_UVD_JRBC_RB_CNTL__RB_NO_FETCH__SHIFT 0x0 +#define UVD_JRBC3_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN__SHIFT 0x1 +#define UVD_JRBC3_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER__SHIFT 0x4 +#define UVD_JRBC3_UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK 0x00000001L +#define UVD_JRBC3_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK 0x00000002L +#define UVD_JRBC3_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER_MASK 0x0007FFF0L +//UVD_JRBC3_UVD_JRBC_IB_SIZE +#define UVD_JRBC3_UVD_JRBC_IB_SIZE__IB_SIZE__SHIFT 0x4 +#define UVD_JRBC3_UVD_JRBC_IB_SIZE__IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC3_UVD_JRBC_URGENT_CNTL +#define UVD_JRBC3_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK__SHIFT 0x0 +#define UVD_JRBC3_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK_MASK 0x00000003L +//UVD_JRBC3_UVD_JRBC_RB_REF_DATA +#define UVD_JRBC3_UVD_JRBC_RB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC3_UVD_JRBC_RB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC3_UVD_JRBC_RB_COND_RD_TIMER +#define UVD_JRBC3_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC3_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC3_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC3_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC3_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC3_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC3_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC3_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC3_UVD_JRBC_SOFT_RESET +#define UVD_JRBC3_UVD_JRBC_SOFT_RESET__RESET__SHIFT 0x0 +#define UVD_JRBC3_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS__SHIFT 0x11 +#define UVD_JRBC3_UVD_JRBC_SOFT_RESET__RESET_MASK 0x00000001L +#define UVD_JRBC3_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS_MASK 0x00020000L +//UVD_JRBC3_UVD_JRBC_STATUS +#define UVD_JRBC3_UVD_JRBC_STATUS__RB_JOB_DONE__SHIFT 0x0 +#define UVD_JRBC3_UVD_JRBC_STATUS__IB_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC3_UVD_JRBC_STATUS__RB_ILLEGAL_CMD__SHIFT 0x2 +#define UVD_JRBC3_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT__SHIFT 0x3 +#define UVD_JRBC3_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT__SHIFT 0x4 +#define UVD_JRBC3_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT__SHIFT 0x5 +#define UVD_JRBC3_UVD_JRBC_STATUS__IB_ILLEGAL_CMD__SHIFT 0x6 +#define UVD_JRBC3_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT__SHIFT 0x7 +#define UVD_JRBC3_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT__SHIFT 0x8 +#define UVD_JRBC3_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT__SHIFT 0x9 +#define UVD_JRBC3_UVD_JRBC_STATUS__RB_TRAP_STATUS__SHIFT 0xa +#define UVD_JRBC3_UVD_JRBC_STATUS__PREEMPT_STATUS__SHIFT 0xb +#define UVD_JRBC3_UVD_JRBC_STATUS__IB_TRAP_STATUS__SHIFT 0xc +#define UVD_JRBC3_UVD_JRBC_STATUS__INT_EN__SHIFT 0x10 +#define UVD_JRBC3_UVD_JRBC_STATUS__INT_ACK__SHIFT 0x11 +#define UVD_JRBC3_UVD_JRBC_STATUS__RB_JOB_DONE_MASK 0x00000001L +#define UVD_JRBC3_UVD_JRBC_STATUS__IB_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC3_UVD_JRBC_STATUS__RB_ILLEGAL_CMD_MASK 0x00000004L +#define UVD_JRBC3_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT_MASK 0x00000008L +#define UVD_JRBC3_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT_MASK 0x00000010L +#define UVD_JRBC3_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT_MASK 0x00000020L +#define UVD_JRBC3_UVD_JRBC_STATUS__IB_ILLEGAL_CMD_MASK 0x00000040L +#define UVD_JRBC3_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT_MASK 0x00000080L +#define UVD_JRBC3_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT_MASK 0x00000100L +#define UVD_JRBC3_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT_MASK 0x00000200L +#define UVD_JRBC3_UVD_JRBC_STATUS__RB_TRAP_STATUS_MASK 0x00000400L +#define UVD_JRBC3_UVD_JRBC_STATUS__PREEMPT_STATUS_MASK 0x00000800L +#define UVD_JRBC3_UVD_JRBC_STATUS__IB_TRAP_STATUS_MASK 0x00001000L +#define UVD_JRBC3_UVD_JRBC_STATUS__INT_EN_MASK 0x00010000L +#define UVD_JRBC3_UVD_JRBC_STATUS__INT_ACK_MASK 0x00020000L +//UVD_JRBC3_UVD_JRBC_RB_RPTR +#define UVD_JRBC3_UVD_JRBC_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_JRBC3_UVD_JRBC_RB_RPTR__RB_RPTR_MASK 0x007FFFF0L +//UVD_JRBC3_UVD_JRBC_RB_BUF_STATUS +#define UVD_JRBC3_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC3_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC3_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC3_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC3_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC3_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC3_UVD_JRBC_IB_BUF_STATUS +#define UVD_JRBC3_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC3_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC3_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC3_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC3_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC3_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC3_UVD_JRBC_IB_SIZE_UPDATE +#define UVD_JRBC3_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE__SHIFT 0x4 +#define UVD_JRBC3_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC3_UVD_JRBC_IB_COND_RD_TIMER +#define UVD_JRBC3_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC3_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC3_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC3_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC3_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC3_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC3_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC3_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC3_UVD_JRBC_IB_REF_DATA +#define UVD_JRBC3_UVD_JRBC_IB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC3_UVD_JRBC_IB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC3_UVD_JPEG_PREEMPT_CMD +#define UVD_JRBC3_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN__SHIFT 0x0 +#define UVD_JRBC3_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC3_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD__SHIFT 0x2 +#define UVD_JRBC3_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN_MASK 0x00000001L +#define UVD_JRBC3_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC3_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD_MASK 0x00000004L +//UVD_JRBC3_UVD_JPEG_PREEMPT_FENCE_DATA0 +#define UVD_JRBC3_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0__SHIFT 0x0 +#define UVD_JRBC3_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0_MASK 0xFFFFFFFFL +//UVD_JRBC3_UVD_JPEG_PREEMPT_FENCE_DATA1 +#define UVD_JRBC3_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1__SHIFT 0x0 +#define UVD_JRBC3_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1_MASK 0xFFFFFFFFL +//UVD_JRBC3_UVD_JRBC_RB_SIZE +#define UVD_JRBC3_UVD_JRBC_RB_SIZE__RB_SIZE__SHIFT 0x4 +#define UVD_JRBC3_UVD_JRBC_RB_SIZE__RB_SIZE_MASK 0x00FFFFF0L +//UVD_JRBC3_UVD_JRBC_SCRATCH0 +#define UVD_JRBC3_UVD_JRBC_SCRATCH0__SCRATCH0__SHIFT 0x0 +#define UVD_JRBC3_UVD_JRBC_SCRATCH0__SCRATCH0_MASK 0xFFFFFFFFL + + +// addressBlock: aid_uvd0_uvd_jrbc4_uvd_jrbc_dec +//UVD_JRBC4_UVD_JRBC_RB_WPTR +#define UVD_JRBC4_UVD_JRBC_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_JRBC4_UVD_JRBC_RB_WPTR__RB_WPTR_MASK 0x007FFFF0L +//UVD_JRBC4_UVD_JRBC_RB_CNTL +#define UVD_JRBC4_UVD_JRBC_RB_CNTL__RB_NO_FETCH__SHIFT 0x0 +#define UVD_JRBC4_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN__SHIFT 0x1 +#define UVD_JRBC4_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER__SHIFT 0x4 +#define UVD_JRBC4_UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK 0x00000001L +#define UVD_JRBC4_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK 0x00000002L +#define UVD_JRBC4_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER_MASK 0x0007FFF0L +//UVD_JRBC4_UVD_JRBC_IB_SIZE +#define UVD_JRBC4_UVD_JRBC_IB_SIZE__IB_SIZE__SHIFT 0x4 +#define UVD_JRBC4_UVD_JRBC_IB_SIZE__IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC4_UVD_JRBC_URGENT_CNTL +#define UVD_JRBC4_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK__SHIFT 0x0 +#define UVD_JRBC4_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK_MASK 0x00000003L +//UVD_JRBC4_UVD_JRBC_RB_REF_DATA +#define UVD_JRBC4_UVD_JRBC_RB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC4_UVD_JRBC_RB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC4_UVD_JRBC_RB_COND_RD_TIMER +#define UVD_JRBC4_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC4_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC4_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC4_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC4_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC4_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC4_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC4_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC4_UVD_JRBC_SOFT_RESET +#define UVD_JRBC4_UVD_JRBC_SOFT_RESET__RESET__SHIFT 0x0 +#define UVD_JRBC4_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS__SHIFT 0x11 +#define UVD_JRBC4_UVD_JRBC_SOFT_RESET__RESET_MASK 0x00000001L +#define UVD_JRBC4_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS_MASK 0x00020000L +//UVD_JRBC4_UVD_JRBC_STATUS +#define UVD_JRBC4_UVD_JRBC_STATUS__RB_JOB_DONE__SHIFT 0x0 +#define UVD_JRBC4_UVD_JRBC_STATUS__IB_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC4_UVD_JRBC_STATUS__RB_ILLEGAL_CMD__SHIFT 0x2 +#define UVD_JRBC4_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT__SHIFT 0x3 +#define UVD_JRBC4_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT__SHIFT 0x4 +#define UVD_JRBC4_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT__SHIFT 0x5 +#define UVD_JRBC4_UVD_JRBC_STATUS__IB_ILLEGAL_CMD__SHIFT 0x6 +#define UVD_JRBC4_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT__SHIFT 0x7 +#define UVD_JRBC4_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT__SHIFT 0x8 +#define UVD_JRBC4_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT__SHIFT 0x9 +#define UVD_JRBC4_UVD_JRBC_STATUS__RB_TRAP_STATUS__SHIFT 0xa +#define UVD_JRBC4_UVD_JRBC_STATUS__PREEMPT_STATUS__SHIFT 0xb +#define UVD_JRBC4_UVD_JRBC_STATUS__IB_TRAP_STATUS__SHIFT 0xc +#define UVD_JRBC4_UVD_JRBC_STATUS__INT_EN__SHIFT 0x10 +#define UVD_JRBC4_UVD_JRBC_STATUS__INT_ACK__SHIFT 0x11 +#define UVD_JRBC4_UVD_JRBC_STATUS__RB_JOB_DONE_MASK 0x00000001L +#define UVD_JRBC4_UVD_JRBC_STATUS__IB_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC4_UVD_JRBC_STATUS__RB_ILLEGAL_CMD_MASK 0x00000004L +#define UVD_JRBC4_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT_MASK 0x00000008L +#define UVD_JRBC4_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT_MASK 0x00000010L +#define UVD_JRBC4_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT_MASK 0x00000020L +#define UVD_JRBC4_UVD_JRBC_STATUS__IB_ILLEGAL_CMD_MASK 0x00000040L +#define UVD_JRBC4_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT_MASK 0x00000080L +#define UVD_JRBC4_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT_MASK 0x00000100L +#define UVD_JRBC4_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT_MASK 0x00000200L +#define UVD_JRBC4_UVD_JRBC_STATUS__RB_TRAP_STATUS_MASK 0x00000400L +#define UVD_JRBC4_UVD_JRBC_STATUS__PREEMPT_STATUS_MASK 0x00000800L +#define UVD_JRBC4_UVD_JRBC_STATUS__IB_TRAP_STATUS_MASK 0x00001000L +#define UVD_JRBC4_UVD_JRBC_STATUS__INT_EN_MASK 0x00010000L +#define UVD_JRBC4_UVD_JRBC_STATUS__INT_ACK_MASK 0x00020000L +//UVD_JRBC4_UVD_JRBC_RB_RPTR +#define UVD_JRBC4_UVD_JRBC_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_JRBC4_UVD_JRBC_RB_RPTR__RB_RPTR_MASK 0x007FFFF0L +//UVD_JRBC4_UVD_JRBC_RB_BUF_STATUS +#define UVD_JRBC4_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC4_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC4_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC4_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC4_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC4_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC4_UVD_JRBC_IB_BUF_STATUS +#define UVD_JRBC4_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC4_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC4_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC4_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC4_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC4_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC4_UVD_JRBC_IB_SIZE_UPDATE +#define UVD_JRBC4_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE__SHIFT 0x4 +#define UVD_JRBC4_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC4_UVD_JRBC_IB_COND_RD_TIMER +#define UVD_JRBC4_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC4_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC4_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC4_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC4_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC4_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC4_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC4_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC4_UVD_JRBC_IB_REF_DATA +#define UVD_JRBC4_UVD_JRBC_IB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC4_UVD_JRBC_IB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC4_UVD_JPEG_PREEMPT_CMD +#define UVD_JRBC4_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN__SHIFT 0x0 +#define UVD_JRBC4_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC4_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD__SHIFT 0x2 +#define UVD_JRBC4_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN_MASK 0x00000001L +#define UVD_JRBC4_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC4_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD_MASK 0x00000004L +//UVD_JRBC4_UVD_JPEG_PREEMPT_FENCE_DATA0 +#define UVD_JRBC4_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0__SHIFT 0x0 +#define UVD_JRBC4_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0_MASK 0xFFFFFFFFL +//UVD_JRBC4_UVD_JPEG_PREEMPT_FENCE_DATA1 +#define UVD_JRBC4_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1__SHIFT 0x0 +#define UVD_JRBC4_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1_MASK 0xFFFFFFFFL +//UVD_JRBC4_UVD_JRBC_RB_SIZE +#define UVD_JRBC4_UVD_JRBC_RB_SIZE__RB_SIZE__SHIFT 0x4 +#define UVD_JRBC4_UVD_JRBC_RB_SIZE__RB_SIZE_MASK 0x00FFFFF0L +//UVD_JRBC4_UVD_JRBC_SCRATCH0 +#define UVD_JRBC4_UVD_JRBC_SCRATCH0__SCRATCH0__SHIFT 0x0 +#define UVD_JRBC4_UVD_JRBC_SCRATCH0__SCRATCH0_MASK 0xFFFFFFFFL + + +// addressBlock: aid_uvd0_uvd_jrbc5_uvd_jrbc_dec +//UVD_JRBC5_UVD_JRBC_RB_WPTR +#define UVD_JRBC5_UVD_JRBC_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_JRBC5_UVD_JRBC_RB_WPTR__RB_WPTR_MASK 0x007FFFF0L +//UVD_JRBC5_UVD_JRBC_RB_CNTL +#define UVD_JRBC5_UVD_JRBC_RB_CNTL__RB_NO_FETCH__SHIFT 0x0 +#define UVD_JRBC5_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN__SHIFT 0x1 +#define UVD_JRBC5_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER__SHIFT 0x4 +#define UVD_JRBC5_UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK 0x00000001L +#define UVD_JRBC5_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK 0x00000002L +#define UVD_JRBC5_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER_MASK 0x0007FFF0L +//UVD_JRBC5_UVD_JRBC_IB_SIZE +#define UVD_JRBC5_UVD_JRBC_IB_SIZE__IB_SIZE__SHIFT 0x4 +#define UVD_JRBC5_UVD_JRBC_IB_SIZE__IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC5_UVD_JRBC_URGENT_CNTL +#define UVD_JRBC5_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK__SHIFT 0x0 +#define UVD_JRBC5_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK_MASK 0x00000003L +//UVD_JRBC5_UVD_JRBC_RB_REF_DATA +#define UVD_JRBC5_UVD_JRBC_RB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC5_UVD_JRBC_RB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC5_UVD_JRBC_RB_COND_RD_TIMER +#define UVD_JRBC5_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC5_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC5_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC5_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC5_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC5_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC5_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC5_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC5_UVD_JRBC_SOFT_RESET +#define UVD_JRBC5_UVD_JRBC_SOFT_RESET__RESET__SHIFT 0x0 +#define UVD_JRBC5_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS__SHIFT 0x11 +#define UVD_JRBC5_UVD_JRBC_SOFT_RESET__RESET_MASK 0x00000001L +#define UVD_JRBC5_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS_MASK 0x00020000L +//UVD_JRBC5_UVD_JRBC_STATUS +#define UVD_JRBC5_UVD_JRBC_STATUS__RB_JOB_DONE__SHIFT 0x0 +#define UVD_JRBC5_UVD_JRBC_STATUS__IB_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC5_UVD_JRBC_STATUS__RB_ILLEGAL_CMD__SHIFT 0x2 +#define UVD_JRBC5_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT__SHIFT 0x3 +#define UVD_JRBC5_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT__SHIFT 0x4 +#define UVD_JRBC5_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT__SHIFT 0x5 +#define UVD_JRBC5_UVD_JRBC_STATUS__IB_ILLEGAL_CMD__SHIFT 0x6 +#define UVD_JRBC5_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT__SHIFT 0x7 +#define UVD_JRBC5_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT__SHIFT 0x8 +#define UVD_JRBC5_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT__SHIFT 0x9 +#define UVD_JRBC5_UVD_JRBC_STATUS__RB_TRAP_STATUS__SHIFT 0xa +#define UVD_JRBC5_UVD_JRBC_STATUS__PREEMPT_STATUS__SHIFT 0xb +#define UVD_JRBC5_UVD_JRBC_STATUS__IB_TRAP_STATUS__SHIFT 0xc +#define UVD_JRBC5_UVD_JRBC_STATUS__INT_EN__SHIFT 0x10 +#define UVD_JRBC5_UVD_JRBC_STATUS__INT_ACK__SHIFT 0x11 +#define UVD_JRBC5_UVD_JRBC_STATUS__RB_JOB_DONE_MASK 0x00000001L +#define UVD_JRBC5_UVD_JRBC_STATUS__IB_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC5_UVD_JRBC_STATUS__RB_ILLEGAL_CMD_MASK 0x00000004L +#define UVD_JRBC5_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT_MASK 0x00000008L +#define UVD_JRBC5_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT_MASK 0x00000010L +#define UVD_JRBC5_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT_MASK 0x00000020L +#define UVD_JRBC5_UVD_JRBC_STATUS__IB_ILLEGAL_CMD_MASK 0x00000040L +#define UVD_JRBC5_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT_MASK 0x00000080L +#define UVD_JRBC5_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT_MASK 0x00000100L +#define UVD_JRBC5_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT_MASK 0x00000200L +#define UVD_JRBC5_UVD_JRBC_STATUS__RB_TRAP_STATUS_MASK 0x00000400L +#define UVD_JRBC5_UVD_JRBC_STATUS__PREEMPT_STATUS_MASK 0x00000800L +#define UVD_JRBC5_UVD_JRBC_STATUS__IB_TRAP_STATUS_MASK 0x00001000L +#define UVD_JRBC5_UVD_JRBC_STATUS__INT_EN_MASK 0x00010000L +#define UVD_JRBC5_UVD_JRBC_STATUS__INT_ACK_MASK 0x00020000L +//UVD_JRBC5_UVD_JRBC_RB_RPTR +#define UVD_JRBC5_UVD_JRBC_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_JRBC5_UVD_JRBC_RB_RPTR__RB_RPTR_MASK 0x007FFFF0L +//UVD_JRBC5_UVD_JRBC_RB_BUF_STATUS +#define UVD_JRBC5_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC5_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC5_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC5_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC5_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC5_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC5_UVD_JRBC_IB_BUF_STATUS +#define UVD_JRBC5_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC5_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC5_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC5_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC5_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC5_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC5_UVD_JRBC_IB_SIZE_UPDATE +#define UVD_JRBC5_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE__SHIFT 0x4 +#define UVD_JRBC5_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC5_UVD_JRBC_IB_COND_RD_TIMER +#define UVD_JRBC5_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC5_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC5_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC5_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC5_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC5_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC5_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC5_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC5_UVD_JRBC_IB_REF_DATA +#define UVD_JRBC5_UVD_JRBC_IB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC5_UVD_JRBC_IB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC5_UVD_JPEG_PREEMPT_CMD +#define UVD_JRBC5_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN__SHIFT 0x0 +#define UVD_JRBC5_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC5_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD__SHIFT 0x2 +#define UVD_JRBC5_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN_MASK 0x00000001L +#define UVD_JRBC5_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC5_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD_MASK 0x00000004L +//UVD_JRBC5_UVD_JPEG_PREEMPT_FENCE_DATA0 +#define UVD_JRBC5_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0__SHIFT 0x0 +#define UVD_JRBC5_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0_MASK 0xFFFFFFFFL +//UVD_JRBC5_UVD_JPEG_PREEMPT_FENCE_DATA1 +#define UVD_JRBC5_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1__SHIFT 0x0 +#define UVD_JRBC5_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1_MASK 0xFFFFFFFFL +//UVD_JRBC5_UVD_JRBC_RB_SIZE +#define UVD_JRBC5_UVD_JRBC_RB_SIZE__RB_SIZE__SHIFT 0x4 +#define UVD_JRBC5_UVD_JRBC_RB_SIZE__RB_SIZE_MASK 0x00FFFFF0L +//UVD_JRBC5_UVD_JRBC_SCRATCH0 +#define UVD_JRBC5_UVD_JRBC_SCRATCH0__SCRATCH0__SHIFT 0x0 +#define UVD_JRBC5_UVD_JRBC_SCRATCH0__SCRATCH0_MASK 0xFFFFFFFFL + + +// addressBlock: aid_uvd0_uvd_jrbc6_uvd_jrbc_dec +//UVD_JRBC6_UVD_JRBC_RB_WPTR +#define UVD_JRBC6_UVD_JRBC_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_JRBC6_UVD_JRBC_RB_WPTR__RB_WPTR_MASK 0x007FFFF0L +//UVD_JRBC6_UVD_JRBC_RB_CNTL +#define UVD_JRBC6_UVD_JRBC_RB_CNTL__RB_NO_FETCH__SHIFT 0x0 +#define UVD_JRBC6_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN__SHIFT 0x1 +#define UVD_JRBC6_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER__SHIFT 0x4 +#define UVD_JRBC6_UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK 0x00000001L +#define UVD_JRBC6_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK 0x00000002L +#define UVD_JRBC6_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER_MASK 0x0007FFF0L +//UVD_JRBC6_UVD_JRBC_IB_SIZE +#define UVD_JRBC6_UVD_JRBC_IB_SIZE__IB_SIZE__SHIFT 0x4 +#define UVD_JRBC6_UVD_JRBC_IB_SIZE__IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC6_UVD_JRBC_URGENT_CNTL +#define UVD_JRBC6_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK__SHIFT 0x0 +#define UVD_JRBC6_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK_MASK 0x00000003L +//UVD_JRBC6_UVD_JRBC_RB_REF_DATA +#define UVD_JRBC6_UVD_JRBC_RB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC6_UVD_JRBC_RB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC6_UVD_JRBC_RB_COND_RD_TIMER +#define UVD_JRBC6_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC6_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC6_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC6_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC6_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC6_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC6_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC6_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC6_UVD_JRBC_SOFT_RESET +#define UVD_JRBC6_UVD_JRBC_SOFT_RESET__RESET__SHIFT 0x0 +#define UVD_JRBC6_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS__SHIFT 0x11 +#define UVD_JRBC6_UVD_JRBC_SOFT_RESET__RESET_MASK 0x00000001L +#define UVD_JRBC6_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS_MASK 0x00020000L +//UVD_JRBC6_UVD_JRBC_STATUS +#define UVD_JRBC6_UVD_JRBC_STATUS__RB_JOB_DONE__SHIFT 0x0 +#define UVD_JRBC6_UVD_JRBC_STATUS__IB_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC6_UVD_JRBC_STATUS__RB_ILLEGAL_CMD__SHIFT 0x2 +#define UVD_JRBC6_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT__SHIFT 0x3 +#define UVD_JRBC6_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT__SHIFT 0x4 +#define UVD_JRBC6_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT__SHIFT 0x5 +#define UVD_JRBC6_UVD_JRBC_STATUS__IB_ILLEGAL_CMD__SHIFT 0x6 +#define UVD_JRBC6_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT__SHIFT 0x7 +#define UVD_JRBC6_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT__SHIFT 0x8 +#define UVD_JRBC6_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT__SHIFT 0x9 +#define UVD_JRBC6_UVD_JRBC_STATUS__RB_TRAP_STATUS__SHIFT 0xa +#define UVD_JRBC6_UVD_JRBC_STATUS__PREEMPT_STATUS__SHIFT 0xb +#define UVD_JRBC6_UVD_JRBC_STATUS__IB_TRAP_STATUS__SHIFT 0xc +#define UVD_JRBC6_UVD_JRBC_STATUS__INT_EN__SHIFT 0x10 +#define UVD_JRBC6_UVD_JRBC_STATUS__INT_ACK__SHIFT 0x11 +#define UVD_JRBC6_UVD_JRBC_STATUS__RB_JOB_DONE_MASK 0x00000001L +#define UVD_JRBC6_UVD_JRBC_STATUS__IB_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC6_UVD_JRBC_STATUS__RB_ILLEGAL_CMD_MASK 0x00000004L +#define UVD_JRBC6_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT_MASK 0x00000008L +#define UVD_JRBC6_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT_MASK 0x00000010L +#define UVD_JRBC6_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT_MASK 0x00000020L +#define UVD_JRBC6_UVD_JRBC_STATUS__IB_ILLEGAL_CMD_MASK 0x00000040L +#define UVD_JRBC6_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT_MASK 0x00000080L +#define UVD_JRBC6_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT_MASK 0x00000100L +#define UVD_JRBC6_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT_MASK 0x00000200L +#define UVD_JRBC6_UVD_JRBC_STATUS__RB_TRAP_STATUS_MASK 0x00000400L +#define UVD_JRBC6_UVD_JRBC_STATUS__PREEMPT_STATUS_MASK 0x00000800L +#define UVD_JRBC6_UVD_JRBC_STATUS__IB_TRAP_STATUS_MASK 0x00001000L +#define UVD_JRBC6_UVD_JRBC_STATUS__INT_EN_MASK 0x00010000L +#define UVD_JRBC6_UVD_JRBC_STATUS__INT_ACK_MASK 0x00020000L +//UVD_JRBC6_UVD_JRBC_RB_RPTR +#define UVD_JRBC6_UVD_JRBC_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_JRBC6_UVD_JRBC_RB_RPTR__RB_RPTR_MASK 0x007FFFF0L +//UVD_JRBC6_UVD_JRBC_RB_BUF_STATUS +#define UVD_JRBC6_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC6_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC6_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC6_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC6_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC6_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC6_UVD_JRBC_IB_BUF_STATUS +#define UVD_JRBC6_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC6_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC6_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC6_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC6_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC6_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC6_UVD_JRBC_IB_SIZE_UPDATE +#define UVD_JRBC6_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE__SHIFT 0x4 +#define UVD_JRBC6_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC6_UVD_JRBC_IB_COND_RD_TIMER +#define UVD_JRBC6_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC6_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC6_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC6_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC6_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC6_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC6_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC6_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC6_UVD_JRBC_IB_REF_DATA +#define UVD_JRBC6_UVD_JRBC_IB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC6_UVD_JRBC_IB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC6_UVD_JPEG_PREEMPT_CMD +#define UVD_JRBC6_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN__SHIFT 0x0 +#define UVD_JRBC6_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC6_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD__SHIFT 0x2 +#define UVD_JRBC6_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN_MASK 0x00000001L +#define UVD_JRBC6_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC6_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD_MASK 0x00000004L +//UVD_JRBC6_UVD_JPEG_PREEMPT_FENCE_DATA0 +#define UVD_JRBC6_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0__SHIFT 0x0 +#define UVD_JRBC6_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0_MASK 0xFFFFFFFFL +//UVD_JRBC6_UVD_JPEG_PREEMPT_FENCE_DATA1 +#define UVD_JRBC6_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1__SHIFT 0x0 +#define UVD_JRBC6_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1_MASK 0xFFFFFFFFL +//UVD_JRBC6_UVD_JRBC_RB_SIZE +#define UVD_JRBC6_UVD_JRBC_RB_SIZE__RB_SIZE__SHIFT 0x4 +#define UVD_JRBC6_UVD_JRBC_RB_SIZE__RB_SIZE_MASK 0x00FFFFF0L +//UVD_JRBC6_UVD_JRBC_SCRATCH0 +#define UVD_JRBC6_UVD_JRBC_SCRATCH0__SCRATCH0__SHIFT 0x0 +#define UVD_JRBC6_UVD_JRBC_SCRATCH0__SCRATCH0_MASK 0xFFFFFFFFL + + +// addressBlock: aid_uvd0_uvd_jrbc7_uvd_jrbc_dec +//UVD_JRBC7_UVD_JRBC_RB_WPTR +#define UVD_JRBC7_UVD_JRBC_RB_WPTR__RB_WPTR__SHIFT 0x4 +#define UVD_JRBC7_UVD_JRBC_RB_WPTR__RB_WPTR_MASK 0x007FFFF0L +//UVD_JRBC7_UVD_JRBC_RB_CNTL +#define UVD_JRBC7_UVD_JRBC_RB_CNTL__RB_NO_FETCH__SHIFT 0x0 +#define UVD_JRBC7_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN__SHIFT 0x1 +#define UVD_JRBC7_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER__SHIFT 0x4 +#define UVD_JRBC7_UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK 0x00000001L +#define UVD_JRBC7_UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK 0x00000002L +#define UVD_JRBC7_UVD_JRBC_RB_CNTL__RB_PRE_WRITE_TIMER_MASK 0x0007FFF0L +//UVD_JRBC7_UVD_JRBC_IB_SIZE +#define UVD_JRBC7_UVD_JRBC_IB_SIZE__IB_SIZE__SHIFT 0x4 +#define UVD_JRBC7_UVD_JRBC_IB_SIZE__IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC7_UVD_JRBC_URGENT_CNTL +#define UVD_JRBC7_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK__SHIFT 0x0 +#define UVD_JRBC7_UVD_JRBC_URGENT_CNTL__CMD_READ_REQ_PRIORITY_MARK_MASK 0x00000003L +//UVD_JRBC7_UVD_JRBC_RB_REF_DATA +#define UVD_JRBC7_UVD_JRBC_RB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC7_UVD_JRBC_RB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC7_UVD_JRBC_RB_COND_RD_TIMER +#define UVD_JRBC7_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC7_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC7_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC7_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC7_UVD_JRBC_RB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC7_UVD_JRBC_RB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC7_UVD_JRBC_RB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC7_UVD_JRBC_RB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC7_UVD_JRBC_SOFT_RESET +#define UVD_JRBC7_UVD_JRBC_SOFT_RESET__RESET__SHIFT 0x0 +#define UVD_JRBC7_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS__SHIFT 0x11 +#define UVD_JRBC7_UVD_JRBC_SOFT_RESET__RESET_MASK 0x00000001L +#define UVD_JRBC7_UVD_JRBC_SOFT_RESET__SCLK_RESET_STATUS_MASK 0x00020000L +//UVD_JRBC7_UVD_JRBC_STATUS +#define UVD_JRBC7_UVD_JRBC_STATUS__RB_JOB_DONE__SHIFT 0x0 +#define UVD_JRBC7_UVD_JRBC_STATUS__IB_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC7_UVD_JRBC_STATUS__RB_ILLEGAL_CMD__SHIFT 0x2 +#define UVD_JRBC7_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT__SHIFT 0x3 +#define UVD_JRBC7_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT__SHIFT 0x4 +#define UVD_JRBC7_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT__SHIFT 0x5 +#define UVD_JRBC7_UVD_JRBC_STATUS__IB_ILLEGAL_CMD__SHIFT 0x6 +#define UVD_JRBC7_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT__SHIFT 0x7 +#define UVD_JRBC7_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT__SHIFT 0x8 +#define UVD_JRBC7_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT__SHIFT 0x9 +#define UVD_JRBC7_UVD_JRBC_STATUS__RB_TRAP_STATUS__SHIFT 0xa +#define UVD_JRBC7_UVD_JRBC_STATUS__PREEMPT_STATUS__SHIFT 0xb +#define UVD_JRBC7_UVD_JRBC_STATUS__IB_TRAP_STATUS__SHIFT 0xc +#define UVD_JRBC7_UVD_JRBC_STATUS__INT_EN__SHIFT 0x10 +#define UVD_JRBC7_UVD_JRBC_STATUS__INT_ACK__SHIFT 0x11 +#define UVD_JRBC7_UVD_JRBC_STATUS__RB_JOB_DONE_MASK 0x00000001L +#define UVD_JRBC7_UVD_JRBC_STATUS__IB_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC7_UVD_JRBC_STATUS__RB_ILLEGAL_CMD_MASK 0x00000004L +#define UVD_JRBC7_UVD_JRBC_STATUS__RB_COND_REG_RD_TIMEOUT_MASK 0x00000008L +#define UVD_JRBC7_UVD_JRBC_STATUS__RB_MEM_WR_TIMEOUT_MASK 0x00000010L +#define UVD_JRBC7_UVD_JRBC_STATUS__RB_MEM_RD_TIMEOUT_MASK 0x00000020L +#define UVD_JRBC7_UVD_JRBC_STATUS__IB_ILLEGAL_CMD_MASK 0x00000040L +#define UVD_JRBC7_UVD_JRBC_STATUS__IB_COND_REG_RD_TIMEOUT_MASK 0x00000080L +#define UVD_JRBC7_UVD_JRBC_STATUS__IB_MEM_WR_TIMEOUT_MASK 0x00000100L +#define UVD_JRBC7_UVD_JRBC_STATUS__IB_MEM_RD_TIMEOUT_MASK 0x00000200L +#define UVD_JRBC7_UVD_JRBC_STATUS__RB_TRAP_STATUS_MASK 0x00000400L +#define UVD_JRBC7_UVD_JRBC_STATUS__PREEMPT_STATUS_MASK 0x00000800L +#define UVD_JRBC7_UVD_JRBC_STATUS__IB_TRAP_STATUS_MASK 0x00001000L +#define UVD_JRBC7_UVD_JRBC_STATUS__INT_EN_MASK 0x00010000L +#define UVD_JRBC7_UVD_JRBC_STATUS__INT_ACK_MASK 0x00020000L +//UVD_JRBC7_UVD_JRBC_RB_RPTR +#define UVD_JRBC7_UVD_JRBC_RB_RPTR__RB_RPTR__SHIFT 0x4 +#define UVD_JRBC7_UVD_JRBC_RB_RPTR__RB_RPTR_MASK 0x007FFFF0L +//UVD_JRBC7_UVD_JRBC_RB_BUF_STATUS +#define UVD_JRBC7_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC7_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC7_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC7_UVD_JRBC_RB_BUF_STATUS__RB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC7_UVD_JRBC_RB_BUF_STATUS__RB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC7_UVD_JRBC_RB_BUF_STATUS__RB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC7_UVD_JRBC_IB_BUF_STATUS +#define UVD_JRBC7_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID__SHIFT 0x0 +#define UVD_JRBC7_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR__SHIFT 0x10 +#define UVD_JRBC7_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR__SHIFT 0x18 +#define UVD_JRBC7_UVD_JRBC_IB_BUF_STATUS__IB_BUF_VALID_MASK 0x0000FFFFL +#define UVD_JRBC7_UVD_JRBC_IB_BUF_STATUS__IB_BUF_RD_ADDR_MASK 0x000F0000L +#define UVD_JRBC7_UVD_JRBC_IB_BUF_STATUS__IB_BUF_WR_ADDR_MASK 0x03000000L +//UVD_JRBC7_UVD_JRBC_IB_SIZE_UPDATE +#define UVD_JRBC7_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE__SHIFT 0x4 +#define UVD_JRBC7_UVD_JRBC_IB_SIZE_UPDATE__REMAIN_IB_SIZE_MASK 0x007FFFF0L +//UVD_JRBC7_UVD_JRBC_IB_COND_RD_TIMER +#define UVD_JRBC7_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT__SHIFT 0x0 +#define UVD_JRBC7_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT__SHIFT 0x10 +#define UVD_JRBC7_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN__SHIFT 0x18 +#define UVD_JRBC7_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN__SHIFT 0x19 +#define UVD_JRBC7_UVD_JRBC_IB_COND_RD_TIMER__RETRY_TIMER_CNT_MASK 0x0000FFFFL +#define UVD_JRBC7_UVD_JRBC_IB_COND_RD_TIMER__RETRY_INTERVAL_CNT_MASK 0x00FF0000L +#define UVD_JRBC7_UVD_JRBC_IB_COND_RD_TIMER__CONTINUOUS_POLL_EN_MASK 0x01000000L +#define UVD_JRBC7_UVD_JRBC_IB_COND_RD_TIMER__MEM_TIMEOUT_EN_MASK 0x02000000L +//UVD_JRBC7_UVD_JRBC_IB_REF_DATA +#define UVD_JRBC7_UVD_JRBC_IB_REF_DATA__REF_DATA__SHIFT 0x0 +#define UVD_JRBC7_UVD_JRBC_IB_REF_DATA__REF_DATA_MASK 0xFFFFFFFFL +//UVD_JRBC7_UVD_JPEG_PREEMPT_CMD +#define UVD_JRBC7_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN__SHIFT 0x0 +#define UVD_JRBC7_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE__SHIFT 0x1 +#define UVD_JRBC7_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD__SHIFT 0x2 +#define UVD_JRBC7_UVD_JPEG_PREEMPT_CMD__PREEMPT_EN_MASK 0x00000001L +#define UVD_JRBC7_UVD_JPEG_PREEMPT_CMD__WAIT_JPEG_JOB_DONE_MASK 0x00000002L +#define UVD_JRBC7_UVD_JPEG_PREEMPT_CMD__PREEMPT_FENCE_CMD_MASK 0x00000004L +//UVD_JRBC7_UVD_JPEG_PREEMPT_FENCE_DATA0 +#define UVD_JRBC7_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0__SHIFT 0x0 +#define UVD_JRBC7_UVD_JPEG_PREEMPT_FENCE_DATA0__PREEMPT_FENCE_DATA0_MASK 0xFFFFFFFFL +//UVD_JRBC7_UVD_JPEG_PREEMPT_FENCE_DATA1 +#define UVD_JRBC7_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1__SHIFT 0x0 +#define UVD_JRBC7_UVD_JPEG_PREEMPT_FENCE_DATA1__PREEMPT_FENCE_DATA1_MASK 0xFFFFFFFFL +//UVD_JRBC7_UVD_JRBC_RB_SIZE +#define UVD_JRBC7_UVD_JRBC_RB_SIZE__RB_SIZE__SHIFT 0x4 +#define UVD_JRBC7_UVD_JRBC_RB_SIZE__RB_SIZE_MASK 0x00FFFFF0L +//UVD_JRBC7_UVD_JRBC_SCRATCH0 +#define UVD_JRBC7_UVD_JRBC_SCRATCH0__SCRATCH0__SHIFT 0x0 +#define UVD_JRBC7_UVD_JRBC_SCRATCH0__SCRATCH0_MASK 0xFFFFFFFFL + + +// addressBlock: aid_uvd0_uvd_jmi1_uvd_jmi_dec +//UVD_JMI1_UVD_JPEG_DEC_PF_CTRL +#define UVD_JMI1_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS__SHIFT 0x0 +#define UVD_JMI1_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING__SHIFT 0x1 +#define UVD_JMI1_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS_MASK 0x00000001L +#define UVD_JMI1_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING_MASK 0x00000002L +//UVD_JMI1_UVD_LMI_JRBC_CTRL +#define UVD_JMI1_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI1_UVD_LMI_JRBC_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI1_UVD_LMI_JRBC_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI1_UVD_LMI_JRBC_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI1_UVD_LMI_JRBC_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI1_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI1_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI1_UVD_LMI_JRBC_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI1_UVD_LMI_JRBC_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI1_UVD_LMI_JRBC_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI1_UVD_LMI_JRBC_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI1_UVD_LMI_JPEG_CTRL +#define UVD_JMI1_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI1_UVD_LMI_JPEG_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI1_UVD_LMI_JPEG_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI1_UVD_LMI_JPEG_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI1_UVD_LMI_JPEG_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI1_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI1_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI1_UVD_LMI_JPEG_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI1_UVD_LMI_JPEG_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI1_UVD_LMI_JPEG_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI1_UVD_LMI_JPEG_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI1_JPEG_LMI_DROP +#define UVD_JMI1_JPEG_LMI_DROP__JPEG_WR_DROP__SHIFT 0x0 +#define UVD_JMI1_JPEG_LMI_DROP__JRBC_WR_DROP__SHIFT 0x1 +#define UVD_JMI1_JPEG_LMI_DROP__JPEG_RD_DROP__SHIFT 0x2 +#define UVD_JMI1_JPEG_LMI_DROP__JRBC_RD_DROP__SHIFT 0x3 +#define UVD_JMI1_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP__SHIFT 0x4 +#define UVD_JMI1_JPEG_LMI_DROP__JPEG_WR_DROP_MASK 0x00000001L +#define UVD_JMI1_JPEG_LMI_DROP__JRBC_WR_DROP_MASK 0x00000002L +#define UVD_JMI1_JPEG_LMI_DROP__JPEG_RD_DROP_MASK 0x00000004L +#define UVD_JMI1_JPEG_LMI_DROP__JRBC_RD_DROP_MASK 0x00000008L +#define UVD_JMI1_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP_MASK 0x00000010L +//UVD_JMI1_UVD_LMI_JRBC_IB_VMID +#define UVD_JMI1_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID__SHIFT 0x4 +#define UVD_JMI1_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI1_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI1_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI1_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI1_UVD_LMI_JRBC_RB_VMID +#define UVD_JMI1_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID__SHIFT 0x4 +#define UVD_JMI1_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI1_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI1_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI1_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI1_UVD_LMI_JPEG_VMID +#define UVD_JMI1_UVD_LMI_JPEG_VMID__JPEG_RD_VMID__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JPEG_VMID__JPEG_WR_VMID__SHIFT 0x4 +#define UVD_JMI1_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID__SHIFT 0x8 +#define UVD_JMI1_UVD_LMI_JPEG_VMID__JPEG_RD_VMID_MASK 0x0000000FL +#define UVD_JMI1_UVD_LMI_JPEG_VMID__JPEG_WR_VMID_MASK 0x000000F0L +#define UVD_JMI1_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID_MASK 0x00000F00L +//UVD_JMI1_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW +#define UVD_JMI1_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH +#define UVD_JMI1_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JRBC_RB_64BIT_BAR_LOW +#define UVD_JMI1_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH +#define UVD_JMI1_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI1_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI1_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JPEG_PREEMPT_VMID +#define UVD_JMI1_UVD_LMI_JPEG_PREEMPT_VMID__VMID__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JPEG_PREEMPT_VMID__VMID_MASK 0x0000000FL +//UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP__SHIFT 0x0 +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP__SHIFT 0x2 +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP__SHIFT 0x4 +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP__SHIFT 0x6 +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP__SHIFT 0x8 +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP__SHIFT 0xa +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP__SHIFT 0xc +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP__SHIFT 0xe +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP__SHIFT 0x10 +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP_MASK 0x00000003L +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP_MASK 0x0000000CL +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP_MASK 0x00000030L +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP_MASK 0x000000C0L +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP_MASK 0x00000300L +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP_MASK 0x00000C00L +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP_MASK 0x00003000L +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP_MASK 0x0000C000L +#define UVD_JMI1_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP_MASK 0x00030000L +//UVD_JMI1_UVD_JMI_ATOMIC_CNTL +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en__SHIFT 0x0 +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL__atomic_max_burst__SHIFT 0x1 +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop__SHIFT 0x5 +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en__SHIFT 0x6 +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG__SHIFT 0x7 +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE__SHIFT 0xb +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en_MASK 0x00000001L +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL__atomic_max_burst_MASK 0x0000001EL +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop_MASK 0x00000020L +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en_MASK 0x00000040L +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG_MASK 0x00000780L +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE_MASK 0x00000800L +//UVD_JMI1_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW +#define UVD_JMI1_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI1_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH +#define UVD_JMI1_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI1_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JPEG_READ_64BIT_BAR_LOW +#define UVD_JMI1_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH +#define UVD_JMI1_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW +#define UVD_JMI1_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH +#define UVD_JMI1_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JRBC_IB_64BIT_BAR_LOW +#define UVD_JMI1_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH +#define UVD_JMI1_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI1_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI1_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI1_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI1_UVD_JMI_ATOMIC_CNTL2 +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap__SHIFT 0x10 +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP__SHIFT 0x18 +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap_MASK 0x00FF0000L +#define UVD_JMI1_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP_MASK 0xFF000000L + + +// addressBlock: aid_uvd0_uvd_jmi2_uvd_jmi_dec +//UVD_JMI2_UVD_JPEG_DEC_PF_CTRL +#define UVD_JMI2_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS__SHIFT 0x0 +#define UVD_JMI2_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING__SHIFT 0x1 +#define UVD_JMI2_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS_MASK 0x00000001L +#define UVD_JMI2_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING_MASK 0x00000002L +//UVD_JMI2_UVD_LMI_JRBC_CTRL +#define UVD_JMI2_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI2_UVD_LMI_JRBC_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI2_UVD_LMI_JRBC_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI2_UVD_LMI_JRBC_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI2_UVD_LMI_JRBC_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI2_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI2_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI2_UVD_LMI_JRBC_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI2_UVD_LMI_JRBC_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI2_UVD_LMI_JRBC_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI2_UVD_LMI_JRBC_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI2_UVD_LMI_JPEG_CTRL +#define UVD_JMI2_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI2_UVD_LMI_JPEG_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI2_UVD_LMI_JPEG_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI2_UVD_LMI_JPEG_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI2_UVD_LMI_JPEG_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI2_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI2_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI2_UVD_LMI_JPEG_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI2_UVD_LMI_JPEG_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI2_UVD_LMI_JPEG_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI2_UVD_LMI_JPEG_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI2_JPEG_LMI_DROP +#define UVD_JMI2_JPEG_LMI_DROP__JPEG_WR_DROP__SHIFT 0x0 +#define UVD_JMI2_JPEG_LMI_DROP__JRBC_WR_DROP__SHIFT 0x1 +#define UVD_JMI2_JPEG_LMI_DROP__JPEG_RD_DROP__SHIFT 0x2 +#define UVD_JMI2_JPEG_LMI_DROP__JRBC_RD_DROP__SHIFT 0x3 +#define UVD_JMI2_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP__SHIFT 0x4 +#define UVD_JMI2_JPEG_LMI_DROP__JPEG_WR_DROP_MASK 0x00000001L +#define UVD_JMI2_JPEG_LMI_DROP__JRBC_WR_DROP_MASK 0x00000002L +#define UVD_JMI2_JPEG_LMI_DROP__JPEG_RD_DROP_MASK 0x00000004L +#define UVD_JMI2_JPEG_LMI_DROP__JRBC_RD_DROP_MASK 0x00000008L +#define UVD_JMI2_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP_MASK 0x00000010L +//UVD_JMI2_UVD_LMI_JRBC_IB_VMID +#define UVD_JMI2_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID__SHIFT 0x4 +#define UVD_JMI2_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI2_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI2_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI2_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI2_UVD_LMI_JRBC_RB_VMID +#define UVD_JMI2_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID__SHIFT 0x4 +#define UVD_JMI2_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI2_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI2_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI2_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI2_UVD_LMI_JPEG_VMID +#define UVD_JMI2_UVD_LMI_JPEG_VMID__JPEG_RD_VMID__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JPEG_VMID__JPEG_WR_VMID__SHIFT 0x4 +#define UVD_JMI2_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID__SHIFT 0x8 +#define UVD_JMI2_UVD_LMI_JPEG_VMID__JPEG_RD_VMID_MASK 0x0000000FL +#define UVD_JMI2_UVD_LMI_JPEG_VMID__JPEG_WR_VMID_MASK 0x000000F0L +#define UVD_JMI2_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID_MASK 0x00000F00L +//UVD_JMI2_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW +#define UVD_JMI2_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH +#define UVD_JMI2_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JRBC_RB_64BIT_BAR_LOW +#define UVD_JMI2_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH +#define UVD_JMI2_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI2_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI2_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JPEG_PREEMPT_VMID +#define UVD_JMI2_UVD_LMI_JPEG_PREEMPT_VMID__VMID__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JPEG_PREEMPT_VMID__VMID_MASK 0x0000000FL +//UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP__SHIFT 0x0 +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP__SHIFT 0x2 +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP__SHIFT 0x4 +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP__SHIFT 0x6 +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP__SHIFT 0x8 +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP__SHIFT 0xa +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP__SHIFT 0xc +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP__SHIFT 0xe +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP__SHIFT 0x10 +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP_MASK 0x00000003L +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP_MASK 0x0000000CL +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP_MASK 0x00000030L +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP_MASK 0x000000C0L +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP_MASK 0x00000300L +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP_MASK 0x00000C00L +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP_MASK 0x00003000L +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP_MASK 0x0000C000L +#define UVD_JMI2_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP_MASK 0x00030000L +//UVD_JMI2_UVD_JMI_ATOMIC_CNTL +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en__SHIFT 0x0 +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL__atomic_max_burst__SHIFT 0x1 +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop__SHIFT 0x5 +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en__SHIFT 0x6 +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG__SHIFT 0x7 +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE__SHIFT 0xb +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en_MASK 0x00000001L +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL__atomic_max_burst_MASK 0x0000001EL +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop_MASK 0x00000020L +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en_MASK 0x00000040L +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG_MASK 0x00000780L +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE_MASK 0x00000800L +//UVD_JMI2_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW +#define UVD_JMI2_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI2_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH +#define UVD_JMI2_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI2_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JPEG_READ_64BIT_BAR_LOW +#define UVD_JMI2_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH +#define UVD_JMI2_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW +#define UVD_JMI2_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH +#define UVD_JMI2_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JRBC_IB_64BIT_BAR_LOW +#define UVD_JMI2_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH +#define UVD_JMI2_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI2_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI2_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI2_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI2_UVD_JMI_ATOMIC_CNTL2 +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap__SHIFT 0x10 +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP__SHIFT 0x18 +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap_MASK 0x00FF0000L +#define UVD_JMI2_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP_MASK 0xFF000000L + + +// addressBlock: aid_uvd0_uvd_jmi3_uvd_jmi_dec +//UVD_JMI3_UVD_JPEG_DEC_PF_CTRL +#define UVD_JMI3_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS__SHIFT 0x0 +#define UVD_JMI3_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING__SHIFT 0x1 +#define UVD_JMI3_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS_MASK 0x00000001L +#define UVD_JMI3_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING_MASK 0x00000002L +//UVD_JMI3_UVD_LMI_JRBC_CTRL +#define UVD_JMI3_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI3_UVD_LMI_JRBC_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI3_UVD_LMI_JRBC_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI3_UVD_LMI_JRBC_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI3_UVD_LMI_JRBC_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI3_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI3_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI3_UVD_LMI_JRBC_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI3_UVD_LMI_JRBC_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI3_UVD_LMI_JRBC_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI3_UVD_LMI_JRBC_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI3_UVD_LMI_JPEG_CTRL +#define UVD_JMI3_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI3_UVD_LMI_JPEG_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI3_UVD_LMI_JPEG_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI3_UVD_LMI_JPEG_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI3_UVD_LMI_JPEG_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI3_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI3_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI3_UVD_LMI_JPEG_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI3_UVD_LMI_JPEG_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI3_UVD_LMI_JPEG_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI3_UVD_LMI_JPEG_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI3_JPEG_LMI_DROP +#define UVD_JMI3_JPEG_LMI_DROP__JPEG_WR_DROP__SHIFT 0x0 +#define UVD_JMI3_JPEG_LMI_DROP__JRBC_WR_DROP__SHIFT 0x1 +#define UVD_JMI3_JPEG_LMI_DROP__JPEG_RD_DROP__SHIFT 0x2 +#define UVD_JMI3_JPEG_LMI_DROP__JRBC_RD_DROP__SHIFT 0x3 +#define UVD_JMI3_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP__SHIFT 0x4 +#define UVD_JMI3_JPEG_LMI_DROP__JPEG_WR_DROP_MASK 0x00000001L +#define UVD_JMI3_JPEG_LMI_DROP__JRBC_WR_DROP_MASK 0x00000002L +#define UVD_JMI3_JPEG_LMI_DROP__JPEG_RD_DROP_MASK 0x00000004L +#define UVD_JMI3_JPEG_LMI_DROP__JRBC_RD_DROP_MASK 0x00000008L +#define UVD_JMI3_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP_MASK 0x00000010L +//UVD_JMI3_UVD_LMI_JRBC_IB_VMID +#define UVD_JMI3_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID__SHIFT 0x4 +#define UVD_JMI3_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI3_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI3_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI3_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI3_UVD_LMI_JRBC_RB_VMID +#define UVD_JMI3_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID__SHIFT 0x4 +#define UVD_JMI3_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI3_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI3_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI3_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI3_UVD_LMI_JPEG_VMID +#define UVD_JMI3_UVD_LMI_JPEG_VMID__JPEG_RD_VMID__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JPEG_VMID__JPEG_WR_VMID__SHIFT 0x4 +#define UVD_JMI3_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID__SHIFT 0x8 +#define UVD_JMI3_UVD_LMI_JPEG_VMID__JPEG_RD_VMID_MASK 0x0000000FL +#define UVD_JMI3_UVD_LMI_JPEG_VMID__JPEG_WR_VMID_MASK 0x000000F0L +#define UVD_JMI3_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID_MASK 0x00000F00L +//UVD_JMI3_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW +#define UVD_JMI3_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH +#define UVD_JMI3_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JRBC_RB_64BIT_BAR_LOW +#define UVD_JMI3_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH +#define UVD_JMI3_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI3_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI3_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JPEG_PREEMPT_VMID +#define UVD_JMI3_UVD_LMI_JPEG_PREEMPT_VMID__VMID__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JPEG_PREEMPT_VMID__VMID_MASK 0x0000000FL +//UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP__SHIFT 0x0 +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP__SHIFT 0x2 +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP__SHIFT 0x4 +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP__SHIFT 0x6 +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP__SHIFT 0x8 +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP__SHIFT 0xa +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP__SHIFT 0xc +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP__SHIFT 0xe +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP__SHIFT 0x10 +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP_MASK 0x00000003L +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP_MASK 0x0000000CL +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP_MASK 0x00000030L +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP_MASK 0x000000C0L +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP_MASK 0x00000300L +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP_MASK 0x00000C00L +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP_MASK 0x00003000L +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP_MASK 0x0000C000L +#define UVD_JMI3_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP_MASK 0x00030000L +//UVD_JMI3_UVD_JMI_ATOMIC_CNTL +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en__SHIFT 0x0 +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL__atomic_max_burst__SHIFT 0x1 +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop__SHIFT 0x5 +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en__SHIFT 0x6 +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG__SHIFT 0x7 +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE__SHIFT 0xb +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en_MASK 0x00000001L +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL__atomic_max_burst_MASK 0x0000001EL +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop_MASK 0x00000020L +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en_MASK 0x00000040L +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG_MASK 0x00000780L +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE_MASK 0x00000800L +//UVD_JMI3_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW +#define UVD_JMI3_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI3_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH +#define UVD_JMI3_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI3_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JPEG_READ_64BIT_BAR_LOW +#define UVD_JMI3_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH +#define UVD_JMI3_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW +#define UVD_JMI3_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH +#define UVD_JMI3_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JRBC_IB_64BIT_BAR_LOW +#define UVD_JMI3_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH +#define UVD_JMI3_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI3_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI3_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI3_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI3_UVD_JMI_ATOMIC_CNTL2 +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap__SHIFT 0x10 +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP__SHIFT 0x18 +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap_MASK 0x00FF0000L +#define UVD_JMI3_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP_MASK 0xFF000000L + + +// addressBlock: aid_uvd0_uvd_jmi4_uvd_jmi_dec +//UVD_JMI4_UVD_JPEG_DEC_PF_CTRL +#define UVD_JMI4_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS__SHIFT 0x0 +#define UVD_JMI4_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING__SHIFT 0x1 +#define UVD_JMI4_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS_MASK 0x00000001L +#define UVD_JMI4_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING_MASK 0x00000002L +//UVD_JMI4_UVD_LMI_JRBC_CTRL +#define UVD_JMI4_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI4_UVD_LMI_JRBC_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI4_UVD_LMI_JRBC_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI4_UVD_LMI_JRBC_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI4_UVD_LMI_JRBC_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI4_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI4_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI4_UVD_LMI_JRBC_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI4_UVD_LMI_JRBC_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI4_UVD_LMI_JRBC_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI4_UVD_LMI_JRBC_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI4_UVD_LMI_JPEG_CTRL +#define UVD_JMI4_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI4_UVD_LMI_JPEG_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI4_UVD_LMI_JPEG_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI4_UVD_LMI_JPEG_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI4_UVD_LMI_JPEG_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI4_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI4_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI4_UVD_LMI_JPEG_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI4_UVD_LMI_JPEG_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI4_UVD_LMI_JPEG_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI4_UVD_LMI_JPEG_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI4_JPEG_LMI_DROP +#define UVD_JMI4_JPEG_LMI_DROP__JPEG_WR_DROP__SHIFT 0x0 +#define UVD_JMI4_JPEG_LMI_DROP__JRBC_WR_DROP__SHIFT 0x1 +#define UVD_JMI4_JPEG_LMI_DROP__JPEG_RD_DROP__SHIFT 0x2 +#define UVD_JMI4_JPEG_LMI_DROP__JRBC_RD_DROP__SHIFT 0x3 +#define UVD_JMI4_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP__SHIFT 0x4 +#define UVD_JMI4_JPEG_LMI_DROP__JPEG_WR_DROP_MASK 0x00000001L +#define UVD_JMI4_JPEG_LMI_DROP__JRBC_WR_DROP_MASK 0x00000002L +#define UVD_JMI4_JPEG_LMI_DROP__JPEG_RD_DROP_MASK 0x00000004L +#define UVD_JMI4_JPEG_LMI_DROP__JRBC_RD_DROP_MASK 0x00000008L +#define UVD_JMI4_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP_MASK 0x00000010L +//UVD_JMI4_UVD_LMI_JRBC_IB_VMID +#define UVD_JMI4_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID__SHIFT 0x4 +#define UVD_JMI4_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI4_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI4_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI4_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI4_UVD_LMI_JRBC_RB_VMID +#define UVD_JMI4_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID__SHIFT 0x4 +#define UVD_JMI4_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI4_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI4_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI4_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI4_UVD_LMI_JPEG_VMID +#define UVD_JMI4_UVD_LMI_JPEG_VMID__JPEG_RD_VMID__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JPEG_VMID__JPEG_WR_VMID__SHIFT 0x4 +#define UVD_JMI4_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID__SHIFT 0x8 +#define UVD_JMI4_UVD_LMI_JPEG_VMID__JPEG_RD_VMID_MASK 0x0000000FL +#define UVD_JMI4_UVD_LMI_JPEG_VMID__JPEG_WR_VMID_MASK 0x000000F0L +#define UVD_JMI4_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID_MASK 0x00000F00L +//UVD_JMI4_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW +#define UVD_JMI4_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH +#define UVD_JMI4_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JRBC_RB_64BIT_BAR_LOW +#define UVD_JMI4_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH +#define UVD_JMI4_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI4_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI4_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JPEG_PREEMPT_VMID +#define UVD_JMI4_UVD_LMI_JPEG_PREEMPT_VMID__VMID__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JPEG_PREEMPT_VMID__VMID_MASK 0x0000000FL +//UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP__SHIFT 0x0 +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP__SHIFT 0x2 +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP__SHIFT 0x4 +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP__SHIFT 0x6 +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP__SHIFT 0x8 +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP__SHIFT 0xa +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP__SHIFT 0xc +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP__SHIFT 0xe +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP__SHIFT 0x10 +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP_MASK 0x00000003L +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP_MASK 0x0000000CL +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP_MASK 0x00000030L +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP_MASK 0x000000C0L +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP_MASK 0x00000300L +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP_MASK 0x00000C00L +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP_MASK 0x00003000L +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP_MASK 0x0000C000L +#define UVD_JMI4_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP_MASK 0x00030000L +//UVD_JMI4_UVD_JMI_ATOMIC_CNTL +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en__SHIFT 0x0 +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL__atomic_max_burst__SHIFT 0x1 +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop__SHIFT 0x5 +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en__SHIFT 0x6 +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG__SHIFT 0x7 +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE__SHIFT 0xb +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en_MASK 0x00000001L +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL__atomic_max_burst_MASK 0x0000001EL +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop_MASK 0x00000020L +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en_MASK 0x00000040L +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG_MASK 0x00000780L +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE_MASK 0x00000800L +//UVD_JMI4_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW +#define UVD_JMI4_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI4_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH +#define UVD_JMI4_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI4_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JPEG_READ_64BIT_BAR_LOW +#define UVD_JMI4_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH +#define UVD_JMI4_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW +#define UVD_JMI4_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH +#define UVD_JMI4_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JRBC_IB_64BIT_BAR_LOW +#define UVD_JMI4_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH +#define UVD_JMI4_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI4_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI4_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI4_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI4_UVD_JMI_ATOMIC_CNTL2 +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap__SHIFT 0x10 +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP__SHIFT 0x18 +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap_MASK 0x00FF0000L +#define UVD_JMI4_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP_MASK 0xFF000000L + + +// addressBlock: aid_uvd0_uvd_jmi5_uvd_jmi_dec +//UVD_JMI5_UVD_JPEG_DEC_PF_CTRL +#define UVD_JMI5_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS__SHIFT 0x0 +#define UVD_JMI5_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING__SHIFT 0x1 +#define UVD_JMI5_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS_MASK 0x00000001L +#define UVD_JMI5_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING_MASK 0x00000002L +//UVD_JMI5_UVD_LMI_JRBC_CTRL +#define UVD_JMI5_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI5_UVD_LMI_JRBC_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI5_UVD_LMI_JRBC_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI5_UVD_LMI_JRBC_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI5_UVD_LMI_JRBC_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI5_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI5_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI5_UVD_LMI_JRBC_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI5_UVD_LMI_JRBC_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI5_UVD_LMI_JRBC_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI5_UVD_LMI_JRBC_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI5_UVD_LMI_JPEG_CTRL +#define UVD_JMI5_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI5_UVD_LMI_JPEG_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI5_UVD_LMI_JPEG_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI5_UVD_LMI_JPEG_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI5_UVD_LMI_JPEG_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI5_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI5_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI5_UVD_LMI_JPEG_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI5_UVD_LMI_JPEG_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI5_UVD_LMI_JPEG_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI5_UVD_LMI_JPEG_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI5_JPEG_LMI_DROP +#define UVD_JMI5_JPEG_LMI_DROP__JPEG_WR_DROP__SHIFT 0x0 +#define UVD_JMI5_JPEG_LMI_DROP__JRBC_WR_DROP__SHIFT 0x1 +#define UVD_JMI5_JPEG_LMI_DROP__JPEG_RD_DROP__SHIFT 0x2 +#define UVD_JMI5_JPEG_LMI_DROP__JRBC_RD_DROP__SHIFT 0x3 +#define UVD_JMI5_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP__SHIFT 0x4 +#define UVD_JMI5_JPEG_LMI_DROP__JPEG_WR_DROP_MASK 0x00000001L +#define UVD_JMI5_JPEG_LMI_DROP__JRBC_WR_DROP_MASK 0x00000002L +#define UVD_JMI5_JPEG_LMI_DROP__JPEG_RD_DROP_MASK 0x00000004L +#define UVD_JMI5_JPEG_LMI_DROP__JRBC_RD_DROP_MASK 0x00000008L +#define UVD_JMI5_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP_MASK 0x00000010L +//UVD_JMI5_UVD_LMI_JRBC_IB_VMID +#define UVD_JMI5_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID__SHIFT 0x4 +#define UVD_JMI5_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI5_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI5_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI5_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI5_UVD_LMI_JRBC_RB_VMID +#define UVD_JMI5_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID__SHIFT 0x4 +#define UVD_JMI5_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI5_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI5_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI5_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI5_UVD_LMI_JPEG_VMID +#define UVD_JMI5_UVD_LMI_JPEG_VMID__JPEG_RD_VMID__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JPEG_VMID__JPEG_WR_VMID__SHIFT 0x4 +#define UVD_JMI5_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID__SHIFT 0x8 +#define UVD_JMI5_UVD_LMI_JPEG_VMID__JPEG_RD_VMID_MASK 0x0000000FL +#define UVD_JMI5_UVD_LMI_JPEG_VMID__JPEG_WR_VMID_MASK 0x000000F0L +#define UVD_JMI5_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID_MASK 0x00000F00L +//UVD_JMI5_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW +#define UVD_JMI5_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH +#define UVD_JMI5_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JRBC_RB_64BIT_BAR_LOW +#define UVD_JMI5_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH +#define UVD_JMI5_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI5_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI5_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JPEG_PREEMPT_VMID +#define UVD_JMI5_UVD_LMI_JPEG_PREEMPT_VMID__VMID__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JPEG_PREEMPT_VMID__VMID_MASK 0x0000000FL +//UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP__SHIFT 0x0 +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP__SHIFT 0x2 +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP__SHIFT 0x4 +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP__SHIFT 0x6 +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP__SHIFT 0x8 +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP__SHIFT 0xa +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP__SHIFT 0xc +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP__SHIFT 0xe +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP__SHIFT 0x10 +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP_MASK 0x00000003L +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP_MASK 0x0000000CL +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP_MASK 0x00000030L +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP_MASK 0x000000C0L +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP_MASK 0x00000300L +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP_MASK 0x00000C00L +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP_MASK 0x00003000L +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP_MASK 0x0000C000L +#define UVD_JMI5_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP_MASK 0x00030000L +//UVD_JMI5_UVD_JMI_ATOMIC_CNTL +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en__SHIFT 0x0 +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL__atomic_max_burst__SHIFT 0x1 +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop__SHIFT 0x5 +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en__SHIFT 0x6 +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG__SHIFT 0x7 +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE__SHIFT 0xb +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en_MASK 0x00000001L +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL__atomic_max_burst_MASK 0x0000001EL +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop_MASK 0x00000020L +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en_MASK 0x00000040L +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG_MASK 0x00000780L +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE_MASK 0x00000800L +//UVD_JMI5_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW +#define UVD_JMI5_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI5_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH +#define UVD_JMI5_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI5_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JPEG_READ_64BIT_BAR_LOW +#define UVD_JMI5_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH +#define UVD_JMI5_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW +#define UVD_JMI5_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH +#define UVD_JMI5_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JRBC_IB_64BIT_BAR_LOW +#define UVD_JMI5_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH +#define UVD_JMI5_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI5_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI5_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI5_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI5_UVD_JMI_ATOMIC_CNTL2 +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap__SHIFT 0x10 +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP__SHIFT 0x18 +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap_MASK 0x00FF0000L +#define UVD_JMI5_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP_MASK 0xFF000000L + + +// addressBlock: aid_uvd0_uvd_jmi6_uvd_jmi_dec +//UVD_JMI6_UVD_JPEG_DEC_PF_CTRL +#define UVD_JMI6_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS__SHIFT 0x0 +#define UVD_JMI6_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING__SHIFT 0x1 +#define UVD_JMI6_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS_MASK 0x00000001L +#define UVD_JMI6_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING_MASK 0x00000002L +//UVD_JMI6_UVD_LMI_JRBC_CTRL +#define UVD_JMI6_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI6_UVD_LMI_JRBC_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI6_UVD_LMI_JRBC_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI6_UVD_LMI_JRBC_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI6_UVD_LMI_JRBC_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI6_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI6_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI6_UVD_LMI_JRBC_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI6_UVD_LMI_JRBC_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI6_UVD_LMI_JRBC_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI6_UVD_LMI_JRBC_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI6_UVD_LMI_JPEG_CTRL +#define UVD_JMI6_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI6_UVD_LMI_JPEG_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI6_UVD_LMI_JPEG_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI6_UVD_LMI_JPEG_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI6_UVD_LMI_JPEG_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI6_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI6_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI6_UVD_LMI_JPEG_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI6_UVD_LMI_JPEG_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI6_UVD_LMI_JPEG_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI6_UVD_LMI_JPEG_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI6_JPEG_LMI_DROP +#define UVD_JMI6_JPEG_LMI_DROP__JPEG_WR_DROP__SHIFT 0x0 +#define UVD_JMI6_JPEG_LMI_DROP__JRBC_WR_DROP__SHIFT 0x1 +#define UVD_JMI6_JPEG_LMI_DROP__JPEG_RD_DROP__SHIFT 0x2 +#define UVD_JMI6_JPEG_LMI_DROP__JRBC_RD_DROP__SHIFT 0x3 +#define UVD_JMI6_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP__SHIFT 0x4 +#define UVD_JMI6_JPEG_LMI_DROP__JPEG_WR_DROP_MASK 0x00000001L +#define UVD_JMI6_JPEG_LMI_DROP__JRBC_WR_DROP_MASK 0x00000002L +#define UVD_JMI6_JPEG_LMI_DROP__JPEG_RD_DROP_MASK 0x00000004L +#define UVD_JMI6_JPEG_LMI_DROP__JRBC_RD_DROP_MASK 0x00000008L +#define UVD_JMI6_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP_MASK 0x00000010L +//UVD_JMI6_UVD_LMI_JRBC_IB_VMID +#define UVD_JMI6_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID__SHIFT 0x4 +#define UVD_JMI6_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI6_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI6_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI6_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI6_UVD_LMI_JRBC_RB_VMID +#define UVD_JMI6_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID__SHIFT 0x4 +#define UVD_JMI6_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI6_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI6_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI6_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI6_UVD_LMI_JPEG_VMID +#define UVD_JMI6_UVD_LMI_JPEG_VMID__JPEG_RD_VMID__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JPEG_VMID__JPEG_WR_VMID__SHIFT 0x4 +#define UVD_JMI6_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID__SHIFT 0x8 +#define UVD_JMI6_UVD_LMI_JPEG_VMID__JPEG_RD_VMID_MASK 0x0000000FL +#define UVD_JMI6_UVD_LMI_JPEG_VMID__JPEG_WR_VMID_MASK 0x000000F0L +#define UVD_JMI6_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID_MASK 0x00000F00L +//UVD_JMI6_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW +#define UVD_JMI6_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH +#define UVD_JMI6_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JRBC_RB_64BIT_BAR_LOW +#define UVD_JMI6_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH +#define UVD_JMI6_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI6_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI6_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JPEG_PREEMPT_VMID +#define UVD_JMI6_UVD_LMI_JPEG_PREEMPT_VMID__VMID__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JPEG_PREEMPT_VMID__VMID_MASK 0x0000000FL +//UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP__SHIFT 0x0 +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP__SHIFT 0x2 +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP__SHIFT 0x4 +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP__SHIFT 0x6 +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP__SHIFT 0x8 +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP__SHIFT 0xa +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP__SHIFT 0xc +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP__SHIFT 0xe +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP__SHIFT 0x10 +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP_MASK 0x00000003L +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP_MASK 0x0000000CL +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP_MASK 0x00000030L +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP_MASK 0x000000C0L +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP_MASK 0x00000300L +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP_MASK 0x00000C00L +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP_MASK 0x00003000L +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP_MASK 0x0000C000L +#define UVD_JMI6_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP_MASK 0x00030000L +//UVD_JMI6_UVD_JMI_ATOMIC_CNTL +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en__SHIFT 0x0 +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL__atomic_max_burst__SHIFT 0x1 +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop__SHIFT 0x5 +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en__SHIFT 0x6 +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG__SHIFT 0x7 +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE__SHIFT 0xb +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en_MASK 0x00000001L +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL__atomic_max_burst_MASK 0x0000001EL +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop_MASK 0x00000020L +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en_MASK 0x00000040L +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG_MASK 0x00000780L +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE_MASK 0x00000800L +//UVD_JMI6_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW +#define UVD_JMI6_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI6_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH +#define UVD_JMI6_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI6_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JPEG_READ_64BIT_BAR_LOW +#define UVD_JMI6_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH +#define UVD_JMI6_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW +#define UVD_JMI6_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH +#define UVD_JMI6_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JRBC_IB_64BIT_BAR_LOW +#define UVD_JMI6_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH +#define UVD_JMI6_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI6_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI6_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI6_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI6_UVD_JMI_ATOMIC_CNTL2 +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap__SHIFT 0x10 +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP__SHIFT 0x18 +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap_MASK 0x00FF0000L +#define UVD_JMI6_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP_MASK 0xFF000000L + + +// addressBlock: aid_uvd0_uvd_jmi7_uvd_jmi_dec +//UVD_JMI7_UVD_JPEG_DEC_PF_CTRL +#define UVD_JMI7_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS__SHIFT 0x0 +#define UVD_JMI7_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING__SHIFT 0x1 +#define UVD_JMI7_UVD_JPEG_DEC_PF_CTRL__DEC_PF_HANDLING_DIS_MASK 0x00000001L +#define UVD_JMI7_UVD_JPEG_DEC_PF_CTRL__DEC_PF_SW_GATING_MASK 0x00000002L +//UVD_JMI7_UVD_LMI_JRBC_CTRL +#define UVD_JMI7_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI7_UVD_LMI_JRBC_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI7_UVD_LMI_JRBC_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI7_UVD_LMI_JRBC_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI7_UVD_LMI_JRBC_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI7_UVD_LMI_JRBC_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI7_UVD_LMI_JRBC_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI7_UVD_LMI_JRBC_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI7_UVD_LMI_JRBC_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI7_UVD_LMI_JRBC_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI7_UVD_LMI_JRBC_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI7_UVD_LMI_JPEG_CTRL +#define UVD_JMI7_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN__SHIFT 0x1 +#define UVD_JMI7_UVD_LMI_JPEG_CTRL__RD_MAX_BURST__SHIFT 0x4 +#define UVD_JMI7_UVD_LMI_JPEG_CTRL__WR_MAX_BURST__SHIFT 0x8 +#define UVD_JMI7_UVD_LMI_JPEG_CTRL__RD_SWAP__SHIFT 0x14 +#define UVD_JMI7_UVD_LMI_JPEG_CTRL__WR_SWAP__SHIFT 0x16 +#define UVD_JMI7_UVD_LMI_JPEG_CTRL__ARB_RD_WAIT_EN_MASK 0x00000001L +#define UVD_JMI7_UVD_LMI_JPEG_CTRL__ARB_WR_WAIT_EN_MASK 0x00000002L +#define UVD_JMI7_UVD_LMI_JPEG_CTRL__RD_MAX_BURST_MASK 0x000000F0L +#define UVD_JMI7_UVD_LMI_JPEG_CTRL__WR_MAX_BURST_MASK 0x00000F00L +#define UVD_JMI7_UVD_LMI_JPEG_CTRL__RD_SWAP_MASK 0x00300000L +#define UVD_JMI7_UVD_LMI_JPEG_CTRL__WR_SWAP_MASK 0x00C00000L +//UVD_JMI7_JPEG_LMI_DROP +#define UVD_JMI7_JPEG_LMI_DROP__JPEG_WR_DROP__SHIFT 0x0 +#define UVD_JMI7_JPEG_LMI_DROP__JRBC_WR_DROP__SHIFT 0x1 +#define UVD_JMI7_JPEG_LMI_DROP__JPEG_RD_DROP__SHIFT 0x2 +#define UVD_JMI7_JPEG_LMI_DROP__JRBC_RD_DROP__SHIFT 0x3 +#define UVD_JMI7_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP__SHIFT 0x4 +#define UVD_JMI7_JPEG_LMI_DROP__JPEG_WR_DROP_MASK 0x00000001L +#define UVD_JMI7_JPEG_LMI_DROP__JRBC_WR_DROP_MASK 0x00000002L +#define UVD_JMI7_JPEG_LMI_DROP__JPEG_RD_DROP_MASK 0x00000004L +#define UVD_JMI7_JPEG_LMI_DROP__JRBC_RD_DROP_MASK 0x00000008L +#define UVD_JMI7_JPEG_LMI_DROP__JPEG_ATOMIC_WR_DROP_MASK 0x00000010L +//UVD_JMI7_UVD_LMI_JRBC_IB_VMID +#define UVD_JMI7_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID__SHIFT 0x4 +#define UVD_JMI7_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI7_UVD_LMI_JRBC_IB_VMID__IB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI7_UVD_LMI_JRBC_IB_VMID__IB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI7_UVD_LMI_JRBC_IB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI7_UVD_LMI_JRBC_RB_VMID +#define UVD_JMI7_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID__SHIFT 0x4 +#define UVD_JMI7_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID__SHIFT 0x8 +#define UVD_JMI7_UVD_LMI_JRBC_RB_VMID__RB_WR_VMID_MASK 0x0000000FL +#define UVD_JMI7_UVD_LMI_JRBC_RB_VMID__RB_RD_VMID_MASK 0x000000F0L +#define UVD_JMI7_UVD_LMI_JRBC_RB_VMID__MEM_RD_VMID_MASK 0x00000F00L +//UVD_JMI7_UVD_LMI_JPEG_VMID +#define UVD_JMI7_UVD_LMI_JPEG_VMID__JPEG_RD_VMID__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JPEG_VMID__JPEG_WR_VMID__SHIFT 0x4 +#define UVD_JMI7_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID__SHIFT 0x8 +#define UVD_JMI7_UVD_LMI_JPEG_VMID__JPEG_RD_VMID_MASK 0x0000000FL +#define UVD_JMI7_UVD_LMI_JPEG_VMID__JPEG_WR_VMID_MASK 0x000000F0L +#define UVD_JMI7_UVD_LMI_JPEG_VMID__ATOMIC_USER0_WR_VMID_MASK 0x00000F00L +//UVD_JMI7_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW +#define UVD_JMI7_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH +#define UVD_JMI7_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JPEG_PREEMPT_FENCE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JRBC_RB_64BIT_BAR_LOW +#define UVD_JMI7_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JRBC_RB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH +#define UVD_JMI7_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI7_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI7_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JPEG_PREEMPT_VMID +#define UVD_JMI7_UVD_LMI_JPEG_PREEMPT_VMID__VMID__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JPEG_PREEMPT_VMID__VMID_MASK 0x0000000FL +//UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP__SHIFT 0x0 +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP__SHIFT 0x2 +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP__SHIFT 0x4 +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP__SHIFT 0x6 +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP__SHIFT 0x8 +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP__SHIFT 0xa +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP__SHIFT 0xc +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP__SHIFT 0xe +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP__SHIFT 0x10 +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__RB_MC_SWAP_MASK 0x00000003L +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__IB_MC_SWAP_MASK 0x0000000CL +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_WR_MC_SWAP_MASK 0x00000030L +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_WR_MC_SWAP_MASK 0x000000C0L +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__RB_MEM_RD_MC_SWAP_MASK 0x00000300L +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__IB_MEM_RD_MC_SWAP_MASK 0x00000C00L +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__PREEMPT_WR_MC_SWAP_MASK 0x00003000L +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__JPEG_RD_MC_SWAP_MASK 0x0000C000L +#define UVD_JMI7_UVD_JMI_DEC_SWAP_CNTL__JPEG_WR_MC_SWAP_MASK 0x00030000L +//UVD_JMI7_UVD_JMI_ATOMIC_CNTL +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en__SHIFT 0x0 +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL__atomic_max_burst__SHIFT 0x1 +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop__SHIFT 0x5 +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en__SHIFT 0x6 +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG__SHIFT 0x7 +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE__SHIFT 0xb +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL__atomic_arb_wait_en_MASK 0x00000001L +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL__atomic_max_burst_MASK 0x0000001EL +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL__atomic_wr_drop_MASK 0x00000020L +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL__atomic_wr_clamping_en_MASK 0x00000040L +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL__ATOMIC_WR_URG_MASK 0x00000780L +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL__ATOMIC_SW_GATE_MASK 0x00000800L +//UVD_JMI7_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW +#define UVD_JMI7_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI7_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH +#define UVD_JMI7_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI7_UVD_JMI_ATOMIC_USER0_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JPEG_READ_64BIT_BAR_LOW +#define UVD_JMI7_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JPEG_READ_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH +#define UVD_JMI7_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JPEG_READ_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW +#define UVD_JMI7_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JPEG_WRITE_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH +#define UVD_JMI7_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JPEG_WRITE_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JRBC_IB_64BIT_BAR_LOW +#define UVD_JMI7_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JRBC_IB_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH +#define UVD_JMI7_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JRBC_IB_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW +#define UVD_JMI7_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_LOW__BITS_31_0_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH +#define UVD_JMI7_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32__SHIFT 0x0 +#define UVD_JMI7_UVD_LMI_JRBC_IB_MEM_WR_64BIT_BAR_HIGH__BITS_63_32_MASK 0xFFFFFFFFL +//UVD_JMI7_UVD_JMI_ATOMIC_CNTL2 +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap__SHIFT 0x10 +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP__SHIFT 0x18 +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL2__atomic_uvd_swap_MASK 0x00FF0000L +#define UVD_JMI7_UVD_JMI_ATOMIC_CNTL2__ATOMIC_MC_SWAP_MASK 0xFF000000L + + +// addressBlock: uvdctxind +//UVD_CGC_MEM_CTRL +#define UVD_CGC_MEM_CTRL__LMI_MC_LS_EN__SHIFT 0x0 +#define UVD_CGC_MEM_CTRL__MPC_LS_EN__SHIFT 0x1 +#define UVD_CGC_MEM_CTRL__MPRD_LS_EN__SHIFT 0x2 +#define UVD_CGC_MEM_CTRL__WCB_LS_EN__SHIFT 0x3 +#define UVD_CGC_MEM_CTRL__UDEC_RE_LS_EN__SHIFT 0x4 +#define UVD_CGC_MEM_CTRL__UDEC_CM_LS_EN__SHIFT 0x5 +#define UVD_CGC_MEM_CTRL__UDEC_IT_LS_EN__SHIFT 0x6 +#define UVD_CGC_MEM_CTRL__UDEC_DB_LS_EN__SHIFT 0x7 +#define UVD_CGC_MEM_CTRL__UDEC_MP_LS_EN__SHIFT 0x8 +#define UVD_CGC_MEM_CTRL__SYS_LS_EN__SHIFT 0x9 +#define UVD_CGC_MEM_CTRL__VCPU_LS_EN__SHIFT 0xa +#define UVD_CGC_MEM_CTRL__MIF_LS_EN__SHIFT 0xc +#define UVD_CGC_MEM_CTRL__LCM_LS_EN__SHIFT 0xd +#define UVD_CGC_MEM_CTRL__MMSCH_LS_EN__SHIFT 0xe +#define UVD_CGC_MEM_CTRL__MPC1_LS_EN__SHIFT 0xf +#define UVD_CGC_MEM_CTRL__LS_SET_DELAY__SHIFT 0x10 +#define UVD_CGC_MEM_CTRL__LS_CLEAR_DELAY__SHIFT 0x14 +#define UVD_CGC_MEM_CTRL__LMI_MC_LS_EN_MASK 0x00000001L +#define UVD_CGC_MEM_CTRL__MPC_LS_EN_MASK 0x00000002L +#define UVD_CGC_MEM_CTRL__MPRD_LS_EN_MASK 0x00000004L +#define UVD_CGC_MEM_CTRL__WCB_LS_EN_MASK 0x00000008L +#define UVD_CGC_MEM_CTRL__UDEC_RE_LS_EN_MASK 0x00000010L +#define UVD_CGC_MEM_CTRL__UDEC_CM_LS_EN_MASK 0x00000020L +#define UVD_CGC_MEM_CTRL__UDEC_IT_LS_EN_MASK 0x00000040L +#define UVD_CGC_MEM_CTRL__UDEC_DB_LS_EN_MASK 0x00000080L +#define UVD_CGC_MEM_CTRL__UDEC_MP_LS_EN_MASK 0x00000100L +#define UVD_CGC_MEM_CTRL__SYS_LS_EN_MASK 0x00000200L +#define UVD_CGC_MEM_CTRL__VCPU_LS_EN_MASK 0x00000400L +#define UVD_CGC_MEM_CTRL__MIF_LS_EN_MASK 0x00001000L +#define UVD_CGC_MEM_CTRL__LCM_LS_EN_MASK 0x00002000L +#define UVD_CGC_MEM_CTRL__MMSCH_LS_EN_MASK 0x00004000L +#define UVD_CGC_MEM_CTRL__MPC1_LS_EN_MASK 0x00008000L +#define UVD_CGC_MEM_CTRL__LS_SET_DELAY_MASK 0x000F0000L +#define UVD_CGC_MEM_CTRL__LS_CLEAR_DELAY_MASK 0x00F00000L +//UVD_CGC_CTRL2 +#define UVD_CGC_CTRL2__DYN_OCLK_RAMP_EN__SHIFT 0x0 +#define UVD_CGC_CTRL2__DYN_RCLK_RAMP_EN__SHIFT 0x1 +#define UVD_CGC_CTRL2__GATER_DIV_ID__SHIFT 0x2 +#define UVD_CGC_CTRL2__DYN_OCLK_RAMP_EN_MASK 0x00000001L +#define UVD_CGC_CTRL2__DYN_RCLK_RAMP_EN_MASK 0x00000002L +#define UVD_CGC_CTRL2__GATER_DIV_ID_MASK 0x0000001CL +//UVD_CGC_MEM_DS_CTRL +#define UVD_CGC_MEM_DS_CTRL__LMI_MC_DS_EN__SHIFT 0x0 +#define UVD_CGC_MEM_DS_CTRL__MPC_DS_EN__SHIFT 0x1 +#define UVD_CGC_MEM_DS_CTRL__MPRD_DS_EN__SHIFT 0x2 +#define UVD_CGC_MEM_DS_CTRL__WCB_DS_EN__SHIFT 0x3 +#define UVD_CGC_MEM_DS_CTRL__UDEC_RE_DS_EN__SHIFT 0x4 +#define UVD_CGC_MEM_DS_CTRL__UDEC_CM_DS_EN__SHIFT 0x5 +#define UVD_CGC_MEM_DS_CTRL__UDEC_IT_DS_EN__SHIFT 0x6 +#define UVD_CGC_MEM_DS_CTRL__UDEC_DB_DS_EN__SHIFT 0x7 +#define UVD_CGC_MEM_DS_CTRL__UDEC_MP_DS_EN__SHIFT 0x8 +#define UVD_CGC_MEM_DS_CTRL__SYS_DS_EN__SHIFT 0x9 +#define UVD_CGC_MEM_DS_CTRL__VCPU_DS_EN__SHIFT 0xa +#define UVD_CGC_MEM_DS_CTRL__MIF_DS_EN__SHIFT 0xc +#define UVD_CGC_MEM_DS_CTRL__LCM_DS_EN__SHIFT 0xd +#define UVD_CGC_MEM_DS_CTRL__MMSCH_DS_EN__SHIFT 0xe +#define UVD_CGC_MEM_DS_CTRL__MPC1_DS_EN__SHIFT 0xf +#define UVD_CGC_MEM_DS_CTRL__LMI_MC_DS_EN_MASK 0x00000001L +#define UVD_CGC_MEM_DS_CTRL__MPC_DS_EN_MASK 0x00000002L +#define UVD_CGC_MEM_DS_CTRL__MPRD_DS_EN_MASK 0x00000004L +#define UVD_CGC_MEM_DS_CTRL__WCB_DS_EN_MASK 0x00000008L +#define UVD_CGC_MEM_DS_CTRL__UDEC_RE_DS_EN_MASK 0x00000010L +#define UVD_CGC_MEM_DS_CTRL__UDEC_CM_DS_EN_MASK 0x00000020L +#define UVD_CGC_MEM_DS_CTRL__UDEC_IT_DS_EN_MASK 0x00000040L +#define UVD_CGC_MEM_DS_CTRL__UDEC_DB_DS_EN_MASK 0x00000080L +#define UVD_CGC_MEM_DS_CTRL__UDEC_MP_DS_EN_MASK 0x00000100L +#define UVD_CGC_MEM_DS_CTRL__SYS_DS_EN_MASK 0x00000200L +#define UVD_CGC_MEM_DS_CTRL__VCPU_DS_EN_MASK 0x00000400L +#define UVD_CGC_MEM_DS_CTRL__MIF_DS_EN_MASK 0x00001000L +#define UVD_CGC_MEM_DS_CTRL__LCM_DS_EN_MASK 0x00002000L +#define UVD_CGC_MEM_DS_CTRL__MMSCH_DS_EN_MASK 0x00004000L +#define UVD_CGC_MEM_DS_CTRL__MPC1_DS_EN_MASK 0x00008000L +//UVD_CGC_MEM_SD_CTRL +#define UVD_CGC_MEM_SD_CTRL__LMI_MC_SD_EN__SHIFT 0x0 +#define UVD_CGC_MEM_SD_CTRL__MPC_SD_EN__SHIFT 0x1 +#define UVD_CGC_MEM_SD_CTRL__MPRD_SD_EN__SHIFT 0x2 +#define UVD_CGC_MEM_SD_CTRL__WCB_SD_EN__SHIFT 0x3 +#define UVD_CGC_MEM_SD_CTRL__UDEC_RE_SD_EN__SHIFT 0x4 +#define UVD_CGC_MEM_SD_CTRL__UDEC_CM_SD_EN__SHIFT 0x5 +#define UVD_CGC_MEM_SD_CTRL__UDEC_IT_SD_EN__SHIFT 0x6 +#define UVD_CGC_MEM_SD_CTRL__UDEC_DB_SD_EN__SHIFT 0x7 +#define UVD_CGC_MEM_SD_CTRL__UDEC_MP_SD_EN__SHIFT 0x8 +#define UVD_CGC_MEM_SD_CTRL__SYS_SD_EN__SHIFT 0x9 +#define UVD_CGC_MEM_SD_CTRL__VCPU_SD_EN__SHIFT 0xa +#define UVD_CGC_MEM_SD_CTRL__MIF_SD_EN__SHIFT 0xc +#define UVD_CGC_MEM_SD_CTRL__LCM_SD_EN__SHIFT 0xd +#define UVD_CGC_MEM_SD_CTRL__MMSCH_SD_EN__SHIFT 0xe +#define UVD_CGC_MEM_SD_CTRL__MPC1_SD_EN__SHIFT 0xf +#define UVD_CGC_MEM_SD_CTRL__LMI_MC_SD_EN_MASK 0x00000001L +#define UVD_CGC_MEM_SD_CTRL__MPC_SD_EN_MASK 0x00000002L +#define UVD_CGC_MEM_SD_CTRL__MPRD_SD_EN_MASK 0x00000004L +#define UVD_CGC_MEM_SD_CTRL__WCB_SD_EN_MASK 0x00000008L +#define UVD_CGC_MEM_SD_CTRL__UDEC_RE_SD_EN_MASK 0x00000010L +#define UVD_CGC_MEM_SD_CTRL__UDEC_CM_SD_EN_MASK 0x00000020L +#define UVD_CGC_MEM_SD_CTRL__UDEC_IT_SD_EN_MASK 0x00000040L +#define UVD_CGC_MEM_SD_CTRL__UDEC_DB_SD_EN_MASK 0x00000080L +#define UVD_CGC_MEM_SD_CTRL__UDEC_MP_SD_EN_MASK 0x00000100L +#define UVD_CGC_MEM_SD_CTRL__SYS_SD_EN_MASK 0x00000200L +#define UVD_CGC_MEM_SD_CTRL__VCPU_SD_EN_MASK 0x00000400L +#define UVD_CGC_MEM_SD_CTRL__MIF_SD_EN_MASK 0x00001000L +#define UVD_CGC_MEM_SD_CTRL__LCM_SD_EN_MASK 0x00002000L +#define UVD_CGC_MEM_SD_CTRL__MMSCH_SD_EN_MASK 0x00004000L +#define UVD_CGC_MEM_SD_CTRL__MPC1_SD_EN_MASK 0x00008000L +//UVD_SW_SCRATCH_00 +#define UVD_SW_SCRATCH_00__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_00__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_01 +#define UVD_SW_SCRATCH_01__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_01__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_02 +#define UVD_SW_SCRATCH_02__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_02__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_03 +#define UVD_SW_SCRATCH_03__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_03__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_04 +#define UVD_SW_SCRATCH_04__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_04__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_05 +#define UVD_SW_SCRATCH_05__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_05__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_06 +#define UVD_SW_SCRATCH_06__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_06__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_07 +#define UVD_SW_SCRATCH_07__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_07__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_08 +#define UVD_SW_SCRATCH_08__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_08__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_09 +#define UVD_SW_SCRATCH_09__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_09__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_10 +#define UVD_SW_SCRATCH_10__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_10__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_11 +#define UVD_SW_SCRATCH_11__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_11__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_12 +#define UVD_SW_SCRATCH_12__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_12__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_13 +#define UVD_SW_SCRATCH_13__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_13__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_14 +#define UVD_SW_SCRATCH_14__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_14__DATA_MASK 0xFFFFFFFFL +//UVD_SW_SCRATCH_15 +#define UVD_SW_SCRATCH_15__DATA__SHIFT 0x0 +#define UVD_SW_SCRATCH_15__DATA_MASK 0xFFFFFFFFL +//UVD_IH_SEM_CTRL +#define UVD_IH_SEM_CTRL__IH_STALL_EN__SHIFT 0x0 +#define UVD_IH_SEM_CTRL__SEM_STALL_EN__SHIFT 0x1 +#define UVD_IH_SEM_CTRL__IH_STATUS_CLEAN__SHIFT 0x2 +#define UVD_IH_SEM_CTRL__SEM_STATUS_CLEAN__SHIFT 0x3 +#define UVD_IH_SEM_CTRL__IH_VMID__SHIFT 0x4 +#define UVD_IH_SEM_CTRL__IH_USER_DATA__SHIFT 0x8 +#define UVD_IH_SEM_CTRL__IH_RINGID__SHIFT 0x14 +#define UVD_IH_SEM_CTRL__IH_STALL_EN_MASK 0x00000001L +#define UVD_IH_SEM_CTRL__SEM_STALL_EN_MASK 0x00000002L +#define UVD_IH_SEM_CTRL__IH_STATUS_CLEAN_MASK 0x00000004L +#define UVD_IH_SEM_CTRL__SEM_STATUS_CLEAN_MASK 0x00000008L +#define UVD_IH_SEM_CTRL__IH_VMID_MASK 0x000000F0L +#define UVD_IH_SEM_CTRL__IH_USER_DATA_MASK 0x000FFF00L +#define UVD_IH_SEM_CTRL__IH_RINGID_MASK 0x0FF00000L + + +// addressBlock: lmi_adp_indirect +//UVD_LMI_CRC0 +#define UVD_LMI_CRC0__CRC32__SHIFT 0x0 +#define UVD_LMI_CRC0__CRC32_MASK 0xFFFFFFFFL +//UVD_LMI_CRC1 +#define UVD_LMI_CRC1__CRC32__SHIFT 0x0 +#define UVD_LMI_CRC1__CRC32_MASK 0xFFFFFFFFL +//UVD_LMI_CRC2 +#define UVD_LMI_CRC2__CRC32__SHIFT 0x0 +#define UVD_LMI_CRC2__CRC32_MASK 0xFFFFFFFFL +//UVD_LMI_CRC3 +#define UVD_LMI_CRC3__CRC32__SHIFT 0x0 +#define UVD_LMI_CRC3__CRC32_MASK 0xFFFFFFFFL +//UVD_LMI_CRC10 +#define UVD_LMI_CRC10__CRC32__SHIFT 0x0 +#define UVD_LMI_CRC10__CRC32_MASK 0xFFFFFFFFL +//UVD_LMI_CRC11 +#define UVD_LMI_CRC11__CRC32__SHIFT 0x0 +#define UVD_LMI_CRC11__CRC32_MASK 0xFFFFFFFFL +//UVD_LMI_CRC12 +#define UVD_LMI_CRC12__CRC32__SHIFT 0x0 +#define UVD_LMI_CRC12__CRC32_MASK 0xFFFFFFFFL +//UVD_LMI_CRC13 +#define UVD_LMI_CRC13__CRC32__SHIFT 0x0 +#define UVD_LMI_CRC13__CRC32_MASK 0xFFFFFFFFL +//UVD_LMI_CRC14 +#define UVD_LMI_CRC14__CRC32__SHIFT 0x0 +#define UVD_LMI_CRC14__CRC32_MASK 0xFFFFFFFFL +//UVD_LMI_CRC15 +#define UVD_LMI_CRC15__CRC32__SHIFT 0x0 +#define UVD_LMI_CRC15__CRC32_MASK 0xFFFFFFFFL +//UVD_LMI_SWAP_CNTL2 +#define UVD_LMI_SWAP_CNTL2__SCPU_R_MC_SWAP__SHIFT 0x0 +#define UVD_LMI_SWAP_CNTL2__SCPU_W_MC_SWAP__SHIFT 0x2 +#define UVD_LMI_SWAP_CNTL2__ATOMIC_MC_SWAP__SHIFT 0x4 +#define UVD_LMI_SWAP_CNTL2__CENC_MC_SWAP__SHIFT 0xc +#define UVD_LMI_SWAP_CNTL2__FBC_KEY_MC_SWAP__SHIFT 0xe +#define UVD_LMI_SWAP_CNTL2__SCPU_R_MC_SWAP_MASK 0x00000003L +#define UVD_LMI_SWAP_CNTL2__SCPU_W_MC_SWAP_MASK 0x0000000CL +#define UVD_LMI_SWAP_CNTL2__ATOMIC_MC_SWAP_MASK 0x00000FF0L +#define UVD_LMI_SWAP_CNTL2__CENC_MC_SWAP_MASK 0x00003000L +#define UVD_LMI_SWAP_CNTL2__FBC_KEY_MC_SWAP_MASK 0x0000C000L +//UVD_MEMCHECK_SYS_INT_EN +#define UVD_MEMCHECK_SYS_INT_EN__RE_ERR_EN__SHIFT 0x0 +#define UVD_MEMCHECK_SYS_INT_EN__IT_ERR_EN__SHIFT 0x1 +#define UVD_MEMCHECK_SYS_INT_EN__MP_ERR_EN__SHIFT 0x2 +#define UVD_MEMCHECK_SYS_INT_EN__DB_ERR_EN__SHIFT 0x3 +#define UVD_MEMCHECK_SYS_INT_EN__DBW_ERR_EN__SHIFT 0x4 +#define UVD_MEMCHECK_SYS_INT_EN__CM_ERR_EN__SHIFT 0x5 +#define UVD_MEMCHECK_SYS_INT_EN__MIF_REF_ERR_EN__SHIFT 0x6 +#define UVD_MEMCHECK_SYS_INT_EN__VCPU_ERR_EN__SHIFT 0x7 +#define UVD_MEMCHECK_SYS_INT_EN__MIF_DBW_ERR_EN__SHIFT 0x8 +#define UVD_MEMCHECK_SYS_INT_EN__MIF_CM_COLOC_ERR_EN__SHIFT 0x9 +#define UVD_MEMCHECK_SYS_INT_EN__MIF_BSP0_ERR_EN__SHIFT 0xa +#define UVD_MEMCHECK_SYS_INT_EN__MIF_BSP1_ERR_EN__SHIFT 0xb +#define UVD_MEMCHECK_SYS_INT_EN__SRE_ERR_EN__SHIFT 0xc +#define UVD_MEMCHECK_SYS_INT_EN__IT_RD_ERR_EN__SHIFT 0xf +#define UVD_MEMCHECK_SYS_INT_EN__CM_RD_ERR_EN__SHIFT 0x10 +#define UVD_MEMCHECK_SYS_INT_EN__DB_RD_ERR_EN__SHIFT 0x11 +#define UVD_MEMCHECK_SYS_INT_EN__MIF_RD_ERR_EN__SHIFT 0x12 +#define UVD_MEMCHECK_SYS_INT_EN__IDCT_RD_ERR_EN__SHIFT 0x13 +#define UVD_MEMCHECK_SYS_INT_EN__MPC_RD_ERR_EN__SHIFT 0x14 +#define UVD_MEMCHECK_SYS_INT_EN__LBSI_RD_ERR_EN__SHIFT 0x15 +#define UVD_MEMCHECK_SYS_INT_EN__RBC_RD_ERR_EN__SHIFT 0x18 +#define UVD_MEMCHECK_SYS_INT_EN__MIF_BSP2_ERR_EN__SHIFT 0x1b +#define UVD_MEMCHECK_SYS_INT_EN__MIF_BSP3_ERR_EN__SHIFT 0x1c +#define UVD_MEMCHECK_SYS_INT_EN__MIF_SCLR_ERR_EN__SHIFT 0x1d +#define UVD_MEMCHECK_SYS_INT_EN__MIF_SCLR2_ERR_EN__SHIFT 0x1e +#define UVD_MEMCHECK_SYS_INT_EN__PREF_ERR_EN__SHIFT 0x1f +#define UVD_MEMCHECK_SYS_INT_EN__RE_ERR_EN_MASK 0x00000001L +#define UVD_MEMCHECK_SYS_INT_EN__IT_ERR_EN_MASK 0x00000002L +#define UVD_MEMCHECK_SYS_INT_EN__MP_ERR_EN_MASK 0x00000004L +#define UVD_MEMCHECK_SYS_INT_EN__DB_ERR_EN_MASK 0x00000008L +#define UVD_MEMCHECK_SYS_INT_EN__DBW_ERR_EN_MASK 0x00000010L +#define UVD_MEMCHECK_SYS_INT_EN__CM_ERR_EN_MASK 0x00000020L +#define UVD_MEMCHECK_SYS_INT_EN__MIF_REF_ERR_EN_MASK 0x00000040L +#define UVD_MEMCHECK_SYS_INT_EN__VCPU_ERR_EN_MASK 0x00000080L +#define UVD_MEMCHECK_SYS_INT_EN__MIF_DBW_ERR_EN_MASK 0x00000100L +#define UVD_MEMCHECK_SYS_INT_EN__MIF_CM_COLOC_ERR_EN_MASK 0x00000200L +#define UVD_MEMCHECK_SYS_INT_EN__MIF_BSP0_ERR_EN_MASK 0x00000400L +#define UVD_MEMCHECK_SYS_INT_EN__MIF_BSP1_ERR_EN_MASK 0x00000800L +#define UVD_MEMCHECK_SYS_INT_EN__SRE_ERR_EN_MASK 0x00001000L +#define UVD_MEMCHECK_SYS_INT_EN__IT_RD_ERR_EN_MASK 0x00008000L +#define UVD_MEMCHECK_SYS_INT_EN__CM_RD_ERR_EN_MASK 0x00010000L +#define UVD_MEMCHECK_SYS_INT_EN__DB_RD_ERR_EN_MASK 0x00020000L +#define UVD_MEMCHECK_SYS_INT_EN__MIF_RD_ERR_EN_MASK 0x00040000L +#define UVD_MEMCHECK_SYS_INT_EN__IDCT_RD_ERR_EN_MASK 0x00080000L +#define UVD_MEMCHECK_SYS_INT_EN__MPC_RD_ERR_EN_MASK 0x00100000L +#define UVD_MEMCHECK_SYS_INT_EN__LBSI_RD_ERR_EN_MASK 0x00200000L +#define UVD_MEMCHECK_SYS_INT_EN__RBC_RD_ERR_EN_MASK 0x01000000L +#define UVD_MEMCHECK_SYS_INT_EN__MIF_BSP2_ERR_EN_MASK 0x08000000L +#define UVD_MEMCHECK_SYS_INT_EN__MIF_BSP3_ERR_EN_MASK 0x10000000L +#define UVD_MEMCHECK_SYS_INT_EN__MIF_SCLR_ERR_EN_MASK 0x20000000L +#define UVD_MEMCHECK_SYS_INT_EN__MIF_SCLR2_ERR_EN_MASK 0x40000000L +#define UVD_MEMCHECK_SYS_INT_EN__PREF_ERR_EN_MASK 0x80000000L +//UVD_MEMCHECK_SYS_INT_STAT +#define UVD_MEMCHECK_SYS_INT_STAT__RE_LO_ERR__SHIFT 0x0 +#define UVD_MEMCHECK_SYS_INT_STAT__RE_HI_ERR__SHIFT 0x1 +#define UVD_MEMCHECK_SYS_INT_STAT__IT_LO_ERR__SHIFT 0x2 +#define UVD_MEMCHECK_SYS_INT_STAT__IT_HI_ERR__SHIFT 0x3 +#define UVD_MEMCHECK_SYS_INT_STAT__MP_LO_ERR__SHIFT 0x4 +#define UVD_MEMCHECK_SYS_INT_STAT__MP_HI_ERR__SHIFT 0x5 +#define UVD_MEMCHECK_SYS_INT_STAT__DB_LO_ERR__SHIFT 0x6 +#define UVD_MEMCHECK_SYS_INT_STAT__DB_HI_ERR__SHIFT 0x7 +#define UVD_MEMCHECK_SYS_INT_STAT__DBW_LO_ERR__SHIFT 0x8 +#define UVD_MEMCHECK_SYS_INT_STAT__DBW_HI_ERR__SHIFT 0x9 +#define UVD_MEMCHECK_SYS_INT_STAT__CM_LO_ERR__SHIFT 0xa +#define UVD_MEMCHECK_SYS_INT_STAT__CM_HI_ERR__SHIFT 0xb +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_REF_LO_ERR__SHIFT 0xc +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_REF_HI_ERR__SHIFT 0xd +#define UVD_MEMCHECK_SYS_INT_STAT__VCPU_LO_ERR__SHIFT 0xe +#define UVD_MEMCHECK_SYS_INT_STAT__VCPU_HI_ERR__SHIFT 0xf +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_DBW_LO_ERR__SHIFT 0x10 +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_DBW_HI_ERR__SHIFT 0x11 +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_CM_COLOC_LO_ERR__SHIFT 0x12 +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_CM_COLOC_HI_ERR__SHIFT 0x13 +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_BSP0_LO_ERR__SHIFT 0x14 +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_BSP0_HI_ERR__SHIFT 0x15 +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_BSP1_LO_ERR__SHIFT 0x16 +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_BSP1_HI_ERR__SHIFT 0x17 +#define UVD_MEMCHECK_SYS_INT_STAT__SRE_LO_ERR__SHIFT 0x18 +#define UVD_MEMCHECK_SYS_INT_STAT__SRE_HI_ERR__SHIFT 0x19 +#define UVD_MEMCHECK_SYS_INT_STAT__IT_RD_LO_ERR__SHIFT 0x1e +#define UVD_MEMCHECK_SYS_INT_STAT__IT_RD_HI_ERR__SHIFT 0x1f +#define UVD_MEMCHECK_SYS_INT_STAT__RE_LO_ERR_MASK 0x00000001L +#define UVD_MEMCHECK_SYS_INT_STAT__RE_HI_ERR_MASK 0x00000002L +#define UVD_MEMCHECK_SYS_INT_STAT__IT_LO_ERR_MASK 0x00000004L +#define UVD_MEMCHECK_SYS_INT_STAT__IT_HI_ERR_MASK 0x00000008L +#define UVD_MEMCHECK_SYS_INT_STAT__MP_LO_ERR_MASK 0x00000010L +#define UVD_MEMCHECK_SYS_INT_STAT__MP_HI_ERR_MASK 0x00000020L +#define UVD_MEMCHECK_SYS_INT_STAT__DB_LO_ERR_MASK 0x00000040L +#define UVD_MEMCHECK_SYS_INT_STAT__DB_HI_ERR_MASK 0x00000080L +#define UVD_MEMCHECK_SYS_INT_STAT__DBW_LO_ERR_MASK 0x00000100L +#define UVD_MEMCHECK_SYS_INT_STAT__DBW_HI_ERR_MASK 0x00000200L +#define UVD_MEMCHECK_SYS_INT_STAT__CM_LO_ERR_MASK 0x00000400L +#define UVD_MEMCHECK_SYS_INT_STAT__CM_HI_ERR_MASK 0x00000800L +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_REF_LO_ERR_MASK 0x00001000L +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_REF_HI_ERR_MASK 0x00002000L +#define UVD_MEMCHECK_SYS_INT_STAT__VCPU_LO_ERR_MASK 0x00004000L +#define UVD_MEMCHECK_SYS_INT_STAT__VCPU_HI_ERR_MASK 0x00008000L +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_DBW_LO_ERR_MASK 0x00010000L +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_DBW_HI_ERR_MASK 0x00020000L +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_CM_COLOC_LO_ERR_MASK 0x00040000L +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_CM_COLOC_HI_ERR_MASK 0x00080000L +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_BSP0_LO_ERR_MASK 0x00100000L +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_BSP0_HI_ERR_MASK 0x00200000L +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_BSP1_LO_ERR_MASK 0x00400000L +#define UVD_MEMCHECK_SYS_INT_STAT__MIF_BSP1_HI_ERR_MASK 0x00800000L +#define UVD_MEMCHECK_SYS_INT_STAT__SRE_LO_ERR_MASK 0x01000000L +#define UVD_MEMCHECK_SYS_INT_STAT__SRE_HI_ERR_MASK 0x02000000L +#define UVD_MEMCHECK_SYS_INT_STAT__IT_RD_LO_ERR_MASK 0x40000000L +#define UVD_MEMCHECK_SYS_INT_STAT__IT_RD_HI_ERR_MASK 0x80000000L +//UVD_MEMCHECK_SYS_INT_ACK +#define UVD_MEMCHECK_SYS_INT_ACK__RE_LO_ACK__SHIFT 0x0 +#define UVD_MEMCHECK_SYS_INT_ACK__RE_HI_ACK__SHIFT 0x1 +#define UVD_MEMCHECK_SYS_INT_ACK__IT_LO_ACK__SHIFT 0x2 +#define UVD_MEMCHECK_SYS_INT_ACK__IT_HI_ACK__SHIFT 0x3 +#define UVD_MEMCHECK_SYS_INT_ACK__MP_LO_ACK__SHIFT 0x4 +#define UVD_MEMCHECK_SYS_INT_ACK__MP_HI_ACK__SHIFT 0x5 +#define UVD_MEMCHECK_SYS_INT_ACK__DB_LO_ACK__SHIFT 0x6 +#define UVD_MEMCHECK_SYS_INT_ACK__DB_HI_ACK__SHIFT 0x7 +#define UVD_MEMCHECK_SYS_INT_ACK__DBW_LO_ACK__SHIFT 0x8 +#define UVD_MEMCHECK_SYS_INT_ACK__DBW_HI_ACK__SHIFT 0x9 +#define UVD_MEMCHECK_SYS_INT_ACK__CM_LO_ACK__SHIFT 0xa +#define UVD_MEMCHECK_SYS_INT_ACK__CM_HI_ACK__SHIFT 0xb +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_REF_LO_ACK__SHIFT 0xc +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_REF_HI_ACK__SHIFT 0xd +#define UVD_MEMCHECK_SYS_INT_ACK__VCPU_LO_ACK__SHIFT 0xe +#define UVD_MEMCHECK_SYS_INT_ACK__VCPU_HI_ACK__SHIFT 0xf +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_DBW_LO_ACK__SHIFT 0x10 +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_DBW_HI_ACK__SHIFT 0x11 +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_CM_COLOC_LO_ACK__SHIFT 0x12 +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_CM_COLOC_HI_ACK__SHIFT 0x13 +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_BSP0_LO_ACK__SHIFT 0x14 +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_BSP0_HI_ACK__SHIFT 0x15 +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_BSP1_LO_ACK__SHIFT 0x16 +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_BSP1_HI_ACK__SHIFT 0x17 +#define UVD_MEMCHECK_SYS_INT_ACK__SRE_LO_ACK__SHIFT 0x18 +#define UVD_MEMCHECK_SYS_INT_ACK__SRE_HI_ACK__SHIFT 0x19 +#define UVD_MEMCHECK_SYS_INT_ACK__IT_RD_LO_ACK__SHIFT 0x1e +#define UVD_MEMCHECK_SYS_INT_ACK__IT_RD_HI_ACK__SHIFT 0x1f +#define UVD_MEMCHECK_SYS_INT_ACK__RE_LO_ACK_MASK 0x00000001L +#define UVD_MEMCHECK_SYS_INT_ACK__RE_HI_ACK_MASK 0x00000002L +#define UVD_MEMCHECK_SYS_INT_ACK__IT_LO_ACK_MASK 0x00000004L +#define UVD_MEMCHECK_SYS_INT_ACK__IT_HI_ACK_MASK 0x00000008L +#define UVD_MEMCHECK_SYS_INT_ACK__MP_LO_ACK_MASK 0x00000010L +#define UVD_MEMCHECK_SYS_INT_ACK__MP_HI_ACK_MASK 0x00000020L +#define UVD_MEMCHECK_SYS_INT_ACK__DB_LO_ACK_MASK 0x00000040L +#define UVD_MEMCHECK_SYS_INT_ACK__DB_HI_ACK_MASK 0x00000080L +#define UVD_MEMCHECK_SYS_INT_ACK__DBW_LO_ACK_MASK 0x00000100L +#define UVD_MEMCHECK_SYS_INT_ACK__DBW_HI_ACK_MASK 0x00000200L +#define UVD_MEMCHECK_SYS_INT_ACK__CM_LO_ACK_MASK 0x00000400L +#define UVD_MEMCHECK_SYS_INT_ACK__CM_HI_ACK_MASK 0x00000800L +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_REF_LO_ACK_MASK 0x00001000L +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_REF_HI_ACK_MASK 0x00002000L +#define UVD_MEMCHECK_SYS_INT_ACK__VCPU_LO_ACK_MASK 0x00004000L +#define UVD_MEMCHECK_SYS_INT_ACK__VCPU_HI_ACK_MASK 0x00008000L +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_DBW_LO_ACK_MASK 0x00010000L +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_DBW_HI_ACK_MASK 0x00020000L +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_CM_COLOC_LO_ACK_MASK 0x00040000L +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_CM_COLOC_HI_ACK_MASK 0x00080000L +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_BSP0_LO_ACK_MASK 0x00100000L +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_BSP0_HI_ACK_MASK 0x00200000L +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_BSP1_LO_ACK_MASK 0x00400000L +#define UVD_MEMCHECK_SYS_INT_ACK__MIF_BSP1_HI_ACK_MASK 0x00800000L +#define UVD_MEMCHECK_SYS_INT_ACK__SRE_LO_ACK_MASK 0x01000000L +#define UVD_MEMCHECK_SYS_INT_ACK__SRE_HI_ACK_MASK 0x02000000L +#define UVD_MEMCHECK_SYS_INT_ACK__IT_RD_LO_ACK_MASK 0x40000000L +#define UVD_MEMCHECK_SYS_INT_ACK__IT_RD_HI_ACK_MASK 0x80000000L +//UVD_MEMCHECK_VCPU_INT_EN +#define UVD_MEMCHECK_VCPU_INT_EN__RE_ERR_EN__SHIFT 0x0 +#define UVD_MEMCHECK_VCPU_INT_EN__IT_ERR_EN__SHIFT 0x1 +#define UVD_MEMCHECK_VCPU_INT_EN__MP_ERR_EN__SHIFT 0x2 +#define UVD_MEMCHECK_VCPU_INT_EN__DB_ERR_EN__SHIFT 0x3 +#define UVD_MEMCHECK_VCPU_INT_EN__DBW_ERR_EN__SHIFT 0x4 +#define UVD_MEMCHECK_VCPU_INT_EN__CM_ERR_EN__SHIFT 0x5 +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_REF_ERR_EN__SHIFT 0x6 +#define UVD_MEMCHECK_VCPU_INT_EN__VCPU_ERR_EN__SHIFT 0x7 +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_DBW_ERR_EN__SHIFT 0x8 +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_CM_COLOC_ERR_EN__SHIFT 0x9 +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_BSP0_ERR_EN__SHIFT 0xa +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_BSP1_ERR_EN__SHIFT 0xb +#define UVD_MEMCHECK_VCPU_INT_EN__SRE_ERR_EN__SHIFT 0xc +#define UVD_MEMCHECK_VCPU_INT_EN__IT_RD_ERR_EN__SHIFT 0xf +#define UVD_MEMCHECK_VCPU_INT_EN__CM_RD_ERR_EN__SHIFT 0x10 +#define UVD_MEMCHECK_VCPU_INT_EN__DB_RD_ERR_EN__SHIFT 0x11 +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_RD_ERR_EN__SHIFT 0x12 +#define UVD_MEMCHECK_VCPU_INT_EN__IDCT_RD_ERR_EN__SHIFT 0x13 +#define UVD_MEMCHECK_VCPU_INT_EN__MPC_RD_ERR_EN__SHIFT 0x14 +#define UVD_MEMCHECK_VCPU_INT_EN__LBSI_RD_ERR_EN__SHIFT 0x15 +#define UVD_MEMCHECK_VCPU_INT_EN__RBC_RD_ERR_EN__SHIFT 0x18 +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_BSP2_ERR_EN__SHIFT 0x19 +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_BSP3_ERR_EN__SHIFT 0x1a +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_SCLR_ERR_EN__SHIFT 0x1b +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_SCLR2_ERR_EN__SHIFT 0x1c +#define UVD_MEMCHECK_VCPU_INT_EN__PREF_ERR_EN__SHIFT 0x1d +#define UVD_MEMCHECK_VCPU_INT_EN__RE_ERR_EN_MASK 0x00000001L +#define UVD_MEMCHECK_VCPU_INT_EN__IT_ERR_EN_MASK 0x00000002L +#define UVD_MEMCHECK_VCPU_INT_EN__MP_ERR_EN_MASK 0x00000004L +#define UVD_MEMCHECK_VCPU_INT_EN__DB_ERR_EN_MASK 0x00000008L +#define UVD_MEMCHECK_VCPU_INT_EN__DBW_ERR_EN_MASK 0x00000010L +#define UVD_MEMCHECK_VCPU_INT_EN__CM_ERR_EN_MASK 0x00000020L +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_REF_ERR_EN_MASK 0x00000040L +#define UVD_MEMCHECK_VCPU_INT_EN__VCPU_ERR_EN_MASK 0x00000080L +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_DBW_ERR_EN_MASK 0x00000100L +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_CM_COLOC_ERR_EN_MASK 0x00000200L +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_BSP0_ERR_EN_MASK 0x00000400L +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_BSP1_ERR_EN_MASK 0x00000800L +#define UVD_MEMCHECK_VCPU_INT_EN__SRE_ERR_EN_MASK 0x00001000L +#define UVD_MEMCHECK_VCPU_INT_EN__IT_RD_ERR_EN_MASK 0x00008000L +#define UVD_MEMCHECK_VCPU_INT_EN__CM_RD_ERR_EN_MASK 0x00010000L +#define UVD_MEMCHECK_VCPU_INT_EN__DB_RD_ERR_EN_MASK 0x00020000L +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_RD_ERR_EN_MASK 0x00040000L +#define UVD_MEMCHECK_VCPU_INT_EN__IDCT_RD_ERR_EN_MASK 0x00080000L +#define UVD_MEMCHECK_VCPU_INT_EN__MPC_RD_ERR_EN_MASK 0x00100000L +#define UVD_MEMCHECK_VCPU_INT_EN__LBSI_RD_ERR_EN_MASK 0x00200000L +#define UVD_MEMCHECK_VCPU_INT_EN__RBC_RD_ERR_EN_MASK 0x01000000L +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_BSP2_ERR_EN_MASK 0x02000000L +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_BSP3_ERR_EN_MASK 0x04000000L +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_SCLR_ERR_EN_MASK 0x08000000L +#define UVD_MEMCHECK_VCPU_INT_EN__MIF_SCLR2_ERR_EN_MASK 0x10000000L +#define UVD_MEMCHECK_VCPU_INT_EN__PREF_ERR_EN_MASK 0x20000000L +//UVD_MEMCHECK_VCPU_INT_STAT +#define UVD_MEMCHECK_VCPU_INT_STAT__RE_LO_ERR__SHIFT 0x0 +#define UVD_MEMCHECK_VCPU_INT_STAT__RE_HI_ERR__SHIFT 0x1 +#define UVD_MEMCHECK_VCPU_INT_STAT__IT_LO_ERR__SHIFT 0x2 +#define UVD_MEMCHECK_VCPU_INT_STAT__IT_HI_ERR__SHIFT 0x3 +#define UVD_MEMCHECK_VCPU_INT_STAT__MP_LO_ERR__SHIFT 0x4 +#define UVD_MEMCHECK_VCPU_INT_STAT__MP_HI_ERR__SHIFT 0x5 +#define UVD_MEMCHECK_VCPU_INT_STAT__DB_LO_ERR__SHIFT 0x6 +#define UVD_MEMCHECK_VCPU_INT_STAT__DB_HI_ERR__SHIFT 0x7 +#define UVD_MEMCHECK_VCPU_INT_STAT__DBW_LO_ERR__SHIFT 0x8 +#define UVD_MEMCHECK_VCPU_INT_STAT__DBW_HI_ERR__SHIFT 0x9 +#define UVD_MEMCHECK_VCPU_INT_STAT__CM_LO_ERR__SHIFT 0xa +#define UVD_MEMCHECK_VCPU_INT_STAT__CM_HI_ERR__SHIFT 0xb +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_REF_LO_ERR__SHIFT 0xc +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_REF_HI_ERR__SHIFT 0xd +#define UVD_MEMCHECK_VCPU_INT_STAT__VCPU_LO_ERR__SHIFT 0xe +#define UVD_MEMCHECK_VCPU_INT_STAT__VCPU_HI_ERR__SHIFT 0xf +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_DBW_LO_ERR__SHIFT 0x10 +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_DBW_HI_ERR__SHIFT 0x11 +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_CM_COLOC_LO_ERR__SHIFT 0x12 +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_CM_COLOC_HI_ERR__SHIFT 0x13 +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_BSP0_LO_ERR__SHIFT 0x14 +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_BSP0_HI_ERR__SHIFT 0x15 +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_BSP1_LO_ERR__SHIFT 0x16 +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_BSP1_HI_ERR__SHIFT 0x17 +#define UVD_MEMCHECK_VCPU_INT_STAT__SRE_LO_ERR__SHIFT 0x18 +#define UVD_MEMCHECK_VCPU_INT_STAT__SRE_HI_ERR__SHIFT 0x19 +#define UVD_MEMCHECK_VCPU_INT_STAT__IT_RD_LO_ERR__SHIFT 0x1e +#define UVD_MEMCHECK_VCPU_INT_STAT__IT_RD_HI_ERR__SHIFT 0x1f +#define UVD_MEMCHECK_VCPU_INT_STAT__RE_LO_ERR_MASK 0x00000001L +#define UVD_MEMCHECK_VCPU_INT_STAT__RE_HI_ERR_MASK 0x00000002L +#define UVD_MEMCHECK_VCPU_INT_STAT__IT_LO_ERR_MASK 0x00000004L +#define UVD_MEMCHECK_VCPU_INT_STAT__IT_HI_ERR_MASK 0x00000008L +#define UVD_MEMCHECK_VCPU_INT_STAT__MP_LO_ERR_MASK 0x00000010L +#define UVD_MEMCHECK_VCPU_INT_STAT__MP_HI_ERR_MASK 0x00000020L +#define UVD_MEMCHECK_VCPU_INT_STAT__DB_LO_ERR_MASK 0x00000040L +#define UVD_MEMCHECK_VCPU_INT_STAT__DB_HI_ERR_MASK 0x00000080L +#define UVD_MEMCHECK_VCPU_INT_STAT__DBW_LO_ERR_MASK 0x00000100L +#define UVD_MEMCHECK_VCPU_INT_STAT__DBW_HI_ERR_MASK 0x00000200L +#define UVD_MEMCHECK_VCPU_INT_STAT__CM_LO_ERR_MASK 0x00000400L +#define UVD_MEMCHECK_VCPU_INT_STAT__CM_HI_ERR_MASK 0x00000800L +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_REF_LO_ERR_MASK 0x00001000L +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_REF_HI_ERR_MASK 0x00002000L +#define UVD_MEMCHECK_VCPU_INT_STAT__VCPU_LO_ERR_MASK 0x00004000L +#define UVD_MEMCHECK_VCPU_INT_STAT__VCPU_HI_ERR_MASK 0x00008000L +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_DBW_LO_ERR_MASK 0x00010000L +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_DBW_HI_ERR_MASK 0x00020000L +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_CM_COLOC_LO_ERR_MASK 0x00040000L +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_CM_COLOC_HI_ERR_MASK 0x00080000L +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_BSP0_LO_ERR_MASK 0x00100000L +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_BSP0_HI_ERR_MASK 0x00200000L +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_BSP1_LO_ERR_MASK 0x00400000L +#define UVD_MEMCHECK_VCPU_INT_STAT__MIF_BSP1_HI_ERR_MASK 0x00800000L +#define UVD_MEMCHECK_VCPU_INT_STAT__SRE_LO_ERR_MASK 0x01000000L +#define UVD_MEMCHECK_VCPU_INT_STAT__SRE_HI_ERR_MASK 0x02000000L +#define UVD_MEMCHECK_VCPU_INT_STAT__IT_RD_LO_ERR_MASK 0x40000000L +#define UVD_MEMCHECK_VCPU_INT_STAT__IT_RD_HI_ERR_MASK 0x80000000L +//UVD_MEMCHECK_VCPU_INT_ACK +#define UVD_MEMCHECK_VCPU_INT_ACK__RE_LO_ACK__SHIFT 0x0 +#define UVD_MEMCHECK_VCPU_INT_ACK__RE_HI_ACK__SHIFT 0x1 +#define UVD_MEMCHECK_VCPU_INT_ACK__IT_LO_ACK__SHIFT 0x2 +#define UVD_MEMCHECK_VCPU_INT_ACK__IT_HI_ACK__SHIFT 0x3 +#define UVD_MEMCHECK_VCPU_INT_ACK__MP_LO_ACK__SHIFT 0x4 +#define UVD_MEMCHECK_VCPU_INT_ACK__MP_HI_ACK__SHIFT 0x5 +#define UVD_MEMCHECK_VCPU_INT_ACK__DB_LO_ACK__SHIFT 0x6 +#define UVD_MEMCHECK_VCPU_INT_ACK__DB_HI_ACK__SHIFT 0x7 +#define UVD_MEMCHECK_VCPU_INT_ACK__DBW_LO_ACK__SHIFT 0x8 +#define UVD_MEMCHECK_VCPU_INT_ACK__DBW_HI_ACK__SHIFT 0x9 +#define UVD_MEMCHECK_VCPU_INT_ACK__CM_LO_ACK__SHIFT 0xa +#define UVD_MEMCHECK_VCPU_INT_ACK__CM_HI_ACK__SHIFT 0xb +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_REF_LO_ACK__SHIFT 0xc +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_REF_HI_ACK__SHIFT 0xd +#define UVD_MEMCHECK_VCPU_INT_ACK__VCPU_LO_ACK__SHIFT 0xe +#define UVD_MEMCHECK_VCPU_INT_ACK__VCPU_HI_ACK__SHIFT 0xf +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_DBW_LO_ACK__SHIFT 0x10 +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_DBW_HI_ACK__SHIFT 0x11 +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_CM_COLOC_LO_ACK__SHIFT 0x12 +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_CM_COLOC_HI_ACK__SHIFT 0x13 +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_BSP0_LO_ACK__SHIFT 0x14 +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_BSP0_HI_ACK__SHIFT 0x15 +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_BSP1_LO_ACK__SHIFT 0x16 +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_BSP1_HI_ACK__SHIFT 0x17 +#define UVD_MEMCHECK_VCPU_INT_ACK__SRE_LO_ACK__SHIFT 0x18 +#define UVD_MEMCHECK_VCPU_INT_ACK__SRE_HI_ACK__SHIFT 0x19 +#define UVD_MEMCHECK_VCPU_INT_ACK__IT_RD_LO_ACK__SHIFT 0x1e +#define UVD_MEMCHECK_VCPU_INT_ACK__IT_RD_HI_ACK__SHIFT 0x1f +#define UVD_MEMCHECK_VCPU_INT_ACK__RE_LO_ACK_MASK 0x00000001L +#define UVD_MEMCHECK_VCPU_INT_ACK__RE_HI_ACK_MASK 0x00000002L +#define UVD_MEMCHECK_VCPU_INT_ACK__IT_LO_ACK_MASK 0x00000004L +#define UVD_MEMCHECK_VCPU_INT_ACK__IT_HI_ACK_MASK 0x00000008L +#define UVD_MEMCHECK_VCPU_INT_ACK__MP_LO_ACK_MASK 0x00000010L +#define UVD_MEMCHECK_VCPU_INT_ACK__MP_HI_ACK_MASK 0x00000020L +#define UVD_MEMCHECK_VCPU_INT_ACK__DB_LO_ACK_MASK 0x00000040L +#define UVD_MEMCHECK_VCPU_INT_ACK__DB_HI_ACK_MASK 0x00000080L +#define UVD_MEMCHECK_VCPU_INT_ACK__DBW_LO_ACK_MASK 0x00000100L +#define UVD_MEMCHECK_VCPU_INT_ACK__DBW_HI_ACK_MASK 0x00000200L +#define UVD_MEMCHECK_VCPU_INT_ACK__CM_LO_ACK_MASK 0x00000400L +#define UVD_MEMCHECK_VCPU_INT_ACK__CM_HI_ACK_MASK 0x00000800L +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_REF_LO_ACK_MASK 0x00001000L +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_REF_HI_ACK_MASK 0x00002000L +#define UVD_MEMCHECK_VCPU_INT_ACK__VCPU_LO_ACK_MASK 0x00004000L +#define UVD_MEMCHECK_VCPU_INT_ACK__VCPU_HI_ACK_MASK 0x00008000L +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_DBW_LO_ACK_MASK 0x00010000L +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_DBW_HI_ACK_MASK 0x00020000L +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_CM_COLOC_LO_ACK_MASK 0x00040000L +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_CM_COLOC_HI_ACK_MASK 0x00080000L +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_BSP0_LO_ACK_MASK 0x00100000L +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_BSP0_HI_ACK_MASK 0x00200000L +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_BSP1_LO_ACK_MASK 0x00400000L +#define UVD_MEMCHECK_VCPU_INT_ACK__MIF_BSP1_HI_ACK_MASK 0x00800000L +#define UVD_MEMCHECK_VCPU_INT_ACK__SRE_LO_ACK_MASK 0x01000000L +#define UVD_MEMCHECK_VCPU_INT_ACK__SRE_HI_ACK_MASK 0x02000000L +#define UVD_MEMCHECK_VCPU_INT_ACK__IT_RD_LO_ACK_MASK 0x40000000L +#define UVD_MEMCHECK_VCPU_INT_ACK__IT_RD_HI_ACK_MASK 0x80000000L +//UVD_MEMCHECK2_SYS_INT_STAT +#define UVD_MEMCHECK2_SYS_INT_STAT__CM_RD_LO_ERR__SHIFT 0x0 +#define UVD_MEMCHECK2_SYS_INT_STAT__CM_RD_HI_ERR__SHIFT 0x1 +#define UVD_MEMCHECK2_SYS_INT_STAT__DB_RD_LO_ERR__SHIFT 0x2 +#define UVD_MEMCHECK2_SYS_INT_STAT__DB_RD_HI_ERR__SHIFT 0x3 +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_RD_LO_ERR__SHIFT 0x4 +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_RD_HI_ERR__SHIFT 0x5 +#define UVD_MEMCHECK2_SYS_INT_STAT__IDCT_RD_LO_ERR__SHIFT 0x6 +#define UVD_MEMCHECK2_SYS_INT_STAT__IDCT_RD_HI_ERR__SHIFT 0x7 +#define UVD_MEMCHECK2_SYS_INT_STAT__MPC_RD_LO_ERR__SHIFT 0x8 +#define UVD_MEMCHECK2_SYS_INT_STAT__MPC_RD_HI_ERR__SHIFT 0x9 +#define UVD_MEMCHECK2_SYS_INT_STAT__LBSI_RD_LO_ERR__SHIFT 0xa +#define UVD_MEMCHECK2_SYS_INT_STAT__LBSI_RD_HI_ERR__SHIFT 0xb +#define UVD_MEMCHECK2_SYS_INT_STAT__RBC_RD_LO_ERR__SHIFT 0x10 +#define UVD_MEMCHECK2_SYS_INT_STAT__RBC_RD_HI_ERR__SHIFT 0x11 +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_BSP2_LO_ERR__SHIFT 0x16 +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_BSP2_HI_ERR__SHIFT 0x17 +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_BSP3_LO_ERR__SHIFT 0x18 +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_BSP3_HI_ERR__SHIFT 0x19 +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_SCLR_LO_ERR__SHIFT 0x1a +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_SCLR_HI_ERR__SHIFT 0x1b +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_SCLR2_LO_ERR__SHIFT 0x1c +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_SCLR2_HI_ERR__SHIFT 0x1d +#define UVD_MEMCHECK2_SYS_INT_STAT__PREF_LO_ERR__SHIFT 0x1e +#define UVD_MEMCHECK2_SYS_INT_STAT__PREF_HI_ERR__SHIFT 0x1f +#define UVD_MEMCHECK2_SYS_INT_STAT__CM_RD_LO_ERR_MASK 0x00000001L +#define UVD_MEMCHECK2_SYS_INT_STAT__CM_RD_HI_ERR_MASK 0x00000002L +#define UVD_MEMCHECK2_SYS_INT_STAT__DB_RD_LO_ERR_MASK 0x00000004L +#define UVD_MEMCHECK2_SYS_INT_STAT__DB_RD_HI_ERR_MASK 0x00000008L +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_RD_LO_ERR_MASK 0x00000010L +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_RD_HI_ERR_MASK 0x00000020L +#define UVD_MEMCHECK2_SYS_INT_STAT__IDCT_RD_LO_ERR_MASK 0x00000040L +#define UVD_MEMCHECK2_SYS_INT_STAT__IDCT_RD_HI_ERR_MASK 0x00000080L +#define UVD_MEMCHECK2_SYS_INT_STAT__MPC_RD_LO_ERR_MASK 0x00000100L +#define UVD_MEMCHECK2_SYS_INT_STAT__MPC_RD_HI_ERR_MASK 0x00000200L +#define UVD_MEMCHECK2_SYS_INT_STAT__LBSI_RD_LO_ERR_MASK 0x00000400L +#define UVD_MEMCHECK2_SYS_INT_STAT__LBSI_RD_HI_ERR_MASK 0x00000800L +#define UVD_MEMCHECK2_SYS_INT_STAT__RBC_RD_LO_ERR_MASK 0x00010000L +#define UVD_MEMCHECK2_SYS_INT_STAT__RBC_RD_HI_ERR_MASK 0x00020000L +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_BSP2_LO_ERR_MASK 0x00400000L +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_BSP2_HI_ERR_MASK 0x00800000L +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_BSP3_LO_ERR_MASK 0x01000000L +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_BSP3_HI_ERR_MASK 0x02000000L +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_SCLR_LO_ERR_MASK 0x04000000L +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_SCLR_HI_ERR_MASK 0x08000000L +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_SCLR2_LO_ERR_MASK 0x10000000L +#define UVD_MEMCHECK2_SYS_INT_STAT__MIF_SCLR2_HI_ERR_MASK 0x20000000L +#define UVD_MEMCHECK2_SYS_INT_STAT__PREF_LO_ERR_MASK 0x40000000L +#define UVD_MEMCHECK2_SYS_INT_STAT__PREF_HI_ERR_MASK 0x80000000L +//UVD_MEMCHECK2_SYS_INT_ACK +#define UVD_MEMCHECK2_SYS_INT_ACK__CM_RD_LO_ACK__SHIFT 0x0 +#define UVD_MEMCHECK2_SYS_INT_ACK__CM_RD_HI_ACK__SHIFT 0x1 +#define UVD_MEMCHECK2_SYS_INT_ACK__DB_RD_LO_ACK__SHIFT 0x2 +#define UVD_MEMCHECK2_SYS_INT_ACK__DB_RD_HI_ACK__SHIFT 0x3 +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_RD_LO_ACK__SHIFT 0x4 +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_RD_HI_ACK__SHIFT 0x5 +#define UVD_MEMCHECK2_SYS_INT_ACK__IDCT_RD_LO_ACK__SHIFT 0x6 +#define UVD_MEMCHECK2_SYS_INT_ACK__IDCT_RD_HI_ACK__SHIFT 0x7 +#define UVD_MEMCHECK2_SYS_INT_ACK__MPC_RD_LO_ACK__SHIFT 0x8 +#define UVD_MEMCHECK2_SYS_INT_ACK__MPC_RD_HI_ACK__SHIFT 0x9 +#define UVD_MEMCHECK2_SYS_INT_ACK__LBSI_RD_LO_ACK__SHIFT 0xa +#define UVD_MEMCHECK2_SYS_INT_ACK__LBSI_RD_HI_ACK__SHIFT 0xb +#define UVD_MEMCHECK2_SYS_INT_ACK__RBC_RD_LO_ACK__SHIFT 0x10 +#define UVD_MEMCHECK2_SYS_INT_ACK__RBC_RD_HI_ACK__SHIFT 0x11 +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_BSP2_LO_ACK__SHIFT 0x16 +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_BSP2_HI_ACK__SHIFT 0x17 +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_BSP3_LO_ACK__SHIFT 0x18 +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_BSP3_HI_ACK__SHIFT 0x19 +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_SCLR_LO_ACK__SHIFT 0x1a +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_SCLR_HI_ACK__SHIFT 0x1b +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_SCLR2_LO_ACK__SHIFT 0x1c +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_SCLR2_HI_ACK__SHIFT 0x1d +#define UVD_MEMCHECK2_SYS_INT_ACK__PREF_LO_ACK__SHIFT 0x1e +#define UVD_MEMCHECK2_SYS_INT_ACK__PREF_HI_ACK__SHIFT 0x1f +#define UVD_MEMCHECK2_SYS_INT_ACK__CM_RD_LO_ACK_MASK 0x00000001L +#define UVD_MEMCHECK2_SYS_INT_ACK__CM_RD_HI_ACK_MASK 0x00000002L +#define UVD_MEMCHECK2_SYS_INT_ACK__DB_RD_LO_ACK_MASK 0x00000004L +#define UVD_MEMCHECK2_SYS_INT_ACK__DB_RD_HI_ACK_MASK 0x00000008L +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_RD_LO_ACK_MASK 0x00000010L +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_RD_HI_ACK_MASK 0x00000020L +#define UVD_MEMCHECK2_SYS_INT_ACK__IDCT_RD_LO_ACK_MASK 0x00000040L +#define UVD_MEMCHECK2_SYS_INT_ACK__IDCT_RD_HI_ACK_MASK 0x00000080L +#define UVD_MEMCHECK2_SYS_INT_ACK__MPC_RD_LO_ACK_MASK 0x00000100L +#define UVD_MEMCHECK2_SYS_INT_ACK__MPC_RD_HI_ACK_MASK 0x00000200L +#define UVD_MEMCHECK2_SYS_INT_ACK__LBSI_RD_LO_ACK_MASK 0x00000400L +#define UVD_MEMCHECK2_SYS_INT_ACK__LBSI_RD_HI_ACK_MASK 0x00000800L +#define UVD_MEMCHECK2_SYS_INT_ACK__RBC_RD_LO_ACK_MASK 0x00010000L +#define UVD_MEMCHECK2_SYS_INT_ACK__RBC_RD_HI_ACK_MASK 0x00020000L +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_BSP2_LO_ACK_MASK 0x00400000L +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_BSP2_HI_ACK_MASK 0x00800000L +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_BSP3_LO_ACK_MASK 0x01000000L +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_BSP3_HI_ACK_MASK 0x02000000L +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_SCLR_LO_ACK_MASK 0x04000000L +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_SCLR_HI_ACK_MASK 0x08000000L +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_SCLR2_LO_ACK_MASK 0x10000000L +#define UVD_MEMCHECK2_SYS_INT_ACK__MIF_SCLR2_HI_ACK_MASK 0x20000000L +#define UVD_MEMCHECK2_SYS_INT_ACK__PREF_LO_ACK_MASK 0x40000000L +#define UVD_MEMCHECK2_SYS_INT_ACK__PREF_HI_ACK_MASK 0x80000000L +//UVD_MEMCHECK2_VCPU_INT_STAT +#define UVD_MEMCHECK2_VCPU_INT_STAT__CM_RD_LO_ERR__SHIFT 0x0 +#define UVD_MEMCHECK2_VCPU_INT_STAT__CM_RD_HI_ERR__SHIFT 0x1 +#define UVD_MEMCHECK2_VCPU_INT_STAT__DB_RD_LO_ERR__SHIFT 0x2 +#define UVD_MEMCHECK2_VCPU_INT_STAT__DB_RD_HI_ERR__SHIFT 0x3 +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_RD_LO_ERR__SHIFT 0x4 +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_RD_HI_ERR__SHIFT 0x5 +#define UVD_MEMCHECK2_VCPU_INT_STAT__IDCT_RD_LO_ERR__SHIFT 0x6 +#define UVD_MEMCHECK2_VCPU_INT_STAT__IDCT_RD_HI_ERR__SHIFT 0x7 +#define UVD_MEMCHECK2_VCPU_INT_STAT__MPC_RD_LO_ERR__SHIFT 0x8 +#define UVD_MEMCHECK2_VCPU_INT_STAT__MPC_RD_HI_ERR__SHIFT 0x9 +#define UVD_MEMCHECK2_VCPU_INT_STAT__LBSI_RD_LO_ERR__SHIFT 0xa +#define UVD_MEMCHECK2_VCPU_INT_STAT__LBSI_RD_HI_ERR__SHIFT 0xb +#define UVD_MEMCHECK2_VCPU_INT_STAT__RBC_RD_LO_ERR__SHIFT 0x10 +#define UVD_MEMCHECK2_VCPU_INT_STAT__RBC_RD_HI_ERR__SHIFT 0x11 +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_BSP2_LO_ERR__SHIFT 0x12 +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_BSP2_HI_ERR__SHIFT 0x13 +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_BSP3_LO_ERR__SHIFT 0x14 +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_BSP3_HI_ERR__SHIFT 0x15 +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_SCLR_LO_ERR__SHIFT 0x16 +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_SCLR_HI_ERR__SHIFT 0x17 +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_SCLR2_LO_ERR__SHIFT 0x18 +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_SCLR2_HI_ERR__SHIFT 0x19 +#define UVD_MEMCHECK2_VCPU_INT_STAT__PREF_LO_ERR__SHIFT 0x1a +#define UVD_MEMCHECK2_VCPU_INT_STAT__PREF_HI_ERR__SHIFT 0x1b +#define UVD_MEMCHECK2_VCPU_INT_STAT__CM_RD_LO_ERR_MASK 0x00000001L +#define UVD_MEMCHECK2_VCPU_INT_STAT__CM_RD_HI_ERR_MASK 0x00000002L +#define UVD_MEMCHECK2_VCPU_INT_STAT__DB_RD_LO_ERR_MASK 0x00000004L +#define UVD_MEMCHECK2_VCPU_INT_STAT__DB_RD_HI_ERR_MASK 0x00000008L +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_RD_LO_ERR_MASK 0x00000010L +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_RD_HI_ERR_MASK 0x00000020L +#define UVD_MEMCHECK2_VCPU_INT_STAT__IDCT_RD_LO_ERR_MASK 0x00000040L +#define UVD_MEMCHECK2_VCPU_INT_STAT__IDCT_RD_HI_ERR_MASK 0x00000080L +#define UVD_MEMCHECK2_VCPU_INT_STAT__MPC_RD_LO_ERR_MASK 0x00000100L +#define UVD_MEMCHECK2_VCPU_INT_STAT__MPC_RD_HI_ERR_MASK 0x00000200L +#define UVD_MEMCHECK2_VCPU_INT_STAT__LBSI_RD_LO_ERR_MASK 0x00000400L +#define UVD_MEMCHECK2_VCPU_INT_STAT__LBSI_RD_HI_ERR_MASK 0x00000800L +#define UVD_MEMCHECK2_VCPU_INT_STAT__RBC_RD_LO_ERR_MASK 0x00010000L +#define UVD_MEMCHECK2_VCPU_INT_STAT__RBC_RD_HI_ERR_MASK 0x00020000L +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_BSP2_LO_ERR_MASK 0x00040000L +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_BSP2_HI_ERR_MASK 0x00080000L +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_BSP3_LO_ERR_MASK 0x00100000L +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_BSP3_HI_ERR_MASK 0x00200000L +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_SCLR_LO_ERR_MASK 0x00400000L +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_SCLR_HI_ERR_MASK 0x00800000L +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_SCLR2_LO_ERR_MASK 0x01000000L +#define UVD_MEMCHECK2_VCPU_INT_STAT__MIF_SCLR2_HI_ERR_MASK 0x02000000L +#define UVD_MEMCHECK2_VCPU_INT_STAT__PREF_LO_ERR_MASK 0x04000000L +#define UVD_MEMCHECK2_VCPU_INT_STAT__PREF_HI_ERR_MASK 0x08000000L +//UVD_MEMCHECK2_VCPU_INT_ACK +#define UVD_MEMCHECK2_VCPU_INT_ACK__CM_RD_LO_ACK__SHIFT 0x0 +#define UVD_MEMCHECK2_VCPU_INT_ACK__CM_RD_HI_ACK__SHIFT 0x1 +#define UVD_MEMCHECK2_VCPU_INT_ACK__DB_RD_LO_ACK__SHIFT 0x2 +#define UVD_MEMCHECK2_VCPU_INT_ACK__DB_RD_HI_ACK__SHIFT 0x3 +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_RD_LO_ACK__SHIFT 0x4 +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_RD_HI_ACK__SHIFT 0x5 +#define UVD_MEMCHECK2_VCPU_INT_ACK__IDCT_RD_LO_ACK__SHIFT 0x6 +#define UVD_MEMCHECK2_VCPU_INT_ACK__IDCT_RD_HI_ACK__SHIFT 0x7 +#define UVD_MEMCHECK2_VCPU_INT_ACK__MPC_RD_LO_ACK__SHIFT 0x8 +#define UVD_MEMCHECK2_VCPU_INT_ACK__MPC_RD_HI_ACK__SHIFT 0x9 +#define UVD_MEMCHECK2_VCPU_INT_ACK__LBSI_RD_LO_ACK__SHIFT 0xa +#define UVD_MEMCHECK2_VCPU_INT_ACK__LBSI_RD_HI_ACK__SHIFT 0xb +#define UVD_MEMCHECK2_VCPU_INT_ACK__RBC_RD_LO_ACK__SHIFT 0x10 +#define UVD_MEMCHECK2_VCPU_INT_ACK__RBC_RD_HI_ACK__SHIFT 0x11 +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_BSP2_LO_ACK__SHIFT 0x12 +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_BSP2_HI_ACK__SHIFT 0x13 +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_BSP3_LO_ACK__SHIFT 0x14 +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_BSP3_HI_ACK__SHIFT 0x15 +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_SCLR_LO_ACK__SHIFT 0x16 +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_SCLR_HI_ACK__SHIFT 0x17 +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_SCLR2_LO_ACK__SHIFT 0x18 +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_SCLR2_HI_ACK__SHIFT 0x19 +#define UVD_MEMCHECK2_VCPU_INT_ACK__PREF_LO_ACK__SHIFT 0x1a +#define UVD_MEMCHECK2_VCPU_INT_ACK__PREF_HI_ACK__SHIFT 0x1b +#define UVD_MEMCHECK2_VCPU_INT_ACK__CM_RD_LO_ACK_MASK 0x00000001L +#define UVD_MEMCHECK2_VCPU_INT_ACK__CM_RD_HI_ACK_MASK 0x00000002L +#define UVD_MEMCHECK2_VCPU_INT_ACK__DB_RD_LO_ACK_MASK 0x00000004L +#define UVD_MEMCHECK2_VCPU_INT_ACK__DB_RD_HI_ACK_MASK 0x00000008L +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_RD_LO_ACK_MASK 0x00000010L +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_RD_HI_ACK_MASK 0x00000020L +#define UVD_MEMCHECK2_VCPU_INT_ACK__IDCT_RD_LO_ACK_MASK 0x00000040L +#define UVD_MEMCHECK2_VCPU_INT_ACK__IDCT_RD_HI_ACK_MASK 0x00000080L +#define UVD_MEMCHECK2_VCPU_INT_ACK__MPC_RD_LO_ACK_MASK 0x00000100L +#define UVD_MEMCHECK2_VCPU_INT_ACK__MPC_RD_HI_ACK_MASK 0x00000200L +#define UVD_MEMCHECK2_VCPU_INT_ACK__LBSI_RD_LO_ACK_MASK 0x00000400L +#define UVD_MEMCHECK2_VCPU_INT_ACK__LBSI_RD_HI_ACK_MASK 0x00000800L +#define UVD_MEMCHECK2_VCPU_INT_ACK__RBC_RD_LO_ACK_MASK 0x00010000L +#define UVD_MEMCHECK2_VCPU_INT_ACK__RBC_RD_HI_ACK_MASK 0x00020000L +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_BSP2_LO_ACK_MASK 0x00040000L +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_BSP2_HI_ACK_MASK 0x00080000L +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_BSP3_LO_ACK_MASK 0x00100000L +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_BSP3_HI_ACK_MASK 0x00200000L +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_SCLR_LO_ACK_MASK 0x00400000L +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_SCLR_HI_ACK_MASK 0x00800000L +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_SCLR2_LO_ACK_MASK 0x01000000L +#define UVD_MEMCHECK2_VCPU_INT_ACK__MIF_SCLR2_HI_ACK_MASK 0x02000000L +#define UVD_MEMCHECK2_VCPU_INT_ACK__PREF_LO_ACK_MASK 0x04000000L +#define UVD_MEMCHECK2_VCPU_INT_ACK__PREF_HI_ACK_MASK 0x08000000L + + +#endif diff --git a/drivers/gpu/drm/amd/include/atomfirmware.h b/drivers/gpu/drm/amd/include/atomfirmware.h index bbe1337a8cee..e68c1e280322 100644 --- a/drivers/gpu/drm/amd/include/atomfirmware.h +++ b/drivers/gpu/drm/amd/include/atomfirmware.h @@ -182,6 +182,7 @@ enum atom_dgpu_vram_type { ATOM_DGPU_VRAM_TYPE_HBM2 = 0x60, ATOM_DGPU_VRAM_TYPE_HBM2E = 0x61, ATOM_DGPU_VRAM_TYPE_GDDR6 = 0x70, + ATOM_DGPU_VRAM_TYPE_HBM3 = 0x80, }; enum atom_dp_vs_preemph_def{ diff --git a/drivers/gpu/drm/amd/include/discovery.h b/drivers/gpu/drm/amd/include/discovery.h index f150404ffc68..f43e29722ef7 100644 --- a/drivers/gpu/drm/amd/include/discovery.h +++ b/drivers/gpu/drm/amd/include/discovery.h @@ -79,7 +79,14 @@ typedef struct ip_discovery_header uint32_t id; /* Table ID */ uint16_t num_dies; /* Number of Dies */ die_info die_info[16]; /* list die information for up to 16 dies */ - uint16_t padding[1]; /* padding */ + union { + uint16_t padding[1]; /* version <= 3 */ + struct { /* version == 4 */ + uint8_t base_addr_64_bit : 1; /* ip structures are using 64 bit base address */ + uint8_t reserved : 7; + uint8_t reserved2; + }; + }; } ip_discovery_header; typedef struct ip @@ -115,9 +122,29 @@ typedef struct ip_v3 uint8_t sub_revision : 4; /* HCID Sub-Revision */ uint8_t variant : 4; /* HW variant */ #endif - uint32_t base_address[1]; /* Base Address list. Corresponds to the num_base_address field*/ + uint32_t base_address[]; /* Base Address list. Corresponds to the num_base_address field*/ } ip_v3; +typedef struct ip_v4 { + uint16_t hw_id; /* Hardware ID */ + uint8_t instance_number; /* Instance number for the IP */ + uint8_t num_base_address; /* Number of base addresses*/ + uint8_t major; /* Hardware ID.major version */ + uint8_t minor; /* Hardware ID.minor version */ + uint8_t revision; /* Hardware ID.revision version */ +#if defined(LITTLEENDIAN_CPU) + uint8_t sub_revision : 4; /* HCID Sub-Revision */ + uint8_t variant : 4; /* HW variant */ +#elif defined(BIGENDIAN_CPU) + uint8_t variant : 4; /* HW variant */ + uint8_t sub_revision : 4; /* HCID Sub-Revision */ +#endif + union { + DECLARE_FLEX_ARRAY(uint32_t, base_address); /* 32-bit Base Address list. Corresponds to the num_base_address field*/ + DECLARE_FLEX_ARRAY(uint64_t, base_address_64); /* 64-bit Base Address list. Corresponds to the num_base_address field*/ + } __packed; +} ip_v4; + typedef struct die_header { uint16_t die_id; @@ -134,6 +161,7 @@ typedef struct ip_structure { ip *ip_list; ip_v3 *ip_v3_list; + ip_v4 *ip_v4_list; }; /* IP list. Variable size*/ } die; } ip_structure; diff --git a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h index 5cb3e8634739..d0df3381539f 100644 --- a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h +++ b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h @@ -230,28 +230,30 @@ struct kfd2kgd_calls { /* Register access functions */ void (*program_sh_mem_settings)(struct amdgpu_device *adev, uint32_t vmid, uint32_t sh_mem_config, uint32_t sh_mem_ape1_base, - uint32_t sh_mem_ape1_limit, uint32_t sh_mem_bases); + uint32_t sh_mem_ape1_limit, uint32_t sh_mem_bases, + uint32_t inst); int (*set_pasid_vmid_mapping)(struct amdgpu_device *adev, u32 pasid, - unsigned int vmid); + unsigned int vmid, uint32_t inst); - int (*init_interrupts)(struct amdgpu_device *adev, uint32_t pipe_id); + int (*init_interrupts)(struct amdgpu_device *adev, uint32_t pipe_id, + uint32_t inst); int (*hqd_load)(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, uint32_t __user *wptr, uint32_t wptr_shift, uint32_t wptr_mask, - struct mm_struct *mm); + struct mm_struct *mm, uint32_t inst); int (*hiq_mqd_load)(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id, uint32_t queue_id, - uint32_t doorbell_off); + uint32_t doorbell_off, uint32_t inst); int (*hqd_sdma_load)(struct amdgpu_device *adev, void *mqd, uint32_t __user *wptr, struct mm_struct *mm); int (*hqd_dump)(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t queue_id, - uint32_t (**dump)[2], uint32_t *n_regs); + uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst); int (*hqd_sdma_dump)(struct amdgpu_device *adev, uint32_t engine_id, uint32_t queue_id, @@ -259,12 +261,12 @@ struct kfd2kgd_calls { bool (*hqd_is_occupied)(struct amdgpu_device *adev, uint64_t queue_address, uint32_t pipe_id, - uint32_t queue_id); + uint32_t queue_id, uint32_t inst); int (*hqd_destroy)(struct amdgpu_device *adev, void *mqd, enum kfd_preempt_type reset_type, unsigned int timeout, uint32_t pipe_id, - uint32_t queue_id); + uint32_t queue_id, uint32_t inst); bool (*hqd_sdma_is_occupied)(struct amdgpu_device *adev, void *mqd); @@ -273,7 +275,7 @@ struct kfd2kgd_calls { int (*wave_control_execute)(struct amdgpu_device *adev, uint32_t gfx_index_val, - uint32_t sq_cmd); + uint32_t sq_cmd, uint32_t inst); bool (*get_atc_vmid_pasid_mapping_info)(struct amdgpu_device *adev, uint8_t vmid, uint16_t *p_pasid); @@ -289,10 +291,45 @@ struct kfd2kgd_calls { uint32_t vmid, uint64_t page_table_base); uint32_t (*read_vmid_from_vmfault_reg)(struct amdgpu_device *adev); + uint32_t (*enable_debug_trap)(struct amdgpu_device *adev, + bool restore_dbg_registers, + uint32_t vmid); + uint32_t (*disable_debug_trap)(struct amdgpu_device *adev, + bool keep_trap_enabled, + uint32_t vmid); + int (*validate_trap_override_request)(struct amdgpu_device *adev, + uint32_t trap_override, + uint32_t *trap_mask_supported); + uint32_t (*set_wave_launch_trap_override)(struct amdgpu_device *adev, + uint32_t vmid, + uint32_t trap_override, + uint32_t trap_mask_bits, + uint32_t trap_mask_request, + uint32_t *trap_mask_prev, + uint32_t kfd_dbg_trap_cntl_prev); + uint32_t (*set_wave_launch_mode)(struct amdgpu_device *adev, + uint8_t wave_launch_mode, + uint32_t vmid); + uint32_t (*set_address_watch)(struct amdgpu_device *adev, + uint64_t watch_address, + uint32_t watch_address_mask, + uint32_t watch_id, + uint32_t watch_mode, + uint32_t debug_vmid); + uint32_t (*clear_address_watch)(struct amdgpu_device *adev, + uint32_t watch_id); + void (*get_iq_wait_times)(struct amdgpu_device *adev, + uint32_t *wait_times); + void (*build_grace_period_packet_info)(struct amdgpu_device *adev, + uint32_t wait_times, + uint32_t grace_period, + uint32_t *reg_offset, + uint32_t *reg_data); void (*get_cu_occupancy)(struct amdgpu_device *adev, int pasid, - int *wave_cnt, int *max_waves_per_cu); + int *wave_cnt, int *max_waves_per_cu, uint32_t inst); void (*program_trap_handler_settings)(struct amdgpu_device *adev, - uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr); + uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr, + uint32_t inst); }; #endif /* KGD_KFD_INTERFACE_H_INCLUDED */ diff --git a/drivers/gpu/drm/amd/include/mes_v11_api_def.h b/drivers/gpu/drm/amd/include/mes_v11_api_def.h index dc694cb246d9..0997e999416a 100644 --- a/drivers/gpu/drm/amd/include/mes_v11_api_def.h +++ b/drivers/gpu/drm/amd/include/mes_v11_api_def.h @@ -274,7 +274,8 @@ union MESAPI__ADD_QUEUE { uint32_t is_kfd_process : 1; uint32_t trap_en : 1; uint32_t is_aql_queue : 1; - uint32_t reserved : 20; + uint32_t skip_process_ctx_clear : 1; + uint32_t reserved : 19; }; struct MES_API_STATUS api_status; uint64_t tma_addr; @@ -523,6 +524,7 @@ enum MESAPI_MISC_OPCODE { MESAPI_MISC__QUERY_STATUS, MESAPI_MISC__READ_REG, MESAPI_MISC__WAIT_REG_MEM, + MESAPI_MISC__SET_SHADER_DEBUGGER, MESAPI_MISC__MAX, }; @@ -561,6 +563,21 @@ struct QUERY_STATUS { uint32_t context_id; }; +struct SET_SHADER_DEBUGGER { + uint64_t process_context_addr; + union { + struct { + uint32_t single_memop : 1; /* SQ_DEBUG.single_memop */ + uint32_t single_alu_op : 1; /* SQ_DEBUG.single_alu_op */ + uint32_t reserved : 30; + }; + uint32_t u32all; + } flags; + uint32_t spi_gdbg_per_vmid_cntl; + uint32_t tcp_watch_cntl[4]; /* TCP_WATCHx_CNTL */ + uint32_t trap_en; +}; + union MESAPI__MISC { struct { union MES_API_HEADER header; @@ -573,6 +590,9 @@ union MESAPI__MISC { struct QUERY_STATUS query_status; struct READ_REG read_reg; struct WAIT_REG_MEM wait_reg_mem; + struct SET_SHADER_DEBUGGER set_shader_debugger; + enum MES_AMD_PRIORITY_LEVEL queue_sch_level; + uint32_t data[MISC_DATA_MAX_SIZE_IN_DWORDS]; }; }; diff --git a/drivers/gpu/drm/amd/include/v9_structs.h b/drivers/gpu/drm/amd/include/v9_structs.h index a0c672889fe4..a2f81b9c38af 100644 --- a/drivers/gpu/drm/amd/include/v9_structs.h +++ b/drivers/gpu/drm/amd/include/v9_structs.h @@ -196,10 +196,20 @@ struct v9_mqd { uint32_t compute_wave_restore_addr_lo; uint32_t compute_wave_restore_addr_hi; uint32_t compute_wave_restore_control; - uint32_t compute_static_thread_mgmt_se4; - uint32_t compute_static_thread_mgmt_se5; - uint32_t compute_static_thread_mgmt_se6; - uint32_t compute_static_thread_mgmt_se7; + union { + struct { + uint32_t compute_static_thread_mgmt_se4; + uint32_t compute_static_thread_mgmt_se5; + uint32_t compute_static_thread_mgmt_se6; + uint32_t compute_static_thread_mgmt_se7; + }; + struct { + uint32_t compute_current_logic_xcc_id; // offset: 39 (0x27) + uint32_t compute_restart_cg_tg_id; // offset: 40 (0x28) + uint32_t compute_tg_chunk_size; // offset: 41 (0x29) + uint32_t compute_restore_tg_chunk_size; // offset: 42 (0x2A) + }; + }; uint32_t reserved_43; uint32_t reserved_44; uint32_t reserved_45; @@ -382,8 +392,16 @@ struct v9_mqd { uint32_t iqtimer_pkt_dw29; uint32_t iqtimer_pkt_dw30; uint32_t iqtimer_pkt_dw31; - uint32_t reserved_225; - uint32_t reserved_226; + union { + struct { + uint32_t reserved_225; + uint32_t reserved_226; + }; + struct { + uint32_t pm4_target_xcc_in_xcp; // offset: 225 (0xE1) + uint32_t cp_mqd_stride_size; // offset: 226 (0xE2) + }; + }; uint32_t reserved_227; uint32_t set_resources_header; uint32_t set_resources_dw1; diff --git a/drivers/gpu/drm/amd/pm/amdgpu_pm.c b/drivers/gpu/drm/amd/pm/amdgpu_pm.c index 58c2246918fd..a57952b93e73 100644 --- a/drivers/gpu/drm/amd/pm/amdgpu_pm.c +++ b/drivers/gpu/drm/amd/pm/amdgpu_pm.c @@ -678,7 +678,12 @@ static ssize_t amdgpu_set_pp_table(struct device *dev, * clock labeled OD_MCLK * * - three <frequency, voltage> points labeled OD_VDDC_CURVE. - * They can be used to calibrate the sclk voltage curve. + * They can be used to calibrate the sclk voltage curve. This is + * available for Vega20 and NV1X. + * + * - voltage offset for the six anchor points of the v/f curve labeled + * OD_VDDC_CURVE. They can be used to calibrate the v/f curve. This + * is only availabe for some SMU13 ASICs. * * - voltage offset(in mV) applied on target voltage calculation. * This is available for Sienna Cichlid, Navy Flounder and Dimgrey @@ -719,12 +724,19 @@ static ssize_t amdgpu_set_pp_table(struct device *dev, * E.g., "p 2 0 800" would set the minimum core clock on core * 2 to 800Mhz. * - * For sclk voltage curve, enter the new values by writing a - * string that contains "vc point clock voltage" to the file. The - * points are indexed by 0, 1 and 2. E.g., "vc 0 300 600" will - * update point1 with clock set as 300Mhz and voltage as - * 600mV. "vc 2 1000 1000" will update point3 with clock set - * as 1000Mhz and voltage 1000mV. + * For sclk voltage curve, + * - For NV1X, enter the new values by writing a string that + * contains "vc point clock voltage" to the file. The points + * are indexed by 0, 1 and 2. E.g., "vc 0 300 600" will update + * point1 with clock set as 300Mhz and voltage as 600mV. "vc 2 + * 1000 1000" will update point3 with clock set as 1000Mhz and + * voltage 1000mV. + * - For SMU13 ASICs, enter the new values by writing a string that + * contains "vc anchor_point_index voltage_offset" to the file. + * There are total six anchor points defined on the v/f curve with + * index as 0 - 5. + * - "vc 0 10" will update the voltage offset for point1 as 10mv. + * - "vc 5 -10" will update the voltage offset for point6 as -10mv. * * To update the voltage offset applied for gfxclk/voltage calculation, * enter the new value by writing a string that contains "vo offset". @@ -871,13 +883,11 @@ static ssize_t amdgpu_get_pp_od_clk_voltage(struct device *dev, } if (ret == -ENOENT) { size = amdgpu_dpm_print_clock_levels(adev, OD_SCLK, buf); - if (size > 0) { - size += amdgpu_dpm_print_clock_levels(adev, OD_MCLK, buf + size); - size += amdgpu_dpm_print_clock_levels(adev, OD_VDDC_CURVE, buf + size); - size += amdgpu_dpm_print_clock_levels(adev, OD_VDDGFX_OFFSET, buf + size); - size += amdgpu_dpm_print_clock_levels(adev, OD_RANGE, buf + size); - size += amdgpu_dpm_print_clock_levels(adev, OD_CCLK, buf + size); - } + size += amdgpu_dpm_print_clock_levels(adev, OD_MCLK, buf + size); + size += amdgpu_dpm_print_clock_levels(adev, OD_VDDC_CURVE, buf + size); + size += amdgpu_dpm_print_clock_levels(adev, OD_VDDGFX_OFFSET, buf + size); + size += amdgpu_dpm_print_clock_levels(adev, OD_RANGE, buf + size); + size += amdgpu_dpm_print_clock_levels(adev, OD_CCLK, buf + size); } if (size == 0) @@ -3362,7 +3372,8 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj, return 0; /* Skip crit temp on APU */ - if ((adev->flags & AMD_IS_APU) && (adev->family >= AMDGPU_FAMILY_CZ) && + if ((((adev->flags & AMD_IS_APU) && (adev->family >= AMDGPU_FAMILY_CZ)) || + (gc_ver == IP_VERSION(9, 4, 3))) && (attr == &sensor_dev_attr_temp1_crit.dev_attr.attr || attr == &sensor_dev_attr_temp1_crit_hyst.dev_attr.attr)) return 0; @@ -3395,9 +3406,10 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj, attr == &sensor_dev_attr_pwm1_enable.dev_attr.attr)) /* can't manage state */ effective_mode &= ~S_IWUSR; - /* In the case of APUs, this is only implemented on Vangogh */ + /* not implemented yet for APUs other than GC 10.3.1 (vangogh) and 9.4.3 */ if (((adev->family == AMDGPU_FAMILY_SI) || - ((adev->flags & AMD_IS_APU) && (gc_ver != IP_VERSION(10, 3, 1)))) && + ((adev->flags & AMD_IS_APU) && (gc_ver != IP_VERSION(10, 3, 1)) && + (gc_ver != IP_VERSION(9, 4, 3)))) && (attr == &sensor_dev_attr_power1_cap_max.dev_attr.attr || attr == &sensor_dev_attr_power1_cap_min.dev_attr.attr || attr == &sensor_dev_attr_power1_cap.dev_attr.attr || @@ -3426,36 +3438,48 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj, return 0; if ((adev->family == AMDGPU_FAMILY_SI || /* not implemented yet */ - adev->family == AMDGPU_FAMILY_KV) && /* not implemented yet */ + adev->family == AMDGPU_FAMILY_KV || /* not implemented yet */ + (gc_ver == IP_VERSION(9, 4, 3))) && (attr == &sensor_dev_attr_in0_input.dev_attr.attr || attr == &sensor_dev_attr_in0_label.dev_attr.attr)) return 0; - /* only APUs have vddnb */ - if (!(adev->flags & AMD_IS_APU) && + /* only APUs other than gc 9,4,3 have vddnb */ + if ((!(adev->flags & AMD_IS_APU) || (gc_ver == IP_VERSION(9, 4, 3))) && (attr == &sensor_dev_attr_in1_input.dev_attr.attr || attr == &sensor_dev_attr_in1_label.dev_attr.attr)) return 0; - /* no mclk on APUs */ - if ((adev->flags & AMD_IS_APU) && + /* no mclk on APUs other than gc 9,4,3*/ + if (((adev->flags & AMD_IS_APU) && (gc_ver != IP_VERSION(9, 4, 3))) && (attr == &sensor_dev_attr_freq2_input.dev_attr.attr || attr == &sensor_dev_attr_freq2_label.dev_attr.attr)) return 0; - /* only SOC15 dGPUs support hotspot and mem temperatures */ if (((adev->flags & AMD_IS_APU) || gc_ver < IP_VERSION(9, 0, 0)) && + (gc_ver != IP_VERSION(9, 4, 3)) && + (attr == &sensor_dev_attr_temp2_input.dev_attr.attr || + attr == &sensor_dev_attr_temp2_label.dev_attr.attr || + attr == &sensor_dev_attr_temp3_input.dev_attr.attr || + attr == &sensor_dev_attr_temp3_label.dev_attr.attr)) + return 0; + + /* hotspot temperature for gc 9,4,3*/ + if ((gc_ver == IP_VERSION(9, 4, 3)) && + (attr == &sensor_dev_attr_temp1_input.dev_attr.attr || + attr == &sensor_dev_attr_temp1_label.dev_attr.attr)) + return 0; + + /* only SOC15 dGPUs support hotspot and mem temperatures */ + if (((adev->flags & AMD_IS_APU) || gc_ver < IP_VERSION(9, 0, 0) || + (gc_ver == IP_VERSION(9, 4, 3))) && (attr == &sensor_dev_attr_temp2_crit.dev_attr.attr || attr == &sensor_dev_attr_temp2_crit_hyst.dev_attr.attr || attr == &sensor_dev_attr_temp3_crit.dev_attr.attr || attr == &sensor_dev_attr_temp3_crit_hyst.dev_attr.attr || attr == &sensor_dev_attr_temp1_emergency.dev_attr.attr || attr == &sensor_dev_attr_temp2_emergency.dev_attr.attr || - attr == &sensor_dev_attr_temp3_emergency.dev_attr.attr || - attr == &sensor_dev_attr_temp2_input.dev_attr.attr || - attr == &sensor_dev_attr_temp3_input.dev_attr.attr || - attr == &sensor_dev_attr_temp2_label.dev_attr.attr || - attr == &sensor_dev_attr_temp3_label.dev_attr.attr)) + attr == &sensor_dev_attr_temp3_emergency.dev_attr.attr)) return 0; /* only Vangogh has fast PPT limit and power labels */ diff --git a/drivers/gpu/drm/amd/pm/inc/smu_v13_0_0_pptable.h b/drivers/gpu/drm/amd/pm/inc/smu_v13_0_0_pptable.h index 566a0da59e53..1dc7a065a6d4 100644 --- a/drivers/gpu/drm/amd/pm/inc/smu_v13_0_0_pptable.h +++ b/drivers/gpu/drm/amd/pm/inc/smu_v13_0_0_pptable.h @@ -38,13 +38,12 @@ #define SMU_13_0_0_PP_THERMALCONTROLLER_NONE 0 #define SMU_13_0_0_PP_THERMALCONTROLLER_NAVI21 28 -#define SMU_13_0_0_PP_OVERDRIVE_VERSION 0x81 // OverDrive 8 Table Version 0.2 +#define SMU_13_0_0_PP_OVERDRIVE_VERSION 0x83 // OverDrive 8 Table Version 0.2 #define SMU_13_0_0_PP_POWERSAVINGCLOCK_VERSION 0x01 // Power Saving Clock Table Version 1.00 enum SMU_13_0_0_ODFEATURE_CAP { SMU_13_0_0_ODCAP_GFXCLK_LIMITS = 0, - SMU_13_0_0_ODCAP_GFXCLK_CURVE, SMU_13_0_0_ODCAP_UCLK_LIMITS, SMU_13_0_0_ODCAP_POWER_LIMIT, SMU_13_0_0_ODCAP_FAN_ACOUSTIC_LIMIT, @@ -59,13 +58,13 @@ enum SMU_13_0_0_ODFEATURE_CAP SMU_13_0_0_ODCAP_FAN_CURVE, SMU_13_0_0_ODCAP_AUTO_FAN_ACOUSTIC_LIMIT, SMU_13_0_0_ODCAP_POWER_MODE, + SMU_13_0_0_ODCAP_PER_ZONE_GFX_VOLTAGE_OFFSET, SMU_13_0_0_ODCAP_COUNT, }; enum SMU_13_0_0_ODFEATURE_ID { SMU_13_0_0_ODFEATURE_GFXCLK_LIMITS = 1 << SMU_13_0_0_ODCAP_GFXCLK_LIMITS, //GFXCLK Limit feature - SMU_13_0_0_ODFEATURE_GFXCLK_CURVE = 1 << SMU_13_0_0_ODCAP_GFXCLK_CURVE, //GFXCLK Curve feature SMU_13_0_0_ODFEATURE_UCLK_LIMITS = 1 << SMU_13_0_0_ODCAP_UCLK_LIMITS, //UCLK Limit feature SMU_13_0_0_ODFEATURE_POWER_LIMIT = 1 << SMU_13_0_0_ODCAP_POWER_LIMIT, //Power Limit feature SMU_13_0_0_ODFEATURE_FAN_ACOUSTIC_LIMIT = 1 << SMU_13_0_0_ODCAP_FAN_ACOUSTIC_LIMIT, //Fan Acoustic RPM feature @@ -80,6 +79,7 @@ enum SMU_13_0_0_ODFEATURE_ID SMU_13_0_0_ODFEATURE_FAN_CURVE = 1 << SMU_13_0_0_ODCAP_FAN_CURVE, //Fan Curve feature SMU_13_0_0_ODFEATURE_AUTO_FAN_ACOUSTIC_LIMIT = 1 << SMU_13_0_0_ODCAP_AUTO_FAN_ACOUSTIC_LIMIT, //Auto Fan Acoustic RPM feature SMU_13_0_0_ODFEATURE_POWER_MODE = 1 << SMU_13_0_0_ODCAP_POWER_MODE, //Optimized GPU Power Mode feature + SMU_13_0_0_ODFEATURE_PER_ZONE_GFX_VOLTAGE_OFFSET = 1 << SMU_13_0_0_ODCAP_PER_ZONE_GFX_VOLTAGE_OFFSET, //Perzone voltage offset feature SMU_13_0_0_ODFEATURE_COUNT = 16, }; @@ -89,10 +89,6 @@ enum SMU_13_0_0_ODSETTING_ID { SMU_13_0_0_ODSETTING_GFXCLKFMAX = 0, SMU_13_0_0_ODSETTING_GFXCLKFMIN, - SMU_13_0_0_ODSETTING_CUSTOM_GFX_VF_CURVE_A, - SMU_13_0_0_ODSETTING_CUSTOM_GFX_VF_CURVE_B, - SMU_13_0_0_ODSETTING_CUSTOM_GFX_VF_CURVE_C, - SMU_13_0_0_ODSETTING_CUSTOM_CURVE_VFT_FMIN, SMU_13_0_0_ODSETTING_UCLKFMIN, SMU_13_0_0_ODSETTING_UCLKFMAX, SMU_13_0_0_ODSETTING_POWERPERCENTAGE, @@ -117,6 +113,12 @@ enum SMU_13_0_0_ODSETTING_ID SMU_13_0_0_ODSETTING_FAN_CURVE_SPEED_5, SMU_13_0_0_ODSETTING_AUTO_FAN_ACOUSTIC_LIMIT, SMU_13_0_0_ODSETTING_POWER_MODE, + SMU_13_0_0_ODSETTING_PER_ZONE_GFX_VOLTAGE_OFFSET_POINT_1, + SMU_13_0_0_ODSETTING_PER_ZONE_GFX_VOLTAGE_OFFSET_POINT_2, + SMU_13_0_0_ODSETTING_PER_ZONE_GFX_VOLTAGE_OFFSET_POINT_3, + SMU_13_0_0_ODSETTING_PER_ZONE_GFX_VOLTAGE_OFFSET_POINT_4, + SMU_13_0_0_ODSETTING_PER_ZONE_GFX_VOLTAGE_OFFSET_POINT_5, + SMU_13_0_0_ODSETTING_PER_ZONE_GFX_VOLTAGE_OFFSET_POINT_6, SMU_13_0_0_ODSETTING_COUNT, }; #define SMU_13_0_0_MAX_ODSETTING 64 //Maximum Number of ODSettings diff --git a/drivers/gpu/drm/amd/pm/legacy-dpm/kv_dpm.c b/drivers/gpu/drm/amd/pm/legacy-dpm/kv_dpm.c index f5e08b60f66e..36c831b280ed 100644 --- a/drivers/gpu/drm/amd/pm/legacy-dpm/kv_dpm.c +++ b/drivers/gpu/drm/amd/pm/legacy-dpm/kv_dpm.c @@ -508,19 +508,19 @@ static int kv_enable_didt(struct amdgpu_device *adev, bool enable) pi->caps_db_ramping || pi->caps_td_ramping || pi->caps_tcp_ramping) { - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); if (enable) { ret = kv_program_pt_config_registers(adev, didt_config_kv); if (ret) { - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return ret; } } kv_do_enable_didt(adev, enable); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } return 0; diff --git a/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c b/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c index d6d9e3b1b2c0..02e69ccff3ba 100644 --- a/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c +++ b/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c @@ -6925,23 +6925,6 @@ static int si_dpm_enable(struct amdgpu_device *adev) return 0; } -static int si_set_temperature_range(struct amdgpu_device *adev) -{ - int ret; - - ret = si_thermal_enable_alert(adev, false); - if (ret) - return ret; - ret = si_thermal_set_temperature_range(adev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX); - if (ret) - return ret; - ret = si_thermal_enable_alert(adev, true); - if (ret) - return ret; - - return ret; -} - static void si_dpm_disable(struct amdgpu_device *adev) { struct rv7xx_power_info *pi = rv770_get_pi(adev); @@ -7626,18 +7609,6 @@ static int si_dpm_process_interrupt(struct amdgpu_device *adev, static int si_dpm_late_init(void *handle) { - int ret; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - if (!adev->pm.dpm_enabled) - return 0; - - ret = si_set_temperature_range(adev); - if (ret) - return ret; -#if 0 //TODO ? - si_dpm_powergate_uvd(adev, true); -#endif return 0; } diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_powertune.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_powertune.c index 32a5a00fd8ae..21be23ec3c79 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_powertune.c +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_powertune.c @@ -973,7 +973,7 @@ int smu7_enable_didt_config(struct pp_hwmgr *hwmgr) PP_CAP(PHM_PlatformCaps_TDRamping) || PP_CAP(PHM_PlatformCaps_TCPRamping)) { - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); mutex_lock(&adev->grbm_idx_mutex); value = 0; value2 = cgs_read_register(hwmgr->device, mmGRBM_GFX_INDEX); @@ -1048,13 +1048,13 @@ int smu7_enable_didt_config(struct pp_hwmgr *hwmgr) } mutex_unlock(&adev->grbm_idx_mutex); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } return 0; error: mutex_unlock(&adev->grbm_idx_mutex); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return result; } @@ -1068,7 +1068,7 @@ int smu7_disable_didt_config(struct pp_hwmgr *hwmgr) PP_CAP(PHM_PlatformCaps_TDRamping) || PP_CAP(PHM_PlatformCaps_TCPRamping)) { - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); result = smu7_enable_didt(hwmgr, false); PP_ASSERT_WITH_CODE((result == 0), @@ -1081,12 +1081,12 @@ int smu7_disable_didt_config(struct pp_hwmgr *hwmgr) PP_ASSERT_WITH_CODE((0 == result), "Failed to disable DPM DIDT.", goto error); } - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } return 0; error: - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return result; } diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_powertune.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_powertune.c index 9757d47dd6b8..309a9d3bc1b7 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_powertune.c +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_powertune.c @@ -915,7 +915,7 @@ static int vega10_enable_cac_driving_se_didt_config(struct pp_hwmgr *hwmgr) num_se = adev->gfx.config.max_shader_engines; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); mutex_lock(&adev->grbm_idx_mutex); for (count = 0; count < num_se; count++) { @@ -940,7 +940,7 @@ static int vega10_enable_cac_driving_se_didt_config(struct pp_hwmgr *hwmgr) vega10_didt_set_mask(hwmgr, true); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return 0; } @@ -949,11 +949,11 @@ static int vega10_disable_cac_driving_se_didt_config(struct pp_hwmgr *hwmgr) { struct amdgpu_device *adev = hwmgr->adev; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); vega10_didt_set_mask(hwmgr, false); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return 0; } @@ -966,7 +966,7 @@ static int vega10_enable_psm_gc_didt_config(struct pp_hwmgr *hwmgr) num_se = adev->gfx.config.max_shader_engines; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); mutex_lock(&adev->grbm_idx_mutex); for (count = 0; count < num_se; count++) { @@ -985,7 +985,7 @@ static int vega10_enable_psm_gc_didt_config(struct pp_hwmgr *hwmgr) vega10_didt_set_mask(hwmgr, true); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); vega10_program_gc_didt_config_registers(hwmgr, GCDiDtDroopCtrlConfig_vega10); if (PP_CAP(PHM_PlatformCaps_GCEDC)) @@ -1002,11 +1002,11 @@ static int vega10_disable_psm_gc_didt_config(struct pp_hwmgr *hwmgr) struct amdgpu_device *adev = hwmgr->adev; uint32_t data; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); vega10_didt_set_mask(hwmgr, false); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); if (PP_CAP(PHM_PlatformCaps_GCEDC)) { data = 0x00000000; @@ -1027,7 +1027,7 @@ static int vega10_enable_se_edc_config(struct pp_hwmgr *hwmgr) num_se = adev->gfx.config.max_shader_engines; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); mutex_lock(&adev->grbm_idx_mutex); for (count = 0; count < num_se; count++) { @@ -1048,7 +1048,7 @@ static int vega10_enable_se_edc_config(struct pp_hwmgr *hwmgr) vega10_didt_set_mask(hwmgr, true); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return 0; } @@ -1057,11 +1057,11 @@ static int vega10_disable_se_edc_config(struct pp_hwmgr *hwmgr) { struct amdgpu_device *adev = hwmgr->adev; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); vega10_didt_set_mask(hwmgr, false); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return 0; } @@ -1075,7 +1075,7 @@ static int vega10_enable_psm_gc_edc_config(struct pp_hwmgr *hwmgr) num_se = adev->gfx.config.max_shader_engines; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); vega10_program_gc_didt_config_registers(hwmgr, AvfsPSMResetConfig_vega10); @@ -1096,7 +1096,7 @@ static int vega10_enable_psm_gc_edc_config(struct pp_hwmgr *hwmgr) vega10_didt_set_mask(hwmgr, true); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); vega10_program_gc_didt_config_registers(hwmgr, PSMGCEDCDroopCtrlConfig_vega10); @@ -1116,11 +1116,11 @@ static int vega10_disable_psm_gc_edc_config(struct pp_hwmgr *hwmgr) struct amdgpu_device *adev = hwmgr->adev; uint32_t data; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); vega10_didt_set_mask(hwmgr, false); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); if (PP_CAP(PHM_PlatformCaps_GCEDC)) { data = 0x00000000; @@ -1138,7 +1138,7 @@ static int vega10_enable_se_edc_force_stall_config(struct pp_hwmgr *hwmgr) struct amdgpu_device *adev = hwmgr->adev; int result; - amdgpu_gfx_rlc_enter_safe_mode(adev); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); mutex_lock(&adev->grbm_idx_mutex); WREG32_SOC15(GC, 0, mmGRBM_GFX_INDEX, 0xE0000000); @@ -1151,7 +1151,7 @@ static int vega10_enable_se_edc_force_stall_config(struct pp_hwmgr *hwmgr) vega10_didt_set_mask(hwmgr, false); - amdgpu_gfx_rlc_exit_safe_mode(adev); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return 0; } diff --git a/drivers/gpu/drm/amd/pm/powerplay/inc/hwmgr.h b/drivers/gpu/drm/amd/pm/powerplay/inc/hwmgr.h index 5ce433e2c16a..f1580a26a850 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/inc/hwmgr.h +++ b/drivers/gpu/drm/amd/pm/powerplay/inc/hwmgr.h @@ -359,7 +359,7 @@ struct pp_hwmgr_func { int (*set_ppfeature_status)(struct pp_hwmgr *hwmgr, uint64_t ppfeature_masks); int (*set_mp1_state)(struct pp_hwmgr *hwmgr, enum pp_mp1_state mp1_state); int (*asic_reset)(struct pp_hwmgr *hwmgr, enum SMU_ASIC_RESET_MODE mode); - int (*smu_i2c_bus_access)(struct pp_hwmgr *hwmgr, bool aquire); + int (*smu_i2c_bus_access)(struct pp_hwmgr *hwmgr, bool acquire); int (*set_df_cstate)(struct pp_hwmgr *hwmgr, enum pp_df_cstate state); int (*set_xgmi_pstate)(struct pp_hwmgr *hwmgr, uint32_t pstate); int (*disable_power_features_for_compute_performance)(struct pp_hwmgr *hwmgr, diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c index 5633c5797e85..4dea79a0c5b5 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c +++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c @@ -733,6 +733,24 @@ static int smu_late_init(void *handle) return ret; } + /* + * Explicitly notify PMFW the power mode the system in. Since + * the PMFW may boot the ASIC with a different mode. + * For those supporting ACDC switch via gpio, PMFW will + * handle the switch automatically. Driver involvement + * is unnecessary. + */ + if (!smu->dc_controlled_by_gpio) { + ret = smu_set_power_source(smu, + adev->pm.ac_power ? SMU_POWER_SOURCE_AC : + SMU_POWER_SOURCE_DC); + if (ret) { + dev_err(adev->dev, "Failed to switch to %s mode!\n", + adev->pm.ac_power ? "AC" : "DC"); + return ret; + } + } + if ((adev->ip_versions[MP1_HWIP][0] == IP_VERSION(13, 0, 1)) || (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(13, 0, 3))) return 0; @@ -804,11 +822,20 @@ static int smu_init_fb_allocations(struct smu_context *smu) } } + driver_table->domain = AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT; /* VRAM allocation for driver table */ for (i = 0; i < SMU_TABLE_COUNT; i++) { if (tables[i].size == 0) continue; + /* If one of the tables has VRAM domain restriction, keep it in + * VRAM + */ + if ((tables[i].domain & + (AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT)) == + AMDGPU_GEM_DOMAIN_VRAM) + driver_table->domain = AMDGPU_GEM_DOMAIN_VRAM; + if (i == SMU_TABLE_PMSTATUSLOG) continue; @@ -818,7 +845,6 @@ static int smu_init_fb_allocations(struct smu_context *smu) driver_table->size = max_table_size; driver_table->align = PAGE_SIZE; - driver_table->domain = AMDGPU_GEM_DOMAIN_VRAM; ret = amdgpu_bo_create_kernel(adev, driver_table->size, diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_aldebaran.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_aldebaran.h index 90200f31ff52..cddf45eebee8 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_aldebaran.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_aldebaran.h @@ -24,6 +24,8 @@ #ifndef SMU13_DRIVER_IF_ALDEBARAN_H #define SMU13_DRIVER_IF_ALDEBARAN_H +#define SMU13_DRIVER_IF_VERSION_ALDE 0x08 + #define NUM_VCLK_DPM_LEVELS 8 #define NUM_DCLK_DPM_LEVELS 8 #define NUM_SOCCLK_DPM_LEVELS 8 diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_0.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_0.h index b686fb68a6e7..9dd1ed5b8940 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_0.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_0.h @@ -24,8 +24,10 @@ #ifndef SMU13_DRIVER_IF_V13_0_0_H #define SMU13_DRIVER_IF_V13_0_0_H +#define SMU13_0_0_DRIVER_IF_VERSION 0x3D + //Increment this version if SkuTable_t or BoardTable_t change -#define PPTABLE_VERSION 0x26 +#define PPTABLE_VERSION 0x2B #define NUM_GFXCLK_DPM_LEVELS 16 #define NUM_SOCCLK_DPM_LEVELS 8 @@ -94,7 +96,7 @@ #define FEATURE_ATHUB_MMHUB_PG_BIT 48 #define FEATURE_SOC_PCC_BIT 49 #define FEATURE_EDC_PWRBRK_BIT 50 -#define FEATURE_SPARE_51_BIT 51 +#define FEATURE_BOMXCO_SVI3_PROG_BIT 51 #define FEATURE_SPARE_52_BIT 52 #define FEATURE_SPARE_53_BIT 53 #define FEATURE_SPARE_54_BIT 54 @@ -310,6 +312,7 @@ typedef enum { I2C_CONTROLLER_PROTOCOL_VR_IR35217, I2C_CONTROLLER_PROTOCOL_TMP_MAX31875, I2C_CONTROLLER_PROTOCOL_INA3221, + I2C_CONTROLLER_PROTOCOL_TMP_MAX6604, I2C_CONTROLLER_PROTOCOL_COUNT, } I2cControllerProtocol_e; @@ -568,6 +571,7 @@ typedef enum { } POWER_SOURCE_e; typedef enum { + MEM_VENDOR_PLACEHOLDER0, MEM_VENDOR_SAMSUNG, MEM_VENDOR_INFINEON, MEM_VENDOR_ELPIDA, @@ -577,7 +581,6 @@ typedef enum { MEM_VENDOR_MOSEL, MEM_VENDOR_WINBOND, MEM_VENDOR_ESMT, - MEM_VENDOR_PLACEHOLDER0, MEM_VENDOR_PLACEHOLDER1, MEM_VENDOR_PLACEHOLDER2, MEM_VENDOR_PLACEHOLDER3, @@ -665,7 +668,14 @@ typedef enum { #define PP_NUM_RTAVFS_PWL_ZONES 5 - +#define PP_OD_FEATURE_GFX_VF_CURVE_BIT 0 +#define PP_OD_FEATURE_PPT_BIT 2 +#define PP_OD_FEATURE_FAN_CURVE_BIT 3 +#define PP_OD_FEATURE_GFXCLK_BIT 7 +#define PP_OD_FEATURE_UCLK_BIT 8 +#define PP_OD_FEATURE_ZERO_FAN_BIT 9 +#define PP_OD_FEATURE_TEMPERATURE_BIT 10 +#define PP_OD_FEATURE_COUNT 13 // VBIOS or PPLIB configures telemetry slope and offset. Only slope expected to be set for SVI3 // Slope Q1.7, Offset Q1.2 @@ -687,10 +697,8 @@ typedef struct { //Voltage control int16_t VoltageOffsetPerZoneBoundary[PP_NUM_OD_VF_CURVE_POINTS]; - uint16_t VddGfxVmax; // in mV - uint8_t IdlePwrSavingFeaturesCtrl; - uint8_t RuntimePwrSavingFeaturesCtrl; + uint32_t Reserved; //Frequency changes int16_t GfxclkFmin; // MHz @@ -727,10 +735,9 @@ typedef struct { uint32_t FeatureCtrlMask; int16_t VoltageOffsetPerZoneBoundary; - uint16_t VddGfxVmax; // in mV + uint16_t Reserved1; - uint8_t IdlePwrSavingFeaturesCtrl; - uint8_t RuntimePwrSavingFeaturesCtrl; + uint16_t Reserved2; int16_t GfxclkFmin; // MHz int16_t GfxclkFmax; // MHz @@ -806,6 +813,9 @@ typedef enum { #define INVALID_BOARD_GPIO 0xFF +#define MARKETING_BASE_CLOCKS 0 +#define MARKETING_GAME_CLOCKS 1 +#define MARKETING_BOOST_CLOCKS 2 typedef struct { //PLL 0 @@ -1096,10 +1106,15 @@ typedef struct { uint16_t DcsExitHysteresis; //The min amount of time power credit accumulator should have a value > 0 before SMU exits the DCS throttling phase. uint16_t DcsTimeout; //This is the amount of time SMU FW waits for RLC to put GFX into GFXOFF before reverting to the fallback mechanism of throttling GFXCLK to Fmin. + uint8_t FoptEnabled; + uint8_t DcsSpare2[3]; + uint32_t DcsFoptM; //Tuning paramters to shift Fopt calculation + uint32_t DcsFoptB; //Tuning paramters to shift Fopt calculation - uint32_t DcsSpare[16]; + uint32_t DcsSpare[11]; // UCLK section + uint16_t ShadowFreqTableUclk[NUM_UCLK_DPM_LEVELS]; // In MHz uint8_t UseStrobeModeOptimizations; //Set to indicate that FW should use strobe mode optimizations uint8_t PaddingMem[3]; @@ -1245,8 +1260,13 @@ typedef struct { QuadraticInt_t qFeffCoeffBaseClock[POWER_SOURCE_COUNT]; QuadraticInt_t qFeffCoeffBoostClock[POWER_SOURCE_COUNT]; + uint16_t TemperatureLimit_Hynix; // In degrees Celsius. Memory temperature limit associated with Hynix + uint16_t TemperatureLimit_Micron; // In degrees Celsius. Memory temperature limit associated with Micron + uint16_t TemperatureFwCtfLimit_Hynix; + uint16_t TemperatureFwCtfLimit_Micron; + // SECTION: Sku Reserved - uint32_t Spare[43]; + uint32_t Spare[41]; // Padding for MMHUB - do not modify this uint32_t MmHubPadding[8]; @@ -1318,8 +1338,9 @@ typedef struct { // UCLK Spread Spectrum uint8_t UclkSpreadPercent[MEM_VENDOR_COUNT]; + uint8_t GfxclkSpreadEnable; + // FCLK Spread Spectrum - uint8_t FclkSpreadPadding; uint8_t FclkSpreadPercent; // Q4.4 uint16_t FclkSpreadFreq; // kHz @@ -1444,6 +1465,8 @@ typedef struct { uint8_t ThrottlingPercentage[THROTTLER_COUNT]; + uint8_t VmaxThrottlingPercentage; + uint8_t Padding1[3]; //metrics for D3hot entry/exit and driver ARM msgs uint32_t D3HotEntryCountPerMode[D3HOT_SEQUENCE_COUNT]; @@ -1463,7 +1486,7 @@ typedef struct { typedef struct { SmuMetrics_t SmuMetrics; - uint32_t Spare[30]; + uint32_t Spare[29]; // Padding - ignore uint32_t MmHubPadding[8]; // SMU internal use diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_4.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_4.h index 2162ecd1057d..fee9293b3f97 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_4.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_4.h @@ -27,7 +27,7 @@ // *** IMPORTANT *** // SMU TEAM: Always increment the interface version if // any structure is changed in this file -#define PMFW_DRIVER_IF_VERSION 8 +#define SMU13_0_4_DRIVER_IF_VERSION 8 typedef struct { int32_t value; diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_5.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_5.h index aa971412b434..7589faa0232d 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_5.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_5.h @@ -23,7 +23,7 @@ #ifndef __SMU13_DRIVER_IF_V13_0_5_H__ #define __SMU13_DRIVER_IF_V13_0_5_H__ -#define PMFW_DRIVER_IF_VERSION 4 +#define SMU13_0_5_DRIVER_IF_VERSION 4 // Throttler Status Bitmask #define THROTTLER_STATUS_BIT_SPL 0 diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_6.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_6.h index be596777cd2c..ca4a5e99ccd1 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_6.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_6.h @@ -26,7 +26,7 @@ // *** IMPORTANT *** // PMFW TEAM: Always increment the interface version if // anything is changed in this file -#define SMU13_0_6_DRIVER_IF_VERSION 0x08042022 +#define SMU13_0_6_DRIVER_IF_VERSION 0x08042024 //I2C Interface #define NUM_I2C_CONTROLLERS 8 @@ -106,7 +106,7 @@ typedef enum { } UCLK_DPM_MODE_e; typedef struct { - //0-26 SOC, 27-29 SOCIO + //0-23 SOC, 24-26 SOCIO, 27-29 SOC uint16_t avgPsmCount[30]; uint16_t minPsmCount[30]; float avgPsmVoltage[30]; @@ -121,6 +121,17 @@ typedef struct { float minPsmVoltage[30]; } AvfsDebugTableXcd_t; +// Defines used for IH-based thermal interrupts to GFX driver - A/X only +#define IH_INTERRUPT_ID_TO_DRIVER 0xFE +#define IH_INTERRUPT_CONTEXT_ID_THERMAL_THROTTLING 0x7 + +//thermal over-temp mask defines for IH interrupt to host +#define THROTTLER_PROCHOT_BIT 0 +#define THROTTLER_PPT_BIT 1 +#define THROTTLER_THERMAL_SOCKET_BIT 2//AID, XCD, CCD throttling +#define THROTTLER_THERMAL_VR_BIT 3//VRHOT +#define THROTTLER_THERMAL_HBM_BIT 4 + // These defines are used with the following messages: // SMC_MSG_TransferTableDram2Smu // SMC_MSG_TransferTableSmu2Dram diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_7.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_7.h index 4c46a0392451..62b7c0daff68 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_7.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_7.h @@ -25,7 +25,7 @@ // *** IMPORTANT *** // PMFW TEAM: Always increment the interface version on any change to this file -#define SMU13_DRIVER_IF_VERSION 0x35 +#define SMU13_0_7_DRIVER_IF_VERSION 0x35 //Increment this version if SkuTable_t or BoardTable_t change #define PPTABLE_VERSION 0x27 @@ -683,18 +683,12 @@ typedef struct { #define PP_OD_FEATURE_GFX_VF_CURVE_BIT 0 -#define PP_OD_FEATURE_VMAX_BIT 1 #define PP_OD_FEATURE_PPT_BIT 2 #define PP_OD_FEATURE_FAN_CURVE_BIT 3 -#define PP_OD_FEATURE_FREQ_DETER_BIT 4 -#define PP_OD_FEATURE_FULL_CTRL_BIT 5 -#define PP_OD_FEATURE_TDC_BIT 6 #define PP_OD_FEATURE_GFXCLK_BIT 7 #define PP_OD_FEATURE_UCLK_BIT 8 #define PP_OD_FEATURE_ZERO_FAN_BIT 9 #define PP_OD_FEATURE_TEMPERATURE_BIT 10 -#define PP_OD_FEATURE_POWER_FEATURE_CTRL_BIT 11 -#define PP_OD_FEATURE_ASIC_TDC_BIT 12 #define PP_OD_FEATURE_COUNT 13 typedef enum { @@ -713,10 +707,8 @@ typedef struct { //Voltage control int16_t VoltageOffsetPerZoneBoundary[PP_NUM_OD_VF_CURVE_POINTS]; - uint16_t VddGfxVmax; // in mV - uint8_t IdlePwrSavingFeaturesCtrl; - uint8_t RuntimePwrSavingFeaturesCtrl; + uint32_t Reserved; //Frequency changes int16_t GfxclkFmin; // MHz @@ -741,12 +733,7 @@ typedef struct { uint8_t MaxOpTemp; uint8_t Padding[4]; - uint16_t GfxVoltageFullCtrlMode; - uint16_t GfxclkFullCtrlMode; - uint16_t UclkFullCtrlMode; - int16_t AsicTdc; - - uint32_t Spare[10]; + uint32_t Spare[12]; uint32_t MmHubPadding[8]; // SMU internal use. Adding here instead of external as a workaround } OverDriveTable_t; @@ -759,10 +746,9 @@ typedef struct { uint32_t FeatureCtrlMask; int16_t VoltageOffsetPerZoneBoundary; - uint16_t VddGfxVmax; // in mV + uint16_t Reserved1; - uint8_t IdlePwrSavingFeaturesCtrl; - uint8_t RuntimePwrSavingFeaturesCtrl; + uint16_t Reserved2; int16_t GfxclkFmin; // MHz int16_t GfxclkFmax; // MHz @@ -785,12 +771,7 @@ typedef struct { uint8_t MaxOpTemp; uint8_t Padding[4]; - uint16_t GfxVoltageFullCtrlMode; - uint16_t GfxclkFullCtrlMode; - uint16_t UclkFullCtrlMode; - int16_t AsicTdc; - - uint32_t Spare[10]; + uint32_t Spare[12]; } OverDriveLimits_t; diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_yellow_carp.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_yellow_carp.h index 25540cb28208..7417634827ad 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_yellow_carp.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_yellow_carp.h @@ -26,7 +26,7 @@ // *** IMPORTANT *** // SMU TEAM: Always increment the interface version if // any structure is changed in this file -#define SMU13_DRIVER_IF_VERSION 4 +#define SMU13_YELLOW_CARP_DRIVER_IF_VERSION 4 typedef struct { int32_t value; diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_pmfw.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_pmfw.h index bdccbb4a6276..252aef190c5c 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_pmfw.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_pmfw.h @@ -123,9 +123,9 @@ typedef enum { VOLTAGE_GUARDBAND_COUNT } GFX_GUARDBAND_e; -#define SMU_METRICS_TABLE_VERSION 0x1 +#define SMU_METRICS_TABLE_VERSION 0x5 -typedef struct { +typedef struct __attribute__((packed, aligned(4))) { uint32_t AccumulationCounter; //TEMPERATURE @@ -198,11 +198,20 @@ typedef struct { uint32_t SocketThmResidencyAcc; uint32_t VrThmResidencyAcc; uint32_t HbmThmResidencyAcc; + uint32_t spare; + + // New Items at end to maintain driver compatibility + uint32_t GfxclkFrequency[8]; + + //PSNs + uint64_t PublicSerialNumber_AID[4]; + uint64_t PublicSerialNumber_XCD[8]; + uint64_t PublicSerialNumber_CCD[12]; } MetricsTable_t; -#define SMU_VF_METRICS_TABLE_VERSION 0x1 +#define SMU_VF_METRICS_TABLE_VERSION 0x3 -typedef struct { +typedef struct __attribute__((packed, aligned(4))) { uint32_t AccumulationCounter; uint32_t InstGfxclk_TargFreq; uint64_t AccGfxclk_TargFreq; diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_ppsmc.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_ppsmc.h index b838e8db395a..ae4f44c4b877 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_ppsmc.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_ppsmc.h @@ -82,7 +82,8 @@ #define PPSMC_MSG_SetSoftMaxGfxClk 0x31 #define PPSMC_MSG_GetMinGfxDpmFreq 0x32 #define PPSMC_MSG_GetMaxGfxDpmFreq 0x33 -#define PPSMC_Message_Count 0x34 +#define PPSMC_MSG_PrepareForDriverUnload 0x34 +#define PPSMC_Message_Count 0x35 //PPSMC Reset Types for driver msg argument #define PPSMC_RESET_TYPE_DRIVER_MODE_1_RESET 0x1 diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h index df3baaab0037..6a0ac0bbaace 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h @@ -25,17 +25,6 @@ #include "amdgpu_smu.h" -#define SMU13_DRIVER_IF_VERSION_INV 0xFFFFFFFF -#define SMU13_DRIVER_IF_VERSION_YELLOW_CARP 0x04 -#define SMU13_DRIVER_IF_VERSION_ALDE 0x08 -#define SMU13_DRIVER_IF_VERSION_SMU_V13_0_0_0 0x37 -#define SMU13_DRIVER_IF_VERSION_SMU_V13_0_4 0x08 -#define SMU13_DRIVER_IF_VERSION_SMU_V13_0_5 0x04 -#define SMU13_DRIVER_IF_VERSION_SMU_V13_0_0_10 0x32 -#define SMU13_DRIVER_IF_VERSION_SMU_V13_0_7 0x37 -#define SMU13_DRIVER_IF_VERSION_SMU_V13_0_10 0x1D -#define SMU13_DRIVER_IF_VERSION_SMU_V13_0_6 0x0 - #define SMU13_MODE1_RESET_WAIT_TIME_IN_MS 500 //500ms /* MP Apertures */ @@ -62,6 +51,8 @@ #define CTF_OFFSET_HOTSPOT 5 #define CTF_OFFSET_MEM 5 +#define SMU_13_VCLK_SHIFT 16 + extern const int pmfw_decoded_link_speed[5]; extern const int pmfw_decoded_link_width[7]; @@ -130,6 +121,7 @@ struct smu_13_0_power_context { uint32_t power_source; uint8_t in_power_limit_boost_mode; enum smu_13_0_power_state power_state; + atomic_t throttle_status; }; #if defined(SWSMU_CODE_LAYER_L2) || defined(SWSMU_CODE_LAYER_L3) diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0_7_pptable.h b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0_7_pptable.h index 478862ded0bd..eadbe0149cae 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0_7_pptable.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0_7_pptable.h @@ -38,13 +38,12 @@ #define SMU_13_0_7_PP_THERMALCONTROLLER_NONE 0 #define SMU_13_0_7_PP_THERMALCONTROLLER_NAVI21 28 -#define SMU_13_0_7_PP_OVERDRIVE_VERSION 0x81 // OverDrive 8 Table Version 0.2 +#define SMU_13_0_7_PP_OVERDRIVE_VERSION 0x83 // OverDrive 8 Table Version 0.2 #define SMU_13_0_7_PP_POWERSAVINGCLOCK_VERSION 0x01 // Power Saving Clock Table Version 1.00 enum SMU_13_0_7_ODFEATURE_CAP { SMU_13_0_7_ODCAP_GFXCLK_LIMITS = 0, - SMU_13_0_7_ODCAP_GFXCLK_CURVE, SMU_13_0_7_ODCAP_UCLK_LIMITS, SMU_13_0_7_ODCAP_POWER_LIMIT, SMU_13_0_7_ODCAP_FAN_ACOUSTIC_LIMIT, @@ -59,13 +58,13 @@ enum SMU_13_0_7_ODFEATURE_CAP SMU_13_0_7_ODCAP_FAN_CURVE, SMU_13_0_7_ODCAP_AUTO_FAN_ACOUSTIC_LIMIT, SMU_13_0_7_ODCAP_POWER_MODE, + SMU_13_0_7_ODCAP_PER_ZONE_GFX_VOLTAGE_OFFSET, SMU_13_0_7_ODCAP_COUNT, }; enum SMU_13_0_7_ODFEATURE_ID { SMU_13_0_7_ODFEATURE_GFXCLK_LIMITS = 1 << SMU_13_0_7_ODCAP_GFXCLK_LIMITS, //GFXCLK Limit feature - SMU_13_0_7_ODFEATURE_GFXCLK_CURVE = 1 << SMU_13_0_7_ODCAP_GFXCLK_CURVE, //GFXCLK Curve feature SMU_13_0_7_ODFEATURE_UCLK_LIMITS = 1 << SMU_13_0_7_ODCAP_UCLK_LIMITS, //UCLK Limit feature SMU_13_0_7_ODFEATURE_POWER_LIMIT = 1 << SMU_13_0_7_ODCAP_POWER_LIMIT, //Power Limit feature SMU_13_0_7_ODFEATURE_FAN_ACOUSTIC_LIMIT = 1 << SMU_13_0_7_ODCAP_FAN_ACOUSTIC_LIMIT, //Fan Acoustic RPM feature @@ -80,6 +79,7 @@ enum SMU_13_0_7_ODFEATURE_ID SMU_13_0_7_ODFEATURE_FAN_CURVE = 1 << SMU_13_0_7_ODCAP_FAN_CURVE, //Fan Curve feature SMU_13_0_7_ODFEATURE_AUTO_FAN_ACOUSTIC_LIMIT = 1 << SMU_13_0_7_ODCAP_AUTO_FAN_ACOUSTIC_LIMIT, //Auto Fan Acoustic RPM feature SMU_13_0_7_ODFEATURE_POWER_MODE = 1 << SMU_13_0_7_ODCAP_POWER_MODE, //Optimized GPU Power Mode feature + SMU_13_0_7_ODFEATURE_PER_ZONE_GFX_VOLTAGE_OFFSET = 1 << SMU_13_0_7_ODCAP_PER_ZONE_GFX_VOLTAGE_OFFSET, //Perzone voltage offset feature SMU_13_0_7_ODFEATURE_COUNT = 16, }; @@ -89,10 +89,6 @@ enum SMU_13_0_7_ODSETTING_ID { SMU_13_0_7_ODSETTING_GFXCLKFMAX = 0, SMU_13_0_7_ODSETTING_GFXCLKFMIN, - SMU_13_0_7_ODSETTING_CUSTOM_GFX_VF_CURVE_A, - SMU_13_0_7_ODSETTING_CUSTOM_GFX_VF_CURVE_B, - SMU_13_0_7_ODSETTING_CUSTOM_GFX_VF_CURVE_C, - SMU_13_0_7_ODSETTING_CUSTOM_CURVE_VFT_FMIN, SMU_13_0_7_ODSETTING_UCLKFMIN, SMU_13_0_7_ODSETTING_UCLKFMAX, SMU_13_0_7_ODSETTING_POWERPERCENTAGE, @@ -117,6 +113,12 @@ enum SMU_13_0_7_ODSETTING_ID SMU_13_0_7_ODSETTING_FAN_CURVE_SPEED_5, SMU_13_0_7_ODSETTING_AUTO_FAN_ACOUSTIC_LIMIT, SMU_13_0_7_ODSETTING_POWER_MODE, + SMU_13_0_7_ODSETTING_PER_ZONE_GFX_VOLTAGE_OFFSET_POINT_1, + SMU_13_0_7_ODSETTING_PER_ZONE_GFX_VOLTAGE_OFFSET_POINT_2, + SMU_13_0_7_ODSETTING_PER_ZONE_GFX_VOLTAGE_OFFSET_POINT_3, + SMU_13_0_7_ODSETTING_PER_ZONE_GFX_VOLTAGE_OFFSET_POINT_4, + SMU_13_0_7_ODSETTING_PER_ZONE_GFX_VOLTAGE_OFFSET_POINT_5, + SMU_13_0_7_ODSETTING_PER_ZONE_GFX_VOLTAGE_OFFSET_POINT_6, SMU_13_0_7_ODSETTING_COUNT, }; #define SMU_13_0_7_MAX_ODSETTING 64 //Maximum Number of ODSettings diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c index c4000518dc56..275f708db636 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c @@ -3413,26 +3413,8 @@ static int navi10_post_smu_init(struct smu_context *smu) return 0; ret = navi10_run_umc_cdr_workaround(smu); - if (ret) { + if (ret) dev_err(adev->dev, "Failed to apply umc cdr workaround!\n"); - return ret; - } - - if (!smu->dc_controlled_by_gpio) { - /* - * For Navi1X, manually switch it to AC mode as PMFW - * may boot it with DC mode. - */ - ret = smu_v11_0_set_power_source(smu, - adev->pm.ac_power ? - SMU_POWER_SOURCE_AC : - SMU_POWER_SOURCE_DC); - if (ret) { - dev_err(adev->dev, "Failed to switch to %s mode!\n", - adev->pm.ac_power ? "AC" : "DC"); - return ret; - } - } return ret; } diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c index 75f18681e984..f7ed3e655e39 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c @@ -431,7 +431,13 @@ static int sienna_cichlid_append_powerplay_table(struct smu_context *smu) { struct atom_smc_dpm_info_v4_9 *smc_dpm_table; int index, ret; - I2cControllerConfig_t *table_member; + PPTable_beige_goby_t *ppt_beige_goby; + PPTable_t *ppt; + + if (smu->adev->ip_versions[MP1_HWIP][0] == IP_VERSION(11, 0, 13)) + ppt_beige_goby = smu->smu_table.driver_pptable; + else + ppt = smu->smu_table.driver_pptable; index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1, smc_dpm_info); @@ -440,9 +446,13 @@ static int sienna_cichlid_append_powerplay_table(struct smu_context *smu) (uint8_t **)&smc_dpm_table); if (ret) return ret; - GET_PPTABLE_MEMBER(I2cControllers, &table_member); - memcpy(table_member, smc_dpm_table->I2cControllers, - sizeof(*smc_dpm_table) - sizeof(smc_dpm_table->table_header)); + + if (smu->adev->ip_versions[MP1_HWIP][0] == IP_VERSION(11, 0, 13)) + smu_memcpy_trailing(ppt_beige_goby, I2cControllers, BoardReserved, + smc_dpm_table, I2cControllers); + else + smu_memcpy_trailing(ppt, I2cControllers, BoardReserved, + smc_dpm_table, I2cControllers); return 0; } @@ -2067,33 +2077,94 @@ static int sienna_cichlid_display_disable_memory_clock_switch(struct smu_context return ret; } +static void sienna_cichlid_get_override_pcie_settings(struct smu_context *smu, + uint32_t *gen_speed_override, + uint32_t *lane_width_override) +{ + struct amdgpu_device *adev = smu->adev; + + *gen_speed_override = 0xff; + *lane_width_override = 0xff; + + switch (adev->pdev->device) { + case 0x73A0: + case 0x73A1: + case 0x73A2: + case 0x73A3: + case 0x73AB: + case 0x73AE: + /* Bit 7:0: PCIE lane width, 1 to 7 corresponds is x1 to x32 */ + *lane_width_override = 6; + break; + case 0x73E0: + case 0x73E1: + case 0x73E3: + *lane_width_override = 4; + break; + case 0x7420: + case 0x7421: + case 0x7422: + case 0x7423: + case 0x7424: + *lane_width_override = 3; + break; + default: + break; + } +} + +#define MAX(a, b) ((a) > (b) ? (a) : (b)) + static int sienna_cichlid_update_pcie_parameters(struct smu_context *smu, uint32_t pcie_gen_cap, uint32_t pcie_width_cap) { struct smu_11_0_dpm_context *dpm_context = smu->smu_dpm.dpm_context; - - uint32_t smu_pcie_arg; + struct smu_11_0_pcie_table *pcie_table = &dpm_context->dpm_tables.pcie_table; + uint32_t gen_speed_override, lane_width_override; uint8_t *table_member1, *table_member2; + uint32_t min_gen_speed, max_gen_speed; + uint32_t min_lane_width, max_lane_width; + uint32_t smu_pcie_arg; int ret, i; GET_PPTABLE_MEMBER(PcieGenSpeed, &table_member1); GET_PPTABLE_MEMBER(PcieLaneCount, &table_member2); - /* lclk dpm table setup */ - for (i = 0; i < MAX_PCIE_CONF; i++) { - dpm_context->dpm_tables.pcie_table.pcie_gen[i] = table_member1[i]; - dpm_context->dpm_tables.pcie_table.pcie_lane[i] = table_member2[i]; + sienna_cichlid_get_override_pcie_settings(smu, + &gen_speed_override, + &lane_width_override); + + /* PCIE gen speed override */ + if (gen_speed_override != 0xff) { + min_gen_speed = MIN(pcie_gen_cap, gen_speed_override); + max_gen_speed = MIN(pcie_gen_cap, gen_speed_override); + } else { + min_gen_speed = MAX(0, table_member1[0]); + max_gen_speed = MIN(pcie_gen_cap, table_member1[1]); + min_gen_speed = min_gen_speed > max_gen_speed ? + max_gen_speed : min_gen_speed; + } + pcie_table->pcie_gen[0] = min_gen_speed; + pcie_table->pcie_gen[1] = max_gen_speed; + + /* PCIE lane width override */ + if (lane_width_override != 0xff) { + min_lane_width = MIN(pcie_width_cap, lane_width_override); + max_lane_width = MIN(pcie_width_cap, lane_width_override); + } else { + min_lane_width = MAX(1, table_member2[0]); + max_lane_width = MIN(pcie_width_cap, table_member2[1]); + min_lane_width = min_lane_width > max_lane_width ? + max_lane_width : min_lane_width; } + pcie_table->pcie_lane[0] = min_lane_width; + pcie_table->pcie_lane[1] = max_lane_width; for (i = 0; i < NUM_LINK_LEVELS; i++) { - smu_pcie_arg = (i << 16) | - ((table_member1[i] <= pcie_gen_cap) ? - (table_member1[i] << 8) : - (pcie_gen_cap << 8)) | - ((table_member2[i] <= pcie_width_cap) ? - table_member2[i] : - pcie_width_cap); + smu_pcie_arg = (i << 16 | + pcie_table->pcie_gen[i] << 8 | + pcie_table->pcie_lane[i]); ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_OverridePcieParameters, @@ -2101,11 +2172,6 @@ static int sienna_cichlid_update_pcie_parameters(struct smu_context *smu, NULL); if (ret) return ret; - - if (table_member1[i] > pcie_gen_cap) - dpm_context->dpm_tables.pcie_table.pcie_gen[i] = pcie_gen_cap; - if (table_member2[i] > pcie_width_cap) - dpm_context->dpm_tables.pcie_table.pcie_lane[i] = pcie_width_cap; } return 0; diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c index 7433dcaa16e0..067b4e0b026c 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c @@ -582,7 +582,7 @@ static int vangogh_print_legacy_clk_levels(struct smu_context *smu, DpmClocks_t *clk_table = smu->smu_table.clocks_table; SmuMetrics_legacy_t metrics; struct smu_dpm_context *smu_dpm_ctx = &(smu->smu_dpm); - int i, size = 0, ret = 0; + int i, idx, size = 0, ret = 0; uint32_t cur_value = 0, value = 0, count = 0; bool cur_value_match_level = false; @@ -656,7 +656,8 @@ static int vangogh_print_legacy_clk_levels(struct smu_context *smu, case SMU_MCLK: case SMU_FCLK: for (i = 0; i < count; i++) { - ret = vangogh_get_dpm_clk_limited(smu, clk_type, i, &value); + idx = (clk_type == SMU_FCLK || clk_type == SMU_MCLK) ? (count - i - 1) : i; + ret = vangogh_get_dpm_clk_limited(smu, clk_type, idx, &value); if (ret) return ret; if (!value) @@ -683,7 +684,7 @@ static int vangogh_print_clk_levels(struct smu_context *smu, DpmClocks_t *clk_table = smu->smu_table.clocks_table; SmuMetrics_t metrics; struct smu_dpm_context *smu_dpm_ctx = &(smu->smu_dpm); - int i, size = 0, ret = 0; + int i, idx, size = 0, ret = 0; uint32_t cur_value = 0, value = 0, count = 0; bool cur_value_match_level = false; uint32_t min, max; @@ -765,7 +766,8 @@ static int vangogh_print_clk_levels(struct smu_context *smu, case SMU_MCLK: case SMU_FCLK: for (i = 0; i < count; i++) { - ret = vangogh_get_dpm_clk_limited(smu, clk_type, i, &value); + idx = (clk_type == SMU_FCLK || clk_type == SMU_MCLK) ? (count - i - 1) : i; + ret = vangogh_get_dpm_clk_limited(smu, clk_type, idx, &value); if (ret) return ret; if (!value) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c index 5cdc07165480..8a8ba25c9ad7 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c @@ -494,7 +494,7 @@ static int renoir_set_fine_grain_gfx_freq_parameters(struct smu_context *smu) static int renoir_print_clk_levels(struct smu_context *smu, enum smu_clk_type clk_type, char *buf) { - int i, size = 0, ret = 0; + int i, idx, size = 0, ret = 0; uint32_t cur_value = 0, value = 0, count = 0, min = 0, max = 0; SmuMetrics_t metrics; struct smu_dpm_context *smu_dpm_ctx = &(smu->smu_dpm); @@ -594,7 +594,8 @@ static int renoir_print_clk_levels(struct smu_context *smu, case SMU_VCLK: case SMU_DCLK: for (i = 0; i < count; i++) { - ret = renoir_get_dpm_clk_limited(smu, clk_type, i, &value); + idx = (clk_type == SMU_FCLK || clk_type == SMU_MCLK) ? (count - i - 1) : i; + ret = renoir_get_dpm_clk_limited(smu, clk_type, idx, &value); if (ret) return ret; if (!value) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c index d30ec3005ea1..e80f122d8aec 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c @@ -2147,5 +2147,6 @@ void aldebaran_set_ppt_funcs(struct smu_context *smu) smu->clock_map = aldebaran_clk_map; smu->feature_map = aldebaran_feature_mask_map; smu->table_map = aldebaran_table_map; + smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_ALDE; smu_v13_0_set_smu_mailbox_registers(smu); } diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c index 393c6a7b9609..e52c563f0dac 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c @@ -269,45 +269,10 @@ int smu_v13_0_check_fw_version(struct smu_context *smu) smu_major = (smu_version >> 16) & 0xff; smu_minor = (smu_version >> 8) & 0xff; smu_debug = (smu_version >> 0) & 0xff; - if (smu->is_apu) + if (smu->is_apu || + adev->ip_versions[MP1_HWIP][0] == IP_VERSION(13, 0, 6)) adev->pm.fw_version = smu_version; - switch (adev->ip_versions[MP1_HWIP][0]) { - case IP_VERSION(13, 0, 2): - smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_ALDE; - break; - case IP_VERSION(13, 0, 0): - smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_SMU_V13_0_0_0; - break; - case IP_VERSION(13, 0, 10): - smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_SMU_V13_0_0_10; - break; - case IP_VERSION(13, 0, 7): - smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_SMU_V13_0_7; - break; - case IP_VERSION(13, 0, 1): - case IP_VERSION(13, 0, 3): - case IP_VERSION(13, 0, 8): - smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_YELLOW_CARP; - break; - case IP_VERSION(13, 0, 4): - case IP_VERSION(13, 0, 11): - smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_SMU_V13_0_4; - break; - case IP_VERSION(13, 0, 5): - smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_SMU_V13_0_5; - break; - case IP_VERSION(13, 0, 6): - smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_SMU_V13_0_6; - adev->pm.fw_version = smu_version; - break; - default: - dev_err(adev->dev, "smu unsupported IP version: 0x%x.\n", - adev->ip_versions[MP1_HWIP][0]); - smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_INV; - break; - } - /* only for dGPU w/ SMU13*/ if (adev->pm.fw) dev_dbg(smu->adev->dev, "smu fw reported program %d, version = 0x%08x (%d.%d.%d)\n", @@ -502,17 +467,26 @@ int smu_v13_0_init_smc_tables(struct smu_context *smu) ret = -ENOMEM; goto err3_out; } + + smu_table->user_overdrive_table = + kzalloc(tables[SMU_TABLE_OVERDRIVE].size, GFP_KERNEL); + if (!smu_table->user_overdrive_table) { + ret = -ENOMEM; + goto err4_out; + } } smu_table->combo_pptable = kzalloc(tables[SMU_TABLE_COMBO_PPTABLE].size, GFP_KERNEL); if (!smu_table->combo_pptable) { ret = -ENOMEM; - goto err4_out; + goto err5_out; } return 0; +err5_out: + kfree(smu_table->user_overdrive_table); err4_out: kfree(smu_table->boot_overdrive_table); err3_out: @@ -532,12 +506,14 @@ int smu_v13_0_fini_smc_tables(struct smu_context *smu) kfree(smu_table->gpu_metrics_table); kfree(smu_table->combo_pptable); + kfree(smu_table->user_overdrive_table); kfree(smu_table->boot_overdrive_table); kfree(smu_table->overdrive_table); kfree(smu_table->max_sustainable_clocks); kfree(smu_table->driver_pptable); smu_table->gpu_metrics_table = NULL; smu_table->combo_pptable = NULL; + smu_table->user_overdrive_table = NULL; smu_table->boot_overdrive_table = NULL; smu_table->overdrive_table = NULL; smu_table->max_sustainable_clocks = NULL; @@ -573,11 +549,11 @@ int smu_v13_0_init_power(struct smu_context *smu) if (smu_power->power_context || smu_power->power_context_size != 0) return -EINVAL; - smu_power->power_context = kzalloc(sizeof(struct smu_13_0_dpm_context), + smu_power->power_context = kzalloc(sizeof(struct smu_13_0_power_context), GFP_KERNEL); if (!smu_power->power_context) return -ENOMEM; - smu_power->power_context_size = sizeof(struct smu_13_0_dpm_context); + smu_power->power_context_size = sizeof(struct smu_13_0_power_context); return 0; } diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c index 09405ef1e3c8..a6083957ae51 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c @@ -46,7 +46,6 @@ #include "asic_reg/mp/mp_13_0_0_sh_mask.h" #include "smu_cmn.h" #include "amdgpu_ras.h" -#include "umc_v8_10.h" /* * DO NOT use these for err/warn/info/debug messages. @@ -237,6 +236,7 @@ static struct cmn2asic_mapping smu_v13_0_0_table_map[SMU_TABLE_COUNT] = { [SMU_TABLE_COMBO_PPTABLE] = {1, TABLE_COMBO_PPTABLE}, TAB_MAP(I2C_COMMANDS), TAB_MAP(ECCINFO), + TAB_MAP(OVERDRIVE), }; static struct cmn2asic_mapping smu_v13_0_0_pwr_src_map[SMU_POWER_SOURCE_COUNT] = { @@ -331,6 +331,11 @@ static int smu_v13_0_0_check_powerplay_table(struct smu_context *smu) struct smu_13_0_0_powerplay_table *powerplay_table = table_context->power_play_table; struct smu_baco_context *smu_baco = &smu->smu_baco; + PPTable_t *pptable = smu->smu_table.driver_pptable; + const OverDriveLimits_t * const overdrive_upperlimits = + &pptable->SkuTable.OverDriveLimitsBasicMax; + const OverDriveLimits_t * const overdrive_lowerlimits = + &pptable->SkuTable.OverDriveLimitsMin; if (powerplay_table->platform_caps & SMU_13_0_0_PP_PLATFORM_CAP_HARDWAREDC) smu->dc_controlled_by_gpio = true; @@ -342,6 +347,10 @@ static int smu_v13_0_0_check_powerplay_table(struct smu_context *smu) if (powerplay_table->platform_caps & SMU_13_0_0_PP_PLATFORM_CAP_MACO) smu_baco->maco_support = true; + if (!overdrive_lowerlimits->FeatureCtrlMask || + !overdrive_upperlimits->FeatureCtrlMask) + smu->od_enabled = false; + table_context->thermal_controller_type = powerplay_table->thermal_controller_type; @@ -461,7 +470,7 @@ static int smu_v13_0_0_tables_init(struct smu_context *smu) PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM); SMU_TABLE_INIT(tables, SMU_TABLE_I2C_COMMANDS, sizeof(SwI2cRequest_t), PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM); - SMU_TABLE_INIT(tables, SMU_TABLE_OVERDRIVE, sizeof(OverDriveTable_t), + SMU_TABLE_INIT(tables, SMU_TABLE_OVERDRIVE, sizeof(OverDriveTableExternal_t), PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM); SMU_TABLE_INIT(tables, SMU_TABLE_PMSTATUSLOG, SMU13_TOOL_SIZE, PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM); @@ -1022,17 +1031,119 @@ static int smu_v13_0_0_get_current_clk_freq_by_table(struct smu_context *smu, value); } +static bool smu_v13_0_0_is_od_feature_supported(struct smu_context *smu, + int od_feature_bit) +{ + PPTable_t *pptable = smu->smu_table.driver_pptable; + const OverDriveLimits_t * const overdrive_upperlimits = + &pptable->SkuTable.OverDriveLimitsBasicMax; + + return overdrive_upperlimits->FeatureCtrlMask & (1U << od_feature_bit); +} + +static void smu_v13_0_0_get_od_setting_limits(struct smu_context *smu, + int od_feature_bit, + bool lower_boundary, + int32_t *min, + int32_t *max) +{ + PPTable_t *pptable = smu->smu_table.driver_pptable; + const OverDriveLimits_t * const overdrive_upperlimits = + &pptable->SkuTable.OverDriveLimitsBasicMax; + const OverDriveLimits_t * const overdrive_lowerlimits = + &pptable->SkuTable.OverDriveLimitsMin; + int32_t od_min_setting, od_max_setting; + + switch (od_feature_bit) { + case PP_OD_FEATURE_GFXCLK_BIT: + if (lower_boundary) { + od_min_setting = overdrive_lowerlimits->GfxclkFmin; + od_max_setting = overdrive_upperlimits->GfxclkFmin; + } else { + od_min_setting = overdrive_lowerlimits->GfxclkFmax; + od_max_setting = overdrive_upperlimits->GfxclkFmax; + } + break; + case PP_OD_FEATURE_UCLK_BIT: + if (lower_boundary) { + od_min_setting = overdrive_lowerlimits->UclkFmin; + od_max_setting = overdrive_upperlimits->UclkFmin; + } else { + od_min_setting = overdrive_lowerlimits->UclkFmax; + od_max_setting = overdrive_upperlimits->UclkFmax; + } + break; + case PP_OD_FEATURE_GFX_VF_CURVE_BIT: + od_min_setting = overdrive_lowerlimits->VoltageOffsetPerZoneBoundary; + od_max_setting = overdrive_upperlimits->VoltageOffsetPerZoneBoundary; + break; + default: + break; + } + + if (min) + *min = od_min_setting; + if (max) + *max = od_max_setting; +} + +static void smu_v13_0_0_dump_od_table(struct smu_context *smu, + OverDriveTableExternal_t *od_table) +{ + struct amdgpu_device *adev = smu->adev; + + dev_dbg(adev->dev, "OD: Gfxclk: (%d, %d)\n", od_table->OverDriveTable.GfxclkFmin, + od_table->OverDriveTable.GfxclkFmax); + dev_dbg(adev->dev, "OD: Uclk: (%d, %d)\n", od_table->OverDriveTable.UclkFmin, + od_table->OverDriveTable.UclkFmax); +} + +static int smu_v13_0_0_get_overdrive_table(struct smu_context *smu, + OverDriveTableExternal_t *od_table) +{ + int ret = 0; + + ret = smu_cmn_update_table(smu, + SMU_TABLE_OVERDRIVE, + 0, + (void *)od_table, + false); + if (ret) + dev_err(smu->adev->dev, "Failed to get overdrive table!\n"); + + return ret; +} + +static int smu_v13_0_0_upload_overdrive_table(struct smu_context *smu, + OverDriveTableExternal_t *od_table) +{ + int ret = 0; + + ret = smu_cmn_update_table(smu, + SMU_TABLE_OVERDRIVE, + 0, + (void *)od_table, + true); + if (ret) + dev_err(smu->adev->dev, "Failed to upload overdrive table!\n"); + + return ret; +} + static int smu_v13_0_0_print_clk_levels(struct smu_context *smu, enum smu_clk_type clk_type, char *buf) { struct smu_dpm_context *smu_dpm = &smu->smu_dpm; struct smu_13_0_dpm_context *dpm_context = smu_dpm->dpm_context; + OverDriveTableExternal_t *od_table = + (OverDriveTableExternal_t *)smu->smu_table.overdrive_table; struct smu_13_0_dpm_table *single_dpm_table; struct smu_13_0_pcie_table *pcie_table; const int link_width[] = {0, 1, 2, 4, 8, 12, 16}; uint32_t gen_speed, lane_width; int i, curr_freq, size = 0; + int32_t min_value, max_value; int ret = 0; smu_cmn_get_sysfs_buf(&buf, &size); @@ -1149,6 +1260,89 @@ static int smu_v13_0_0_print_clk_levels(struct smu_context *smu, "*" : ""); break; + case SMU_OD_SCLK: + if (!smu_v13_0_0_is_od_feature_supported(smu, + PP_OD_FEATURE_GFXCLK_BIT)) + break; + + size += sysfs_emit_at(buf, size, "OD_SCLK:\n"); + size += sysfs_emit_at(buf, size, "0: %uMhz\n1: %uMhz\n", + od_table->OverDriveTable.GfxclkFmin, + od_table->OverDriveTable.GfxclkFmax); + break; + + case SMU_OD_MCLK: + if (!smu_v13_0_0_is_od_feature_supported(smu, + PP_OD_FEATURE_UCLK_BIT)) + break; + + size += sysfs_emit_at(buf, size, "OD_MCLK:\n"); + size += sysfs_emit_at(buf, size, "0: %uMhz\n1: %uMHz\n", + od_table->OverDriveTable.UclkFmin, + od_table->OverDriveTable.UclkFmax); + break; + + case SMU_OD_VDDC_CURVE: + if (!smu_v13_0_0_is_od_feature_supported(smu, + PP_OD_FEATURE_GFX_VF_CURVE_BIT)) + break; + + size += sysfs_emit_at(buf, size, "OD_VDDC_CURVE:\n"); + for (i = 0; i < PP_NUM_OD_VF_CURVE_POINTS; i++) + size += sysfs_emit_at(buf, size, "%d: %dmv\n", + i, + od_table->OverDriveTable.VoltageOffsetPerZoneBoundary[i]); + break; + + case SMU_OD_RANGE: + if (!smu_v13_0_0_is_od_feature_supported(smu, PP_OD_FEATURE_GFXCLK_BIT) && + !smu_v13_0_0_is_od_feature_supported(smu, PP_OD_FEATURE_UCLK_BIT) && + !smu_v13_0_0_is_od_feature_supported(smu, PP_OD_FEATURE_GFX_VF_CURVE_BIT)) + break; + + size += sysfs_emit_at(buf, size, "%s:\n", "OD_RANGE"); + + if (smu_v13_0_0_is_od_feature_supported(smu, PP_OD_FEATURE_GFXCLK_BIT)) { + smu_v13_0_0_get_od_setting_limits(smu, + PP_OD_FEATURE_GFXCLK_BIT, + true, + &min_value, + NULL); + smu_v13_0_0_get_od_setting_limits(smu, + PP_OD_FEATURE_GFXCLK_BIT, + false, + NULL, + &max_value); + size += sysfs_emit_at(buf, size, "SCLK: %7uMhz %10uMhz\n", + min_value, max_value); + } + + if (smu_v13_0_0_is_od_feature_supported(smu, PP_OD_FEATURE_UCLK_BIT)) { + smu_v13_0_0_get_od_setting_limits(smu, + PP_OD_FEATURE_UCLK_BIT, + true, + &min_value, + NULL); + smu_v13_0_0_get_od_setting_limits(smu, + PP_OD_FEATURE_UCLK_BIT, + false, + NULL, + &max_value); + size += sysfs_emit_at(buf, size, "MCLK: %7uMhz %10uMhz\n", + min_value, max_value); + } + + if (smu_v13_0_0_is_od_feature_supported(smu, PP_OD_FEATURE_GFX_VF_CURVE_BIT)) { + smu_v13_0_0_get_od_setting_limits(smu, + PP_OD_FEATURE_GFX_VF_CURVE_BIT, + true, + &min_value, + &max_value); + size += sysfs_emit_at(buf, size, "VDDC_CURVE: %7dmv %10dmv\n", + min_value, max_value); + } + break; + default: break; } @@ -1156,6 +1350,222 @@ static int smu_v13_0_0_print_clk_levels(struct smu_context *smu, return size; } +static int smu_v13_0_0_od_edit_dpm_table(struct smu_context *smu, + enum PP_OD_DPM_TABLE_COMMAND type, + long input[], + uint32_t size) +{ + struct smu_table_context *table_context = &smu->smu_table; + OverDriveTableExternal_t *od_table = + (OverDriveTableExternal_t *)table_context->overdrive_table; + struct amdgpu_device *adev = smu->adev; + uint32_t offset_of_featurectrlmask; + int32_t minimum, maximum; + uint32_t feature_ctrlmask; + int i, ret = 0; + + switch (type) { + case PP_OD_EDIT_SCLK_VDDC_TABLE: + if (!smu_v13_0_0_is_od_feature_supported(smu, PP_OD_FEATURE_GFXCLK_BIT)) { + dev_warn(adev->dev, "GFXCLK_LIMITS setting not supported!\n"); + return -ENOTSUPP; + } + + for (i = 0; i < size; i += 2) { + if (i + 2 > size) { + dev_info(adev->dev, "invalid number of input parameters %d\n", size); + return -EINVAL; + } + + switch (input[i]) { + case 0: + smu_v13_0_0_get_od_setting_limits(smu, + PP_OD_FEATURE_GFXCLK_BIT, + true, + &minimum, + &maximum); + if (input[i + 1] < minimum || + input[i + 1] > maximum) { + dev_info(adev->dev, "GfxclkFmin (%ld) must be within [%u, %u]!\n", + input[i + 1], minimum, maximum); + return -EINVAL; + } + + od_table->OverDriveTable.GfxclkFmin = input[i + 1]; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_GFXCLK_BIT; + break; + + case 1: + smu_v13_0_0_get_od_setting_limits(smu, + PP_OD_FEATURE_GFXCLK_BIT, + false, + &minimum, + &maximum); + if (input[i + 1] < minimum || + input[i + 1] > maximum) { + dev_info(adev->dev, "GfxclkFmax (%ld) must be within [%u, %u]!\n", + input[i + 1], minimum, maximum); + return -EINVAL; + } + + od_table->OverDriveTable.GfxclkFmax = input[i + 1]; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_GFXCLK_BIT; + break; + + default: + dev_info(adev->dev, "Invalid SCLK_VDDC_TABLE index: %ld\n", input[i]); + dev_info(adev->dev, "Supported indices: [0:min,1:max]\n"); + return -EINVAL; + } + } + + if (od_table->OverDriveTable.GfxclkFmin > od_table->OverDriveTable.GfxclkFmax) { + dev_err(adev->dev, + "Invalid setting: GfxclkFmin(%u) is bigger than GfxclkFmax(%u)\n", + (uint32_t)od_table->OverDriveTable.GfxclkFmin, + (uint32_t)od_table->OverDriveTable.GfxclkFmax); + return -EINVAL; + } + break; + + case PP_OD_EDIT_MCLK_VDDC_TABLE: + if (!smu_v13_0_0_is_od_feature_supported(smu, PP_OD_FEATURE_UCLK_BIT)) { + dev_warn(adev->dev, "UCLK_LIMITS setting not supported!\n"); + return -ENOTSUPP; + } + + for (i = 0; i < size; i += 2) { + if (i + 2 > size) { + dev_info(adev->dev, "invalid number of input parameters %d\n", size); + return -EINVAL; + } + + switch (input[i]) { + case 0: + smu_v13_0_0_get_od_setting_limits(smu, + PP_OD_FEATURE_UCLK_BIT, + true, + &minimum, + &maximum); + if (input[i + 1] < minimum || + input[i + 1] > maximum) { + dev_info(adev->dev, "UclkFmin (%ld) must be within [%u, %u]!\n", + input[i + 1], minimum, maximum); + return -EINVAL; + } + + od_table->OverDriveTable.UclkFmin = input[i + 1]; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_UCLK_BIT; + break; + + case 1: + smu_v13_0_0_get_od_setting_limits(smu, + PP_OD_FEATURE_UCLK_BIT, + false, + &minimum, + &maximum); + if (input[i + 1] < minimum || + input[i + 1] > maximum) { + dev_info(adev->dev, "UclkFmax (%ld) must be within [%u, %u]!\n", + input[i + 1], minimum, maximum); + return -EINVAL; + } + + od_table->OverDriveTable.UclkFmax = input[i + 1]; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_UCLK_BIT; + break; + + default: + dev_info(adev->dev, "Invalid MCLK_VDDC_TABLE index: %ld\n", input[i]); + dev_info(adev->dev, "Supported indices: [0:min,1:max]\n"); + return -EINVAL; + } + } + + if (od_table->OverDriveTable.UclkFmin > od_table->OverDriveTable.UclkFmax) { + dev_err(adev->dev, + "Invalid setting: UclkFmin(%u) is bigger than UclkFmax(%u)\n", + (uint32_t)od_table->OverDriveTable.UclkFmin, + (uint32_t)od_table->OverDriveTable.UclkFmax); + return -EINVAL; + } + break; + + case PP_OD_EDIT_VDDC_CURVE: + if (!smu_v13_0_0_is_od_feature_supported(smu, PP_OD_FEATURE_GFX_VF_CURVE_BIT)) { + dev_warn(adev->dev, "VF curve setting not supported!\n"); + return -ENOTSUPP; + } + + if (input[0] >= PP_NUM_OD_VF_CURVE_POINTS || + input[0] < 0) + return -EINVAL; + + smu_v13_0_0_get_od_setting_limits(smu, + PP_OD_FEATURE_GFX_VF_CURVE_BIT, + true, + &minimum, + &maximum); + if (input[1] < minimum || + input[1] > maximum) { + dev_info(adev->dev, "Voltage offset (%ld) must be within [%d, %d]!\n", + input[1], minimum, maximum); + return -EINVAL; + } + + od_table->OverDriveTable.VoltageOffsetPerZoneBoundary[input[0]] = input[1]; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_GFX_VF_CURVE_BIT; + break; + + case PP_OD_RESTORE_DEFAULT_TABLE: + feature_ctrlmask = od_table->OverDriveTable.FeatureCtrlMask; + memcpy(od_table, + table_context->boot_overdrive_table, + sizeof(OverDriveTableExternal_t)); + od_table->OverDriveTable.FeatureCtrlMask = feature_ctrlmask; + fallthrough; + + case PP_OD_COMMIT_DPM_TABLE: + /* + * The member below instructs PMFW the settings focused in + * this single operation. + * `uint32_t FeatureCtrlMask;` + * It does not contain actual informations about user's custom + * settings. Thus we do not cache it. + */ + offset_of_featurectrlmask = offsetof(OverDriveTable_t, FeatureCtrlMask); + if (memcmp((u8 *)od_table + offset_of_featurectrlmask, + table_context->user_overdrive_table + offset_of_featurectrlmask, + sizeof(OverDriveTableExternal_t) - offset_of_featurectrlmask)) { + smu_v13_0_0_dump_od_table(smu, od_table); + + ret = smu_v13_0_0_upload_overdrive_table(smu, od_table); + if (ret) { + dev_err(adev->dev, "Failed to upload overdrive table!\n"); + return ret; + } + + od_table->OverDriveTable.FeatureCtrlMask = 0; + memcpy(table_context->user_overdrive_table + offset_of_featurectrlmask, + (u8 *)od_table + offset_of_featurectrlmask, + sizeof(OverDriveTableExternal_t) - offset_of_featurectrlmask); + + if (!memcmp(table_context->user_overdrive_table, + table_context->boot_overdrive_table, + sizeof(OverDriveTableExternal_t))) + smu->user_dpm_profile.user_od = false; + else + smu->user_dpm_profile.user_od = true; + } + break; + + default: + return -ENOSYS; + } + + return ret; +} + static int smu_v13_0_0_force_clk_levels(struct smu_context *smu, enum smu_clk_type clk_type, uint32_t mask) @@ -1384,6 +1794,78 @@ static ssize_t smu_v13_0_0_get_gpu_metrics(struct smu_context *smu, return sizeof(struct gpu_metrics_v1_3); } +static int smu_v13_0_0_set_default_od_settings(struct smu_context *smu) +{ + OverDriveTableExternal_t *od_table = + (OverDriveTableExternal_t *)smu->smu_table.overdrive_table; + OverDriveTableExternal_t *boot_od_table = + (OverDriveTableExternal_t *)smu->smu_table.boot_overdrive_table; + OverDriveTableExternal_t *user_od_table = + (OverDriveTableExternal_t *)smu->smu_table.user_overdrive_table; + OverDriveTableExternal_t user_od_table_bak; + int ret = 0; + int i; + + ret = smu_v13_0_0_get_overdrive_table(smu, boot_od_table); + if (ret) + return ret; + + smu_v13_0_0_dump_od_table(smu, boot_od_table); + + memcpy(od_table, + boot_od_table, + sizeof(OverDriveTableExternal_t)); + + /* + * For S3/S4/Runpm resume, we need to setup those overdrive tables again, + * but we have to preserve user defined values in "user_od_table". + */ + if (!smu->adev->in_suspend) { + memcpy(user_od_table, + boot_od_table, + sizeof(OverDriveTableExternal_t)); + smu->user_dpm_profile.user_od = false; + } else if (smu->user_dpm_profile.user_od) { + memcpy(&user_od_table_bak, + user_od_table, + sizeof(OverDriveTableExternal_t)); + memcpy(user_od_table, + boot_od_table, + sizeof(OverDriveTableExternal_t)); + user_od_table->OverDriveTable.GfxclkFmin = + user_od_table_bak.OverDriveTable.GfxclkFmin; + user_od_table->OverDriveTable.GfxclkFmax = + user_od_table_bak.OverDriveTable.GfxclkFmax; + user_od_table->OverDriveTable.UclkFmin = + user_od_table_bak.OverDriveTable.UclkFmin; + user_od_table->OverDriveTable.UclkFmax = + user_od_table_bak.OverDriveTable.UclkFmax; + for (i = 0; i < PP_NUM_OD_VF_CURVE_POINTS; i++) + user_od_table->OverDriveTable.VoltageOffsetPerZoneBoundary[i] = + user_od_table_bak.OverDriveTable.VoltageOffsetPerZoneBoundary[i]; + } + + return 0; +} + +static int smu_v13_0_0_restore_user_od_settings(struct smu_context *smu) +{ + struct smu_table_context *table_context = &smu->smu_table; + OverDriveTableExternal_t *od_table = table_context->overdrive_table; + OverDriveTableExternal_t *user_od_table = table_context->user_overdrive_table; + int res; + + user_od_table->OverDriveTable.FeatureCtrlMask = 1U << PP_OD_FEATURE_GFXCLK_BIT | + 1U << PP_OD_FEATURE_UCLK_BIT | + 1U << PP_OD_FEATURE_GFX_VF_CURVE_BIT; + res = smu_v13_0_0_upload_overdrive_table(smu, user_od_table); + user_od_table->OverDriveTable.FeatureCtrlMask = 0; + if (res == 0) + memcpy(od_table, user_od_table, sizeof(OverDriveTableExternal_t)); + + return res; +} + static int smu_v13_0_0_populate_umd_state_clk(struct smu_context *smu) { struct smu_13_0_dpm_context *dpm_context = @@ -1696,10 +2178,39 @@ static int smu_v13_0_0_set_power_profile_mode(struct smu_context *smu, } } - /* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */ - workload_type = smu_cmn_to_asic_specific_index(smu, + if (smu->power_profile_mode == PP_SMC_POWER_PROFILE_COMPUTE && + (((smu->adev->pdev->device == 0x744C) && (smu->adev->pdev->revision == 0xC8)) || + ((smu->adev->pdev->device == 0x744C) && (smu->adev->pdev->revision == 0xCC)))) { + ret = smu_cmn_update_table(smu, + SMU_TABLE_ACTIVITY_MONITOR_COEFF, + WORKLOAD_PPLIB_COMPUTE_BIT, + (void *)(&activity_monitor_external), + false); + if (ret) { + dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__); + return ret; + } + + ret = smu_cmn_update_table(smu, + SMU_TABLE_ACTIVITY_MONITOR_COEFF, + WORKLOAD_PPLIB_CUSTOM_BIT, + (void *)(&activity_monitor_external), + true); + if (ret) { + dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__); + return ret; + } + + workload_type = smu_cmn_to_asic_specific_index(smu, + CMN2ASIC_MAPPING_WORKLOAD, + PP_SMC_POWER_PROFILE_CUSTOM); + } else { + /* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */ + workload_type = smu_cmn_to_asic_specific_index(smu, CMN2ASIC_MAPPING_WORKLOAD, smu->power_profile_mode); + } + if (workload_type < 0) return -EINVAL; @@ -2097,7 +2608,7 @@ static ssize_t smu_v13_0_0_get_ecc_info(struct smu_context *smu, ecc_table = (EccInfoTable_t *)smu_table->ecc_table; - for (i = 0; i < UMC_V8_10_TOTAL_CHANNEL_NUM(adev); i++) { + for (i = 0; i < ARRAY_SIZE(ecc_table->EccInfo); i++) { ecc_info_per_channel = &(eccinfo->ecc[i]); ecc_info_per_channel->ce_count_lo_chip = ecc_table->EccInfo[i].ce_count_lo_chip; @@ -2150,6 +2661,9 @@ static const struct pptable_funcs smu_v13_0_0_ppt_funcs = { .notify_memory_pool_location = smu_v13_0_notify_memory_pool_location, .get_gpu_metrics = smu_v13_0_0_get_gpu_metrics, .set_soft_freq_limited_range = smu_v13_0_set_soft_freq_limited_range, + .set_default_od_settings = smu_v13_0_0_set_default_od_settings, + .restore_user_od_settings = smu_v13_0_0_restore_user_od_settings, + .od_edit_dpm_table = smu_v13_0_0_od_edit_dpm_table, .init_pptable_microcode = smu_v13_0_init_pptable_microcode, .populate_umd_state_clk = smu_v13_0_0_populate_umd_state_clk, .set_performance_level = smu_v13_0_set_performance_level, @@ -2199,5 +2713,6 @@ void smu_v13_0_0_set_ppt_funcs(struct smu_context *smu) smu->table_map = smu_v13_0_0_table_map; smu->pwr_src_map = smu_v13_0_0_pwr_src_map; smu->workload_map = smu_v13_0_0_workload_map; + smu->smc_driver_if_version = SMU13_0_0_DRIVER_IF_VERSION; smu_v13_0_0_set_smu_mailbox_registers(smu); } diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_4_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_4_ppt.c index 8fa9a36c38b6..ef37dda9908f 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_4_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_4_ppt.c @@ -54,6 +54,10 @@ #define FEATURE_MASK(feature) (1ULL << feature) +#define SMU_13_0_4_UMD_PSTATE_GFXCLK 938 +#define SMU_13_0_4_UMD_PSTATE_SOCCLK 938 +#define SMU_13_0_4_UMD_PSTATE_FCLK 1875 + #define SMC_DPM_FEATURE ( \ FEATURE_MASK(FEATURE_CCLK_DPM_BIT) | \ FEATURE_MASK(FEATURE_VCN_DPM_BIT) | \ @@ -478,7 +482,7 @@ static int smu_v13_0_4_get_dpm_level_count(struct smu_context *smu, static int smu_v13_0_4_print_clk_levels(struct smu_context *smu, enum smu_clk_type clk_type, char *buf) { - int i, size = 0, ret = 0; + int i, idx, size = 0, ret = 0; uint32_t cur_value = 0, value = 0, count = 0; uint32_t min, max; @@ -512,7 +516,8 @@ static int smu_v13_0_4_print_clk_levels(struct smu_context *smu, break; for (i = 0; i < count; i++) { - ret = smu_v13_0_4_get_dpm_freq_by_index(smu, clk_type, i, &value); + idx = (clk_type == SMU_FCLK || clk_type == SMU_MCLK) ? (count - i - 1) : i; + ret = smu_v13_0_4_get_dpm_freq_by_index(smu, clk_type, idx, &value); if (ret) break; @@ -830,6 +835,8 @@ static int smu_v13_0_4_set_soft_freq_limited_range(struct smu_context *smu, uint32_t max) { enum smu_message_type msg_set_min, msg_set_max; + uint32_t min_clk = min; + uint32_t max_clk = max; int ret = 0; if (!smu_v13_0_4_clk_dpm_is_enabled(smu, clk_type)) @@ -858,12 +865,17 @@ static int smu_v13_0_4_set_soft_freq_limited_range(struct smu_context *smu, return -EINVAL; } - ret = smu_cmn_send_smc_msg_with_param(smu, msg_set_min, min, NULL); + if (clk_type == SMU_VCLK) { + min_clk = min << SMU_13_VCLK_SHIFT; + max_clk = max << SMU_13_VCLK_SHIFT; + } + + ret = smu_cmn_send_smc_msg_with_param(smu, msg_set_min, min_clk, NULL); if (ret) return ret; return smu_cmn_send_smc_msg_with_param(smu, msg_set_max, - max, NULL); + max_clk, NULL); } static int smu_v13_0_4_force_clk_levels(struct smu_context *smu, @@ -900,6 +912,50 @@ static int smu_v13_0_4_force_clk_levels(struct smu_context *smu, return ret; } +static int smu_v13_0_4_get_dpm_profile_freq(struct smu_context *smu, + enum amd_dpm_forced_level level, + enum smu_clk_type clk_type, + uint32_t *min_clk, + uint32_t *max_clk) +{ + int ret = 0; + uint32_t clk_limit = 0; + + switch (clk_type) { + case SMU_GFXCLK: + case SMU_SCLK: + clk_limit = SMU_13_0_4_UMD_PSTATE_GFXCLK; + if (level == AMD_DPM_FORCED_LEVEL_PROFILE_PEAK) + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_SCLK, NULL, &clk_limit); + else if (level == AMD_DPM_FORCED_LEVEL_PROFILE_MIN_SCLK) + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_SCLK, &clk_limit, NULL); + break; + case SMU_SOCCLK: + clk_limit = SMU_13_0_4_UMD_PSTATE_SOCCLK; + if (level == AMD_DPM_FORCED_LEVEL_PROFILE_PEAK) + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_SOCCLK, NULL, &clk_limit); + break; + case SMU_FCLK: + clk_limit = SMU_13_0_4_UMD_PSTATE_FCLK; + if (level == AMD_DPM_FORCED_LEVEL_PROFILE_PEAK) + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_FCLK, NULL, &clk_limit); + else if (level == AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK) + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_FCLK, &clk_limit, NULL); + break; + case SMU_VCLK: + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_VCLK, NULL, &clk_limit); + break; + case SMU_DCLK: + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_DCLK, NULL, &clk_limit); + break; + default: + ret = -EINVAL; + break; + } + *min_clk = *max_clk = clk_limit; + return ret; +} + static int smu_v13_0_4_set_performance_level(struct smu_context *smu, enum amd_dpm_forced_level level) { @@ -907,6 +963,8 @@ static int smu_v13_0_4_set_performance_level(struct smu_context *smu, uint32_t sclk_min = 0, sclk_max = 0; uint32_t fclk_min = 0, fclk_max = 0; uint32_t socclk_min = 0, socclk_max = 0; + uint32_t vclk_min = 0, vclk_max = 0; + uint32_t dclk_min = 0, dclk_max = 0; int ret = 0; switch (level) { @@ -914,28 +972,42 @@ static int smu_v13_0_4_set_performance_level(struct smu_context *smu, smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_SCLK, NULL, &sclk_max); smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_FCLK, NULL, &fclk_max); smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_SOCCLK, NULL, &socclk_max); + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_VCLK, NULL, &vclk_max); + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_DCLK, NULL, &dclk_max); sclk_min = sclk_max; fclk_min = fclk_max; socclk_min = socclk_max; + vclk_min = vclk_max; + dclk_min = dclk_max; break; case AMD_DPM_FORCED_LEVEL_LOW: smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_SCLK, &sclk_min, NULL); smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_FCLK, &fclk_min, NULL); smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_SOCCLK, &socclk_min, NULL); + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_VCLK, &vclk_min, NULL); + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_DCLK, &dclk_min, NULL); sclk_max = sclk_min; fclk_max = fclk_min; socclk_max = socclk_min; + vclk_max = vclk_min; + dclk_max = dclk_min; break; case AMD_DPM_FORCED_LEVEL_AUTO: smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_SCLK, &sclk_min, &sclk_max); smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_FCLK, &fclk_min, &fclk_max); smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_SOCCLK, &socclk_min, &socclk_max); + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_VCLK, &vclk_min, &vclk_max); + smu_v13_0_4_get_dpm_ultimate_freq(smu, SMU_DCLK, &dclk_min, &dclk_max); break; case AMD_DPM_FORCED_LEVEL_PROFILE_STANDARD: case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_SCLK: case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK: case AMD_DPM_FORCED_LEVEL_PROFILE_PEAK: - /* Temporarily do nothing since the optimal clocks haven't been provided yet */ + smu_v13_0_4_get_dpm_profile_freq(smu, level, SMU_SCLK, &sclk_min, &sclk_max); + smu_v13_0_4_get_dpm_profile_freq(smu, level, SMU_FCLK, &fclk_min, &fclk_max); + smu_v13_0_4_get_dpm_profile_freq(smu, level, SMU_SOCCLK, &socclk_min, &socclk_max); + smu_v13_0_4_get_dpm_profile_freq(smu, level, SMU_VCLK, &vclk_min, &vclk_max); + smu_v13_0_4_get_dpm_profile_freq(smu, level, SMU_DCLK, &dclk_min, &dclk_max); break; case AMD_DPM_FORCED_LEVEL_MANUAL: case AMD_DPM_FORCED_LEVEL_PROFILE_EXIT: @@ -975,6 +1047,23 @@ static int smu_v13_0_4_set_performance_level(struct smu_context *smu, return ret; } + if (vclk_min && vclk_max) { + ret = smu_v13_0_4_set_soft_freq_limited_range(smu, + SMU_VCLK, + vclk_min, + vclk_max); + if (ret) + return ret; + } + + if (dclk_min && dclk_max) { + ret = smu_v13_0_4_set_soft_freq_limited_range(smu, + SMU_DCLK, + dclk_min, + dclk_max); + if (ret) + return ret; + } return ret; } @@ -1043,6 +1132,7 @@ void smu_v13_0_4_set_ppt_funcs(struct smu_context *smu) smu->message_map = smu_v13_0_4_message_map; smu->feature_map = smu_v13_0_4_feature_mask_map; smu->table_map = smu_v13_0_4_table_map; + smu->smc_driver_if_version = SMU13_0_4_DRIVER_IF_VERSION; smu->is_apu = true; if (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(13, 0, 4)) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_5_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_5_ppt.c index 66445964efbd..42f110602eb1 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_5_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_5_ppt.c @@ -831,6 +831,8 @@ static int smu_v13_0_5_set_soft_freq_limited_range(struct smu_context *smu, uint32_t max) { enum smu_message_type msg_set_min, msg_set_max; + uint32_t min_clk = min; + uint32_t max_clk = max; int ret = 0; if (!smu_v13_0_5_clk_dpm_is_enabled(smu, clk_type)) @@ -851,11 +853,16 @@ static int smu_v13_0_5_set_soft_freq_limited_range(struct smu_context *smu, return -EINVAL; } - ret = smu_cmn_send_smc_msg_with_param(smu, msg_set_min, min, NULL); + if (clk_type == SMU_VCLK) { + min_clk = min << SMU_13_VCLK_SHIFT; + max_clk = max << SMU_13_VCLK_SHIFT; + } + + ret = smu_cmn_send_smc_msg_with_param(smu, msg_set_min, min_clk, NULL); if (ret) goto out; - ret = smu_cmn_send_smc_msg_with_param(smu, msg_set_max, max, NULL); + ret = smu_cmn_send_smc_msg_with_param(smu, msg_set_max, max_clk, NULL); if (ret) goto out; @@ -866,7 +873,7 @@ out: static int smu_v13_0_5_print_clk_levels(struct smu_context *smu, enum smu_clk_type clk_type, char *buf) { - int i, size = 0, ret = 0; + int i, idx, size = 0, ret = 0; uint32_t cur_value = 0, value = 0, count = 0; uint32_t min = 0, max = 0; @@ -898,7 +905,8 @@ static int smu_v13_0_5_print_clk_levels(struct smu_context *smu, goto print_clk_out; for (i = 0; i < count; i++) { - ret = smu_v13_0_5_get_dpm_freq_by_index(smu, clk_type, i, &value); + idx = (clk_type == SMU_MCLK) ? (count - i - 1) : i; + ret = smu_v13_0_5_get_dpm_freq_by_index(smu, clk_type, idx, &value); if (ret) goto print_clk_out; @@ -970,31 +978,79 @@ force_level_out: return ret; } +static int smu_v13_0_5_get_dpm_profile_freq(struct smu_context *smu, + enum amd_dpm_forced_level level, + enum smu_clk_type clk_type, + uint32_t *min_clk, + uint32_t *max_clk) +{ + int ret = 0; + uint32_t clk_limit = 0; + + switch (clk_type) { + case SMU_GFXCLK: + case SMU_SCLK: + clk_limit = SMU_13_0_5_UMD_PSTATE_GFXCLK; + if (level == AMD_DPM_FORCED_LEVEL_PROFILE_PEAK) + smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_SCLK, NULL, &clk_limit); + else if (level == AMD_DPM_FORCED_LEVEL_PROFILE_MIN_SCLK) + smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_SCLK, &clk_limit, NULL); + break; + case SMU_VCLK: + smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_VCLK, NULL, &clk_limit); + break; + case SMU_DCLK: + smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_DCLK, NULL, &clk_limit); + break; + default: + ret = -EINVAL; + break; + } + *min_clk = *max_clk = clk_limit; + return ret; +} + static int smu_v13_0_5_set_performance_level(struct smu_context *smu, enum amd_dpm_forced_level level) { struct amdgpu_device *adev = smu->adev; uint32_t sclk_min = 0, sclk_max = 0; + uint32_t vclk_min = 0, vclk_max = 0; + uint32_t dclk_min = 0, dclk_max = 0; int ret = 0; switch (level) { case AMD_DPM_FORCED_LEVEL_HIGH: smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_SCLK, NULL, &sclk_max); + smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_VCLK, NULL, &vclk_max); + smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_DCLK, NULL, &dclk_max); sclk_min = sclk_max; + vclk_min = vclk_max; + dclk_min = dclk_max; break; case AMD_DPM_FORCED_LEVEL_LOW: smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_SCLK, &sclk_min, NULL); + smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_VCLK, &vclk_min, NULL); + smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_DCLK, &dclk_min, NULL); sclk_max = sclk_min; + vclk_max = vclk_min; + dclk_max = dclk_min; break; case AMD_DPM_FORCED_LEVEL_AUTO: smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_SCLK, &sclk_min, &sclk_max); + smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_VCLK, &vclk_min, &vclk_max); + smu_v13_0_5_get_dpm_ultimate_freq(smu, SMU_DCLK, &dclk_min, &dclk_max); break; case AMD_DPM_FORCED_LEVEL_PROFILE_STANDARD: case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_SCLK: - case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK: case AMD_DPM_FORCED_LEVEL_PROFILE_PEAK: - /* Temporarily do nothing since the optimal clocks haven't been provided yet */ + smu_v13_0_5_get_dpm_profile_freq(smu, level, SMU_SCLK, &sclk_min, &sclk_max); + smu_v13_0_5_get_dpm_profile_freq(smu, level, SMU_VCLK, &vclk_min, &vclk_max); + smu_v13_0_5_get_dpm_profile_freq(smu, level, SMU_DCLK, &dclk_min, &dclk_max); break; + case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK: + dev_err(adev->dev, "The performance level profile_min_mclk is not supported."); + return -EOPNOTSUPP; case AMD_DPM_FORCED_LEVEL_MANUAL: case AMD_DPM_FORCED_LEVEL_PROFILE_EXIT: return 0; @@ -1015,6 +1071,23 @@ static int smu_v13_0_5_set_performance_level(struct smu_context *smu, smu->gfx_actual_soft_max_freq = sclk_max; } + if (vclk_min && vclk_max) { + ret = smu_v13_0_5_set_soft_freq_limited_range(smu, + SMU_VCLK, + vclk_min, + vclk_max); + if (ret) + return ret; + } + + if (dclk_min && dclk_max) { + ret = smu_v13_0_5_set_soft_freq_limited_range(smu, + SMU_DCLK, + dclk_min, + dclk_max); + if (ret) + return ret; + } return ret; } @@ -1068,6 +1141,7 @@ void smu_v13_0_5_set_ppt_funcs(struct smu_context *smu) smu->feature_map = smu_v13_0_5_feature_mask_map; smu->table_map = smu_v13_0_5_table_map; smu->is_apu = true; + smu->smc_driver_if_version = SMU13_0_5_DRIVER_IF_VERSION; smu->param_reg = SOC15_REG_OFFSET(MP1, 0, mmMP1_C2PMSG_34); smu->msg_reg = SOC15_REG_OFFSET(MP1, 0, mmMP1_C2PMSG_2); smu->resp_reg = SOC15_REG_OFFSET(MP1, 0, mmMP1_C2PMSG_33); diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_5_ppt.h b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_5_ppt.h index 40bc0f8e6d61..263cd651855e 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_5_ppt.h +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_5_ppt.h @@ -24,6 +24,6 @@ #define __SMU_V13_0_5_PPT_H__ extern void smu_v13_0_5_set_ppt_funcs(struct smu_context *smu); -#define SMU_13_0_5_UMD_PSTATE_GFXCLK 1100 +#define SMU_13_0_5_UMD_PSTATE_GFXCLK 700 #endif diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c index ea8f3d6fb98b..a92ea4601ea4 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c @@ -80,7 +80,10 @@ /* possible frequency drift (1Mhz) */ #define EPSILON 1 -#define smnPCIE_ESM_CTRL 0x111003D0 +#define smnPCIE_ESM_CTRL 0x193D0 +#define smnPCIE_LC_LINK_WIDTH_CNTL 0x1ab40288 +#define PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_RD_MASK 0x00000070L +#define PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_RD__SHIFT 0x4 static const struct cmn2asic_msg_mapping smu_v13_0_6_message_map[SMU_MSG_MAX_COUNT] = { MSG_MAP(TestMessage, PPSMC_MSG_TestMessage, 0), @@ -122,6 +125,7 @@ static const struct cmn2asic_msg_mapping smu_v13_0_6_message_map[SMU_MSG_MAX_COU MSG_MAP(GetMaxGfxclkFrequency, PPSMC_MSG_GetMaxGfxDpmFreq, 0), MSG_MAP(SetSoftMinGfxclk, PPSMC_MSG_SetSoftMinGfxClk, 0), MSG_MAP(SetSoftMaxGfxClk, PPSMC_MSG_SetSoftMaxGfxClk, 0), + MSG_MAP(PrepareMp1ForUnload, PPSMC_MSG_PrepareForDriverUnload, 0), }; static const struct cmn2asic_mapping smu_v13_0_6_clk_map[SMU_CLK_COUNT] = { @@ -171,18 +175,12 @@ static const struct cmn2asic_mapping smu_v13_0_6_table_map[SMU_TABLE_COUNT] = { TAB_MAP(I2C_COMMANDS), }; -#define THROTTLER_PROCHOT_GFX_BIT 0 -#define THROTTLER_PPT_BIT 1 -#define THROTTLER_TEMP_SOC_BIT 2 -#define THROTTLER_TEMP_VR_GFX_BIT 3 -#define THROTTLER_TEMP_HBM_BIT 4 - static const uint8_t smu_v13_0_6_throttler_map[] = { [THROTTLER_PPT_BIT] = (SMU_THROTTLER_PPT0_BIT), - [THROTTLER_TEMP_SOC_BIT] = (SMU_THROTTLER_TEMP_GPU_BIT), - [THROTTLER_TEMP_HBM_BIT] = (SMU_THROTTLER_TEMP_MEM_BIT), - [THROTTLER_TEMP_VR_GFX_BIT] = (SMU_THROTTLER_TEMP_VR_GFX_BIT), - [THROTTLER_PROCHOT_GFX_BIT] = (SMU_THROTTLER_PROCHOT_GFX_BIT), + [THROTTLER_THERMAL_SOCKET_BIT] = (SMU_THROTTLER_TEMP_GPU_BIT), + [THROTTLER_THERMAL_HBM_BIT] = (SMU_THROTTLER_TEMP_MEM_BIT), + [THROTTLER_THERMAL_VR_BIT] = (SMU_THROTTLER_TEMP_VR_GFX_BIT), + [THROTTLER_PROCHOT_BIT] = (SMU_THROTTLER_PROCHOT_GFX_BIT), }; struct PPTable_t { @@ -197,10 +195,12 @@ struct PPTable_t { uint32_t LclkFrequencyTable[4]; uint32_t MaxLclkDpmRange; uint32_t MinLclkDpmRange; + uint64_t PublicSerialNumber_AID; bool Init; }; #define SMUQ10_TO_UINT(x) ((x) >> 10) +#define SMUQ16_TO_UINT(x) ((x) >> 16) struct smu_v13_0_6_dpm_map { enum smu_clk_type clk_type; @@ -220,10 +220,12 @@ static int smu_v13_0_6_tables_init(struct smu_context *smu) PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM); SMU_TABLE_INIT(tables, SMU_TABLE_SMU_METRICS, sizeof(MetricsTable_t), - PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM); + PAGE_SIZE, + AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT); SMU_TABLE_INIT(tables, SMU_TABLE_I2C_COMMANDS, sizeof(SwI2cRequest_t), - PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM); + PAGE_SIZE, + AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT); smu_table->metrics_table = kzalloc(sizeof(MetricsTable_t), GFP_KERNEL); if (!smu_table->metrics_table) @@ -355,6 +357,9 @@ static int smu_v13_0_6_setup_driver_pptable(struct smu_context *smu) SMUQ10_TO_UINT(metrics->LclkFrequencyTable[i]); } + /* use AID0 serial number by default */ + pptable->PublicSerialNumber_AID = metrics->PublicSerialNumber_AID[0]; + pptable->Init = true; } @@ -385,7 +390,7 @@ static int smu_v13_0_6_get_dpm_ultimate_freq(struct smu_context *smu, break; case SMU_SOCCLK: if (pptable->Init) - clock_limit = pptable->UclkFrequencyTable[0]; + clock_limit = pptable->SocclkFrequencyTable[0]; break; case SMU_FCLK: if (pptable->Init) @@ -638,16 +643,14 @@ static int smu_v13_0_6_freqs_in_same_level(int32_t frequency1, return (abs(frequency1 - frequency2) <= EPSILON); } -static uint32_t smu_v13_0_6_get_throttler_status(struct smu_context *smu, - MetricsTable_t *metrics) +static uint32_t smu_v13_0_6_get_throttler_status(struct smu_context *smu) { + struct smu_power_context *smu_power = &smu->smu_power; + struct smu_13_0_power_context *power_context = smu_power->power_context; uint32_t throttler_status = 0; - throttler_status |= metrics->ProchotResidencyAcc > 0 ? 1U << THROTTLER_PROCHOT_GFX_BIT : 0; - throttler_status |= metrics->PptResidencyAcc > 0 ? 1U << THROTTLER_PPT_BIT : 0; - throttler_status |= metrics->SocketThmResidencyAcc > 0 ? 1U << THROTTLER_TEMP_SOC_BIT : 0; - throttler_status |= metrics->VrThmResidencyAcc > 0 ? 1U << THROTTLER_TEMP_VR_GFX_BIT : 0; - throttler_status |= metrics->HbmThmResidencyAcc > 0 ? 1U << THROTTLER_TEMP_HBM_BIT : 0; + throttler_status = atomic_read(&power_context->throttle_status); + dev_dbg(smu->adev->dev, "SMU Throttler status: %u", throttler_status); return throttler_status; } @@ -658,7 +661,10 @@ static int smu_v13_0_6_get_smu_metrics_data(struct smu_context *smu, { struct smu_table_context *smu_table = &smu->smu_table; MetricsTable_t *metrics = (MetricsTable_t *)smu_table->metrics_table; + struct amdgpu_device *adev = smu->adev; + uint32_t smu_version; int ret = 0; + int xcc_id; ret = smu_v13_0_6_get_metrics_table(smu, NULL, false); if (ret) @@ -668,7 +674,13 @@ static int smu_v13_0_6_get_smu_metrics_data(struct smu_context *smu, switch (member) { case METRICS_CURR_GFXCLK: case METRICS_AVERAGE_GFXCLK: - *value = 0; + smu_cmn_get_smc_version(smu, NULL, &smu_version); + if (smu_version >= 0x552F00) { + xcc_id = GET_INST(GC, 0); + *value = SMUQ10_TO_UINT(metrics->GfxclkFrequency[xcc_id]); + } else { + *value = 0; + } break; case METRICS_CURR_SOCCLK: case METRICS_AVERAGE_SOCCLK: @@ -708,9 +720,6 @@ static int smu_v13_0_6_get_smu_metrics_data(struct smu_context *smu, case METRICS_TEMPERATURE_VRSOC: *value = SMUQ10_TO_UINT(metrics->MaxVrTemperature); break; - case METRICS_THROTTLER_STATUS: - *value = smu_v13_0_6_get_throttler_status(smu, metrics); - break; default: *value = UINT_MAX; break; @@ -1246,21 +1255,6 @@ static int smu_v13_0_6_get_power_limit(struct smu_context *smu, uint32_t power_limit = 0; int ret; - if (!smu_cmn_feature_is_enabled(smu, SMU_FEATURE_PPT_BIT)) { - if (current_power_limit) - *current_power_limit = 0; - if (default_power_limit) - *default_power_limit = 0; - if (max_power_limit) - *max_power_limit = 0; - - dev_warn( - smu->adev->dev, - "PPT feature is not enabled, power values can't be fetched."); - - return 0; - } - ret = smu_cmn_send_smc_msg(smu, SMU_MSG_GetPptLimit, &power_limit); if (ret) { @@ -1287,16 +1281,147 @@ static int smu_v13_0_6_set_power_limit(struct smu_context *smu, return smu_v13_0_set_power_limit(smu, limit_type, limit); } -static int smu_v13_0_6_system_features_control(struct smu_context *smu, - bool enable) +static int smu_v13_0_6_irq_process(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + struct amdgpu_iv_entry *entry) { - int ret; + struct smu_context *smu = adev->powerplay.pp_handle; + struct smu_power_context *smu_power = &smu->smu_power; + struct smu_13_0_power_context *power_context = smu_power->power_context; + uint32_t client_id = entry->client_id; + uint32_t ctxid = entry->src_data[0]; + uint32_t src_id = entry->src_id; + uint32_t data; + + if (client_id == SOC15_IH_CLIENTID_MP1) { + if (src_id == IH_INTERRUPT_ID_TO_DRIVER) { + /* ACK SMUToHost interrupt */ + data = RREG32_SOC15(MP1, 0, regMP1_SMN_IH_SW_INT_CTRL); + data = REG_SET_FIELD(data, MP1_SMN_IH_SW_INT_CTRL, INT_ACK, 1); + WREG32_SOC15(MP1, 0, regMP1_SMN_IH_SW_INT_CTRL, data); + /* + * ctxid is used to distinguish different events for SMCToHost + * interrupt. + */ + switch (ctxid) { + case IH_INTERRUPT_CONTEXT_ID_THERMAL_THROTTLING: + /* + * Increment the throttle interrupt counter + */ + atomic64_inc(&smu->throttle_int_counter); + + if (!atomic_read(&adev->throttling_logging_enabled)) + return 0; + + /* This uses the new method which fixes the + * incorrect throttling status reporting + * through metrics table. For older FWs, + * it will be ignored. + */ + if (__ratelimit(&adev->throttling_logging_rs)) { + atomic_set( + &power_context->throttle_status, + entry->src_data[1]); + schedule_work(&smu->throttling_logging_work); + } + + break; + } + } + } + + return 0; +} + +static int smu_v13_0_6_set_irq_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned tyep, + enum amdgpu_interrupt_state state) +{ + uint32_t val = 0; + + switch (state) { + case AMDGPU_IRQ_STATE_DISABLE: + /* For MP1 SW irqs */ + val = RREG32_SOC15(MP1, 0, regMP1_SMN_IH_SW_INT_CTRL); + val = REG_SET_FIELD(val, MP1_SMN_IH_SW_INT_CTRL, INT_MASK, 1); + WREG32_SOC15(MP1, 0, regMP1_SMN_IH_SW_INT_CTRL, val); - /* Nothing to be done for APU */ - if (smu->adev->flags & AMD_IS_APU) + break; + case AMDGPU_IRQ_STATE_ENABLE: + /* For MP1 SW irqs */ + val = RREG32_SOC15(MP1, 0, regMP1_SMN_IH_SW_INT); + val = REG_SET_FIELD(val, MP1_SMN_IH_SW_INT, ID, 0xFE); + val = REG_SET_FIELD(val, MP1_SMN_IH_SW_INT, VALID, 0); + WREG32_SOC15(MP1, 0, regMP1_SMN_IH_SW_INT, val); + + val = RREG32_SOC15(MP1, 0, regMP1_SMN_IH_SW_INT_CTRL); + val = REG_SET_FIELD(val, MP1_SMN_IH_SW_INT_CTRL, INT_MASK, 0); + WREG32_SOC15(MP1, 0, regMP1_SMN_IH_SW_INT_CTRL, val); + + break; + default: + break; + } + + return 0; +} + +static const struct amdgpu_irq_src_funcs smu_v13_0_6_irq_funcs = +{ + .set = smu_v13_0_6_set_irq_state, + .process = smu_v13_0_6_irq_process, +}; + +static int smu_v13_0_6_register_irq_handler(struct smu_context *smu) +{ + struct amdgpu_device *adev = smu->adev; + struct amdgpu_irq_src *irq_src = &smu->irq_source; + int ret = 0; + + if (amdgpu_sriov_vf(adev)) return 0; - ret = smu_v13_0_system_features_control(smu, enable); + irq_src->num_types = 1; + irq_src->funcs = &smu_v13_0_6_irq_funcs; + + ret = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_MP1, + IH_INTERRUPT_ID_TO_DRIVER, + irq_src); + if (ret) + return ret; + + return ret; +} + +static int smu_v13_0_6_notify_unload(struct smu_context *smu) +{ + uint32_t smu_version; + + smu_cmn_get_smc_version(smu, NULL, &smu_version); + if (smu_version <= 0x553500) + return 0; + + dev_dbg(smu->adev->dev, "Notify PMFW about driver unload"); + /* Ignore return, just intimate FW that driver is not going to be there */ + smu_cmn_send_smc_msg(smu, SMU_MSG_PrepareMp1ForUnload, NULL); + + return 0; +} + +static int smu_v13_0_6_system_features_control(struct smu_context *smu, + bool enable) +{ + struct amdgpu_device *adev = smu->adev; + int ret = 0; + + if (enable) { + if (!(adev->flags & AMD_IS_APU)) + ret = smu_v13_0_system_features_control(smu, enable); + } else { + /* Notify FW that the device is no longer driver managed */ + smu_v13_0_6_notify_unload(smu); + } return ret; } @@ -1737,19 +1862,11 @@ static void smu_v13_0_6_i2c_control_fini(struct smu_context *smu) static void smu_v13_0_6_get_unique_id(struct smu_context *smu) { struct amdgpu_device *adev = smu->adev; - //SmuMetrics_t *metrics = smu->smu_table.metrics_table; - uint32_t upper32 = 0, lower32 = 0; - int ret; - - ret = smu_cmn_get_metrics_table(smu, NULL, false); - if (ret) - goto out; - - //upper32 = metrics->PublicSerialNumUpper32; - //lower32 = metrics->PublicSerialNumLower32; + struct smu_table_context *smu_table = &smu->smu_table; + struct PPTable_t *pptable = + (struct PPTable_t *)smu_table->driver_pptable; -out: - adev->unique_id = ((uint64_t)upper32 << 32) | lower32; + adev->unique_id = pptable->PublicSerialNumber_AID; if (adev->serial[0] == '\0') sprintf(adev->serial, "%016llx", adev->unique_id); } @@ -1774,37 +1891,35 @@ static int smu_v13_0_6_allow_xgmi_power_down(struct smu_context *smu, bool en) en ? 0 : 1, NULL); } -static const struct throttling_logging_label { - uint32_t feature_mask; - const char *label; -} logging_label[] = { - { (1U << THROTTLER_TEMP_HBM_BIT), "HBM" }, - { (1U << THROTTLER_TEMP_SOC_BIT), "SOC" }, - { (1U << THROTTLER_TEMP_VR_GFX_BIT), "VR limit" }, +static const char *const throttling_logging_label[] = { + [THROTTLER_PROCHOT_BIT] = "Prochot", + [THROTTLER_PPT_BIT] = "PPT", + [THROTTLER_THERMAL_SOCKET_BIT] = "SOC", + [THROTTLER_THERMAL_VR_BIT] = "VR", + [THROTTLER_THERMAL_HBM_BIT] = "HBM" }; + static void smu_v13_0_6_log_thermal_throttling_event(struct smu_context *smu) { - int ret; int throttler_idx, throtting_events = 0, buf_idx = 0; struct amdgpu_device *adev = smu->adev; uint32_t throttler_status; char log_buf[256]; - ret = smu_v13_0_6_get_smu_metrics_data(smu, METRICS_THROTTLER_STATUS, - &throttler_status); - if (ret) + throttler_status = smu_v13_0_6_get_throttler_status(smu); + if (!throttler_status) return; memset(log_buf, 0, sizeof(log_buf)); - for (throttler_idx = 0; throttler_idx < ARRAY_SIZE(logging_label); + for (throttler_idx = 0; + throttler_idx < ARRAY_SIZE(throttling_logging_label); throttler_idx++) { - if (throttler_status & - logging_label[throttler_idx].feature_mask) { + if (throttler_status & (1U << throttler_idx)) { throtting_events++; - buf_idx += snprintf(log_buf + buf_idx, - sizeof(log_buf) - buf_idx, "%s%s", - throtting_events > 1 ? " and " : "", - logging_label[throttler_idx].label); + buf_idx += snprintf( + log_buf + buf_idx, sizeof(log_buf) - buf_idx, + "%s%s", throtting_events > 1 ? " and " : "", + throttling_logging_label[throttler_idx]); if (buf_idx >= sizeof(log_buf)) { dev_err(adev->dev, "buffer overflow!\n"); log_buf[sizeof(log_buf) - 1] = '\0'; @@ -1813,16 +1928,24 @@ static void smu_v13_0_6_log_thermal_throttling_event(struct smu_context *smu) } } - dev_warn( - adev->dev, - "WARN: GPU thermal throttling temperature reached, expect performance decrease. %s.\n", - log_buf); + dev_warn(adev->dev, + "WARN: GPU is throttled, expect performance decrease. %s.\n", + log_buf); kgd2kfd_smi_event_throttle( smu->adev->kfd.dev, smu_cmn_get_indep_throttler_status(throttler_status, smu_v13_0_6_throttler_map)); } +static int +smu_v13_0_6_get_current_pcie_link_width_level(struct smu_context *smu) +{ + struct amdgpu_device *adev = smu->adev; + + return REG_GET_FIELD(RREG32_PCIE(smnPCIE_LC_LINK_WIDTH_CNTL), + PCIE_LC_LINK_WIDTH_CNTL, LC_LINK_WIDTH_RD); +} + static int smu_v13_0_6_get_current_pcie_link_speed(struct smu_context *smu) { struct amdgpu_device *adev = smu->adev; @@ -1841,8 +1964,12 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table struct smu_table_context *smu_table = &smu->smu_table; struct gpu_metrics_v1_3 *gpu_metrics = (struct gpu_metrics_v1_3 *)smu_table->gpu_metrics_table; + struct amdgpu_device *adev = smu->adev; + int ret = 0, inst0, xcc0; MetricsTable_t *metrics; - int i, ret = 0; + + inst0 = adev->sdma.instance[0].aid_id; + xcc0 = GET_INST(GC, 0); metrics = kzalloc(sizeof(MetricsTable_t), GFP_KERNEL); ret = smu_v13_0_6_get_metrics_table(smu, metrics, true); @@ -1851,51 +1978,59 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table smu_cmn_init_soft_gpu_metrics(gpu_metrics, 1, 3); - /* TODO: Decide on how to fill in zero value fields */ - gpu_metrics->temperature_edge = 0; - gpu_metrics->temperature_hotspot = 0; - gpu_metrics->temperature_mem = 0; - gpu_metrics->temperature_vrgfx = 0; - gpu_metrics->temperature_vrsoc = 0; - gpu_metrics->temperature_vrmem = 0; - - gpu_metrics->average_gfx_activity = 0; - gpu_metrics->average_umc_activity = 0; - gpu_metrics->average_mm_activity = 0; - - gpu_metrics->average_socket_power = 0; - gpu_metrics->energy_accumulator = 0; - - gpu_metrics->average_gfxclk_frequency = 0; - gpu_metrics->average_socclk_frequency = 0; - gpu_metrics->average_uclk_frequency = 0; - gpu_metrics->average_vclk0_frequency = 0; - gpu_metrics->average_dclk0_frequency = 0; - - gpu_metrics->current_gfxclk = 0; - gpu_metrics->current_socclk = 0; - gpu_metrics->current_uclk = 0; - gpu_metrics->current_vclk0 = 0; - gpu_metrics->current_dclk0 = 0; - + gpu_metrics->temperature_hotspot = + SMUQ10_TO_UINT(metrics->MaxSocketTemperature); + /* Individual HBM stack temperature is not reported */ + gpu_metrics->temperature_mem = + SMUQ10_TO_UINT(metrics->MaxHbmTemperature); + /* Reports max temperature of all voltage rails */ + gpu_metrics->temperature_vrsoc = + SMUQ10_TO_UINT(metrics->MaxVrTemperature); + + gpu_metrics->average_gfx_activity = + SMUQ10_TO_UINT(metrics->SocketGfxBusy); + gpu_metrics->average_umc_activity = + SMUQ10_TO_UINT(metrics->DramBandwidthUtilization); + + gpu_metrics->average_socket_power = + SMUQ10_TO_UINT(metrics->SocketPower); + gpu_metrics->energy_accumulator = + SMUQ16_TO_UINT(metrics->SocketEnergyAcc); + + gpu_metrics->current_gfxclk = + SMUQ10_TO_UINT(metrics->GfxclkFrequency[xcc0]); + gpu_metrics->current_socclk = + SMUQ10_TO_UINT(metrics->SocclkFrequency[inst0]); + gpu_metrics->current_uclk = SMUQ10_TO_UINT(metrics->UclkFrequency); + gpu_metrics->current_vclk0 = + SMUQ10_TO_UINT(metrics->VclkFrequency[inst0]); + gpu_metrics->current_dclk0 = + SMUQ10_TO_UINT(metrics->DclkFrequency[inst0]); + + gpu_metrics->average_gfxclk_frequency = gpu_metrics->current_gfxclk; + gpu_metrics->average_socclk_frequency = gpu_metrics->current_socclk; + gpu_metrics->average_uclk_frequency = gpu_metrics->current_uclk; + gpu_metrics->average_vclk0_frequency = gpu_metrics->current_vclk0; + gpu_metrics->average_dclk0_frequency = gpu_metrics->current_dclk0; + + /* Throttle status is not reported through metrics now */ gpu_metrics->throttle_status = 0; - gpu_metrics->indep_throttle_status = smu_cmn_get_indep_throttler_status( - gpu_metrics->throttle_status, smu_v13_0_6_throttler_map); - gpu_metrics->current_fan_speed = 0; - - gpu_metrics->pcie_link_width = 0; - gpu_metrics->pcie_link_speed = smu_v13_0_6_get_current_pcie_link_speed(smu); + if (!(adev->flags & AMD_IS_APU)) { + gpu_metrics->pcie_link_width = + smu_v13_0_6_get_current_pcie_link_width_level(smu); + gpu_metrics->pcie_link_speed = + smu_v13_0_6_get_current_pcie_link_speed(smu); + } gpu_metrics->system_clock_counter = ktime_get_boottime_ns(); - gpu_metrics->gfx_activity_acc = 0; - gpu_metrics->mem_activity_acc = 0; + gpu_metrics->gfx_activity_acc = + SMUQ10_TO_UINT(metrics->SocketGfxBusyAcc); + gpu_metrics->mem_activity_acc = + SMUQ10_TO_UINT(metrics->DramBandwidthUtilizationAcc); - for (i = 0; i < NUM_HBM_INSTANCES; i++) - gpu_metrics->temperature_hbm[i] = 0; - - gpu_metrics->firmware_timestamp = 0; + gpu_metrics->firmware_timestamp = metrics->Timestamp; *table = (void *)gpu_metrics; kfree(metrics); @@ -1905,27 +2040,27 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table static int smu_v13_0_6_mode2_reset(struct smu_context *smu) { - u32 smu_version; int ret = 0, index; struct amdgpu_device *adev = smu->adev; int timeout = 10; - smu_cmn_get_smc_version(smu, NULL, &smu_version); - index = smu_cmn_to_asic_specific_index(smu, CMN2ASIC_MAPPING_MSG, SMU_MSG_GfxDeviceDriverReset); mutex_lock(&smu->message_lock); + ret = smu_cmn_send_msg_without_waiting(smu, (uint16_t)index, SMU_RESET_MODE_2); + /* This is similar to FLR, wait till max FLR timeout */ msleep(100); + dev_dbg(smu->adev->dev, "restore config space...\n"); /* Restore the config space saved during init */ amdgpu_device_load_pci_state(adev->pdev); dev_dbg(smu->adev->dev, "wait for reset ack\n"); - while (ret == -ETIME && timeout) { + do { ret = smu_cmn_wait_for_response(smu); /* Wait a bit more time for getting ACK */ if (ret == -ETIME) { @@ -1934,16 +2069,14 @@ static int smu_v13_0_6_mode2_reset(struct smu_context *smu) continue; } - if (ret != 1) { + if (ret) { dev_err(adev->dev, - "failed to send mode2 message \tparam: 0x%08x response %#x\n", + "failed to send mode2 message \tparam: 0x%08x error code %d\n", SMU_RESET_MODE_2, ret); goto out; } - } + } while (ret == -ETIME && timeout); - if (ret == 1) - ret = 0; out: mutex_unlock(&smu->message_lock); @@ -2032,11 +2165,9 @@ static const struct pptable_funcs smu_v13_0_6_ppt_funcs = { .feature_is_enabled = smu_cmn_feature_is_enabled, .set_power_limit = smu_v13_0_6_set_power_limit, .set_xgmi_pstate = smu_v13_0_set_xgmi_pstate, - /* TODO: Thermal limits unknown, skip these for now - .register_irq_handler = smu_v13_0_register_irq_handler, + .register_irq_handler = smu_v13_0_6_register_irq_handler, .enable_thermal_alert = smu_v13_0_enable_thermal_alert, .disable_thermal_alert = smu_v13_0_disable_thermal_alert, - */ .setup_pptable = smu_v13_0_6_setup_pptable, .baco_is_support = smu_v13_0_6_is_baco_supported, .get_dpm_ultimate_freq = smu_v13_0_6_get_dpm_ultimate_freq, @@ -2065,5 +2196,6 @@ void smu_v13_0_6_set_ppt_funcs(struct smu_context *smu) smu->clock_map = smu_v13_0_6_clk_map; smu->feature_map = smu_v13_0_6_feature_mask_map; smu->table_map = smu_v13_0_6_table_map; + smu->smc_driver_if_version = SMU13_0_6_DRIVER_IF_VERSION; smu_v13_0_set_smu_mailbox_registers(smu); } diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c index 3d9ff46706fb..cda4e818aab7 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c @@ -125,6 +125,7 @@ static struct cmn2asic_msg_mapping smu_v13_0_7_message_map[SMU_MSG_MAX_COUNT] = MSG_MAP(ArmD3, PPSMC_MSG_ArmD3, 0), MSG_MAP(AllowGpo, PPSMC_MSG_SetGpoAllow, 0), MSG_MAP(GetPptLimit, PPSMC_MSG_GetPptLimit, 0), + MSG_MAP(NotifyPowerSource, PPSMC_MSG_NotifyPowerSource, 0), }; static struct cmn2asic_mapping smu_v13_0_7_clk_map[SMU_CLK_COUNT] = { @@ -205,6 +206,7 @@ static struct cmn2asic_mapping smu_v13_0_7_table_map[SMU_TABLE_COUNT] = { TAB_MAP(DRIVER_SMU_CONFIG), TAB_MAP(ACTIVITY_MONITOR_COEFF), [SMU_TABLE_COMBO_PPTABLE] = {1, TABLE_COMBO_PPTABLE}, + TAB_MAP(OVERDRIVE), }; static struct cmn2asic_mapping smu_v13_0_7_pwr_src_map[SMU_POWER_SOURCE_COUNT] = { @@ -321,6 +323,10 @@ static int smu_v13_0_7_check_powerplay_table(struct smu_context *smu) struct smu_baco_context *smu_baco = &smu->smu_baco; PPTable_t *smc_pptable = table_context->driver_pptable; BoardTable_t *BoardTable = &smc_pptable->BoardTable; + const OverDriveLimits_t * const overdrive_upperlimits = + &smc_pptable->SkuTable.OverDriveLimitsBasicMax; + const OverDriveLimits_t * const overdrive_lowerlimits = + &smc_pptable->SkuTable.OverDriveLimitsMin; if (powerplay_table->platform_caps & SMU_13_0_7_PP_PLATFORM_CAP_HARDWAREDC) smu->dc_controlled_by_gpio = true; @@ -332,6 +338,10 @@ static int smu_v13_0_7_check_powerplay_table(struct smu_context *smu) if (smu_baco->platform_support && (BoardTable->HsrEnabled || BoardTable->VddqOffEnabled)) smu_baco->maco_support = true; + if (!overdrive_lowerlimits->FeatureCtrlMask || + !overdrive_upperlimits->FeatureCtrlMask) + smu->od_enabled = false; + table_context->thermal_controller_type = powerplay_table->thermal_controller_type; @@ -478,7 +488,7 @@ static int smu_v13_0_7_tables_init(struct smu_context *smu) PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM); SMU_TABLE_INIT(tables, SMU_TABLE_I2C_COMMANDS, sizeof(SwI2cRequest_t), PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM); - SMU_TABLE_INIT(tables, SMU_TABLE_OVERDRIVE, sizeof(OverDriveTable_t), + SMU_TABLE_INIT(tables, SMU_TABLE_OVERDRIVE, sizeof(OverDriveTableExternal_t), PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM); SMU_TABLE_INIT(tables, SMU_TABLE_PMSTATUSLOG, SMU13_TOOL_SIZE, PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM); @@ -1012,16 +1022,118 @@ static int smu_v13_0_7_get_current_clk_freq_by_table(struct smu_context *smu, value); } +static bool smu_v13_0_7_is_od_feature_supported(struct smu_context *smu, + int od_feature_bit) +{ + PPTable_t *pptable = smu->smu_table.driver_pptable; + const OverDriveLimits_t * const overdrive_upperlimits = + &pptable->SkuTable.OverDriveLimitsBasicMax; + + return overdrive_upperlimits->FeatureCtrlMask & (1U << od_feature_bit); +} + +static void smu_v13_0_7_get_od_setting_limits(struct smu_context *smu, + int od_feature_bit, + bool lower_boundary, + int32_t *min, + int32_t *max) +{ + PPTable_t *pptable = smu->smu_table.driver_pptable; + const OverDriveLimits_t * const overdrive_upperlimits = + &pptable->SkuTable.OverDriveLimitsBasicMax; + const OverDriveLimits_t * const overdrive_lowerlimits = + &pptable->SkuTable.OverDriveLimitsMin; + int32_t od_min_setting, od_max_setting; + + switch (od_feature_bit) { + case PP_OD_FEATURE_GFXCLK_BIT: + if (lower_boundary) { + od_min_setting = overdrive_lowerlimits->GfxclkFmin; + od_max_setting = overdrive_upperlimits->GfxclkFmin; + } else { + od_min_setting = overdrive_lowerlimits->GfxclkFmax; + od_max_setting = overdrive_upperlimits->GfxclkFmax; + } + break; + case PP_OD_FEATURE_UCLK_BIT: + if (lower_boundary) { + od_min_setting = overdrive_lowerlimits->UclkFmin; + od_max_setting = overdrive_upperlimits->UclkFmin; + } else { + od_min_setting = overdrive_lowerlimits->UclkFmax; + od_max_setting = overdrive_upperlimits->UclkFmax; + } + break; + case PP_OD_FEATURE_GFX_VF_CURVE_BIT: + od_min_setting = overdrive_lowerlimits->VoltageOffsetPerZoneBoundary; + od_max_setting = overdrive_upperlimits->VoltageOffsetPerZoneBoundary; + break; + default: + break; + } + + if (min) + *min = od_min_setting; + if (max) + *max = od_max_setting; +} + +static void smu_v13_0_7_dump_od_table(struct smu_context *smu, + OverDriveTableExternal_t *od_table) +{ + struct amdgpu_device *adev = smu->adev; + + dev_dbg(adev->dev, "OD: Gfxclk: (%d, %d)\n", od_table->OverDriveTable.GfxclkFmin, + od_table->OverDriveTable.GfxclkFmax); + dev_dbg(adev->dev, "OD: Uclk: (%d, %d)\n", od_table->OverDriveTable.UclkFmin, + od_table->OverDriveTable.UclkFmax); +} + +static int smu_v13_0_7_get_overdrive_table(struct smu_context *smu, + OverDriveTableExternal_t *od_table) +{ + int ret = 0; + + ret = smu_cmn_update_table(smu, + SMU_TABLE_OVERDRIVE, + 0, + (void *)od_table, + false); + if (ret) + dev_err(smu->adev->dev, "Failed to get overdrive table!\n"); + + return ret; +} + +static int smu_v13_0_7_upload_overdrive_table(struct smu_context *smu, + OverDriveTableExternal_t *od_table) +{ + int ret = 0; + + ret = smu_cmn_update_table(smu, + SMU_TABLE_OVERDRIVE, + 0, + (void *)od_table, + true); + if (ret) + dev_err(smu->adev->dev, "Failed to upload overdrive table!\n"); + + return ret; +} + static int smu_v13_0_7_print_clk_levels(struct smu_context *smu, enum smu_clk_type clk_type, char *buf) { struct smu_dpm_context *smu_dpm = &smu->smu_dpm; struct smu_13_0_dpm_context *dpm_context = smu_dpm->dpm_context; + OverDriveTableExternal_t *od_table = + (OverDriveTableExternal_t *)smu->smu_table.overdrive_table; struct smu_13_0_dpm_table *single_dpm_table; struct smu_13_0_pcie_table *pcie_table; uint32_t gen_speed, lane_width; int i, curr_freq, size = 0; + int32_t min_value, max_value; int ret = 0; smu_cmn_get_sysfs_buf(&buf, &size); @@ -1138,6 +1250,89 @@ static int smu_v13_0_7_print_clk_levels(struct smu_context *smu, "*" : ""); break; + case SMU_OD_SCLK: + if (!smu_v13_0_7_is_od_feature_supported(smu, + PP_OD_FEATURE_GFXCLK_BIT)) + break; + + size += sysfs_emit_at(buf, size, "OD_SCLK:\n"); + size += sysfs_emit_at(buf, size, "0: %uMhz\n1: %uMhz\n", + od_table->OverDriveTable.GfxclkFmin, + od_table->OverDriveTable.GfxclkFmax); + break; + + case SMU_OD_MCLK: + if (!smu_v13_0_7_is_od_feature_supported(smu, + PP_OD_FEATURE_UCLK_BIT)) + break; + + size += sysfs_emit_at(buf, size, "OD_MCLK:\n"); + size += sysfs_emit_at(buf, size, "0: %uMhz\n1: %uMHz\n", + od_table->OverDriveTable.UclkFmin, + od_table->OverDriveTable.UclkFmax); + break; + + case SMU_OD_VDDC_CURVE: + if (!smu_v13_0_7_is_od_feature_supported(smu, + PP_OD_FEATURE_GFX_VF_CURVE_BIT)) + break; + + size += sysfs_emit_at(buf, size, "OD_VDDC_CURVE:\n"); + for (i = 0; i < PP_NUM_OD_VF_CURVE_POINTS; i++) + size += sysfs_emit_at(buf, size, "%d: %dmv\n", + i, + od_table->OverDriveTable.VoltageOffsetPerZoneBoundary[i]); + break; + + case SMU_OD_RANGE: + if (!smu_v13_0_7_is_od_feature_supported(smu, PP_OD_FEATURE_GFXCLK_BIT) && + !smu_v13_0_7_is_od_feature_supported(smu, PP_OD_FEATURE_UCLK_BIT) && + !smu_v13_0_7_is_od_feature_supported(smu, PP_OD_FEATURE_GFX_VF_CURVE_BIT)) + break; + + size += sysfs_emit_at(buf, size, "%s:\n", "OD_RANGE"); + + if (smu_v13_0_7_is_od_feature_supported(smu, PP_OD_FEATURE_GFXCLK_BIT)) { + smu_v13_0_7_get_od_setting_limits(smu, + PP_OD_FEATURE_GFXCLK_BIT, + true, + &min_value, + NULL); + smu_v13_0_7_get_od_setting_limits(smu, + PP_OD_FEATURE_GFXCLK_BIT, + false, + NULL, + &max_value); + size += sysfs_emit_at(buf, size, "SCLK: %7uMhz %10uMhz\n", + min_value, max_value); + } + + if (smu_v13_0_7_is_od_feature_supported(smu, PP_OD_FEATURE_UCLK_BIT)) { + smu_v13_0_7_get_od_setting_limits(smu, + PP_OD_FEATURE_UCLK_BIT, + true, + &min_value, + NULL); + smu_v13_0_7_get_od_setting_limits(smu, + PP_OD_FEATURE_UCLK_BIT, + false, + NULL, + &max_value); + size += sysfs_emit_at(buf, size, "MCLK: %7uMhz %10uMhz\n", + min_value, max_value); + } + + if (smu_v13_0_7_is_od_feature_supported(smu, PP_OD_FEATURE_GFX_VF_CURVE_BIT)) { + smu_v13_0_7_get_od_setting_limits(smu, + PP_OD_FEATURE_GFX_VF_CURVE_BIT, + true, + &min_value, + &max_value); + size += sysfs_emit_at(buf, size, "VDDC_CURVE: %7dmv %10dmv\n", + min_value, max_value); + } + break; + default: break; } @@ -1145,6 +1340,222 @@ static int smu_v13_0_7_print_clk_levels(struct smu_context *smu, return size; } +static int smu_v13_0_7_od_edit_dpm_table(struct smu_context *smu, + enum PP_OD_DPM_TABLE_COMMAND type, + long input[], + uint32_t size) +{ + struct smu_table_context *table_context = &smu->smu_table; + OverDriveTableExternal_t *od_table = + (OverDriveTableExternal_t *)table_context->overdrive_table; + struct amdgpu_device *adev = smu->adev; + uint32_t offset_of_featurectrlmask; + int32_t minimum, maximum; + uint32_t feature_ctrlmask; + int i, ret = 0; + + switch (type) { + case PP_OD_EDIT_SCLK_VDDC_TABLE: + if (!smu_v13_0_7_is_od_feature_supported(smu, PP_OD_FEATURE_GFXCLK_BIT)) { + dev_warn(adev->dev, "GFXCLK_LIMITS setting not supported!\n"); + return -ENOTSUPP; + } + + for (i = 0; i < size; i += 2) { + if (i + 2 > size) { + dev_info(adev->dev, "invalid number of input parameters %d\n", size); + return -EINVAL; + } + + switch (input[i]) { + case 0: + smu_v13_0_7_get_od_setting_limits(smu, + PP_OD_FEATURE_GFXCLK_BIT, + true, + &minimum, + &maximum); + if (input[i + 1] < minimum || + input[i + 1] > maximum) { + dev_info(adev->dev, "GfxclkFmin (%ld) must be within [%u, %u]!\n", + input[i + 1], minimum, maximum); + return -EINVAL; + } + + od_table->OverDriveTable.GfxclkFmin = input[i + 1]; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_GFXCLK_BIT; + break; + + case 1: + smu_v13_0_7_get_od_setting_limits(smu, + PP_OD_FEATURE_GFXCLK_BIT, + false, + &minimum, + &maximum); + if (input[i + 1] < minimum || + input[i + 1] > maximum) { + dev_info(adev->dev, "GfxclkFmax (%ld) must be within [%u, %u]!\n", + input[i + 1], minimum, maximum); + return -EINVAL; + } + + od_table->OverDriveTable.GfxclkFmax = input[i + 1]; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_GFXCLK_BIT; + break; + + default: + dev_info(adev->dev, "Invalid SCLK_VDDC_TABLE index: %ld\n", input[i]); + dev_info(adev->dev, "Supported indices: [0:min,1:max]\n"); + return -EINVAL; + } + } + + if (od_table->OverDriveTable.GfxclkFmin > od_table->OverDriveTable.GfxclkFmax) { + dev_err(adev->dev, + "Invalid setting: GfxclkFmin(%u) is bigger than GfxclkFmax(%u)\n", + (uint32_t)od_table->OverDriveTable.GfxclkFmin, + (uint32_t)od_table->OverDriveTable.GfxclkFmax); + return -EINVAL; + } + break; + + case PP_OD_EDIT_MCLK_VDDC_TABLE: + if (!smu_v13_0_7_is_od_feature_supported(smu, PP_OD_FEATURE_UCLK_BIT)) { + dev_warn(adev->dev, "UCLK_LIMITS setting not supported!\n"); + return -ENOTSUPP; + } + + for (i = 0; i < size; i += 2) { + if (i + 2 > size) { + dev_info(adev->dev, "invalid number of input parameters %d\n", size); + return -EINVAL; + } + + switch (input[i]) { + case 0: + smu_v13_0_7_get_od_setting_limits(smu, + PP_OD_FEATURE_UCLK_BIT, + true, + &minimum, + &maximum); + if (input[i + 1] < minimum || + input[i + 1] > maximum) { + dev_info(adev->dev, "UclkFmin (%ld) must be within [%u, %u]!\n", + input[i + 1], minimum, maximum); + return -EINVAL; + } + + od_table->OverDriveTable.UclkFmin = input[i + 1]; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_UCLK_BIT; + break; + + case 1: + smu_v13_0_7_get_od_setting_limits(smu, + PP_OD_FEATURE_UCLK_BIT, + false, + &minimum, + &maximum); + if (input[i + 1] < minimum || + input[i + 1] > maximum) { + dev_info(adev->dev, "UclkFmax (%ld) must be within [%u, %u]!\n", + input[i + 1], minimum, maximum); + return -EINVAL; + } + + od_table->OverDriveTable.UclkFmax = input[i + 1]; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_UCLK_BIT; + break; + + default: + dev_info(adev->dev, "Invalid MCLK_VDDC_TABLE index: %ld\n", input[i]); + dev_info(adev->dev, "Supported indices: [0:min,1:max]\n"); + return -EINVAL; + } + } + + if (od_table->OverDriveTable.UclkFmin > od_table->OverDriveTable.UclkFmax) { + dev_err(adev->dev, + "Invalid setting: UclkFmin(%u) is bigger than UclkFmax(%u)\n", + (uint32_t)od_table->OverDriveTable.UclkFmin, + (uint32_t)od_table->OverDriveTable.UclkFmax); + return -EINVAL; + } + break; + + case PP_OD_EDIT_VDDC_CURVE: + if (!smu_v13_0_7_is_od_feature_supported(smu, PP_OD_FEATURE_GFX_VF_CURVE_BIT)) { + dev_warn(adev->dev, "VF curve setting not supported!\n"); + return -ENOTSUPP; + } + + if (input[0] >= PP_NUM_OD_VF_CURVE_POINTS || + input[0] < 0) + return -EINVAL; + + smu_v13_0_7_get_od_setting_limits(smu, + PP_OD_FEATURE_GFX_VF_CURVE_BIT, + true, + &minimum, + &maximum); + if (input[1] < minimum || + input[1] > maximum) { + dev_info(adev->dev, "Voltage offset (%ld) must be within [%d, %d]!\n", + input[1], minimum, maximum); + return -EINVAL; + } + + od_table->OverDriveTable.VoltageOffsetPerZoneBoundary[input[0]] = input[1]; + od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_GFX_VF_CURVE_BIT; + break; + + case PP_OD_RESTORE_DEFAULT_TABLE: + feature_ctrlmask = od_table->OverDriveTable.FeatureCtrlMask; + memcpy(od_table, + table_context->boot_overdrive_table, + sizeof(OverDriveTableExternal_t)); + od_table->OverDriveTable.FeatureCtrlMask = feature_ctrlmask; + fallthrough; + + case PP_OD_COMMIT_DPM_TABLE: + /* + * The member below instructs PMFW the settings focused in + * this single operation. + * `uint32_t FeatureCtrlMask;` + * It does not contain actual informations about user's custom + * settings. Thus we do not cache it. + */ + offset_of_featurectrlmask = offsetof(OverDriveTable_t, FeatureCtrlMask); + if (memcmp((u8 *)od_table + offset_of_featurectrlmask, + table_context->user_overdrive_table + offset_of_featurectrlmask, + sizeof(OverDriveTableExternal_t) - offset_of_featurectrlmask)) { + smu_v13_0_7_dump_od_table(smu, od_table); + + ret = smu_v13_0_7_upload_overdrive_table(smu, od_table); + if (ret) { + dev_err(adev->dev, "Failed to upload overdrive table!\n"); + return ret; + } + + od_table->OverDriveTable.FeatureCtrlMask = 0; + memcpy(table_context->user_overdrive_table + offset_of_featurectrlmask, + (u8 *)od_table + offset_of_featurectrlmask, + sizeof(OverDriveTableExternal_t) - offset_of_featurectrlmask); + + if (!memcmp(table_context->user_overdrive_table, + table_context->boot_overdrive_table, + sizeof(OverDriveTableExternal_t))) + smu->user_dpm_profile.user_od = false; + else + smu->user_dpm_profile.user_od = true; + } + break; + + default: + return -ENOSYS; + } + + return ret; +} + static int smu_v13_0_7_force_clk_levels(struct smu_context *smu, enum smu_clk_type clk_type, uint32_t mask) @@ -1370,6 +1781,78 @@ static ssize_t smu_v13_0_7_get_gpu_metrics(struct smu_context *smu, return sizeof(struct gpu_metrics_v1_3); } +static int smu_v13_0_7_set_default_od_settings(struct smu_context *smu) +{ + OverDriveTableExternal_t *od_table = + (OverDriveTableExternal_t *)smu->smu_table.overdrive_table; + OverDriveTableExternal_t *boot_od_table = + (OverDriveTableExternal_t *)smu->smu_table.boot_overdrive_table; + OverDriveTableExternal_t *user_od_table = + (OverDriveTableExternal_t *)smu->smu_table.user_overdrive_table; + OverDriveTableExternal_t user_od_table_bak; + int ret = 0; + int i; + + ret = smu_v13_0_7_get_overdrive_table(smu, boot_od_table); + if (ret) + return ret; + + smu_v13_0_7_dump_od_table(smu, boot_od_table); + + memcpy(od_table, + boot_od_table, + sizeof(OverDriveTableExternal_t)); + + /* + * For S3/S4/Runpm resume, we need to setup those overdrive tables again, + * but we have to preserve user defined values in "user_od_table". + */ + if (!smu->adev->in_suspend) { + memcpy(user_od_table, + boot_od_table, + sizeof(OverDriveTableExternal_t)); + smu->user_dpm_profile.user_od = false; + } else if (smu->user_dpm_profile.user_od) { + memcpy(&user_od_table_bak, + user_od_table, + sizeof(OverDriveTableExternal_t)); + memcpy(user_od_table, + boot_od_table, + sizeof(OverDriveTableExternal_t)); + user_od_table->OverDriveTable.GfxclkFmin = + user_od_table_bak.OverDriveTable.GfxclkFmin; + user_od_table->OverDriveTable.GfxclkFmax = + user_od_table_bak.OverDriveTable.GfxclkFmax; + user_od_table->OverDriveTable.UclkFmin = + user_od_table_bak.OverDriveTable.UclkFmin; + user_od_table->OverDriveTable.UclkFmax = + user_od_table_bak.OverDriveTable.UclkFmax; + for (i = 0; i < PP_NUM_OD_VF_CURVE_POINTS; i++) + user_od_table->OverDriveTable.VoltageOffsetPerZoneBoundary[i] = + user_od_table_bak.OverDriveTable.VoltageOffsetPerZoneBoundary[i]; + } + + return 0; +} + +static int smu_v13_0_7_restore_user_od_settings(struct smu_context *smu) +{ + struct smu_table_context *table_context = &smu->smu_table; + OverDriveTableExternal_t *od_table = table_context->overdrive_table; + OverDriveTableExternal_t *user_od_table = table_context->user_overdrive_table; + int res; + + user_od_table->OverDriveTable.FeatureCtrlMask = 1U << PP_OD_FEATURE_GFXCLK_BIT | + 1U << PP_OD_FEATURE_UCLK_BIT | + 1U << PP_OD_FEATURE_GFX_VF_CURVE_BIT; + res = smu_v13_0_7_upload_overdrive_table(smu, user_od_table); + user_od_table->OverDriveTable.FeatureCtrlMask = 0; + if (res == 0) + memcpy(od_table, user_od_table, sizeof(OverDriveTableExternal_t)); + + return res; +} + static int smu_v13_0_7_populate_umd_state_clk(struct smu_context *smu) { struct smu_13_0_dpm_context *dpm_context = @@ -1759,6 +2242,9 @@ static const struct pptable_funcs smu_v13_0_7_ppt_funcs = { .notify_memory_pool_location = smu_v13_0_notify_memory_pool_location, .get_gpu_metrics = smu_v13_0_7_get_gpu_metrics, .set_soft_freq_limited_range = smu_v13_0_set_soft_freq_limited_range, + .set_default_od_settings = smu_v13_0_7_set_default_od_settings, + .restore_user_od_settings = smu_v13_0_7_restore_user_od_settings, + .od_edit_dpm_table = smu_v13_0_7_od_edit_dpm_table, .set_performance_level = smu_v13_0_set_performance_level, .gfx_off_control = smu_v13_0_gfx_off_control, .get_fan_speed_pwm = smu_v13_0_7_get_fan_speed_pwm, @@ -1770,6 +2256,7 @@ static const struct pptable_funcs smu_v13_0_7_ppt_funcs = { .enable_mgpu_fan_boost = smu_v13_0_7_enable_mgpu_fan_boost, .get_power_limit = smu_v13_0_7_get_power_limit, .set_power_limit = smu_v13_0_set_power_limit, + .set_power_source = smu_v13_0_set_power_source, .get_power_profile_mode = smu_v13_0_7_get_power_profile_mode, .set_power_profile_mode = smu_v13_0_7_set_power_profile_mode, .set_tool_table_location = smu_v13_0_set_tool_table_location, @@ -1796,5 +2283,6 @@ void smu_v13_0_7_set_ppt_funcs(struct smu_context *smu) smu->table_map = smu_v13_0_7_table_map; smu->pwr_src_map = smu_v13_0_7_pwr_src_map; smu->workload_map = smu_v13_0_7_workload_map; + smu->smc_driver_if_version = SMU13_0_7_DRIVER_IF_VERSION; smu_v13_0_set_smu_mailbox_registers(smu); } diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c index 04e56b0b3033..a1be2029ba4a 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c @@ -47,6 +47,14 @@ #define SMUIO_GFX_MISC_CNTL__PWR_GFXOFF_STATUS_MASK 0x00000006L #define SMUIO_GFX_MISC_CNTL__PWR_GFXOFF_STATUS__SHIFT 0x1L +#define SMU_13_0_8_UMD_PSTATE_GFXCLK 533 +#define SMU_13_0_8_UMD_PSTATE_SOCCLK 533 +#define SMU_13_0_8_UMD_PSTATE_FCLK 800 + +#define SMU_13_0_1_UMD_PSTATE_GFXCLK 700 +#define SMU_13_0_1_UMD_PSTATE_SOCCLK 678 +#define SMU_13_0_1_UMD_PSTATE_FCLK 1800 + #define FEATURE_MASK(feature) (1ULL << feature) #define SMC_DPM_FEATURE ( \ FEATURE_MASK(FEATURE_CCLK_DPM_BIT) | \ @@ -957,6 +965,9 @@ static int yellow_carp_set_soft_freq_limited_range(struct smu_context *smu, uint32_t max) { enum smu_message_type msg_set_min, msg_set_max; + uint32_t min_clk = min; + uint32_t max_clk = max; + int ret = 0; if (!yellow_carp_clk_dpm_is_enabled(smu, clk_type)) @@ -985,11 +996,17 @@ static int yellow_carp_set_soft_freq_limited_range(struct smu_context *smu, return -EINVAL; } - ret = smu_cmn_send_smc_msg_with_param(smu, msg_set_min, min, NULL); + if (clk_type == SMU_VCLK) { + min_clk = min << SMU_13_VCLK_SHIFT; + max_clk = max << SMU_13_VCLK_SHIFT; + } + + ret = smu_cmn_send_smc_msg_with_param(smu, msg_set_min, min_clk, NULL); + if (ret) goto out; - ret = smu_cmn_send_smc_msg_with_param(smu, msg_set_max, max, NULL); + ret = smu_cmn_send_smc_msg_with_param(smu, msg_set_max, max_clk, NULL); if (ret) goto out; @@ -997,12 +1014,49 @@ out: return ret; } +static uint32_t yellow_carp_get_umd_pstate_clk_default(struct smu_context *smu, + enum smu_clk_type clk_type) +{ + uint32_t clk_limit = 0; + struct amdgpu_device *adev = smu->adev; + + switch (clk_type) { + case SMU_GFXCLK: + case SMU_SCLK: + if ((adev->ip_versions[MP1_HWIP][0]) == IP_VERSION(13, 0, 8)) + clk_limit = SMU_13_0_8_UMD_PSTATE_GFXCLK; + if ((adev->ip_versions[MP1_HWIP][0]) == IP_VERSION(13, 0, 1) || + (adev->ip_versions[MP1_HWIP][0]) == IP_VERSION(13, 0, 3)) + clk_limit = SMU_13_0_1_UMD_PSTATE_GFXCLK; + break; + case SMU_SOCCLK: + if ((adev->ip_versions[MP1_HWIP][0]) == IP_VERSION(13, 0, 8)) + clk_limit = SMU_13_0_8_UMD_PSTATE_SOCCLK; + if ((adev->ip_versions[MP1_HWIP][0]) == IP_VERSION(13, 0, 1) || + (adev->ip_versions[MP1_HWIP][0]) == IP_VERSION(13, 0, 3)) + clk_limit = SMU_13_0_1_UMD_PSTATE_SOCCLK; + break; + case SMU_FCLK: + if ((adev->ip_versions[MP1_HWIP][0]) == IP_VERSION(13, 0, 8)) + clk_limit = SMU_13_0_8_UMD_PSTATE_FCLK; + if ((adev->ip_versions[MP1_HWIP][0]) == IP_VERSION(13, 0, 1) || + (adev->ip_versions[MP1_HWIP][0]) == IP_VERSION(13, 0, 3)) + clk_limit = SMU_13_0_1_UMD_PSTATE_FCLK; + break; + default: + break; + } + + return clk_limit; +} + static int yellow_carp_print_clk_levels(struct smu_context *smu, enum smu_clk_type clk_type, char *buf) { - int i, size = 0, ret = 0; + int i, idx, size = 0, ret = 0; uint32_t cur_value = 0, value = 0, count = 0; uint32_t min, max; + uint32_t clk_limit = 0; smu_cmn_get_sysfs_buf(&buf, &size); @@ -1033,7 +1087,8 @@ static int yellow_carp_print_clk_levels(struct smu_context *smu, goto print_clk_out; for (i = 0; i < count; i++) { - ret = yellow_carp_get_dpm_freq_by_index(smu, clk_type, i, &value); + idx = (clk_type == SMU_FCLK || clk_type == SMU_MCLK) ? (count - i - 1) : i; + ret = yellow_carp_get_dpm_freq_by_index(smu, clk_type, idx, &value); if (ret) goto print_clk_out; @@ -1043,6 +1098,7 @@ static int yellow_carp_print_clk_levels(struct smu_context *smu, break; case SMU_GFXCLK: case SMU_SCLK: + clk_limit = yellow_carp_get_umd_pstate_clk_default(smu, clk_type); ret = yellow_carp_get_current_clk_freq(smu, clk_type, &cur_value); if (ret) goto print_clk_out; @@ -1057,7 +1113,7 @@ static int yellow_carp_print_clk_levels(struct smu_context *smu, size += sysfs_emit_at(buf, size, "0: %uMhz %s\n", min, i == 0 ? "*" : ""); size += sysfs_emit_at(buf, size, "1: %uMhz %s\n", - i == 1 ? cur_value : YELLOW_CARP_UMD_PSTATE_GFXCLK, + i == 1 ? cur_value : clk_limit, i == 1 ? "*" : ""); size += sysfs_emit_at(buf, size, "2: %uMhz %s\n", max, i == 2 ? "*" : ""); @@ -1106,6 +1162,49 @@ force_level_out: return ret; } +static int yellow_carp_get_dpm_profile_freq(struct smu_context *smu, + enum amd_dpm_forced_level level, + enum smu_clk_type clk_type, + uint32_t *min_clk, + uint32_t *max_clk) +{ + int ret = 0; + uint32_t clk_limit = 0; + + clk_limit = yellow_carp_get_umd_pstate_clk_default(smu, clk_type); + + switch (clk_type) { + case SMU_GFXCLK: + case SMU_SCLK: + if (level == AMD_DPM_FORCED_LEVEL_PROFILE_PEAK) + yellow_carp_get_dpm_ultimate_freq(smu, SMU_SCLK, NULL, &clk_limit); + else if (level == AMD_DPM_FORCED_LEVEL_PROFILE_MIN_SCLK) + yellow_carp_get_dpm_ultimate_freq(smu, SMU_SCLK, &clk_limit, NULL); + break; + case SMU_SOCCLK: + if (level == AMD_DPM_FORCED_LEVEL_PROFILE_PEAK) + yellow_carp_get_dpm_ultimate_freq(smu, SMU_SOCCLK, NULL, &clk_limit); + break; + case SMU_FCLK: + if (level == AMD_DPM_FORCED_LEVEL_PROFILE_PEAK) + yellow_carp_get_dpm_ultimate_freq(smu, SMU_FCLK, NULL, &clk_limit); + else if (level == AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK) + yellow_carp_get_dpm_ultimate_freq(smu, SMU_FCLK, &clk_limit, NULL); + break; + case SMU_VCLK: + yellow_carp_get_dpm_ultimate_freq(smu, SMU_VCLK, NULL, &clk_limit); + break; + case SMU_DCLK: + yellow_carp_get_dpm_ultimate_freq(smu, SMU_DCLK, NULL, &clk_limit); + break; + default: + ret = -EINVAL; + break; + } + *min_clk = *max_clk = clk_limit; + return ret; +} + static int yellow_carp_set_performance_level(struct smu_context *smu, enum amd_dpm_forced_level level) { @@ -1113,6 +1212,9 @@ static int yellow_carp_set_performance_level(struct smu_context *smu, uint32_t sclk_min = 0, sclk_max = 0; uint32_t fclk_min = 0, fclk_max = 0; uint32_t socclk_min = 0, socclk_max = 0; + uint32_t vclk_min = 0, vclk_max = 0; + uint32_t dclk_min = 0, dclk_max = 0; + int ret = 0; switch (level) { @@ -1120,28 +1222,42 @@ static int yellow_carp_set_performance_level(struct smu_context *smu, yellow_carp_get_dpm_ultimate_freq(smu, SMU_SCLK, NULL, &sclk_max); yellow_carp_get_dpm_ultimate_freq(smu, SMU_FCLK, NULL, &fclk_max); yellow_carp_get_dpm_ultimate_freq(smu, SMU_SOCCLK, NULL, &socclk_max); + yellow_carp_get_dpm_ultimate_freq(smu, SMU_VCLK, NULL, &vclk_max); + yellow_carp_get_dpm_ultimate_freq(smu, SMU_DCLK, NULL, &dclk_max); sclk_min = sclk_max; fclk_min = fclk_max; socclk_min = socclk_max; + vclk_min = vclk_max; + dclk_min = dclk_max; break; case AMD_DPM_FORCED_LEVEL_LOW: yellow_carp_get_dpm_ultimate_freq(smu, SMU_SCLK, &sclk_min, NULL); yellow_carp_get_dpm_ultimate_freq(smu, SMU_FCLK, &fclk_min, NULL); yellow_carp_get_dpm_ultimate_freq(smu, SMU_SOCCLK, &socclk_min, NULL); + yellow_carp_get_dpm_ultimate_freq(smu, SMU_VCLK, &vclk_min, NULL); + yellow_carp_get_dpm_ultimate_freq(smu, SMU_DCLK, &dclk_min, NULL); sclk_max = sclk_min; fclk_max = fclk_min; socclk_max = socclk_min; + vclk_max = vclk_min; + dclk_max = dclk_min; break; case AMD_DPM_FORCED_LEVEL_AUTO: yellow_carp_get_dpm_ultimate_freq(smu, SMU_SCLK, &sclk_min, &sclk_max); yellow_carp_get_dpm_ultimate_freq(smu, SMU_FCLK, &fclk_min, &fclk_max); yellow_carp_get_dpm_ultimate_freq(smu, SMU_SOCCLK, &socclk_min, &socclk_max); + yellow_carp_get_dpm_ultimate_freq(smu, SMU_VCLK, &vclk_min, &vclk_max); + yellow_carp_get_dpm_ultimate_freq(smu, SMU_DCLK, &dclk_min, &dclk_max); break; case AMD_DPM_FORCED_LEVEL_PROFILE_STANDARD: case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_SCLK: case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK: case AMD_DPM_FORCED_LEVEL_PROFILE_PEAK: - /* Temporarily do nothing since the optimal clocks haven't been provided yet */ + yellow_carp_get_dpm_profile_freq(smu, level, SMU_SCLK, &sclk_min, &sclk_max); + yellow_carp_get_dpm_profile_freq(smu, level, SMU_FCLK, &fclk_min, &fclk_max); + yellow_carp_get_dpm_profile_freq(smu, level, SMU_SOCCLK, &socclk_min, &socclk_max); + yellow_carp_get_dpm_profile_freq(smu, level, SMU_VCLK, &vclk_min, &vclk_max); + yellow_carp_get_dpm_profile_freq(smu, level, SMU_DCLK, &dclk_min, &dclk_max); break; case AMD_DPM_FORCED_LEVEL_MANUAL: case AMD_DPM_FORCED_LEVEL_PROFILE_EXIT: @@ -1181,6 +1297,24 @@ static int yellow_carp_set_performance_level(struct smu_context *smu, return ret; } + if (vclk_min && vclk_max) { + ret = yellow_carp_set_soft_freq_limited_range(smu, + SMU_VCLK, + vclk_min, + vclk_max); + if (ret) + return ret; + } + + if (dclk_min && dclk_max) { + ret = yellow_carp_set_soft_freq_limited_range(smu, + SMU_DCLK, + dclk_min, + dclk_max); + if (ret) + return ret; + } + return ret; } @@ -1234,5 +1368,6 @@ void yellow_carp_set_ppt_funcs(struct smu_context *smu) smu->feature_map = yellow_carp_feature_mask_map; smu->table_map = yellow_carp_table_map; smu->is_apu = true; + smu->smc_driver_if_version = SMU13_YELLOW_CARP_DRIVER_IF_VERSION; smu_v13_0_set_smu_mailbox_registers(smu); } diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.h b/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.h index a9205a8ea3ad..b3ad8352c68a 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.h +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.h @@ -24,6 +24,5 @@ #define __YELLOW_CARP_PPT_H__ extern void yellow_carp_set_ppt_funcs(struct smu_context *smu); -#define YELLOW_CARP_UMD_PSTATE_GFXCLK 1100 #endif |