diff options
| author | Sean Paul <seanpaul@chromium.org> | 2018-10-24 14:26:04 -0400 |
|---|---|---|
| committer | Sean Paul <seanpaul@chromium.org> | 2018-10-24 14:26:04 -0400 |
| commit | 6542e9adc0da1e23d81ff9314265a029b961906d (patch) | |
| tree | 35ba360a7150d8b042a9fd75ab54ef83b34a2b95 /drivers/gpu/drm | |
| parent | 2b02a05bdc3a62d36e0d0b015351897109e25991 (diff) | |
| parent | f2bfc71aee75feff33ca659322b72ffeed5a243d (diff) | |
Merge drm/drm-next into drm-misc-next
4.19 is out, Lyude asked for a backmerge, and it's been a while. All
very good reasons on their own :-)
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Diffstat (limited to 'drivers/gpu/drm')
402 files changed, 12811 insertions, 8557 deletions
diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig index 736b7e67e4ec..4385f00e1d05 100644 --- a/drivers/gpu/drm/Kconfig +++ b/drivers/gpu/drm/Kconfig @@ -110,6 +110,26 @@ config DRM_FBDEV_OVERALLOC is 100. Typical values for double buffering will be 200, triple buffering 300. +config DRM_FBDEV_LEAK_PHYS_SMEM + bool "Shamelessly allow leaking of fbdev physical address (DANGEROUS)" + depends on DRM_FBDEV_EMULATION && EXPERT + default n + help + In order to keep user-space compatibility, we want in certain + use-cases to keep leaking the fbdev physical address to the + user-space program handling the fbdev buffer. + This affects, not only, Amlogic, Allwinner or Rockchip devices + with ARM Mali GPUs using an userspace Blob. + This option is not supported by upstream developers and should be + removed as soon as possible and be considered as a broken and + legacy behaviour from a modern fbdev device driver. + + Please send any bug reports when using this to your proprietary + software vendor that requires this. + + If in doubt, say "N" or spread the word to your closed source + library vendor. + config DRM_LOAD_EDID_FIRMWARE bool "Allow to specify an EDID data set instead of probing for it" depends on DRM diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 6cb35e3dab30..d0102cfc8efb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -81,6 +81,23 @@ #include "amdgpu_bo_list.h" #include "amdgpu_gem.h" +#define MAX_GPU_INSTANCE 16 + +struct amdgpu_gpu_instance +{ + struct amdgpu_device *adev; + int mgpu_fan_enabled; +}; + +struct amdgpu_mgpu_info +{ + struct amdgpu_gpu_instance gpu_ins[MAX_GPU_INSTANCE]; + struct mutex mutex; + uint32_t num_gpu; + uint32_t num_dgpu; + uint32_t num_apu; +}; + /* * Modules parameters. */ @@ -134,6 +151,7 @@ extern int amdgpu_compute_multipipe; extern int amdgpu_gpu_recovery; extern int amdgpu_emu_mode; extern uint amdgpu_smu_memory_pool_size; +extern struct amdgpu_mgpu_info mgpu_info; #ifdef CONFIG_DRM_AMDGPU_SI extern int amdgpu_si_support; @@ -146,6 +164,7 @@ extern int amdgpu_cik_support; #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) /* AMDGPU_IB_POOL_SIZE must be a power of 2 */ #define AMDGPU_IB_POOL_SIZE 16 #define AMDGPU_DEBUGFS_MAX_COMPONENTS 32 @@ -408,16 +427,25 @@ typedef enum _AMDGPU_DOORBELL64_ASSIGNMENT AMDGPU_DOORBELL64_GFX_RING0 = 0x8b, /* - * Other graphics doorbells can be allocated here: from 0x8c to 0xef + * Other graphics doorbells can be allocated here: from 0x8c to 0xdf * Graphics voltage island aperture 1 - * default non-graphics QWORD index is 0xF0 - 0xFF inclusive + * default non-graphics QWORD index is 0xe0 - 0xFF inclusive */ - /* sDMA engines */ - AMDGPU_DOORBELL64_sDMA_ENGINE0 = 0xF0, - AMDGPU_DOORBELL64_sDMA_HI_PRI_ENGINE0 = 0xF1, - AMDGPU_DOORBELL64_sDMA_ENGINE1 = 0xF2, - AMDGPU_DOORBELL64_sDMA_HI_PRI_ENGINE1 = 0xF3, + /* sDMA engines reserved from 0xe0 -oxef */ + AMDGPU_DOORBELL64_sDMA_ENGINE0 = 0xE0, + AMDGPU_DOORBELL64_sDMA_HI_PRI_ENGINE0 = 0xE1, + AMDGPU_DOORBELL64_sDMA_ENGINE1 = 0xE8, + AMDGPU_DOORBELL64_sDMA_HI_PRI_ENGINE1 = 0xE9, + + /* For vega10 sriov, the sdma doorbell must be fixed as follow + * to keep the same setting with host driver, or it will + * happen conflicts + */ + AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE0 = 0xF0, + AMDGPU_VEGA10_DOORBELL64_sDMA_HI_PRI_ENGINE0 = 0xF1, + AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE1 = 0xF2, + AMDGPU_VEGA10_DOORBELL64_sDMA_HI_PRI_ENGINE1 = 0xF3, /* Interrupt handler */ AMDGPU_DOORBELL64_IH = 0xF4, /* For legacy interrupt ring buffer */ @@ -588,31 +616,6 @@ void amdgpu_benchmark(struct amdgpu_device *adev, int test_number); */ void amdgpu_test_moves(struct amdgpu_device *adev); - -/* - * amdgpu smumgr functions - */ -struct amdgpu_smumgr_funcs { - int (*check_fw_load_finish)(struct amdgpu_device *adev, uint32_t fwtype); - int (*request_smu_load_fw)(struct amdgpu_device *adev); - int (*request_smu_specific_fw)(struct amdgpu_device *adev, uint32_t fwtype); -}; - -/* - * amdgpu smumgr - */ -struct amdgpu_smumgr { - struct amdgpu_bo *toc_buf; - struct amdgpu_bo *smu_buf; - /* asic priv smu data */ - void *priv; - spinlock_t smu_lock; - /* smumgr functions */ - const struct amdgpu_smumgr_funcs *smumgr_funcs; - /* ucode loading complete flag */ - uint32_t fw_flags; -}; - /* * ASIC specific register table accessible by UMD */ @@ -948,9 +951,6 @@ struct amdgpu_device { u32 cg_flags; u32 pg_flags; - /* amdgpu smumgr */ - struct amdgpu_smumgr smu; - /* gfx */ struct amdgpu_gfx gfx; @@ -1015,6 +1015,9 @@ struct amdgpu_device { bool has_hw_reset; u8 reset_magic[AMDGPU_RESET_MAGIC_NUM]; + /* s3/s4 mask */ + bool in_suspend; + /* record last mm index being written through WREG32*/ unsigned long last_mm_index; bool in_gpu_reset; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c index 6488e90ec948..7f0afc526419 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c @@ -359,7 +359,9 @@ out: * * Checks the acpi event and if it matches an atif event, * handles it. - * Returns NOTIFY code + * + * Returns: + * NOTIFY_BAD or NOTIFY_DONE, depending on the event. */ static int amdgpu_atif_handler(struct amdgpu_device *adev, struct acpi_bus_event *event) @@ -373,11 +375,16 @@ static int amdgpu_atif_handler(struct amdgpu_device *adev, if (strcmp(event->device_class, ACPI_VIDEO_CLASS) != 0) return NOTIFY_DONE; + /* Is this actually our event? */ if (!atif || !atif->notification_cfg.enabled || - event->type != atif->notification_cfg.command_code) - /* Not our event */ - return NOTIFY_DONE; + event->type != atif->notification_cfg.command_code) { + /* These events will generate keypresses otherwise */ + if (event->type == ACPI_VIDEO_NOTIFY_PROBE) + return NOTIFY_BAD; + else + return NOTIFY_DONE; + } if (atif->functions.sbios_requests) { struct atif_sbios_requests req; @@ -386,7 +393,7 @@ static int amdgpu_atif_handler(struct amdgpu_device *adev, count = amdgpu_atif_get_sbios_requests(atif, &req); if (count <= 0) - return NOTIFY_DONE; + return NOTIFY_BAD; DRM_DEBUG_DRIVER("ATIF: %d pending SBIOS requests\n", count); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index 0f9947edb12a..c31a8849e9f8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -76,6 +76,7 @@ void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev) kfd2kgd = amdgpu_amdkfd_gfx_8_0_get_functions(); break; case CHIP_VEGA10: + case CHIP_VEGA20: case CHIP_RAVEN: kfd2kgd = amdgpu_amdkfd_gfx_9_0_get_functions(); break; @@ -123,7 +124,7 @@ static void amdgpu_doorbell_get_kfd_info(struct amdgpu_device *adev, void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) { - int i; + int i, n; int last_valid_bit; if (adev->kfd) { struct kgd2kfd_shared_resources gpu_resources = { @@ -162,7 +163,15 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) &gpu_resources.doorbell_physical_address, &gpu_resources.doorbell_aperture_size, &gpu_resources.doorbell_start_offset); - if (adev->asic_type >= CHIP_VEGA10) { + + if (adev->asic_type < CHIP_VEGA10) { + kgd2kfd->device_init(adev->kfd, &gpu_resources); + return; + } + + n = (adev->asic_type < CHIP_VEGA20) ? 2 : 8; + + for (i = 0; i < n; i += 2) { /* On SOC15 the BIF is involved in routing * doorbells using the low 12 bits of the * address. Communicate the assignments to @@ -170,20 +179,31 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) * process in case of 64-bit doorbells so we * can use each doorbell assignment twice. */ - gpu_resources.sdma_doorbell[0][0] = - AMDGPU_DOORBELL64_sDMA_ENGINE0; - gpu_resources.sdma_doorbell[0][1] = - AMDGPU_DOORBELL64_sDMA_ENGINE0 + 0x200; - gpu_resources.sdma_doorbell[1][0] = - AMDGPU_DOORBELL64_sDMA_ENGINE1; - gpu_resources.sdma_doorbell[1][1] = - AMDGPU_DOORBELL64_sDMA_ENGINE1 + 0x200; - /* Doorbells 0x0f0-0ff and 0x2f0-2ff are reserved for - * SDMA, IH and VCN. So don't use them for the CP. - */ - gpu_resources.reserved_doorbell_mask = 0x1f0; - gpu_resources.reserved_doorbell_val = 0x0f0; + if (adev->asic_type == CHIP_VEGA10) { + gpu_resources.sdma_doorbell[0][i] = + AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE0 + (i >> 1); + gpu_resources.sdma_doorbell[0][i+1] = + AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE0 + 0x200 + (i >> 1); + gpu_resources.sdma_doorbell[1][i] = + AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE1 + (i >> 1); + gpu_resources.sdma_doorbell[1][i+1] = + AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE1 + 0x200 + (i >> 1); + } else { + gpu_resources.sdma_doorbell[0][i] = + AMDGPU_DOORBELL64_sDMA_ENGINE0 + (i >> 1); + gpu_resources.sdma_doorbell[0][i+1] = + AMDGPU_DOORBELL64_sDMA_ENGINE0 + 0x200 + (i >> 1); + gpu_resources.sdma_doorbell[1][i] = + AMDGPU_DOORBELL64_sDMA_ENGINE1 + (i >> 1); + gpu_resources.sdma_doorbell[1][i+1] = + AMDGPU_DOORBELL64_sDMA_ENGINE1 + 0x200 + (i >> 1); + } } + /* Doorbells 0x0e0-0ff and 0x2e0-2ff are reserved for + * SDMA, IH and VCN. So don't use them for the CP. + */ + gpu_resources.reserved_doorbell_mask = 0x1e0; + gpu_resources.reserved_doorbell_val = 0x0e0; kgd2kfd->device_init(adev->kfd, &gpu_resources); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h index 056fc6ef6c63..8e0d4f7196b4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h @@ -174,7 +174,7 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev, struct amdgpu_vm *vm); void amdgpu_amdkfd_gpuvm_destroy_process_vm(struct kgd_dev *kgd, void *vm); void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *vm); -uint32_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *vm); +uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *vm); int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( struct kgd_dev *kgd, uint64_t va, uint64_t size, void *vm, struct kgd_mem **mem, 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 b2e45c8e2e0d..244d9834a381 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c @@ -142,7 +142,7 @@ static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type); static void set_scratch_backing_va(struct kgd_dev *kgd, uint64_t va, uint32_t vmid); static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, - uint32_t page_table_base); + uint64_t page_table_base); static int invalidate_tlbs(struct kgd_dev *kgd, uint16_t pasid); static int invalidate_tlbs_vmid(struct kgd_dev *kgd, uint16_t vmid); static uint32_t read_vmid_from_vmfault_reg(struct kgd_dev *kgd); @@ -874,7 +874,7 @@ static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type) } static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, - uint32_t page_table_base) + uint64_t page_table_base) { struct amdgpu_device *adev = get_amdgpu_device(kgd); @@ -882,7 +882,8 @@ static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, pr_err("trying to set page table base for wrong VMID\n"); return; } - WREG32(mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8, page_table_base); + WREG32(mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8, + lower_32_bits(page_table_base)); } static int invalidate_tlbs(struct kgd_dev *kgd, uint16_t pasid) 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 ea7c18ce7754..9f149914ad6c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c @@ -45,8 +45,6 @@ enum hqd_dequeue_request_type { RESET_WAVES }; -struct vi_sdma_mqd; - /* * Register access functions */ @@ -100,7 +98,7 @@ static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type); static void set_scratch_backing_va(struct kgd_dev *kgd, uint64_t va, uint32_t vmid); static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, - uint32_t page_table_base); + uint64_t page_table_base); static int invalidate_tlbs(struct kgd_dev *kgd, uint16_t pasid); static int invalidate_tlbs_vmid(struct kgd_dev *kgd, uint16_t vmid); @@ -282,7 +280,8 @@ static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id) lock_srbm(kgd, mec, pipe, 0, 0); - WREG32(mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK); + WREG32(mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK | + CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK); unlock_srbm(kgd); @@ -834,7 +833,7 @@ static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type) } static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, - uint32_t page_table_base) + uint64_t page_table_base) { struct amdgpu_device *adev = get_amdgpu_device(kgd); @@ -842,7 +841,8 @@ static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, pr_err("trying to set page table base for wrong VMID\n"); return; } - WREG32(mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8, page_table_base); + WREG32(mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8, + lower_32_bits(page_table_base)); } static int invalidate_tlbs(struct kgd_dev *kgd, uint16_t pasid) 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 c9176537550b..42cb4c4e0929 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c @@ -138,7 +138,7 @@ static bool get_atc_vmid_pasid_mapping_valid(struct kgd_dev *kgd, static uint16_t get_atc_vmid_pasid_mapping_pasid(struct kgd_dev *kgd, uint8_t vmid); static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, - uint32_t page_table_base); + uint64_t page_table_base); static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type); static void set_scratch_backing_va(struct kgd_dev *kgd, uint64_t va, uint32_t vmid); @@ -1013,11 +1013,10 @@ static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type) } static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid, - uint32_t page_table_base) + uint64_t page_table_base) { struct amdgpu_device *adev = get_amdgpu_device(kgd); - uint64_t base = (uint64_t)page_table_base << PAGE_SHIFT | - AMDGPU_PTE_VALID; + uint64_t base = page_table_base | AMDGPU_PTE_VALID; if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) { pr_err("trying to set page table base for wrong VMID %u\n", diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index 6ee9dc476c86..df0a059565f9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -1131,11 +1131,15 @@ void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *vm) amdgpu_vm_release_compute(adev, avm); } -uint32_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *vm) +uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *vm) { struct amdgpu_vm *avm = (struct amdgpu_vm *)vm; + struct amdgpu_bo *pd = avm->root.base.bo; + struct amdgpu_device *adev = amdgpu_ttm_adev(pd->tbo.bdev); - return avm->pd_phys_addr >> AMDGPU_GPU_PAGE_SHIFT; + if (adev->asic_type < CHIP_VEGA10) + return avm->pd_phys_addr >> AMDGPU_GPU_PAGE_SHIFT; + return avm->pd_phys_addr; } int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c index f5fb93795a69..dd9a4fb9ce39 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c @@ -826,21 +826,13 @@ int amdgpu_debugfs_regs_init(struct amdgpu_device *adev) { struct drm_minor *minor = adev->ddev->primary; struct dentry *ent, *root = minor->debugfs_root; - unsigned i, j; + unsigned int i; for (i = 0; i < ARRAY_SIZE(debugfs_regs); i++) { ent = debugfs_create_file(debugfs_regs_names[i], S_IFREG | S_IRUGO, root, adev, debugfs_regs[i]); - if (IS_ERR(ent)) { - for (j = 0; j < i; j++) { - debugfs_remove(adev->debugfs_regs[i]); - adev->debugfs_regs[i] = NULL; - } - return PTR_ERR(ent); - } - - if (!i) + if (!i && !IS_ERR_OR_NULL(ent)) i_size_write(ent->d_inode, adev->rmmio_size); adev->debugfs_regs[i] = ent; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index bd79d0a31942..1e4dd09a5072 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -1525,6 +1525,92 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) return 0; } +static int amdgpu_device_ip_hw_init_phase1(struct amdgpu_device *adev) +{ + int i, r; + + for (i = 0; i < adev->num_ip_blocks; i++) { + if (!adev->ip_blocks[i].status.sw) + continue; + if (adev->ip_blocks[i].status.hw) + continue; + if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON || + adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH) { + r = adev->ip_blocks[i].version->funcs->hw_init(adev); + if (r) { + DRM_ERROR("hw_init of IP block <%s> failed %d\n", + adev->ip_blocks[i].version->funcs->name, r); + return r; + } + adev->ip_blocks[i].status.hw = true; + } + } + + return 0; +} + +static int amdgpu_device_ip_hw_init_phase2(struct amdgpu_device *adev) +{ + int i, r; + + for (i = 0; i < adev->num_ip_blocks; i++) { + if (!adev->ip_blocks[i].status.sw) + continue; + if (adev->ip_blocks[i].status.hw) + continue; + r = adev->ip_blocks[i].version->funcs->hw_init(adev); + if (r) { + DRM_ERROR("hw_init of IP block <%s> failed %d\n", + adev->ip_blocks[i].version->funcs->name, r); + return r; + } + adev->ip_blocks[i].status.hw = true; + } + + return 0; +} + +static int amdgpu_device_fw_loading(struct amdgpu_device *adev) +{ + int r = 0; + int i; + + if (adev->asic_type >= CHIP_VEGA10) { + for (i = 0; i < adev->num_ip_blocks; i++) { + if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP) { + if (adev->in_gpu_reset || adev->in_suspend) { + if (amdgpu_sriov_vf(adev) && adev->in_gpu_reset) + break; /* sriov gpu reset, psp need to do hw_init before IH because of hw limit */ + r = adev->ip_blocks[i].version->funcs->resume(adev); + if (r) { + DRM_ERROR("resume of IP block <%s> failed %d\n", + adev->ip_blocks[i].version->funcs->name, r); + return r; + } + } else { + r = adev->ip_blocks[i].version->funcs->hw_init(adev); + if (r) { + DRM_ERROR("hw_init of IP block <%s> failed %d\n", + adev->ip_blocks[i].version->funcs->name, r); + return r; + } + } + adev->ip_blocks[i].status.hw = true; + } + } + } + + if (adev->powerplay.pp_funcs->load_firmware) { + r = adev->powerplay.pp_funcs->load_firmware(adev->powerplay.pp_handle); + if (r) { + pr_err("firmware loading failed\n"); + return r; + } + } + + return 0; +} + /** * amdgpu_device_ip_init - run init for hardware IPs * @@ -1581,19 +1667,21 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) } } - for (i = 0; i < adev->num_ip_blocks; i++) { - if (!adev->ip_blocks[i].status.sw) - continue; - if (adev->ip_blocks[i].status.hw) - continue; - r = adev->ip_blocks[i].version->funcs->hw_init((void *)adev); - if (r) { - DRM_ERROR("hw_init of IP block <%s> failed %d\n", - adev->ip_blocks[i].version->funcs->name, r); - return r; - } - adev->ip_blocks[i].status.hw = true; - } + r = amdgpu_ucode_create_bo(adev); /* create ucode bo when sw_init complete*/ + if (r) + return r; + + r = amdgpu_device_ip_hw_init_phase1(adev); + if (r) + return r; + + r = amdgpu_device_fw_loading(adev); + if (r) + return r; + + r = amdgpu_device_ip_hw_init_phase2(adev); + if (r) + return r; amdgpu_xgmi_add_device(adev); amdgpu_amdkfd_device_init(adev); @@ -1656,7 +1744,7 @@ static int amdgpu_device_set_cg_state(struct amdgpu_device *adev, for (j = 0; j < adev->num_ip_blocks; j++) { i = state == AMD_CG_STATE_GATE ? j : adev->num_ip_blocks - j - 1; - if (!adev->ip_blocks[i].status.valid) + if (!adev->ip_blocks[i].status.late_initialized) continue; /* skip CG for VCE/UVD, it's handled specially */ if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_UVD && @@ -1686,7 +1774,7 @@ static int amdgpu_device_set_pg_state(struct amdgpu_device *adev, enum amd_power for (j = 0; j < adev->num_ip_blocks; j++) { i = state == AMD_PG_STATE_GATE ? j : adev->num_ip_blocks - j - 1; - if (!adev->ip_blocks[i].status.valid) + if (!adev->ip_blocks[i].status.late_initialized) continue; /* skip CG for VCE/UVD, it's handled specially */ if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_UVD && @@ -1723,7 +1811,7 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev) int i = 0, r; for (i = 0; i < adev->num_ip_blocks; i++) { - if (!adev->ip_blocks[i].status.valid) + if (!adev->ip_blocks[i].status.hw) continue; if (adev->ip_blocks[i].version->funcs->late_init) { r = adev->ip_blocks[i].version->funcs->late_init((void *)adev); @@ -1732,8 +1820,8 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev) adev->ip_blocks[i].version->funcs->name, r); return r; } - adev->ip_blocks[i].status.late_initialized = true; } + adev->ip_blocks[i].status.late_initialized = true; } amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE); @@ -1803,6 +1891,7 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev) continue; if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC) { + amdgpu_ucode_free_bo(adev); amdgpu_free_static_csa(adev); amdgpu_device_wb_fini(adev); amdgpu_device_vram_scratch_fini(adev); @@ -1833,6 +1922,43 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev) return 0; } +static int amdgpu_device_enable_mgpu_fan_boost(void) +{ + struct amdgpu_gpu_instance *gpu_ins; + struct amdgpu_device *adev; + int i, ret = 0; + + mutex_lock(&mgpu_info.mutex); + + /* + * MGPU fan boost feature should be enabled + * only when there are two or more dGPUs in + * the system + */ + if (mgpu_info.num_dgpu < 2) + goto out; + + for (i = 0; i < mgpu_info.num_dgpu; i++) { + gpu_ins = &(mgpu_info.gpu_ins[i]); + adev = gpu_ins->adev; + if (!(adev->flags & AMD_IS_APU) && + !gpu_ins->mgpu_fan_enabled && + adev->powerplay.pp_funcs && + adev->powerplay.pp_funcs->enable_mgpu_fan_boost) { + ret = amdgpu_dpm_enable_mgpu_fan_boost(adev); + if (ret) + break; + + gpu_ins->mgpu_fan_enabled = 1; + } + } + +out: + mutex_unlock(&mgpu_info.mutex); + + return ret; +} + /** * amdgpu_device_ip_late_init_func_handler - work handler for ib test * @@ -1847,6 +1973,10 @@ static void amdgpu_device_ip_late_init_func_handler(struct work_struct *work) r = amdgpu_ib_ring_tests(adev); if (r) DRM_ERROR("ib ring test failed (%d).\n", r); + + r = amdgpu_device_enable_mgpu_fan_boost(); + if (r) + DRM_ERROR("enable mgpu fan boost failed (%d).\n", r); } static void amdgpu_device_delay_enable_gfx_off(struct work_struct *work) @@ -2082,7 +2212,8 @@ static int amdgpu_device_ip_resume_phase2(struct amdgpu_device *adev) continue; if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON || adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC || - adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH) + adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH || + adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP) continue; r = adev->ip_blocks[i].version->funcs->resume(adev); if (r) { @@ -2114,6 +2245,11 @@ static int amdgpu_device_ip_resume(struct amdgpu_device *adev) r = amdgpu_device_ip_resume_phase1(adev); if (r) return r; + + r = amdgpu_device_fw_loading(adev); + if (r) + return r; + r = amdgpu_device_ip_resume_phase2(adev); return r; @@ -2608,6 +2744,7 @@ int amdgpu_device_suspend(struct drm_device *dev, bool suspend, bool fbcon) if (dev->switch_power_state == DRM_SWITCH_POWER_OFF) return 0; + adev->in_suspend = true; drm_kms_helper_poll_disable(dev); if (fbcon) @@ -2793,6 +2930,8 @@ int amdgpu_device_resume(struct drm_device *dev, bool resume, bool fbcon) #ifdef CONFIG_PM dev->dev->power.disable_depth--; #endif + adev->in_suspend = false; + return 0; } @@ -3061,6 +3200,10 @@ retry: if (r) goto out; + r = amdgpu_device_fw_loading(adev); + if (r) + return r; + r = amdgpu_device_ip_resume_phase2(adev); if (r) goto out; @@ -3117,6 +3260,10 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev, /* we need recover gart prior to run SMC/CP/SDMA resume */ amdgpu_gtt_mgr_recover(&adev->mman.bdev.man[TTM_PL_TT]); + r = amdgpu_device_fw_loading(adev); + if (r) + return r; + /* now we are okay to resume SMC/CP/SDMA */ r = amdgpu_device_ip_reinit_late_sriov(adev); if (r) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h index ff24e1cc5b65..f972cd156795 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h @@ -278,6 +278,9 @@ enum amdgpu_pcie_gen { #define amdgpu_dpm_get_fan_speed_rpm(adev, s) \ ((adev)->powerplay.pp_funcs->get_fan_speed_rpm)((adev)->powerplay.pp_handle, (s)) +#define amdgpu_dpm_set_fan_speed_rpm(adev, s) \ + ((adev)->powerplay.pp_funcs->set_fan_speed_rpm)((adev)->powerplay.pp_handle, (s)) + #define amdgpu_dpm_get_sclk(adev, l) \ ((adev)->powerplay.pp_funcs->get_sclk((adev)->powerplay.pp_handle, (l))) @@ -357,6 +360,10 @@ enum amdgpu_pcie_gen { ((adev)->powerplay.pp_funcs->odn_edit_dpm_table(\ (adev)->powerplay.pp_handle, type, parameter, size)) +#define amdgpu_dpm_enable_mgpu_fan_boost(adev) \ + ((adev)->powerplay.pp_funcs->enable_mgpu_fan_boost(\ + (adev)->powerplay.pp_handle)) + struct amdgpu_dpm { struct amdgpu_ps *ps; /* number of valid power states */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 723f0f7754bd..28781414d71c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -127,6 +127,9 @@ int amdgpu_compute_multipipe = -1; int amdgpu_gpu_recovery = -1; /* auto */ int amdgpu_emu_mode = 0; uint amdgpu_smu_memory_pool_size = 0; +struct amdgpu_mgpu_info mgpu_info = { + .mutex = __MUTEX_INITIALIZER(mgpu_info.mutex), +}; /** * DOC: vramlimit (int) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c index 176f28777f5e..5448cf27654e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c @@ -196,6 +196,19 @@ int amdgpu_fence_emit_polling(struct amdgpu_ring *ring, uint32_t *s) } /** + * amdgpu_fence_schedule_fallback - schedule fallback check + * + * @ring: pointer to struct amdgpu_ring + * + * Start a timer as fallback to our interrupts. + */ +static void amdgpu_fence_schedule_fallback(struct amdgpu_ring *ring) +{ + mod_timer(&ring->fence_drv.fallback_timer, + jiffies + AMDGPU_FENCE_JIFFIES_TIMEOUT); +} + +/** * amdgpu_fence_process - check for fence activity * * @ring: pointer to struct amdgpu_ring @@ -203,8 +216,10 @@ int amdgpu_fence_emit_polling(struct amdgpu_ring *ring, uint32_t *s) * Checks the current fence value and calculates the last * signalled fence value. Wakes the fence queue if the * sequence number has increased. + * + * Returns true if fence was processed */ -void amdgpu_fence_process(struct amdgpu_ring *ring) +bool amdgpu_fence_process(struct amdgpu_ring *ring) { struct amdgpu_fence_driver *drv = &ring->fence_drv; uint32_t seq, last_seq; @@ -216,8 +231,12 @@ void amdgpu_fence_process(struct amdgpu_ring *ring) } while (atomic_cmpxchg(&drv->last_seq, last_seq, seq) != last_seq); + if (del_timer(&ring->fence_drv.fallback_timer) && + seq != ring->fence_drv.sync_seq) + amdgpu_fence_schedule_fallback(ring); + if (unlikely(seq == last_seq)) - return; + return false; last_seq &= drv->num_fences_mask; seq &= drv->num_fences_mask; @@ -244,6 +263,24 @@ void amdgpu_fence_process(struct amdgpu_ring *ring) dma_fence_put(fence); } while (last_seq != seq); + + return true; +} + +/** + * amdgpu_fence_fallback - fallback for hardware interrupts + * + * @work: delayed work item + * + * Checks for fence activity. + */ +static void amdgpu_fence_fallback(struct timer_list *t) +{ + struct amdgpu_ring *ring = from_timer(ring, t, + fence_drv.fallback_timer); + + if (amdgpu_fence_process(ring)) + DRM_WARN("Fence fallback timer expired on ring %s\n", ring->name); } /** @@ -393,6 +430,8 @@ int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring, atomic_set(&ring->fence_drv.last_seq, 0); ring->fence_drv.initialized = false; + timer_setup(&ring->fence_drv.fallback_timer, amdgpu_fence_fallback, 0); + ring->fence_drv.num_fences_mask = num_hw_submission * 2 - 1; spin_lock_init(&ring->fence_drv.lock); ring->fence_drv.fences = kcalloc(num_hw_submission * 2, sizeof(void *), @@ -468,6 +507,7 @@ void amdgpu_fence_driver_fini(struct amdgpu_device *adev) amdgpu_irq_put(adev, ring->fence_drv.irq_src, ring->fence_drv.irq_type); drm_sched_fini(&ring->sched); + del_timer_sync(&ring->fence_drv.fallback_timer); for (j = 0; j <= ring->fence_drv.num_fences_mask; ++j) dma_fence_put(ring->fence_drv.fences[j]); kfree(ring->fence_drv.fences); @@ -561,6 +601,27 @@ static const char *amdgpu_fence_get_timeline_name(struct dma_fence *f) } /** + * amdgpu_fence_enable_signaling - enable signalling on fence + * @fence: fence + * + * This function is called with fence_queue lock held, and adds a callback + * to fence_queue that checks if this fence is signaled, and if so it + * signals the fence and removes itself. + */ +static bool amdgpu_fence_enable_signaling(struct dma_fence *f) +{ + struct amdgpu_fence *fence = to_amdgpu_fence(f); + struct amdgpu_ring *ring = fence->ring; + + if (!timer_pending(&ring->fence_drv.fallback_timer)) + amdgpu_fence_schedule_fallback(ring); + + DMA_FENCE_TRACE(&fence->base, "armed on ring %i!\n", ring->idx); + + return true; +} + +/** * amdgpu_fence_free - free up the fence memory * * @rcu: RCU callback head @@ -590,6 +651,7 @@ static void amdgpu_fence_release(struct dma_fence *f) static const struct dma_fence_ops amdgpu_fence_ops = { .get_driver_name = amdgpu_fence_get_driver_name, .get_timeline_name = amdgpu_fence_get_timeline_name, + .enable_signaling = amdgpu_fence_enable_signaling, .release = amdgpu_fence_release, }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h index f172e92c463c..b61b5c11aead 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h @@ -297,8 +297,7 @@ struct amdgpu_gfx { /* reset mask */ uint32_t grbm_soft_reset; uint32_t srbm_soft_reset; - /* s3/s4 mask */ - bool in_suspend; + /* NGG */ struct amdgpu_ngg ngg; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c index 9a5b252784a1..d73367cab4f3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c @@ -146,6 +146,8 @@ void amdgpu_gmc_gart_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc) { const uint64_t four_gb = 0x100000000ULL; u64 size_af, size_bf; + /*To avoid the hole, limit the max mc address to AMDGPU_GMC_HOLE_START*/ + u64 max_mc_address = min(adev->gmc.mc_mask, AMDGPU_GMC_HOLE_START - 1); mc->gart_size += adev->pm.smu_prv_buffer_size; @@ -153,7 +155,7 @@ void amdgpu_gmc_gart_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc) * the GART base on a 4GB boundary as well. */ size_bf = mc->fb_start; - size_af = adev->gmc.mc_mask + 1 - ALIGN(mc->fb_end + 1, four_gb); + size_af = max_mc_address + 1 - ALIGN(mc->fb_end + 1, four_gb); if (mc->gart_size > max(size_bf, size_af)) { dev_warn(adev->dev, "limiting GART\n"); @@ -164,7 +166,7 @@ void amdgpu_gmc_gart_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc) (size_af < mc->gart_size)) mc->gart_start = 0; else - mc->gart_start = mc->mc_mask - mc->gart_size + 1; + mc->gart_start = max_mc_address - mc->gart_size + 1; mc->gart_start &= ~(four_gb - 1); mc->gart_end = mc->gart_start + mc->gart_size - 1; @@ -200,16 +202,13 @@ void amdgpu_gmc_agp_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc) } if (size_bf > size_af) { - mc->agp_start = mc->fb_start > mc->gart_start ? - mc->gart_end + 1 : 0; + mc->agp_start = (mc->fb_start - size_bf) & sixteen_gb_mask; mc->agp_size = size_bf; } else { - mc->agp_start = (mc->fb_start > mc->gart_start ? - mc->fb_end : mc->gart_end) + 1, + mc->agp_start = ALIGN(mc->fb_end + 1, sixteen_gb); mc->agp_size = size_af; } - mc->agp_start = ALIGN(mc->agp_start, sixteen_gb); mc->agp_end = mc->agp_start + mc->agp_size - 1; dev_info(adev->dev, "AGP: %lluM 0x%016llX - 0x%016llX\n", mc->agp_size >> 20, mc->agp_start, mc->agp_end); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c index 47817e00f54f..b8963b725dfa 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c @@ -354,6 +354,14 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev) if (!ring || !ring->ready) continue; + /* skip IB tests for KIQ in general for the below reasons: + * 1. We never submit IBs to the KIQ + * 2. KIQ doesn't use the EOP interrupts, + * we use some other CP interrupt. + */ + if (ring->funcs->type == AMDGPU_RING_TYPE_KIQ) + continue; + /* MM engine need more time */ if (ring->funcs->type == AMDGPU_RING_TYPE_UVD || ring->funcs->type == AMDGPU_RING_TYPE_VCE || diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c index 3a072a7a39f0..df9b173c3d0b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c @@ -574,7 +574,7 @@ void amdgpu_vmid_mgr_init(struct amdgpu_device *adev) /* skip over VMID 0, since it is the system VM */ for (j = 1; j < id_mgr->num_ids; ++j) { amdgpu_vmid_reset(adev, i, j); - amdgpu_sync_create(&id_mgr->ids[i].active); + amdgpu_sync_create(&id_mgr->ids[j].active); list_add_tail(&id_mgr->ids[j].list, &id_mgr->ids_lru); } } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c index 4ed86218cef3..8af67f649660 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c @@ -24,46 +24,21 @@ #include <drm/drmP.h> #include "amdgpu.h" #include "amdgpu_ih.h" -#include "amdgpu_amdkfd.h" - -/** - * amdgpu_ih_ring_alloc - allocate memory for the IH ring - * - * @adev: amdgpu_device pointer - * - * Allocate a ring buffer for the interrupt controller. - * Returns 0 for success, errors for failure. - */ -static int amdgpu_ih_ring_alloc(struct amdgpu_device *adev) -{ - int r; - - /* Allocate ring buffer */ - if (adev->irq.ih.ring_obj == NULL) { - r = amdgpu_bo_create_kernel(adev, adev->irq.ih.ring_size, - PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT, - &adev->irq.ih.ring_obj, - &adev->irq.ih.gpu_addr, - (void **)&adev->irq.ih.ring); - if (r) { - DRM_ERROR("amdgpu: failed to create ih ring buffer (%d).\n", r); - return r; - } - } - return 0; -} /** * amdgpu_ih_ring_init - initialize the IH state * * @adev: amdgpu_device pointer + * @ih: ih ring to initialize + * @ring_size: ring size to allocate + * @use_bus_addr: true when we can use dma_alloc_coherent * * Initializes the IH state and allocates a buffer * for the IH ring buffer. * Returns 0 for success, errors for failure. */ -int amdgpu_ih_ring_init(struct amdgpu_device *adev, unsigned ring_size, - bool use_bus_addr) +int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, + unsigned ring_size, bool use_bus_addr) { u32 rb_bufsz; int r; @@ -71,70 +46,76 @@ int amdgpu_ih_ring_init(struct amdgpu_device *adev, unsigned ring_size, /* Align ring size */ rb_bufsz = order_base_2(ring_size / 4); ring_size = (1 << rb_bufsz) * 4; - adev->irq.ih.ring_size = ring_size; - adev->irq.ih.ptr_mask = adev->irq.ih.ring_size - 1; - adev->irq.ih.rptr = 0; - adev->irq.ih.use_bus_addr = use_bus_addr; - - if (adev->irq.ih.use_bus_addr) { - if (!adev->irq.ih.ring) { - /* add 8 bytes for the rptr/wptr shadows and - * add them to the end of the ring allocation. - */ - adev->irq.ih.ring = pci_alloc_consistent(adev->pdev, - adev->irq.ih.ring_size + 8, - &adev->irq.ih.rb_dma_addr); - if (adev->irq.ih.ring == NULL) - return -ENOMEM; - memset((void *)adev->irq.ih.ring, 0, adev->irq.ih.ring_size + 8); - adev->irq.ih.wptr_offs = (adev->irq.ih.ring_size / 4) + 0; - adev->irq.ih.rptr_offs = (adev->irq.ih.ring_size / 4) + 1; - } - return 0; + ih->ring_size = ring_size; + ih->ptr_mask = ih->ring_size - 1; + ih->rptr = 0; + ih->use_bus_addr = use_bus_addr; + + if (use_bus_addr) { + if (ih->ring) + return 0; + + /* add 8 bytes for the rptr/wptr shadows and + * add them to the end of the ring allocation. + */ + ih->ring = dma_alloc_coherent(adev->dev, ih->ring_size + 8, + &ih->rb_dma_addr, GFP_KERNEL); + if (ih->ring == NULL) + return -ENOMEM; + + memset((void *)ih->ring, 0, ih->ring_size + 8); + ih->wptr_offs = (ih->ring_size / 4) + 0; + ih->rptr_offs = (ih->ring_size / 4) + 1; } else { - r = amdgpu_device_wb_get(adev, &adev->irq.ih.wptr_offs); + r = amdgpu_device_wb_get(adev, &ih->wptr_offs); + if (r) + return r; + + r = amdgpu_device_wb_get(adev, &ih->rptr_offs); if (r) { - dev_err(adev->dev, "(%d) ih wptr_offs wb alloc failed\n", r); + amdgpu_device_wb_free(adev, ih->wptr_offs); return r; } - r = amdgpu_device_wb_get(adev, &adev->irq.ih.rptr_offs); + r = amdgpu_bo_create_kernel(adev, ih->ring_size, PAGE_SIZE, + AMDGPU_GEM_DOMAIN_GTT, + &ih->ring_obj, &ih->gpu_addr, + (void **)&ih->ring); if (r) { - amdgpu_device_wb_free(adev, adev->irq.ih.wptr_offs); - dev_err(adev->dev, "(%d) ih rptr_offs wb alloc failed\n", r); + amdgpu_device_wb_free(adev, ih->rptr_offs); + amdgpu_device_wb_free(adev, ih->wptr_offs); return r; } - - return amdgpu_ih_ring_alloc(adev); } + return 0; } /** * amdgpu_ih_ring_fini - tear down the IH state * * @adev: amdgpu_device pointer + * @ih: ih ring to tear down * * Tears down the IH state and frees buffer * used for the IH ring buffer. */ -void amdgpu_ih_ring_fini(struct amdgpu_device *adev) +void amdgpu_ih_ring_fini(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih) { - if (adev->irq.ih.use_bus_addr) { - if (adev->irq.ih.ring) { - /* add 8 bytes for the rptr/wptr shadows and - * add them to the end of the ring allocation. - */ - pci_free_consistent(adev->pdev, adev->irq.ih.ring_size + 8, - (void *)adev->irq.ih.ring, - adev->irq.ih.rb_dma_addr); - adev->irq.ih.ring = NULL; - } + if (ih->use_bus_addr) { + if (!ih->ring) + return; + + /* add 8 bytes for the rptr/wptr shadows and + * add them to the end of the ring allocation. + */ + dma_free_coherent(adev->dev, ih->ring_size + 8, + (void *)ih->ring, ih->rb_dma_addr); + ih->ring = NULL; } else { - amdgpu_bo_free_kernel(&adev->irq.ih.ring_obj, - &adev->irq.ih.gpu_addr, - (void **)&adev->irq.ih.ring); - amdgpu_device_wb_free(adev, adev->irq.ih.wptr_offs); - amdgpu_device_wb_free(adev, adev->irq.ih.rptr_offs); + amdgpu_bo_free_kernel(&ih->ring_obj, &ih->gpu_addr, + (void **)&ih->ring); + amdgpu_device_wb_free(adev, ih->wptr_offs); + amdgpu_device_wb_free(adev, ih->rptr_offs); } } @@ -142,56 +123,43 @@ void amdgpu_ih_ring_fini(struct amdgpu_device *adev) * amdgpu_ih_process - interrupt handler * * @adev: amdgpu_device pointer + * @ih: ih ring to process * * Interrupt hander (VI), walk the IH ring. * Returns irq process return code. */ -int amdgpu_ih_process(struct amdgpu_device *adev) +int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, + void (*callback)(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih)) { - struct amdgpu_iv_entry entry; u32 wptr; - if (!adev->irq.ih.enabled || adev->shutdown) + if (!ih->enabled || adev->shutdown) return IRQ_NONE; wptr = amdgpu_ih_get_wptr(adev); restart_ih: /* is somebody else already processing irqs? */ - if (atomic_xchg(&adev->irq.ih.lock, 1)) + if (atomic_xchg(&ih->lock, 1)) return IRQ_NONE; - DRM_DEBUG("%s: rptr %d, wptr %d\n", __func__, adev->irq.ih.rptr, wptr); + DRM_DEBUG("%s: rptr %d, wptr %d\n", __func__, ih->rptr, wptr); /* Order reading of wptr vs. reading of IH ring data */ rmb(); - while (adev->irq.ih.rptr != wptr) { - u32 ring_index = adev->irq.ih.rptr >> 2; - - /* Prescreening of high-frequency interrupts */ - if (!amdgpu_ih_prescreen_iv(adev)) { - adev->irq.ih.rptr &= adev->irq.ih.ptr_mask; - continue; - } - - /* Before dispatching irq to IP blocks, send it to amdkfd */ - amdgpu_amdkfd_interrupt(adev, - (const void *) &adev->irq.ih.ring[ring_index]); - - entry.iv_entry = (const uint32_t *) - &adev->irq.ih.ring[ring_index]; - amdgpu_ih_decode_iv(adev, &entry); - adev->irq.ih.rptr &= adev->irq.ih.ptr_mask; - - amdgpu_irq_dispatch(adev, &entry); + while (ih->rptr != wptr) { + callback(adev, ih); + ih->rptr &= ih->ptr_mask; } + amdgpu_ih_set_rptr(adev); - atomic_set(&adev->irq.ih.lock, 0); + atomic_set(&ih->lock, 0); /* make sure wptr hasn't changed while processing */ wptr = amdgpu_ih_get_wptr(adev); - if (wptr != adev->irq.ih.rptr) + if (wptr != ih->rptr) goto restart_ih; return IRQ_HANDLED; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h index 0d5b3f5201d2..9ce8c93ec19b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h @@ -24,12 +24,8 @@ #ifndef __AMDGPU_IH_H__ #define __AMDGPU_IH_H__ -#include "soc15_ih_clientid.h" - struct amdgpu_device; - -#define AMDGPU_IH_CLIENTID_LEGACY 0 -#define AMDGPU_IH_CLIENTID_MAX SOC15_IH_CLIENTID_MAX +struct amdgpu_iv_entry; /* * R6xx+ IH ring @@ -51,22 +47,6 @@ struct amdgpu_ih_ring { dma_addr_t rb_dma_addr; /* only used when use_bus_addr = true */ }; -#define AMDGPU_IH_SRC_DATA_MAX_SIZE_DW 4 - -struct amdgpu_iv_entry { - unsigned client_id; - unsigned src_id; - unsigned ring_id; - unsigned vmid; - unsigned vmid_src; - uint64_t timestamp; - unsigned timestamp_src; - unsigned pasid; - unsigned pasid_src; - unsigned src_data[AMDGPU_IH_SRC_DATA_MAX_SIZE_DW]; - const uint32_t *iv_entry; -}; - /* provided by the ih block */ struct amdgpu_ih_funcs { /* ring read/write ptr handling, called from interrupt context */ @@ -82,9 +62,11 @@ struct amdgpu_ih_funcs { #define amdgpu_ih_decode_iv(adev, iv) (adev)->irq.ih_funcs->decode_iv((adev), (iv)) #define amdgpu_ih_set_rptr(adev) (adev)->irq.ih_funcs->set_rptr((adev)) -int amdgpu_ih_ring_init(struct amdgpu_device *adev, unsigned ring_size, - bool use_bus_addr); -void amdgpu_ih_ring_fini(struct amdgpu_device *adev); -int amdgpu_ih_process(struct amdgpu_device *adev); +int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, + unsigned ring_size, bool use_bus_addr); +void amdgpu_ih_ring_fini(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih); +int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, + void (*callback)(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih)); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c index b927e8798534..52c17f6219a7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c @@ -51,6 +51,7 @@ #include "atom.h" #include "amdgpu_connectors.h" #include "amdgpu_trace.h" +#include "amdgpu_amdkfd.h" #include <linux/pm_runtime.h> @@ -123,7 +124,7 @@ void amdgpu_irq_disable_all(struct amdgpu_device *adev) int r; spin_lock_irqsave(&adev->irq.lock, irqflags); - for (i = 0; i < AMDGPU_IH_CLIENTID_MAX; ++i) { + for (i = 0; i < AMDGPU_IRQ_CLIENTID_MAX; ++i) { if (!adev->irq.client[i].sources) continue; @@ -147,6 +148,34 @@ void amdgpu_irq_disable_all(struct amdgpu_device *adev) } /** + * amdgpu_irq_callback - callback from the IH ring + * + * @adev: amdgpu device pointer + * @ih: amdgpu ih ring + * + * Callback from IH ring processing to handle the entry at the current position + * and advance the read pointer. + */ +static void amdgpu_irq_callback(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) +{ + u32 ring_index = ih->rptr >> 2; + struct amdgpu_iv_entry entry; + + /* Prescreening of high-frequency interrupts */ + if (!amdgpu_ih_prescreen_iv(adev)) + return; + + /* Before dispatching irq to IP blocks, send it to amdkfd */ + amdgpu_amdkfd_interrupt(adev, (const void *) &ih->ring[ring_index]); + + entry.iv_entry = (const uint32_t *)&ih->ring[ring_index]; + amdgpu_ih_decode_iv(adev, &entry); + + amdgpu_irq_dispatch(adev, &entry); +} + +/** * amdgpu_irq_handler - IRQ handler * * @irq: IRQ number (unused) @@ -163,7 +192,7 @@ irqreturn_t amdgpu_irq_handler(int irq, void *arg) struct amdgpu_device *adev = dev->dev_private; irqreturn_t ret; - ret = amdgpu_ih_process(adev); + ret = amdgpu_ih_process(adev, &adev->irq.ih, amdgpu_irq_callback); if (ret == IRQ_HANDLED) pm_runtime_mark_last_busy(dev->dev); return ret; @@ -273,7 +302,7 @@ void amdgpu_irq_fini(struct amdgpu_device *adev) cancel_work_sync(&adev->reset_work); } - for (i = 0; i < AMDGPU_IH_CLIENTID_MAX; ++i) { + for (i = 0; i < AMDGPU_IRQ_CLIENTID_MAX; ++i) { if (!adev->irq.client[i].sources) continue; @@ -313,7 +342,7 @@ int amdgpu_irq_add_id(struct amdgpu_device *adev, unsigned client_id, unsigned src_id, struct amdgpu_irq_src *source) { - if (client_id >= AMDGPU_IH_CLIENTID_MAX) + if (client_id >= AMDGPU_IRQ_CLIENTID_MAX) return -EINVAL; if (src_id >= AMDGPU_MAX_IRQ_SRC_ID) @@ -367,7 +396,7 @@ void amdgpu_irq_dispatch(struct amdgpu_device *adev, trace_amdgpu_iv(entry); - if (client_id >= AMDGPU_IH_CLIENTID_MAX) { + if (client_id >= AMDGPU_IRQ_CLIENTID_MAX) { DRM_DEBUG("Invalid client_id in IV: %d\n", client_id); return; } @@ -440,7 +469,7 @@ void amdgpu_irq_gpu_reset_resume_helper(struct amdgpu_device *adev) { int i, j, k; - for (i = 0; i < AMDGPU_IH_CLIENTID_MAX; ++i) { + for (i = 0; i < AMDGPU_IRQ_CLIENTID_MAX; ++i) { if (!adev->irq.client[i].sources) continue; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h index 3375ad778edc..f6ce171cb8aa 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h @@ -25,19 +25,38 @@ #define __AMDGPU_IRQ_H__ #include <linux/irqdomain.h> +#include "soc15_ih_clientid.h" #include "amdgpu_ih.h" -#define AMDGPU_MAX_IRQ_SRC_ID 0x100 +#define AMDGPU_MAX_IRQ_SRC_ID 0x100 #define AMDGPU_MAX_IRQ_CLIENT_ID 0x100 +#define AMDGPU_IRQ_CLIENTID_LEGACY 0 +#define AMDGPU_IRQ_CLIENTID_MAX SOC15_IH_CLIENTID_MAX + +#define AMDGPU_IRQ_SRC_DATA_MAX_SIZE_DW 4 + struct amdgpu_device; -struct amdgpu_iv_entry; enum amdgpu_interrupt_state { AMDGPU_IRQ_STATE_DISABLE, AMDGPU_IRQ_STATE_ENABLE, }; +struct amdgpu_iv_entry { + unsigned client_id; + unsigned src_id; + unsigned ring_id; + unsigned vmid; + unsigned vmid_src; + uint64_t timestamp; + unsigned timestamp_src; + unsigned pasid; + unsigned pasid_src; + unsigned src_data[AMDGPU_IRQ_SRC_DATA_MAX_SIZE_DW]; + const uint32_t *iv_entry; +}; + struct amdgpu_irq_src { unsigned num_types; atomic_t *enabled_types; @@ -63,7 +82,7 @@ struct amdgpu_irq { bool installed; spinlock_t lock; /* interrupt sources */ - struct amdgpu_irq_client client[AMDGPU_IH_CLIENTID_MAX]; + struct amdgpu_irq_client client[AMDGPU_IRQ_CLIENTID_MAX]; /* status, etc. */ bool msi_enabled; /* msi enabled */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index a64056dadc58..81732a84c2ab 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -40,6 +40,30 @@ #include "amdgpu_gem.h" #include "amdgpu_display.h" +static void amdgpu_unregister_gpu_instance(struct amdgpu_device *adev) +{ + struct amdgpu_gpu_instance *gpu_instance; + int i; + + mutex_lock(&mgpu_info.mutex); + + for (i = 0; i < mgpu_info.num_gpu; i++) { + gpu_instance = &(mgpu_info.gpu_ins[i]); + if (gpu_instance->adev == adev) { + mgpu_info.gpu_ins[i] = + mgpu_info.gpu_ins[mgpu_info.num_gpu - 1]; + mgpu_info.num_gpu--; + if (adev->flags & AMD_IS_APU) + mgpu_info.num_apu--; + else + mgpu_info.num_dgpu--; + break; + } + } + + mutex_unlock(&mgpu_info.mutex); +} + /** * amdgpu_driver_unload_kms - Main unload function for KMS. * @@ -55,6 +79,8 @@ void amdgpu_driver_unload_kms(struct drm_device *dev) if (adev == NULL) return; + amdgpu_unregister_gpu_instance(adev); + if (adev->rmmio == NULL) goto done_free; @@ -75,6 +101,31 @@ done_free: dev->dev_private = NULL; } +static void amdgpu_register_gpu_instance(struct amdgpu_device *adev) +{ + struct amdgpu_gpu_instance *gpu_instance; + + mutex_lock(&mgpu_info.mutex); + + if (mgpu_info.num_gpu >= MAX_GPU_INSTANCE) { + DRM_ERROR("Cannot register more gpu instance\n"); + mutex_unlock(&mgpu_info.mutex); + return; + } + + gpu_instance = &(mgpu_info.gpu_ins[mgpu_info.num_gpu]); + gpu_instance->adev = adev; + gpu_instance->mgpu_fan_enabled = 0; + + mgpu_info.num_gpu++; + if (adev->flags & AMD_IS_APU) + mgpu_info.num_apu++; + else + mgpu_info.num_dgpu++; + + mutex_unlock(&mgpu_info.mutex); +} + /** * amdgpu_driver_load_kms - Main load function for KMS. * @@ -169,6 +220,7 @@ int amdgpu_driver_load_kms(struct drm_device *dev, unsigned long flags) pm_runtime_put_autosuspend(dev->dev); } + amdgpu_register_gpu_instance(adev); out: if (r) { /* balance pm_runtime_get_sync in amdgpu_driver_unload_kms */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c index 8c334fc808c2..94055a485e01 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c @@ -1120,12 +1120,19 @@ static ssize_t amdgpu_hwmon_set_pwm1(struct device *dev, struct amdgpu_device *adev = dev_get_drvdata(dev); int err; u32 value; + u32 pwm_mode; /* Can't adjust fan when the card is off */ if ((adev->flags & AMD_IS_PX) && (adev->ddev->switch_power_state != DRM_SWITCH_POWER_ON)) return -EINVAL; + pwm_mode = amdgpu_dpm_get_fan_control_mode(adev); + if (pwm_mode != AMD_FAN_CTRL_MANUAL) { + pr_info("manual fan speed control should be enabled first\n"); + return -EINVAL; + } + err = kstrtou32(buf, 10, &value); if (err) return err; @@ -1187,6 +1194,148 @@ static ssize_t amdgpu_hwmon_get_fan1_input(struct device *dev, return sprintf(buf, "%i\n", speed); } +static ssize_t amdgpu_hwmon_get_fan1_min(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct amdgpu_device *adev = dev_get_drvdata(dev); + u32 min_rpm = 0; + u32 size = sizeof(min_rpm); + int r; + + if (!adev->powerplay.pp_funcs->read_sensor) + return -EINVAL; + + r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_MIN_FAN_RPM, + (void *)&min_rpm, &size); + if (r) + return r; + + return snprintf(buf, PAGE_SIZE, "%d\n", min_rpm); +} + +static ssize_t amdgpu_hwmon_get_fan1_max(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct amdgpu_device *adev = dev_get_drvdata(dev); + u32 max_rpm = 0; + u32 size = sizeof(max_rpm); + int r; + + if (!adev->powerplay.pp_funcs->read_sensor) + return -EINVAL; + + r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_MAX_FAN_RPM, + (void *)&max_rpm, &size); + if (r) + return r; + + return snprintf(buf, PAGE_SIZE, "%d\n", max_rpm); +} + +static ssize_t amdgpu_hwmon_get_fan1_target(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct amdgpu_device *adev = dev_get_drvdata(dev); + int err; + u32 rpm = 0; + + /* Can't adjust fan when the card is off */ + if ((adev->flags & AMD_IS_PX) && + (adev->ddev->switch_power_state != DRM_SWITCH_POWER_ON)) + return -EINVAL; + + if (adev->powerplay.pp_funcs->get_fan_speed_rpm) { + err = amdgpu_dpm_get_fan_speed_rpm(adev, &rpm); + if (err) + return err; + } + + return sprintf(buf, "%i\n", rpm); +} + +static ssize_t amdgpu_hwmon_set_fan1_target(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + struct amdgpu_device *adev = dev_get_drvdata(dev); + int err; + u32 value; + u32 pwm_mode; + + pwm_mode = amdgpu_dpm_get_fan_control_mode(adev); + if (pwm_mode != AMD_FAN_CTRL_MANUAL) + return -ENODATA; + + /* Can't adjust fan when the card is off */ + if ((adev->flags & AMD_IS_PX) && + (adev->ddev->switch_power_state != DRM_SWITCH_POWER_ON)) + return -EINVAL; + + err = kstrtou32(buf, 10, &value); + if (err) + return err; + + if (adev->powerplay.pp_funcs->set_fan_speed_rpm) { + err = amdgpu_dpm_set_fan_speed_rpm(adev, value); + if (err) + return err; + } + + return count; +} + +static ssize_t amdgpu_hwmon_get_fan1_enable(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct amdgpu_device *adev = dev_get_drvdata(dev); + u32 pwm_mode = 0; + + if (!adev->powerplay.pp_funcs->get_fan_control_mode) + return -EINVAL; + + pwm_mode = amdgpu_dpm_get_fan_control_mode(adev); + + return sprintf(buf, "%i\n", pwm_mode == AMD_FAN_CTRL_AUTO ? 0 : 1); +} + +static ssize_t amdgpu_hwmon_set_fan1_enable(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + struct amdgpu_device *adev = dev_get_drvdata(dev); + int err; + int value; + u32 pwm_mode; + + /* Can't adjust fan when the card is off */ + if ((adev->flags & AMD_IS_PX) && + (adev->ddev->switch_power_state != DRM_SWITCH_POWER_ON)) + return -EINVAL; + + if (!adev->powerplay.pp_funcs->set_fan_control_mode) + return -EINVAL; + + err = kstrtoint(buf, 10, &value); + if (err) + return err; + + if (value == 0) + pwm_mode = AMD_FAN_CTRL_AUTO; + else if (value == 1) + pwm_mode = AMD_FAN_CTRL_MANUAL; + else + return -EINVAL; + + amdgpu_dpm_set_fan_control_mode(adev, pwm_mode); + + return count; +} + static ssize_t amdgpu_hwmon_show_vddgfx(struct device *dev, struct device_attribute *attr, char *buf) @@ -1406,8 +1555,16 @@ static ssize_t amdgpu_hwmon_set_power_cap(struct device *dev, * * - pwm1_max: pulse width modulation fan control maximum level (255) * + * - fan1_min: an minimum value Unit: revolution/min (RPM) + * + * - fan1_max: an maxmum value Unit: revolution/max (RPM) + * * - fan1_input: fan speed in RPM * + * - fan[1-*]_target: Desired fan speed Unit: revolution/min (RPM) + * + * - fan[1-*]_enable: Enable or disable the sensors.1: Enable 0: Disable + * * You can use hwmon tools like sensors to view this information on your system. * */ @@ -1420,6 +1577,10 @@ static SENSOR_DEVICE_ATTR(pwm1_enable, S_IRUGO | S_IWUSR, amdgpu_hwmon_get_pwm1_ static SENSOR_DEVICE_ATTR(pwm1_min, S_IRUGO, amdgpu_hwmon_get_pwm1_min, NULL, 0); static SENSOR_DEVICE_ATTR(pwm1_max, S_IRUGO, amdgpu_hwmon_get_pwm1_max, NULL, 0); static SENSOR_DEVICE_ATTR(fan1_input, S_IRUGO, amdgpu_hwmon_get_fan1_input, NULL, 0); +static SENSOR_DEVICE_ATTR(fan1_min, S_IRUGO, amdgpu_hwmon_get_fan1_min, NULL, 0); +static SENSOR_DEVICE_ATTR(fan1_max, S_IRUGO, amdgpu_hwmon_get_fan1_max, NULL, 0); +static SENSOR_DEVICE_ATTR(fan1_target, S_IRUGO | S_IWUSR, amdgpu_hwmon_get_fan1_target, amdgpu_hwmon_set_fan1_target, 0); +static SENSOR_DEVICE_ATTR(fan1_enable, S_IRUGO | S_IWUSR, amdgpu_hwmon_get_fan1_enable, amdgpu_hwmon_set_fan1_enable, 0); static SENSOR_DEVICE_ATTR(in0_input, S_IRUGO, amdgpu_hwmon_show_vddgfx, NULL, 0); static SENSOR_DEVICE_ATTR(in0_label, S_IRUGO, amdgpu_hwmon_show_vddgfx_label, NULL, 0); static SENSOR_DEVICE_ATTR(in1_input, S_IRUGO, amdgpu_hwmon_show_vddnb, NULL, 0); @@ -1438,6 +1599,10 @@ static struct attribute *hwmon_attributes[] = { &sensor_dev_attr_pwm1_min.dev_attr.attr, &sensor_dev_attr_pwm1_max.dev_attr.attr, &sensor_dev_attr_fan1_input.dev_attr.attr, + &sensor_dev_attr_fan1_min.dev_attr.attr, + &sensor_dev_attr_fan1_max.dev_attr.attr, + &sensor_dev_attr_fan1_target.dev_attr.attr, + &sensor_dev_attr_fan1_enable.dev_attr.attr, &sensor_dev_attr_in0_input.dev_attr.attr, &sensor_dev_attr_in0_label.dev_attr.attr, &sensor_dev_attr_in1_input.dev_attr.attr, @@ -1456,13 +1621,16 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj, struct amdgpu_device *adev = dev_get_drvdata(dev); umode_t effective_mode = attr->mode; - /* Skip fan attributes if fan is not present */ if (adev->pm.no_fan && (attr == &sensor_dev_attr_pwm1.dev_attr.attr || attr == &sensor_dev_attr_pwm1_enable.dev_attr.attr || attr == &sensor_dev_attr_pwm1_max.dev_attr.attr || attr == &sensor_dev_attr_pwm1_min.dev_attr.attr || - attr == &sensor_dev_attr_fan1_input.dev_attr.attr)) + attr == &sensor_dev_attr_fan1_input.dev_attr.attr || + attr == &sensor_dev_attr_fan1_min.dev_attr.attr || + attr == &sensor_dev_attr_fan1_max.dev_attr.attr || + attr == &sensor_dev_attr_fan1_target.dev_attr.attr || + attr == &sensor_dev_attr_fan1_enable.dev_attr.attr)) return 0; /* Skip limit attributes if DPM is not enabled */ @@ -1472,7 +1640,12 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj, attr == &sensor_dev_attr_pwm1.dev_attr.attr || attr == &sensor_dev_attr_pwm1_enable.dev_attr.attr || attr == &sensor_dev_attr_pwm1_max.dev_attr.attr || - attr == &sensor_dev_attr_pwm1_min.dev_attr.attr)) + attr == &sensor_dev_attr_pwm1_min.dev_attr.attr || + attr == &sensor_dev_attr_fan1_input.dev_attr.attr || + attr == &sensor_dev_attr_fan1_min.dev_attr.attr || + attr == &sensor_dev_attr_fan1_max.dev_attr.attr || + attr == &sensor_dev_attr_fan1_target.dev_attr.attr || + attr == &sensor_dev_attr_fan1_enable.dev_attr.attr)) return 0; /* mask fan attributes if we have no bindings for this asic to expose */ @@ -1497,10 +1670,18 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj, /* hide max/min values if we can't both query and manage the fan */ if ((!adev->powerplay.pp_funcs->set_fan_speed_percent && !adev->powerplay.pp_funcs->get_fan_speed_percent) && + (!adev->powerplay.pp_funcs->set_fan_speed_rpm && + !adev->powerplay.pp_funcs->get_fan_speed_rpm) && (attr == &sensor_dev_attr_pwm1_max.dev_attr.attr || attr == &sensor_dev_attr_pwm1_min.dev_attr.attr)) return 0; + if ((!adev->powerplay.pp_funcs->set_fan_speed_rpm && + !adev->powerplay.pp_funcs->get_fan_speed_rpm) && + (attr == &sensor_dev_attr_fan1_max.dev_attr.attr || + attr == &sensor_dev_attr_fan1_min.dev_attr.attr)) + return 0; + /* only APUs have vddnb */ if (!(adev->flags & AMD_IS_APU) && (attr == &sensor_dev_attr_in1_input.dev_attr.attr || @@ -1976,6 +2157,7 @@ void amdgpu_pm_compute_clocks(struct amdgpu_device *adev) static int amdgpu_debugfs_pm_info_pp(struct seq_file *m, struct amdgpu_device *adev) { uint32_t value; + uint64_t value64; uint32_t query = 0; int size; @@ -2014,6 +2196,10 @@ static int amdgpu_debugfs_pm_info_pp(struct seq_file *m, struct amdgpu_device *a seq_printf(m, "GPU Load: %u %%\n", value); seq_printf(m, "\n"); + /* SMC feature mask */ + if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_ENABLED_SMC_FEATURES_MASK, (void *)&value64, &size)) + seq_printf(m, "SMC Feature Mask: 0x%016llx\n", value64); + /* UVD clocks */ if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_UVD_POWER, (void *)&value, &size)) { if (!value) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index bd397d2916fb..25d2f3e757f1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c @@ -452,8 +452,6 @@ static int psp_hw_fini(void *handle) if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) return 0; - amdgpu_ucode_fini_bo(adev); - psp_ring_destroy(psp, PSP_RING_TYPE__KM); amdgpu_bo_free_kernel(&psp->tmr_bo, &psp->tmr_mc_addr, &psp->tmr_buf); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index 44fc665e4577..4caa301ce454 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -77,6 +77,7 @@ struct amdgpu_fence_driver { bool initialized; struct amdgpu_irq_src *irq_src; unsigned irq_type; + struct timer_list fallback_timer; unsigned num_fences_mask; spinlock_t lock; struct dma_fence **fences; @@ -96,7 +97,7 @@ void amdgpu_fence_driver_resume(struct amdgpu_device *adev); int amdgpu_fence_emit(struct amdgpu_ring *ring, struct dma_fence **fence, unsigned flags); int amdgpu_fence_emit_polling(struct amdgpu_ring *ring, uint32_t *s); -void amdgpu_fence_process(struct amdgpu_ring *ring); +bool amdgpu_fence_process(struct amdgpu_ring *ring); int amdgpu_fence_wait_empty(struct amdgpu_ring *ring); signed long amdgpu_fence_wait_polling(struct amdgpu_ring *ring, uint32_t wait_seq, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h index d17503f0df8e..500113ec65ca 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h @@ -46,10 +46,6 @@ struct amdgpu_sdma_instance { struct amdgpu_sdma { struct amdgpu_sdma_instance instance[AMDGPU_MAX_SDMA_INSTANCES]; -#ifdef CONFIG_DRM_AMDGPU_SI - //SI DMA has a difference trap irq number for the second engine - struct amdgpu_irq_src trap_irq_1; -#endif struct amdgpu_irq_src trap_irq; struct amdgpu_irq_src illegal_inst_irq; int num_instances; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h index 2e87414422f9..e9bf70e2ac51 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h @@ -103,7 +103,7 @@ TRACE_EVENT(amdgpu_iv, __entry->src_data[2] = iv->src_data[2]; __entry->src_data[3] = iv->src_data[3]; ), - TP_printk("client_id:%u src_id:%u ring:%u vmid:%u timestamp: %llu pasid:%u src_data: %08x %08x %08x %08x\n", + TP_printk("client_id:%u src_id:%u ring:%u vmid:%u timestamp: %llu pasid:%u src_data: %08x %08x %08x %08x", __entry->client_id, __entry->src_id, __entry->ring_id, __entry->vmid, __entry->timestamp, __entry->pasid, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace_points.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace_points.c index b160b958e5fe..f212402570a5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace_points.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace_points.c @@ -1,4 +1,4 @@ -// SPDX-License-Identifier: GPL-2.0 +// SPDX-License-Identifier: MIT /* Copyright Red Hat Inc 2010. * * Permission is hereby granted, free of charge, to any person obtaining a diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c index 1fa8bc337859..7b33867036e7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c @@ -297,10 +297,7 @@ amdgpu_ucode_get_load_type(struct amdgpu_device *adev, int load_type) case CHIP_POLARIS11: case CHIP_POLARIS12: case CHIP_VEGAM: - if (!load_type) - return AMDGPU_FW_LOAD_DIRECT; - else - return AMDGPU_FW_LOAD_SMU; + return AMDGPU_FW_LOAD_SMU; case CHIP_VEGA10: case CHIP_RAVEN: case CHIP_VEGA12: @@ -423,32 +420,41 @@ static int amdgpu_ucode_patch_jt(struct amdgpu_firmware_info *ucode, return 0; } -int amdgpu_ucode_init_bo(struct amdgpu_device *adev) +int amdgpu_ucode_create_bo(struct amdgpu_device *adev) { - uint64_t fw_offset = 0; - int i, err; - struct amdgpu_firmware_info *ucode = NULL; - const struct common_firmware_header *header = NULL; - - if (!adev->firmware.fw_size) { - dev_warn(adev->dev, "No ip firmware need to load\n"); - return 0; - } - - if (!adev->in_gpu_reset) { - err = amdgpu_bo_create_kernel(adev, adev->firmware.fw_size, PAGE_SIZE, - amdgpu_sriov_vf(adev) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT, - &adev->firmware.fw_buf, - &adev->firmware.fw_buf_mc, - &adev->firmware.fw_buf_ptr); - if (err) { + if (adev->firmware.load_type != AMDGPU_FW_LOAD_DIRECT) { + amdgpu_bo_create_kernel(adev, adev->firmware.fw_size, PAGE_SIZE, + amdgpu_sriov_vf(adev) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT, + &adev->firmware.fw_buf, + &adev->firmware.fw_buf_mc, + &adev->firmware.fw_buf_ptr); + if (!adev->firmware.fw_buf) { dev_err(adev->dev, "failed to create kernel buffer for firmware.fw_buf\n"); - goto failed; + return -ENOMEM; + } else if (amdgpu_sriov_vf(adev)) { + memset(adev->firmware.fw_buf_ptr, 0, adev->firmware.fw_size); } } + return 0; +} + +void amdgpu_ucode_free_bo(struct amdgpu_device *adev) +{ + if (adev->firmware.load_type != AMDGPU_FW_LOAD_DIRECT) + amdgpu_bo_free_kernel(&adev->firmware.fw_buf, + &adev->firmware.fw_buf_mc, + &adev->firmware.fw_buf_ptr); +} - memset(adev->firmware.fw_buf_ptr, 0, adev->firmware.fw_size); +int amdgpu_ucode_init_bo(struct amdgpu_device *adev) +{ + uint64_t fw_offset = 0; + int i; + struct amdgpu_firmware_info *ucode = NULL; + /* for baremetal, the ucode is allocated in gtt, so don't need to fill the bo when reset/suspend */ + if (!amdgpu_sriov_vf(adev) && (adev->in_gpu_reset || adev->in_suspend)) + return 0; /* * if SMU loaded firmware, it needn't add SMC, UVD, and VCE * ucode info here @@ -465,7 +471,6 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev) for (i = 0; i < adev->firmware.max_ucodes; i++) { ucode = &adev->firmware.ucode[i]; if (ucode->fw) { - header = (const struct common_firmware_header *)ucode->fw->data; amdgpu_ucode_init_single_fw(adev, ucode, adev->firmware.fw_buf_mc + fw_offset, adev->firmware.fw_buf_ptr + fw_offset); if (i == AMDGPU_UCODE_ID_CP_MEC1 && @@ -480,33 +485,4 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev) } } return 0; - -failed: - if (err) - adev->firmware.load_type = AMDGPU_FW_LOAD_DIRECT; - - return err; -} - -int amdgpu_ucode_fini_bo(struct amdgpu_device *adev) -{ - int i; - struct amdgpu_firmware_info *ucode = NULL; - - if (!adev->firmware.fw_size) - return 0; - - for (i = 0; i < adev->firmware.max_ucodes; i++) { - ucode = &adev->firmware.ucode[i]; - if (ucode->fw) { - ucode->mc_addr = 0; - ucode->kaddr = NULL; - } - } - - amdgpu_bo_free_kernel(&adev->firmware.fw_buf, - &adev->firmware.fw_buf_mc, - &adev->firmware.fw_buf_ptr); - - return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h index 8f3f1117728c..aa6641b944a0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h @@ -276,8 +276,10 @@ void amdgpu_ucode_print_gpu_info_hdr(const struct common_firmware_header *hdr); int amdgpu_ucode_validate(const struct firmware *fw); bool amdgpu_ucode_hdr_version(union amdgpu_firmware_header *hdr, uint16_t hdr_major, uint16_t hdr_minor); + int amdgpu_ucode_init_bo(struct amdgpu_device *adev); -int amdgpu_ucode_fini_bo(struct amdgpu_device *adev); +int amdgpu_ucode_create_bo(struct amdgpu_device *adev); +void amdgpu_ucode_free_bo(struct amdgpu_device *adev); enum amdgpu_firmware_load_type amdgpu_ucode_get_load_type(struct amdgpu_device *adev, int load_type); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c index 0cc5190f4f36..5f3f54073818 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c @@ -258,6 +258,8 @@ int amdgpu_vce_suspend(struct amdgpu_device *adev) { int i; + cancel_delayed_work_sync(&adev->vce.idle_work); + if (adev->vce.vcpu_bo == NULL) return 0; @@ -268,7 +270,6 @@ int amdgpu_vce_suspend(struct amdgpu_device *adev) if (i == AMDGPU_MAX_VCE_HANDLES) return 0; - cancel_delayed_work_sync(&adev->vce.idle_work); /* TODO: suspending running encoding sessions isn't supported */ return -EINVAL; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index a73674f9a0f5..27da13df2f11 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -36,6 +36,7 @@ #include "soc15_common.h" #include "vcn/vcn_1_0_offset.h" +#include "vcn/vcn_1_0_sh_mask.h" /* 1 second timeout */ #define VCN_IDLE_TIMEOUT msecs_to_jiffies(1000) @@ -120,8 +121,7 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev) version_major, version_minor, family_id); } - bo_size = AMDGPU_VCN_STACK_SIZE + AMDGPU_VCN_HEAP_SIZE - + AMDGPU_VCN_SESSION_SIZE * 40; + bo_size = AMDGPU_VCN_STACK_SIZE + AMDGPU_VCN_CONTEXT_SIZE; if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) bo_size += AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8); r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE, @@ -162,11 +162,11 @@ int amdgpu_vcn_suspend(struct amdgpu_device *adev) unsigned size; void *ptr; + cancel_delayed_work_sync(&adev->vcn.idle_work); + if (adev->vcn.vcpu_bo == NULL) return 0; - cancel_delayed_work_sync(&adev->vcn.idle_work); - size = amdgpu_bo_size(adev->vcn.vcpu_bo); ptr = adev->vcn.cpu_addr; @@ -212,18 +212,161 @@ int amdgpu_vcn_resume(struct amdgpu_device *adev) return 0; } +static int amdgpu_vcn_pause_dpg_mode(struct amdgpu_device *adev, + struct dpg_pause_state *new_state) +{ + int ret_code; + uint32_t reg_data = 0; + uint32_t reg_data2 = 0; + struct amdgpu_ring *ring; + + /* pause/unpause if state is changed */ + if (adev->vcn.pause_state.fw_based != new_state->fw_based) { + DRM_DEBUG("dpg pause state changed %d:%d -> %d:%d", + adev->vcn.pause_state.fw_based, adev->vcn.pause_state.jpeg, + new_state->fw_based, new_state->jpeg); + + reg_data = RREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE) & + (~UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK); + + if (new_state->fw_based == VCN_DPG_STATE__PAUSE) { + ret_code = 0; + + if (!(reg_data & UVD_DPG_PAUSE__JPEG_PAUSE_DPG_ACK_MASK)) + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS, + UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF, + UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code); + + if (!ret_code) { + /* pause DPG non-jpeg */ + reg_data |= UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK; + WREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE, reg_data); + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_DPG_PAUSE, + UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK, + UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK, ret_code); + + /* Restore */ + ring = &adev->vcn.ring_enc[0]; + WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_LO, ring->gpu_addr); + WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(UVD, 0, mmUVD_RB_SIZE, ring->ring_size / 4); + WREG32_SOC15(UVD, 0, mmUVD_RB_RPTR, lower_32_bits(ring->wptr)); + WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR, lower_32_bits(ring->wptr)); + + ring = &adev->vcn.ring_enc[1]; + WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_LO2, ring->gpu_addr); + WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(UVD, 0, mmUVD_RB_SIZE2, ring->ring_size / 4); + WREG32_SOC15(UVD, 0, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr)); + WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr)); + + ring = &adev->vcn.ring_dec; + WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, + RREG32_SOC15(UVD, 0, mmUVD_SCRATCH2)); + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS, + UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON, + UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code); + } + } else { + /* unpause dpg non-jpeg, no need to wait */ + reg_data &= ~UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK; + WREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE, reg_data); + } + adev->vcn.pause_state.fw_based = new_state->fw_based; + } + + /* pause/unpause if state is changed */ + if (adev->vcn.pause_state.jpeg != new_state->jpeg) { + DRM_DEBUG("dpg pause state changed %d:%d -> %d:%d", + adev->vcn.pause_state.fw_based, adev->vcn.pause_state.jpeg, + new_state->fw_based, new_state->jpeg); + + reg_data = RREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE) & + (~UVD_DPG_PAUSE__JPEG_PAUSE_DPG_ACK_MASK); + + if (new_state->jpeg == VCN_DPG_STATE__PAUSE) { + ret_code = 0; + + if (!(reg_data & UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK)) + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS, + UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF, + UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code); + + if (!ret_code) { + /* Make sure JPRG Snoop is disabled before sending the pause */ + reg_data2 = RREG32_SOC15(UVD, 0, mmUVD_POWER_STATUS); + reg_data2 |= UVD_POWER_STATUS__JRBC_SNOOP_DIS_MASK; + WREG32_SOC15(UVD, 0, mmUVD_POWER_STATUS, reg_data2); + + /* pause DPG jpeg */ + reg_data |= UVD_DPG_PAUSE__JPEG_PAUSE_DPG_REQ_MASK; + WREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE, reg_data); + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_DPG_PAUSE, + UVD_DPG_PAUSE__JPEG_PAUSE_DPG_ACK_MASK, + UVD_DPG_PAUSE__JPEG_PAUSE_DPG_ACK_MASK, ret_code); + + /* Restore */ + ring = &adev->vcn.ring_jpeg; + 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 | + UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK); + WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_64BIT_BAR_LOW, + lower_32_bits(ring->gpu_addr)); + WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_64BIT_BAR_HIGH, + upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_RPTR, ring->wptr); + WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR, ring->wptr); + WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, + UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK); + + ring = &adev->vcn.ring_dec; + WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, + RREG32_SOC15(UVD, 0, mmUVD_SCRATCH2)); + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS, + UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON, + UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code); + } + } else { + /* unpause dpg jpeg, no need to wait */ + reg_data &= ~UVD_DPG_PAUSE__JPEG_PAUSE_DPG_REQ_MASK; + WREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE, reg_data); + } + adev->vcn.pause_state.jpeg = new_state->jpeg; + } + + return 0; +} + static void amdgpu_vcn_idle_work_handler(struct work_struct *work) { struct amdgpu_device *adev = container_of(work, struct amdgpu_device, vcn.idle_work.work); - unsigned fences = amdgpu_fence_count_emitted(&adev->vcn.ring_dec); - unsigned i; + unsigned int fences = 0; + unsigned int i; for (i = 0; i < adev->vcn.num_enc_rings; ++i) { fences += amdgpu_fence_count_emitted(&adev->vcn.ring_enc[i]); } + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { + struct dpg_pause_state new_state; + + if (fences) + new_state.fw_based = VCN_DPG_STATE__PAUSE; + else + new_state.fw_based = VCN_DPG_STATE__UNPAUSE; + + if (amdgpu_fence_count_emitted(&adev->vcn.ring_jpeg)) + new_state.jpeg = VCN_DPG_STATE__PAUSE; + else + new_state.jpeg = VCN_DPG_STATE__UNPAUSE; + + amdgpu_vcn_pause_dpg_mode(adev, &new_state); + } + fences += amdgpu_fence_count_emitted(&adev->vcn.ring_jpeg); + fences += amdgpu_fence_count_emitted(&adev->vcn.ring_dec); if (fences == 0) { amdgpu_gfx_off_ctrl(adev, true); @@ -250,6 +393,22 @@ void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring) amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN, AMD_PG_STATE_UNGATE); } + + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { + struct dpg_pause_state new_state; + + if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC) + new_state.fw_based = VCN_DPG_STATE__PAUSE; + else + new_state.fw_based = adev->vcn.pause_state.fw_based; + + if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_JPEG) + new_state.jpeg = VCN_DPG_STATE__PAUSE; + else + new_state.jpeg = adev->vcn.pause_state.jpeg; + + amdgpu_vcn_pause_dpg_mode(adev, &new_state); + } } void amdgpu_vcn_ring_end_use(struct amdgpu_ring *ring) @@ -264,7 +423,7 @@ int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring) unsigned i; int r; - WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID), 0xCAFEDEAD); + WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9), 0xCAFEDEAD); r = amdgpu_ring_alloc(ring, 3); if (r) { DRM_ERROR("amdgpu: cp failed to lock ring %d (%d).\n", @@ -272,11 +431,11 @@ int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring) return r; } amdgpu_ring_write(ring, - PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID), 0)); + PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9), 0)); amdgpu_ring_write(ring, 0xDEADBEEF); amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID)); + tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9)); if (tmp == 0xDEADBEEF) break; DRM_UDELAY(1); @@ -616,7 +775,7 @@ int amdgpu_vcn_jpeg_ring_test_ring(struct amdgpu_ring *ring) unsigned i; int r; - WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID), 0xCAFEDEAD); + WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9), 0xCAFEDEAD); r = amdgpu_ring_alloc(ring, 3); if (r) { @@ -626,12 +785,12 @@ int amdgpu_vcn_jpeg_ring_test_ring(struct amdgpu_ring *ring) } amdgpu_ring_write(ring, - PACKETJ(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID), 0, 0, 0)); + PACKETJ(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9), 0, 0, 0)); amdgpu_ring_write(ring, 0xDEADBEEF); amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID)); + tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9)); if (tmp == 0xDEADBEEF) break; DRM_UDELAY(1); @@ -665,7 +824,7 @@ static int amdgpu_vcn_jpeg_set_reg(struct amdgpu_ring *ring, uint32_t handle, ib = &job->ibs[0]; - ib->ptr[0] = PACKETJ(SOC15_REG_OFFSET(UVD, 0, mmUVD_JPEG_PITCH), 0, 0, PACKETJ_TYPE0); + ib->ptr[0] = PACKETJ(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9), 0, 0, PACKETJ_TYPE0); ib->ptr[1] = 0xDEADBEEF; for (i = 2; i < 16; i += 2) { ib->ptr[i] = PACKETJ(0, 0, 0, PACKETJ_TYPE6); @@ -714,7 +873,7 @@ int amdgpu_vcn_jpeg_ring_test_ib(struct amdgpu_ring *ring, long timeout) r = 0; for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_JPEG_PITCH)); + tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9)); if (tmp == 0xDEADBEEF) break; DRM_UDELAY(1); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index 0b0b8638d73f..a0ad19af9080 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -24,9 +24,9 @@ #ifndef __AMDGPU_VCN_H__ #define __AMDGPU_VCN_H__ -#define AMDGPU_VCN_STACK_SIZE (200*1024) -#define AMDGPU_VCN_HEAP_SIZE (256*1024) -#define AMDGPU_VCN_SESSION_SIZE (50*1024) +#define AMDGPU_VCN_STACK_SIZE (128*1024) +#define AMDGPU_VCN_CONTEXT_SIZE (512*1024) + #define AMDGPU_VCN_FIRMWARE_OFFSET 256 #define AMDGPU_VCN_MAX_ENC_RINGS 3 @@ -56,6 +56,16 @@ enum engine_status_constants { UVD_STATUS__RBC_BUSY = 0x1, }; +enum internal_dpg_state { + VCN_DPG_STATE__UNPAUSE = 0, + VCN_DPG_STATE__PAUSE, +}; + +struct dpg_pause_state { + enum internal_dpg_state fw_based; + enum internal_dpg_state jpeg; +}; + struct amdgpu_vcn { struct amdgpu_bo *vcpu_bo; void *cpu_addr; @@ -69,6 +79,8 @@ struct amdgpu_vcn { struct amdgpu_ring ring_jpeg; struct amdgpu_irq_src irq; unsigned num_enc_rings; + enum amd_powergating_state cur_state; + struct dpg_pause_state pause_state; }; int amdgpu_vcn_sw_init(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c index d2469453dca2..79220a91abe3 100644 --- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c @@ -6277,12 +6277,12 @@ static int ci_dpm_sw_init(void *handle) int ret; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - ret = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 230, + ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 230, &adev->pm.dpm.thermal.irq); if (ret) return ret; - ret = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 231, + ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 231, &adev->pm.dpm.thermal.irq); if (ret) return ret; diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c index 78ab939ae5d8..f41f5f57e9f3 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik.c +++ b/drivers/gpu/drm/amd/amdgpu/cik.c @@ -2002,6 +2002,8 @@ int cik_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &cik_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v7_0_ip_block); amdgpu_device_ip_block_add(adev, &cik_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v7_2_ip_block); + amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block); if (amdgpu_dpm == -1) amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); else @@ -2014,8 +2016,6 @@ int cik_set_ip_blocks(struct amdgpu_device *adev) #endif else amdgpu_device_ip_block_add(adev, &dce_v8_2_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v7_2_ip_block); - amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block); amdgpu_device_ip_block_add(adev, &uvd_v4_2_ip_block); amdgpu_device_ip_block_add(adev, &vce_v2_0_ip_block); break; @@ -2023,6 +2023,8 @@ int cik_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &cik_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v7_0_ip_block); amdgpu_device_ip_block_add(adev, &cik_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v7_3_ip_block); + amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block); if (amdgpu_dpm == -1) amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); else @@ -2035,8 +2037,6 @@ int cik_set_ip_blocks(struct amdgpu_device *adev) #endif else amdgpu_device_ip_block_add(adev, &dce_v8_5_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v7_3_ip_block); - amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block); amdgpu_device_ip_block_add(adev, &uvd_v4_2_ip_block); amdgpu_device_ip_block_add(adev, &vce_v2_0_ip_block); break; @@ -2044,6 +2044,8 @@ int cik_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &cik_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v7_0_ip_block); amdgpu_device_ip_block_add(adev, &cik_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v7_1_ip_block); + amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block); amdgpu_device_ip_block_add(adev, &kv_smu_ip_block); if (adev->enable_virtual_display) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); @@ -2053,8 +2055,7 @@ int cik_set_ip_blocks(struct amdgpu_device *adev) #endif else amdgpu_device_ip_block_add(adev, &dce_v8_1_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v7_1_ip_block); - amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block); + amdgpu_device_ip_block_add(adev, &uvd_v4_2_ip_block); amdgpu_device_ip_block_add(adev, &vce_v2_0_ip_block); break; @@ -2063,6 +2064,8 @@ int cik_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &cik_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v7_0_ip_block); amdgpu_device_ip_block_add(adev, &cik_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v7_2_ip_block); + amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block); amdgpu_device_ip_block_add(adev, &kv_smu_ip_block); if (adev->enable_virtual_display) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); @@ -2072,8 +2075,6 @@ int cik_set_ip_blocks(struct amdgpu_device *adev) #endif else amdgpu_device_ip_block_add(adev, &dce_v8_3_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v7_2_ip_block); - amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block); amdgpu_device_ip_block_add(adev, &uvd_v4_2_ip_block); amdgpu_device_ip_block_add(adev, &vce_v2_0_ip_block); break; diff --git a/drivers/gpu/drm/amd/amdgpu/cik_ih.c b/drivers/gpu/drm/amd/amdgpu/cik_ih.c index 44d10c2172f6..b5775c6a857b 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_ih.c @@ -276,7 +276,7 @@ static void cik_ih_decode_iv(struct amdgpu_device *adev, dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]); dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]); - entry->client_id = AMDGPU_IH_CLIENTID_LEGACY; + entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY; entry->src_id = dw[0] & 0xff; entry->src_data[0] = dw[1] & 0xfffffff; entry->ring_id = dw[2] & 0xff; @@ -318,7 +318,7 @@ static int cik_ih_sw_init(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = amdgpu_ih_ring_init(adev, 64 * 1024, false); + r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, false); if (r) return r; @@ -332,7 +332,7 @@ static int cik_ih_sw_fini(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; amdgpu_irq_fini(adev); - amdgpu_ih_ring_fini(adev); + amdgpu_ih_ring_fini(adev, &adev->irq.ih); amdgpu_irq_remove_domain(adev); return 0; @@ -468,8 +468,7 @@ static const struct amdgpu_ih_funcs cik_ih_funcs = { static void cik_ih_set_interrupt_funcs(struct amdgpu_device *adev) { - if (adev->irq.ih_funcs == NULL) - adev->irq.ih_funcs = &cik_ih_funcs; + adev->irq.ih_funcs = &cik_ih_funcs; } const struct amdgpu_ip_block_version cik_ih_ip_block = diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c index 154b1499b07e..b918c8886b75 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c @@ -970,19 +970,19 @@ static int cik_sdma_sw_init(void *handle) } /* SDMA trap event */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 224, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 224, &adev->sdma.trap_irq); if (r) return r; /* SDMA Privileged inst */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 241, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 241, &adev->sdma.illegal_inst_irq); if (r) return r; /* SDMA Privileged inst */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 247, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 247, &adev->sdma.illegal_inst_irq); if (r) return r; @@ -1370,10 +1370,8 @@ static const struct amdgpu_buffer_funcs cik_sdma_buffer_funcs = { static void cik_sdma_set_buffer_funcs(struct amdgpu_device *adev) { - if (adev->mman.buffer_funcs == NULL) { - adev->mman.buffer_funcs = &cik_sdma_buffer_funcs; - adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring; - } + adev->mman.buffer_funcs = &cik_sdma_buffer_funcs; + adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring; } static const struct amdgpu_vm_pte_funcs cik_sdma_vm_pte_funcs = { @@ -1389,15 +1387,13 @@ static void cik_sdma_set_vm_pte_funcs(struct amdgpu_device *adev) struct drm_gpu_scheduler *sched; unsigned i; - if (adev->vm_manager.vm_pte_funcs == NULL) { - adev->vm_manager.vm_pte_funcs = &cik_sdma_vm_pte_funcs; - for (i = 0; i < adev->sdma.num_instances; i++) { - sched = &adev->sdma.instance[i].ring.sched; - adev->vm_manager.vm_pte_rqs[i] = - &sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL]; - } - adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances; + adev->vm_manager.vm_pte_funcs = &cik_sdma_vm_pte_funcs; + for (i = 0; i < adev->sdma.num_instances; i++) { + sched = &adev->sdma.instance[i].ring.sched; + adev->vm_manager.vm_pte_rqs[i] = + &sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL]; } + adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances; } const struct amdgpu_ip_block_version cik_sdma_ip_block = diff --git a/drivers/gpu/drm/amd/amdgpu/cz_ih.c b/drivers/gpu/drm/amd/amdgpu/cz_ih.c index 960c29e17da6..df5ac4d85a00 100644 --- a/drivers/gpu/drm/amd/amdgpu/cz_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/cz_ih.c @@ -255,7 +255,7 @@ static void cz_ih_decode_iv(struct amdgpu_device *adev, dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]); dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]); - entry->client_id = AMDGPU_IH_CLIENTID_LEGACY; + entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY; entry->src_id = dw[0] & 0xff; entry->src_data[0] = dw[1] & 0xfffffff; entry->ring_id = dw[2] & 0xff; @@ -297,7 +297,7 @@ static int cz_ih_sw_init(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = amdgpu_ih_ring_init(adev, 64 * 1024, false); + r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, false); if (r) return r; @@ -311,7 +311,7 @@ static int cz_ih_sw_fini(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; amdgpu_irq_fini(adev); - amdgpu_ih_ring_fini(adev); + amdgpu_ih_ring_fini(adev, &adev->irq.ih); amdgpu_irq_remove_domain(adev); return 0; @@ -449,8 +449,7 @@ static const struct amdgpu_ih_funcs cz_ih_funcs = { static void cz_ih_set_interrupt_funcs(struct amdgpu_device *adev) { - if (adev->irq.ih_funcs == NULL) - adev->irq.ih_funcs = &cz_ih_funcs; + adev->irq.ih_funcs = &cz_ih_funcs; } const struct amdgpu_ip_block_version cz_ih_ip_block = diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c index 3916aa6cc4ec..4cfecdce29a3 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c @@ -2746,19 +2746,19 @@ static int dce_v10_0_sw_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; for (i = 0; i < adev->mode_info.num_crtc; i++) { - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i + 1, &adev->crtc_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + 1, &adev->crtc_irq); if (r) return r; } for (i = VISLANDS30_IV_SRCID_D1_GRPH_PFLIP; i < 20; i += 2) { - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i, &adev->pageflip_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i, &adev->pageflip_irq); if (r) return r; } /* HPD hotplug */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq); if (r) return r; @@ -3570,8 +3570,7 @@ static const struct amdgpu_display_funcs dce_v10_0_display_funcs = { static void dce_v10_0_set_display_funcs(struct amdgpu_device *adev) { - if (adev->mode_info.funcs == NULL) - adev->mode_info.funcs = &dce_v10_0_display_funcs; + adev->mode_info.funcs = &dce_v10_0_display_funcs; } static const struct amdgpu_irq_src_funcs dce_v10_0_crtc_irq_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c index 4ffb612a4e53..7c868916d90f 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c @@ -2867,19 +2867,19 @@ static int dce_v11_0_sw_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; for (i = 0; i < adev->mode_info.num_crtc; i++) { - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i + 1, &adev->crtc_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + 1, &adev->crtc_irq); if (r) return r; } for (i = VISLANDS30_IV_SRCID_D1_GRPH_PFLIP; i < 20; i += 2) { - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i, &adev->pageflip_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i, &adev->pageflip_irq); if (r) return r; } /* HPD hotplug */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq); if (r) return r; @@ -3702,8 +3702,7 @@ static const struct amdgpu_display_funcs dce_v11_0_display_funcs = { static void dce_v11_0_set_display_funcs(struct amdgpu_device *adev) { - if (adev->mode_info.funcs == NULL) - adev->mode_info.funcs = &dce_v11_0_display_funcs; + adev->mode_info.funcs = &dce_v11_0_display_funcs; } static const struct amdgpu_irq_src_funcs dce_v11_0_crtc_irq_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c index 480c5348a14f..17eaaba36017 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c @@ -2616,19 +2616,19 @@ static int dce_v6_0_sw_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; for (i = 0; i < adev->mode_info.num_crtc; i++) { - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i + 1, &adev->crtc_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + 1, &adev->crtc_irq); if (r) return r; } for (i = 8; i < 20; i += 2) { - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i, &adev->pageflip_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i, &adev->pageflip_irq); if (r) return r; } /* HPD hotplug */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 42, &adev->hpd_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 42, &adev->hpd_irq); if (r) return r; @@ -3376,8 +3376,7 @@ static const struct amdgpu_display_funcs dce_v6_0_display_funcs = { static void dce_v6_0_set_display_funcs(struct amdgpu_device *adev) { - if (adev->mode_info.funcs == NULL) - adev->mode_info.funcs = &dce_v6_0_display_funcs; + adev->mode_info.funcs = &dce_v6_0_display_funcs; } static const struct amdgpu_irq_src_funcs dce_v6_0_crtc_irq_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c index 797196476c94..8c0576978d36 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c @@ -2643,19 +2643,19 @@ static int dce_v8_0_sw_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; for (i = 0; i < adev->mode_info.num_crtc; i++) { - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i + 1, &adev->crtc_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + 1, &adev->crtc_irq); if (r) return r; } for (i = 8; i < 20; i += 2) { - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i, &adev->pageflip_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i, &adev->pageflip_irq); if (r) return r; } /* HPD hotplug */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 42, &adev->hpd_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 42, &adev->hpd_irq); if (r) return r; @@ -3458,8 +3458,7 @@ static const struct amdgpu_display_funcs dce_v8_0_display_funcs = { static void dce_v8_0_set_display_funcs(struct amdgpu_device *adev) { - if (adev->mode_info.funcs == NULL) - adev->mode_info.funcs = &dce_v8_0_display_funcs; + adev->mode_info.funcs = &dce_v8_0_display_funcs; } static const struct amdgpu_irq_src_funcs dce_v8_0_crtc_irq_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/dce_virtual.c b/drivers/gpu/drm/amd/amdgpu/dce_virtual.c index 15257634a53a..fdace004544d 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_virtual.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_virtual.c @@ -372,7 +372,7 @@ static int dce_virtual_sw_init(void *handle) int r, i; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SMU_DISP_TIMER2_TRIGGER, &adev->crtc_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SMU_DISP_TIMER2_TRIGGER, &adev->crtc_irq); if (r) return r; @@ -649,8 +649,7 @@ static const struct amdgpu_display_funcs dce_virtual_display_funcs = { static void dce_virtual_set_display_funcs(struct amdgpu_device *adev) { - if (adev->mode_info.funcs == NULL) - adev->mode_info.funcs = &dce_virtual_display_funcs; + adev->mode_info.funcs = &dce_virtual_display_funcs; } static int dce_virtual_pageflip(struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c index de184a886057..d76eb27945dc 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c @@ -1552,7 +1552,7 @@ static void gfx_v6_0_config_init(struct amdgpu_device *adev) adev->gfx.config.double_offchip_lds_buf = 0; } -static void gfx_v6_0_gpu_init(struct amdgpu_device *adev) +static void gfx_v6_0_constants_init(struct amdgpu_device *adev) { u32 gb_addr_config = 0; u32 mc_shared_chmap, mc_arb_ramcfg; @@ -3094,15 +3094,15 @@ static int gfx_v6_0_sw_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; int i, r; - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 181, &adev->gfx.eop_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 181, &adev->gfx.eop_irq); if (r) return r; - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 184, &adev->gfx.priv_reg_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 184, &adev->gfx.priv_reg_irq); if (r) return r; - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 185, &adev->gfx.priv_inst_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 185, &adev->gfx.priv_inst_irq); if (r) return r; @@ -3175,7 +3175,7 @@ static int gfx_v6_0_hw_init(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - gfx_v6_0_gpu_init(adev); + gfx_v6_0_constants_init(adev); r = gfx_v6_0_rlc_resume(adev); if (r) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c index fc39ebbc9d9f..0e72bc09939a 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c @@ -1886,14 +1886,14 @@ static void gfx_v7_0_config_init(struct amdgpu_device *adev) } /** - * gfx_v7_0_gpu_init - setup the 3D engine + * gfx_v7_0_constants_init - setup the 3D engine * * @adev: amdgpu_device pointer * - * Configures the 3D engine and tiling configuration - * registers so that the 3D engine is usable. + * init the gfx constants such as the 3D engine, tiling configuration + * registers, maximum number of quad pipes, render backends... */ -static void gfx_v7_0_gpu_init(struct amdgpu_device *adev) +static void gfx_v7_0_constants_init(struct amdgpu_device *adev) { u32 sh_mem_cfg, sh_static_mem_cfg, sh_mem_base; u32 tmp; @@ -4516,18 +4516,18 @@ static int gfx_v7_0_sw_init(void *handle) adev->gfx.mec.num_queue_per_pipe = 8; /* EOP Event */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 181, &adev->gfx.eop_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 181, &adev->gfx.eop_irq); if (r) return r; /* Privileged reg */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 184, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 184, &adev->gfx.priv_reg_irq); if (r) return r; /* Privileged inst */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 185, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 185, &adev->gfx.priv_inst_irq); if (r) return r; @@ -4624,7 +4624,7 @@ static int gfx_v7_0_hw_init(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - gfx_v7_0_gpu_init(adev); + gfx_v7_0_constants_init(adev); /* init rlc */ r = gfx_v7_0_rlc_resume(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index 470dc80f4fe7..3d0f277a6523 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -1173,64 +1173,61 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev) } } - if (adev->firmware.load_type == AMDGPU_FW_LOAD_SMU) { - info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_PFP]; - info->ucode_id = AMDGPU_UCODE_ID_CP_PFP; - info->fw = adev->gfx.pfp_fw; - header = (const struct common_firmware_header *)info->fw->data; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); - - info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_ME]; - info->ucode_id = AMDGPU_UCODE_ID_CP_ME; - info->fw = adev->gfx.me_fw; - header = (const struct common_firmware_header *)info->fw->data; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); - - info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_CE]; - info->ucode_id = AMDGPU_UCODE_ID_CP_CE; - info->fw = adev->gfx.ce_fw; - header = (const struct common_firmware_header *)info->fw->data; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); + info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_PFP]; + info->ucode_id = AMDGPU_UCODE_ID_CP_PFP; + info->fw = adev->gfx.pfp_fw; + header = (const struct common_firmware_header *)info->fw->data; + adev->firmware.fw_size += + ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); + + info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_ME]; + info->ucode_id = AMDGPU_UCODE_ID_CP_ME; + info->fw = adev->gfx.me_fw; + header = (const struct common_firmware_header *)info->fw->data; + adev->firmware.fw_size += + ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); + + info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_CE]; + info->ucode_id = AMDGPU_UCODE_ID_CP_CE; + info->fw = adev->gfx.ce_fw; + header = (const struct common_firmware_header *)info->fw->data; + adev->firmware.fw_size += + ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); + + info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_G]; + info->ucode_id = AMDGPU_UCODE_ID_RLC_G; + info->fw = adev->gfx.rlc_fw; + header = (const struct common_firmware_header *)info->fw->data; + adev->firmware.fw_size += + ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); + + info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC1]; + info->ucode_id = AMDGPU_UCODE_ID_CP_MEC1; + info->fw = adev->gfx.mec_fw; + header = (const struct common_firmware_header *)info->fw->data; + adev->firmware.fw_size += + ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); + + /* we need account JT in */ + cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data; + adev->firmware.fw_size += + ALIGN(le32_to_cpu(cp_hdr->jt_size) << 2, PAGE_SIZE); - info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_G]; - info->ucode_id = AMDGPU_UCODE_ID_RLC_G; - info->fw = adev->gfx.rlc_fw; - header = (const struct common_firmware_header *)info->fw->data; + if (amdgpu_sriov_vf(adev)) { + info = &adev->firmware.ucode[AMDGPU_UCODE_ID_STORAGE]; + info->ucode_id = AMDGPU_UCODE_ID_STORAGE; + info->fw = adev->gfx.mec_fw; adev->firmware.fw_size += - ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); + ALIGN(le32_to_cpu(64 * PAGE_SIZE), PAGE_SIZE); + } - info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC1]; - info->ucode_id = AMDGPU_UCODE_ID_CP_MEC1; - info->fw = adev->gfx.mec_fw; + if (adev->gfx.mec2_fw) { + info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC2]; + info->ucode_id = AMDGPU_UCODE_ID_CP_MEC2; + info->fw = adev->gfx.mec2_fw; header = (const struct common_firmware_header *)info->fw->data; adev->firmware.fw_size += ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); - - /* we need account JT in */ - cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(cp_hdr->jt_size) << 2, PAGE_SIZE); - - if (amdgpu_sriov_vf(adev)) { - info = &adev->firmware.ucode[AMDGPU_UCODE_ID_STORAGE]; - info->ucode_id = AMDGPU_UCODE_ID_STORAGE; - info->fw = adev->gfx.mec_fw; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(64 * PAGE_SIZE), PAGE_SIZE); - } - - if (adev->gfx.mec2_fw) { - info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC2]; - info->ucode_id = AMDGPU_UCODE_ID_CP_MEC2; - info->fw = adev->gfx.mec2_fw; - header = (const struct common_firmware_header *)info->fw->data; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); - } - } out: @@ -2048,36 +2045,31 @@ static int gfx_v8_0_sw_init(void *handle) adev->gfx.mec.num_pipe_per_mec = 4; adev->gfx.mec.num_queue_per_pipe = 8; - /* KIQ event */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_INT_IB2, &adev->gfx.kiq.irq); - if (r) - return r; - /* EOP Event */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_END_OF_PIPE, &adev->gfx.eop_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_END_OF_PIPE, &adev->gfx.eop_irq); if (r) return r; /* Privileged reg */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_PRIV_REG_FAULT, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_PRIV_REG_FAULT, &adev->gfx.priv_reg_irq); if (r) return r; /* Privileged inst */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_PRIV_INSTR_FAULT, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_PRIV_INSTR_FAULT, &adev->gfx.priv_inst_irq); if (r) return r; /* Add CP EDC/ECC irq */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_ECC_ERROR, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_ECC_ERROR, &adev->gfx.cp_ecc_error_irq); if (r) return r; /* SQ interrupts. */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SQ_INTERRUPT_MSG, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SQ_INTERRUPT_MSG, &adev->gfx.sq_irq); if (r) { DRM_ERROR("amdgpu_irq_add() for SQ failed: %d\n", r); @@ -3835,7 +3827,7 @@ static void gfx_v8_0_config_init(struct amdgpu_device *adev) } } -static void gfx_v8_0_gpu_init(struct amdgpu_device *adev) +static void gfx_v8_0_constants_init(struct amdgpu_device *adev) { u32 tmp, sh_static_mem_cfg; int i; @@ -4181,65 +4173,11 @@ static void gfx_v8_0_rlc_start(struct amdgpu_device *adev) udelay(50); } -static int gfx_v8_0_rlc_load_microcode(struct amdgpu_device *adev) -{ - const struct rlc_firmware_header_v2_0 *hdr; - const __le32 *fw_data; - unsigned i, fw_size; - - if (!adev->gfx.rlc_fw) - return -EINVAL; - - hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data; - amdgpu_ucode_print_rlc_hdr(&hdr->header); - - fw_data = (const __le32 *)(adev->gfx.rlc_fw->data + - le32_to_cpu(hdr->header.ucode_array_offset_bytes)); - fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4; - - WREG32(mmRLC_GPM_UCODE_ADDR, 0); - for (i = 0; i < fw_size; i++) - WREG32(mmRLC_GPM_UCODE_DATA, le32_to_cpup(fw_data++)); - WREG32(mmRLC_GPM_UCODE_ADDR, adev->gfx.rlc_fw_version); - - return 0; -} - static int gfx_v8_0_rlc_resume(struct amdgpu_device *adev) { - int r; - u32 tmp; - gfx_v8_0_rlc_stop(adev); - - /* disable CG */ - tmp = RREG32(mmRLC_CGCG_CGLS_CTRL); - tmp &= ~(RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK | - RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK); - WREG32(mmRLC_CGCG_CGLS_CTRL, tmp); - if (adev->asic_type == CHIP_POLARIS11 || - adev->asic_type == CHIP_POLARIS10 || - adev->asic_type == CHIP_POLARIS12 || - adev->asic_type == CHIP_VEGAM) { - tmp = RREG32(mmRLC_CGCG_CGLS_CTRL_3D); - tmp &= ~0x3; - WREG32(mmRLC_CGCG_CGLS_CTRL_3D, tmp); - } - - /* disable PG */ - WREG32(mmRLC_PG_CNTL, 0); - gfx_v8_0_rlc_reset(adev); gfx_v8_0_init_pg(adev); - - - if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) { - /* legacy rlc firmware loading */ - r = gfx_v8_0_rlc_load_microcode(adev); - if (r) - return r; - } - gfx_v8_0_rlc_start(adev); return 0; @@ -4265,63 +4203,6 @@ static void gfx_v8_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable) udelay(50); } -static int gfx_v8_0_cp_gfx_load_microcode(struct amdgpu_device *adev) -{ - const struct gfx_firmware_header_v1_0 *pfp_hdr; - const struct gfx_firmware_header_v1_0 *ce_hdr; - const struct gfx_firmware_header_v1_0 *me_hdr; - const __le32 *fw_data; - unsigned i, fw_size; - - if (!adev->gfx.me_fw || !adev->gfx.pfp_fw || !adev->gfx.ce_fw) - return -EINVAL; - - pfp_hdr = (const struct gfx_firmware_header_v1_0 *) - adev->gfx.pfp_fw->data; - ce_hdr = (const struct gfx_firmware_header_v1_0 *) - adev->gfx.ce_fw->data; - me_hdr = (const struct gfx_firmware_header_v1_0 *) - adev->gfx.me_fw->data; - - amdgpu_ucode_print_gfx_hdr(&pfp_hdr->header); - amdgpu_ucode_print_gfx_hdr(&ce_hdr->header); - amdgpu_ucode_print_gfx_hdr(&me_hdr->header); - - gfx_v8_0_cp_gfx_enable(adev, false); - - /* PFP */ - fw_data = (const __le32 *) - (adev->gfx.pfp_fw->data + - le32_to_cpu(pfp_hdr->header.ucode_array_offset_bytes)); - fw_size = le32_to_cpu(pfp_hdr->header.ucode_size_bytes) / 4; - WREG32(mmCP_PFP_UCODE_ADDR, 0); - for (i = 0; i < fw_size; i++) - WREG32(mmCP_PFP_UCODE_DATA, le32_to_cpup(fw_data++)); - WREG32(mmCP_PFP_UCODE_ADDR, adev->gfx.pfp_fw_version); - - /* CE */ - fw_data = (const __le32 *) - (adev->gfx.ce_fw->data + - le32_to_cpu(ce_hdr->header.ucode_array_offset_bytes)); - fw_size = le32_to_cpu(ce_hdr->header.ucode_size_bytes) / 4; - WREG32(mmCP_CE_UCODE_ADDR, 0); - for (i = 0; i < fw_size; i++) - WREG32(mmCP_CE_UCODE_DATA, le32_to_cpup(fw_data++)); - WREG32(mmCP_CE_UCODE_ADDR, adev->gfx.ce_fw_version); - - /* ME */ - fw_data = (const __le32 *) - (adev->gfx.me_fw->data + - le32_to_cpu(me_hdr->header.ucode_array_offset_bytes)); - fw_size = le32_to_cpu(me_hdr->header.ucode_size_bytes) / 4; - WREG32(mmCP_ME_RAM_WADDR, 0); - for (i = 0; i < fw_size; i++) - WREG32(mmCP_ME_RAM_DATA, le32_to_cpup(fw_data++)); - WREG32(mmCP_ME_RAM_WADDR, adev->gfx.me_fw_version); - - return 0; -} - static u32 gfx_v8_0_get_csb_size(struct amdgpu_device *adev) { u32 count = 0; @@ -4521,52 +4402,6 @@ static void gfx_v8_0_cp_compute_enable(struct amdgpu_device *adev, bool enable) udelay(50); } -static int gfx_v8_0_cp_compute_load_microcode(struct amdgpu_device *adev) -{ - const struct gfx_firmware_header_v1_0 *mec_hdr; - const __le32 *fw_data; - unsigned i, fw_size; - - if (!adev->gfx.mec_fw) - return -EINVAL; - - gfx_v8_0_cp_compute_enable(adev, false); - - 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)); - fw_size = le32_to_cpu(mec_hdr->header.ucode_size_bytes) / 4; - - /* MEC1 */ - WREG32(mmCP_MEC_ME1_UCODE_ADDR, 0); - for (i = 0; i < fw_size; i++) - WREG32(mmCP_MEC_ME1_UCODE_DATA, le32_to_cpup(fw_data+i)); - WREG32(mmCP_MEC_ME1_UCODE_ADDR, adev->gfx.mec_fw_version); - - /* Loading MEC2 firmware is only necessary if MEC2 should run different microcode than MEC1. */ - if (adev->gfx.mec2_fw) { - const struct gfx_firmware_header_v1_0 *mec2_hdr; - - mec2_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec2_fw->data; - amdgpu_ucode_print_gfx_hdr(&mec2_hdr->header); - - fw_data = (const __le32 *) - (adev->gfx.mec2_fw->data + - le32_to_cpu(mec2_hdr->header.ucode_array_offset_bytes)); - fw_size = le32_to_cpu(mec2_hdr->header.ucode_size_bytes) / 4; - - WREG32(mmCP_MEC_ME2_UCODE_ADDR, 0); - for (i = 0; i < fw_size; i++) - WREG32(mmCP_MEC_ME2_UCODE_DATA, le32_to_cpup(fw_data+i)); - WREG32(mmCP_MEC_ME2_UCODE_ADDR, adev->gfx.mec2_fw_version); - } - - return 0; -} - /* KIQ functions */ static void gfx_v8_0_kiq_setting(struct amdgpu_ring *ring) { @@ -4892,7 +4727,7 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring) struct vi_mqd *mqd = ring->mqd_ptr; int mqd_idx = ring - &adev->gfx.compute_ring[0]; - if (!adev->in_gpu_reset && !adev->gfx.in_suspend) { + if (!adev->in_gpu_reset && !adev->in_suspend) { memset((void *)mqd, 0, sizeof(struct vi_mqd_allocation)); ((struct vi_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF; ((struct vi_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF; @@ -5000,17 +4835,6 @@ static int gfx_v8_0_cp_resume(struct amdgpu_device *adev) if (!(adev->flags & AMD_IS_APU)) gfx_v8_0_enable_gui_idle_interrupt(adev, false); - if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) { - /* legacy firmware loading */ - r = gfx_v8_0_cp_gfx_load_microcode(adev); - if (r) - return r; - - r = gfx_v8_0_cp_compute_load_microcode(adev); - if (r) - return r; - } - r = gfx_v8_0_kiq_resume(adev); if (r) return r; @@ -5039,7 +4863,7 @@ static int gfx_v8_0_hw_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; gfx_v8_0_init_golden_registers(adev); - gfx_v8_0_gpu_init(adev); + gfx_v8_0_constants_init(adev); r = gfx_v8_0_rlc_resume(adev); if (r) @@ -5080,6 +4904,55 @@ static int gfx_v8_0_kcq_disable(struct amdgpu_device *adev) return r; } +static bool gfx_v8_0_is_idle(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + if (REG_GET_FIELD(RREG32(mmGRBM_STATUS), GRBM_STATUS, GUI_ACTIVE) + || RREG32(mmGRBM_STATUS2) != 0x8) + return false; + else + return true; +} + +static bool gfx_v8_0_rlc_is_idle(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + if (RREG32(mmGRBM_STATUS2) != 0x8) + return false; + else + return true; +} + +static int gfx_v8_0_wait_for_rlc_idle(void *handle) +{ + unsigned int i; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + for (i = 0; i < adev->usec_timeout; i++) { + if (gfx_v8_0_rlc_is_idle(handle)) + return 0; + + udelay(1); + } + return -ETIMEDOUT; +} + +static int gfx_v8_0_wait_for_idle(void *handle) +{ + unsigned int i; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + for (i = 0; i < adev->usec_timeout; i++) { + if (gfx_v8_0_is_idle(handle)) + return 0; + + udelay(1); + } + return -ETIMEDOUT; +} + static int gfx_v8_0_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; @@ -5098,51 +4971,27 @@ static int gfx_v8_0_hw_fini(void *handle) pr_debug("For SRIOV client, shouldn't do anything.\n"); return 0; } - gfx_v8_0_cp_enable(adev, false); - gfx_v8_0_rlc_stop(adev); - + adev->gfx.rlc.funcs->enter_safe_mode(adev); + if (!gfx_v8_0_wait_for_idle(adev)) + gfx_v8_0_cp_enable(adev, false); + else + pr_err("cp is busy, skip halt cp\n"); + if (!gfx_v8_0_wait_for_rlc_idle(adev)) + gfx_v8_0_rlc_stop(adev); + else + pr_err("rlc is busy, skip halt rlc\n"); + adev->gfx.rlc.funcs->exit_safe_mode(adev); return 0; } static int gfx_v8_0_suspend(void *handle) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - adev->gfx.in_suspend = true; - return gfx_v8_0_hw_fini(adev); + return gfx_v8_0_hw_fini(handle); } static int gfx_v8_0_resume(void *handle) { - int r; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - r = gfx_v8_0_hw_init(adev); - adev->gfx.in_suspend = false; - return r; -} - -static bool gfx_v8_0_is_idle(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - if (REG_GET_FIELD(RREG32(mmGRBM_STATUS), GRBM_STATUS, GUI_ACTIVE)) - return false; - else - return true; -} - -static int gfx_v8_0_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_v8_0_is_idle(handle)) - return 0; - - udelay(1); - } - return -ETIMEDOUT; + return gfx_v8_0_hw_init(handle); } static bool gfx_v8_0_check_soft_reset(void *handle) @@ -7013,52 +6862,6 @@ static int gfx_v8_0_sq_irq(struct amdgpu_device *adev, return 0; } -static int gfx_v8_0_kiq_set_interrupt_state(struct amdgpu_device *adev, - struct amdgpu_irq_src *src, - unsigned int type, - enum amdgpu_interrupt_state state) -{ - struct amdgpu_ring *ring = &(adev->gfx.kiq.ring); - - switch (type) { - case AMDGPU_CP_KIQ_IRQ_DRIVER0: - WREG32_FIELD(CPC_INT_CNTL, GENERIC2_INT_ENABLE, - state == AMDGPU_IRQ_STATE_DISABLE ? 0 : 1); - if (ring->me == 1) - WREG32_FIELD_OFFSET(CP_ME1_PIPE0_INT_CNTL, - ring->pipe, - GENERIC2_INT_ENABLE, - state == AMDGPU_IRQ_STATE_DISABLE ? 0 : 1); - else - WREG32_FIELD_OFFSET(CP_ME2_PIPE0_INT_CNTL, - ring->pipe, - GENERIC2_INT_ENABLE, - state == AMDGPU_IRQ_STATE_DISABLE ? 0 : 1); - break; - default: - BUG(); /* kiq only support GENERIC2_INT now */ - break; - } - return 0; -} - -static int gfx_v8_0_kiq_irq(struct amdgpu_device *adev, - struct amdgpu_irq_src *source, - struct amdgpu_iv_entry *entry) -{ - u8 me_id, pipe_id, queue_id; - struct amdgpu_ring *ring = &(adev->gfx.kiq.ring); - - me_id = (entry->ring_id & 0x0c) >> 2; - pipe_id = (entry->ring_id & 0x03) >> 0; - queue_id = (entry->ring_id & 0x70) >> 4; - DRM_DEBUG("IH: CPC GENERIC2_INT, me:%d, pipe:%d, queue:%d\n", - me_id, pipe_id, queue_id); - - amdgpu_fence_process(ring); - return 0; -} - static const struct amd_ip_funcs gfx_v8_0_ip_funcs = { .name = "gfx_v8_0", .early_init = gfx_v8_0_early_init, @@ -7209,11 +7012,6 @@ static const struct amdgpu_irq_src_funcs gfx_v8_0_priv_inst_irq_funcs = { .process = gfx_v8_0_priv_inst_irq, }; -static const struct amdgpu_irq_src_funcs gfx_v8_0_kiq_irq_funcs = { - .set = gfx_v8_0_kiq_set_interrupt_state, - .process = gfx_v8_0_kiq_irq, -}; - static const struct amdgpu_irq_src_funcs gfx_v8_0_cp_ecc_error_irq_funcs = { .set = gfx_v8_0_set_cp_ecc_int_state, .process = gfx_v8_0_cp_ecc_error_irq, @@ -7235,9 +7033,6 @@ static void gfx_v8_0_set_irq_funcs(struct amdgpu_device *adev) adev->gfx.priv_inst_irq.num_types = 1; adev->gfx.priv_inst_irq.funcs = &gfx_v8_0_priv_inst_irq_funcs; - adev->gfx.kiq.irq.num_types = AMDGPU_CP_KIQ_IRQ_LAST; - adev->gfx.kiq.irq.funcs = &gfx_v8_0_kiq_irq_funcs; - adev->gfx.cp_ecc_error_irq.num_types = 1; adev->gfx.cp_ecc_error_irq.funcs = &gfx_v8_0_cp_ecc_error_irq_funcs; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index f369d9603435..6d7baf59d6e1 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -97,6 +97,7 @@ MODULE_FIRMWARE("amdgpu/raven2_rlc.bin"); static const struct soc15_reg_golden golden_settings_gc_9_0[] = { SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG2, 0xf00fffff, 0x00000400), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG3, 0x80000000, 0x80000000), SOC15_REG_GOLDEN_VALUE(GC, 0, mmGB_GPU_ID, 0x0000000f, 0x00000000), SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_BINNER_EVENT_CNTL_3, 0x00000003, 0x82400024), SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_ENHANCE, 0x3fffffff, 0x00000001), @@ -133,7 +134,10 @@ static const struct soc15_reg_golden golden_settings_gc_9_0_vg10[] = SOC15_REG_GOLDEN_VALUE(GC, 0, mmRMI_UTCL1_CNTL2, 0x00030000, 0x00020000), SOC15_REG_GOLDEN_VALUE(GC, 0, mmSPI_CONFIG_CNTL_1, 0x0000000f, 0x01000107), SOC15_REG_GOLDEN_VALUE(GC, 0, mmTD_CNTL, 0x00001800, 0x00000800), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080) + SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC1_F32_INT_DIS, 0x00000000, 0x00000800), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC2_F32_INT_DIS, 0x00000000, 0x00000800), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000) }; static const struct soc15_reg_golden golden_settings_gc_9_0_vg20[] = @@ -173,7 +177,10 @@ static const struct soc15_reg_golden golden_settings_gc_9_1[] = SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_LO, 0xffffffff, 0x00003120), SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_CACHE_INVALIDATION, 0x3fff3af3, 0x19200000), SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_GS_MAX_WAVE_ID, 0x00000fff, 0x000000ff), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080) + SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC1_F32_INT_DIS, 0x00000000, 0x00000800), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC2_F32_INT_DIS, 0x00000000, 0x00000800), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000) }; static const struct soc15_reg_golden golden_settings_gc_9_1_rv1[] = @@ -247,7 +254,10 @@ static const struct soc15_reg_golden golden_settings_gc_9_2_1_vg12[] = SOC15_REG_GOLDEN_VALUE(GC, 0, mmSPI_CONFIG_CNTL_1, 0xffff03ff, 0x01000107), SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_HI, 0xffffffff, 0x00000000), SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_LO, 0xffffffff, 0x76325410), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmTD_CNTL, 0x01bd9f33, 0x01000000) + SOC15_REG_GOLDEN_VALUE(GC, 0, mmTD_CNTL, 0x01bd9f33, 0x01000000), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC1_F32_INT_DIS, 0x00000000, 0x00000800), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC2_F32_INT_DIS, 0x00000000, 0x00000800), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000) }; static const u32 GFX_RLC_SRM_INDEX_CNTL_ADDR_OFFSETS[] = @@ -908,6 +918,50 @@ static void gfx_v9_0_get_csb_buffer(struct amdgpu_device *adev, buffer[count++] = cpu_to_le32(0); } +static void gfx_v9_0_init_always_on_cu_mask(struct amdgpu_device *adev) +{ + struct amdgpu_cu_info *cu_info = &adev->gfx.cu_info; + uint32_t pg_always_on_cu_num = 2; + uint32_t always_on_cu_num; + uint32_t i, j, k; + uint32_t mask, cu_bitmap, counter; + + if (adev->flags & AMD_IS_APU) + always_on_cu_num = 4; + else if (adev->asic_type == CHIP_VEGA12) + always_on_cu_num = 8; + else + always_on_cu_num = 12; + + 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; + cu_bitmap = 0; + counter = 0; + gfx_v9_0_select_se_sh(adev, i, j, 0xffffffff); + + for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) { + if (cu_info->bitmap[i][j] & mask) { + if (counter == pg_always_on_cu_num) + WREG32_SOC15(GC, 0, mmRLC_PG_ALWAYS_ON_CU_MASK, cu_bitmap); + if (counter < always_on_cu_num) + cu_bitmap |= mask; + else + break; + counter++; + } + mask <<= 1; + } + + WREG32_SOC15(GC, 0, mmRLC_LB_ALWAYS_ACTIVE_CU_MASK, cu_bitmap); + cu_info->ao_cu_bitmap[i][j] = cu_bitmap; + } + } + gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + mutex_unlock(&adev->grbm_idx_mutex); +} + static void gfx_v9_0_init_lbpw(struct amdgpu_device *adev) { uint32_t data; @@ -941,8 +995,59 @@ static void gfx_v9_0_init_lbpw(struct amdgpu_device *adev) data |= 0x00C00000; WREG32_SOC15(GC, 0, mmRLC_GPM_GENERAL_7, data); - /* set RLC_LB_ALWAYS_ACTIVE_CU_MASK = 0xFFF */ - WREG32_SOC15(GC, 0, mmRLC_LB_ALWAYS_ACTIVE_CU_MASK, 0xFFF); + /* + * RLC_LB_ALWAYS_ACTIVE_CU_MASK = 0xF (4 CUs AON for Raven), + * programmed in gfx_v9_0_init_always_on_cu_mask() + */ + + /* set RLC_LB_CNTL = 0x8000_0095, 31 bit is reserved, + * but used for RLC_LB_CNTL configuration */ + data = RLC_LB_CNTL__LB_CNT_SPIM_ACTIVE_MASK; + data |= REG_SET_FIELD(data, RLC_LB_CNTL, CU_MASK_USED_OFF_HYST, 0x09); + data |= REG_SET_FIELD(data, RLC_LB_CNTL, RESERVED, 0x80000); + WREG32_SOC15(GC, 0, mmRLC_LB_CNTL, data); + mutex_unlock(&adev->grbm_idx_mutex); + + gfx_v9_0_init_always_on_cu_mask(adev); +} + +static void gfx_v9_4_init_lbpw(struct amdgpu_device *adev) +{ + uint32_t data; + + /* set mmRLC_LB_THR_CONFIG_1/2/3/4 */ + WREG32_SOC15(GC, 0, mmRLC_LB_THR_CONFIG_1, 0x0000007F); + WREG32_SOC15(GC, 0, mmRLC_LB_THR_CONFIG_2, 0x033388F8); + WREG32_SOC15(GC, 0, mmRLC_LB_THR_CONFIG_3, 0x00000077); + WREG32_SOC15(GC, 0, mmRLC_LB_THR_CONFIG_4, (0x10 | 0x27 << 8 | 0x02FA << 16)); + + /* set mmRLC_LB_CNTR_INIT = 0x0000_0000 */ + WREG32_SOC15(GC, 0, mmRLC_LB_CNTR_INIT, 0x00000000); + + /* set mmRLC_LB_CNTR_MAX = 0x0000_0500 */ + WREG32_SOC15(GC, 0, mmRLC_LB_CNTR_MAX, 0x00000800); + + mutex_lock(&adev->grbm_idx_mutex); + /* set mmRLC_LB_INIT_CU_MASK thru broadcast mode to enable all SE/SH*/ + gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + WREG32_SOC15(GC, 0, mmRLC_LB_INIT_CU_MASK, 0xffffffff); + + /* set mmRLC_LB_PARAMS = 0x003F_1006 */ + data = REG_SET_FIELD(0, RLC_LB_PARAMS, FIFO_SAMPLES, 0x0003); + data |= REG_SET_FIELD(data, RLC_LB_PARAMS, PG_IDLE_SAMPLES, 0x0010); + data |= REG_SET_FIELD(data, RLC_LB_PARAMS, PG_IDLE_SAMPLE_INTERVAL, 0x033F); + WREG32_SOC15(GC, 0, mmRLC_LB_PARAMS, data); + + /* set mmRLC_GPM_GENERAL_7[31-16] = 0x00C0 */ + data = RREG32_SOC15(GC, 0, mmRLC_GPM_GENERAL_7); + data &= 0x0000FFFF; + data |= 0x00C00000; + WREG32_SOC15(GC, 0, mmRLC_GPM_GENERAL_7, data); + + /* + * RLC_LB_ALWAYS_ACTIVE_CU_MASK = 0xFFF (12 CUs AON), + * programmed in gfx_v9_0_init_always_on_cu_mask() + */ /* set RLC_LB_CNTL = 0x8000_0095, 31 bit is reserved, * but used for RLC_LB_CNTL configuration */ @@ -951,6 +1056,8 @@ static void gfx_v9_0_init_lbpw(struct amdgpu_device *adev) data |= REG_SET_FIELD(data, RLC_LB_CNTL, RESERVED, 0x80000); WREG32_SOC15(GC, 0, mmRLC_LB_CNTL, data); mutex_unlock(&adev->grbm_idx_mutex); + + gfx_v9_0_init_always_on_cu_mask(adev); } static void gfx_v9_0_enable_lbpw(struct amdgpu_device *adev, bool enable) @@ -1084,8 +1191,17 @@ static int gfx_v9_0_rlc_init(struct amdgpu_device *adev) rv_init_cp_jump_table(adev); amdgpu_bo_kunmap(adev->gfx.rlc.cp_table_obj); amdgpu_bo_unreserve(adev->gfx.rlc.cp_table_obj); + } + switch (adev->asic_type) { + case CHIP_RAVEN: gfx_v9_0_init_lbpw(adev); + break; + case CHIP_VEGA20: + gfx_v9_4_init_lbpw(adev); + break; + default: + break; } return 0; @@ -1605,11 +1721,6 @@ static int gfx_v9_0_sw_init(void *handle) adev->gfx.mec.num_pipe_per_mec = 4; adev->gfx.mec.num_queue_per_pipe = 8; - /* KIQ event */ - r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP, GFX_9_0__SRCID__CP_IB2_INTERRUPT_PKT, &adev->gfx.kiq.irq); - if (r) - return r; - /* 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) @@ -1847,7 +1958,7 @@ static void gfx_v9_0_init_compute_vmid(struct amdgpu_device *adev) mutex_unlock(&adev->srbm_mutex); } -static void gfx_v9_0_gpu_init(struct amdgpu_device *adev) +static void gfx_v9_0_constants_init(struct amdgpu_device *adev) { u32 tmp; int i; @@ -2403,7 +2514,8 @@ static int gfx_v9_0_rlc_resume(struct amdgpu_device *adev) return r; } - if (adev->asic_type == CHIP_RAVEN) { + if (adev->asic_type == CHIP_RAVEN || + adev->asic_type == CHIP_VEGA20) { if (amdgpu_lbpw != 0) gfx_v9_0_enable_lbpw(adev, true); else @@ -3091,7 +3203,7 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring) struct v9_mqd *mqd = ring->mqd_ptr; int mqd_idx = ring - &adev->gfx.compute_ring[0]; - if (!adev->in_gpu_reset && !adev->gfx.in_suspend) { + if (!adev->in_gpu_reset && !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; @@ -3235,7 +3347,7 @@ static int gfx_v9_0_hw_init(void *handle) gfx_v9_0_init_golden_registers(adev); - gfx_v9_0_gpu_init(adev); + gfx_v9_0_constants_init(adev); r = gfx_v9_0_csb_vram_pin(adev); if (r) @@ -3310,7 +3422,7 @@ static int gfx_v9_0_hw_fini(void *handle) /* Use deinitialize sequence from CAIL when unbinding device from driver, * otherwise KIQ is hanging when binding back */ - if (!adev->in_gpu_reset && !adev->gfx.in_suspend) { + if (!adev->in_gpu_reset && !adev->in_suspend) { mutex_lock(&adev->srbm_mutex); soc15_grbm_select(adev, adev->gfx.kiq.ring.me, adev->gfx.kiq.ring.pipe, @@ -3330,20 +3442,12 @@ static int gfx_v9_0_hw_fini(void *handle) static int gfx_v9_0_suspend(void *handle) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - adev->gfx.in_suspend = true; - return gfx_v9_0_hw_fini(adev); + return gfx_v9_0_hw_fini(handle); } static int gfx_v9_0_resume(void *handle) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - int r; - - r = gfx_v9_0_hw_init(adev); - adev->gfx.in_suspend = false; - return r; + return gfx_v9_0_hw_init(handle); } static bool gfx_v9_0_is_idle(void *handle) @@ -4609,68 +4713,6 @@ static int gfx_v9_0_priv_inst_irq(struct amdgpu_device *adev, return 0; } -static int gfx_v9_0_kiq_set_interrupt_state(struct amdgpu_device *adev, - struct amdgpu_irq_src *src, - unsigned int type, - enum amdgpu_interrupt_state state) -{ - uint32_t tmp, target; - struct amdgpu_ring *ring = &(adev->gfx.kiq.ring); - - if (ring->me == 1) - target = SOC15_REG_OFFSET(GC, 0, mmCP_ME1_PIPE0_INT_CNTL); - else - target = SOC15_REG_OFFSET(GC, 0, mmCP_ME2_PIPE0_INT_CNTL); - target += ring->pipe; - - switch (type) { - case AMDGPU_CP_KIQ_IRQ_DRIVER0: - if (state == AMDGPU_IRQ_STATE_DISABLE) { - tmp = RREG32_SOC15(GC, 0, mmCPC_INT_CNTL); - tmp = REG_SET_FIELD(tmp, CPC_INT_CNTL, - GENERIC2_INT_ENABLE, 0); - WREG32_SOC15(GC, 0, mmCPC_INT_CNTL, tmp); - - tmp = RREG32(target); - tmp = REG_SET_FIELD(tmp, CP_ME2_PIPE0_INT_CNTL, - GENERIC2_INT_ENABLE, 0); - WREG32(target, tmp); - } else { - tmp = RREG32_SOC15(GC, 0, mmCPC_INT_CNTL); - tmp = REG_SET_FIELD(tmp, CPC_INT_CNTL, - GENERIC2_INT_ENABLE, 1); - WREG32_SOC15(GC, 0, mmCPC_INT_CNTL, tmp); - - tmp = RREG32(target); - tmp = REG_SET_FIELD(tmp, CP_ME2_PIPE0_INT_CNTL, - GENERIC2_INT_ENABLE, 1); - WREG32(target, tmp); - } - break; - default: - BUG(); /* kiq only support GENERIC2_INT now */ - break; - } - return 0; -} - -static int gfx_v9_0_kiq_irq(struct amdgpu_device *adev, - struct amdgpu_irq_src *source, - struct amdgpu_iv_entry *entry) -{ - u8 me_id, pipe_id, queue_id; - struct amdgpu_ring *ring = &(adev->gfx.kiq.ring); - - me_id = (entry->ring_id & 0x0c) >> 2; - pipe_id = (entry->ring_id & 0x03) >> 0; - queue_id = (entry->ring_id & 0x70) >> 4; - DRM_DEBUG("IH: CPC GENERIC2_INT, me:%d, pipe:%d, queue:%d\n", - me_id, pipe_id, queue_id); - - amdgpu_fence_process(ring); - return 0; -} - static const struct amd_ip_funcs gfx_v9_0_ip_funcs = { .name = "gfx_v9_0", .early_init = gfx_v9_0_early_init, @@ -4819,11 +4861,6 @@ static void gfx_v9_0_set_ring_funcs(struct amdgpu_device *adev) adev->gfx.compute_ring[i].funcs = &gfx_v9_0_ring_funcs_compute; } -static const struct amdgpu_irq_src_funcs gfx_v9_0_kiq_irq_funcs = { - .set = gfx_v9_0_kiq_set_interrupt_state, - .process = gfx_v9_0_kiq_irq, -}; - static const struct amdgpu_irq_src_funcs gfx_v9_0_eop_irq_funcs = { .set = gfx_v9_0_set_eop_interrupt_state, .process = gfx_v9_0_eop_irq, @@ -4849,9 +4886,6 @@ static void gfx_v9_0_set_irq_funcs(struct amdgpu_device *adev) adev->gfx.priv_inst_irq.num_types = 1; adev->gfx.priv_inst_irq.funcs = &gfx_v9_0_priv_inst_irq_funcs; - - adev->gfx.kiq.irq.num_types = AMDGPU_CP_KIQ_IRQ_LAST; - adev->gfx.kiq.irq.funcs = &gfx_v9_0_kiq_irq_funcs; } static void gfx_v9_0_set_rlc_funcs(struct amdgpu_device *adev) @@ -4871,7 +4905,20 @@ static void gfx_v9_0_set_rlc_funcs(struct amdgpu_device *adev) static void gfx_v9_0_set_gds_init(struct amdgpu_device *adev) { /* init asci gds info */ - adev->gds.mem.total_size = RREG32_SOC15(GC, 0, mmGDS_VMID0_SIZE); + switch (adev->asic_type) { + case CHIP_VEGA10: + case CHIP_VEGA12: + case CHIP_VEGA20: + adev->gds.mem.total_size = 0x10000; + break; + case CHIP_RAVEN: + adev->gds.mem.total_size = 0x1000; + break; + default: + adev->gds.mem.total_size = 0x10000; + break; + } + adev->gds.gws.total_size = 64; adev->gds.oa.total_size = 16; diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c index 65f58ebcf835..ceb7847b504f 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c @@ -82,7 +82,8 @@ static void gfxhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev) * to get rid of the VM fault and hardware hang. */ WREG32_SOC15(GC, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR, - (max(adev->gmc.vram_end, adev->gmc.agp_end) >> 18) + 0x1); + max((adev->gmc.vram_end >> 18) + 0x1, + adev->gmc.agp_end >> 18)); else WREG32_SOC15(GC, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR, max(adev->gmc.vram_end, adev->gmc.agp_end) >> 18); diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c index 4411463ca719..e1c2b4e9c7b2 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c @@ -859,11 +859,11 @@ static int gmc_v6_0_sw_init(void *handle) adev->gmc.vram_type = gmc_v6_0_convert_vram_type(tmp); } - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 146, &adev->gmc.vm_fault); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 146, &adev->gmc.vm_fault); if (r) return r; - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 147, &adev->gmc.vm_fault); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 147, &adev->gmc.vm_fault); if (r) return r; @@ -1180,8 +1180,7 @@ static const struct amdgpu_irq_src_funcs gmc_v6_0_irq_funcs = { static void gmc_v6_0_set_gmc_funcs(struct amdgpu_device *adev) { - if (adev->gmc.gmc_funcs == NULL) - adev->gmc.gmc_funcs = &gmc_v6_0_gmc_funcs; + adev->gmc.gmc_funcs = &gmc_v6_0_gmc_funcs; } static void gmc_v6_0_set_irq_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c index ae776ce9a415..910c4ce19cb3 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c @@ -991,11 +991,11 @@ static int gmc_v7_0_sw_init(void *handle) adev->gmc.vram_type = gmc_v7_0_convert_vram_type(tmp); } - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_PAGE_INV_FAULT, &adev->gmc.vm_fault); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_PAGE_INV_FAULT, &adev->gmc.vm_fault); if (r) return r; - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_MEM_PROT_FAULT, &adev->gmc.vm_fault); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_MEM_PROT_FAULT, &adev->gmc.vm_fault); if (r) return r; @@ -1388,8 +1388,7 @@ static const struct amdgpu_irq_src_funcs gmc_v7_0_irq_funcs = { static void gmc_v7_0_set_gmc_funcs(struct amdgpu_device *adev) { - if (adev->gmc.gmc_funcs == NULL) - adev->gmc.gmc_funcs = &gmc_v7_0_gmc_funcs; + adev->gmc.gmc_funcs = &gmc_v7_0_gmc_funcs; } static void gmc_v7_0_set_irq_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index 53ae49b8bde8..1d3265c97b70 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -1095,11 +1095,11 @@ static int gmc_v8_0_sw_init(void *handle) adev->gmc.vram_type = gmc_v8_0_convert_vram_type(tmp); } - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_PAGE_INV_FAULT, &adev->gmc.vm_fault); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_PAGE_INV_FAULT, &adev->gmc.vm_fault); if (r) return r; - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_MEM_PROT_FAULT, &adev->gmc.vm_fault); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_MEM_PROT_FAULT, &adev->gmc.vm_fault); if (r) return r; @@ -1733,8 +1733,7 @@ static const struct amdgpu_irq_src_funcs gmc_v8_0_irq_funcs = { static void gmc_v8_0_set_gmc_funcs(struct amdgpu_device *adev) { - if (adev->gmc.gmc_funcs == NULL) - adev->gmc.gmc_funcs = &gmc_v8_0_gmc_funcs; + adev->gmc.gmc_funcs = &gmc_v8_0_gmc_funcs; } static void gmc_v8_0_set_irq_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c index aad3c7c5fb3a..f35d7a554ad5 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c @@ -593,8 +593,7 @@ static const struct amdgpu_gmc_funcs gmc_v9_0_gmc_funcs = { static void gmc_v9_0_set_gmc_funcs(struct amdgpu_device *adev) { - if (adev->gmc.gmc_funcs == NULL) - adev->gmc.gmc_funcs = &gmc_v9_0_gmc_funcs; + adev->gmc.gmc_funcs = &gmc_v9_0_gmc_funcs; } static int gmc_v9_0_early_init(void *handle) diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c index 842c4b677b4d..cf0fc61aebe6 100644 --- a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c @@ -255,7 +255,7 @@ static void iceland_ih_decode_iv(struct amdgpu_device *adev, dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]); dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]); - entry->client_id = AMDGPU_IH_CLIENTID_LEGACY; + entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY; entry->src_id = dw[0] & 0xff; entry->src_data[0] = dw[1] & 0xfffffff; entry->ring_id = dw[2] & 0xff; @@ -297,7 +297,7 @@ static int iceland_ih_sw_init(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = amdgpu_ih_ring_init(adev, 64 * 1024, false); + r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, false); if (r) return r; @@ -311,7 +311,7 @@ static int iceland_ih_sw_fini(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; amdgpu_irq_fini(adev); - amdgpu_ih_ring_fini(adev); + amdgpu_ih_ring_fini(adev, &adev->irq.ih); amdgpu_irq_remove_domain(adev); return 0; @@ -447,8 +447,7 @@ static const struct amdgpu_ih_funcs iceland_ih_funcs = { static void iceland_ih_set_interrupt_funcs(struct amdgpu_device *adev) { - if (adev->irq.ih_funcs == NULL) - adev->irq.ih_funcs = &iceland_ih_funcs; + adev->irq.ih_funcs = &iceland_ih_funcs; } const struct amdgpu_ip_block_version iceland_ih_ip_block = diff --git a/drivers/gpu/drm/amd/amdgpu/kv_dpm.c b/drivers/gpu/drm/amd/amdgpu/kv_dpm.c index cb79a93c2eb7..d0e478f43443 100644 --- a/drivers/gpu/drm/amd/amdgpu/kv_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/kv_dpm.c @@ -2995,12 +2995,12 @@ static int kv_dpm_sw_init(void *handle) int ret; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - ret = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 230, + ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 230, &adev->pm.dpm.thermal.irq); if (ret) return ret; - ret = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 231, + ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 231, &adev->pm.dpm.thermal.irq); if (ret) return ret; diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c index 80698b5ffa4a..14649f8475f3 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c @@ -100,7 +100,8 @@ static void mmhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev) * to get rid of the VM fault and hardware hang. */ WREG32_SOC15(MMHUB, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR, - (max(adev->gmc.vram_end, adev->gmc.agp_end) >> 18) + 0x1); + max((adev->gmc.vram_end >> 18) + 0x1, + adev->gmc.agp_end >> 18)); else WREG32_SOC15(MMHUB, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR, max(adev->gmc.vram_end, adev->gmc.agp_end) >> 18); diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c index 842567b53df5..64e875d528dd 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c @@ -580,11 +580,11 @@ int xgpu_vi_mailbox_add_irq_id(struct amdgpu_device *adev) { int r; - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 135, &adev->virt.rcv_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 135, &adev->virt.rcv_irq); if (r) return r; - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 138, &adev->virt.ack_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 138, &adev->virt.ack_irq); if (r) { amdgpu_irq_put(adev, &adev->virt.rcv_irq, 0); return r; diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c index 9217af00bc8d..3f3fac2d50cd 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c @@ -306,11 +306,8 @@ static int psp_v11_0_ring_stop(struct psp_context *psp, enum psp_ring_type ring_type) { int ret = 0; - struct psp_ring *ring; struct amdgpu_device *adev = psp->adev; - ring = &psp->km_ring; - /* Write the ring destroy command to C2PMSG_64 */ WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, GFX_CTRL_CMD_ID_DESTROY_RINGS); diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c index c403bdf8ad70..2d4770e173dd 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c @@ -504,41 +504,6 @@ static int sdma_v2_4_rlc_resume(struct amdgpu_device *adev) return 0; } -/** - * sdma_v2_4_load_microcode - load the sDMA ME ucode - * - * @adev: amdgpu_device pointer - * - * Loads the sDMA0/1 ucode. - * Returns 0 for success, -EINVAL if the ucode is not available. - */ -static int sdma_v2_4_load_microcode(struct amdgpu_device *adev) -{ - const struct sdma_firmware_header_v1_0 *hdr; - const __le32 *fw_data; - u32 fw_size; - int i, j; - - /* halt the MEs */ - sdma_v2_4_enable(adev, false); - - for (i = 0; i < adev->sdma.num_instances; i++) { - if (!adev->sdma.instance[i].fw) - return -EINVAL; - hdr = (const struct sdma_firmware_header_v1_0 *)adev->sdma.instance[i].fw->data; - amdgpu_ucode_print_sdma_hdr(&hdr->header); - fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4; - fw_data = (const __le32 *) - (adev->sdma.instance[i].fw->data + - le32_to_cpu(hdr->header.ucode_array_offset_bytes)); - WREG32(mmSDMA0_UCODE_ADDR + sdma_offsets[i], 0); - for (j = 0; j < fw_size; j++) - WREG32(mmSDMA0_UCODE_DATA + sdma_offsets[i], le32_to_cpup(fw_data++)); - WREG32(mmSDMA0_UCODE_ADDR + sdma_offsets[i], adev->sdma.instance[i].fw_version); - } - - return 0; -} /** * sdma_v2_4_start - setup and start the async dma engines @@ -552,13 +517,6 @@ static int sdma_v2_4_start(struct amdgpu_device *adev) { int r; - - if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) { - r = sdma_v2_4_load_microcode(adev); - if (r) - return r; - } - /* halt the engine before programing */ sdma_v2_4_enable(adev, false); @@ -898,19 +856,19 @@ static int sdma_v2_4_sw_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; /* SDMA trap event */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_TRAP, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_TRAP, &adev->sdma.trap_irq); if (r) return r; /* SDMA Privileged inst */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 241, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 241, &adev->sdma.illegal_inst_irq); if (r) return r; /* SDMA Privileged inst */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_SRBM_WRITE, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_SRBM_WRITE, &adev->sdma.illegal_inst_irq); if (r) return r; @@ -1296,10 +1254,8 @@ static const struct amdgpu_buffer_funcs sdma_v2_4_buffer_funcs = { static void sdma_v2_4_set_buffer_funcs(struct amdgpu_device *adev) { - if (adev->mman.buffer_funcs == NULL) { - adev->mman.buffer_funcs = &sdma_v2_4_buffer_funcs; - adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring; - } + adev->mman.buffer_funcs = &sdma_v2_4_buffer_funcs; + adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring; } static const struct amdgpu_vm_pte_funcs sdma_v2_4_vm_pte_funcs = { @@ -1315,15 +1271,13 @@ static void sdma_v2_4_set_vm_pte_funcs(struct amdgpu_device *adev) struct drm_gpu_scheduler *sched; unsigned i; - if (adev->vm_manager.vm_pte_funcs == NULL) { - adev->vm_manager.vm_pte_funcs = &sdma_v2_4_vm_pte_funcs; - for (i = 0; i < adev->sdma.num_instances; i++) { - sched = &adev->sdma.instance[i].ring.sched; - adev->vm_manager.vm_pte_rqs[i] = - &sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL]; - } - adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances; + adev->vm_manager.vm_pte_funcs = &sdma_v2_4_vm_pte_funcs; + for (i = 0; i < adev->sdma.num_instances; i++) { + sched = &adev->sdma.instance[i].ring.sched; + adev->vm_manager.vm_pte_rqs[i] = + &sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL]; } + adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances; } const struct amdgpu_ip_block_version sdma_v2_4_ip_block = diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c index 2677d6a1bf42..6fb3edaba0ec 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c @@ -318,14 +318,13 @@ static int sdma_v3_0_init_microcode(struct amdgpu_device *adev) if (adev->sdma.instance[i].feature_version >= 20) adev->sdma.instance[i].burst_nop = true; - if (adev->firmware.load_type == AMDGPU_FW_LOAD_SMU) { - info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA0 + i]; - info->ucode_id = AMDGPU_UCODE_ID_SDMA0 + i; - info->fw = adev->sdma.instance[i].fw; - header = (const struct common_firmware_header *)info->fw->data; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); - } + info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA0 + i]; + info->ucode_id = AMDGPU_UCODE_ID_SDMA0 + i; + info->fw = adev->sdma.instance[i].fw; + header = (const struct common_firmware_header *)info->fw->data; + adev->firmware.fw_size += + ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE); + } out: if (err) { @@ -778,42 +777,6 @@ static int sdma_v3_0_rlc_resume(struct amdgpu_device *adev) } /** - * sdma_v3_0_load_microcode - load the sDMA ME ucode - * - * @adev: amdgpu_device pointer - * - * Loads the sDMA0/1 ucode. - * Returns 0 for success, -EINVAL if the ucode is not available. - */ -static int sdma_v3_0_load_microcode(struct amdgpu_device *adev) -{ - const struct sdma_firmware_header_v1_0 *hdr; - const __le32 *fw_data; - u32 fw_size; - int i, j; - - /* halt the MEs */ - sdma_v3_0_enable(adev, false); - - for (i = 0; i < adev->sdma.num_instances; i++) { - if (!adev->sdma.instance[i].fw) - return -EINVAL; - hdr = (const struct sdma_firmware_header_v1_0 *)adev->sdma.instance[i].fw->data; - amdgpu_ucode_print_sdma_hdr(&hdr->header); - fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4; - fw_data = (const __le32 *) - (adev->sdma.instance[i].fw->data + - le32_to_cpu(hdr->header.ucode_array_offset_bytes)); - WREG32(mmSDMA0_UCODE_ADDR + sdma_offsets[i], 0); - for (j = 0; j < fw_size; j++) - WREG32(mmSDMA0_UCODE_DATA + sdma_offsets[i], le32_to_cpup(fw_data++)); - WREG32(mmSDMA0_UCODE_ADDR + sdma_offsets[i], adev->sdma.instance[i].fw_version); - } - - return 0; -} - -/** * sdma_v3_0_start - setup and start the async dma engines * * @adev: amdgpu_device pointer @@ -825,12 +788,6 @@ static int sdma_v3_0_start(struct amdgpu_device *adev) { int r; - if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) { - r = sdma_v3_0_load_microcode(adev); - if (r) - return r; - } - /* disable sdma engine before programing it */ sdma_v3_0_ctx_switch_enable(adev, false); sdma_v3_0_enable(adev, false); @@ -1177,19 +1134,19 @@ static int sdma_v3_0_sw_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; /* SDMA trap event */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_TRAP, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_TRAP, &adev->sdma.trap_irq); if (r) return r; /* SDMA Privileged inst */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 241, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 241, &adev->sdma.illegal_inst_irq); if (r) return r; /* SDMA Privileged inst */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_SRBM_WRITE, + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_SRBM_WRITE, &adev->sdma.illegal_inst_irq); if (r) return r; @@ -1736,10 +1693,8 @@ static const struct amdgpu_buffer_funcs sdma_v3_0_buffer_funcs = { static void sdma_v3_0_set_buffer_funcs(struct amdgpu_device *adev) { - if (adev->mman.buffer_funcs == NULL) { - adev->mman.buffer_funcs = &sdma_v3_0_buffer_funcs; - adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring; - } + adev->mman.buffer_funcs = &sdma_v3_0_buffer_funcs; + adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring; } static const struct amdgpu_vm_pte_funcs sdma_v3_0_vm_pte_funcs = { @@ -1755,15 +1710,13 @@ static void sdma_v3_0_set_vm_pte_funcs(struct amdgpu_device *adev) struct drm_gpu_scheduler *sched; unsigned i; - if (adev->vm_manager.vm_pte_funcs == NULL) { - adev->vm_manager.vm_pte_funcs = &sdma_v3_0_vm_pte_funcs; - for (i = 0; i < adev->sdma.num_instances; i++) { - sched = &adev->sdma.instance[i].ring.sched; - adev->vm_manager.vm_pte_rqs[i] = - &sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL]; - } - adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances; + adev->vm_manager.vm_pte_funcs = &sdma_v3_0_vm_pte_funcs; + for (i = 0; i < adev->sdma.num_instances; i++) { + sched = &adev->sdma.instance[i].ring.sched; + adev->vm_manager.vm_pte_rqs[i] = + &sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL]; } + adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances; } const struct amdgpu_ip_block_version sdma_v3_0_ip_block = diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c index 2ea1f0d8f5be..04fa3d972636 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c @@ -148,6 +148,7 @@ static const struct soc15_reg_golden golden_settings_sdma0_4_2[] = SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_RLC7_RB_RPTR_ADDR_LO, 0xfffffffd, 0x00000001), SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_RLC7_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000), SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_UTCL1_PAGE, 0x000003ff, 0x000003c0), + SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_UTCL1_WATERMK, 0xFE000000, 0x00000000), }; static const struct soc15_reg_golden golden_settings_sdma1_4_2[] = { @@ -177,6 +178,7 @@ static const struct soc15_reg_golden golden_settings_sdma1_4_2[] = { SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_RLC7_RB_RPTR_ADDR_LO, 0xfffffffd, 0x00000001), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_RLC7_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_UTCL1_PAGE, 0x000003ff, 0x000003c0), + SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_UTCL1_WATERMK, 0xFE000000, 0x00000000), }; static const struct soc15_reg_golden golden_settings_sdma_rv1[] = @@ -818,7 +820,7 @@ sdma_v4_1_update_power_gating(struct amdgpu_device *adev, bool enable) uint32_t def, data; if (enable && (adev->pg_flags & AMD_PG_SUPPORT_SDMA)) { - /* disable idle interrupt */ + /* enable idle interrupt */ def = data = RREG32(SOC15_REG_OFFSET(SDMA0, 0, mmSDMA0_CNTL)); data |= SDMA0_CNTL__CTXEMPTY_INT_ENABLE_MASK; @@ -1320,9 +1322,15 @@ static int sdma_v4_0_sw_init(void *handle) DRM_INFO("use_doorbell being set to: [%s]\n", ring->use_doorbell?"true":"false"); - ring->doorbell_index = (i == 0) ? - (AMDGPU_DOORBELL64_sDMA_ENGINE0 << 1) //get DWORD offset - : (AMDGPU_DOORBELL64_sDMA_ENGINE1 << 1); // get DWORD offset + if (adev->asic_type == CHIP_VEGA10) + ring->doorbell_index = (i == 0) ? + (AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE0 << 1) //get DWORD offset + : (AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE1 << 1); // get DWORD offset + else + ring->doorbell_index = (i == 0) ? + (AMDGPU_DOORBELL64_sDMA_ENGINE0 << 1) //get DWORD offset + : (AMDGPU_DOORBELL64_sDMA_ENGINE1 << 1); // get DWORD offset + sprintf(ring->name, "sdma%d", i); r = amdgpu_ring_init(adev, ring, 1024, @@ -1358,6 +1366,9 @@ static int sdma_v4_0_hw_init(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; + if (adev->asic_type == CHIP_RAVEN && adev->powerplay.pp_funcs->set_powergating_by_smu) + amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, false); + sdma_v4_0_init_golden_registers(adev); r = sdma_v4_0_start(adev); @@ -1375,6 +1386,9 @@ static int sdma_v4_0_hw_fini(void *handle) sdma_v4_0_ctx_switch_enable(adev, false); sdma_v4_0_enable(adev, false); + if (adev->asic_type == CHIP_RAVEN && adev->powerplay.pp_funcs->set_powergating_by_smu) + amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, true); + return 0; } @@ -1801,10 +1815,8 @@ static const struct amdgpu_buffer_funcs sdma_v4_0_buffer_funcs = { static void sdma_v4_0_set_buffer_funcs(struct amdgpu_device *adev) { - if (adev->mman.buffer_funcs == NULL) { - adev->mman.buffer_funcs = &sdma_v4_0_buffer_funcs; - adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring; - } + adev->mman.buffer_funcs = &sdma_v4_0_buffer_funcs; + adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring; } static const struct amdgpu_vm_pte_funcs sdma_v4_0_vm_pte_funcs = { @@ -1820,15 +1832,13 @@ static void sdma_v4_0_set_vm_pte_funcs(struct amdgpu_device *adev) struct drm_gpu_scheduler *sched; unsigned i; - if (adev->vm_manager.vm_pte_funcs == NULL) { - adev->vm_manager.vm_pte_funcs = &sdma_v4_0_vm_pte_funcs; - for (i = 0; i < adev->sdma.num_instances; i++) { - sched = &adev->sdma.instance[i].ring.sched; - adev->vm_manager.vm_pte_rqs[i] = - &sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL]; - } - adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances; + adev->vm_manager.vm_pte_funcs = &sdma_v4_0_vm_pte_funcs; + for (i = 0; i < adev->sdma.num_instances; i++) { + sched = &adev->sdma.instance[i].ring.sched; + adev->vm_manager.vm_pte_rqs[i] = + &sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL]; } + adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances; } const struct amdgpu_ip_block_version sdma_v4_0_ip_block = { diff --git a/drivers/gpu/drm/amd/amdgpu/si.c b/drivers/gpu/drm/amd/amdgpu/si.c index c364ef94cc36..f8408f88cd37 100644 --- a/drivers/gpu/drm/amd/amdgpu/si.c +++ b/drivers/gpu/drm/amd/amdgpu/si.c @@ -2057,13 +2057,13 @@ int si_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &si_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v6_0_ip_block); amdgpu_device_ip_block_add(adev, &si_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v6_0_ip_block); + amdgpu_device_ip_block_add(adev, &si_dma_ip_block); amdgpu_device_ip_block_add(adev, &si_smu_ip_block); if (adev->enable_virtual_display) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); else amdgpu_device_ip_block_add(adev, &dce_v6_0_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v6_0_ip_block); - amdgpu_device_ip_block_add(adev, &si_dma_ip_block); /* amdgpu_device_ip_block_add(adev, &uvd_v3_1_ip_block); */ /* amdgpu_device_ip_block_add(adev, &vce_v1_0_ip_block); */ break; @@ -2071,13 +2071,14 @@ int si_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &si_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v6_0_ip_block); amdgpu_device_ip_block_add(adev, &si_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v6_0_ip_block); + amdgpu_device_ip_block_add(adev, &si_dma_ip_block); amdgpu_device_ip_block_add(adev, &si_smu_ip_block); if (adev->enable_virtual_display) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); else amdgpu_device_ip_block_add(adev, &dce_v6_4_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v6_0_ip_block); - amdgpu_device_ip_block_add(adev, &si_dma_ip_block); + /* amdgpu_device_ip_block_add(adev, &uvd_v3_1_ip_block); */ /* amdgpu_device_ip_block_add(adev, &vce_v1_0_ip_block); */ break; @@ -2085,11 +2086,11 @@ int si_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &si_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v6_0_ip_block); amdgpu_device_ip_block_add(adev, &si_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v6_0_ip_block); + amdgpu_device_ip_block_add(adev, &si_dma_ip_block); amdgpu_device_ip_block_add(adev, &si_smu_ip_block); if (adev->enable_virtual_display) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v6_0_ip_block); - amdgpu_device_ip_block_add(adev, &si_dma_ip_block); break; default: BUG(); diff --git a/drivers/gpu/drm/amd/amdgpu/si_dma.c b/drivers/gpu/drm/amd/amdgpu/si_dma.c index fafaf259b17b..adbaea6da0d7 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_dma.c +++ b/drivers/gpu/drm/amd/amdgpu/si_dma.c @@ -502,12 +502,14 @@ static int si_dma_sw_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; /* DMA0 trap event */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 224, &adev->sdma.trap_irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 224, + &adev->sdma.trap_irq); if (r) return r; /* DMA1 trap event */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 244, &adev->sdma.trap_irq_1); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 244, + &adev->sdma.trap_irq); if (r) return r; @@ -649,17 +651,10 @@ static int si_dma_process_trap_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) { - amdgpu_fence_process(&adev->sdma.instance[0].ring); - - return 0; -} - -static int si_dma_process_trap_irq_1(struct amdgpu_device *adev, - struct amdgpu_irq_src *source, - struct amdgpu_iv_entry *entry) -{ - amdgpu_fence_process(&adev->sdma.instance[1].ring); - + if (entry->src_id == 224) + amdgpu_fence_process(&adev->sdma.instance[0].ring); + else + amdgpu_fence_process(&adev->sdma.instance[1].ring); return 0; } @@ -786,11 +781,6 @@ static const struct amdgpu_irq_src_funcs si_dma_trap_irq_funcs = { .process = si_dma_process_trap_irq, }; -static const struct amdgpu_irq_src_funcs si_dma_trap_irq_funcs_1 = { - .set = si_dma_set_trap_irq_state, - .process = si_dma_process_trap_irq_1, -}; - static const struct amdgpu_irq_src_funcs si_dma_illegal_inst_irq_funcs = { .process = si_dma_process_illegal_inst_irq, }; @@ -799,7 +789,6 @@ static void si_dma_set_irq_funcs(struct amdgpu_device *adev) { adev->sdma.trap_irq.num_types = AMDGPU_SDMA_IRQ_LAST; adev->sdma.trap_irq.funcs = &si_dma_trap_irq_funcs; - adev->sdma.trap_irq_1.funcs = &si_dma_trap_irq_funcs_1; adev->sdma.illegal_inst_irq.funcs = &si_dma_illegal_inst_irq_funcs; } @@ -863,10 +852,8 @@ static const struct amdgpu_buffer_funcs si_dma_buffer_funcs = { static void si_dma_set_buffer_funcs(struct amdgpu_device *adev) { - if (adev->mman.buffer_funcs == NULL) { - adev->mman.buffer_funcs = &si_dma_buffer_funcs; - adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring; - } + adev->mman.buffer_funcs = &si_dma_buffer_funcs; + adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring; } static const struct amdgpu_vm_pte_funcs si_dma_vm_pte_funcs = { @@ -882,15 +869,13 @@ static void si_dma_set_vm_pte_funcs(struct amdgpu_device *adev) struct drm_gpu_scheduler *sched; unsigned i; - if (adev->vm_manager.vm_pte_funcs == NULL) { - adev->vm_manager.vm_pte_funcs = &si_dma_vm_pte_funcs; - for (i = 0; i < adev->sdma.num_instances; i++) { - sched = &adev->sdma.instance[i].ring.sched; - adev->vm_manager.vm_pte_rqs[i] = - &sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL]; - } - adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances; + adev->vm_manager.vm_pte_funcs = &si_dma_vm_pte_funcs; + for (i = 0; i < adev->sdma.num_instances; i++) { + sched = &adev->sdma.instance[i].ring.sched; + adev->vm_manager.vm_pte_rqs[i] = + &sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL]; } + adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances; } const struct amdgpu_ip_block_version si_dma_ip_block = diff --git a/drivers/gpu/drm/amd/amdgpu/si_dpm.c b/drivers/gpu/drm/amd/amdgpu/si_dpm.c index 1de96995e690..da58040fdbdc 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/si_dpm.c @@ -7687,11 +7687,11 @@ static int si_dpm_sw_init(void *handle) int ret; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - ret = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 230, &adev->pm.dpm.thermal.irq); + ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 230, &adev->pm.dpm.thermal.irq); if (ret) return ret; - ret = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 231, &adev->pm.dpm.thermal.irq); + ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 231, &adev->pm.dpm.thermal.irq); if (ret) return ret; diff --git a/drivers/gpu/drm/amd/amdgpu/si_ih.c b/drivers/gpu/drm/amd/amdgpu/si_ih.c index 60dad63098a2..b3d7d9f83202 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/si_ih.c @@ -142,7 +142,7 @@ static void si_ih_decode_iv(struct amdgpu_device *adev, dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]); dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]); - entry->client_id = AMDGPU_IH_CLIENTID_LEGACY; + entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY; entry->src_id = dw[0] & 0xff; entry->src_data[0] = dw[1] & 0xfffffff; entry->ring_id = dw[2] & 0xff; @@ -170,7 +170,7 @@ static int si_ih_sw_init(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = amdgpu_ih_ring_init(adev, 64 * 1024, false); + r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, false); if (r) return r; @@ -182,7 +182,7 @@ static int si_ih_sw_fini(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; amdgpu_irq_fini(adev); - amdgpu_ih_ring_fini(adev); + amdgpu_ih_ring_fini(adev, &adev->irq.ih); return 0; } @@ -308,8 +308,7 @@ static const struct amdgpu_ih_funcs si_ih_funcs = { static void si_ih_set_interrupt_funcs(struct amdgpu_device *adev) { - if (adev->irq.ih_funcs == NULL) - adev->irq.ih_funcs = &si_ih_funcs; + adev->irq.ih_funcs = &si_ih_funcs; } const struct amdgpu_ip_block_version si_ih_ip_block = diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c index 138c4810a3de..bf5e6a413dee 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.c +++ b/drivers/gpu/drm/amd/amdgpu/soc15.c @@ -529,6 +529,8 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); else amdgpu_device_ip_block_add(adev, &psp_v3_1_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); + amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); if (!amdgpu_sriov_vf(adev)) amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) @@ -539,8 +541,6 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev) #else # warning "Enable CONFIG_DRM_AMD_DC for display support on SOC15." #endif - amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); if (!(adev->asic_type == CHIP_VEGA20 && amdgpu_sriov_vf(adev))) { amdgpu_device_ip_block_add(adev, &uvd_v7_0_ip_block); amdgpu_device_ip_block_add(adev, &vce_v4_0_ip_block); @@ -551,6 +551,8 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block); amdgpu_device_ip_block_add(adev, &vega10_ih_ip_block); amdgpu_device_ip_block_add(adev, &psp_v10_0_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); + amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); @@ -560,8 +562,6 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev) #else # warning "Enable CONFIG_DRM_AMD_DC for display support on SOC15." #endif - amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); amdgpu_device_ip_block_add(adev, &vcn_v1_0_ip_block); break; default: @@ -739,7 +739,8 @@ static int soc15_common_early_init(void *handle) adev->pg_flags = AMD_PG_SUPPORT_SDMA | AMD_PG_SUPPORT_MMHUB | - AMD_PG_SUPPORT_VCN; + AMD_PG_SUPPORT_VCN | + AMD_PG_SUPPORT_VCN_DPG; } else { adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_MGLS | diff --git a/drivers/gpu/drm/amd/amdgpu/soc15_common.h b/drivers/gpu/drm/amd/amdgpu/soc15_common.h index f5d602540673..958b10a57073 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15_common.h +++ b/drivers/gpu/drm/amd/amdgpu/soc15_common.h @@ -57,13 +57,33 @@ loop--; \ if (!loop) { \ DRM_ERROR("Register(%d) [%s] failed to reach value 0x%08x != 0x%08x\n", \ - inst, #reg, expected_value, (tmp_ & (mask))); \ + inst, #reg, (unsigned)expected_value, (unsigned)(tmp_ & (mask))); \ ret = -ETIMEDOUT; \ break; \ } \ } \ } while (0) +#define RREG32_SOC15_DPG_MODE(ip, inst, reg, mask, sram_sel) \ + ({ WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_MASK, mask); \ + WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_CTL, \ + UVD_DPG_LMA_CTL__MASK_EN_MASK | \ + ((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) \ + << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT) | \ + (sram_sel << UVD_DPG_LMA_CTL__SRAM_SEL__SHIFT)); \ + RREG32_SOC15(ip, inst, mmUVD_DPG_LMA_DATA); }) + +#define WREG32_SOC15_DPG_MODE(ip, inst, reg, value, mask, sram_sel) \ + do { \ + WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_DATA, value); \ + WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_MASK, mask); \ + WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_CTL, \ + UVD_DPG_LMA_CTL__READ_WRITE_MASK | \ + ((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) \ + << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT) | \ + (sram_sel << UVD_DPG_LMA_CTL__SRAM_SEL__SHIFT)); \ + } while (0) + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c index 52853d8a8fdd..3abffd06b5c7 100644 --- a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c @@ -266,7 +266,7 @@ static void tonga_ih_decode_iv(struct amdgpu_device *adev, dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]); dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]); - entry->client_id = AMDGPU_IH_CLIENTID_LEGACY; + entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY; entry->src_id = dw[0] & 0xff; entry->src_data[0] = dw[1] & 0xfffffff; entry->ring_id = dw[2] & 0xff; @@ -317,7 +317,7 @@ static int tonga_ih_sw_init(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = amdgpu_ih_ring_init(adev, 64 * 1024, true); + r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, true); if (r) return r; @@ -334,7 +334,7 @@ static int tonga_ih_sw_fini(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; amdgpu_irq_fini(adev); - amdgpu_ih_ring_fini(adev); + amdgpu_ih_ring_fini(adev, &adev->irq.ih); amdgpu_irq_remove_domain(adev); return 0; @@ -513,8 +513,7 @@ static const struct amdgpu_ih_funcs tonga_ih_funcs = { static void tonga_ih_set_interrupt_funcs(struct amdgpu_device *adev) { - if (adev->irq.ih_funcs == NULL) - adev->irq.ih_funcs = &tonga_ih_funcs; + adev->irq.ih_funcs = &tonga_ih_funcs; } const struct amdgpu_ip_block_version tonga_ih_ip_block = diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c index 8a926d1df939..1fc17bf39fed 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c @@ -108,7 +108,7 @@ static int uvd_v4_2_sw_init(void *handle) int r; /* UVD TRAP */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 124, &adev->uvd.inst->irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 124, &adev->uvd.inst->irq); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c index 50248059412e..fde6ad5ac9ab 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c @@ -105,7 +105,7 @@ static int uvd_v5_0_sw_init(void *handle) int r; /* UVD TRAP */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_UVD_SYSTEM_MESSAGE, &adev->uvd.inst->irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_UVD_SYSTEM_MESSAGE, &adev->uvd.inst->irq); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c index 6ae82cc2e55e..7a5b40275e8e 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c @@ -274,7 +274,7 @@ err: */ static int uvd_v6_0_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, - bool direct, struct dma_fence **fence) + struct dma_fence **fence) { const unsigned ib_size_dw = 16; struct amdgpu_job *job; @@ -310,11 +310,7 @@ static int uvd_v6_0_enc_get_destroy_msg(struct amdgpu_ring *ring, for (i = ib->length_dw; i < ib_size_dw; ++i) ib->ptr[i] = 0x0; - if (direct) - r = amdgpu_job_submit_direct(job, ring, &f); - else - r = amdgpu_job_submit(job, &ring->adev->vce.entity, - AMDGPU_FENCE_OWNER_UNDEFINED, &f); + r = amdgpu_job_submit_direct(job, ring, &f); if (r) goto err; @@ -345,7 +341,7 @@ static int uvd_v6_0_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout) goto error; } - r = uvd_v6_0_enc_get_destroy_msg(ring, 1, true, &fence); + r = uvd_v6_0_enc_get_destroy_msg(ring, 1, &fence); if (r) { DRM_ERROR("amdgpu: failed to get destroy ib (%ld).\n", r); goto error; @@ -393,14 +389,14 @@ static int uvd_v6_0_sw_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; /* UVD TRAP */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_UVD_SYSTEM_MESSAGE, &adev->uvd.inst->irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_UVD_SYSTEM_MESSAGE, &adev->uvd.inst->irq); if (r) return r; /* UVD ENC TRAP */ if (uvd_v6_0_enc_support(adev)) { for (i = 0; i < adev->uvd.num_enc_rings; ++i) { - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i + VISLANDS30_IV_SRCID_UVD_ENC_GEN_PURP, &adev->uvd.inst->irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + VISLANDS30_IV_SRCID_UVD_ENC_GEN_PURP, &adev->uvd.inst->irq); if (r) return r; } diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c index a289f6a20b6b..58b39afcfb86 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c @@ -280,8 +280,8 @@ err: * * Close up a stream for HW test or if userspace failed to do so */ -int uvd_v7_0_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, - bool direct, struct dma_fence **fence) +static int uvd_v7_0_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, + struct dma_fence **fence) { const unsigned ib_size_dw = 16; struct amdgpu_job *job; @@ -317,11 +317,7 @@ int uvd_v7_0_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, for (i = ib->length_dw; i < ib_size_dw; ++i) ib->ptr[i] = 0x0; - if (direct) - r = amdgpu_job_submit_direct(job, ring, &f); - else - r = amdgpu_job_submit(job, &ring->adev->vce.entity, - AMDGPU_FENCE_OWNER_UNDEFINED, &f); + r = amdgpu_job_submit_direct(job, ring, &f); if (r) goto err; @@ -352,7 +348,7 @@ static int uvd_v7_0_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout) goto error; } - r = uvd_v7_0_enc_get_destroy_msg(ring, 1, true, &fence); + r = uvd_v7_0_enc_get_destroy_msg(ring, 1, &fence); if (r) { DRM_ERROR("amdgpu: (%d)failed to get destroy ib (%ld).\n", ring->me, r); goto error; diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c index 7eaa54ba016b..ea28828360d3 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c @@ -417,7 +417,7 @@ static int vce_v2_0_sw_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; /* VCE */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 167, &adev->vce.irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 167, &adev->vce.irq); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c index c8390f9adfd6..6dbd39730070 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c @@ -423,7 +423,7 @@ static int vce_v3_0_sw_init(void *handle) int r, i; /* VCE */ - r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_VCE_TRAP, &adev->vce.irq); + r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_VCE_TRAP, &adev->vce.irq); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c index 2664bb2c47c3..eae90922fdbe 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c @@ -37,6 +37,11 @@ #include "ivsrcid/vcn/irqsrcs_vcn_1_0.h" +#define mmUVD_RBC_XX_IB_REG_CHECK 0x05ab +#define mmUVD_RBC_XX_IB_REG_CHECK_BASE_IDX 1 +#define mmUVD_REG_XX_MASK 0x05ac +#define mmUVD_REG_XX_MASK_BASE_IDX 1 + static int vcn_v1_0_stop(struct amdgpu_device *adev); static void vcn_v1_0_set_dec_ring_funcs(struct amdgpu_device *adev); static void vcn_v1_0_set_enc_ring_funcs(struct amdgpu_device *adev); @@ -198,7 +203,8 @@ static int vcn_v1_0_hw_init(void *handle) done: if (!r) - DRM_INFO("VCN decode and encode initialized successfully.\n"); + DRM_INFO("VCN decode and encode initialized successfully(under %s).\n", + (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)?"DPG Mode":"SPG Mode"); return r; } @@ -266,17 +272,18 @@ static int vcn_v1_0_resume(void *handle) } /** - * vcn_v1_0_mc_resume - memory controller programming + * vcn_v1_0_mc_resume_spg_mode - memory controller programming * * @adev: amdgpu_device pointer * * Let the VCN memory controller know it's offsets */ -static void vcn_v1_0_mc_resume(struct amdgpu_device *adev) +static void vcn_v1_0_mc_resume_spg_mode(struct amdgpu_device *adev) { uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4); uint32_t offset; + /* cache window 0: fw */ if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW, (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_lo)); @@ -296,20 +303,21 @@ static void vcn_v1_0_mc_resume(struct amdgpu_device *adev) WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE0, size); + /* cache window 1: stack */ WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW, lower_32_bits(adev->vcn.gpu_addr + offset)); WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH, upper_32_bits(adev->vcn.gpu_addr + offset)); WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET1, 0); - WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_HEAP_SIZE); + WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE); + /* cache window 2: context */ WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW, - lower_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_HEAP_SIZE)); + lower_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_STACK_SIZE)); WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH, - upper_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_HEAP_SIZE)); + upper_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_STACK_SIZE)); WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET2, 0); - WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE2, - AMDGPU_VCN_STACK_SIZE + (AMDGPU_VCN_SESSION_SIZE * 40)); + WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE); WREG32_SOC15(UVD, 0, mmUVD_UDEC_ADDR_CONFIG, adev->gfx.config.gb_addr_config); @@ -317,6 +325,96 @@ static void vcn_v1_0_mc_resume(struct amdgpu_device *adev) adev->gfx.config.gb_addr_config); WREG32_SOC15(UVD, 0, mmUVD_UDEC_DBW_ADDR_CONFIG, adev->gfx.config.gb_addr_config); + WREG32_SOC15(UVD, 0, mmUVD_UDEC_DBW_UV_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(UVD, 0, mmUVD_MIF_CURR_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(UVD, 0, mmUVD_MIF_CURR_UV_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(UVD, 0, mmUVD_MIF_RECON1_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(UVD, 0, mmUVD_MIF_RECON1_UV_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(UVD, 0, mmUVD_MIF_REF_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(UVD, 0, mmUVD_MIF_REF_UV_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(UVD, 0, mmUVD_JPEG_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(UVD, 0, mmUVD_JPEG_UV_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); +} + +static void vcn_v1_0_mc_resume_dpg_mode(struct amdgpu_device *adev) +{ + uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4); + uint32_t offset; + + /* cache window 0: fw */ + if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW, + (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_lo), + 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH, + (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_hi), + 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0, 0, + 0xFFFFFFFF, 0); + offset = 0; + } else { + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW, + lower_32_bits(adev->vcn.gpu_addr), 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH, + upper_32_bits(adev->vcn.gpu_addr), 0xFFFFFFFF, 0); + offset = size; + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0, + AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0xFFFFFFFF, 0); + } + + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_SIZE0, size, 0xFFFFFFFF, 0); + + /* cache window 1: stack */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW, + lower_32_bits(adev->vcn.gpu_addr + offset), 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH, + upper_32_bits(adev->vcn.gpu_addr + offset), 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_OFFSET1, 0, + 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE, + 0xFFFFFFFF, 0); + + /* cache window 2: context */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW, + lower_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), + 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH, + upper_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), + 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_OFFSET2, 0, 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE, + 0xFFFFFFFF, 0); + + /* VCN global tiling registers */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_UDEC_ADDR_CONFIG, + adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_UDEC_DB_ADDR_CONFIG, + adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_UDEC_DBW_ADDR_CONFIG, + adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_UDEC_DBW_UV_ADDR_CONFIG, + adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_CURR_ADDR_CONFIG, + adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_CURR_UV_ADDR_CONFIG, + adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_RECON1_ADDR_CONFIG, + adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_RECON1_UV_ADDR_CONFIG, + adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_REF_ADDR_CONFIG, + adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_REF_UV_ADDR_CONFIG, + adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0); } /** @@ -519,6 +617,60 @@ static void vcn_v1_0_enable_clock_gating(struct amdgpu_device *adev) WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL, data); } +static void vcn_v1_0_clock_gating_dpg_mode(struct amdgpu_device *adev, uint8_t sram_sel) +{ + uint32_t reg_data = 0; + + /* disable JPEG CGC */ + if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG) + reg_data = 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + else + reg_data = 0 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + reg_data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT; + reg_data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT; + WREG32_SOC15_DPG_MODE(UVD, 0, mmJPEG_CGC_CTRL, reg_data, 0xFFFFFFFF, sram_sel); + + WREG32_SOC15_DPG_MODE(UVD, 0, mmJPEG_CGC_GATE, 0, 0xFFFFFFFF, sram_sel); + + /* enable sw clock gating control */ + if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG) + reg_data = 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + else + 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__UDEC_RE_MODE_MASK | + UVD_CGC_CTRL__UDEC_CM_MODE_MASK | + UVD_CGC_CTRL__UDEC_IT_MODE_MASK | + UVD_CGC_CTRL__UDEC_DB_MODE_MASK | + UVD_CGC_CTRL__UDEC_MP_MODE_MASK | + UVD_CGC_CTRL__SYS_MODE_MASK | + UVD_CGC_CTRL__UDEC_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 | + UVD_CGC_CTRL__SCPU_MODE_MASK); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_CGC_CTRL, reg_data, 0xFFFFFFFF, sram_sel); + + /* turn off clock gating */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_CGC_GATE, 0, 0xFFFFFFFF, sram_sel); + + /* turn on SUVD clock gating */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_SUVD_CGC_GATE, 1, 0xFFFFFFFF, sram_sel); + + /* turn on sw mode in UVD_SUVD_CGC_CTRL */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_SUVD_CGC_CTRL, 0, 0xFFFFFFFF, sram_sel); +} + static void vcn_1_0_disable_static_power_gating(struct amdgpu_device *adev) { uint32_t data = 0; @@ -614,7 +766,7 @@ static void vcn_1_0_enable_static_power_gating(struct amdgpu_device *adev) * * Setup and start the VCN block */ -static int vcn_v1_0_start(struct amdgpu_device *adev) +static int vcn_v1_0_start_spg_mode(struct amdgpu_device *adev) { struct amdgpu_ring *ring = &adev->vcn.ring_dec; uint32_t rb_bufsz, tmp; @@ -625,41 +777,24 @@ static int vcn_v1_0_start(struct amdgpu_device *adev) lmi_swap_cntl = 0; vcn_1_0_disable_static_power_gating(adev); + + tmp = RREG32_SOC15(UVD, 0, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY; + WREG32_SOC15(UVD, 0, mmUVD_STATUS, tmp); + /* disable clock gating */ vcn_v1_0_disable_clock_gating(adev); - vcn_v1_0_mc_resume(adev); - /* disable interupt */ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_MASTINT_EN), 0, ~UVD_MASTINT_EN__VCPU_EN_MASK); - /* stall UMC and register bus before resetting VCPU */ - WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2), - UVD_LMI_CTRL2__STALL_ARB_UMC_MASK, - ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); - mdelay(1); - - /* put LMI, VCPU, RBC etc... into reset */ - WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET, - UVD_SOFT_RESET__LMI_SOFT_RESET_MASK | - UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK | - UVD_SOFT_RESET__LBSI_SOFT_RESET_MASK | - UVD_SOFT_RESET__RBC_SOFT_RESET_MASK | - UVD_SOFT_RESET__CSM_SOFT_RESET_MASK | - UVD_SOFT_RESET__CXW_SOFT_RESET_MASK | - UVD_SOFT_RESET__TAP_SOFT_RESET_MASK | - UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK); - mdelay(5); - /* initialize VCN memory controller */ - WREG32_SOC15(UVD, 0, mmUVD_LMI_CTRL, - (0x40 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) | - UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | - UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | - UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK | - UVD_LMI_CTRL__REQ_MODE_MASK | - 0x00100000L); + tmp = RREG32_SOC15(UVD, 0, mmUVD_LMI_CTRL); + WREG32_SOC15(UVD, 0, mmUVD_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); #ifdef __BIG_ENDIAN /* swap (8 in 32) RB and IB */ @@ -667,41 +802,61 @@ static int vcn_v1_0_start(struct amdgpu_device *adev) #endif WREG32_SOC15(UVD, 0, mmUVD_LMI_SWAP_CNTL, lmi_swap_cntl); - WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXA0, 0x40c2040); - WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXA1, 0x0); - WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXB0, 0x40c2040); - WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXB1, 0x0); - WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_ALU, 0); - WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUX, 0x88); + tmp = RREG32_SOC15(UVD, 0, mmUVD_MPC_CNTL); + tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK; + tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT; + WREG32_SOC15(UVD, 0, mmUVD_MPC_CNTL, tmp); + + WREG32_SOC15(UVD, 0, mmUVD_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))); + + WREG32_SOC15(UVD, 0, mmUVD_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))); - /* take all subblocks out of reset, except VCPU */ - WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET, - UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK); - mdelay(5); + WREG32_SOC15(UVD, 0, mmUVD_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_v1_0_mc_resume_spg_mode(adev); + + WREG32_SOC15(UVD, 0, mmUVD_REG_XX_MASK, 0x10); + WREG32_SOC15(UVD, 0, mmUVD_RBC_XX_IB_REG_CHECK, + RREG32_SOC15(UVD, 0, mmUVD_RBC_XX_IB_REG_CHECK) | 0x3); /* enable VCPU clock */ - WREG32_SOC15(UVD, 0, mmUVD_VCPU_CNTL, - UVD_VCPU_CNTL__CLK_EN_MASK); + WREG32_SOC15(UVD, 0, mmUVD_VCPU_CNTL, UVD_VCPU_CNTL__CLK_EN_MASK); + + /* boot up the VCPU */ + WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET), 0, + ~UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK); /* enable UMC */ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2), 0, ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); - /* boot up the VCPU */ - WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET, 0); - mdelay(10); + tmp = RREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET); + tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET, tmp); for (i = 0; i < 10; ++i) { uint32_t status; for (j = 0; j < 100; ++j) { status = RREG32_SOC15(UVD, 0, mmUVD_STATUS); - if (status & 2) + if (status & UVD_STATUS__IDLE) break; mdelay(10); } r = 0; - if (status & 2) + if (status & UVD_STATUS__IDLE) break; DRM_ERROR("VCN decode not responding, trying to reset the VCPU!!!\n"); @@ -721,24 +876,22 @@ static int vcn_v1_0_start(struct amdgpu_device *adev) } /* enable master interrupt */ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_MASTINT_EN), - (UVD_MASTINT_EN__VCPU_EN_MASK|UVD_MASTINT_EN__SYS_EN_MASK), - ~(UVD_MASTINT_EN__VCPU_EN_MASK|UVD_MASTINT_EN__SYS_EN_MASK)); + UVD_MASTINT_EN__VCPU_EN_MASK, ~UVD_MASTINT_EN__VCPU_EN_MASK); /* enable system interrupt for JRBC, TODO: move to set interrupt*/ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SYS_INT_EN), UVD_SYS_INT_EN__UVD_JRBC_EN_MASK, ~UVD_SYS_INT_EN__UVD_JRBC_EN_MASK); - /* clear the bit 4 of VCN_STATUS */ - WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_STATUS), 0, - ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); + /* clear the busy bit of UVD_STATUS */ + tmp = RREG32_SOC15(UVD, 0, mmUVD_STATUS) & ~UVD_STATUS__UVD_BUSY; + WREG32_SOC15(UVD, 0, mmUVD_STATUS, tmp); /* force RBC into idle state */ rb_bufsz = order_base_2(ring->ring_size); tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz); tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1); tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1); - tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_WPTR_POLL_EN, 0); tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1); tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1); WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_CNTL, tmp); @@ -759,6 +912,8 @@ static int vcn_v1_0_start(struct amdgpu_device *adev) /* Initialize the ring buffer's read and write pointers */ WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR, 0); + WREG32_SOC15(UVD, 0, mmUVD_SCRATCH2, 0); + ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR); WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr)); @@ -782,12 +937,13 @@ static int vcn_v1_0_start(struct amdgpu_device *adev) ring = &adev->vcn.ring_jpeg; WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_VMID, 0); - WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L)); + WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK | + UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK); WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_64BIT_BAR_LOW, lower_32_bits(ring->gpu_addr)); WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_64BIT_BAR_HIGH, upper_32_bits(ring->gpu_addr)); WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_RPTR, 0); WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR, 0); - WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, 0x00000002L); + WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK); /* initialize wptr */ ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR); @@ -799,6 +955,166 @@ static int vcn_v1_0_start(struct amdgpu_device *adev) return 0; } +static int vcn_v1_0_start_dpg_mode(struct amdgpu_device *adev) +{ + struct amdgpu_ring *ring = &adev->vcn.ring_dec; + uint32_t rb_bufsz, tmp; + uint32_t lmi_swap_cntl; + + /* disable byte swapping */ + lmi_swap_cntl = 0; + + vcn_1_0_enable_static_power_gating(adev); + + /* enable dynamic power gating mode */ + tmp = RREG32_SOC15(UVD, 0, mmUVD_POWER_STATUS); + tmp |= UVD_POWER_STATUS__UVD_PG_MODE_MASK; + tmp |= UVD_POWER_STATUS__UVD_PG_EN_MASK; + WREG32_SOC15(UVD, 0, mmUVD_POWER_STATUS, tmp); + + /* enable clock gating */ + vcn_v1_0_clock_gating_dpg_mode(adev, 0); + + /* enable VCPU clock */ + tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT); + tmp |= UVD_VCPU_CNTL__CLK_EN_MASK; + tmp |= UVD_VCPU_CNTL__MIF_WR_LOW_THRESHOLD_BP_MASK; + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CNTL, tmp, 0xFFFFFFFF, 0); + + /* disable interupt */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MASTINT_EN, + 0, UVD_MASTINT_EN__VCPU_EN_MASK, 0); + + /* initialize VCN memory controller */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_CTRL, + (8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) | + UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | + UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__REQ_MODE_MASK | + UVD_LMI_CTRL__CRC_RESET_MASK | + UVD_LMI_CTRL__MASK_MC_URGENT_MASK | + 0x00100000L, 0xFFFFFFFF, 0); + +#ifdef __BIG_ENDIAN + /* swap (8 in 32) RB and IB */ + lmi_swap_cntl = 0xa; +#endif + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_SWAP_CNTL, lmi_swap_cntl, 0xFFFFFFFF, 0); + + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MPC_CNTL, + 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT, 0xFFFFFFFF, 0); + + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_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)), 0xFFFFFFFF, 0); + + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_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)), 0xFFFFFFFF, 0); + + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_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)), 0xFFFFFFFF, 0); + + vcn_v1_0_mc_resume_dpg_mode(adev); + + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_REG_XX_MASK, 0x10, 0xFFFFFFFF, 0); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_RBC_XX_IB_REG_CHECK, 0x3, 0xFFFFFFFF, 0); + + /* boot up the VCPU */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_SOFT_RESET, 0, 0xFFFFFFFF, 0); + + /* enable UMC */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_CTRL2, + 0x1F << UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM__SHIFT, + 0xFFFFFFFF, 0); + + /* enable master interrupt */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MASTINT_EN, + UVD_MASTINT_EN__VCPU_EN_MASK, UVD_MASTINT_EN__VCPU_EN_MASK, 0); + + vcn_v1_0_clock_gating_dpg_mode(adev, 1); + /* setup mmUVD_LMI_CTRL */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_CTRL, + (8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) | + UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | + UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__REQ_MODE_MASK | + UVD_LMI_CTRL__CRC_RESET_MASK | + UVD_LMI_CTRL__MASK_MC_URGENT_MASK | + 0x00100000L, 0xFFFFFFFF, 1); + + tmp = adev->gfx.config.gb_addr_config; + /* setup VCN global tiling registers */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_JPEG_ADDR_CONFIG, tmp, 0xFFFFFFFF, 1); + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_JPEG_UV_ADDR_CONFIG, tmp, 0xFFFFFFFF, 1); + + /* enable System Interrupt for JRBC */ + WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_SYS_INT_EN, + UVD_SYS_INT_EN__UVD_JRBC_EN_MASK, 0xFFFFFFFF, 1); + + /* force RBC into idle state */ + rb_bufsz = order_base_2(ring->ring_size); + tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz); + tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1); + tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1); + tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1); + tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1); + WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_CNTL, tmp); + + /* set the write pointer delay */ + WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR_CNTL, 0); + + /* set the wb address */ + WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR_ADDR, + (upper_32_bits(ring->gpu_addr) >> 2)); + + /* programm the RB_BASE for ring buffer */ + WREG32_SOC15(UVD, 0, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW, + lower_32_bits(ring->gpu_addr)); + WREG32_SOC15(UVD, 0, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH, + upper_32_bits(ring->gpu_addr)); + + /* Initialize the ring buffer's read and write pointers */ + WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR, 0); + + WREG32_SOC15(UVD, 0, mmUVD_SCRATCH2, 0); + + ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR); + WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, + lower_32_bits(ring->wptr)); + + WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_CNTL), 0, + ~UVD_RBC_RB_CNTL__RB_NO_FETCH_MASK); + + /* initialize wptr */ + ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR); + + /* copy patch commands to the jpeg ring */ + vcn_v1_0_jpeg_ring_set_patch_ring(ring, + (ring->wptr + ring->max_dw * amdgpu_sched_hw_submission)); + + return 0; +} + +static int vcn_v1_0_start(struct amdgpu_device *adev) +{ + int r; + + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + r = vcn_v1_0_start_dpg_mode(adev); + else + r = vcn_v1_0_start_spg_mode(adev); + return r; +} + /** * vcn_v1_0_stop - stop VCN block * @@ -806,41 +1122,90 @@ static int vcn_v1_0_start(struct amdgpu_device *adev) * * stop the VCN block */ -static int vcn_v1_0_stop(struct amdgpu_device *adev) +static int vcn_v1_0_stop_spg_mode(struct amdgpu_device *adev) { - /* force RBC into idle state */ - WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_CNTL, 0x11010101); + int ret_code, tmp; - /* Stall UMC and register bus before resetting VCPU */ - WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2), - UVD_LMI_CTRL2__STALL_ARB_UMC_MASK, - ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); - mdelay(1); + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7, ret_code); + + 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; + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_LMI_STATUS, tmp, tmp, ret_code); /* put VCPU into reset */ - WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET, - UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK); - mdelay(5); + WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET), + UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK, + ~UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK); + + tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK | + UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_LMI_STATUS, tmp, tmp, ret_code); /* disable VCPU clock */ - WREG32_SOC15(UVD, 0, mmUVD_VCPU_CNTL, 0x0); + WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__CLK_EN_MASK); - /* Unstall UMC and register bus */ - WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2), 0, - ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); + /* reset LMI UMC/LMI */ + WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET), + UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK, + ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK); + + WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET), + UVD_SOFT_RESET__LMI_SOFT_RESET_MASK, + ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK); - WREG32_SOC15(VCN, 0, mmUVD_STATUS, 0); + WREG32_SOC15(UVD, 0, mmUVD_STATUS, 0); vcn_v1_0_enable_clock_gating(adev); vcn_1_0_enable_static_power_gating(adev); return 0; } +static int vcn_v1_0_stop_dpg_mode(struct amdgpu_device *adev) +{ + int ret_code = 0; + + /* Wait for power status to be UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF */ + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS, + UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF, + UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code); + + if (!ret_code) { + int tmp = RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR) & 0x7FFFFFFF; + /* wait for read ptr to be equal to write ptr */ + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_RBC_RB_RPTR, tmp, 0xFFFFFFFF, ret_code); + + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS, + UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF, + UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code); + } + + /* disable dynamic power gating mode */ + WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_POWER_STATUS), 0, + ~UVD_POWER_STATUS__UVD_PG_MODE_MASK); + + return 0; +} + +static int vcn_v1_0_stop(struct amdgpu_device *adev) +{ + int r; + + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + r = vcn_v1_0_stop_dpg_mode(adev); + else + r = vcn_v1_0_stop_spg_mode(adev); + + return r; +} + static bool vcn_v1_0_is_idle(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - return (RREG32_SOC15(VCN, 0, mmUVD_STATUS) == 0x2); + return (RREG32_SOC15(VCN, 0, mmUVD_STATUS) == UVD_STATUS__IDLE); } static int vcn_v1_0_wait_for_idle(void *handle) @@ -848,7 +1213,8 @@ static int vcn_v1_0_wait_for_idle(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; int ret = 0; - SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_STATUS, 0x2, 0x2, ret); + SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_STATUS, UVD_STATUS__IDLE, + UVD_STATUS__IDLE, ret); return ret; } @@ -910,6 +1276,10 @@ static void vcn_v1_0_dec_ring_set_wptr(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + WREG32_SOC15(UVD, 0, mmUVD_SCRATCH2, + lower_32_bits(ring->wptr) | 0x80000000); + WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr)); } @@ -1633,12 +2003,20 @@ static int vcn_v1_0_set_powergating_state(void *handle, * revisit this when there is a cleaner line between * the smc and the hw blocks */ + int ret; struct amdgpu_device *adev = (struct amdgpu_device *)handle; + if(state == adev->vcn.cur_state) + return 0; + if (state == AMD_PG_STATE_GATE) - return vcn_v1_0_stop(adev); + ret = vcn_v1_0_stop(adev); else - return vcn_v1_0_start(adev); + ret = vcn_v1_0_start(adev); + + if(!ret) + adev->vcn.cur_state = state; + return ret; } static const struct amd_ip_funcs vcn_v1_0_ip_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c index acbe5a770207..a99f71797aa3 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c @@ -380,7 +380,7 @@ static int vega10_ih_sw_init(void *handle) int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = amdgpu_ih_ring_init(adev, 256 * 1024, true); + r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 256 * 1024, true); if (r) return r; @@ -397,7 +397,7 @@ static int vega10_ih_sw_fini(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; amdgpu_irq_fini(adev); - amdgpu_ih_ring_fini(adev); + amdgpu_ih_ring_fini(adev, &adev->irq.ih); return 0; } @@ -494,8 +494,7 @@ static const struct amdgpu_ih_funcs vega10_ih_funcs = { static void vega10_ih_set_interrupt_funcs(struct amdgpu_device *adev) { - if (adev->irq.ih_funcs == NULL) - adev->irq.ih_funcs = &vega10_ih_funcs; + adev->irq.ih_funcs = &vega10_ih_funcs; } const struct amdgpu_ip_block_version vega10_ih_ip_block = diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c index 88b57a5e9489..07880d35e9de 100644 --- a/drivers/gpu/drm/amd/amdgpu/vi.c +++ b/drivers/gpu/drm/amd/amdgpu/vi.c @@ -1596,16 +1596,18 @@ int vi_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &vi_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v7_4_ip_block); amdgpu_device_ip_block_add(adev, &iceland_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block); + amdgpu_device_ip_block_add(adev, &sdma_v2_4_ip_block); amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); if (adev->enable_virtual_display) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v2_4_ip_block); break; case CHIP_FIJI: amdgpu_device_ip_block_add(adev, &vi_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v8_5_ip_block); amdgpu_device_ip_block_add(adev, &tonga_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block); + amdgpu_device_ip_block_add(adev, &sdma_v3_0_ip_block); amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); @@ -1615,8 +1617,6 @@ int vi_set_ip_blocks(struct amdgpu_device *adev) #endif else amdgpu_device_ip_block_add(adev, &dce_v10_1_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v3_0_ip_block); if (!amdgpu_sriov_vf(adev)) { amdgpu_device_ip_block_add(adev, &uvd_v6_0_ip_block); amdgpu_device_ip_block_add(adev, &vce_v3_0_ip_block); @@ -1626,6 +1626,8 @@ int vi_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &vi_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v8_0_ip_block); amdgpu_device_ip_block_add(adev, &tonga_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block); + amdgpu_device_ip_block_add(adev, &sdma_v3_0_ip_block); amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); @@ -1635,8 +1637,6 @@ int vi_set_ip_blocks(struct amdgpu_device *adev) #endif else amdgpu_device_ip_block_add(adev, &dce_v10_0_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v3_0_ip_block); if (!amdgpu_sriov_vf(adev)) { amdgpu_device_ip_block_add(adev, &uvd_v5_0_ip_block); amdgpu_device_ip_block_add(adev, &vce_v3_0_ip_block); @@ -1649,6 +1649,8 @@ int vi_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &vi_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v8_1_ip_block); amdgpu_device_ip_block_add(adev, &tonga_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block); + amdgpu_device_ip_block_add(adev, &sdma_v3_1_ip_block); amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); if (adev->enable_virtual_display) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); @@ -1658,8 +1660,6 @@ int vi_set_ip_blocks(struct amdgpu_device *adev) #endif else amdgpu_device_ip_block_add(adev, &dce_v11_2_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v3_1_ip_block); amdgpu_device_ip_block_add(adev, &uvd_v6_3_ip_block); amdgpu_device_ip_block_add(adev, &vce_v3_4_ip_block); break; @@ -1667,6 +1667,8 @@ int vi_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &vi_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v8_0_ip_block); amdgpu_device_ip_block_add(adev, &cz_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block); + amdgpu_device_ip_block_add(adev, &sdma_v3_0_ip_block); amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); if (adev->enable_virtual_display) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); @@ -1676,8 +1678,6 @@ int vi_set_ip_blocks(struct amdgpu_device *adev) #endif else amdgpu_device_ip_block_add(adev, &dce_v11_0_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v3_0_ip_block); amdgpu_device_ip_block_add(adev, &uvd_v6_0_ip_block); amdgpu_device_ip_block_add(adev, &vce_v3_1_ip_block); #if defined(CONFIG_DRM_AMD_ACP) @@ -1688,6 +1688,8 @@ int vi_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &vi_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v8_0_ip_block); amdgpu_device_ip_block_add(adev, &cz_ih_ip_block); + amdgpu_device_ip_block_add(adev, &gfx_v8_1_ip_block); + amdgpu_device_ip_block_add(adev, &sdma_v3_0_ip_block); amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); if (adev->enable_virtual_display) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); @@ -1697,8 +1699,6 @@ int vi_set_ip_blocks(struct amdgpu_device *adev) #endif else amdgpu_device_ip_block_add(adev, &dce_v11_0_ip_block); - amdgpu_device_ip_block_add(adev, &gfx_v8_1_ip_block); - amdgpu_device_ip_block_add(adev, &sdma_v3_0_ip_block); amdgpu_device_ip_block_add(adev, &uvd_v6_2_ip_block); amdgpu_device_ip_block_add(adev, &vce_v3_4_ip_block); #if defined(CONFIG_DRM_AMD_ACP) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c index 758398bdb39b..14d5b5fa822d 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c @@ -447,6 +447,24 @@ static int kfd_ioctl_set_cu_mask(struct file *filp, struct kfd_process *p, return retval; } +static int kfd_ioctl_get_queue_wave_state(struct file *filep, + struct kfd_process *p, void *data) +{ + struct kfd_ioctl_get_queue_wave_state_args *args = data; + int r; + + mutex_lock(&p->mutex); + + r = pqm_get_wave_state(&p->pqm, args->queue_id, + (void __user *)args->ctl_stack_address, + &args->ctl_stack_used_size, + &args->save_area_used_size); + + mutex_unlock(&p->mutex); + + return r; +} + static int kfd_ioctl_set_memory_policy(struct file *filep, struct kfd_process *p, void *data) { @@ -1615,6 +1633,9 @@ static const struct amdkfd_ioctl_desc amdkfd_ioctls[] = { AMDKFD_IOCTL_DEF(AMDKFD_IOC_SET_CU_MASK, kfd_ioctl_set_cu_mask, 0), + AMDKFD_IOCTL_DEF(AMDKFD_IOC_GET_QUEUE_WAVE_STATE, + kfd_ioctl_get_queue_wave_state, 0) + }; #define AMDKFD_CORE_IOCTL_COUNT ARRAY_SIZE(amdkfd_ioctls) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c index d4560f1869bd..56412b0e7e1c 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c @@ -647,6 +647,7 @@ static int kfd_fill_gpu_cache_info(struct kfd_dev *kdev, num_of_cache_types = ARRAY_SIZE(polaris11_cache_info); break; case CHIP_VEGA10: + case CHIP_VEGA20: pcache_info = vega10_cache_info; num_of_cache_types = ARRAY_SIZE(vega10_cache_info); break; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c index 9b4e6ad4a7df..a9f18ea7e354 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c @@ -53,6 +53,7 @@ static const struct kfd_device_info kaveri_device_info = { .needs_iommu_device = true, .needs_pci_atomics = false, .num_sdma_engines = 2, + .num_sdma_queues_per_engine = 2, }; static const struct kfd_device_info carrizo_device_info = { @@ -69,6 +70,7 @@ static const struct kfd_device_info carrizo_device_info = { .needs_iommu_device = true, .needs_pci_atomics = false, .num_sdma_engines = 2, + .num_sdma_queues_per_engine = 2, }; static const struct kfd_device_info raven_device_info = { @@ -84,6 +86,7 @@ static const struct kfd_device_info raven_device_info = { .needs_iommu_device = true, .needs_pci_atomics = true, .num_sdma_engines = 1, + .num_sdma_queues_per_engine = 2, }; #endif @@ -101,6 +104,7 @@ static const struct kfd_device_info hawaii_device_info = { .needs_iommu_device = false, .needs_pci_atomics = false, .num_sdma_engines = 2, + .num_sdma_queues_per_engine = 2, }; static const struct kfd_device_info tonga_device_info = { @@ -116,21 +120,7 @@ static const struct kfd_device_info tonga_device_info = { .needs_iommu_device = false, .needs_pci_atomics = true, .num_sdma_engines = 2, -}; - -static const struct kfd_device_info tonga_vf_device_info = { - .asic_family = CHIP_TONGA, - .max_pasid_bits = 16, - .max_no_of_hqd = 24, - .doorbell_size = 4, - .ih_ring_entry_size = 4 * sizeof(uint32_t), - .event_interrupt_class = &event_interrupt_class_cik, - .num_of_watch_points = 4, - .mqd_size_aligned = MQD_SIZE_ALIGNED, - .supports_cwsr = false, - .needs_iommu_device = false, - .needs_pci_atomics = false, - .num_sdma_engines = 2, + .num_sdma_queues_per_engine = 2, }; static const struct kfd_device_info fiji_device_info = { @@ -146,6 +136,7 @@ static const struct kfd_device_info fiji_device_info = { .needs_iommu_device = false, .needs_pci_atomics = true, .num_sdma_engines = 2, + .num_sdma_queues_per_engine = 2, }; static const struct kfd_device_info fiji_vf_device_info = { @@ -161,6 +152,7 @@ static const struct kfd_device_info fiji_vf_device_info = { .needs_iommu_device = false, .needs_pci_atomics = false, .num_sdma_engines = 2, + .num_sdma_queues_per_engine = 2, }; @@ -177,6 +169,7 @@ static const struct kfd_device_info polaris10_device_info = { .needs_iommu_device = false, .needs_pci_atomics = true, .num_sdma_engines = 2, + .num_sdma_queues_per_engine = 2, }; static const struct kfd_device_info polaris10_vf_device_info = { @@ -192,6 +185,7 @@ static const struct kfd_device_info polaris10_vf_device_info = { .needs_iommu_device = false, .needs_pci_atomics = false, .num_sdma_engines = 2, + .num_sdma_queues_per_engine = 2, }; static const struct kfd_device_info polaris11_device_info = { @@ -207,6 +201,7 @@ static const struct kfd_device_info polaris11_device_info = { .needs_iommu_device = false, .needs_pci_atomics = true, .num_sdma_engines = 2, + .num_sdma_queues_per_engine = 2, }; static const struct kfd_device_info vega10_device_info = { @@ -222,6 +217,7 @@ static const struct kfd_device_info vega10_device_info = { .needs_iommu_device = false, .needs_pci_atomics = false, .num_sdma_engines = 2, + .num_sdma_queues_per_engine = 2, }; static const struct kfd_device_info vega10_vf_device_info = { @@ -237,8 +233,24 @@ static const struct kfd_device_info vega10_vf_device_info = { .needs_iommu_device = false, .needs_pci_atomics = false, .num_sdma_engines = 2, + .num_sdma_queues_per_engine = 2, }; +static const struct kfd_device_info vega20_device_info = { + .asic_family = CHIP_VEGA20, + .max_pasid_bits = 16, + .max_no_of_hqd = 24, + .doorbell_size = 8, + .ih_ring_entry_size = 8 * sizeof(uint32_t), + .event_interrupt_class = &event_interrupt_class_v9, + .num_of_watch_points = 4, + .mqd_size_aligned = MQD_SIZE_ALIGNED, + .supports_cwsr = true, + .needs_iommu_device = false, + .needs_pci_atomics = false, + .num_sdma_engines = 2, + .num_sdma_queues_per_engine = 8, +}; struct kfd_deviceid { unsigned short did; @@ -293,7 +305,6 @@ static const struct kfd_deviceid supported_devices[] = { { 0x6928, &tonga_device_info }, /* Tonga */ { 0x6929, &tonga_device_info }, /* Tonga */ { 0x692B, &tonga_device_info }, /* Tonga */ - { 0x692F, &tonga_vf_device_info }, /* Tonga vf */ { 0x6938, &tonga_device_info }, /* Tonga */ { 0x6939, &tonga_device_info }, /* Tonga */ { 0x7300, &fiji_device_info }, /* Fiji */ @@ -328,6 +339,12 @@ static const struct kfd_deviceid supported_devices[] = { { 0x6868, &vega10_device_info }, /* Vega10 */ { 0x686C, &vega10_vf_device_info }, /* Vega10 vf*/ { 0x687F, &vega10_device_info }, /* Vega10 */ + { 0x66a0, &vega20_device_info }, /* Vega20 */ + { 0x66a1, &vega20_device_info }, /* Vega20 */ + { 0x66a2, &vega20_device_info }, /* Vega20 */ + { 0x66a3, &vega20_device_info }, /* Vega20 */ + { 0x66a7, &vega20_device_info }, /* Vega20 */ + { 0x66af, &vega20_device_info } /* Vega20 */ }; static int kfd_gtt_sa_init(struct kfd_dev *kfd, unsigned int buf_size, @@ -366,6 +383,10 @@ struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, return NULL; } + kfd = kzalloc(sizeof(*kfd), GFP_KERNEL); + if (!kfd) + return NULL; + /* Allow BIF to recode atomics to PCIe 3.0 AtomicOps. * 32 and 64-bit requests are possible and must be * supported. @@ -377,12 +398,10 @@ struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, dev_info(kfd_device, "skipped device %x:%x, PCI rejects atomics\n", pdev->vendor, pdev->device); + kfree(kfd); return NULL; - } - - kfd = kzalloc(sizeof(*kfd), GFP_KERNEL); - if (!kfd) - return NULL; + } else if (!ret) + kfd->pci_atomic_requested = true; kfd->kgd = kgd; kfd->device_info = device_info; @@ -419,6 +438,10 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd, { unsigned int size; + kfd->mec_fw_version = kfd->kfd2kgd->get_fw_version(kfd->kgd, + KGD_ENGINE_MEC1); + kfd->sdma_fw_version = kfd->kfd2kgd->get_fw_version(kfd->kgd, + KGD_ENGINE_SDMA1); kfd->shared_resources = *gpu_resources; kfd->vm_info.first_vmid_kfd = ffs(gpu_resources->compute_vmid_bitmap)-1; 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 ec0d62a16e53..a3b933967171 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c @@ -109,7 +109,7 @@ static unsigned int get_num_sdma_engines(struct device_queue_manager *dqm) unsigned int get_num_sdma_queues(struct device_queue_manager *dqm) { return dqm->dev->device_info->num_sdma_engines - * KFD_SDMA_QUEUES_PER_ENGINE; + * dqm->dev->device_info->num_sdma_queues_per_engine; } void program_sh_mem_settings(struct device_queue_manager *dqm, @@ -358,8 +358,8 @@ static int create_compute_queue_nocpsch(struct device_queue_manager *dqm, struct queue *q, struct qcm_process_device *qpd) { - int retval; struct mqd_manager *mqd_mgr; + int retval; mqd_mgr = dqm->ops.get_mqd_manager(dqm, KFD_MQD_TYPE_COMPUTE); if (!mqd_mgr) @@ -387,8 +387,12 @@ static int create_compute_queue_nocpsch(struct device_queue_manager *dqm, if (!q->properties.is_active) return 0; - retval = mqd_mgr->load_mqd(mqd_mgr, q->mqd, q->pipe, q->queue, - &q->properties, q->process->mm); + if (WARN(q->process->mm != current->mm, + "should only run in user thread")) + retval = -EFAULT; + else + retval = mqd_mgr->load_mqd(mqd_mgr, q->mqd, q->pipe, q->queue, + &q->properties, current->mm); if (retval) goto out_uninit_mqd; @@ -545,9 +549,15 @@ static int update_queue(struct device_queue_manager *dqm, struct queue *q) retval = map_queues_cpsch(dqm); else if (q->properties.is_active && (q->properties.type == KFD_QUEUE_TYPE_COMPUTE || - q->properties.type == KFD_QUEUE_TYPE_SDMA)) - retval = mqd_mgr->load_mqd(mqd_mgr, q->mqd, q->pipe, q->queue, - &q->properties, q->process->mm); + q->properties.type == KFD_QUEUE_TYPE_SDMA)) { + if (WARN(q->process->mm != current->mm, + "should only run in user thread")) + retval = -EFAULT; + else + retval = mqd_mgr->load_mqd(mqd_mgr, q->mqd, + q->pipe, q->queue, + &q->properties, current->mm); + } out_unlock: dqm_unlock(dqm); @@ -653,10 +663,11 @@ out: static int restore_process_queues_nocpsch(struct device_queue_manager *dqm, struct qcm_process_device *qpd) { + struct mm_struct *mm = NULL; struct queue *q; struct mqd_manager *mqd_mgr; struct kfd_process_device *pdd; - uint32_t pd_base; + uint64_t pd_base; int retval = 0; pdd = qpd_to_pdd(qpd); @@ -676,7 +687,7 @@ static int restore_process_queues_nocpsch(struct device_queue_manager *dqm, /* Update PD Base in QPD */ qpd->page_table_base = pd_base; - pr_debug("Updated PD address to 0x%08x\n", pd_base); + pr_debug("Updated PD address to 0x%llx\n", pd_base); if (!list_empty(&qpd->queues_list)) { dqm->dev->kfd2kgd->set_vm_context_page_table_base( @@ -686,6 +697,15 @@ static int restore_process_queues_nocpsch(struct device_queue_manager *dqm, kfd_flush_tlb(pdd); } + /* Take a safe reference to the mm_struct, which may otherwise + * disappear even while the kfd_process is still referenced. + */ + mm = get_task_mm(pdd->process->lead_thread); + if (!mm) { + retval = -EFAULT; + goto out; + } + /* activate all active queues on the qpd */ list_for_each_entry(q, &qpd->queues_list, list) { if (!q->properties.is_evicted) @@ -700,14 +720,15 @@ static int restore_process_queues_nocpsch(struct device_queue_manager *dqm, q->properties.is_evicted = false; q->properties.is_active = true; retval = mqd_mgr->load_mqd(mqd_mgr, q->mqd, q->pipe, - q->queue, &q->properties, - q->process->mm); + q->queue, &q->properties, mm); if (retval) goto out; dqm->queue_count++; } qpd->evicted = 0; out: + if (mm) + mmput(mm); dqm_unlock(dqm); return retval; } @@ -717,7 +738,7 @@ static int restore_process_queues_cpsch(struct device_queue_manager *dqm, { struct queue *q; struct kfd_process_device *pdd; - uint32_t pd_base; + uint64_t pd_base; int retval = 0; pdd = qpd_to_pdd(qpd); @@ -737,7 +758,7 @@ static int restore_process_queues_cpsch(struct device_queue_manager *dqm, /* Update PD Base in QPD */ qpd->page_table_base = pd_base; - pr_debug("Updated PD address to 0x%08x\n", pd_base); + pr_debug("Updated PD address to 0x%llx\n", pd_base); /* activate all active queues on the qpd */ list_for_each_entry(q, &qpd->queues_list, list) { @@ -761,7 +782,7 @@ static int register_process(struct device_queue_manager *dqm, { struct device_process_node *n; struct kfd_process_device *pdd; - uint32_t pd_base; + uint64_t pd_base; int retval; n = kzalloc(sizeof(*n), GFP_KERNEL); @@ -779,6 +800,7 @@ static int register_process(struct device_queue_manager *dqm, /* Update PD Base in QPD */ qpd->page_table_base = pd_base; + pr_debug("Updated PD address to 0x%llx\n", pd_base); retval = dqm->asic_ops.update_qpd(dqm, qpd); @@ -1342,9 +1364,6 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm, { int retval; struct mqd_manager *mqd_mgr; - bool preempt_all_queues; - - preempt_all_queues = false; retval = 0; @@ -1528,6 +1547,41 @@ static int process_termination_nocpsch(struct device_queue_manager *dqm, return retval; } +static int get_wave_state(struct device_queue_manager *dqm, + struct queue *q, + void __user *ctl_stack, + u32 *ctl_stack_used_size, + u32 *save_area_used_size) +{ + struct mqd_manager *mqd; + int r; + + dqm_lock(dqm); + + if (q->properties.type != KFD_QUEUE_TYPE_COMPUTE || + q->properties.is_active || !q->device->cwsr_enabled) { + r = -EINVAL; + goto dqm_unlock; + } + + mqd = dqm->ops.get_mqd_manager(dqm, KFD_MQD_TYPE_COMPUTE); + if (!mqd) { + r = -ENOMEM; + goto dqm_unlock; + } + + if (!mqd->get_wave_state) { + r = -EINVAL; + goto dqm_unlock; + } + + r = mqd->get_wave_state(mqd, q->mqd, ctl_stack, ctl_stack_used_size, + save_area_used_size); + +dqm_unlock: + dqm_unlock(dqm); + return r; +} static int process_termination_cpsch(struct device_queue_manager *dqm, struct qcm_process_device *qpd) @@ -1649,6 +1703,7 @@ struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev) dqm->ops.process_termination = process_termination_cpsch; dqm->ops.evict_process_queues = evict_process_queues_cpsch; dqm->ops.restore_process_queues = restore_process_queues_cpsch; + dqm->ops.get_wave_state = get_wave_state; break; case KFD_SCHED_POLICY_NO_HWS: /* initialize dqm for no cp scheduling */ @@ -1668,6 +1723,7 @@ struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev) dqm->ops.evict_process_queues = evict_process_queues_nocpsch; dqm->ops.restore_process_queues = restore_process_queues_nocpsch; + dqm->ops.get_wave_state = get_wave_state; break; default: pr_err("Invalid scheduling policy %d\n", dqm->sched_policy); @@ -1695,6 +1751,7 @@ struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev) break; case CHIP_VEGA10: + case CHIP_VEGA20: case CHIP_RAVEN: device_queue_manager_init_v9(&dqm->asic_ops); break; @@ -1806,7 +1863,9 @@ int dqm_debugfs_hqds(struct seq_file *m, void *data) } for (pipe = 0; pipe < get_num_sdma_engines(dqm); pipe++) { - for (queue = 0; queue < KFD_SDMA_QUEUES_PER_ENGINE; queue++) { + for (queue = 0; + queue < dqm->dev->device_info->num_sdma_queues_per_engine; + queue++) { r = dqm->dev->kfd2kgd->hqd_sdma_dump( dqm->dev->kgd, pipe, queue, &dump, &n_regs); if (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 00da3169a004..70e38a2e23b9 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h @@ -33,7 +33,6 @@ #define KFD_UNMAP_LATENCY_MS (4000) #define QUEUE_PREEMPT_DEFAULT_TIMEOUT_MS (2 * KFD_UNMAP_LATENCY_MS + 1000) -#define KFD_SDMA_QUEUES_PER_ENGINE (2) struct device_process_node { struct qcm_process_device *qpd; @@ -82,6 +81,8 @@ struct device_process_node { * * @restore_process_queues: Restore all evicted queues queues of a process * + * @get_wave_state: Retrieves context save state and optionally copies the + * control stack, if kept in the MQD, to the given userspace address. */ struct device_queue_manager_ops { @@ -137,6 +138,12 @@ struct device_queue_manager_ops { struct qcm_process_device *qpd); int (*restore_process_queues)(struct device_queue_manager *dqm, struct qcm_process_device *qpd); + + int (*get_wave_state)(struct device_queue_manager *dqm, + struct queue *q, + void __user *ctl_stack, + u32 *ctl_stack_used_size, + u32 *save_area_used_size); }; struct device_queue_manager_asic_ops { diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c index 97d5423c5673..3d66cec414af 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c @@ -400,6 +400,7 @@ int kfd_init_apertures(struct kfd_process *process) kfd_init_apertures_vi(pdd, id); break; case CHIP_VEGA10: + case CHIP_VEGA20: case CHIP_RAVEN: kfd_init_apertures_v9(pdd, id); break; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c index 9f84b4d9fb88..6c31f7370193 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c @@ -322,6 +322,7 @@ struct kernel_queue *kernel_queue_init(struct kfd_dev *dev, break; case CHIP_VEGA10: + case CHIP_VEGA20: case CHIP_RAVEN: kernel_queue_init_v9(&kq->ops_asic_specific); break; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue_v9.c index 684a3bf07efd..33830b1a5a54 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue_v9.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue_v9.c @@ -71,8 +71,7 @@ static int pm_map_process_v9(struct packet_manager *pm, uint32_t *buffer, struct qcm_process_device *qpd) { struct pm4_mes_map_process *packet; - uint64_t vm_page_table_base_addr = - (uint64_t)(qpd->page_table_base) << 12; + uint64_t vm_page_table_base_addr = qpd->page_table_base; packet = (struct pm4_mes_map_process *)buffer; memset(buffer, 0, sizeof(struct pm4_mes_map_process)); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c index 3bc25ab84f34..e33019a7a883 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c @@ -39,6 +39,7 @@ struct mqd_manager *mqd_manager_init(enum KFD_MQD_TYPE type, case CHIP_POLARIS11: return mqd_manager_init_vi_tonga(type, dev); case CHIP_VEGA10: + case CHIP_VEGA20: case CHIP_RAVEN: return mqd_manager_init_v9(type, dev); default: diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h index 4e84052d4e21..f8261313ae7b 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h @@ -43,6 +43,9 @@ * * @is_occupied: Checks if the relevant HQD slot is occupied. * + * @get_wave_state: Retrieves context save state and optionally copies the + * control stack, if kept in the MQD, to the given userspace address. + * * @mqd_mutex: Mqd manager mutex. * * @dev: The kfd device structure coupled with this module. @@ -85,6 +88,11 @@ struct mqd_manager { uint64_t queue_address, uint32_t pipe_id, uint32_t queue_id); + int (*get_wave_state)(struct mqd_manager *mm, void *mqd, + void __user *ctl_stack, + u32 *ctl_stack_used_size, + u32 *save_area_used_size); + #if defined(CONFIG_DEBUG_FS) int (*debugfs_show_mqd)(struct seq_file *m, void *data); #endif 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 0cedb37cf513..f381c1cb27bd 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c @@ -266,6 +266,28 @@ static bool is_occupied(struct mqd_manager *mm, void *mqd, pipe_id, queue_id); } +static int get_wave_state(struct mqd_manager *mm, void *mqd, + void __user *ctl_stack, + u32 *ctl_stack_used_size, + u32 *save_area_used_size) +{ + struct v9_mqd *m; + + /* Control stack is located one page after MQD. */ + void *mqd_ctl_stack = (void *)((uintptr_t)mqd + PAGE_SIZE); + + m = get_mqd(mqd); + + *ctl_stack_used_size = m->cp_hqd_cntl_stack_size - + m->cp_hqd_cntl_stack_offset; + *save_area_used_size = m->cp_hqd_wg_state_offset; + + if (copy_to_user(ctl_stack, mqd_ctl_stack, m->cp_hqd_cntl_stack_size)) + return -EFAULT; + + return 0; +} + static int init_mqd_hiq(struct mqd_manager *mm, void **mqd, struct kfd_mem_obj **mqd_mem_obj, uint64_t *gart_addr, struct queue_properties *q) @@ -435,6 +457,7 @@ struct mqd_manager *mqd_manager_init_v9(enum KFD_MQD_TYPE type, mqd->update_mqd = update_mqd; mqd->destroy_mqd = destroy_mqd; mqd->is_occupied = is_occupied; + mqd->get_wave_state = get_wave_state; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd; #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 b81fda3754da..6469b3456f00 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c @@ -269,6 +269,28 @@ static bool is_occupied(struct mqd_manager *mm, void *mqd, pipe_id, queue_id); } +static int get_wave_state(struct mqd_manager *mm, void *mqd, + void __user *ctl_stack, + u32 *ctl_stack_used_size, + u32 *save_area_used_size) +{ + struct vi_mqd *m; + + m = get_mqd(mqd); + + *ctl_stack_used_size = m->cp_hqd_cntl_stack_size - + m->cp_hqd_cntl_stack_offset; + *save_area_used_size = m->cp_hqd_wg_state_offset - + m->cp_hqd_cntl_stack_size; + + /* Control stack is not copied to user mode for GFXv8 because + * it's part of the context save area that is already + * accessible to user mode + */ + + return 0; +} + static int init_mqd_hiq(struct mqd_manager *mm, void **mqd, struct kfd_mem_obj **mqd_mem_obj, uint64_t *gart_addr, struct queue_properties *q) @@ -436,6 +458,7 @@ struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type, mqd->update_mqd = update_mqd; mqd->destroy_mqd = destroy_mqd; mqd->is_occupied = is_occupied; + mqd->get_wave_state = get_wave_state; #if defined(CONFIG_DEBUG_FS) mqd->debugfs_show_mqd = debugfs_show_mqd; #endif diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c index 1092631765cb..c6080ed3b6a7 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c @@ -229,6 +229,7 @@ int pm_init(struct packet_manager *pm, struct device_queue_manager *dqm) pm->pmf = &kfd_vi_pm_funcs; break; case CHIP_VEGA10: + case CHIP_VEGA20: case CHIP_RAVEN: pm->pmf = &kfd_v9_pm_funcs; break; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h index b0064b08aa11..53ff86d45d91 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h @@ -176,6 +176,7 @@ struct kfd_device_info { bool needs_iommu_device; bool needs_pci_atomics; unsigned int num_sdma_engines; + unsigned int num_sdma_queues_per_engine; }; struct kfd_mem_obj { @@ -247,6 +248,10 @@ struct kfd_dev { /* Debug manager */ struct kfd_dbgmgr *dbgmgr; + /* Firmware versions */ + uint16_t mec_fw_version; + uint16_t sdma_fw_version; + /* Maximum process number mapped to HW scheduler */ unsigned int max_proc_per_quantum; @@ -257,6 +262,8 @@ struct kfd_dev { /* xGMI */ uint64_t hive_id; + + bool pci_atomic_requested; }; /* KGD2KFD callbacks */ @@ -500,11 +507,11 @@ struct qcm_process_device { * All the memory management data should be here too */ uint64_t gds_context_area; + uint64_t page_table_base; uint32_t sh_mem_config; uint32_t sh_mem_bases; uint32_t sh_mem_ape1_base; uint32_t sh_mem_ape1_limit; - uint32_t page_table_base; uint32_t gds_size; uint32_t num_gws; uint32_t num_oac; @@ -856,6 +863,11 @@ int pqm_set_cu_mask(struct process_queue_manager *pqm, unsigned int qid, struct queue_properties *p); struct kernel_queue *pqm_get_kernel_queue(struct process_queue_manager *pqm, unsigned int qid); +int pqm_get_wave_state(struct process_queue_manager *pqm, + unsigned int qid, + void __user *ctl_stack, + u32 *ctl_stack_used_size, + u32 *save_area_used_size); int amdkfd_fence_wait_timeout(unsigned int *fence_addr, unsigned int fence_value, 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 c8cad9c078ae..fcaaf93681ac 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c @@ -408,6 +408,28 @@ struct kernel_queue *pqm_get_kernel_queue( return NULL; } +int pqm_get_wave_state(struct process_queue_manager *pqm, + unsigned int qid, + void __user *ctl_stack, + u32 *ctl_stack_used_size, + u32 *save_area_used_size) +{ + struct process_queue_node *pqn; + + pqn = get_queue_by_qid(pqm, qid); + if (!pqn) { + pr_debug("amdkfd: No queue %d exists for operation\n", + qid); + return -EFAULT; + } + + return pqn->q->device->dqm->ops.get_wave_state(pqn->q->device->dqm, + pqn->q, + ctl_stack, + ctl_stack_used_size, + save_area_used_size); +} + #if defined(CONFIG_DEBUG_FS) int pqm_debugfs_mqds(struct seq_file *m, void *data) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c index 0dff66be8d7a..e3843c5929ed 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c @@ -482,11 +482,11 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr, (unsigned long long int) 0); sysfs_show_32bit_prop(buffer, "fw_version", - dev->gpu->kfd2kgd->get_fw_version( - dev->gpu->kgd, - KGD_ENGINE_MEC1)); + dev->gpu->mec_fw_version); sysfs_show_32bit_prop(buffer, "capability", dev->node_props.capability); + sysfs_show_32bit_prop(buffer, "sdma_fw_version", + dev->gpu->sdma_fw_version); } return sysfs_show_32bit_prop(buffer, "max_engine_clk_ccompute", @@ -1127,17 +1127,40 @@ static void kfd_fill_mem_clk_max_info(struct kfd_topology_device *dev) static void kfd_fill_iolink_non_crat_info(struct kfd_topology_device *dev) { - struct kfd_iolink_properties *link; + struct kfd_iolink_properties *link, *cpu_link; + struct kfd_topology_device *cpu_dev; + uint32_t cap; + uint32_t cpu_flag = CRAT_IOLINK_FLAGS_ENABLED; + uint32_t flag = CRAT_IOLINK_FLAGS_ENABLED; if (!dev || !dev->gpu) return; - /* GPU only creates direck links so apply flags setting to all */ - if (dev->gpu->device_info->asic_family == CHIP_HAWAII) - list_for_each_entry(link, &dev->io_link_props, list) - link->flags = CRAT_IOLINK_FLAGS_ENABLED | - CRAT_IOLINK_FLAGS_NO_ATOMICS_32_BIT | - CRAT_IOLINK_FLAGS_NO_ATOMICS_64_BIT; + pcie_capability_read_dword(dev->gpu->pdev, + PCI_EXP_DEVCAP2, &cap); + + if (!(cap & (PCI_EXP_DEVCAP2_ATOMIC_COMP32 | + PCI_EXP_DEVCAP2_ATOMIC_COMP64))) + cpu_flag |= CRAT_IOLINK_FLAGS_NO_ATOMICS_32_BIT | + CRAT_IOLINK_FLAGS_NO_ATOMICS_64_BIT; + + if (!dev->gpu->pci_atomic_requested || + dev->gpu->device_info->asic_family == CHIP_HAWAII) + flag |= CRAT_IOLINK_FLAGS_NO_ATOMICS_32_BIT | + CRAT_IOLINK_FLAGS_NO_ATOMICS_64_BIT; + + /* GPU only creates direct links so apply flags setting to all */ + list_for_each_entry(link, &dev->io_link_props, list) { + link->flags = flag; + cpu_dev = kfd_topology_device_by_proximity_domain( + link->node_to); + if (cpu_dev) { + list_for_each_entry(cpu_link, + &cpu_dev->io_link_props, list) + if (cpu_link->node_to == link->node_from) + cpu_link->flags = cpu_flag; + } + } } int kfd_topology_add_device(struct kfd_dev *gpu) @@ -1255,6 +1278,7 @@ int kfd_topology_add_device(struct kfd_dev *gpu) HSA_CAP_DOORBELL_TYPE_TOTALBITS_MASK); break; case CHIP_VEGA10: + case CHIP_VEGA20: case CHIP_RAVEN: dev->node_props.capability |= ((HSA_CAP_DOORBELL_TYPE_2_0 << HSA_CAP_DOORBELL_TYPE_TOTALBITS_SHIFT) & 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 2be1e3033ce4..e5294d1a3049 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -338,14 +338,6 @@ static int dm_set_powergating_state(void *handle, /* Prototypes of private functions */ static int dm_early_init(void* handle); -static void hotplug_notify_work_func(struct work_struct *work) -{ - struct amdgpu_display_manager *dm = container_of(work, struct amdgpu_display_manager, mst_hotplug_work); - struct drm_device *dev = dm->ddev; - - drm_kms_helper_hotplug_event(dev); -} - /* Allocate memory for FBC compressed data */ static void amdgpu_dm_fbc_init(struct drm_connector *connector) { @@ -447,8 +439,6 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) goto error; } - INIT_WORK(&adev->dm.mst_hotplug_work, hotplug_notify_work_func); - adev->dm.freesync_module = mod_freesync_create(adev->dm.dc); if (!adev->dm.freesync_module) { DRM_ERROR( @@ -728,6 +718,87 @@ amdgpu_dm_find_first_crtc_matching_connector(struct drm_atomic_state *state, return NULL; } +static void emulated_link_detect(struct dc_link *link) +{ + struct dc_sink_init_data sink_init_data = { 0 }; + struct display_sink_capability sink_caps = { 0 }; + enum dc_edid_status edid_status; + struct dc_context *dc_ctx = link->ctx; + struct dc_sink *sink = NULL; + struct dc_sink *prev_sink = NULL; + + link->type = dc_connection_none; + prev_sink = link->local_sink; + + if (prev_sink != NULL) + dc_sink_retain(prev_sink); + + switch (link->connector_signal) { + case SIGNAL_TYPE_HDMI_TYPE_A: { + sink_caps.transaction_type = DDC_TRANSACTION_TYPE_I2C; + sink_caps.signal = SIGNAL_TYPE_HDMI_TYPE_A; + break; + } + + case SIGNAL_TYPE_DVI_SINGLE_LINK: { + sink_caps.transaction_type = DDC_TRANSACTION_TYPE_I2C; + sink_caps.signal = SIGNAL_TYPE_DVI_SINGLE_LINK; + break; + } + + case SIGNAL_TYPE_DVI_DUAL_LINK: { + sink_caps.transaction_type = DDC_TRANSACTION_TYPE_I2C; + sink_caps.signal = SIGNAL_TYPE_DVI_DUAL_LINK; + break; + } + + case SIGNAL_TYPE_LVDS: { + sink_caps.transaction_type = DDC_TRANSACTION_TYPE_I2C; + sink_caps.signal = SIGNAL_TYPE_LVDS; + break; + } + + case SIGNAL_TYPE_EDP: { + sink_caps.transaction_type = + DDC_TRANSACTION_TYPE_I2C_OVER_AUX; + sink_caps.signal = SIGNAL_TYPE_EDP; + break; + } + + case SIGNAL_TYPE_DISPLAY_PORT: { + sink_caps.transaction_type = + DDC_TRANSACTION_TYPE_I2C_OVER_AUX; + sink_caps.signal = SIGNAL_TYPE_VIRTUAL; + break; + } + + default: + DC_ERROR("Invalid connector type! signal:%d\n", + link->connector_signal); + return; + } + + sink_init_data.link = link; + sink_init_data.sink_signal = sink_caps.signal; + + sink = dc_sink_create(&sink_init_data); + if (!sink) { + DC_ERROR("Failed to create sink!\n"); + return; + } + + link->local_sink = sink; + + edid_status = dm_helpers_read_local_edid( + link->ctx, + link, + sink); + + if (edid_status != EDID_OK) + DC_ERROR("Failed to read EDID"); + +} + static int dm_resume(void *handle) { struct amdgpu_device *adev = handle; @@ -741,6 +812,7 @@ static int dm_resume(void *handle) struct drm_plane *plane; struct drm_plane_state *new_plane_state; struct dm_plane_state *dm_new_plane_state; + enum dc_connection_type new_connection_type = dc_connection_none; int ret; int i; @@ -771,7 +843,13 @@ static int dm_resume(void *handle) continue; mutex_lock(&aconnector->hpd_lock); - dc_link_detect(aconnector->dc_link, DETECT_REASON_HPD); + if (!dc_link_detect_sink(aconnector->dc_link, &new_connection_type)) + DRM_ERROR("KMS: Failed to detect connector\n"); + + if (aconnector->base.force && new_connection_type == dc_connection_none) + emulated_link_detect(aconnector->dc_link); + else + dc_link_detect(aconnector->dc_link, DETECT_REASON_HPD); if (aconnector->fake_enable && aconnector->dc_link->local_sink) aconnector->fake_enable = false; @@ -1020,6 +1098,7 @@ static void handle_hpd_irq(void *param) struct amdgpu_dm_connector *aconnector = (struct amdgpu_dm_connector *)param; struct drm_connector *connector = &aconnector->base; struct drm_device *dev = connector->dev; + enum dc_connection_type new_connection_type = dc_connection_none; /* * In case of failure or MST no need to update connector status or notify the OS @@ -1030,7 +1109,21 @@ static void handle_hpd_irq(void *param) if (aconnector->fake_enable) aconnector->fake_enable = false; - if (dc_link_detect(aconnector->dc_link, DETECT_REASON_HPD)) { + if (!dc_link_detect_sink(aconnector->dc_link, &new_connection_type)) + DRM_ERROR("KMS: Failed to detect connector\n"); + + if (aconnector->base.force && new_connection_type == dc_connection_none) { + emulated_link_detect(aconnector->dc_link); + + + drm_modeset_lock_all(dev); + dm_restore_drm_connector_state(dev, connector); + drm_modeset_unlock_all(dev); + + if (aconnector->base.force == DRM_FORCE_UNSPECIFIED) + drm_kms_helper_hotplug_event(dev); + + } else if (dc_link_detect(aconnector->dc_link, DETECT_REASON_HPD)) { amdgpu_dm_update_connector_after_detect(aconnector); @@ -1130,6 +1223,7 @@ static void handle_hpd_rx_irq(void *param) struct drm_device *dev = connector->dev; struct dc_link *dc_link = aconnector->dc_link; bool is_mst_root_connector = aconnector->mst_mgr.mst_state; + enum dc_connection_type new_connection_type = dc_connection_none; /* * TODO:Temporary add mutex to protect hpd interrupt not have a gpio @@ -1142,7 +1236,24 @@ static void handle_hpd_rx_irq(void *param) if (dc_link_handle_hpd_rx_irq(dc_link, NULL, NULL) && !is_mst_root_connector) { /* Downstream Port status changed. */ - if (dc_link_detect(dc_link, DETECT_REASON_HPDRX)) { + if (!dc_link_detect_sink(dc_link, &new_connection_type)) + DRM_ERROR("KMS: Failed to detect connector\n"); + + if (aconnector->base.force && new_connection_type == dc_connection_none) { + emulated_link_detect(dc_link); + + if (aconnector->fake_enable) + aconnector->fake_enable = false; + + amdgpu_dm_update_connector_after_detect(aconnector); + + + drm_modeset_lock_all(dev); + dm_restore_drm_connector_state(dev, connector); + drm_modeset_unlock_all(dev); + + drm_kms_helper_hotplug_event(dev); + } else if (dc_link_detect(dc_link, DETECT_REASON_HPDRX)) { if (aconnector->fake_enable) aconnector->fake_enable = false; @@ -1214,7 +1325,7 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev) struct dc_interrupt_params int_params = {0}; int r; int i; - unsigned client_id = AMDGPU_IH_CLIENTID_LEGACY; + unsigned client_id = AMDGPU_IRQ_CLIENTID_LEGACY; if (adev->asic_type == CHIP_VEGA10 || adev->asic_type == CHIP_VEGA12 || @@ -1539,6 +1650,7 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) struct amdgpu_mode_info *mode_info = &adev->mode_info; uint32_t link_cnt; int32_t total_overlay_planes, total_primary_planes; + enum dc_connection_type new_connection_type = dc_connection_none; link_cnt = dm->dc->caps.max_links; if (amdgpu_dm_mode_config_init(dm->adev)) { @@ -1605,7 +1717,14 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev) link = dc_get_link_at_index(dm->dc, i); - if (dc_link_detect(link, DETECT_REASON_BOOT)) { + if (!dc_link_detect_sink(link, &new_connection_type)) + DRM_ERROR("KMS: Failed to detect connector\n"); + + if (aconnector->base.force && new_connection_type == dc_connection_none) { + emulated_link_detect(link); + amdgpu_dm_update_connector_after_detect(aconnector); + + } else if (dc_link_detect(link, DETECT_REASON_BOOT)) { amdgpu_dm_update_connector_after_detect(aconnector); register_backlight_device(dm, link); } @@ -2648,7 +2767,7 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector, if (dm_state && dm_state->freesync_capable) stream->ignore_msa_timing_param = true; finish: - if (sink && sink->sink_signal == SIGNAL_TYPE_VIRTUAL) + if (sink && sink->sink_signal == SIGNAL_TYPE_VIRTUAL && aconnector->base.force != DRM_FORCE_ON) dc_sink_release(sink); return stream; @@ -4079,6 +4198,7 @@ static void amdgpu_dm_do_flip(struct drm_crtc *crtc, /* TODO eliminate or rename surface_update */ struct dc_surface_update surface_updates[1] = { {0} }; struct dm_crtc_state *acrtc_state = to_dm_crtc_state(crtc->state); + struct dc_stream_status *stream_status; /* Prepare wait for target vblank early - before the fence-waits */ @@ -4134,7 +4254,19 @@ static void amdgpu_dm_do_flip(struct drm_crtc *crtc, spin_unlock_irqrestore(&crtc->dev->event_lock, flags); - surface_updates->surface = dc_stream_get_status(acrtc_state->stream)->plane_states[0]; + stream_status = dc_stream_get_status(acrtc_state->stream); + if (!stream_status) { + DRM_ERROR("No stream status for CRTC: id=%d\n", + acrtc->crtc_id); + return; + } + + surface_updates->surface = stream_status->plane_states[0]; + if (!surface_updates->surface) { + DRM_ERROR("No surface for CRTC: id=%d\n", + acrtc->crtc_id); + return; + } surface_updates->flip_addr = &addr; dc_commit_updates_for_stream(adev->dm.dc, @@ -4608,12 +4740,18 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) } spin_unlock_irqrestore(&adev->ddev->event_lock, flags); - /* Signal HW programming completion */ - drm_atomic_helper_commit_hw_done(state); if (wait_for_vblank) drm_atomic_helper_wait_for_flip_done(dev, state); + /* + * FIXME: + * Delay hw_done() until flip_done() is signaled. This is to block + * another commit from freeing the CRTC state while we're still + * waiting on flip_done. + */ + drm_atomic_helper_commit_hw_done(state); + drm_atomic_helper_cleanup_planes(dev, state); /* @@ -4797,6 +4935,8 @@ void set_freesync_on_stream(struct amdgpu_display_manager *dm, mod_freesync_build_vrr_infopacket(dm->freesync_module, new_stream, &vrr, + packet_type_fs1, + NULL, &vrr_infopacket); new_crtc_state->adjust = vrr.adjust; 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 d4f1bdf93207..978b34a5011c 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h @@ -108,8 +108,6 @@ struct amdgpu_display_manager { const struct dc_link *backlight_link; - struct work_struct mst_hotplug_work; - struct mod_freesync *freesync_module; /** 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 0ef4a40d2247..9a7ac58eb18e 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 @@ -705,7 +705,8 @@ int connector_debugfs_init(struct amdgpu_dm_connector *connector) int i; struct dentry *ent, *dir = connector->base.debugfs_entry; - if (connector->base.connector_type == DRM_MODE_CONNECTOR_DisplayPort) { + if (connector->base.connector_type == DRM_MODE_CONNECTOR_DisplayPort || + connector->base.connector_type == DRM_MODE_CONNECTOR_eDP) { for (i = 0; i < ARRAY_SIZE(dp_debugfs_entries); i++) { ent = debugfs_create_file(dp_debugfs_entries[i].name, 0644, diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c index a910f01838ab..a212178f2edc 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c @@ -36,17 +36,13 @@ * Private declarations. *****************************************************************************/ -struct handler_common_data { +struct amdgpu_dm_irq_handler_data { struct list_head list; interrupt_handler handler; void *handler_arg; /* DM which this handler belongs to */ struct amdgpu_display_manager *dm; -}; - -struct amdgpu_dm_irq_handler_data { - struct handler_common_data hcd; /* DAL irq source which registered for this interrupt. */ enum dc_irq_source irq_source; }; @@ -61,7 +57,7 @@ struct amdgpu_dm_irq_handler_data { * Private functions. *****************************************************************************/ -static void init_handler_common_data(struct handler_common_data *hcd, +static void init_handler_common_data(struct amdgpu_dm_irq_handler_data *hcd, void (*ih)(void *), void *args, struct amdgpu_display_manager *dm) @@ -85,11 +81,9 @@ static void dm_irq_work_func(struct work_struct *work) struct amdgpu_dm_irq_handler_data *handler_data; list_for_each(entry, handler_list) { - handler_data = - list_entry( - entry, - struct amdgpu_dm_irq_handler_data, - hcd.list); + handler_data = list_entry(entry, + struct amdgpu_dm_irq_handler_data, + list); DRM_DEBUG_KMS("DM_IRQ: work_func: for dal_src=%d\n", handler_data->irq_source); @@ -97,7 +91,7 @@ static void dm_irq_work_func(struct work_struct *work) DRM_DEBUG_KMS("DM_IRQ: schedule_work: for dal_src=%d\n", handler_data->irq_source); - handler_data->hcd.handler(handler_data->hcd.handler_arg); + handler_data->handler(handler_data->handler_arg); } /* Call a DAL subcomponent which registered for interrupt notification @@ -137,11 +131,11 @@ static struct list_head *remove_irq_handler(struct amdgpu_device *adev, list_for_each_safe(entry, tmp, hnd_list) { handler = list_entry(entry, struct amdgpu_dm_irq_handler_data, - hcd.list); + list); if (ih == handler) { /* Found our handler. Remove it from the list. */ - list_del(&handler->hcd.list); + list_del(&handler->list); handler_removed = true; break; } @@ -230,8 +224,7 @@ void *amdgpu_dm_irq_register_interrupt(struct amdgpu_device *adev, memset(handler_data, 0, sizeof(*handler_data)); - init_handler_common_data(&handler_data->hcd, ih, handler_args, - &adev->dm); + init_handler_common_data(handler_data, ih, handler_args, &adev->dm); irq_source = int_params->irq_source; @@ -250,7 +243,7 @@ void *amdgpu_dm_irq_register_interrupt(struct amdgpu_device *adev, break; } - list_add_tail(&handler_data->hcd.list, hnd_list); + list_add_tail(&handler_data->list, hnd_list); DM_IRQ_TABLE_UNLOCK(adev, irq_table_flags); @@ -462,15 +455,13 @@ static void amdgpu_dm_irq_immediate_work(struct amdgpu_device *adev, entry, &adev->dm.irq_handler_list_high_tab[irq_source]) { - handler_data = - list_entry( - entry, - struct amdgpu_dm_irq_handler_data, - hcd.list); + handler_data = list_entry(entry, + struct amdgpu_dm_irq_handler_data, + list); /* Call a subcomponent which registered for immediate * interrupt notification */ - handler_data->hcd.handler(handler_data->hcd.handler_arg); + handler_data->handler(handler_data->handler_arg); } DM_IRQ_TABLE_UNLOCK(adev, irq_table_flags); diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c index 6d16b4a0353d..0fab64a2a915 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c @@ -105,6 +105,8 @@ bool dm_pp_apply_display_requirements( adev->powerplay.pp_funcs->display_configuration_change( adev->powerplay.pp_handle, &adev->pm.pm_display_cfg); + + amdgpu_pm_compute_clocks(adev); } return true; diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c index 5e2ea12fbb73..d0fc54f8fb1c 100644 --- a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c +++ b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c @@ -1625,11 +1625,11 @@ void dispclkdppclkdcfclk_deep_sleep_prefetch_parameters_watermarks_and_performan else { v->dsty_after_scaler = 0.0; } - v->v_update_offset_pix =dcn_bw_ceil2(v->htotal[k] / 4.0, 1.0); + v->v_update_offset_pix[k] = dcn_bw_ceil2(v->htotal[k] / 4.0, 1.0); v->total_repeater_delay_time = v->max_inter_dcn_tile_repeaters * (2.0 / v->dppclk + 3.0 / v->dispclk); - v->v_update_width_pix = (14.0 / v->dcf_clk_deep_sleep + 12.0 / v->dppclk + v->total_repeater_delay_time) * v->pixel_clock[k]; - v->v_ready_offset_pix =dcn_bw_max2(150.0 / v->dppclk, v->total_repeater_delay_time + 20.0 / v->dcf_clk_deep_sleep + 10.0 / v->dppclk) * v->pixel_clock[k]; - v->t_setup = (v->v_update_offset_pix + v->v_update_width_pix + v->v_ready_offset_pix) / v->pixel_clock[k]; + v->v_update_width_pix[k] = (14.0 / v->dcf_clk_deep_sleep + 12.0 / v->dppclk + v->total_repeater_delay_time) * v->pixel_clock[k]; + v->v_ready_offset_pix[k] = dcn_bw_max2(150.0 / v->dppclk, v->total_repeater_delay_time + 20.0 / v->dcf_clk_deep_sleep + 10.0 / v->dppclk) * v->pixel_clock[k]; + v->t_setup = (v->v_update_offset_pix[k] + v->v_update_width_pix[k] + v->v_ready_offset_pix[k]) / v->pixel_clock[k]; v->v_startup[k] =dcn_bw_min2(v->v_startup_lines, v->max_vstartup_lines[k]); if (v->prefetch_mode == 0.0) { v->t_wait =dcn_bw_max3(v->dram_clock_change_latency + v->urgent_latency, v->sr_enter_plus_exit_time, v->urgent_latency); diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c index 80ec09eef44f..3208188b7ed4 100644 --- a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +++ b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c @@ -1096,9 +1096,9 @@ bool dcn_validate_bandwidth( if (pipe->top_pipe && pipe->top_pipe->plane_state == pipe->plane_state) continue; - pipe->pipe_dlg_param.vupdate_width = v->v_update_width[input_idx][v->dpp_per_plane[input_idx] == 2 ? 1 : 0]; - pipe->pipe_dlg_param.vupdate_offset = v->v_update_offset[input_idx][v->dpp_per_plane[input_idx] == 2 ? 1 : 0]; - pipe->pipe_dlg_param.vready_offset = v->v_ready_offset[input_idx][v->dpp_per_plane[input_idx] == 2 ? 1 : 0]; + pipe->pipe_dlg_param.vupdate_width = v->v_update_width_pix[input_idx]; + pipe->pipe_dlg_param.vupdate_offset = v->v_update_offset_pix[input_idx]; + pipe->pipe_dlg_param.vready_offset = v->v_ready_offset_pix[input_idx]; pipe->pipe_dlg_param.vstartup_start = v->v_startup[input_idx]; pipe->pipe_dlg_param.htotal = pipe->stream->timing.h_total; @@ -1137,9 +1137,9 @@ bool dcn_validate_bandwidth( TIMING_3D_FORMAT_SIDE_BY_SIDE))) { if (hsplit_pipe && hsplit_pipe->plane_state == pipe->plane_state) { /* update previously split pipe */ - hsplit_pipe->pipe_dlg_param.vupdate_width = v->v_update_width[input_idx][v->dpp_per_plane[input_idx] == 2 ? 1 : 0]; - hsplit_pipe->pipe_dlg_param.vupdate_offset = v->v_update_offset[input_idx][v->dpp_per_plane[input_idx] == 2 ? 1 : 0]; - hsplit_pipe->pipe_dlg_param.vready_offset = v->v_ready_offset[input_idx][v->dpp_per_plane[input_idx] == 2 ? 1 : 0]; + hsplit_pipe->pipe_dlg_param.vupdate_width = v->v_update_width_pix[input_idx]; + hsplit_pipe->pipe_dlg_param.vupdate_offset = v->v_update_offset_pix[input_idx]; + hsplit_pipe->pipe_dlg_param.vready_offset = v->v_ready_offset_pix[input_idx]; hsplit_pipe->pipe_dlg_param.vstartup_start = v->v_startup[input_idx]; hsplit_pipe->pipe_dlg_param.htotal = pipe->stream->timing.h_total; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c index 1c438eedf77a..7c491c91465f 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c @@ -60,6 +60,7 @@ #define DC_LOGGER \ dc->ctx->logger +const static char DC_BUILD_ID[] = "production-build"; /******************************************************************************* * Private functions @@ -460,9 +461,25 @@ void dc_link_set_preferred_link_settings(struct dc *dc, struct dc_link_settings *link_setting, struct dc_link *link) { + int i; + struct pipe_ctx *pipe; + struct dc_stream_state *link_stream; struct dc_link_settings store_settings = *link_setting; - struct dc_stream_state *link_stream = - link->dc->current_state->res_ctx.pipe_ctx[0].stream; + + for (i = 0; i < MAX_PIPES; i++) { + pipe = &dc->current_state->res_ctx.pipe_ctx[i]; + if (pipe->stream && pipe->stream->sink + && pipe->stream->sink->link) { + if (pipe->stream->sink->link == link) + break; + } + } + + /* Stream not found */ + if (i == MAX_PIPES) + return; + + link_stream = link->dc->current_state->res_ctx.pipe_ctx[i].stream; link->preferred_link_setting = store_settings; if (link_stream) @@ -742,6 +759,8 @@ struct dc *dc_create(const struct dc_init_data *init_params) dc->config = init_params->flags; + dc->build_id = DC_BUILD_ID; + DC_LOG_DC("Display Core initialized\n"); @@ -1094,32 +1113,6 @@ static bool is_surface_in_context( return false; } -static unsigned int pixel_format_to_bpp(enum surface_pixel_format format) -{ - switch (format) { - case SURFACE_PIXEL_FORMAT_VIDEO_420_YCbCr: - case SURFACE_PIXEL_FORMAT_VIDEO_420_YCrCb: - return 12; - case SURFACE_PIXEL_FORMAT_GRPH_ARGB1555: - case SURFACE_PIXEL_FORMAT_GRPH_RGB565: - case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCbCr: - case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCrCb: - return 16; - case SURFACE_PIXEL_FORMAT_GRPH_ARGB8888: - case SURFACE_PIXEL_FORMAT_GRPH_ABGR8888: - case SURFACE_PIXEL_FORMAT_GRPH_ARGB2101010: - case SURFACE_PIXEL_FORMAT_GRPH_ABGR2101010: - return 32; - case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616: - case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616F: - case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F: - return 64; - default: - ASSERT_CRITICAL(false); - return -1; - } -} - static enum surface_update_type get_plane_info_update_type(const struct dc_surface_update *u) { union surface_update_flags *update_flags = &u->surface->update_flags; @@ -1153,16 +1146,13 @@ static enum surface_update_type get_plane_info_update_type(const struct dc_surfa || u->plane_info->dcc.grph.meta_pitch != u->surface->dcc.grph.meta_pitch) update_flags->bits.dcc_change = 1; - if (pixel_format_to_bpp(u->plane_info->format) != - pixel_format_to_bpp(u->surface->format)) + if (resource_pixel_format_to_bpp(u->plane_info->format) != + resource_pixel_format_to_bpp(u->surface->format)) /* different bytes per element will require full bandwidth * and DML calculation */ update_flags->bits.bpp_change = 1; - if (u->gamma && dce_use_lut(u->plane_info->format)) - update_flags->bits.gamma_change = 1; - if (memcmp(&u->plane_info->tiling_info, &u->surface->tiling_info, sizeof(union dc_tiling_info)) != 0) { update_flags->bits.swizzle_change = 1; @@ -1179,7 +1169,6 @@ static enum surface_update_type get_plane_info_update_type(const struct dc_surfa if (update_flags->bits.rotation_change || update_flags->bits.stereo_format_change || update_flags->bits.pixel_format_change - || update_flags->bits.gamma_change || update_flags->bits.bpp_change || update_flags->bits.bandwidth_change || update_flags->bits.output_tf_change) @@ -1269,13 +1258,26 @@ static enum surface_update_type det_surface_update(const struct dc *dc, if (u->coeff_reduction_factor) update_flags->bits.coeff_reduction_change = 1; + if (u->gamma) { + enum surface_pixel_format format = SURFACE_PIXEL_FORMAT_GRPH_BEGIN; + + if (u->plane_info) + format = u->plane_info->format; + else if (u->surface) + format = u->surface->format; + + if (dce_use_lut(format)) + update_flags->bits.gamma_change = 1; + } + if (update_flags->bits.in_transfer_func_change) { type = UPDATE_TYPE_MED; elevate_update_type(&overall_type, type); } if (update_flags->bits.input_csc_change - || update_flags->bits.coeff_reduction_change) { + || update_flags->bits.coeff_reduction_change + || update_flags->bits.gamma_change) { type = UPDATE_TYPE_FULL; elevate_update_type(&overall_type, type); } @@ -1379,7 +1381,7 @@ static void notify_display_count_to_smu( * sent as part of pplib_apply_display_requirements. * So just return. */ - if (!pp_smu->set_display_count) + if (!pp_smu || !pp_smu->set_display_count) return; display_count = 0; @@ -1834,3 +1836,16 @@ void dc_link_remove_remote_sink(struct dc_link *link, struct dc_sink *sink) } } } + +void get_clock_requirements_for_state(struct dc_state *state, struct AsicStateEx *info) +{ + info->displayClock = (unsigned int)state->bw.dcn.clk.dispclk_khz; + info->engineClock = (unsigned int)state->bw.dcn.clk.dcfclk_khz; + info->memoryClock = (unsigned int)state->bw.dcn.clk.dramclk_khz; + info->maxSupportedDppClock = (unsigned int)state->bw.dcn.clk.max_supported_dppclk_khz; + info->dppClock = (unsigned int)state->bw.dcn.clk.dppclk_khz; + info->socClock = (unsigned int)state->bw.dcn.clk.socclk_khz; + info->dcfClockDeepSleep = (unsigned int)state->bw.dcn.clk.dcfclk_deep_sleep_khz; + info->fClock = (unsigned int)state->bw.dcn.clk.fclk_khz; + info->phyClock = (unsigned int)state->bw.dcn.clk.phyclk_khz; +}
\ No newline at end of file diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link.c b/drivers/gpu/drm/amd/display/dc/core/dc_link.c index bd58dbae7d3e..fb04a4ad141f 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_link.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_link.c @@ -198,7 +198,7 @@ static bool program_hpd_filter( return result; } -static bool detect_sink(struct dc_link *link, enum dc_connection_type *type) +bool dc_link_detect_sink(struct dc_link *link, enum dc_connection_type *type) { uint32_t is_hpd_high = 0; struct gpio *hpd_pin; @@ -612,7 +612,7 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason) if (link->connector_signal == SIGNAL_TYPE_VIRTUAL) return false; - if (false == detect_sink(link, &new_connection_type)) { + if (false == dc_link_detect_sink(link, &new_connection_type)) { BREAK_TO_DEBUGGER(); return false; } @@ -2559,23 +2559,24 @@ void core_link_enable_stream( pipe_ctx->stream_res.stream_enc, &stream->timing); - resource_build_info_frame(pipe_ctx); - core_dc->hwss.update_info_frame(pipe_ctx); + if (!IS_FPGA_MAXIMUS_DC(core_dc->ctx->dce_environment)) { + resource_build_info_frame(pipe_ctx); + core_dc->hwss.update_info_frame(pipe_ctx); - /* eDP lit up by bios already, no need to enable again. */ - if (pipe_ctx->stream->signal == SIGNAL_TYPE_EDP && - pipe_ctx->stream->apply_edp_fast_boot_optimization) { - pipe_ctx->stream->apply_edp_fast_boot_optimization = false; - pipe_ctx->stream->dpms_off = false; - return; - } + /* eDP lit up by bios already, no need to enable again. */ + if (pipe_ctx->stream->signal == SIGNAL_TYPE_EDP && + pipe_ctx->stream->apply_edp_fast_boot_optimization) { + pipe_ctx->stream->apply_edp_fast_boot_optimization = false; + pipe_ctx->stream->dpms_off = false; + return; + } - if (pipe_ctx->stream->dpms_off) - return; + if (pipe_ctx->stream->dpms_off) + return; - status = enable_link(state, pipe_ctx); + status = enable_link(state, pipe_ctx); - if (status != DC_OK) { + if (status != DC_OK) { DC_LOG_WARNING("enabling link %u failed: %d\n", pipe_ctx->stream->sink->link->link_index, status); @@ -2590,23 +2591,26 @@ void core_link_enable_stream( BREAK_TO_DEBUGGER(); return; } - } + } - core_dc->hwss.enable_audio_stream(pipe_ctx); + core_dc->hwss.enable_audio_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); + /* 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); - core_dc->hwss.enable_stream(pipe_ctx); + core_dc->hwss.enable_stream(pipe_ctx); - if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) - allocate_mst_payload(pipe_ctx); + if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) + allocate_mst_payload(pipe_ctx); + + core_dc->hwss.unblank_stream(pipe_ctx, + &pipe_ctx->stream->sink->link->cur_link_settings); + + } - core_dc->hwss.unblank_stream(pipe_ctx, - &pipe_ctx->stream->sink->link->cur_link_settings); } void core_link_disable_stream(struct pipe_ctx *pipe_ctx, int option) 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 2d6a4300bfa4..b6fe29b9fb65 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c @@ -1975,6 +1975,9 @@ static void calculate_phy_pix_clks(struct dc_stream_state *stream) else stream->phy_pix_clk = stream->timing.pix_clk_khz; + + if (stream->timing.timing_3d_format == TIMING_3D_FORMAT_HW_FRAME_PACKING) + stream->phy_pix_clk *= 2; } enum dc_status resource_map_pool_resources( @@ -2096,6 +2099,14 @@ enum dc_status dc_validate_global_state( if (pipe_ctx->stream != stream) continue; + if (dc->res_pool->funcs->get_default_swizzle_mode && + pipe_ctx->plane_state && + pipe_ctx->plane_state->tiling_info.gfx9.swizzle == DC_SW_UNKNOWN) { + result = dc->res_pool->funcs->get_default_swizzle_mode(pipe_ctx->plane_state); + if (result != DC_OK) + return result; + } + /* Switch to dp clock source only if there is * no non dp stream that shares the same timing * with the dp stream. @@ -2885,3 +2896,32 @@ enum dc_status dc_validate_plane(struct dc *dc, const struct dc_plane_state *pla return res; } + +unsigned int resource_pixel_format_to_bpp(enum surface_pixel_format format) +{ + switch (format) { + case SURFACE_PIXEL_FORMAT_GRPH_PALETA_256_COLORS: + return 8; + case SURFACE_PIXEL_FORMAT_VIDEO_420_YCbCr: + case SURFACE_PIXEL_FORMAT_VIDEO_420_YCrCb: + return 12; + case SURFACE_PIXEL_FORMAT_GRPH_ARGB1555: + case SURFACE_PIXEL_FORMAT_GRPH_RGB565: + case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCbCr: + case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCrCb: + return 16; + case SURFACE_PIXEL_FORMAT_GRPH_ARGB8888: + case SURFACE_PIXEL_FORMAT_GRPH_ABGR8888: + case SURFACE_PIXEL_FORMAT_GRPH_ARGB2101010: + case SURFACE_PIXEL_FORMAT_GRPH_ABGR2101010: + case SURFACE_PIXEL_FORMAT_GRPH_ABGR2101010_XR_BIAS: + return 32; + case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616: + case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616F: + case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F: + return 64; + default: + ASSERT_CRITICAL(false); + return -1; + } +} diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h index 7691139363a9..199527171100 100644 --- a/drivers/gpu/drm/amd/display/dc/dc.h +++ b/drivers/gpu/drm/amd/display/dc/dc.h @@ -38,13 +38,12 @@ #include "inc/compressor.h" #include "dml/display_mode_lib.h" -#define DC_VER "3.1.66" +#define DC_VER "3.1.68" #define MAX_SURFACES 3 #define MAX_STREAMS 6 #define MAX_SINKS_PER_LINK 4 - /******************************************************************************* * Display Core Interfaces ******************************************************************************/ @@ -208,6 +207,7 @@ struct dc_clocks { int dcfclk_deep_sleep_khz; int fclk_khz; int phyclk_khz; + int dramclk_khz; }; struct dc_debug_options { @@ -315,6 +315,8 @@ struct dc { struct compressor *fbc_compressor; struct dc_debug_data debug_data; + + const char *build_id; }; enum frame_buffer_mode { @@ -599,6 +601,8 @@ struct dc_validation_set { enum dc_status dc_validate_plane(struct dc *dc, const struct dc_plane_state *plane_state); +void get_clock_requirements_for_state(struct dc_state *state, struct AsicStateEx *info); + enum dc_status dc_validate_global_state( struct dc *dc, struct dc_state *new_ctx); diff --git a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h index 57f57cf0fe2a..7825e4b5e97c 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h @@ -289,7 +289,8 @@ enum swizzle_mode_values { DC_SW_VAR_S_X = 29, DC_SW_VAR_D_X = 30, DC_SW_VAR_R_X = 31, - DC_SW_MAX + DC_SW_MAX = 32, + DC_SW_UNKNOWN = DC_SW_MAX }; union dc_tiling_info { diff --git a/drivers/gpu/drm/amd/display/dc/dc_link.h b/drivers/gpu/drm/amd/display/dc/dc_link.h index 438fb35d87b8..3bfdccceb524 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_link.h +++ b/drivers/gpu/drm/amd/display/dc/dc_link.h @@ -216,6 +216,7 @@ void dc_link_enable_hpd_filter(struct dc_link *link, bool enable); bool dc_link_is_dp_sink_present(struct dc_link *link); +bool dc_link_detect_sink(struct dc_link *link, enum dc_connection_type *type); /* * DPCD access interfaces */ diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h index 4fb62780a696..6e12d640d020 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_types.h @@ -659,4 +659,16 @@ enum i2c_mot_mode { I2C_MOT_FALSE }; +struct AsicStateEx { + unsigned int memoryClock; + unsigned int displayClock; + unsigned int engineClock; + unsigned int maxSupportedDppClock; + unsigned int dppClock; + unsigned int socClock; + unsigned int dcfClockDeepSleep; + unsigned int fClock; + unsigned int phyClock; +}; + #endif /* DC_TYPES_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c index 3f5b2e6f7553..aaeb7faac0c4 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c @@ -312,7 +312,7 @@ static void process_channel_reply( /* in case HPD is LOW, exit AUX transaction */ if ((sw_status & AUX_SW_STATUS__AUX_SW_HPD_DISCON_MASK)) { - reply->status = AUX_CHANNEL_OPERATION_FAILED_HPD_DISCON; + reply->status = AUX_TRANSACTION_REPLY_HPD_DISCON; return; } diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clocks.c b/drivers/gpu/drm/amd/display/dc/dce/dce_clocks.c index bf6261a1584b..d89a097ba936 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_clocks.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clocks.c @@ -468,6 +468,9 @@ static void dce12_update_clocks(struct dccg *dccg, { struct dm_pp_clock_for_voltage_req clock_voltage_req = {0}; + /* TODO: Investigate why this is needed to fix display corruption. */ + new_clocks->dispclk_khz = new_clocks->dispclk_khz * 115 / 100; + if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, dccg->clks.dispclk_khz)) { clock_voltage_req.clk_type = DM_PP_CLOCK_TYPE_DISPLAY_CLK; clock_voltage_req.clocks_in_khz = new_clocks->dispclk_khz; @@ -661,6 +664,11 @@ static void dce_update_clocks(struct dccg *dccg, bool safe_to_lower) { struct dm_pp_power_level_change_request level_change_req; + struct dce_dccg *clk_dce = TO_DCE_CLOCKS(dccg); + + /* TODO: Investigate why this is needed to fix display corruption. */ + if (!clk_dce->dfs_bypass_active) + new_clocks->dispclk_khz = new_clocks->dispclk_khz * 115 / 100; level_change_req.power_level = dce_get_required_clocks_state(dccg, new_clocks); /* get max clock state from PPLIB */ diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c index 4942590e8b9c..366bc8c2c643 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c @@ -662,21 +662,10 @@ bool dce110_link_encoder_validate_dp_output( const struct dce110_link_encoder *enc110, const struct dc_crtc_timing *crtc_timing) { - /* default RGB only */ - if (crtc_timing->pixel_encoding == PIXEL_ENCODING_RGB) - return true; - - if (enc110->base.features.flags.bits.IS_YCBCR_CAPABLE) - return true; - - /* for DCE 8.x or later DP Y-only feature, - * we need ASIC cap + FeatureSupportDPYonly, not support 666 */ - if (crtc_timing->flags.Y_ONLY && - enc110->base.features.flags.bits.IS_YCBCR_CAPABLE && - crtc_timing->display_color_depth != COLOR_DEPTH_666) - return true; + if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR420) + return false; - return false; + return true; } void dce110_link_encoder_construct( 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 b1cc38827f09..14754a87156c 100644 --- a/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c @@ -551,8 +551,7 @@ static const struct encoder_feature_support link_enc_feature = { .max_hdmi_deep_color = COLOR_DEPTH_121212, .max_hdmi_pixel_clock = 300000, .flags.bits.IS_HBR2_CAPABLE = true, - .flags.bits.IS_TPS3_CAPABLE = true, - .flags.bits.IS_YCBCR_CAPABLE = true + .flags.bits.IS_TPS3_CAPABLE = true }; struct link_encoder *dce100_link_encoder_create( @@ -690,7 +689,9 @@ static void destruct(struct dce110_resource_pool *pool) kfree(DCE110TG_FROM_TG(pool->base.timing_generators[i])); pool->base.timing_generators[i] = NULL; } + } + for (i = 0; i < pool->base.res_cap->num_ddc; i++) { if (pool->base.engines[i] != NULL) dce110_engine_destroy(&pool->base.engines[i]); if (pool->base.hw_i2cs[i] != NULL) { 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 dc1eed5ba996..b75ede5f84f7 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 @@ -1377,26 +1377,13 @@ static enum dc_status apply_single_controller_ctx_to_hw( /* */ dc->hwss.enable_stream_timing(pipe_ctx, context, dc); - /* FPGA does not program backend */ - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - pipe_ctx->stream_res.opp->funcs->opp_set_dyn_expansion( - pipe_ctx->stream_res.opp, - COLOR_SPACE_YCBCR601, - stream->timing.display_color_depth, - pipe_ctx->stream->signal); - - pipe_ctx->stream_res.opp->funcs->opp_program_fmt( - pipe_ctx->stream_res.opp, - &stream->bit_depth_params, - &stream->clamping); - return DC_OK; - } /* TODO: move to stream encoder */ if (pipe_ctx->stream->signal != SIGNAL_TYPE_VIRTUAL) if (DC_OK != bios_parser_crtc_source_select(pipe_ctx)) { BREAK_TO_DEBUGGER(); return DC_ERROR_UNEXPECTED; } + pipe_ctx->stream_res.opp->funcs->opp_set_dyn_expansion( pipe_ctx->stream_res.opp, COLOR_SPACE_YCBCR601, @@ -2550,7 +2537,7 @@ static void pplib_apply_display_requirements( dc->prev_display_config = *pp_display_cfg; } -void dce110_set_bandwidth( +static void dce110_set_bandwidth( struct dc *dc, struct dc_state *context, bool decrease_allowed) diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.h index e4c5db75c4c6..d6db3dbd9015 100644 --- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.h +++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.h @@ -68,11 +68,6 @@ void dce110_fill_display_configs( const struct dc_state *context, struct dm_pp_display_configuration *pp_display_cfg); -void dce110_set_bandwidth( - struct dc *dc, - struct dc_state *context, - bool decrease_allowed); - uint32_t dce110_get_min_vblank_time_us(const struct dc_state *context); void dp_receiver_power_ctrl(struct dc_link *link, bool on); 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 b44cc7042249..de190935f0a4 100644 --- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c @@ -570,8 +570,7 @@ static const struct encoder_feature_support link_enc_feature = { .max_hdmi_deep_color = COLOR_DEPTH_121212, .max_hdmi_pixel_clock = 594000, .flags.bits.IS_HBR2_CAPABLE = true, - .flags.bits.IS_TPS3_CAPABLE = true, - .flags.bits.IS_YCBCR_CAPABLE = true + .flags.bits.IS_TPS3_CAPABLE = true }; static struct link_encoder *dce110_link_encoder_create( @@ -720,7 +719,9 @@ static void destruct(struct dce110_resource_pool *pool) kfree(DCE110TG_FROM_TG(pool->base.timing_generators[i])); pool->base.timing_generators[i] = NULL; } + } + for (i = 0; i < pool->base.res_cap->num_ddc; i++) { if (pool->base.engines[i] != NULL) dce110_engine_destroy(&pool->base.engines[i]); if (pool->base.hw_i2cs[i] != NULL) { 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 0f8332ea1160..3ce79c208ddf 100644 --- a/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c @@ -555,8 +555,7 @@ static const struct encoder_feature_support link_enc_feature = { .flags.bits.IS_HBR2_CAPABLE = true, .flags.bits.IS_HBR3_CAPABLE = true, .flags.bits.IS_TPS3_CAPABLE = true, - .flags.bits.IS_TPS4_CAPABLE = true, - .flags.bits.IS_YCBCR_CAPABLE = true + .flags.bits.IS_TPS4_CAPABLE = true }; struct link_encoder *dce112_link_encoder_create( @@ -694,9 +693,6 @@ static void destruct(struct dce110_resource_pool *pool) if (pool->base.opps[i] != NULL) dce110_opp_destroy(&pool->base.opps[i]); - if (pool->base.engines[i] != NULL) - dce110_engine_destroy(&pool->base.engines[i]); - if (pool->base.transforms[i] != NULL) dce112_transform_destroy(&pool->base.transforms[i]); @@ -712,6 +708,11 @@ static void destruct(struct dce110_resource_pool *pool) kfree(DCE110TG_FROM_TG(pool->base.timing_generators[i])); pool->base.timing_generators[i] = NULL; } + } + + for (i = 0; i < pool->base.res_cap->num_ddc; i++) { + if (pool->base.engines[i] != NULL) + dce110_engine_destroy(&pool->base.engines[i]); if (pool->base.hw_i2cs[i] != NULL) { kfree(pool->base.hw_i2cs[i]); pool->base.hw_i2cs[i] = NULL; 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 5853522a6182..eb0f5f9a973b 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 @@ -244,17 +244,6 @@ static void dce120_update_dchub( dh_data->dchub_info_valid = false; } -static void dce120_set_bandwidth( - struct dc *dc, - struct dc_state *context, - bool decrease_allowed) -{ - if (context->stream_count <= 0) - return; - - dce110_set_bandwidth(dc, context, decrease_allowed); -} - void dce120_hw_sequencer_construct(struct dc *dc) { /* All registers used by dce11.2 match those in dce11 in offset and @@ -263,6 +252,5 @@ void dce120_hw_sequencer_construct(struct dc *dc) dce110_hw_sequencer_construct(dc); dc->hwss.enable_display_power_gating = dce120_enable_display_power_gating; dc->hwss.update_dchub = dce120_update_dchub; - dc->hwss.set_bandwidth = dce120_set_bandwidth; } 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 59055801af44..79ab5f9f9115 100644 --- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c @@ -533,7 +533,9 @@ static void destruct(struct dce110_resource_pool *pool) kfree(DCE110TG_FROM_TG(pool->base.timing_generators[i])); pool->base.timing_generators[i] = NULL; } + } + for (i = 0; i < pool->base.res_cap->num_ddc; i++) { if (pool->base.engines[i] != NULL) dce110_engine_destroy(&pool->base.engines[i]); if (pool->base.hw_i2cs[i] != NULL) { @@ -609,7 +611,6 @@ static const struct encoder_feature_support link_enc_feature = { .flags.bits.IS_HBR3_CAPABLE = true, .flags.bits.IS_TPS3_CAPABLE = true, .flags.bits.IS_TPS4_CAPABLE = true, - .flags.bits.IS_YCBCR_CAPABLE = true }; static struct link_encoder *dce120_link_encoder_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 1dc590ccc5f9..d68f951f9869 100644 --- a/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c @@ -650,8 +650,7 @@ static const struct encoder_feature_support link_enc_feature = { .max_hdmi_deep_color = COLOR_DEPTH_121212, .max_hdmi_pixel_clock = 297000, .flags.bits.IS_HBR2_CAPABLE = true, - .flags.bits.IS_TPS3_CAPABLE = true, - .flags.bits.IS_YCBCR_CAPABLE = true + .flags.bits.IS_TPS3_CAPABLE = true }; struct link_encoder *dce80_link_encoder_create( @@ -739,7 +738,9 @@ static void destruct(struct dce110_resource_pool *pool) kfree(DCE110TG_FROM_TG(pool->base.timing_generators[i])); pool->base.timing_generators[i] = NULL; } + } + for (i = 0; i < pool->base.res_cap->num_ddc; i++) { if (pool->base.engines[i] != NULL) dce110_engine_destroy(&pool->base.engines[i]); if (pool->base.hw_i2cs[i] != NULL) { diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c index 1ea91e153d3a..4254e7e1a509 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c @@ -87,6 +87,23 @@ void hubbub1_wm_read_state(struct hubbub *hubbub, s->dram_clk_chanage = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_D); } +void hubbub1_disable_allow_self_refresh(struct hubbub *hubbub) +{ + REG_UPDATE(DCHUBBUB_ARB_DRAM_STATE_CNTL, + DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE, 0); +} + +bool hububu1_is_allow_self_refresh_enabled(struct hubbub *hubbub) +{ + uint32_t enable = 0; + + REG_GET(DCHUBBUB_ARB_DRAM_STATE_CNTL, + DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE, &enable); + + return true ? false : enable; +} + + bool hubbub1_verify_allow_pstate_change_high( struct hubbub *hubbub) { @@ -116,7 +133,43 @@ bool hubbub1_verify_allow_pstate_change_high( forced_pstate_allow = false; } - /* RV1: + /* RV2: + * dchubbubdebugind, at: 0xB + * description + * 0: Pipe0 Plane0 Allow Pstate Change + * 1: Pipe0 Plane1 Allow Pstate Change + * 2: Pipe0 Cursor0 Allow Pstate Change + * 3: Pipe0 Cursor1 Allow Pstate Change + * 4: Pipe1 Plane0 Allow Pstate Change + * 5: Pipe1 Plane1 Allow Pstate Change + * 6: Pipe1 Cursor0 Allow Pstate Change + * 7: Pipe1 Cursor1 Allow Pstate Change + * 8: Pipe2 Plane0 Allow Pstate Change + * 9: Pipe2 Plane1 Allow Pstate Change + * 10: Pipe2 Cursor0 Allow Pstate Change + * 11: Pipe2 Cursor1 Allow Pstate Change + * 12: Pipe3 Plane0 Allow Pstate Change + * 13: Pipe3 Plane1 Allow Pstate Change + * 14: Pipe3 Cursor0 Allow Pstate Change + * 15: Pipe3 Cursor1 Allow Pstate Change + * 16: Pipe4 Plane0 Allow Pstate Change + * 17: Pipe4 Plane1 Allow Pstate Change + * 18: Pipe4 Cursor0 Allow Pstate Change + * 19: Pipe4 Cursor1 Allow Pstate Change + * 20: Pipe5 Plane0 Allow Pstate Change + * 21: Pipe5 Plane1 Allow Pstate Change + * 22: Pipe5 Cursor0 Allow Pstate Change + * 23: Pipe5 Cursor1 Allow Pstate Change + * 24: Pipe6 Plane0 Allow Pstate Change + * 25: Pipe6 Plane1 Allow Pstate Change + * 26: Pipe6 Cursor0 Allow Pstate Change + * 27: Pipe6 Cursor1 Allow Pstate Change + * 28: WB0 Allow Pstate Change + * 29: WB1 Allow Pstate Change + * 30: Arbiter's allow_pstate_change + * 31: SOC pstate change request" + * + * RV1: * dchubbubdebugind, at: 0x7 * description "3-0: Pipe0 cursor0 QOS * 7-4: Pipe1 cursor0 QOS @@ -140,7 +193,6 @@ bool hubbub1_verify_allow_pstate_change_high( * 31: SOC pstate change request */ - REG_WRITE(DCHUBBUB_TEST_DEBUG_INDEX, hubbub->debug_test_index_pstate); for (i = 0; i < pstate_wait_timeout_us; i++) { @@ -802,5 +854,9 @@ void hubbub1_construct(struct hubbub *hubbub, hubbub->masks = hubbub_mask; hubbub->debug_test_index_pstate = 0x7; +#if defined(CONFIG_DRM_AMD_DC_DCN1_01) + if (ctx->dce_version == DCN_VERSION_1_01) + hubbub->debug_test_index_pstate = 0xB; +#endif } diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.h index d6e596eef4c5..d0f03d152913 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.h +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.h @@ -203,6 +203,10 @@ void hubbub1_program_watermarks( unsigned int refclk_mhz, bool safe_to_lower); +void hubbub1_disable_allow_self_refresh(struct hubbub *hubbub); + +bool hububu1_is_allow_self_refresh_enabled(struct hubbub *hubub); + void hubbub1_toggle_watermark_change_req( struct hubbub *hubbub); 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 6bd4ec39f869..193184affefb 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 @@ -44,6 +44,7 @@ #include "dcn10_hubp.h" #include "dcn10_hubbub.h" #include "dcn10_cm_common.h" +#include "dc_link_dp.h" #define DC_LOGGER_INIT(logger) @@ -996,7 +997,21 @@ static void dcn10_init_hw(struct dc *dc) } else { if (!dcb->funcs->is_accelerated_mode(dcb)) { + bool allow_self_fresh_force_enable = + hububu1_is_allow_self_refresh_enabled(dc->res_pool->hubbub); + bios_golden_init(dc); + + /* WA for making DF sleep when idle after resume from S0i3. + * DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE is set to 1 by + * command table, if DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE = 0 + * before calling command table and it changed to 1 after, + * it should be set back to 0. + */ + if (allow_self_fresh_force_enable == false && + hububu1_is_allow_self_refresh_enabled(dc->res_pool->hubbub)) + hubbub1_disable_allow_self_refresh(dc->res_pool->hubbub); + disable_vga(dc->hwseq); } diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c index 6f675206a136..ba6a8686062f 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c @@ -606,22 +606,10 @@ bool dcn10_link_encoder_validate_dp_output( const struct dcn10_link_encoder *enc10, const struct dc_crtc_timing *crtc_timing) { - /* default RGB only */ - if (crtc_timing->pixel_encoding == PIXEL_ENCODING_RGB) - return true; - - if (enc10->base.features.flags.bits.IS_YCBCR_CAPABLE) - return true; - - /* for DCE 8.x or later DP Y-only feature, - * we need ASIC cap + FeatureSupportDPYonly, not support 666 - */ - if (crtc_timing->flags.Y_ONLY && - enc10->base.features.flags.bits.IS_YCBCR_CAPABLE && - crtc_timing->display_color_depth != COLOR_DEPTH_666) - return true; + if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR420) + return false; - return false; + return true; } void dcn10_link_encoder_construct( 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 411f89218e01..54626682bab2 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c @@ -98,7 +98,6 @@ static uint32_t get_start_vline(struct timing_generator *optc, const struct dc_c struct dc_crtc_timing patched_crtc_timing; int vesa_sync_start; int asic_blank_end; - int interlace_factor; int vertical_line_start; patched_crtc_timing = *dc_crtc_timing; @@ -112,16 +111,13 @@ static uint32_t get_start_vline(struct timing_generator *optc, const struct dc_c vesa_sync_start - patched_crtc_timing.h_border_left; - interlace_factor = patched_crtc_timing.flags.INTERLACE ? 2 : 1; - vesa_sync_start = patched_crtc_timing.v_addressable + patched_crtc_timing.v_border_bottom + patched_crtc_timing.v_front_porch; asic_blank_end = (patched_crtc_timing.v_total - vesa_sync_start - - patched_crtc_timing.v_border_top) - * interlace_factor; + patched_crtc_timing.v_border_top); vertical_line_start = asic_blank_end - optc->dlg_otg_param.vstartup_start + 1; if (vertical_line_start < 0) { @@ -154,7 +150,7 @@ void optc1_program_vline_interrupt( req_delta_lines--; if (req_delta_lines > vsync_line) - start_line = dc_crtc_timing->v_total - (req_delta_lines - vsync_line) - 1; + start_line = dc_crtc_timing->v_total - (req_delta_lines - vsync_line) + 2; else start_line = vsync_line - req_delta_lines; @@ -186,7 +182,6 @@ void optc1_program_timing( uint32_t v_sync_end; uint32_t v_init, v_fp2; uint32_t h_sync_polarity, v_sync_polarity; - uint32_t interlace_factor; uint32_t start_point = 0; uint32_t field_num = 0; uint32_t h_div_2; @@ -237,16 +232,8 @@ void optc1_program_timing( REG_UPDATE(OTG_H_SYNC_A_CNTL, OTG_H_SYNC_A_POL, h_sync_polarity); - /* Load vertical timing */ + v_total = patched_crtc_timing.v_total - 1; - /* CRTC_V_TOTAL = v_total - 1 */ - if (patched_crtc_timing.flags.INTERLACE) { - interlace_factor = 2; - v_total = 2 * patched_crtc_timing.v_total; - } else { - interlace_factor = 1; - v_total = patched_crtc_timing.v_total - 1; - } REG_SET(OTG_V_TOTAL, 0, OTG_V_TOTAL, v_total); @@ -259,7 +246,7 @@ void optc1_program_timing( OTG_V_TOTAL_MIN, v_total); /* v_sync_start = 0, v_sync_end = v_sync_width */ - v_sync_end = patched_crtc_timing.v_sync_width * interlace_factor; + v_sync_end = patched_crtc_timing.v_sync_width; REG_UPDATE_2(OTG_V_SYNC_A, OTG_V_SYNC_A_START, 0, @@ -271,15 +258,13 @@ void optc1_program_timing( asic_blank_end = (patched_crtc_timing.v_total - vesa_sync_start - - patched_crtc_timing.v_border_top) - * interlace_factor; + patched_crtc_timing.v_border_top); /* v_blank_start = v_blank_end + v_active */ asic_blank_start = asic_blank_end + (patched_crtc_timing.v_border_top + patched_crtc_timing.v_addressable + - patched_crtc_timing.v_border_bottom) - * interlace_factor; + patched_crtc_timing.v_border_bottom); REG_UPDATE_2(OTG_V_BLANK_START_END, OTG_V_BLANK_START, asic_blank_start, @@ -301,7 +286,7 @@ void optc1_program_timing( 0 : 1; REG_UPDATE(OTG_V_SYNC_A_CNTL, - OTG_V_SYNC_A_POL, v_sync_polarity); + OTG_V_SYNC_A_POL, v_sync_polarity); v_init = asic_blank_start; if (optc->dlg_otg_param.signal == SIGNAL_TYPE_DISPLAY_PORT || @@ -532,7 +517,6 @@ bool optc1_validate_timing( struct timing_generator *optc, const struct dc_crtc_timing *timing) { - uint32_t interlace_factor; uint32_t v_blank; uint32_t h_blank; uint32_t min_v_blank; @@ -540,10 +524,8 @@ bool optc1_validate_timing( ASSERT(timing != NULL); - interlace_factor = timing->flags.INTERLACE ? 2 : 1; v_blank = (timing->v_total - timing->v_addressable - - timing->v_border_top - timing->v_border_bottom) * - interlace_factor; + timing->v_border_top - timing->v_border_bottom); h_blank = (timing->h_total - timing->h_addressable - timing->h_border_right - 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 cb1b134b8fcb..a71453a15ae3 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c @@ -507,6 +507,18 @@ static const struct resource_caps res_cap = { .num_ddc = 4, }; +#if defined(CONFIG_DRM_AMD_DC_DCN1_01) +static const struct resource_caps rv2_res_cap = { + .num_timing_generator = 3, + .num_opp = 3, + .num_video_plane = 3, + .num_audio = 3, + .num_stream_encoder = 3, + .num_pll = 3, + .num_ddc = 3, +}; +#endif + static const struct dc_debug_options debug_defaults_drv = { .sanity_checks = true, .disable_dmcu = true, @@ -711,8 +723,7 @@ static const struct encoder_feature_support link_enc_feature = { .flags.bits.IS_HBR2_CAPABLE = true, .flags.bits.IS_HBR3_CAPABLE = true, .flags.bits.IS_TPS3_CAPABLE = true, - .flags.bits.IS_TPS4_CAPABLE = true, - .flags.bits.IS_YCBCR_CAPABLE = true + .flags.bits.IS_TPS4_CAPABLE = true }; struct link_encoder *dcn10_link_encoder_create( @@ -897,7 +908,9 @@ static void destruct(struct dcn10_resource_pool *pool) kfree(DCN10TG_FROM_TG(pool->base.timing_generators[i])); pool->base.timing_generators[i] = NULL; } + } + for (i = 0; i < pool->base.res_cap->num_ddc; i++) { if (pool->base.engines[i] != NULL) pool->base.engines[i]->funcs->destroy_engine(&pool->base.engines[i]); if (pool->base.hw_i2cs[i] != NULL) { @@ -1119,6 +1132,24 @@ static enum dc_status dcn10_validate_plane(const struct dc_plane_state *plane_st return DC_OK; } +static enum dc_status dcn10_get_default_swizzle_mode(struct dc_plane_state *plane_state) +{ + enum dc_status result = DC_OK; + + enum surface_pixel_format surf_pix_format = plane_state->format; + unsigned int bpp = resource_pixel_format_to_bpp(surf_pix_format); + + enum swizzle_mode_values swizzle = DC_SW_LINEAR; + + if (bpp == 64) + swizzle = DC_SW_64KB_D; + else + swizzle = DC_SW_64KB_S; + + plane_state->tiling_info.gfx9.swizzle = swizzle; + return result; +} + static const struct dc_cap_funcs cap_funcs = { .get_dcc_compression_cap = dcn10_get_dcc_compression_cap }; @@ -1129,7 +1160,8 @@ static const struct resource_funcs dcn10_res_pool_funcs = { .validate_bandwidth = dcn_validate_bandwidth, .acquire_idle_pipe_for_layer = dcn10_acquire_idle_pipe_for_layer, .validate_plane = dcn10_validate_plane, - .add_stream_to_ctx = dcn10_add_stream_to_ctx + .add_stream_to_ctx = dcn10_add_stream_to_ctx, + .get_default_swizzle_mode = dcn10_get_default_swizzle_mode }; static uint32_t read_pipe_fuses(struct dc_context *ctx) @@ -1152,7 +1184,12 @@ static bool construct( ctx->dc_bios->regs = &bios_regs; - pool->base.res_cap = &res_cap; +#if defined(CONFIG_DRM_AMD_DC_DCN1_01) + if (ctx->dce_version == DCN_VERSION_1_01) + pool->base.res_cap = &rv2_res_cap; + else +#endif + pool->base.res_cap = &res_cap; pool->base.funcs = &dcn10_res_pool_funcs; /* diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/aux_engine_dce110.c b/drivers/gpu/drm/amd/display/dc/i2caux/dce110/aux_engine_dce110.c index 8eee8ace1259..59c3ed43d609 100644 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/aux_engine_dce110.c +++ b/drivers/gpu/drm/amd/display/dc/i2caux/dce110/aux_engine_dce110.c @@ -346,7 +346,7 @@ static void process_channel_reply( /* in case HPD is LOW, exit AUX transaction */ if ((sw_status & AUX_SW_STATUS__AUX_SW_HPD_DISCON_MASK)) { - reply->status = AUX_CHANNEL_OPERATION_FAILED_HPD_DISCON; + reply->status = AUX_TRANSACTION_REPLY_HPD_DISCON; return; } 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 831a1bdf622c..c1976c175b57 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/core_types.h +++ b/drivers/gpu/drm/amd/display/dc/inc/core_types.h @@ -120,6 +120,9 @@ struct resource_funcs { struct dc *dc, struct dc_state *new_ctx, struct dc_stream_state *stream); + enum dc_status (*get_default_swizzle_mode)( + struct dc_plane_state *plane_state); + }; struct audio_support{ diff --git a/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h b/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h index ddbb673caa08..e688eb9b975c 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h +++ b/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h @@ -504,10 +504,10 @@ struct dcn_bw_internal_vars { float prefetch_mode; float dstx_after_scaler; float dsty_after_scaler; - float v_update_offset_pix; + float v_update_offset_pix[number_of_planes_minus_one + 1]; float total_repeater_delay_time; - float v_update_width_pix; - float v_ready_offset_pix; + float v_update_width_pix[number_of_planes_minus_one + 1]; + float v_ready_offset_pix[number_of_planes_minus_one + 1]; float t_setup; float t_wait; float bandwidth_available_for_immediate_flip; diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h b/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h index 58818920ed41..e28e9770e0a3 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h @@ -58,7 +58,6 @@ struct encoder_feature_support { uint32_t IS_HBR3_CAPABLE:1; uint32_t IS_TPS3_CAPABLE:1; uint32_t IS_TPS4_CAPABLE:1; - uint32_t IS_YCBCR_CAPABLE:1; uint32_t HDMI_6GB_EN:1; } bits; uint32_t raw; diff --git a/drivers/gpu/drm/amd/display/dc/inc/resource.h b/drivers/gpu/drm/amd/display/dc/inc/resource.h index 76d00c6dbca9..33b99e3ab10d 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/resource.h +++ b/drivers/gpu/drm/amd/display/dc/inc/resource.h @@ -172,4 +172,7 @@ void update_audio_usage( const struct resource_pool *pool, struct audio *audio, bool acquired); + +unsigned int resource_pixel_format_to_bpp(enum surface_pixel_format format); + #endif /* DRIVERS_GPU_DRM_AMD_DC_DEV_DC_INC_RESOURCE_H_ */ diff --git a/drivers/gpu/drm/amd/display/include/signal_types.h b/drivers/gpu/drm/amd/display/include/signal_types.h index 03476b142d8e..f56d2891475f 100644 --- a/drivers/gpu/drm/amd/display/include/signal_types.h +++ b/drivers/gpu/drm/amd/display/include/signal_types.h @@ -102,4 +102,9 @@ static inline bool dc_is_audio_capable_signal(enum signal_type signal) dc_is_hdmi_signal(signal)); } +static inline bool dc_is_virtual_signal(enum signal_type signal) +{ + return (signal == SIGNAL_TYPE_VIRTUAL); +} + #endif diff --git a/drivers/gpu/drm/amd/display/modules/color/color_gamma.c b/drivers/gpu/drm/amd/display/modules/color/color_gamma.c index 15427f4fc990..cdcefd087487 100644 --- a/drivers/gpu/drm/amd/display/modules/color/color_gamma.c +++ b/drivers/gpu/drm/amd/display/modules/color/color_gamma.c @@ -1069,10 +1069,14 @@ static void build_evenly_distributed_points( struct dividers dividers) { struct gamma_pixel *p = points; - struct gamma_pixel *p_last = p + numberof_points - 1; + struct gamma_pixel *p_last; uint32_t i = 0; + // This function should not gets called with 0 as a parameter + ASSERT(numberof_points > 0); + p_last = p + numberof_points - 1; + do { struct fixed31_32 value = dc_fixpt_from_fraction(i, numberof_points - 1); @@ -1083,7 +1087,7 @@ static void build_evenly_distributed_points( ++p; ++i; - } while (i != numberof_points); + } while (i < numberof_points); p->r = dc_fixpt_div(p_last->r, dividers.divider1); p->g = dc_fixpt_div(p_last->g, dividers.divider1); diff --git a/drivers/gpu/drm/amd/display/modules/freesync/freesync.c b/drivers/gpu/drm/amd/display/modules/freesync/freesync.c index e1688902a1b0..4018c7180d00 100644 --- a/drivers/gpu/drm/amd/display/modules/freesync/freesync.c +++ b/drivers/gpu/drm/amd/display/modules/freesync/freesync.c @@ -480,22 +480,11 @@ bool mod_freesync_get_v_position(struct mod_freesync *mod_freesync, return false; } -void mod_freesync_build_vrr_infopacket(struct mod_freesync *mod_freesync, - const struct dc_stream_state *stream, - const struct mod_vrr_params *vrr, - struct dc_info_packet *infopacket) +static void build_vrr_infopacket_header_v1(enum signal_type signal, + struct dc_info_packet *infopacket, + unsigned int *payload_size) { - /* SPD info packet for FreeSync */ - unsigned char checksum = 0; - unsigned int idx, payload_size = 0; - - /* Check if Freesync is supported. Return if false. If true, - * set the corresponding bit in the info packet - */ - if (!vrr->supported || !vrr->send_vsif) - return; - - if (dc_is_hdmi_signal(stream->signal)) { + if (dc_is_hdmi_signal(signal)) { /* HEADER */ @@ -510,9 +499,9 @@ void mod_freesync_build_vrr_infopacket(struct mod_freesync *mod_freesync, /* HB2 = [Bits 7:5 = 0] [Bits 4:0 = Length = 0x08] */ infopacket->hb2 = 0x08; - payload_size = 0x08; + *payload_size = 0x08; - } else if (dc_is_dp_signal(stream->signal)) { + } else if (dc_is_dp_signal(signal)) { /* HEADER */ @@ -536,9 +525,62 @@ void mod_freesync_build_vrr_infopacket(struct mod_freesync *mod_freesync, */ infopacket->hb3 = 0x04; - payload_size = 0x1B; + *payload_size = 0x1B; } +} + +static void build_vrr_infopacket_header_v2(enum signal_type signal, + struct dc_info_packet *infopacket, + unsigned int *payload_size) +{ + if (dc_is_hdmi_signal(signal)) { + + /* HEADER */ + + /* HB0 = Packet Type = 0x83 (Source Product + * Descriptor InfoFrame) + */ + infopacket->hb0 = DC_HDMI_INFOFRAME_TYPE_SPD; + + /* HB1 = Version = 0x02 */ + infopacket->hb1 = 0x02; + + /* HB2 = [Bits 7:5 = 0] [Bits 4:0 = Length = 0x09] */ + infopacket->hb2 = 0x09; + + *payload_size = 0x0A; + } else if (dc_is_dp_signal(signal)) { + + /* HEADER */ + + /* HB0 = Secondary-data Packet ID = 0 - Only non-zero + * when used to associate audio related info packets + */ + infopacket->hb0 = 0x00; + + /* HB1 = Packet Type = 0x83 (Source Product + * Descriptor InfoFrame) + */ + infopacket->hb1 = DC_HDMI_INFOFRAME_TYPE_SPD; + + /* HB2 = [Bits 7:0 = Least significant eight bits - + * For INFOFRAME, the value must be 1Bh] + */ + infopacket->hb2 = 0x1B; + + /* HB3 = [Bits 7:2 = INFOFRAME SDP Version Number = 0x2] + * [Bits 1:0 = Most significant two bits = 0x00] + */ + infopacket->hb3 = 0x08; + + *payload_size = 0x1B; + } +} + +static void build_vrr_infopacket_data(const struct mod_vrr_params *vrr, + struct dc_info_packet *infopacket) +{ /* PB1 = 0x1A (24bit AMD IEEE OUI (0x00001A) - Byte 0) */ infopacket->sb[1] = 0x1A; @@ -576,15 +618,39 @@ void mod_freesync_build_vrr_infopacket(struct mod_freesync *mod_freesync, */ infopacket->sb[8] = (unsigned char)(vrr->max_refresh_in_uhz / 1000000); - /* PB9 - PB27 = Reserved */ + //FreeSync HDR + infopacket->sb[9] = 0; + infopacket->sb[10] = 0; +} + +static void build_vrr_infopacket_fs2_data(enum color_transfer_func app_tf, + struct dc_info_packet *infopacket) +{ + if (app_tf != transfer_func_unknown) { + infopacket->valid = true; + + infopacket->sb[6] |= 0x08; // PB6 = [Bit 3 = Native Color Active] + + if (app_tf == transfer_func_gamma_22) { + infopacket->sb[9] |= 0x04; // PB6 = [Bit 2 = Gamma 2.2 EOTF Active] + } + } +} + +static void build_vrr_infopacket_checksum(unsigned int *payload_size, + struct dc_info_packet *infopacket) +{ /* Calculate checksum */ + unsigned int idx = 0; + unsigned char checksum = 0; + checksum += infopacket->hb0; checksum += infopacket->hb1; checksum += infopacket->hb2; checksum += infopacket->hb3; - for (idx = 1; idx <= payload_size; idx++) + for (idx = 1; idx <= *payload_size; idx++) checksum += infopacket->sb[idx]; /* PB0 = Checksum (one byte complement) */ @@ -593,6 +659,64 @@ void mod_freesync_build_vrr_infopacket(struct mod_freesync *mod_freesync, infopacket->valid = true; } +static void build_vrr_infopacket_v1(enum signal_type signal, + const struct mod_vrr_params *vrr, + struct dc_info_packet *infopacket) +{ + /* SPD info packet for FreeSync */ + unsigned int payload_size = 0; + + build_vrr_infopacket_header_v1(signal, infopacket, &payload_size); + build_vrr_infopacket_data(vrr, infopacket); + build_vrr_infopacket_checksum(&payload_size, infopacket); + + infopacket->valid = true; +} + +static void build_vrr_infopacket_v2(enum signal_type signal, + const struct mod_vrr_params *vrr, + const enum color_transfer_func *app_tf, + struct dc_info_packet *infopacket) +{ + unsigned int payload_size = 0; + + build_vrr_infopacket_header_v2(signal, infopacket, &payload_size); + build_vrr_infopacket_data(vrr, infopacket); + + if (app_tf != NULL) + build_vrr_infopacket_fs2_data(*app_tf, infopacket); + + build_vrr_infopacket_checksum(&payload_size, infopacket); + + infopacket->valid = true; +} + +void mod_freesync_build_vrr_infopacket(struct mod_freesync *mod_freesync, + const struct dc_stream_state *stream, + const struct mod_vrr_params *vrr, + enum vrr_packet_type packet_type, + const enum color_transfer_func *app_tf, + struct dc_info_packet *infopacket) +{ + /* SPD info packet for FreeSync */ + + /* Check if Freesync is supported. Return if false. If true, + * set the corresponding bit in the info packet + */ + if (!vrr->supported || !vrr->send_vsif) + return; + + switch (packet_type) { + case packet_type_fs2: + build_vrr_infopacket_v2(stream->signal, vrr, app_tf, infopacket); + break; + case packet_type_vrr: + case packet_type_fs1: + default: + build_vrr_infopacket_v1(stream->signal, vrr, infopacket); + } +} + void mod_freesync_build_vrr_params(struct mod_freesync *mod_freesync, const struct dc_stream_state *stream, struct mod_freesync_config *in_config, diff --git a/drivers/gpu/drm/amd/display/modules/inc/mod_freesync.h b/drivers/gpu/drm/amd/display/modules/inc/mod_freesync.h index a0f32cde721c..949a8b62aa98 100644 --- a/drivers/gpu/drm/amd/display/modules/inc/mod_freesync.h +++ b/drivers/gpu/drm/amd/display/modules/inc/mod_freesync.h @@ -54,7 +54,7 @@ #ifndef MOD_FREESYNC_H_ #define MOD_FREESYNC_H_ -#include "dm_services.h" +#include "mod_shared.h" // Access structures struct mod_freesync { @@ -144,6 +144,8 @@ void mod_freesync_get_settings(struct mod_freesync *mod_freesync, void mod_freesync_build_vrr_infopacket(struct mod_freesync *mod_freesync, const struct dc_stream_state *stream, const struct mod_vrr_params *vrr, + enum vrr_packet_type packet_type, + const enum color_transfer_func *app_tf, struct dc_info_packet *infopacket); void mod_freesync_build_vrr_params(struct mod_freesync *mod_freesync, diff --git a/drivers/gpu/drm/amd/display/modules/inc/mod_shared.h b/drivers/gpu/drm/amd/display/modules/inc/mod_shared.h new file mode 100644 index 000000000000..238c431ae483 --- /dev/null +++ b/drivers/gpu/drm/amd/display/modules/inc/mod_shared.h @@ -0,0 +1,49 @@ +/* + * Copyright 2016 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 MOD_SHARED_H_ +#define MOD_SHARED_H_ + +enum color_transfer_func { + transfer_func_unknown, + transfer_func_srgb, + transfer_func_bt709, + transfer_func_pq2084, + transfer_func_pq2084_interim, + transfer_func_linear_0_1, + transfer_func_linear_0_125, + transfer_func_dolbyvision, + transfer_func_gamma_22, + transfer_func_gamma_26 +}; + +enum vrr_packet_type { + packet_type_vrr, + packet_type_fs1, + packet_type_fs2 +}; + +#endif /* MOD_SHARED_H_ */ diff --git a/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c b/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c index 52378fc69079..ff8bfb9b43b0 100644 --- a/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c +++ b/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c @@ -48,9 +48,12 @@ static void mod_build_vsc_infopacket(const struct dc_stream_state *stream, unsigned int i; unsigned int pixelEncoding = 0; unsigned int colorimetryFormat = 0; + bool stereo3dSupport = false; - if (stream->timing.timing_3d_format != TIMING_3D_FORMAT_NONE && stream->view_format != VIEW_3D_FORMAT_NONE) + if (stream->timing.timing_3d_format != TIMING_3D_FORMAT_NONE && stream->view_format != VIEW_3D_FORMAT_NONE) { vscPacketRevision = 1; + stereo3dSupport = true; + } /*VSC packet set to 2 when DP revision >= 1.2*/ if (stream->psr_version != 0) @@ -94,12 +97,59 @@ static void mod_build_vsc_infopacket(const struct dc_stream_state *stream, info_packet->hb2 = 0x01; // 01h = Revision number. VSC SDP supporting 3D stereo only info_packet->hb3 = 0x01; // 01h = VSC SDP supporting 3D stereo only (HB2 = 01h). - if (stream->timing.timing_3d_format == TIMING_3D_FORMAT_INBAND_FA) - info_packet->sb[0] = 0x1; - info_packet->valid = true; } + if (stereo3dSupport) { + /* ==============================================================================================================| + * A. STEREO 3D + * ==============================================================================================================| + * VSC Payload (1 byte) From DP1.2 spec + * + * Bits 3:0 (Stereo Interface Method Code) | Bits 7:4 (Stereo Interface Method Specific Parameter) + * ----------------------------------------------------------------------------------------------------- + * 0 = Non Stereo Video | Must be set to 0x0 + * ----------------------------------------------------------------------------------------------------- + * 1 = Frame/Field Sequential | 0x0: L + R view indication based on MISC1 bit 2:1 + * | 0x1: Right when Stereo Signal = 1 + * | 0x2: Left when Stereo Signal = 1 + * | (others reserved) + * ----------------------------------------------------------------------------------------------------- + * 2 = Stacked Frame | 0x0: Left view is on top and right view on bottom + * | (others reserved) + * ----------------------------------------------------------------------------------------------------- + * 3 = Pixel Interleaved | 0x0: horiz interleaved, right view pixels on even lines + * | 0x1: horiz interleaved, right view pixels on odd lines + * | 0x2: checker board, start with left view pixel + * | 0x3: vertical interleaved, start with left view pixels + * | 0x4: vertical interleaved, start with right view pixels + * | (others reserved) + * ----------------------------------------------------------------------------------------------------- + * 4 = Side-by-side | 0x0: left half represents left eye view + * | 0x1: left half represents right eye view + */ + switch (stream->timing.timing_3d_format) { + case TIMING_3D_FORMAT_HW_FRAME_PACKING: + case TIMING_3D_FORMAT_SW_FRAME_PACKING: + case TIMING_3D_FORMAT_TOP_AND_BOTTOM: + case TIMING_3D_FORMAT_TB_SW_PACKED: + info_packet->sb[0] = 0x02; // Stacked Frame, Left view is on top and right view on bottom. + break; + case TIMING_3D_FORMAT_DP_HDMI_INBAND_FA: + case TIMING_3D_FORMAT_INBAND_FA: + info_packet->sb[0] = 0x01; // Frame/Field Sequential, L + R view indication based on MISC1 bit 2:1 + break; + case TIMING_3D_FORMAT_SIDE_BY_SIDE: + case TIMING_3D_FORMAT_SBS_SW_PACKED: + info_packet->sb[0] = 0x04; // Side-by-side + break; + default: + info_packet->sb[0] = 0x00; // No Stereo Video, Shall be cleared to 0x0. + break; + } + + } + /* 05h = VSC SDP supporting 3D stereo, PSR2, and Pixel Encoding/Colorimetry Format indication. * Added in DP1.3, a DP Source device is allowed to indicate the pixel encoding/colorimetry * format to the DP Sink device with VSC SDP only when the DP Sink device supports it diff --git a/drivers/gpu/drm/amd/display/modules/stats/stats.c b/drivers/gpu/drm/amd/display/modules/stats/stats.c index 3d4c1b1ab8c4..03121ca64fe4 100644 --- a/drivers/gpu/drm/amd/display/modules/stats/stats.c +++ b/drivers/gpu/drm/amd/display/modules/stats/stats.c @@ -186,12 +186,8 @@ void mod_stats_destroy(struct mod_stats *mod_stats) if (mod_stats != NULL) { struct core_stats *core_stats = MOD_STATS_TO_CORE(mod_stats); - if (core_stats->time != NULL) - kfree(core_stats->time); - - if (core_stats->events != NULL) - kfree(core_stats->events); - + kfree(core_stats->time); + kfree(core_stats->events); kfree(core_stats); } } diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h index 86b167ec9863..2083c308007c 100644 --- a/drivers/gpu/drm/amd/include/amd_shared.h +++ b/drivers/gpu/drm/amd/include/amd_shared.h @@ -109,6 +109,7 @@ enum amd_powergating_state { #define AMD_PG_SUPPORT_GFX_PIPELINE (1 << 12) #define AMD_PG_SUPPORT_MMHUB (1 << 13) #define AMD_PG_SUPPORT_VCN (1 << 14) +#define AMD_PG_SUPPORT_VCN_DPG (1 << 15) enum PP_FEATURE_MASK { PP_SCLK_DPM_MASK = 0x1, diff --git a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_9_0_offset.h b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_9_0_offset.h index 4ce090db7ef7..529b37db274c 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_9_0_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_9_0_offset.h @@ -2449,6 +2449,8 @@ #define mmCP_ECC_FIRSTOCCURRENCE_RING2_BASE_IDX 0 #define mmGB_EDC_MODE 0x107e #define mmGB_EDC_MODE_BASE_IDX 0 +#define mmCP_DEBUG 0x107f +#define mmCP_DEBUG_BASE_IDX 0 #define mmCP_CPF_DEBUG 0x1080 #define mmCP_PQ_WPTR_POLL_CNTL 0x1083 #define mmCP_PQ_WPTR_POLL_CNTL_BASE_IDX 0 diff --git a/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_9_0_offset.h b/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_9_0_offset.h index efd2704d0f8f..0d6891095f62 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_9_0_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_9_0_offset.h @@ -175,4 +175,7 @@ #define mmSMUSVI0_PLANE0_CURRENTVID_BASE_IDX 0 #define mmSMUSVI0_PLANE0_CURRENTVID 0x0013 +#define mmSMUSVI0_TEL_PLANE0_BASE_IDX 0 +#define mmSMUSVI0_TEL_PLANE0 0x0004 + #endif diff --git a/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_9_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_9_0_sh_mask.h index 2487ab9621e9..b1d9d8be1119 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_9_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/smuio/smuio_9_0_sh_mask.h @@ -258,4 +258,7 @@ #define SMUSVI0_PLANE0_CURRENTVID__CURRENT_SVI0_PLANE0_VID__SHIFT 0x18 #define SMUSVI0_PLANE0_CURRENTVID__CURRENT_SVI0_PLANE0_VID_MASK 0xFF000000L +#define SMUSVI0_TEL_PLANE0__SVI0_PLANE0_VDDCOR__SHIFT 0x10 +#define SMUSVI0_TEL_PLANE0__SVI0_PLANE0_VDDCOR_MASK 0x01FF0000L + #endif diff --git a/drivers/gpu/drm/amd/include/asic_reg/thm/thm_11_0_2_offset.h b/drivers/gpu/drm/amd/include/asic_reg/thm/thm_11_0_2_offset.h index 510ec3c70626..a9eb57a53e59 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/thm/thm_11_0_2_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/thm/thm_11_0_2_offset.h @@ -26,6 +26,18 @@ #define mmCG_MULT_THERMAL_STATUS 0x005f #define mmCG_MULT_THERMAL_STATUS_BASE_IDX 0 +#define mmCG_FDO_CTRL0 0x0067 +#define mmCG_FDO_CTRL0_BASE_IDX 0 + +#define mmCG_FDO_CTRL1 0x0068 +#define mmCG_FDO_CTRL1_BASE_IDX 0 + +#define mmCG_FDO_CTRL2 0x0069 +#define mmCG_FDO_CTRL2_BASE_IDX 0 + +#define mmCG_TACH_CTRL 0x006a +#define mmCG_TACH_CTRL_BASE_IDX 0 + #define mmTHM_THERMAL_INT_ENA 0x000a #define mmTHM_THERMAL_INT_ENA_BASE_IDX 0 #define mmTHM_THERMAL_INT_CTRL 0x000b diff --git a/drivers/gpu/drm/amd/include/asic_reg/thm/thm_11_0_2_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/thm/thm_11_0_2_sh_mask.h index f69533fa6abf..d130d92aee19 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/thm/thm_11_0_2_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/thm/thm_11_0_2_sh_mask.h @@ -28,6 +28,16 @@ #define CG_MULT_THERMAL_STATUS__CTF_TEMP__SHIFT 0x9 #define CG_MULT_THERMAL_STATUS__ASIC_MAX_TEMP_MASK 0x000001FFL #define CG_MULT_THERMAL_STATUS__CTF_TEMP_MASK 0x0003FE00L +#define CG_FDO_CTRL2__TMIN__SHIFT 0x0 +#define CG_FDO_CTRL2__TMIN_MASK 0x000000FFL +#define CG_FDO_CTRL2__FDO_PWM_MODE__SHIFT 0xb +#define CG_FDO_CTRL2__FDO_PWM_MODE_MASK 0x00003800L +#define CG_FDO_CTRL1__FMAX_DUTY100__SHIFT 0x0 +#define CG_FDO_CTRL1__FMAX_DUTY100_MASK 0x000000FFL +#define CG_FDO_CTRL0__FDO_STATIC_DUTY__SHIFT 0x0 +#define CG_FDO_CTRL0__FDO_STATIC_DUTY_MASK 0x000000FFL +#define CG_TACH_CTRL__TARGET_PERIOD__SHIFT 0x3 +#define CG_TACH_CTRL__TARGET_PERIOD_MASK 0xFFFFFFF8L //THM_THERMAL_INT_ENA #define THM_THERMAL_INT_ENA__THERM_INTH_SET__SHIFT 0x0 diff --git a/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_1_0_offset.h b/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_1_0_offset.h index 216a401028de..442ca7c471a5 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_1_0_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_1_0_offset.h @@ -33,6 +33,14 @@ #define mmUVD_POWER_STATUS_BASE_IDX 1 #define mmCC_UVD_HARVESTING 0x00c7 #define mmCC_UVD_HARVESTING_BASE_IDX 1 +#define mmUVD_DPG_LMA_CTL 0x00d1 +#define mmUVD_DPG_LMA_CTL_BASE_IDX 1 +#define mmUVD_DPG_LMA_DATA 0x00d2 +#define mmUVD_DPG_LMA_DATA_BASE_IDX 1 +#define mmUVD_DPG_LMA_MASK 0x00d3 +#define mmUVD_DPG_LMA_MASK_BASE_IDX 1 +#define mmUVD_DPG_PAUSE 0x00d4 +#define mmUVD_DPG_PAUSE_BASE_IDX 1 #define mmUVD_SCRATCH1 0x00d5 #define mmUVD_SCRATCH1_BASE_IDX 1 #define mmUVD_SCRATCH2 0x00d6 @@ -74,6 +82,18 @@ #define mmUVD_LCM_CGC_CNTRL 0x0123 #define mmUVD_LCM_CGC_CNTRL_BASE_IDX 1 +#define mmUVD_MIF_CURR_UV_ADDR_CONFIG 0x0184 +#define mmUVD_MIF_CURR_UV_ADDR_CONFIG_BASE_IDX 1 +#define mmUVD_MIF_REF_UV_ADDR_CONFIG 0x0185 +#define mmUVD_MIF_REF_UV_ADDR_CONFIG_BASE_IDX 1 +#define mmUVD_MIF_RECON1_UV_ADDR_CONFIG 0x0186 +#define mmUVD_MIF_RECON1_UV_ADDR_CONFIG_BASE_IDX 1 +#define mmUVD_MIF_CURR_ADDR_CONFIG 0x0192 +#define mmUVD_MIF_CURR_ADDR_CONFIG_BASE_IDX 1 +#define mmUVD_MIF_REF_ADDR_CONFIG 0x0193 +#define mmUVD_MIF_REF_ADDR_CONFIG_BASE_IDX 1 +#define mmUVD_MIF_RECON1_ADDR_CONFIG 0x01c5 +#define mmUVD_MIF_RECON1_ADDR_CONFIG_BASE_IDX 1 // addressBlock: uvd_uvdnpdec // base address: 0x20000 @@ -319,6 +339,8 @@ #define mmUVD_LMI_VM_CTRL_BASE_IDX 1 #define mmUVD_LMI_SWAP_CNTL 0x056d #define mmUVD_LMI_SWAP_CNTL_BASE_IDX 1 +#define mmUVD_MPC_CNTL 0x0577 +#define mmUVD_MPC_CNTL_BASE_IDX 1 #define mmUVD_MPC_SET_MUXA0 0x0579 #define mmUVD_MPC_SET_MUXA0_BASE_IDX 1 #define mmUVD_MPC_SET_MUXA1 0x057a diff --git a/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_1_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_1_0_sh_mask.h index 124383dac284..63457f9df4c5 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_1_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_1_0_sh_mask.h @@ -87,6 +87,26 @@ //CC_UVD_HARVESTING #define CC_UVD_HARVESTING__UVD_DISABLE__SHIFT 0x1 #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_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 @@ -965,6 +985,7 @@ #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__RE_OFLD_MIF_WR_REQ_NUM__SHIFT 0x11 #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 @@ -973,6 +994,7 @@ #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__RE_OFLD_MIF_WR_REQ_NUM_MASK 0x01FE0000L //UVD_MASTINT_EN #define UVD_MASTINT_EN__OVERRUN_RST__SHIFT 0x0 #define UVD_MASTINT_EN__VCPU_EN__SHIFT 0x1 @@ -983,6 +1005,7 @@ #define UVD_MASTINT_EN__SYS_EN_MASK 0x00000004L #define UVD_MASTINT_EN__INT_OVERRUN_MASK 0x007FFFF0L //UVD_SYS_INT_EN +#define UVD_SYS_INT_EN__UVD_JRBC_EN__SHIFT 0x4 #define UVD_SYS_INT_EN__UVD_JRBC_EN_MASK 0x00000010L //JPEG_CGC_CTRL #define JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT 0x0 @@ -1024,6 +1047,19 @@ #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__RFU_MASK 0xF8000000L +//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_WRITE_CLEAN_RAW__SHIFT 0x6 +#define UVD_LMI_STATUS__UMC_READ_CLEAN_RAW__SHIFT 0x9 +#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_WRITE_CLEAN_RAW_MASK 0x00000040L +#define UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK 0x00000200L //UVD_LMI_SWAP_CNTL #define UVD_LMI_SWAP_CNTL__RB_MC_SWAP__SHIFT 0x0 #define UVD_LMI_SWAP_CNTL__IB_MC_SWAP__SHIFT 0x2 @@ -1057,6 +1093,9 @@ #define UVD_LMI_SWAP_CNTL__RB_WR_MC_SWAP_MASK 0x0C000000L #define UVD_LMI_SWAP_CNTL__RE_MC_SWAP_MASK 0x30000000L #define UVD_LMI_SWAP_CNTL__MP_MC_SWAP_MASK 0xC0000000L +//UVD_MPC_CNTL +#define UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT 0x3 +#define UVD_MPC_CNTL__REPLACEMENT_MODE_MASK 0x00000038L //UVD_MPC_SET_MUXA0 #define UVD_MPC_SET_MUXA0__VARA_0__SHIFT 0x0 #define UVD_MPC_SET_MUXA0__VARA_1__SHIFT 0x6 @@ -1138,7 +1177,11 @@ #define UVD_VCPU_CACHE_SIZE2__CACHE_SIZE2_MASK 0x001FFFFFL //UVD_VCPU_CNTL #define UVD_VCPU_CNTL__CLK_EN__SHIFT 0x9 +#define UVD_VCPU_CNTL__MIF_WR_LOW_THRESHOLD_BP__SHIFT 0x11 +#define UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT 0x14 #define UVD_VCPU_CNTL__CLK_EN_MASK 0x00000200L +#define UVD_VCPU_CNTL__MIF_WR_LOW_THRESHOLD_BP_MASK 0x00020000L +#define UVD_VCPU_CNTL__PRB_TIMEOUT_VAL_MASK 0x0FF00000L //UVD_SOFT_RESET #define UVD_SOFT_RESET__RBC_SOFT_RESET__SHIFT 0x0 #define UVD_SOFT_RESET__LBSI_SOFT_RESET__SHIFT 0x1 diff --git a/drivers/gpu/drm/amd/include/atomfirmware.h b/drivers/gpu/drm/amd/include/atomfirmware.h index 8ae7adb7329b..d2e7c0fa96c2 100644 --- a/drivers/gpu/drm/amd/include/atomfirmware.h +++ b/drivers/gpu/drm/amd/include/atomfirmware.h @@ -1532,6 +1532,94 @@ struct atom_smc_dpm_info_v4_3 uint32_t boardreserved[10]; }; +struct smudpm_i2ccontrollerconfig_t { + uint32_t enabled; + uint32_t slaveaddress; + uint32_t controllerport; + uint32_t controllername; + uint32_t thermalthrottler; + uint32_t i2cprotocol; + uint32_t i2cspeed; +}; + +struct atom_smc_dpm_info_v4_4 +{ + struct atom_common_table_header table_header; + uint32_t i2c_padding[3]; + + uint16_t maxvoltagestepgfx; + uint16_t maxvoltagestepsoc; + + uint8_t vddgfxvrmapping; + uint8_t vddsocvrmapping; + uint8_t vddmem0vrmapping; + uint8_t vddmem1vrmapping; + + uint8_t gfxulvphasesheddingmask; + uint8_t soculvphasesheddingmask; + uint8_t externalsensorpresent; + uint8_t padding8_v; + + uint16_t gfxmaxcurrent; + uint8_t gfxoffset; + uint8_t padding_telemetrygfx; + + uint16_t socmaxcurrent; + uint8_t socoffset; + uint8_t padding_telemetrysoc; + + uint16_t mem0maxcurrent; + uint8_t mem0offset; + uint8_t padding_telemetrymem0; + + uint16_t mem1maxcurrent; + uint8_t mem1offset; + uint8_t padding_telemetrymem1; + + + uint8_t acdcgpio; + uint8_t acdcpolarity; + uint8_t vr0hotgpio; + uint8_t vr0hotpolarity; + + uint8_t vr1hotgpio; + uint8_t vr1hotpolarity; + uint8_t padding1; + uint8_t padding2; + + + uint8_t ledpin0; + uint8_t ledpin1; + uint8_t ledpin2; + uint8_t padding8_4; + + + uint8_t pllgfxclkspreadenabled; + uint8_t pllgfxclkspreadpercent; + uint16_t pllgfxclkspreadfreq; + + + uint8_t uclkspreadenabled; + uint8_t uclkspreadpercent; + uint16_t uclkspreadfreq; + + + uint8_t fclkspreadenabled; + uint8_t fclkspreadpercent; + uint16_t fclkspreadfreq; + + + uint8_t fllgfxclkspreadenabled; + uint8_t fllgfxclkspreadpercent; + uint16_t fllgfxclkspreadfreq; + + + struct smudpm_i2ccontrollerconfig_t i2ccontrollers[7]; + + + uint32_t boardreserved[10]; +}; + /* *************************************************************************** Data Table asic_profiling_info structure diff --git a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h index f43ed96cfa6c..64ecffd52126 100644 --- a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h +++ b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h @@ -146,10 +146,10 @@ struct kgd2kfd_shared_resources { * is reserved: (D & reserved_doorbell_mask) == reserved_doorbell_val * * KFD currently uses 1024 (= 0x3ff) doorbells per process. If - * doorbells 0x0f0-0x0f7 and 0x2f-0x2f7 are reserved, that means - * mask would be set to 0x1f8 and val set to 0x0f0. + * doorbells 0x0e0-0x0ff and 0x2e0-0x2ff are reserved, that means + * mask would be set to 0x1e0 and val set to 0x0e0. */ - unsigned int sdma_doorbell[2][2]; + unsigned int sdma_doorbell[2][8]; unsigned int reserved_doorbell_mask; unsigned int reserved_doorbell_val; @@ -409,9 +409,9 @@ struct kfd2kgd_calls { struct dma_fence **ef); void (*destroy_process_vm)(struct kgd_dev *kgd, void *vm); void (*release_process_vm)(struct kgd_dev *kgd, void *vm); - uint32_t (*get_process_page_dir)(void *vm); + uint64_t (*get_process_page_dir)(void *vm); void (*set_vm_context_page_table_base)(struct kgd_dev *kgd, - uint32_t vmid, uint32_t page_table_base); + uint32_t vmid, uint64_t page_table_base); int (*alloc_memory_of_gpu)(struct kgd_dev *kgd, uint64_t va, uint64_t size, void *vm, struct kgd_mem **mem, uint64_t *offset, diff --git a/drivers/gpu/drm/amd/include/kgd_pp_interface.h b/drivers/gpu/drm/amd/include/kgd_pp_interface.h index 448dee481a38..980e696989b1 100644 --- a/drivers/gpu/drm/amd/include/kgd_pp_interface.h +++ b/drivers/gpu/drm/amd/include/kgd_pp_interface.h @@ -113,6 +113,9 @@ enum amd_pp_sensors { AMDGPU_PP_SENSOR_GPU_POWER, AMDGPU_PP_SENSOR_STABLE_PSTATE_SCLK, AMDGPU_PP_SENSOR_STABLE_PSTATE_MCLK, + AMDGPU_PP_SENSOR_ENABLED_SMC_FEATURES_MASK, + AMDGPU_PP_SENSOR_MIN_FAN_RPM, + AMDGPU_PP_SENSOR_MAX_FAN_RPM, }; enum amd_pp_task { @@ -227,6 +230,7 @@ struct amd_pm_funcs { enum amd_dpm_forced_level (*get_performance_level)(void *handle); enum amd_pm_state_type (*get_current_power_state)(void *handle); int (*get_fan_speed_rpm)(void *handle, uint32_t *rpm); + int (*set_fan_speed_rpm)(void *handle, uint32_t rpm); int (*get_pp_num_states)(void *handle, struct pp_states_info *data); int (*get_pp_table)(void *handle, char **table); int (*set_pp_table)(void *handle, const char *buf, size_t size); @@ -271,6 +275,7 @@ struct amd_pm_funcs { int (*get_display_mode_validation_clocks)(void *handle, struct amd_pp_simple_clock_info *clocks); int (*notify_smu_enable_pwe)(void *handle); + int (*enable_mgpu_fan_boost)(void *handle); }; #endif diff --git a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c index da4ebff5b74d..e8964cae6b93 100644 --- a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c +++ b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c @@ -109,11 +109,8 @@ static int pp_sw_fini(void *handle) hwmgr_sw_fini(hwmgr); - if (adev->firmware.load_type == AMDGPU_FW_LOAD_SMU) { - release_firmware(adev->pm.fw); - adev->pm.fw = NULL; - amdgpu_ucode_fini_bo(adev); - } + release_firmware(adev->pm.fw); + adev->pm.fw = NULL; return 0; } @@ -124,9 +121,6 @@ static int pp_hw_init(void *handle) struct amdgpu_device *adev = handle; struct pp_hwmgr *hwmgr = adev->powerplay.pp_handle; - if (adev->firmware.load_type == AMDGPU_FW_LOAD_SMU) - amdgpu_ucode_init_bo(adev); - ret = hwmgr_hw_init(hwmgr); if (ret) @@ -273,8 +267,23 @@ const struct amdgpu_ip_block_version pp_smu_ip_block = .funcs = &pp_ip_funcs, }; +/* This interface only be supported On Vi, + * because only smu7/8 can help to load gfx/sdma fw, + * smu need to be enabled before load other ip's fw. + * so call start smu to load smu7 fw and other ip's fw + */ static int pp_dpm_load_fw(void *handle) { + struct pp_hwmgr *hwmgr = handle; + + if (!hwmgr || !hwmgr->smumgr_funcs || !hwmgr->smumgr_funcs->start_smu) + return -EINVAL; + + if (hwmgr->smumgr_funcs->start_smu(hwmgr)) { + pr_err("fw load failed\n"); + return -EINVAL; + } + return 0; } @@ -576,6 +585,24 @@ static int pp_dpm_get_fan_speed_rpm(void *handle, uint32_t *rpm) return ret; } +static int pp_dpm_set_fan_speed_rpm(void *handle, uint32_t rpm) +{ + struct pp_hwmgr *hwmgr = handle; + int ret = 0; + + if (!hwmgr || !hwmgr->pm_en) + return -EINVAL; + + if (hwmgr->hwmgr_func->set_fan_speed_rpm == NULL) { + pr_info("%s was not implemented.\n", __func__); + return 0; + } + mutex_lock(&hwmgr->smu_lock); + ret = hwmgr->hwmgr_func->set_fan_speed_rpm(hwmgr, rpm); + mutex_unlock(&hwmgr->smu_lock); + return ret; +} + static int pp_dpm_get_pp_num_states(void *handle, struct pp_states_info *data) { @@ -813,6 +840,12 @@ static int pp_dpm_read_sensor(void *handle, int idx, case AMDGPU_PP_SENSOR_STABLE_PSTATE_MCLK: *((uint32_t *)value) = hwmgr->pstate_mclk; return 0; + case AMDGPU_PP_SENSOR_MIN_FAN_RPM: + *((uint32_t *)value) = hwmgr->thermal_controller.fanInfo.ulMinRPM; + return 0; + case AMDGPU_PP_SENSOR_MAX_FAN_RPM: + *((uint32_t *)value) = hwmgr->thermal_controller.fanInfo.ulMaxRPM; + return 0; default: mutex_lock(&hwmgr->smu_lock); ret = hwmgr->hwmgr_func->read_sensor(hwmgr, idx, value, size); @@ -861,9 +894,14 @@ static int pp_set_power_profile_mode(void *handle, long *input, uint32_t size) pr_info("%s was not implemented.\n", __func__); return ret; } + + if (hwmgr->dpm_level != AMD_DPM_FORCED_LEVEL_MANUAL) { + pr_info("power profile setting is for manual dpm mode only.\n"); + return ret; + } + mutex_lock(&hwmgr->smu_lock); - if (hwmgr->dpm_level == AMD_DPM_FORCED_LEVEL_MANUAL) - ret = hwmgr->hwmgr_func->set_power_profile_mode(hwmgr, input, size); + ret = hwmgr->hwmgr_func->set_power_profile_mode(hwmgr, input, size); mutex_unlock(&hwmgr->smu_lock); return ret; } @@ -1196,6 +1234,21 @@ static void pp_dpm_powergate_acp(void *handle, bool gate) hwmgr->hwmgr_func->powergate_acp(hwmgr, gate); } +static void pp_dpm_powergate_sdma(void *handle, bool gate) +{ + struct pp_hwmgr *hwmgr = handle; + + if (!hwmgr) + return; + + if (hwmgr->hwmgr_func->powergate_sdma == NULL) { + pr_info("%s was not implemented.\n", __func__); + return; + } + + hwmgr->hwmgr_func->powergate_sdma(hwmgr, gate); +} + static int pp_set_powergating_by_smu(void *handle, uint32_t block_type, bool gate) { @@ -1218,6 +1271,9 @@ static int pp_set_powergating_by_smu(void *handle, case AMD_IP_BLOCK_TYPE_ACP: pp_dpm_powergate_acp(handle, gate); break; + case AMD_IP_BLOCK_TYPE_SDMA: + pp_dpm_powergate_sdma(handle, gate); + break; default: break; } @@ -1243,6 +1299,24 @@ static int pp_notify_smu_enable_pwe(void *handle) return 0; } +static int pp_enable_mgpu_fan_boost(void *handle) +{ + struct pp_hwmgr *hwmgr = handle; + + if (!hwmgr || !hwmgr->pm_en) + return -EINVAL; + + if (hwmgr->hwmgr_func->enable_mgpu_fan_boost == NULL) { + return 0; + } + + mutex_lock(&hwmgr->smu_lock); + hwmgr->hwmgr_func->enable_mgpu_fan_boost(hwmgr); + mutex_unlock(&hwmgr->smu_lock); + + return 0; +} + static const struct amd_pm_funcs pp_dpm_funcs = { .load_firmware = pp_dpm_load_fw, .wait_for_fw_loading_complete = pp_dpm_fw_loading_complete, @@ -1255,6 +1329,7 @@ static const struct amd_pm_funcs pp_dpm_funcs = { .set_fan_speed_percent = pp_dpm_set_fan_speed_percent, .get_fan_speed_percent = pp_dpm_get_fan_speed_percent, .get_fan_speed_rpm = pp_dpm_get_fan_speed_rpm, + .set_fan_speed_rpm = pp_dpm_set_fan_speed_rpm, .get_pp_num_states = pp_dpm_get_pp_num_states, .get_pp_table = pp_dpm_get_pp_table, .set_pp_table = pp_dpm_set_pp_table, @@ -1287,4 +1362,5 @@ static const struct amd_pm_funcs pp_dpm_funcs = { .display_clock_voltage_request = pp_display_clock_voltage_request, .get_display_mode_validation_clocks = pp_get_display_mode_validation_clocks, .notify_smu_enable_pwe = pp_notify_smu_enable_pwe, + .enable_mgpu_fan_boost = pp_enable_mgpu_fan_boost, }; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c index 7500a3e61dba..47ac92369739 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c @@ -89,7 +89,6 @@ int hwmgr_early_init(struct pp_hwmgr *hwmgr) hwmgr_init_default_caps(hwmgr); hwmgr_set_user_specify_caps(hwmgr); hwmgr->fan_ctrl_is_in_default_mode = true; - hwmgr->reload_fw = 1; hwmgr_init_workload_prority(hwmgr); switch (hwmgr->chip_family) { @@ -209,17 +208,6 @@ int hwmgr_hw_init(struct pp_hwmgr *hwmgr) { int ret = 0; - if (!hwmgr || !hwmgr->smumgr_funcs) - return -EINVAL; - - if (hwmgr->smumgr_funcs->start_smu) { - ret = hwmgr->smumgr_funcs->start_smu(hwmgr); - if (ret) { - pr_err("smc start failed\n"); - return -EINVAL; - } - } - if (!hwmgr->pm_en) return 0; @@ -320,13 +308,6 @@ int hwmgr_resume(struct pp_hwmgr *hwmgr) if (!hwmgr) return -EINVAL; - if (hwmgr->smumgr_funcs && hwmgr->smumgr_funcs->start_smu) { - if (hwmgr->smumgr_funcs->start_smu(hwmgr)) { - pr_err("smc start failed\n"); - return -EINVAL; - } - } - if (!hwmgr->pm_en) return 0; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomfwctrl.c b/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomfwctrl.c index d27c1c9df286..4588bddf8b33 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomfwctrl.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomfwctrl.c @@ -488,7 +488,8 @@ int pp_atomfwctrl_get_gpio_information(struct pp_hwmgr *hwmgr, return 0; } -int pp_atomfwctrl_get_clk_information_by_clkid(struct pp_hwmgr *hwmgr, BIOS_CLKID id, uint32_t *frequency) +int pp_atomfwctrl_get_clk_information_by_clkid(struct pp_hwmgr *hwmgr, + uint8_t id, uint32_t *frequency) { struct amdgpu_device *adev = hwmgr->adev; struct atom_get_smu_clock_info_parameters_v3_1 parameters; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomfwctrl.h b/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomfwctrl.h index 22e21668c93a..fe9e8ceef50e 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomfwctrl.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomfwctrl.h @@ -236,7 +236,7 @@ int pp_atomfwctrl_get_vbios_bootup_values(struct pp_hwmgr *hwmgr, int pp_atomfwctrl_get_smc_dpm_information(struct pp_hwmgr *hwmgr, struct pp_atomfwctrl_smc_dpm_parameters *param); int pp_atomfwctrl_get_clk_information_by_clkid(struct pp_hwmgr *hwmgr, - BIOS_CLKID id, uint32_t *frequency); + uint8_t id, uint32_t *frequency); #endif diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu10_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu10_hwmgr.c index 9808bd48b386..dd18cb710391 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu10_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu10_hwmgr.c @@ -552,6 +552,8 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr, { struct smu10_hwmgr *data = hwmgr->backend; struct amdgpu_device *adev = hwmgr->adev; + uint32_t min_sclk = hwmgr->display_config->min_core_set_clock; + uint32_t min_mclk = hwmgr->display_config->min_mem_set_clock/100; if (hwmgr->smu_version < 0x1E3700) { pr_info("smu firmware version too old, can not set dpm level\n"); @@ -563,6 +565,13 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr, (adev->rev_id >= 8)) return 0; + if (min_sclk < data->gfx_min_freq_limit) + min_sclk = data->gfx_min_freq_limit; + + min_sclk /= 100; /* transfer 10KHz to MHz */ + if (min_mclk < data->clock_table.FClocks[0].Freq) + min_mclk = data->clock_table.FClocks[0].Freq; + switch (level) { case AMD_DPM_FORCED_LEVEL_HIGH: case AMD_DPM_FORCED_LEVEL_PROFILE_PEAK: @@ -595,18 +604,18 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr, case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_SCLK: smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetHardMinGfxClk, - data->gfx_min_freq_limit/100); + min_sclk); smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetSoftMaxGfxClk, - data->gfx_min_freq_limit/100); + min_sclk); break; case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK: smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetHardMinFclkByFreq, - SMU10_UMD_PSTATE_MIN_FCLK); + min_mclk); smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetSoftMaxFclkByFreq, - SMU10_UMD_PSTATE_MIN_FCLK); + min_mclk); break; case AMD_DPM_FORCED_LEVEL_PROFILE_STANDARD: smum_send_msg_to_smc_with_parameter(hwmgr, @@ -638,12 +647,12 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr, case AMD_DPM_FORCED_LEVEL_AUTO: smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetHardMinGfxClk, - data->gfx_min_freq_limit/100); + min_sclk); smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetHardMinFclkByFreq, hwmgr->display_config->num_display > 3 ? SMU10_UMD_PSTATE_PEAK_FCLK : - SMU10_UMD_PSTATE_MIN_FCLK); + min_mclk); smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetHardMinSocclkByFreq, @@ -674,10 +683,10 @@ static int smu10_dpm_force_dpm_level(struct pp_hwmgr *hwmgr, data->gfx_min_freq_limit/100); smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetHardMinFclkByFreq, - SMU10_UMD_PSTATE_MIN_FCLK); + min_mclk); smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetSoftMaxFclkByFreq, - SMU10_UMD_PSTATE_MIN_FCLK); + min_mclk); break; case AMD_DPM_FORCED_LEVEL_MANUAL: case AMD_DPM_FORCED_LEVEL_PROFILE_EXIT: @@ -1144,6 +1153,14 @@ static int smu10_powergate_mmhub(struct pp_hwmgr *hwmgr) return smum_send_msg_to_smc(hwmgr, PPSMC_MSG_PowerGateMmHub); } +static int smu10_powergate_sdma(struct pp_hwmgr *hwmgr, bool gate) +{ + if (gate) + return smum_send_msg_to_smc(hwmgr, PPSMC_MSG_PowerDownSdma); + else + return smum_send_msg_to_smc(hwmgr, PPSMC_MSG_PowerUpSdma); +} + static void smu10_powergate_vcn(struct pp_hwmgr *hwmgr, bool bgate) { if (bgate) { @@ -1199,6 +1216,7 @@ static const struct pp_hwmgr_func smu10_hwmgr_funcs = { .smus_notify_pwe = smu10_smus_notify_pwe, .display_clock_voltage_request = smu10_display_clock_voltage_request, .powergate_gfx = smu10_gfx_off_control, + .powergate_sdma = smu10_powergate_sdma, }; int smu10_init_function_pointers(struct pp_hwmgr *hwmgr) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c index 04b7da0e39a6..6c99cbf51c08 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c @@ -4106,17 +4106,17 @@ static int smu7_register_irq_handlers(struct pp_hwmgr *hwmgr) source->funcs = &smu7_irq_funcs; amdgpu_irq_add_id((struct amdgpu_device *)(hwmgr->adev), - AMDGPU_IH_CLIENTID_LEGACY, + AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CG_TSS_THERMAL_LOW_TO_HIGH, source); amdgpu_irq_add_id((struct amdgpu_device *)(hwmgr->adev), - AMDGPU_IH_CLIENTID_LEGACY, + AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CG_TSS_THERMAL_HIGH_TO_LOW, source); /* Register CTF(GPIO_19) interrupt */ amdgpu_irq_add_id((struct amdgpu_device *)(hwmgr->adev), - AMDGPU_IH_CLIENTID_LEGACY, + AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GPIO_19, source); @@ -5035,6 +5035,18 @@ static int smu7_get_performance_level(struct pp_hwmgr *hwmgr, const struct pp_hw return 0; } +static int smu7_power_off_asic(struct pp_hwmgr *hwmgr) +{ + int result; + + result = smu7_disable_dpm_tasks(hwmgr); + PP_ASSERT_WITH_CODE((0 == result), + "[disable_dpm_tasks] Failed to disable DPM!", + ); + + return result; +} + static const struct pp_hwmgr_func smu7_hwmgr_funcs = { .backend_init = &smu7_hwmgr_backend_init, .backend_fini = &smu7_hwmgr_backend_fini, @@ -5092,6 +5104,7 @@ static const struct pp_hwmgr_func smu7_hwmgr_funcs = { .get_power_profile_mode = smu7_get_power_profile_mode, .set_power_profile_mode = smu7_set_power_profile_mode, .get_performance_level = smu7_get_performance_level, + .power_off_asic = smu7_power_off_asic, }; uint8_t smu7_get_sleep_divider_id_from_clock(uint32_t clock, diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_thermal.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_thermal.c index 44527755e747..5bdc0df5a9f4 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_thermal.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_thermal.c @@ -260,6 +260,7 @@ int smu7_fan_ctrl_set_fan_speed_rpm(struct pp_hwmgr *hwmgr, uint32_t speed) if (hwmgr->thermal_controller.fanInfo.bNoFan || (hwmgr->thermal_controller.fanInfo. ucTachometerPulsesPerRevolution == 0) || + speed == 0 || (speed < hwmgr->thermal_controller.fanInfo.ulMinRPM) || (speed > hwmgr->thermal_controller.fanInfo.ulMaxRPM)) return 0; @@ -272,7 +273,7 @@ int smu7_fan_ctrl_set_fan_speed_rpm(struct pp_hwmgr *hwmgr, uint32_t speed) tach_period = 60 * crystal_clock_freq * 10000 / (8 * speed); PHM_WRITE_VFPF_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, - CG_TACH_STATUS, TACH_PERIOD, tach_period); + CG_TACH_CTRL, TARGET_PERIOD, tach_period); return smu7_fan_ctrl_set_static_mode(hwmgr, FDO_PWM_MODE_STATIC_RPM); } diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu8_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu8_hwmgr.c index b8637049198d..fef111ddb736 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu8_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu8_hwmgr.c @@ -880,7 +880,7 @@ static int smu8_set_power_state_tasks(struct pp_hwmgr *hwmgr, const void *input) smu8_update_low_mem_pstate(hwmgr, input); return 0; -}; +} static int smu8_setup_asic_task(struct pp_hwmgr *hwmgr) @@ -934,14 +934,6 @@ static void smu8_reset_cc6_data(struct pp_hwmgr *hwmgr) hw_data->cc6_settings.cpu_pstate_disable = false; } -static int smu8_power_off_asic(struct pp_hwmgr *hwmgr) -{ - smu8_power_up_display_clock_sys_pll(hwmgr); - smu8_clear_nb_dpm_flag(hwmgr); - smu8_reset_cc6_data(hwmgr); - return 0; -}; - static void smu8_program_voting_clients(struct pp_hwmgr *hwmgr) { cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, @@ -1011,6 +1003,17 @@ static void smu8_reset_acp_boot_level(struct pp_hwmgr *hwmgr) data->acp_boot_level = 0xff; } +static int smu8_enable_dpm_tasks(struct pp_hwmgr *hwmgr) +{ + smu8_program_voting_clients(hwmgr); + if (smu8_start_dpm(hwmgr)) + return -EINVAL; + smu8_program_bootup_state(hwmgr); + smu8_reset_acp_boot_level(hwmgr); + + return 0; +} + static int smu8_disable_dpm_tasks(struct pp_hwmgr *hwmgr) { smu8_disable_nb_dpm(hwmgr); @@ -1020,18 +1023,16 @@ static int smu8_disable_dpm_tasks(struct pp_hwmgr *hwmgr) return -EINVAL; return 0; -}; +} -static int smu8_enable_dpm_tasks(struct pp_hwmgr *hwmgr) +static int smu8_power_off_asic(struct pp_hwmgr *hwmgr) { - smu8_program_voting_clients(hwmgr); - if (smu8_start_dpm(hwmgr)) - return -EINVAL; - smu8_program_bootup_state(hwmgr); - smu8_reset_acp_boot_level(hwmgr); - + smu8_disable_dpm_tasks(hwmgr); + smu8_power_up_display_clock_sys_pll(hwmgr); + smu8_clear_nb_dpm_flag(hwmgr); + smu8_reset_cc6_data(hwmgr); return 0; -}; +} static int smu8_apply_state_adjust_rules(struct pp_hwmgr *hwmgr, struct pp_power_state *prequest_ps, @@ -1227,14 +1228,17 @@ static int smu8_dpm_force_dpm_level(struct pp_hwmgr *hwmgr, static int smu8_dpm_powerdown_uvd(struct pp_hwmgr *hwmgr) { - if (PP_CAP(PHM_PlatformCaps_UVDPowerGating)) + if (PP_CAP(PHM_PlatformCaps_UVDPowerGating)) { + smu8_nbdpm_pstate_enable_disable(hwmgr, true, true); return smum_send_msg_to_smc(hwmgr, PPSMC_MSG_UVDPowerOFF); + } return 0; } static int smu8_dpm_powerup_uvd(struct pp_hwmgr *hwmgr) { if (PP_CAP(PHM_PlatformCaps_UVDPowerGating)) { + smu8_nbdpm_pstate_enable_disable(hwmgr, false, true); return smum_send_msg_to_smc_with_parameter( hwmgr, PPSMC_MSG_UVDPowerON, diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu_helper.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu_helper.c index 2aab1b475945..4714b5b59825 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu_helper.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu_helper.c @@ -39,6 +39,50 @@ uint16_t convert_to_vddc(uint8_t vid) return (uint16_t) ((6200 - (vid * 25)) / VOLTAGE_SCALE); } +int phm_copy_clock_limits_array( + struct pp_hwmgr *hwmgr, + uint32_t **pptable_info_array, + const uint32_t *pptable_array, + uint32_t power_saving_clock_count) +{ + uint32_t array_size, i; + uint32_t *table; + + array_size = sizeof(uint32_t) * power_saving_clock_count; + table = kzalloc(array_size, GFP_KERNEL); + if (NULL == table) + return -ENOMEM; + + for (i = 0; i < power_saving_clock_count; i++) + table[i] = le32_to_cpu(pptable_array[i]); + + *pptable_info_array = table; + + return 0; +} + +int phm_copy_overdrive_settings_limits_array( + struct pp_hwmgr *hwmgr, + uint32_t **pptable_info_array, + const uint32_t *pptable_array, + uint32_t od_setting_count) +{ + uint32_t array_size, i; + uint32_t *table; + + array_size = sizeof(uint32_t) * od_setting_count; + table = kzalloc(array_size, GFP_KERNEL); + if (NULL == table) + return -ENOMEM; + + for (i = 0; i < od_setting_count; i++) + table[i] = le32_to_cpu(pptable_array[i]); + + *pptable_info_array = table; + + return 0; +} + uint32_t phm_set_field_to_u32(u32 offset, u32 original_data, u32 field, u32 size) { u32 mask = 0; @@ -545,7 +589,7 @@ int phm_irq_process(struct amdgpu_device *adev, uint32_t client_id = entry->client_id; uint32_t src_id = entry->src_id; - if (client_id == AMDGPU_IH_CLIENTID_LEGACY) { + if (client_id == AMDGPU_IRQ_CLIENTID_LEGACY) { if (src_id == VISLANDS30_IV_SRCID_CG_TSS_THERMAL_LOW_TO_HIGH) pr_warn("GPU over temperature range detected on PCIe %d:%d.%d!\n", PCI_BUS_NUM(adev->pdev->devfn), diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu_helper.h b/drivers/gpu/drm/amd/powerplay/hwmgr/smu_helper.h index 5454289d5226..ad33983a8064 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu_helper.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu_helper.h @@ -47,6 +47,18 @@ struct watermarks { uint32_t padding[7]; }; +int phm_copy_clock_limits_array( + struct pp_hwmgr *hwmgr, + uint32_t **pptable_info_array, + const uint32_t *pptable_array, + uint32_t power_saving_clock_count); + +int phm_copy_overdrive_settings_limits_array( + struct pp_hwmgr *hwmgr, + uint32_t **pptable_info_array, + const uint32_t *pptable_array, + uint32_t od_setting_count); + extern int phm_wait_for_register_unequal(struct pp_hwmgr *hwmgr, uint32_t index, uint32_t value, uint32_t mask); diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c index ca9be583fb62..419a1d77d661 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c @@ -39,6 +39,7 @@ #include "soc15_common.h" #include "pppcielanes.h" #include "vega10_hwmgr.h" +#include "vega10_smumgr.h" #include "vega10_processpptables.h" #include "vega10_pptable.h" #include "vega10_thermal.h" @@ -3713,6 +3714,11 @@ static int vega10_read_sensor(struct pp_hwmgr *hwmgr, int idx, SMUSVI0_PLANE0_CURRENTVID__CURRENT_SVI0_PLANE0_VID__SHIFT; *((uint32_t *)value) = (uint32_t)convert_to_vddc((uint8_t)val_vid); return 0; + case AMDGPU_PP_SENSOR_ENABLED_SMC_FEATURES_MASK: + ret = vega10_get_enabled_smc_features(hwmgr, (uint64_t *)value); + if (!ret) + *size = 8; + break; default: ret = -EINVAL; break; @@ -4940,16 +4946,6 @@ static const struct pp_hwmgr_func vega10_hwmgr_funcs = { .get_performance_level = vega10_get_performance_level, }; -int vega10_enable_smc_features(struct pp_hwmgr *hwmgr, - bool enable, uint32_t feature_mask) -{ - int msg = enable ? PPSMC_MSG_EnableSmuFeatures : - PPSMC_MSG_DisableSmuFeatures; - - return smum_send_msg_to_smc_with_parameter(hwmgr, - msg, feature_mask); -} - int vega10_hwmgr_init(struct pp_hwmgr *hwmgr) { hwmgr->hwmgr_func = &vega10_hwmgr_funcs; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.h b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.h index 339820da9e6a..89870556de1b 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.h @@ -441,7 +441,5 @@ int vega10_update_uvd_dpm(struct pp_hwmgr *hwmgr, bool bgate); int vega10_update_samu_dpm(struct pp_hwmgr *hwmgr, bool bgate); int vega10_update_acp_dpm(struct pp_hwmgr *hwmgr, bool bgate); int vega10_enable_disable_vce_dpm(struct pp_hwmgr *hwmgr, bool enable); -int vega10_enable_smc_features(struct pp_hwmgr *hwmgr, - bool enable, uint32_t feature_mask); #endif /* _VEGA10_HWMGR_H_ */ diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_powertune.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_powertune.c index 22364875a943..2d88abf97e7b 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_powertune.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_powertune.c @@ -23,6 +23,7 @@ #include "hwmgr.h" #include "vega10_hwmgr.h" +#include "vega10_smumgr.h" #include "vega10_powertune.h" #include "vega10_ppsmc.h" #include "vega10_inc.h" diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_processpptables.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_processpptables.c index 16b1a9cf6cf0..b8747a5c9204 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_processpptables.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_processpptables.c @@ -451,23 +451,23 @@ static int get_tdp_table( le16_to_cpu(power_tune_table_v2->usLoadLineResistance); } else { power_tune_table_v3 = (ATOM_Vega10_PowerTune_Table_V3 *)table; - tdp_table->usMaximumPowerDeliveryLimit = power_tune_table_v3->usSocketPowerLimit; - tdp_table->usTDC = power_tune_table_v3->usTdcLimit; - tdp_table->usEDCLimit = power_tune_table_v3->usEdcLimit; - tdp_table->usSoftwareShutdownTemp = power_tune_table_v3->usSoftwareShutdownTemp; - tdp_table->usTemperatureLimitTedge = power_tune_table_v3->usTemperatureLimitTedge; - tdp_table->usTemperatureLimitHotspot = power_tune_table_v3->usTemperatureLimitHotSpot; - tdp_table->usTemperatureLimitLiquid1 = power_tune_table_v3->usTemperatureLimitLiquid1; - tdp_table->usTemperatureLimitLiquid2 = power_tune_table_v3->usTemperatureLimitLiquid2; - tdp_table->usTemperatureLimitHBM = power_tune_table_v3->usTemperatureLimitHBM; - tdp_table->usTemperatureLimitVrVddc = power_tune_table_v3->usTemperatureLimitVrSoc; - tdp_table->usTemperatureLimitVrMvdd = power_tune_table_v3->usTemperatureLimitVrMem; - tdp_table->usTemperatureLimitPlx = power_tune_table_v3->usTemperatureLimitPlx; + tdp_table->usMaximumPowerDeliveryLimit = le16_to_cpu(power_tune_table_v3->usSocketPowerLimit); + tdp_table->usTDC = le16_to_cpu(power_tune_table_v3->usTdcLimit); + tdp_table->usEDCLimit = le16_to_cpu(power_tune_table_v3->usEdcLimit); + tdp_table->usSoftwareShutdownTemp = le16_to_cpu(power_tune_table_v3->usSoftwareShutdownTemp); + tdp_table->usTemperatureLimitTedge = le16_to_cpu(power_tune_table_v3->usTemperatureLimitTedge); + tdp_table->usTemperatureLimitHotspot = le16_to_cpu(power_tune_table_v3->usTemperatureLimitHotSpot); + tdp_table->usTemperatureLimitLiquid1 = le16_to_cpu(power_tune_table_v3->usTemperatureLimitLiquid1); + tdp_table->usTemperatureLimitLiquid2 = le16_to_cpu(power_tune_table_v3->usTemperatureLimitLiquid2); + tdp_table->usTemperatureLimitHBM = le16_to_cpu(power_tune_table_v3->usTemperatureLimitHBM); + tdp_table->usTemperatureLimitVrVddc = le16_to_cpu(power_tune_table_v3->usTemperatureLimitVrSoc); + tdp_table->usTemperatureLimitVrMvdd = le16_to_cpu(power_tune_table_v3->usTemperatureLimitVrMem); + tdp_table->usTemperatureLimitPlx = le16_to_cpu(power_tune_table_v3->usTemperatureLimitPlx); tdp_table->ucLiquid1_I2C_address = power_tune_table_v3->ucLiquid1_I2C_address; tdp_table->ucLiquid2_I2C_address = power_tune_table_v3->ucLiquid2_I2C_address; - tdp_table->usBoostStartTemperature = power_tune_table_v3->usBoostStartTemperature; - tdp_table->usBoostStopTemperature = power_tune_table_v3->usBoostStopTemperature; - tdp_table->ulBoostClock = power_tune_table_v3->ulBoostClock; + tdp_table->usBoostStartTemperature = le16_to_cpu(power_tune_table_v3->usBoostStartTemperature); + tdp_table->usBoostStopTemperature = le16_to_cpu(power_tune_table_v3->usBoostStopTemperature); + tdp_table->ulBoostClock = le32_to_cpu(power_tune_table_v3->ulBoostClock); get_scl_sda_value(power_tune_table_v3->ucLiquid_I2C_Line, &scl, &sda); diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_thermal.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_thermal.c index aa044c1955fe..3f807d6c95ce 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_thermal.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_thermal.c @@ -23,6 +23,7 @@ #include "vega10_thermal.h" #include "vega10_hwmgr.h" +#include "vega10_smumgr.h" #include "vega10_ppsmc.h" #include "vega10_inc.h" #include "soc15_common.h" @@ -311,6 +312,7 @@ int vega10_fan_ctrl_set_fan_speed_rpm(struct pp_hwmgr *hwmgr, uint32_t speed) int result = 0; if (hwmgr->thermal_controller.fanInfo.bNoFan || + speed == 0 || (speed < hwmgr->thermal_controller.fanInfo.ulMinRPM) || (speed > hwmgr->thermal_controller.fanInfo.ulMaxRPM)) return -1; @@ -321,9 +323,9 @@ int vega10_fan_ctrl_set_fan_speed_rpm(struct pp_hwmgr *hwmgr, uint32_t speed) if (!result) { crystal_clock_freq = amdgpu_asic_get_xclk((struct amdgpu_device *)hwmgr->adev); tach_period = 60 * crystal_clock_freq * 10000 / (8 * speed); - WREG32_SOC15(THM, 0, mmCG_TACH_STATUS, - REG_SET_FIELD(RREG32_SOC15(THM, 0, mmCG_TACH_STATUS), - CG_TACH_STATUS, TACH_PERIOD, + WREG32_SOC15(THM, 0, mmCG_TACH_CTRL, + REG_SET_FIELD(RREG32_SOC15(THM, 0, mmCG_TACH_CTRL), + CG_TACH_CTRL, TARGET_PERIOD, tach_period)); } return vega10_fan_ctrl_set_static_mode(hwmgr, FDO_PWM_MODE_STATIC_RPM); diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c index 0789d64246ca..9600e2f226e9 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c @@ -745,8 +745,8 @@ static int vega12_init_smc_table(struct pp_hwmgr *hwmgr) memcpy(pp_table, pptable_information->smc_pptable, sizeof(PPTable_t)); - result = vega12_copy_table_to_smc(hwmgr, - (uint8_t *)pp_table, TABLE_PPTABLE); + result = smum_smc_table_manager(hwmgr, + (uint8_t *)pp_table, TABLE_PPTABLE, false); PP_ASSERT_WITH_CODE(!result, "Failed to upload PPtable!", return result); @@ -1317,7 +1317,11 @@ static int vega12_read_sensor(struct pp_hwmgr *hwmgr, int idx, break; case AMDGPU_PP_SENSOR_GPU_POWER: ret = vega12_get_gpu_power(hwmgr, (uint32_t *)value); - + break; + case AMDGPU_PP_SENSOR_ENABLED_SMC_FEATURES_MASK: + ret = vega12_get_enabled_smc_features(hwmgr, (uint64_t *)value); + if (!ret) + *size = 8; break; default: ret = -EINVAL; @@ -2103,8 +2107,8 @@ static int vega12_display_configuration_changed_task(struct pp_hwmgr *hwmgr) if ((data->water_marks_bitmap & WaterMarksExist) && !(data->water_marks_bitmap & WaterMarksLoaded)) { - result = vega12_copy_table_to_smc(hwmgr, - (uint8_t *)wm_table, TABLE_WATERMARKS); + result = smum_smc_table_manager(hwmgr, + (uint8_t *)wm_table, TABLE_WATERMARKS, false); PP_ASSERT_WITH_CODE(result, "Failed to update WMTABLE!", return EINVAL); data->water_marks_bitmap |= WaterMarksLoaded; } diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_processpptables.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_processpptables.c index cb3a5b1737c8..9817f7a5ed29 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_processpptables.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_processpptables.c @@ -99,50 +99,6 @@ static int set_platform_caps(struct pp_hwmgr *hwmgr, uint32_t powerplay_caps) return 0; } -static int copy_clock_limits_array( - struct pp_hwmgr *hwmgr, - uint32_t **pptable_info_array, - const uint32_t *pptable_array) -{ - uint32_t array_size, i; - uint32_t *table; - - array_size = sizeof(uint32_t) * ATOM_VEGA12_PPCLOCK_COUNT; - - table = kzalloc(array_size, GFP_KERNEL); - if (NULL == table) - return -ENOMEM; - - for (i = 0; i < ATOM_VEGA12_PPCLOCK_COUNT; i++) - table[i] = pptable_array[i]; - - *pptable_info_array = table; - - return 0; -} - -static int copy_overdrive_settings_limits_array( - struct pp_hwmgr *hwmgr, - uint32_t **pptable_info_array, - const uint32_t *pptable_array) -{ - uint32_t array_size, i; - uint32_t *table; - - array_size = sizeof(uint32_t) * ATOM_VEGA12_ODSETTING_COUNT; - - table = kzalloc(array_size, GFP_KERNEL); - if (NULL == table) - return -ENOMEM; - - for (i = 0; i < ATOM_VEGA12_ODSETTING_COUNT; i++) - table[i] = pptable_array[i]; - - *pptable_info_array = table; - - return 0; -} - static int append_vbios_pptable(struct pp_hwmgr *hwmgr, PPTable_t *ppsmc_pptable) { struct pp_atomfwctrl_smc_dpm_parameters smc_dpm_table; @@ -250,14 +206,22 @@ static int init_powerplay_table_information( phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_MicrocodeFanControl); - if (powerplay_table->ODSettingsMax[ATOM_VEGA12_ODSETTING_GFXCLKFMAX] > VEGA12_ENGINECLOCK_HARDMAX) + if (le32_to_cpu(powerplay_table->ODSettingsMax[ATOM_VEGA12_ODSETTING_GFXCLKFMAX]) > VEGA12_ENGINECLOCK_HARDMAX) hwmgr->platform_descriptor.overdriveLimit.engineClock = VEGA12_ENGINECLOCK_HARDMAX; else - hwmgr->platform_descriptor.overdriveLimit.engineClock = powerplay_table->ODSettingsMax[ATOM_VEGA12_ODSETTING_GFXCLKFMAX]; - hwmgr->platform_descriptor.overdriveLimit.memoryClock = powerplay_table->ODSettingsMax[ATOM_VEGA12_ODSETTING_UCLKFMAX]; - - copy_overdrive_settings_limits_array(hwmgr, &pptable_information->od_settings_max, powerplay_table->ODSettingsMax); - copy_overdrive_settings_limits_array(hwmgr, &pptable_information->od_settings_min, powerplay_table->ODSettingsMin); + hwmgr->platform_descriptor.overdriveLimit.engineClock = + le32_to_cpu(powerplay_table->ODSettingsMax[ATOM_VEGA12_ODSETTING_GFXCLKFMAX]); + hwmgr->platform_descriptor.overdriveLimit.memoryClock = + le32_to_cpu(powerplay_table->ODSettingsMax[ATOM_VEGA12_ODSETTING_UCLKFMAX]); + + phm_copy_overdrive_settings_limits_array(hwmgr, + &pptable_information->od_settings_max, + powerplay_table->ODSettingsMax, + ATOM_VEGA12_ODSETTING_COUNT); + phm_copy_overdrive_settings_limits_array(hwmgr, + &pptable_information->od_settings_min, + powerplay_table->ODSettingsMin, + ATOM_VEGA12_ODSETTING_COUNT); /* hwmgr->platformDescriptor.minOverdriveVDDC = 0; hwmgr->platformDescriptor.maxOverdriveVDDC = 0; @@ -267,15 +231,15 @@ static int init_powerplay_table_information( && hwmgr->platform_descriptor.overdriveLimit.memoryClock > 0) phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_ACOverdriveSupport); - pptable_information->us_small_power_limit1 = powerplay_table->usSmallPowerLimit1; - pptable_information->us_small_power_limit2 = powerplay_table->usSmallPowerLimit2; - pptable_information->us_boost_power_limit = powerplay_table->usBoostPowerLimit; - pptable_information->us_od_turbo_power_limit = powerplay_table->usODTurboPowerLimit; - pptable_information->us_od_powersave_power_limit = powerplay_table->usODPowerSavePowerLimit; + pptable_information->us_small_power_limit1 = le16_to_cpu(powerplay_table->usSmallPowerLimit1); + pptable_information->us_small_power_limit2 = le16_to_cpu(powerplay_table->usSmallPowerLimit2); + pptable_information->us_boost_power_limit = le16_to_cpu(powerplay_table->usBoostPowerLimit); + pptable_information->us_od_turbo_power_limit = le16_to_cpu(powerplay_table->usODTurboPowerLimit); + pptable_information->us_od_powersave_power_limit = le16_to_cpu(powerplay_table->usODPowerSavePowerLimit); - pptable_information->us_software_shutdown_temp = powerplay_table->usSoftwareShutdownTemp; + pptable_information->us_software_shutdown_temp = le16_to_cpu(powerplay_table->usSoftwareShutdownTemp); - hwmgr->platform_descriptor.TDPODLimit = (uint16_t)powerplay_table->ODSettingsMax[ATOM_VEGA12_ODSETTING_POWERPERCENTAGE]; + hwmgr->platform_descriptor.TDPODLimit = le32_to_cpu(powerplay_table->ODSettingsMax[ATOM_VEGA12_ODSETTING_POWERPERCENTAGE]); disable_power_control = 0; if (!disable_power_control) { @@ -285,8 +249,8 @@ static int init_powerplay_table_information( PHM_PlatformCaps_PowerControl); } - copy_clock_limits_array(hwmgr, &pptable_information->power_saving_clock_max, powerplay_table->PowerSavingClockMax); - copy_clock_limits_array(hwmgr, &pptable_information->power_saving_clock_min, powerplay_table->PowerSavingClockMin); + phm_copy_clock_limits_array(hwmgr, &pptable_information->power_saving_clock_max, powerplay_table->PowerSavingClockMax, ATOM_VEGA12_PPCLOCK_COUNT); + phm_copy_clock_limits_array(hwmgr, &pptable_information->power_saving_clock_min, powerplay_table->PowerSavingClockMin, ATOM_VEGA12_PPCLOCK_COUNT); pptable_information->smc_pptable = (PPTable_t *)kmalloc(sizeof(PPTable_t), GFP_KERNEL); if (pptable_information->smc_pptable == NULL) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c index d45cbfe8e184..b4dbbb7c334c 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c @@ -46,6 +46,9 @@ #include "ppinterrupt.h" #include "pp_overdriver.h" #include "pp_thermal.h" +#include "soc15_common.h" +#include "smuio/smuio_9_0_offset.h" +#include "smuio/smuio_9_0_sh_mask.h" static void vega20_set_default_registry_data(struct pp_hwmgr *hwmgr) { @@ -461,7 +464,7 @@ static int vega20_get_number_of_dpm_level(struct pp_hwmgr *hwmgr, "[GetNumOfDpmLevel] failed to get dpm levels!", return ret); - vega20_read_arg_from_smc(hwmgr, num_of_levels); + *num_of_levels = smum_get_argument(hwmgr); PP_ASSERT_WITH_CODE(*num_of_levels > 0, "[GetNumOfDpmLevel] number of clk levels is invalid!", return -EINVAL); @@ -481,7 +484,7 @@ static int vega20_get_dpm_frequency_by_index(struct pp_hwmgr *hwmgr, "[GetDpmFreqByIndex] failed to get dpm freq by index!", return ret); - vega20_read_arg_from_smc(hwmgr, clk); + *clk = smum_get_argument(hwmgr); PP_ASSERT_WITH_CODE(*clk, "[GetDpmFreqByIndex] clk value is invalid!", return -EINVAL); @@ -743,8 +746,8 @@ static int vega20_init_smc_table(struct pp_hwmgr *hwmgr) memcpy(pp_table, pptable_information->smc_pptable, sizeof(PPTable_t)); - result = vega20_copy_table_to_smc(hwmgr, - (uint8_t *)pp_table, TABLE_PPTABLE); + result = smum_smc_table_manager(hwmgr, + (uint8_t *)pp_table, TABLE_PPTABLE, false); PP_ASSERT_WITH_CODE(!result, "[InitSMCTable] Failed to upload PPtable!", return result); @@ -1044,7 +1047,7 @@ static int vega20_od8_get_gfx_clock_base_voltage( "[GetBaseVoltage] failed to get GFXCLK AVFS voltage from SMU!", return ret); - vega20_read_arg_from_smc(hwmgr, voltage); + *voltage = smum_get_argument(hwmgr); *voltage = *voltage / VOLTAGE_SCALE; return 0; @@ -1067,7 +1070,7 @@ static int vega20_od8_initialize_default_settings( vega20_od8_set_feature_id(hwmgr); /* Set default values */ - ret = vega20_copy_table_from_smc(hwmgr, (uint8_t *)od_table, TABLE_OVERDRIVE); + ret = smum_smc_table_manager(hwmgr, (uint8_t *)od_table, TABLE_OVERDRIVE, true); PP_ASSERT_WITH_CODE(!ret, "Failed to export over drive table!", return ret); @@ -1195,7 +1198,7 @@ static int vega20_od8_initialize_default_settings( } } - ret = vega20_copy_table_to_smc(hwmgr, (uint8_t *)od_table, TABLE_OVERDRIVE); + ret = smum_smc_table_manager(hwmgr, (uint8_t *)od_table, TABLE_OVERDRIVE, false); PP_ASSERT_WITH_CODE(!ret, "Failed to import over drive table!", return ret); @@ -1214,7 +1217,7 @@ static int vega20_od8_set_settings( struct vega20_od8_single_setting *od8_settings = data->od8_settings.od8_settings_array; - ret = vega20_copy_table_from_smc(hwmgr, (uint8_t *)(&od_table), TABLE_OVERDRIVE); + ret = smum_smc_table_manager(hwmgr, (uint8_t *)(&od_table), TABLE_OVERDRIVE, true); PP_ASSERT_WITH_CODE(!ret, "Failed to export over drive table!", return ret); @@ -1271,7 +1274,7 @@ static int vega20_od8_set_settings( break; } - ret = vega20_copy_table_to_smc(hwmgr, (uint8_t *)(&od_table), TABLE_OVERDRIVE); + ret = smum_smc_table_manager(hwmgr, (uint8_t *)(&od_table), TABLE_OVERDRIVE, false); PP_ASSERT_WITH_CODE(!ret, "Failed to import over drive table!", return ret); @@ -1401,7 +1404,7 @@ static int vega20_get_max_sustainable_clock(struct pp_hwmgr *hwmgr, (clock_select << 16))) == 0, "[GetMaxSustainableClock] Failed to get max DC clock from SMC!", return ret); - vega20_read_arg_from_smc(hwmgr, clock); + *clock = smum_get_argument(hwmgr); /* if DC limit is zero, return AC limit */ if (*clock == 0) { @@ -1410,7 +1413,7 @@ static int vega20_get_max_sustainable_clock(struct pp_hwmgr *hwmgr, (clock_select << 16))) == 0, "[GetMaxSustainableClock] failed to get max AC clock from SMC!", return ret); - vega20_read_arg_from_smc(hwmgr, clock); + *clock = smum_get_argument(hwmgr); } return 0; @@ -1474,6 +1477,19 @@ static int vega20_init_max_sustainable_clocks(struct pp_hwmgr *hwmgr) return 0; } +static int vega20_enable_mgpu_fan_boost(struct pp_hwmgr *hwmgr) +{ + int result; + + result = smum_send_msg_to_smc(hwmgr, + PPSMC_MSG_SetMGpuFanBoostLimitRpm); + PP_ASSERT_WITH_CODE(!result, + "[EnableMgpuFan] Failed to enable mgpu fan boost!", + return result); + + return 0; +} + static void vega20_init_powergate_state(struct pp_hwmgr *hwmgr) { struct vega20_hwmgr *data = @@ -1544,6 +1560,14 @@ static int vega20_enable_dpm_tasks(struct pp_hwmgr *hwmgr) "[EnableDPMTasks] Failed to populate umdpstate clocks!", return result); + result = smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_GetPptLimit, + POWER_SOURCE_AC << 16); + PP_ASSERT_WITH_CODE(!result, + "[GetPptLimit] get default PPT limit failed!", + return result); + hwmgr->power_limit = + hwmgr->default_power_limit = smum_get_argument(hwmgr); + return 0; } @@ -1770,14 +1794,14 @@ static int vega20_get_clock_ranges(struct pp_hwmgr *hwmgr, PPSMC_MSG_GetMaxDpmFreq, (clock_select << 16))) == 0, "[GetClockRanges] Failed to get max clock from SMC!", return ret); - vega20_read_arg_from_smc(hwmgr, clock); + *clock = smum_get_argument(hwmgr); } else { PP_ASSERT_WITH_CODE((ret = smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_GetMinDpmFreq, (clock_select << 16))) == 0, "[GetClockRanges] Failed to get min clock from SMC!", return ret); - vega20_read_arg_from_smc(hwmgr, clock); + *clock = smum_get_argument(hwmgr); } return 0; @@ -1841,7 +1865,7 @@ static int vega20_get_gpu_power(struct pp_hwmgr *hwmgr, int ret = 0; SmuMetrics_t metrics_table; - ret = vega20_copy_table_from_smc(hwmgr, (uint8_t *)&metrics_table, TABLE_SMU_METRICS); + ret = smum_smc_table_manager(hwmgr, (uint8_t *)&metrics_table, TABLE_SMU_METRICS, true); PP_ASSERT_WITH_CODE(!ret, "Failed to export SMU METRICS table!", return ret); @@ -1862,7 +1886,7 @@ static int vega20_get_current_gfx_clk_freq(struct pp_hwmgr *hwmgr, uint32_t *gfx PPSMC_MSG_GetDpmClockFreq, (PPCLK_GFXCLK << 16))) == 0, "[GetCurrentGfxClkFreq] Attempt to get Current GFXCLK Frequency Failed!", return ret); - vega20_read_arg_from_smc(hwmgr, &gfx_clk); + gfx_clk = smum_get_argument(hwmgr); *gfx_freq = gfx_clk * 100; @@ -1880,7 +1904,7 @@ static int vega20_get_current_mclk_freq(struct pp_hwmgr *hwmgr, uint32_t *mclk_f PPSMC_MSG_GetDpmClockFreq, (PPCLK_UCLK << 16))) == 0, "[GetCurrentMClkFreq] Attempt to get Current MCLK Frequency Failed!", return ret); - vega20_read_arg_from_smc(hwmgr, &mem_clk); + mem_clk = smum_get_argument(hwmgr); *mclk_freq = mem_clk * 100; @@ -1893,7 +1917,7 @@ static int vega20_get_current_activity_percent(struct pp_hwmgr *hwmgr, int ret = 0; SmuMetrics_t metrics_table; - ret = vega20_copy_table_from_smc(hwmgr, (uint8_t *)&metrics_table, TABLE_SMU_METRICS); + ret = smum_smc_table_manager(hwmgr, (uint8_t *)&metrics_table, TABLE_SMU_METRICS, true); PP_ASSERT_WITH_CODE(!ret, "Failed to export SMU METRICS table!", return ret); @@ -1907,6 +1931,8 @@ static int vega20_read_sensor(struct pp_hwmgr *hwmgr, int idx, void *value, int *size) { struct vega20_hwmgr *data = (struct vega20_hwmgr *)(hwmgr->backend); + struct amdgpu_device *adev = hwmgr->adev; + uint32_t val_vid; int ret = 0; switch (idx) { @@ -1941,6 +1967,18 @@ static int vega20_read_sensor(struct pp_hwmgr *hwmgr, int idx, *size = 16; ret = vega20_get_gpu_power(hwmgr, (uint32_t *)value); break; + case AMDGPU_PP_SENSOR_VDDGFX: + val_vid = (RREG32_SOC15(SMUIO, 0, mmSMUSVI0_TEL_PLANE0) & + SMUSVI0_TEL_PLANE0__SVI0_PLANE0_VDDCOR_MASK) >> + SMUSVI0_TEL_PLANE0__SVI0_PLANE0_VDDCOR__SHIFT; + *((uint32_t *)value) = + (uint32_t)convert_to_vddc((uint8_t)val_vid); + break; + case AMDGPU_PP_SENSOR_ENABLED_SMC_FEATURES_MASK: + ret = vega20_get_enabled_smc_features(hwmgr, (uint64_t *)value); + if (!ret) + *size = 8; + break; default: ret = -EINVAL; break; @@ -2264,6 +2302,25 @@ static uint32_t vega20_get_fan_control_mode(struct pp_hwmgr *hwmgr) return AMD_FAN_CTRL_AUTO; } +static void vega20_set_fan_control_mode(struct pp_hwmgr *hwmgr, uint32_t mode) +{ + switch (mode) { + case AMD_FAN_CTRL_NONE: + vega20_fan_ctrl_set_fan_speed_percent(hwmgr, 100); + break; + case AMD_FAN_CTRL_MANUAL: + if (PP_CAP(PHM_PlatformCaps_MicrocodeFanControl)) + vega20_fan_ctrl_stop_smc_fan_control(hwmgr); + break; + case AMD_FAN_CTRL_AUTO: + if (PP_CAP(PHM_PlatformCaps_MicrocodeFanControl)) + vega20_fan_ctrl_start_smc_fan_control(hwmgr); + break; + default: + break; + } +} + static int vega20_get_dal_power_level(struct pp_hwmgr *hwmgr, struct amd_pp_simple_clock_info *info) { @@ -2612,18 +2669,18 @@ static int vega20_odn_edit_dpm_table(struct pp_hwmgr *hwmgr, data->gfxclk_overdrive = false; data->memclk_overdrive = false; - ret = vega20_copy_table_from_smc(hwmgr, - (uint8_t *)od_table, - TABLE_OVERDRIVE); + ret = smum_smc_table_manager(hwmgr, + (uint8_t *)od_table, + TABLE_OVERDRIVE, true); PP_ASSERT_WITH_CODE(!ret, "Failed to export overdrive table!", return ret); break; case PP_OD_COMMIT_DPM_TABLE: - ret = vega20_copy_table_to_smc(hwmgr, - (uint8_t *)od_table, - TABLE_OVERDRIVE); + ret = smum_smc_table_manager(hwmgr, + (uint8_t *)od_table, + TABLE_OVERDRIVE, false); PP_ASSERT_WITH_CODE(!ret, "Failed to import overdrive table!", return ret); @@ -2847,8 +2904,8 @@ static int vega20_display_configuration_changed_task(struct pp_hwmgr *hwmgr) if ((data->water_marks_bitmap & WaterMarksExist) && !(data->water_marks_bitmap & WaterMarksLoaded)) { - result = vega20_copy_table_to_smc(hwmgr, - (uint8_t *)wm_table, TABLE_WATERMARKS); + result = smum_smc_table_manager(hwmgr, + (uint8_t *)wm_table, TABLE_WATERMARKS, false); PP_ASSERT_WITH_CODE(!result, "Failed to update WMTABLE!", return result); @@ -3118,6 +3175,34 @@ static int vega20_power_off_asic(struct pp_hwmgr *hwmgr) return result; } +static int conv_power_profile_to_pplib_workload(int power_profile) +{ + int pplib_workload = 0; + + switch (power_profile) { + case PP_SMC_POWER_PROFILE_FULLSCREEN3D: + pplib_workload = WORKLOAD_PPLIB_FULL_SCREEN_3D_BIT; + break; + case PP_SMC_POWER_PROFILE_POWERSAVING: + pplib_workload = WORKLOAD_PPLIB_POWER_SAVING_BIT; + break; + case PP_SMC_POWER_PROFILE_VIDEO: + pplib_workload = WORKLOAD_PPLIB_VIDEO_BIT; + break; + case PP_SMC_POWER_PROFILE_VR: + pplib_workload = WORKLOAD_PPLIB_VR_BIT; + break; + case PP_SMC_POWER_PROFILE_COMPUTE: + pplib_workload = WORKLOAD_PPLIB_COMPUTE_BIT; + break; + case PP_SMC_POWER_PROFILE_CUSTOM: + pplib_workload = WORKLOAD_PPLIB_CUSTOM_BIT; + break; + } + + return pplib_workload; +} + static int vega20_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf) { DpmActivityMonitorCoeffInt_t activity_monitor; @@ -3153,14 +3238,14 @@ static int vega20_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf) for (i = 0; i <= PP_SMC_POWER_PROFILE_CUSTOM; i++) { /* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */ - workload_type = i + 1; + workload_type = conv_power_profile_to_pplib_workload(i); result = vega20_get_activity_monitor_coeff(hwmgr, (uint8_t *)(&activity_monitor), workload_type); PP_ASSERT_WITH_CODE(!result, "[GetPowerProfile] Failed to get activity monitor!", return result); - size += sprintf(buf + size, "%2d(%14s%s)\n", + size += sprintf(buf + size, "%2d %14s%s:\n", i, profile_name[i], (i == hwmgr->power_profile_mode) ? "*" : " "); size += sprintf(buf + size, "%19s %d(%13s) %7d %7d %7d %7d %7d %7d %7d %7d %7d\n", @@ -3226,10 +3311,15 @@ static int vega20_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf) static int vega20_set_power_profile_mode(struct pp_hwmgr *hwmgr, long *input, uint32_t size) { DpmActivityMonitorCoeffInt_t activity_monitor; - int result = 0; + int workload_type, result = 0; hwmgr->power_profile_mode = input[size]; + if (hwmgr->power_profile_mode > PP_SMC_POWER_PROFILE_CUSTOM) { + pr_err("Invalid power profile mode %d\n", hwmgr->power_profile_mode); + return -EINVAL; + } + if (hwmgr->power_profile_mode == PP_SMC_POWER_PROFILE_CUSTOM) { if (size < 10) return -EINVAL; @@ -3296,8 +3386,11 @@ static int vega20_set_power_profile_mode(struct pp_hwmgr *hwmgr, long *input, ui return result); } + /* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */ + workload_type = + conv_power_profile_to_pplib_workload(hwmgr->power_profile_mode); smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetWorkloadMask, - 1 << hwmgr->power_profile_mode); + 1 << workload_type); return 0; } @@ -3427,15 +3520,25 @@ static const struct pp_hwmgr_func vega20_hwmgr_funcs = { .disable_smc_firmware_ctf = vega20_thermal_disable_alert, /* fan control related */ + .get_fan_speed_percent = + vega20_fan_ctrl_get_fan_speed_percent, + .set_fan_speed_percent = + vega20_fan_ctrl_set_fan_speed_percent, .get_fan_speed_info = vega20_fan_ctrl_get_fan_speed_info, .get_fan_speed_rpm = vega20_fan_ctrl_get_fan_speed_rpm, + .set_fan_speed_rpm = + vega20_fan_ctrl_set_fan_speed_rpm, .get_fan_control_mode = vega20_get_fan_control_mode, + .set_fan_control_mode = + vega20_set_fan_control_mode, /* smu memory related */ .notify_cac_buffer_info = vega20_notify_cac_buffer_info, + .enable_mgpu_fan_boost = + vega20_enable_mgpu_fan_boost, }; int vega20_hwmgr_init(struct pp_hwmgr *hwmgr) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_processpptables.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_processpptables.c index 5f1f7a32ac24..e5f7f8230065 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_processpptables.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_processpptables.c @@ -100,9 +100,8 @@ static void dump_pptable(PPTable_t *pptable) pr_info("PpmTemperatureThreshold = %d\n", pptable->PpmTemperatureThreshold); pr_info("MemoryOnPackage = 0x%02x\n", pptable->MemoryOnPackage); - pr_info("padding8_limits[0] = 0x%02x\n", pptable->padding8_limits[0]); - pr_info("padding8_limits[1] = 0x%02x\n", pptable->padding8_limits[1]); - pr_info("padding8_limits[2] = 0x%02x\n", pptable->padding8_limits[2]); + pr_info("padding8_limits = 0x%02x\n", pptable->padding8_limits); + pr_info("Tvr_SocLimit = %d\n", pptable->Tvr_SocLimit); pr_info("UlvVoltageOffsetSoc = %d\n", pptable->UlvVoltageOffsetSoc); pr_info("UlvVoltageOffsetGfx = %d\n", pptable->UlvVoltageOffsetGfx); @@ -417,8 +416,8 @@ static void dump_pptable(PPTable_t *pptable) pr_info("FanGainEdge = %d\n", pptable->FanGainEdge); pr_info("FanGainHotspot = %d\n", pptable->FanGainHotspot); pr_info("FanGainLiquid = %d\n", pptable->FanGainLiquid); - pr_info("FanGainVrVddc = %d\n", pptable->FanGainVrVddc); - pr_info("FanGainVrMvdd = %d\n", pptable->FanGainVrMvdd); + pr_info("FanGainVrGfx = %d\n", pptable->FanGainVrGfx); + pr_info("FanGainVrSoc = %d\n", pptable->FanGainVrSoc); pr_info("FanGainPlx = %d\n", pptable->FanGainPlx); pr_info("FanGainHbm = %d\n", pptable->FanGainHbm); pr_info("FanPwmMin = %d\n", pptable->FanPwmMin); @@ -533,23 +532,20 @@ static void dump_pptable(PPTable_t *pptable) pr_info("MinVoltageUlvGfx = %d\n", pptable->MinVoltageUlvGfx); pr_info("MinVoltageUlvSoc = %d\n", pptable->MinVoltageUlvSoc); - for (i = 0; i < 14; i++) - pr_info("Reserved[%d] = 0x%x\n", i, pptable->Reserved[i]); + pr_info("MGpuFanBoostLimitRpm = %d\n", pptable->MGpuFanBoostLimitRpm); + pr_info("padding16_Fan = %d\n", pptable->padding16_Fan); + + pr_info("FanGainVrMem0 = %d\n", pptable->FanGainVrMem0); + pr_info("FanGainVrMem0 = %d\n", pptable->FanGainVrMem0); - pr_info("Liquid1_I2C_address = 0x%x\n", pptable->Liquid1_I2C_address); - pr_info("Liquid2_I2C_address = 0x%x\n", pptable->Liquid2_I2C_address); - pr_info("Vr_I2C_address = 0x%x\n", pptable->Vr_I2C_address); - pr_info("Plx_I2C_address = 0x%x\n", pptable->Plx_I2C_address); + pr_info("DcBtcGb[AVFS_VOLTAGE_GFX] = 0x%x\n", pptable->DcBtcGb[AVFS_VOLTAGE_GFX]); + pr_info("DcBtcGb[AVFS_VOLTAGE_SOC] = 0x%x\n", pptable->DcBtcGb[AVFS_VOLTAGE_SOC]); - pr_info("Liquid_I2C_LineSCL = 0x%x\n", pptable->Liquid_I2C_LineSCL); - pr_info("Liquid_I2C_LineSDA = 0x%x\n", pptable->Liquid_I2C_LineSDA); - pr_info("Vr_I2C_LineSCL = 0x%x\n", pptable->Vr_I2C_LineSCL); - pr_info("Vr_I2C_LineSDA = 0x%x\n", pptable->Vr_I2C_LineSDA); + for (i = 0; i < 11; i++) + pr_info("Reserved[%d] = 0x%x\n", i, pptable->Reserved[i]); - pr_info("Plx_I2C_LineSCL = 0x%x\n", pptable->Plx_I2C_LineSCL); - pr_info("Plx_I2C_LineSDA = 0x%x\n", pptable->Plx_I2C_LineSDA); - pr_info("VrSensorPresent = 0x%x\n", pptable->VrSensorPresent); - pr_info("LiquidSensorPresent = 0x%x\n", pptable->LiquidSensorPresent); + for (i = 0; i < 3; i++) + pr_info("Padding32[%d] = 0x%x\n", i, pptable->Padding32[i]); pr_info("MaxVoltageStepGfx = 0x%x\n", pptable->MaxVoltageStepGfx); pr_info("MaxVoltageStepSoc = 0x%x\n", pptable->MaxVoltageStepSoc); @@ -611,6 +607,24 @@ static void dump_pptable(PPTable_t *pptable) pr_info("FllGfxclkSpreadPercent = %d\n", pptable->FllGfxclkSpreadPercent); pr_info("FllGfxclkSpreadFreq = %d\n", pptable->FllGfxclkSpreadFreq); + for (i = 0; i < I2C_CONTROLLER_NAME_COUNT; i++) { + pr_info("I2cControllers[%d]:\n", i); + pr_info(" .Enabled = %d\n", + pptable->I2cControllers[i].Enabled); + pr_info(" .SlaveAddress = 0x%x\n", + pptable->I2cControllers[i].SlaveAddress); + pr_info(" .ControllerPort = %d\n", + pptable->I2cControllers[i].ControllerPort); + pr_info(" .ControllerName = %d\n", + pptable->I2cControllers[i].ControllerName); + pr_info(" .ThermalThrottler = %d\n", + pptable->I2cControllers[i].ThermalThrottler); + pr_info(" .I2cProtocol = %d\n", + pptable->I2cControllers[i].I2cProtocol); + pr_info(" .I2cSpeed = %d\n", + pptable->I2cControllers[i].I2cSpeed); + } + for (i = 0; i < 10; i++) pr_info("BoardReserved[%d] = 0x%x\n", i, pptable->BoardReserved[i]); @@ -661,50 +675,6 @@ static int set_platform_caps(struct pp_hwmgr *hwmgr, uint32_t powerplay_caps) return 0; } -static int copy_clock_limits_array( - struct pp_hwmgr *hwmgr, - uint32_t **pptable_info_array, - const uint32_t *pptable_array, - uint32_t power_saving_clock_count) -{ - uint32_t array_size, i; - uint32_t *table; - - array_size = sizeof(uint32_t) * power_saving_clock_count; - table = kzalloc(array_size, GFP_KERNEL); - if (NULL == table) - return -ENOMEM; - - for (i = 0; i < power_saving_clock_count; i++) - table[i] = pptable_array[i]; - - *pptable_info_array = table; - - return 0; -} - -static int copy_overdrive_settings_limits_array( - struct pp_hwmgr *hwmgr, - uint32_t **pptable_info_array, - const uint32_t *pptable_array, - uint32_t od_setting_count) -{ - uint32_t array_size, i; - uint32_t *table; - - array_size = sizeof(uint32_t) * od_setting_count; - table = kzalloc(array_size, GFP_KERNEL); - if (NULL == table) - return -ENOMEM; - - for (i = 0; i < od_setting_count; i++) - table[i] = pptable_array[i]; - - *pptable_info_array = table; - - return 0; -} - static int copy_overdrive_feature_capabilities_array( struct pp_hwmgr *hwmgr, uint8_t **pptable_info_array, @@ -721,7 +691,7 @@ static int copy_overdrive_feature_capabilities_array( return -ENOMEM; for (i = 0; i < od_feature_count; i++) { - table[i] = pptable_array[i]; + table[i] = le32_to_cpu(pptable_array[i]); if (table[i]) od_supported = true; } @@ -737,29 +707,19 @@ static int copy_overdrive_feature_capabilities_array( static int append_vbios_pptable(struct pp_hwmgr *hwmgr, PPTable_t *ppsmc_pptable) { - struct atom_smc_dpm_info_v4_3 *smc_dpm_table; + struct atom_smc_dpm_info_v4_4 *smc_dpm_table; int index = GetIndexIntoMasterDataTable(smc_dpm_info); + int i; PP_ASSERT_WITH_CODE( smc_dpm_table = smu_atom_get_data_table(hwmgr->adev, index, NULL, NULL, NULL), "[appendVbiosPPTable] Failed to retrieve Smc Dpm Table from VBIOS!", return -1); - ppsmc_pptable->Liquid1_I2C_address = smc_dpm_table->liquid1_i2c_address; - ppsmc_pptable->Liquid2_I2C_address = smc_dpm_table->liquid2_i2c_address; - ppsmc_pptable->Vr_I2C_address = smc_dpm_table->vr_i2c_address; - ppsmc_pptable->Plx_I2C_address = smc_dpm_table->plx_i2c_address; - - ppsmc_pptable->Liquid_I2C_LineSCL = smc_dpm_table->liquid_i2c_linescl; - ppsmc_pptable->Liquid_I2C_LineSDA = smc_dpm_table->liquid_i2c_linesda; - ppsmc_pptable->Vr_I2C_LineSCL = smc_dpm_table->vr_i2c_linescl; - ppsmc_pptable->Vr_I2C_LineSDA = smc_dpm_table->vr_i2c_linesda; - - ppsmc_pptable->Plx_I2C_LineSCL = smc_dpm_table->plx_i2c_linescl; - ppsmc_pptable->Plx_I2C_LineSDA = smc_dpm_table->plx_i2c_linesda; - ppsmc_pptable->VrSensorPresent = smc_dpm_table->vrsensorpresent; - ppsmc_pptable->LiquidSensorPresent = smc_dpm_table->liquidsensorpresent; - + memset(ppsmc_pptable->Padding32, + 0, + sizeof(struct atom_smc_dpm_info_v4_4) - + sizeof(struct atom_common_table_header)); ppsmc_pptable->MaxVoltageStepGfx = smc_dpm_table->maxvoltagestepgfx; ppsmc_pptable->MaxVoltageStepSoc = smc_dpm_table->maxvoltagestepsoc; @@ -818,6 +778,24 @@ static int append_vbios_pptable(struct pp_hwmgr *hwmgr, PPTable_t *ppsmc_pptable ppsmc_pptable->FllGfxclkSpreadPercent = smc_dpm_table->fllgfxclkspreadpercent; ppsmc_pptable->FllGfxclkSpreadFreq = smc_dpm_table->fllgfxclkspreadfreq; + if ((smc_dpm_table->table_header.format_revision == 4) && + (smc_dpm_table->table_header.content_revision == 4)) { + for (i = 0; i < I2C_CONTROLLER_NAME_COUNT; i++) { + ppsmc_pptable->I2cControllers[i].Enabled = + smc_dpm_table->i2ccontrollers[i].enabled; + ppsmc_pptable->I2cControllers[i].SlaveAddress = + smc_dpm_table->i2ccontrollers[i].slaveaddress; + ppsmc_pptable->I2cControllers[i].ControllerPort = + smc_dpm_table->i2ccontrollers[i].controllerport; + ppsmc_pptable->I2cControllers[i].ThermalThrottler = + smc_dpm_table->i2ccontrollers[i].thermalthrottler; + ppsmc_pptable->I2cControllers[i].I2cProtocol = + smc_dpm_table->i2ccontrollers[i].i2cprotocol; + ppsmc_pptable->I2cControllers[i].I2cSpeed = + smc_dpm_table->i2ccontrollers[i].i2cspeed; + } + } + return 0; } @@ -834,6 +812,8 @@ static int init_powerplay_table_information( hwmgr->thermal_controller.ucType = powerplay_table->ucThermalControllerType; pptable_information->uc_thermal_controller_type = powerplay_table->ucThermalControllerType; + hwmgr->thermal_controller.fanInfo.ulMinRPM = 0; + hwmgr->thermal_controller.fanInfo.ulMaxRPM = powerplay_table->smcPPTable.FanMaximumRpm; set_hw_cap(hwmgr, ATOM_VEGA20_PP_THERMALCONTROLLER_NONE != hwmgr->thermal_controller.ucType, @@ -842,34 +822,40 @@ static int init_powerplay_table_information( phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_MicrocodeFanControl); if (powerplay_table->OverDrive8Table.ucODTableRevision == 1) { - od_feature_count = (powerplay_table->OverDrive8Table.ODFeatureCount > ATOM_VEGA20_ODFEATURE_COUNT) ? - ATOM_VEGA20_ODFEATURE_COUNT : powerplay_table->OverDrive8Table.ODFeatureCount; - od_setting_count = (powerplay_table->OverDrive8Table.ODSettingCount > ATOM_VEGA20_ODSETTING_COUNT) ? - ATOM_VEGA20_ODSETTING_COUNT : powerplay_table->OverDrive8Table.ODSettingCount; + od_feature_count = + (le32_to_cpu(powerplay_table->OverDrive8Table.ODFeatureCount) > + ATOM_VEGA20_ODFEATURE_COUNT) ? + ATOM_VEGA20_ODFEATURE_COUNT : + le32_to_cpu(powerplay_table->OverDrive8Table.ODFeatureCount); + od_setting_count = + (le32_to_cpu(powerplay_table->OverDrive8Table.ODSettingCount) > + ATOM_VEGA20_ODSETTING_COUNT) ? + ATOM_VEGA20_ODSETTING_COUNT : + le32_to_cpu(powerplay_table->OverDrive8Table.ODSettingCount); copy_overdrive_feature_capabilities_array(hwmgr, &pptable_information->od_feature_capabilities, powerplay_table->OverDrive8Table.ODFeatureCapabilities, od_feature_count); - copy_overdrive_settings_limits_array(hwmgr, + phm_copy_overdrive_settings_limits_array(hwmgr, &pptable_information->od_settings_max, powerplay_table->OverDrive8Table.ODSettingsMax, od_setting_count); - copy_overdrive_settings_limits_array(hwmgr, + phm_copy_overdrive_settings_limits_array(hwmgr, &pptable_information->od_settings_min, powerplay_table->OverDrive8Table.ODSettingsMin, od_setting_count); } - pptable_information->us_small_power_limit1 = powerplay_table->usSmallPowerLimit1; - pptable_information->us_small_power_limit2 = powerplay_table->usSmallPowerLimit2; - pptable_information->us_boost_power_limit = powerplay_table->usBoostPowerLimit; - pptable_information->us_od_turbo_power_limit = powerplay_table->usODTurboPowerLimit; - pptable_information->us_od_powersave_power_limit = powerplay_table->usODPowerSavePowerLimit; + pptable_information->us_small_power_limit1 = le16_to_cpu(powerplay_table->usSmallPowerLimit1); + pptable_information->us_small_power_limit2 = le16_to_cpu(powerplay_table->usSmallPowerLimit2); + pptable_information->us_boost_power_limit = le16_to_cpu(powerplay_table->usBoostPowerLimit); + pptable_information->us_od_turbo_power_limit = le16_to_cpu(powerplay_table->usODTurboPowerLimit); + pptable_information->us_od_powersave_power_limit = le16_to_cpu(powerplay_table->usODPowerSavePowerLimit); - pptable_information->us_software_shutdown_temp = powerplay_table->usSoftwareShutdownTemp; + pptable_information->us_software_shutdown_temp = le16_to_cpu(powerplay_table->usSoftwareShutdownTemp); - hwmgr->platform_descriptor.TDPODLimit = (uint16_t)powerplay_table->OverDrive8Table.ODSettingsMax[ATOM_VEGA20_ODSETTING_POWERPERCENTAGE]; + hwmgr->platform_descriptor.TDPODLimit = le32_to_cpu(powerplay_table->OverDrive8Table.ODSettingsMax[ATOM_VEGA20_ODSETTING_POWERPERCENTAGE]); disable_power_control = 0; if (!disable_power_control && hwmgr->platform_descriptor.TDPODLimit) @@ -877,13 +863,16 @@ static int init_powerplay_table_information( phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_PowerControl); if (powerplay_table->PowerSavingClockTable.ucTableRevision == 1) { - power_saving_clock_count = (powerplay_table->PowerSavingClockTable.PowerSavingClockCount >= ATOM_VEGA20_PPCLOCK_COUNT) ? - ATOM_VEGA20_PPCLOCK_COUNT : powerplay_table->PowerSavingClockTable.PowerSavingClockCount; - copy_clock_limits_array(hwmgr, + power_saving_clock_count = + (le32_to_cpu(powerplay_table->PowerSavingClockTable.PowerSavingClockCount) >= + ATOM_VEGA20_PPCLOCK_COUNT) ? + ATOM_VEGA20_PPCLOCK_COUNT : + le32_to_cpu(powerplay_table->PowerSavingClockTable.PowerSavingClockCount); + phm_copy_clock_limits_array(hwmgr, &pptable_information->power_saving_clock_max, powerplay_table->PowerSavingClockTable.PowerSavingClockMax, power_saving_clock_count); - copy_clock_limits_array(hwmgr, + phm_copy_clock_limits_array(hwmgr, &pptable_information->power_saving_clock_min, powerplay_table->PowerSavingClockTable.PowerSavingClockMin, power_saving_clock_count); @@ -893,7 +882,15 @@ static int init_powerplay_table_information( if (pptable_information->smc_pptable == NULL) return -ENOMEM; - memcpy(pptable_information->smc_pptable, &(powerplay_table->smcPPTable), sizeof(PPTable_t)); + if (powerplay_table->smcPPTable.Version <= 2) + memcpy(pptable_information->smc_pptable, + &(powerplay_table->smcPPTable), + sizeof(PPTable_t) - + sizeof(I2cControllerConfig_t) * I2C_CONTROLLER_NAME_COUNT); + else + memcpy(pptable_information->smc_pptable, + &(powerplay_table->smcPPTable), + sizeof(PPTable_t)); result = append_vbios_pptable(hwmgr, (pptable_information->smc_pptable)); diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_thermal.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_thermal.c index 2984ddd5428c..ede54e87e287 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_thermal.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_thermal.c @@ -29,6 +29,78 @@ #include "soc15_common.h" #include "pp_debug.h" +static int vega20_disable_fan_control_feature(struct pp_hwmgr *hwmgr) +{ + struct vega20_hwmgr *data = hwmgr->backend; + int ret = 0; + + if (data->smu_features[GNLD_FAN_CONTROL].supported) { + ret = vega20_enable_smc_features( + hwmgr, false, + data->smu_features[GNLD_FAN_CONTROL]. + smu_feature_bitmap); + PP_ASSERT_WITH_CODE(!ret, + "Disable FAN CONTROL feature Failed!", + return ret); + data->smu_features[GNLD_FAN_CONTROL].enabled = false; + } + + return ret; +} + +int vega20_fan_ctrl_stop_smc_fan_control(struct pp_hwmgr *hwmgr) +{ + struct vega20_hwmgr *data = hwmgr->backend; + + if (data->smu_features[GNLD_FAN_CONTROL].supported) + return vega20_disable_fan_control_feature(hwmgr); + + return 0; +} + +static int vega20_enable_fan_control_feature(struct pp_hwmgr *hwmgr) +{ + struct vega20_hwmgr *data = hwmgr->backend; + int ret = 0; + + if (data->smu_features[GNLD_FAN_CONTROL].supported) { + ret = vega20_enable_smc_features( + hwmgr, true, + data->smu_features[GNLD_FAN_CONTROL]. + smu_feature_bitmap); + PP_ASSERT_WITH_CODE(!ret, + "Enable FAN CONTROL feature Failed!", + return ret); + data->smu_features[GNLD_FAN_CONTROL].enabled = true; + } + + return ret; +} + +int vega20_fan_ctrl_start_smc_fan_control(struct pp_hwmgr *hwmgr) +{ + struct vega20_hwmgr *data = hwmgr->backend; + + if (data->smu_features[GNLD_FAN_CONTROL].supported) + return vega20_enable_fan_control_feature(hwmgr); + + return 0; +} + +static int vega20_fan_ctrl_set_static_mode(struct pp_hwmgr *hwmgr, uint32_t mode) +{ + struct amdgpu_device *adev = hwmgr->adev; + + WREG32_SOC15(THM, 0, mmCG_FDO_CTRL2, + REG_SET_FIELD(RREG32_SOC15(THM, 0, mmCG_FDO_CTRL2), + CG_FDO_CTRL2, TMIN, 0)); + WREG32_SOC15(THM, 0, mmCG_FDO_CTRL2, + REG_SET_FIELD(RREG32_SOC15(THM, 0, mmCG_FDO_CTRL2), + CG_FDO_CTRL2, FDO_PWM_MODE, mode)); + + return 0; +} + static int vega20_get_current_rpm(struct pp_hwmgr *hwmgr, uint32_t *current_rpm) { int ret = 0; @@ -37,20 +109,67 @@ static int vega20_get_current_rpm(struct pp_hwmgr *hwmgr, uint32_t *current_rpm) PPSMC_MSG_GetCurrentRpm)) == 0, "Attempt to get current RPM from SMC Failed!", return ret); - PP_ASSERT_WITH_CODE((ret = vega20_read_arg_from_smc(hwmgr, - current_rpm)) == 0, - "Attempt to read current RPM from SMC Failed!", - return ret); + *current_rpm = smum_get_argument(hwmgr); return 0; } +int vega20_fan_ctrl_get_fan_speed_percent(struct pp_hwmgr *hwmgr, + uint32_t *speed) +{ + struct vega20_hwmgr *data = (struct vega20_hwmgr *)(hwmgr->backend); + PPTable_t *pp_table = &(data->smc_state_table.pp_table); + uint32_t current_rpm, percent = 0; + int ret = 0; + + ret = vega20_get_current_rpm(hwmgr, ¤t_rpm); + if (ret) + return ret; + + percent = current_rpm * 100 / pp_table->FanMaximumRpm; + + *speed = percent > 100 ? 100 : percent; + + return 0; +} + +int vega20_fan_ctrl_set_fan_speed_percent(struct pp_hwmgr *hwmgr, + uint32_t speed) +{ + struct amdgpu_device *adev = hwmgr->adev; + uint32_t duty100; + uint32_t duty; + uint64_t tmp64; + + if (speed > 100) + speed = 100; + + if (PP_CAP(PHM_PlatformCaps_MicrocodeFanControl)) + vega20_fan_ctrl_stop_smc_fan_control(hwmgr); + + duty100 = REG_GET_FIELD(RREG32_SOC15(THM, 0, mmCG_FDO_CTRL1), + CG_FDO_CTRL1, FMAX_DUTY100); + + if (duty100 == 0) + return -EINVAL; + + tmp64 = (uint64_t)speed * duty100; + do_div(tmp64, 100); + duty = (uint32_t)tmp64; + + WREG32_SOC15(THM, 0, mmCG_FDO_CTRL0, + REG_SET_FIELD(RREG32_SOC15(THM, 0, mmCG_FDO_CTRL0), + CG_FDO_CTRL0, FDO_STATIC_DUTY, duty)); + + return vega20_fan_ctrl_set_static_mode(hwmgr, FDO_PWM_MODE_STATIC); +} + int vega20_fan_ctrl_get_fan_speed_info(struct pp_hwmgr *hwmgr, struct phm_fan_speed_info *fan_speed_info) { memset(fan_speed_info, 0, sizeof(*fan_speed_info)); - fan_speed_info->supports_percent_read = false; - fan_speed_info->supports_percent_write = false; + fan_speed_info->supports_percent_read = true; + fan_speed_info->supports_percent_write = true; fan_speed_info->supports_rpm_read = true; fan_speed_info->supports_rpm_write = true; @@ -64,6 +183,31 @@ int vega20_fan_ctrl_get_fan_speed_rpm(struct pp_hwmgr *hwmgr, uint32_t *speed) return vega20_get_current_rpm(hwmgr, speed); } +int vega20_fan_ctrl_set_fan_speed_rpm(struct pp_hwmgr *hwmgr, uint32_t speed) +{ + struct amdgpu_device *adev = hwmgr->adev; + uint32_t tach_period, crystal_clock_freq; + int result = 0; + + if (!speed) + return -EINVAL; + + if (PP_CAP(PHM_PlatformCaps_MicrocodeFanControl)) { + result = vega20_fan_ctrl_stop_smc_fan_control(hwmgr); + if (result) + return result; + } + + crystal_clock_freq = amdgpu_asic_get_xclk((struct amdgpu_device *)hwmgr->adev); + tach_period = 60 * crystal_clock_freq * 10000 / (8 * speed); + WREG32_SOC15(THM, 0, mmCG_TACH_CTRL, + REG_SET_FIELD(RREG32_SOC15(THM, 0, mmCG_TACH_CTRL), + CG_TACH_CTRL, TARGET_PERIOD, + tach_period)); + + return vega20_fan_ctrl_set_static_mode(hwmgr, FDO_PWM_MODE_STATIC_RPM); +} + /** * Reads the remote temperature from the SIslands thermal controller. * diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_thermal.h b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_thermal.h index 2a6d49fec4e0..2d1769bbd24e 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_thermal.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_thermal.h @@ -50,15 +50,22 @@ struct vega20_temperature { #define FDO_PWM_MODE_STATIC_RPM 5 extern int vega20_thermal_get_temperature(struct pp_hwmgr *hwmgr); -extern int vega20_thermal_stop_thermal_controller(struct pp_hwmgr *hwmgr); extern int vega20_fan_ctrl_get_fan_speed_info(struct pp_hwmgr *hwmgr, struct phm_fan_speed_info *fan_speed_info); -extern int vega20_fan_ctrl_reset_fan_speed_to_default(struct pp_hwmgr *hwmgr); extern int vega20_fan_ctrl_get_fan_speed_rpm(struct pp_hwmgr *hwmgr, uint32_t *speed); +extern int vega20_fan_ctrl_set_fan_speed_rpm(struct pp_hwmgr *hwmgr, + uint32_t speed); +extern int vega20_fan_ctrl_get_fan_speed_percent(struct pp_hwmgr *hwmgr, + uint32_t *speed); +extern int vega20_fan_ctrl_set_fan_speed_percent(struct pp_hwmgr *hwmgr, + uint32_t speed); +extern int vega20_fan_ctrl_stop_smc_fan_control(struct pp_hwmgr *hwmgr); +extern int vega20_fan_ctrl_start_smc_fan_control(struct pp_hwmgr *hwmgr); extern int vega20_thermal_disable_alert(struct pp_hwmgr *hwmgr); extern int vega20_start_thermal_controller(struct pp_hwmgr *hwmgr, struct PP_TemperatureRange *range); +extern int vega20_thermal_stop_thermal_controller(struct pp_hwmgr *hwmgr); #endif diff --git a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h index a6d92128b19c..e5a60aa44b5d 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h +++ b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h @@ -328,6 +328,8 @@ struct pp_hwmgr_func { int (*set_power_limit)(struct pp_hwmgr *hwmgr, uint32_t n); int (*powergate_mmhub)(struct pp_hwmgr *hwmgr); int (*smus_notify_pwe)(struct pp_hwmgr *hwmgr); + int (*powergate_sdma)(struct pp_hwmgr *hwmgr, bool bgate); + int (*enable_mgpu_fan_boost)(struct pp_hwmgr *hwmgr); }; struct pp_table_func { @@ -732,7 +734,6 @@ struct pp_hwmgr { void *smu_backend; const struct pp_smumgr_func *smumgr_funcs; bool is_kicker; - bool reload_fw; enum PP_DAL_POWERLEVEL dal_power_level; struct phm_dynamic_state_info dyn_state; diff --git a/drivers/gpu/drm/amd/powerplay/inc/smu11_driver_if.h b/drivers/gpu/drm/amd/powerplay/inc/smu11_driver_if.h index 71191deb4e76..2998a49960ed 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/smu11_driver_if.h +++ b/drivers/gpu/drm/amd/powerplay/inc/smu11_driver_if.h @@ -27,7 +27,7 @@ // *** IMPORTANT *** // SMU TEAM: Always increment the interface version if // any structure is changed in this file -#define SMU11_DRIVER_IF_VERSION 0x11 +#define SMU11_DRIVER_IF_VERSION 0x12 #define PPTABLE_V20_SMU_VERSION 2 @@ -165,7 +165,7 @@ #define FEATURE_DS_FCLK_MASK (1 << FEATURE_DS_FCLK_BIT ) #define FEATURE_DS_MP1CLK_MASK (1 << FEATURE_DS_MP1CLK_BIT ) #define FEATURE_DS_MP0CLK_MASK (1 << FEATURE_DS_MP0CLK_BIT ) - +#define FEATURE_XGMI_MASK (1 << FEATURE_XGMI_BIT ) #define DPM_OVERRIDE_DISABLE_SOCCLK_PID 0x00000001 #define DPM_OVERRIDE_DISABLE_UCLK_PID 0x00000002 @@ -186,6 +186,9 @@ #define DPM_OVERRIDE_ENABLE_GFXOFF_UCLK_SWITCH 0x00010000 #define DPM_OVERRIDE_ENABLE_GFXOFF_FCLK_SWITCH 0x00020000 +#define I2C_CONTROLLER_ENABLED 1 +#define I2C_CONTROLLER_DISABLED 0 + #define VR_MAPPING_VR_SELECT_MASK 0x01 #define VR_MAPPING_VR_SELECT_SHIFT 0x00 @@ -208,15 +211,17 @@ #define THROTTLER_STATUS_TEMP_HOTSPOT_BIT 2 #define THROTTLER_STATUS_TEMP_HBM_BIT 3 #define THROTTLER_STATUS_TEMP_VR_GFX_BIT 4 -#define THROTTLER_STATUS_TEMP_VR_MEM_BIT 5 -#define THROTTLER_STATUS_TEMP_LIQUID_BIT 6 -#define THROTTLER_STATUS_TEMP_PLX_BIT 7 -#define THROTTLER_STATUS_TEMP_SKIN_BIT 8 -#define THROTTLER_STATUS_TDC_GFX_BIT 9 -#define THROTTLER_STATUS_TDC_SOC_BIT 10 -#define THROTTLER_STATUS_PPT_BIT 11 -#define THROTTLER_STATUS_FIT_BIT 12 -#define THROTTLER_STATUS_PPM_BIT 13 +#define THROTTLER_STATUS_TEMP_VR_SOC_BIT 5 +#define THROTTLER_STATUS_TEMP_VR_MEM0_BIT 6 +#define THROTTLER_STATUS_TEMP_VR_MEM1_BIT 7 +#define THROTTLER_STATUS_TEMP_LIQUID_BIT 8 +#define THROTTLER_STATUS_TEMP_PLX_BIT 9 +#define THROTTLER_STATUS_TEMP_SKIN_BIT 10 +#define THROTTLER_STATUS_TDC_GFX_BIT 11 +#define THROTTLER_STATUS_TDC_SOC_BIT 12 +#define THROTTLER_STATUS_PPT_BIT 13 +#define THROTTLER_STATUS_FIT_BIT 14 +#define THROTTLER_STATUS_PPM_BIT 15 #define TABLE_TRANSFER_OK 0x0 @@ -236,6 +241,58 @@ #define XGMI_STATE_D0 1 #define XGMI_STATE_D3 0 +typedef enum { + I2C_CONTROLLER_PORT_0 = 0, + I2C_CONTROLLER_PORT_1 = 1, +} I2cControllerPort_e; + +typedef enum { + I2C_CONTROLLER_NAME_VR_GFX = 0, + I2C_CONTROLLER_NAME_VR_SOC, + I2C_CONTROLLER_NAME_VR_VDDCI, + I2C_CONTROLLER_NAME_VR_HBM, + I2C_CONTROLLER_NAME_LIQUID_0, + I2C_CONTROLLER_NAME_LIQUID_1, + I2C_CONTROLLER_NAME_PLX, + I2C_CONTROLLER_NAME_COUNT, +} I2cControllerName_e; + +typedef enum { + I2C_CONTROLLER_THROTTLER_TYPE_NONE = 0, + I2C_CONTROLLER_THROTTLER_VR_GFX, + I2C_CONTROLLER_THROTTLER_VR_SOC, + I2C_CONTROLLER_THROTTLER_VR_VDDCI, + I2C_CONTROLLER_THROTTLER_VR_HBM, + I2C_CONTROLLER_THROTTLER_LIQUID_0, + I2C_CONTROLLER_THROTTLER_LIQUID_1, + I2C_CONTROLLER_THROTTLER_PLX, +} I2cControllerThrottler_e; + +typedef enum { + I2C_CONTROLLER_PROTOCOL_VR_XPDE132G5, + I2C_CONTROLLER_PROTOCOL_VR_IR35217, + I2C_CONTROLLER_PROTOCOL_TMP_TMP102A, + I2C_CONTROLLER_PROTOCOL_SPARE_0, + I2C_CONTROLLER_PROTOCOL_SPARE_1, + I2C_CONTROLLER_PROTOCOL_SPARE_2, +} I2cControllerProtocol_e; + +typedef enum { + I2C_CONTROLLER_SPEED_SLOW = 0, + I2C_CONTROLLER_SPEED_FAST = 1, +} I2cControllerSpeed_e; + +typedef struct { + uint32_t Enabled; + uint32_t SlaveAddress; + uint32_t ControllerPort; + uint32_t ControllerName; + + uint32_t ThermalThrottler; + uint32_t I2cProtocol; + uint32_t I2cSpeed; +} I2cControllerConfig_t; + typedef struct { uint32_t a; uint32_t b; @@ -269,6 +326,12 @@ typedef enum { } PPCLK_e; typedef enum { + POWER_SOURCE_AC, + POWER_SOURCE_DC, + POWER_SOURCE_COUNT, +} POWER_SOURCE_e; + +typedef enum { VOLTAGE_MODE_AVFS = 0, VOLTAGE_MODE_AVFS_SS, VOLTAGE_MODE_SS, @@ -328,8 +391,8 @@ typedef struct { uint16_t PpmTemperatureThreshold; uint8_t MemoryOnPackage; - uint8_t padding8_limits[3]; - + uint8_t padding8_limits; + uint16_t Tvr_SocLimit; uint16_t UlvVoltageOffsetSoc; uint16_t UlvVoltageOffsetGfx; @@ -400,8 +463,8 @@ typedef struct { uint16_t FanGainEdge; uint16_t FanGainHotspot; uint16_t FanGainLiquid; - uint16_t FanGainVrVddc; - uint16_t FanGainVrMvdd; + uint16_t FanGainVrGfx; + uint16_t FanGainVrSoc; uint16_t FanGainPlx; uint16_t FanGainHbm; uint16_t FanPwmMin; @@ -438,7 +501,7 @@ typedef struct { uint8_t DcBtcEnabled[AVFS_VOLTAGE_COUNT]; uint8_t Padding8_GfxBtc[2]; - uint16_t DcBtcMin[AVFS_VOLTAGE_COUNT]; + int16_t DcBtcMin[AVFS_VOLTAGE_COUNT]; uint16_t DcBtcMax[AVFS_VOLTAGE_COUNT]; @@ -461,24 +524,14 @@ typedef struct { uint16_t MGpuFanBoostLimitRpm; uint16_t padding16_Fan; - uint32_t Reserved[13]; - + uint16_t FanGainVrMem0; + uint16_t FanGainVrMem1; + uint16_t DcBtcGb[AVFS_VOLTAGE_COUNT]; - uint8_t Liquid1_I2C_address; - uint8_t Liquid2_I2C_address; - uint8_t Vr_I2C_address; - uint8_t Plx_I2C_address; + uint32_t Reserved[11]; - uint8_t Liquid_I2C_LineSCL; - uint8_t Liquid_I2C_LineSDA; - uint8_t Vr_I2C_LineSCL; - uint8_t Vr_I2C_LineSDA; - - uint8_t Plx_I2C_LineSCL; - uint8_t Plx_I2C_LineSDA; - uint8_t VrSensorPresent; - uint8_t LiquidSensorPresent; + uint32_t Padding32[3]; uint16_t MaxVoltageStepGfx; uint16_t MaxVoltageStepSoc; @@ -545,6 +598,8 @@ typedef struct { uint8_t FllGfxclkSpreadPercent; uint16_t FllGfxclkSpreadFreq; + I2cControllerConfig_t I2cControllers[I2C_CONTROLLER_NAME_COUNT]; + uint32_t BoardReserved[10]; @@ -601,7 +656,9 @@ typedef struct { uint16_t TemperatureHotspot ; uint16_t TemperatureHBM ; uint16_t TemperatureVrGfx ; - uint16_t TemperatureVrMem ; + uint16_t TemperatureVrSoc ; + uint16_t TemperatureVrMem0 ; + uint16_t TemperatureVrMem1 ; uint16_t TemperatureLiquid ; uint16_t TemperaturePlx ; uint32_t ThrottlerStatus ; diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/ci_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/ci_smumgr.c index 18643e06bc6f..669bd0c2a16c 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/ci_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/ci_smumgr.c @@ -2269,11 +2269,13 @@ static uint32_t ci_get_offsetof(uint32_t type, uint32_t member) case DRAM_LOG_BUFF_SIZE: return offsetof(SMU7_SoftRegisters, DRAM_LOG_BUFF_SIZE); } + break; case SMU_Discrete_DpmTable: switch (member) { case LowSclkInterruptThreshold: return offsetof(SMU7_Discrete_DpmTable, LowSclkInterruptT); } + break; } pr_debug("can't get the offset of type %x member %x\n", type, member); return 0; diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c index ec14798e87b6..bc8375cbf297 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c @@ -302,16 +302,6 @@ static int fiji_start_smu(struct pp_hwmgr *hwmgr) hwmgr->avfs_supported = false; } - /* To initialize all clock gating before RLC loaded and running.*/ - amdgpu_device_ip_set_clockgating_state(hwmgr->adev, - AMD_IP_BLOCK_TYPE_GFX, AMD_CG_STATE_GATE); - amdgpu_device_ip_set_clockgating_state(hwmgr->adev, - AMD_IP_BLOCK_TYPE_GMC, AMD_CG_STATE_GATE); - amdgpu_device_ip_set_clockgating_state(hwmgr->adev, - AMD_IP_BLOCK_TYPE_SDMA, AMD_CG_STATE_GATE); - amdgpu_device_ip_set_clockgating_state(hwmgr->adev, - AMD_IP_BLOCK_TYPE_COMMON, AMD_CG_STATE_GATE); - /* Setup SoftRegsStart here for register lookup in case * DummyBackEnd is used and ProcessFirmwareHeader is not executed */ @@ -2331,6 +2321,7 @@ static uint32_t fiji_get_offsetof(uint32_t type, uint32_t member) case DRAM_LOG_BUFF_SIZE: return offsetof(SMU73_SoftRegisters, DRAM_LOG_BUFF_SIZE); } + break; case SMU_Discrete_DpmTable: switch (member) { case UvdBootLevel: @@ -2340,6 +2331,7 @@ static uint32_t fiji_get_offsetof(uint32_t type, uint32_t member) case LowSclkInterruptThreshold: return offsetof(SMU73_Discrete_DpmTable, LowSclkInterruptThreshold); } + break; } pr_warn("can't get the offset of type %x member %x\n", type, member); return 0; diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/iceland_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/iceland_smumgr.c index 73aa368a454e..375ccf6ff5f2 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/iceland_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/iceland_smumgr.c @@ -232,26 +232,25 @@ static int iceland_request_smu_load_specific_fw(struct pp_hwmgr *hwmgr, static int iceland_start_smu(struct pp_hwmgr *hwmgr) { + struct iceland_smumgr *priv = hwmgr->smu_backend; int result; - result = iceland_smu_upload_firmware_image(hwmgr); - if (result) - return result; - result = iceland_smu_start_smc(hwmgr); - if (result) - return result; - if (!smu7_is_smc_ram_running(hwmgr)) { - pr_info("smu not running, upload firmware again \n"); result = iceland_smu_upload_firmware_image(hwmgr); if (result) return result; - result = iceland_smu_start_smc(hwmgr); - if (result) - return result; + iceland_smu_start_smc(hwmgr); } + /* Setup SoftRegsStart here to visit the register UcodeLoadStatus + * to check fw loading state + */ + smu7_read_smc_sram_dword(hwmgr, + SMU71_FIRMWARE_HEADER_LOCATION + + offsetof(SMU71_Firmware_Header, SoftRegisters), + &(priv->smu7_data.soft_regs_start), 0x40000); + result = smu7_request_smu_load_fw(hwmgr); return result; @@ -2237,11 +2236,13 @@ static uint32_t iceland_get_offsetof(uint32_t type, uint32_t member) case DRAM_LOG_BUFF_SIZE: return offsetof(SMU71_SoftRegisters, DRAM_LOG_BUFF_SIZE); } + break; case SMU_Discrete_DpmTable: switch (member) { case LowSclkInterruptThreshold: return offsetof(SMU71_Discrete_DpmTable, LowSclkInterruptThreshold); } + break; } pr_warn("can't get the offset of type %x member %x\n", type, member); return 0; @@ -2662,7 +2663,7 @@ const struct pp_smumgr_func iceland_smu_funcs = { .smu_fini = &smu7_smu_fini, .start_smu = &iceland_start_smu, .check_fw_load_finish = &smu7_check_fw_load_finish, - .request_smu_load_fw = &smu7_reload_firmware, + .request_smu_load_fw = &smu7_request_smu_load_fw, .request_smu_load_specific_fw = &iceland_request_smu_load_specific_fw, .send_msg_to_smc = &smu7_send_msg_to_smc, .send_msg_to_smc_with_parameter = &smu7_send_msg_to_smc_with_parameter, diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/smu10_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/smu10_smumgr.c index 6f961dec2088..d0eb8ab50148 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/smu10_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/smu10_smumgr.c @@ -186,40 +186,12 @@ static int smu10_verify_smc_interface(struct pp_hwmgr *hwmgr) return 0; } -/* sdma is disabled by default in vbios, need to re-enable in driver */ -static void smu10_smc_enable_sdma(struct pp_hwmgr *hwmgr) -{ - smu10_send_msg_to_smc(hwmgr, - PPSMC_MSG_PowerUpSdma); -} - -static void smu10_smc_disable_sdma(struct pp_hwmgr *hwmgr) -{ - smu10_send_msg_to_smc(hwmgr, - PPSMC_MSG_PowerDownSdma); -} - -/* vcn is disabled by default in vbios, need to re-enable in driver */ -static void smu10_smc_enable_vcn(struct pp_hwmgr *hwmgr) -{ - smu10_send_msg_to_smc_with_parameter(hwmgr, - PPSMC_MSG_PowerUpVcn, 0); -} - -static void smu10_smc_disable_vcn(struct pp_hwmgr *hwmgr) -{ - smu10_send_msg_to_smc_with_parameter(hwmgr, - PPSMC_MSG_PowerDownVcn, 0); -} - static int smu10_smu_fini(struct pp_hwmgr *hwmgr) { struct smu10_smumgr *priv = (struct smu10_smumgr *)(hwmgr->smu_backend); if (priv) { - smu10_smc_disable_sdma(hwmgr); - smu10_smc_disable_vcn(hwmgr); amdgpu_bo_free_kernel(&priv->smu_tables.entry[SMU10_WMTABLE].handle, &priv->smu_tables.entry[SMU10_WMTABLE].mc_addr, &priv->smu_tables.entry[SMU10_WMTABLE].table); @@ -243,8 +215,7 @@ static int smu10_start_smu(struct pp_hwmgr *hwmgr) if (smu10_verify_smc_interface(hwmgr)) return -EINVAL; - smu10_smc_enable_sdma(hwmgr); - smu10_smc_enable_vcn(hwmgr); + return 0; } diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/smu7_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/smu7_smumgr.c index 186dafc7f166..3f51d545e8ff 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/smu7_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/smu7_smumgr.c @@ -302,44 +302,6 @@ int smu7_write_smc_sram_dword(struct pp_hwmgr *hwmgr, uint32_t smc_addr, uint32_ return 0; } -/* Convert the firmware type to SMU type mask. For MEC, we need to check all MEC related type */ - -static uint32_t smu7_get_mask_for_firmware_type(uint32_t fw_type) -{ - uint32_t result = 0; - - switch (fw_type) { - case UCODE_ID_SDMA0: - result = UCODE_ID_SDMA0_MASK; - break; - case UCODE_ID_SDMA1: - result = UCODE_ID_SDMA1_MASK; - break; - case UCODE_ID_CP_CE: - result = UCODE_ID_CP_CE_MASK; - break; - case UCODE_ID_CP_PFP: - result = UCODE_ID_CP_PFP_MASK; - break; - case UCODE_ID_CP_ME: - result = UCODE_ID_CP_ME_MASK; - break; - case UCODE_ID_CP_MEC: - case UCODE_ID_CP_MEC_JT1: - case UCODE_ID_CP_MEC_JT2: - result = UCODE_ID_CP_MEC_MASK; - break; - case UCODE_ID_RLC_G: - result = UCODE_ID_RLC_G_MASK; - break; - default: - pr_info("UCode type is out of range! \n"); - result = 0; - } - - return result; -} - static int smu7_populate_single_firmware_entry(struct pp_hwmgr *hwmgr, uint32_t fw_type, struct SMU_Entry *entry) @@ -381,10 +343,7 @@ int smu7_request_smu_load_fw(struct pp_hwmgr *hwmgr) uint32_t fw_to_load; int r = 0; - if (!hwmgr->reload_fw) { - pr_info("skip reloading...\n"); - return 0; - } + amdgpu_ucode_init_bo(hwmgr->adev); if (smu_data->soft_regs_start) cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, @@ -467,10 +426,13 @@ int smu7_request_smu_load_fw(struct pp_hwmgr *hwmgr) smu7_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_DRV_DRAM_ADDR_HI, upper_32_bits(smu_data->header_buffer.mc_addr)); smu7_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_DRV_DRAM_ADDR_LO, lower_32_bits(smu_data->header_buffer.mc_addr)); - if (smu7_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_LoadUcodes, fw_to_load)) - pr_err("Fail to Request SMU Load uCode"); + smu7_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_LoadUcodes, fw_to_load); - return r; + r = smu7_check_fw_load_finish(hwmgr, fw_to_load); + if (!r) + return 0; + + pr_err("SMU load firmware failed\n"); failed: kfree(smu_data->toc); @@ -482,13 +444,12 @@ failed: int smu7_check_fw_load_finish(struct pp_hwmgr *hwmgr, uint32_t fw_type) { struct smu7_smumgr *smu_data = (struct smu7_smumgr *)(hwmgr->smu_backend); - uint32_t fw_mask = smu7_get_mask_for_firmware_type(fw_type); uint32_t ret; ret = phm_wait_on_indirect_register(hwmgr, mmSMC_IND_INDEX_11, smu_data->soft_regs_start + smum_get_offsetof(hwmgr, SMU_SoftRegisters, UcodeLoadStatus), - fw_mask, fw_mask); + fw_type, fw_type); return ret; } diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/smu8_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/smu8_smumgr.c index f7e3bc22bb93..f836d30fdd44 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/smu8_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/smu8_smumgr.c @@ -658,11 +658,10 @@ static int smu8_request_smu_load_fw(struct pp_hwmgr *hwmgr) { struct smu8_smumgr *smu8_smu = hwmgr->smu_backend; uint32_t smc_address; + uint32_t fw_to_check = 0; + int ret; - if (!hwmgr->reload_fw) { - pr_info("skip reloading...\n"); - return 0; - } + amdgpu_ucode_init_bo(hwmgr->adev); smu8_smu_populate_firmware_entries(hwmgr); @@ -689,28 +688,9 @@ static int smu8_request_smu_load_fw(struct pp_hwmgr *hwmgr) smu8_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_ExecuteJob, smu8_smu->toc_entry_power_profiling_index); - return smu8_send_msg_to_smc_with_parameter(hwmgr, + smu8_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_ExecuteJob, smu8_smu->toc_entry_initialize_index); -} - -static int smu8_start_smu(struct pp_hwmgr *hwmgr) -{ - int ret = 0; - uint32_t fw_to_check = 0; - struct amdgpu_device *adev = hwmgr->adev; - - uint32_t index = SMN_MP1_SRAM_START_ADDR + - SMU8_FIRMWARE_HEADER_LOCATION + - offsetof(struct SMU8_Firmware_Header, Version); - - - if (hwmgr == NULL || hwmgr->device == NULL) - return -EINVAL; - - cgs_write_register(hwmgr->device, mmMP0PUB_IND_INDEX, index); - hwmgr->smu_version = cgs_read_register(hwmgr->device, mmMP0PUB_IND_DATA); - adev->pm.fw_version = hwmgr->smu_version >> 8; fw_to_check = UCODE_ID_RLC_G_MASK | UCODE_ID_SDMA0_MASK | @@ -724,17 +704,38 @@ static int smu8_start_smu(struct pp_hwmgr *hwmgr) if (hwmgr->chip_id == CHIP_STONEY) fw_to_check &= ~(UCODE_ID_SDMA1_MASK | UCODE_ID_CP_MEC_JT2_MASK); - ret = smu8_request_smu_load_fw(hwmgr); - if (ret) + ret = smu8_check_fw_load_finish(hwmgr, fw_to_check); + if (ret) { pr_err("SMU firmware load failed\n"); - - smu8_check_fw_load_finish(hwmgr, fw_to_check); + return ret; + } ret = smu8_load_mec_firmware(hwmgr); - if (ret) + if (ret) { pr_err("Mec Firmware load failed\n"); + return ret; + } - return ret; + return 0; +} + +static int smu8_start_smu(struct pp_hwmgr *hwmgr) +{ + struct amdgpu_device *adev = hwmgr->adev; + + uint32_t index = SMN_MP1_SRAM_START_ADDR + + SMU8_FIRMWARE_HEADER_LOCATION + + offsetof(struct SMU8_Firmware_Header, Version); + + + if (hwmgr == NULL || hwmgr->device == NULL) + return -EINVAL; + + cgs_write_register(hwmgr->device, mmMP0PUB_IND_INDEX, index); + hwmgr->smu_version = cgs_read_register(hwmgr->device, mmMP0PUB_IND_DATA); + adev->pm.fw_version = hwmgr->smu_version >> 8; + + return smu8_request_smu_load_fw(hwmgr); } static int smu8_smu_init(struct pp_hwmgr *hwmgr) diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/tonga_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/tonga_smumgr.c index ae8378ed32ee..3ed6c5f1e5cf 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/tonga_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/tonga_smumgr.c @@ -192,6 +192,7 @@ static int tonga_start_in_non_protection_mode(struct pp_hwmgr *hwmgr) static int tonga_start_smu(struct pp_hwmgr *hwmgr) { + struct tonga_smumgr *priv = hwmgr->smu_backend; int result; /* Only start SMC if SMC RAM is not running */ @@ -209,6 +210,14 @@ static int tonga_start_smu(struct pp_hwmgr *hwmgr) } } + /* Setup SoftRegsStart here to visit the register UcodeLoadStatus + * to check fw loading state + */ + smu7_read_smc_sram_dword(hwmgr, + SMU72_FIRMWARE_HEADER_LOCATION + + offsetof(SMU72_Firmware_Header, SoftRegisters), + &(priv->smu7_data.soft_regs_start), 0x40000); + result = smu7_request_smu_load_fw(hwmgr); return result; @@ -2619,6 +2628,7 @@ static uint32_t tonga_get_offsetof(uint32_t type, uint32_t member) case DRAM_LOG_BUFF_SIZE: return offsetof(SMU72_SoftRegisters, DRAM_LOG_BUFF_SIZE); } + break; case SMU_Discrete_DpmTable: switch (member) { case UvdBootLevel: @@ -2628,6 +2638,7 @@ static uint32_t tonga_get_offsetof(uint32_t type, uint32_t member) case LowSclkInterruptThreshold: return offsetof(SMU72_Discrete_DpmTable, LowSclkInterruptThreshold); } + break; } pr_warn("can't get the offset of type %x member %x\n", type, member); return 0; diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/vega10_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/vega10_smumgr.c index 5d19115f410c..c81acc3192ad 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/vega10_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/vega10_smumgr.c @@ -88,8 +88,18 @@ static int vega10_copy_table_to_smc(struct pp_hwmgr *hwmgr, return 0; } -static int vega10_get_smc_features(struct pp_hwmgr *hwmgr, - uint32_t *features_enabled) +int vega10_enable_smc_features(struct pp_hwmgr *hwmgr, + bool enable, uint32_t feature_mask) +{ + int msg = enable ? PPSMC_MSG_EnableSmuFeatures : + PPSMC_MSG_DisableSmuFeatures; + + return smum_send_msg_to_smc_with_parameter(hwmgr, + msg, feature_mask); +} + +int vega10_get_enabled_smc_features(struct pp_hwmgr *hwmgr, + uint64_t *features_enabled) { if (features_enabled == NULL) return -EINVAL; @@ -102,9 +112,9 @@ static int vega10_get_smc_features(struct pp_hwmgr *hwmgr, static bool vega10_is_dpm_running(struct pp_hwmgr *hwmgr) { - uint32_t features_enabled = 0; + uint64_t features_enabled = 0; - vega10_get_smc_features(hwmgr, &features_enabled); + vega10_get_enabled_smc_features(hwmgr, &features_enabled); if (features_enabled & SMC_DPM_FEATURES) return true; diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/vega10_smumgr.h b/drivers/gpu/drm/amd/powerplay/smumgr/vega10_smumgr.h index 424e868bc768..bad760f22624 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/vega10_smumgr.h +++ b/drivers/gpu/drm/amd/powerplay/smumgr/vega10_smumgr.h @@ -42,6 +42,10 @@ struct vega10_smumgr { struct smu_table_array smu_tables; }; +int vega10_enable_smc_features(struct pp_hwmgr *hwmgr, + bool enable, uint32_t feature_mask); +int vega10_get_enabled_smc_features(struct pp_hwmgr *hwmgr, + uint64_t *features_enabled); #endif diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/vega12_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/vega12_smumgr.c index 7f0e2109f40d..ddb801517667 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/vega12_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/vega12_smumgr.c @@ -37,8 +37,8 @@ * @param hwmgr the address of the HW manager * @param table_id the driver's table ID to copy from */ -int vega12_copy_table_from_smc(struct pp_hwmgr *hwmgr, - uint8_t *table, int16_t table_id) +static int vega12_copy_table_from_smc(struct pp_hwmgr *hwmgr, + uint8_t *table, int16_t table_id) { struct vega12_smumgr *priv = (struct vega12_smumgr *)(hwmgr->smu_backend); @@ -75,8 +75,8 @@ int vega12_copy_table_from_smc(struct pp_hwmgr *hwmgr, * @param hwmgr the address of the HW manager * @param table_id the table to copy from */ -int vega12_copy_table_to_smc(struct pp_hwmgr *hwmgr, - uint8_t *table, int16_t table_id) +static int vega12_copy_table_to_smc(struct pp_hwmgr *hwmgr, + uint8_t *table, int16_t table_id) { struct vega12_smumgr *priv = (struct vega12_smumgr *)(hwmgr->smu_backend); @@ -351,6 +351,19 @@ static int vega12_start_smu(struct pp_hwmgr *hwmgr) return 0; } +static int vega12_smc_table_manager(struct pp_hwmgr *hwmgr, uint8_t *table, + uint16_t table_id, bool rw) +{ + int ret; + + if (rw) + ret = vega12_copy_table_from_smc(hwmgr, table, table_id); + else + ret = vega12_copy_table_to_smc(hwmgr, table, table_id); + + return ret; +} + const struct pp_smumgr_func vega12_smu_funcs = { .smu_init = &vega12_smu_init, .smu_fini = &vega12_smu_fini, @@ -362,4 +375,5 @@ const struct pp_smumgr_func vega12_smu_funcs = { .upload_pptable_settings = NULL, .is_dpm_running = vega12_is_dpm_running, .get_argument = smu9_get_argument, + .smc_table_manager = vega12_smc_table_manager, }; diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/vega12_smumgr.h b/drivers/gpu/drm/amd/powerplay/smumgr/vega12_smumgr.h index b285cbc04019..aeec965ce81f 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/vega12_smumgr.h +++ b/drivers/gpu/drm/amd/powerplay/smumgr/vega12_smumgr.h @@ -48,10 +48,6 @@ struct vega12_smumgr { #define SMU_FEATURES_HIGH_MASK 0xFFFFFFFF00000000 #define SMU_FEATURES_HIGH_SHIFT 32 -int vega12_copy_table_from_smc(struct pp_hwmgr *hwmgr, - uint8_t *table, int16_t table_id); -int vega12_copy_table_to_smc(struct pp_hwmgr *hwmgr, - uint8_t *table, int16_t table_id); int vega12_enable_smc_features(struct pp_hwmgr *hwmgr, bool enable, uint64_t feature_mask); int vega12_get_enabled_smc_features(struct pp_hwmgr *hwmgr, diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/vega20_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/vega20_smumgr.c index fe7f71079e0e..b7ff7d4d6f44 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/vega20_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/vega20_smumgr.c @@ -148,19 +148,11 @@ static int vega20_send_msg_to_smc_with_parameter(struct pp_hwmgr *hwmgr, return (ret == PPSMC_Result_OK) ? 0 : -EIO; } -/* - * Retrieve an argument from SMC. - * @param hwmgr the address of the powerplay hardware manager. - * @param arg pointer to store the argument from SMC. - * @return Always return 0. - */ -int vega20_read_arg_from_smc(struct pp_hwmgr *hwmgr, uint32_t *arg) +static uint32_t vega20_get_argument(struct pp_hwmgr *hwmgr) { struct amdgpu_device *adev = hwmgr->adev; - *arg = RREG32_SOC15(MP1, 0, mmMP1_SMN_C2PMSG_82); - - return 0; + return RREG32_SOC15(MP1, 0, mmMP1_SMN_C2PMSG_82); } /* @@ -168,8 +160,8 @@ int vega20_read_arg_from_smc(struct pp_hwmgr *hwmgr, uint32_t *arg) * @param hwmgr the address of the HW manager * @param table_id the driver's table ID to copy from */ -int vega20_copy_table_from_smc(struct pp_hwmgr *hwmgr, - uint8_t *table, int16_t table_id) +static int vega20_copy_table_from_smc(struct pp_hwmgr *hwmgr, + uint8_t *table, int16_t table_id) { struct vega20_smumgr *priv = (struct vega20_smumgr *)(hwmgr->smu_backend); @@ -208,8 +200,8 @@ int vega20_copy_table_from_smc(struct pp_hwmgr *hwmgr, * @param hwmgr the address of the HW manager * @param table_id the table to copy from */ -int vega20_copy_table_to_smc(struct pp_hwmgr *hwmgr, - uint8_t *table, int16_t table_id) +static int vega20_copy_table_to_smc(struct pp_hwmgr *hwmgr, + uint8_t *table, int16_t table_id) { struct vega20_smumgr *priv = (struct vega20_smumgr *)(hwmgr->smu_backend); @@ -345,18 +337,12 @@ int vega20_get_enabled_smc_features(struct pp_hwmgr *hwmgr, PPSMC_MSG_GetEnabledSmuFeaturesLow)) == 0, "[GetEnabledSMCFeatures] Attemp to get SMU features Low failed!", return ret); - PP_ASSERT_WITH_CODE((ret = vega20_read_arg_from_smc(hwmgr, - &smc_features_low)) == 0, - "[GetEnabledSMCFeatures] Attemp to read SMU features Low argument failed!", - return ret); + smc_features_low = vega20_get_argument(hwmgr); PP_ASSERT_WITH_CODE((ret = vega20_send_msg_to_smc(hwmgr, PPSMC_MSG_GetEnabledSmuFeaturesHigh)) == 0, "[GetEnabledSMCFeatures] Attemp to get SMU features High failed!", return ret); - PP_ASSERT_WITH_CODE((ret = vega20_read_arg_from_smc(hwmgr, - &smc_features_high)) == 0, - "[GetEnabledSMCFeatures] Attemp to read SMU features High argument failed!", - return ret); + smc_features_high = vega20_get_argument(hwmgr); *features_enabled = ((((uint64_t)smc_features_low << SMU_FEATURES_LOW_SHIFT) & SMU_FEATURES_LOW_MASK) | (((uint64_t)smc_features_high << SMU_FEATURES_HIGH_SHIFT) & SMU_FEATURES_HIGH_MASK)); @@ -574,6 +560,19 @@ static bool vega20_is_dpm_running(struct pp_hwmgr *hwmgr) return false; } +static int vega20_smc_table_manager(struct pp_hwmgr *hwmgr, uint8_t *table, + uint16_t table_id, bool rw) +{ + int ret; + + if (rw) + ret = vega20_copy_table_from_smc(hwmgr, table, table_id); + else + ret = vega20_copy_table_to_smc(hwmgr, table, table_id); + + return ret; +} + const struct pp_smumgr_func vega20_smu_funcs = { .smu_init = &vega20_smu_init, .smu_fini = &vega20_smu_fini, @@ -584,4 +583,6 @@ const struct pp_smumgr_func vega20_smu_funcs = { .download_pptable_settings = NULL, .upload_pptable_settings = NULL, .is_dpm_running = vega20_is_dpm_running, + .get_argument = vega20_get_argument, + .smc_table_manager = vega20_smc_table_manager, }; diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/vega20_smumgr.h b/drivers/gpu/drm/amd/powerplay/smumgr/vega20_smumgr.h index 505eb0d82e3b..77349c3f0162 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/vega20_smumgr.h +++ b/drivers/gpu/drm/amd/powerplay/smumgr/vega20_smumgr.h @@ -47,11 +47,6 @@ struct vega20_smumgr { #define SMU_FEATURES_HIGH_MASK 0xFFFFFFFF00000000 #define SMU_FEATURES_HIGH_SHIFT 32 -int vega20_read_arg_from_smc(struct pp_hwmgr *hwmgr, uint32_t *arg); -int vega20_copy_table_from_smc(struct pp_hwmgr *hwmgr, - uint8_t *table, int16_t table_id); -int vega20_copy_table_to_smc(struct pp_hwmgr *hwmgr, - uint8_t *table, int16_t table_id); int vega20_enable_smc_features(struct pp_hwmgr *hwmgr, bool enable, uint64_t feature_mask); int vega20_get_enabled_smc_features(struct pp_hwmgr *hwmgr, diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/vegam_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/vegam_smumgr.c index 3d415fabbd93..9f71512b2510 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/vegam_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/vegam_smumgr.c @@ -2185,6 +2185,7 @@ static uint32_t vegam_get_offsetof(uint32_t type, uint32_t member) case DRAM_LOG_BUFF_SIZE: return offsetof(SMU75_SoftRegisters, DRAM_LOG_BUFF_SIZE); } + break; case SMU_Discrete_DpmTable: switch (member) { case UvdBootLevel: @@ -2194,6 +2195,7 @@ static uint32_t vegam_get_offsetof(uint32_t type, uint32_t member) case LowSclkInterruptThreshold: return offsetof(SMU75_Discrete_DpmTable, LowSclkInterruptThreshold); } + break; } pr_warn("can't get the offset of type %x member %x\n", type, member); return 0; diff --git a/drivers/gpu/drm/arm/malidp_crtc.c b/drivers/gpu/drm/arm/malidp_crtc.c index ef44202fb43f..e1b72782848c 100644 --- a/drivers/gpu/drm/arm/malidp_crtc.c +++ b/drivers/gpu/drm/arm/malidp_crtc.c @@ -348,19 +348,20 @@ static int malidp_crtc_atomic_check(struct drm_crtc *crtc, /* * check if there is enough rotation memory available for planes - * that need 90° and 270° rotation. Each plane has set its required - * memory size in the ->plane_check() callback, here we only make - * sure that the sums are less that the total usable memory. + * that need 90° and 270° rotion or planes that are compressed. + * Each plane has set its required memory size in the ->plane_check() + * callback, here we only make sure that the sums are less that the + * total usable memory. * * The rotation memory allocation algorithm (for each plane): - * a. If no more rotated planes exist, all remaining rotate - * memory in the bank is available for use by the plane. - * b. If other rotated planes exist, and plane's layer ID is - * DE_VIDEO1, it can use all the memory from first bank if - * secondary rotation memory bank is available, otherwise it can + * a. If no more rotated or compressed planes exist, all remaining + * rotate memory in the bank is available for use by the plane. + * b. If other rotated or compressed planes exist, and plane's + * layer ID is DE_VIDEO1, it can use all the memory from first bank + * if secondary rotation memory bank is available, otherwise it can * use up to half the bank's memory. - * c. If other rotated planes exist, and plane's layer ID is not - * DE_VIDEO1, it can use half of the available memory + * c. If other rotated or compressed planes exist, and plane's layer ID + * is not DE_VIDEO1, it can use half of the available memory. * * Note: this algorithm assumes that the order in which the planes are * checked always has DE_VIDEO1 plane first in the list if it is @@ -372,7 +373,9 @@ static int malidp_crtc_atomic_check(struct drm_crtc *crtc, /* first count the number of rotated planes */ drm_atomic_crtc_state_for_each_plane_state(plane, pstate, state) { - if (pstate->rotation & MALIDP_ROTATED_MASK) + struct drm_framebuffer *fb = pstate->fb; + + if ((pstate->rotation & MALIDP_ROTATED_MASK) || fb->modifier) rotated_planes++; } @@ -388,8 +391,9 @@ static int malidp_crtc_atomic_check(struct drm_crtc *crtc, drm_atomic_crtc_state_for_each_plane_state(plane, pstate, state) { struct malidp_plane *mp = to_malidp_plane(plane); struct malidp_plane_state *ms = to_malidp_plane_state(pstate); + struct drm_framebuffer *fb = pstate->fb; - if (pstate->rotation & MALIDP_ROTATED_MASK) { + if ((pstate->rotation & MALIDP_ROTATED_MASK) || fb->modifier) { /* process current plane */ rotated_planes--; diff --git a/drivers/gpu/drm/arm/malidp_drv.c b/drivers/gpu/drm/arm/malidp_drv.c index 3171ffaadd77..505f316a192e 100644 --- a/drivers/gpu/drm/arm/malidp_drv.c +++ b/drivers/gpu/drm/arm/malidp_drv.c @@ -37,6 +37,8 @@ #include "malidp_hw.h" #define MALIDP_CONF_VALID_TIMEOUT 250 +#define AFBC_HEADER_SIZE 16 +#define AFBC_SUPERBLK_ALIGNMENT 128 static void malidp_write_gamma_table(struct malidp_hw_device *hwdev, u32 data[MALIDP_COEFFTAB_NUM_COEFFS]) @@ -258,8 +260,134 @@ static const struct drm_mode_config_helper_funcs malidp_mode_config_helpers = { .atomic_commit_tail = malidp_atomic_commit_tail, }; +static bool +malidp_verify_afbc_framebuffer_caps(struct drm_device *dev, + const struct drm_mode_fb_cmd2 *mode_cmd) +{ + const struct drm_format_info *info; + + if ((mode_cmd->modifier[0] >> 56) != DRM_FORMAT_MOD_VENDOR_ARM) { + DRM_DEBUG_KMS("Unknown modifier (not Arm)\n"); + return false; + } + + if (mode_cmd->modifier[0] & + ~DRM_FORMAT_MOD_ARM_AFBC(AFBC_MOD_VALID_BITS)) { + DRM_DEBUG_KMS("Unsupported modifiers\n"); + return false; + } + + info = drm_get_format_info(dev, mode_cmd); + if (!info) { + DRM_DEBUG_KMS("Unable to get the format information\n"); + return false; + } + + if (info->num_planes != 1) { + DRM_DEBUG_KMS("AFBC buffers expect one plane\n"); + return false; + } + + if (mode_cmd->offsets[0] != 0) { + DRM_DEBUG_KMS("AFBC buffers' plane offset should be 0\n"); + return false; + } + + switch (mode_cmd->modifier[0] & AFBC_FORMAT_MOD_BLOCK_SIZE_MASK) { + case AFBC_FORMAT_MOD_BLOCK_SIZE_16x16: + if ((mode_cmd->width % 16) || (mode_cmd->height % 16)) { + DRM_DEBUG_KMS("AFBC buffers must be aligned to 16 pixels\n"); + return false; + } + break; + default: + DRM_DEBUG_KMS("Unsupported AFBC block size\n"); + return false; + } + + return true; +} + +static bool +malidp_verify_afbc_framebuffer_size(struct drm_device *dev, + struct drm_file *file, + const struct drm_mode_fb_cmd2 *mode_cmd) +{ + int n_superblocks = 0; + const struct drm_format_info *info; + struct drm_gem_object *objs = NULL; + u32 afbc_superblock_size = 0, afbc_superblock_height = 0; + u32 afbc_superblock_width = 0, afbc_size = 0; + + switch (mode_cmd->modifier[0] & AFBC_FORMAT_MOD_BLOCK_SIZE_MASK) { + case AFBC_FORMAT_MOD_BLOCK_SIZE_16x16: + afbc_superblock_height = 16; + afbc_superblock_width = 16; + break; + default: + DRM_DEBUG_KMS("AFBC superblock size is not supported\n"); + return false; + } + + info = drm_get_format_info(dev, mode_cmd); + + n_superblocks = (mode_cmd->width / afbc_superblock_width) * + (mode_cmd->height / afbc_superblock_height); + + afbc_superblock_size = info->cpp[0] * afbc_superblock_width * + afbc_superblock_height; + + afbc_size = ALIGN(n_superblocks * AFBC_HEADER_SIZE, AFBC_SUPERBLK_ALIGNMENT); + afbc_size += n_superblocks * ALIGN(afbc_superblock_size, AFBC_SUPERBLK_ALIGNMENT); + + if (mode_cmd->width * info->cpp[0] != mode_cmd->pitches[0]) { + DRM_DEBUG_KMS("Invalid value of pitch (=%u) should be same as width (=%u) * cpp (=%u)\n", + mode_cmd->pitches[0], mode_cmd->width, info->cpp[0]); + return false; + } + + objs = drm_gem_object_lookup(file, mode_cmd->handles[0]); + if (!objs) { + DRM_DEBUG_KMS("Failed to lookup GEM object\n"); + return false; + } + + if (objs->size < afbc_size) { + DRM_DEBUG_KMS("buffer size (%zu) too small for AFBC buffer size = %u\n", + objs->size, afbc_size); + drm_gem_object_put_unlocked(objs); + return false; + } + + drm_gem_object_put_unlocked(objs); + + return true; +} + +static bool +malidp_verify_afbc_framebuffer(struct drm_device *dev, struct drm_file *file, + const struct drm_mode_fb_cmd2 *mode_cmd) +{ + if (malidp_verify_afbc_framebuffer_caps(dev, mode_cmd)) + return malidp_verify_afbc_framebuffer_size(dev, file, mode_cmd); + + return false; +} + +struct drm_framebuffer * +malidp_fb_create(struct drm_device *dev, struct drm_file *file, + const struct drm_mode_fb_cmd2 *mode_cmd) +{ + if (mode_cmd->modifier[0]) { + if (!malidp_verify_afbc_framebuffer(dev, file, mode_cmd)) + return ERR_PTR(-EINVAL); + } + + return drm_gem_fb_create(dev, file, mode_cmd); +} + static const struct drm_mode_config_funcs malidp_mode_config_funcs = { - .fb_create = drm_gem_fb_create, + .fb_create = malidp_fb_create, .atomic_check = drm_atomic_helper_check, .atomic_commit = drm_atomic_helper_commit, }; @@ -752,6 +880,7 @@ static int malidp_bind(struct device *dev) drm->irq_enabled = true; ret = drm_vblank_init(drm, drm->mode_config.num_crtc); + drm_crtc_vblank_reset(&malidp->crtc); if (ret < 0) { DRM_ERROR("failed to initialise vblank\n"); goto vblank_fail; diff --git a/drivers/gpu/drm/arm/malidp_drv.h b/drivers/gpu/drm/arm/malidp_drv.h index e3eb0cb1f385..b76c86f18a56 100644 --- a/drivers/gpu/drm/arm/malidp_drv.h +++ b/drivers/gpu/drm/arm/malidp_drv.h @@ -55,6 +55,12 @@ struct malidp_plane { const struct malidp_layer *layer; }; +enum mmu_prefetch_mode { + MALIDP_PREFETCH_MODE_NONE, + MALIDP_PREFETCH_MODE_PARTIAL, + MALIDP_PREFETCH_MODE_FULL, +}; + struct malidp_plane_state { struct drm_plane_state base; @@ -63,6 +69,8 @@ struct malidp_plane_state { /* internal format ID */ u8 format; u8 n_planes; + enum mmu_prefetch_mode mmu_prefetch_mode; + u32 mmu_prefetch_pgsize; }; #define to_malidp_plane(x) container_of(x, struct malidp_plane, base) diff --git a/drivers/gpu/drm/arm/malidp_hw.c b/drivers/gpu/drm/arm/malidp_hw.c index c94a4422e0e9..7aad7dd80d8c 100644 --- a/drivers/gpu/drm/arm/malidp_hw.c +++ b/drivers/gpu/drm/arm/malidp_hw.c @@ -84,16 +84,48 @@ static const struct malidp_format_id malidp550_de_formats[] = { }; static const struct malidp_layer malidp500_layers[] = { - { DE_VIDEO1, MALIDP500_DE_LV_BASE, MALIDP500_DE_LV_PTR_BASE, MALIDP_DE_LV_STRIDE0, MALIDP500_LV_YUV2RGB }, - { DE_GRAPHICS1, MALIDP500_DE_LG1_BASE, MALIDP500_DE_LG1_PTR_BASE, MALIDP_DE_LG_STRIDE, 0 }, - { DE_GRAPHICS2, MALIDP500_DE_LG2_BASE, MALIDP500_DE_LG2_PTR_BASE, MALIDP_DE_LG_STRIDE, 0 }, + /* id, base address, fb pointer address base, stride offset, + * yuv2rgb matrix offset, mmu control register offset, rotation_features + */ + { DE_VIDEO1, MALIDP500_DE_LV_BASE, MALIDP500_DE_LV_PTR_BASE, + MALIDP_DE_LV_STRIDE0, MALIDP500_LV_YUV2RGB, 0, ROTATE_ANY }, + { DE_GRAPHICS1, MALIDP500_DE_LG1_BASE, MALIDP500_DE_LG1_PTR_BASE, + MALIDP_DE_LG_STRIDE, 0, 0, ROTATE_ANY }, + { DE_GRAPHICS2, MALIDP500_DE_LG2_BASE, MALIDP500_DE_LG2_PTR_BASE, + MALIDP_DE_LG_STRIDE, 0, 0, ROTATE_ANY }, }; static const struct malidp_layer malidp550_layers[] = { - { DE_VIDEO1, MALIDP550_DE_LV1_BASE, MALIDP550_DE_LV1_PTR_BASE, MALIDP_DE_LV_STRIDE0, MALIDP550_LV_YUV2RGB }, - { DE_GRAPHICS1, MALIDP550_DE_LG_BASE, MALIDP550_DE_LG_PTR_BASE, MALIDP_DE_LG_STRIDE, 0 }, - { DE_VIDEO2, MALIDP550_DE_LV2_BASE, MALIDP550_DE_LV2_PTR_BASE, MALIDP_DE_LV_STRIDE0, MALIDP550_LV_YUV2RGB }, - { DE_SMART, MALIDP550_DE_LS_BASE, MALIDP550_DE_LS_PTR_BASE, MALIDP550_DE_LS_R1_STRIDE, 0 }, + /* id, base address, fb pointer address base, stride offset, + * yuv2rgb matrix offset, mmu control register offset, rotation_features + */ + { DE_VIDEO1, MALIDP550_DE_LV1_BASE, MALIDP550_DE_LV1_PTR_BASE, + MALIDP_DE_LV_STRIDE0, MALIDP550_LV_YUV2RGB, 0, ROTATE_ANY }, + { DE_GRAPHICS1, MALIDP550_DE_LG_BASE, MALIDP550_DE_LG_PTR_BASE, + MALIDP_DE_LG_STRIDE, 0, 0, ROTATE_ANY }, + { DE_VIDEO2, MALIDP550_DE_LV2_BASE, MALIDP550_DE_LV2_PTR_BASE, + MALIDP_DE_LV_STRIDE0, MALIDP550_LV_YUV2RGB, 0, ROTATE_ANY }, + { DE_SMART, MALIDP550_DE_LS_BASE, MALIDP550_DE_LS_PTR_BASE, + MALIDP550_DE_LS_R1_STRIDE, 0, 0, ROTATE_NONE }, +}; + +static const struct malidp_layer malidp650_layers[] = { + /* id, base address, fb pointer address base, stride offset, + * yuv2rgb matrix offset, mmu control register offset, + * rotation_features + */ + { DE_VIDEO1, MALIDP550_DE_LV1_BASE, MALIDP550_DE_LV1_PTR_BASE, + MALIDP_DE_LV_STRIDE0, MALIDP550_LV_YUV2RGB, + MALIDP650_DE_LV_MMU_CTRL, ROTATE_ANY }, + { DE_GRAPHICS1, MALIDP550_DE_LG_BASE, MALIDP550_DE_LG_PTR_BASE, + MALIDP_DE_LG_STRIDE, 0, MALIDP650_DE_LG_MMU_CTRL, + ROTATE_COMPRESSED }, + { DE_VIDEO2, MALIDP550_DE_LV2_BASE, MALIDP550_DE_LV2_PTR_BASE, + MALIDP_DE_LV_STRIDE0, MALIDP550_LV_YUV2RGB, + MALIDP650_DE_LV_MMU_CTRL, ROTATE_ANY }, + { DE_SMART, MALIDP550_DE_LS_BASE, MALIDP550_DE_LS_PTR_BASE, + MALIDP550_DE_LS_R1_STRIDE, 0, MALIDP650_DE_LS_MMU_CTRL, + ROTATE_NONE }, }; #define SE_N_SCALING_COEFFS 96 @@ -288,10 +320,6 @@ static void malidp500_modeset(struct malidp_hw_device *hwdev, struct videomode * static int malidp500_rotmem_required(struct malidp_hw_device *hwdev, u16 w, u16 h, u32 fmt) { - /* RGB888 or BGR888 can't be rotated */ - if ((fmt == DRM_FORMAT_RGB888) || (fmt == DRM_FORMAT_BGR888)) - return -EINVAL; - /* * Each layer needs enough rotation memory to fit 8 lines * worth of pixel data. Required size is then: @@ -384,7 +412,8 @@ static long malidp500_se_calc_mclk(struct malidp_hw_device *hwdev, static int malidp500_enable_memwrite(struct malidp_hw_device *hwdev, dma_addr_t *addrs, s32 *pitches, - int num_planes, u16 w, u16 h, u32 fmt_id) + int num_planes, u16 w, u16 h, u32 fmt_id, + const s16 *rgb2yuv_coeffs) { u32 base = MALIDP500_SE_MEMWRITE_BASE; u32 de_base = malidp_get_block_base(hwdev, MALIDP_DE_BLOCK); @@ -416,6 +445,16 @@ static int malidp500_enable_memwrite(struct malidp_hw_device *hwdev, malidp_hw_write(hwdev, MALIDP_DE_H_ACTIVE(w) | MALIDP_DE_V_ACTIVE(h), MALIDP500_SE_MEMWRITE_OUT_SIZE); + + if (rgb2yuv_coeffs) { + int i; + + for (i = 0; i < MALIDP_COLORADJ_NUM_COEFFS; i++) { + malidp_hw_write(hwdev, rgb2yuv_coeffs[i], + MALIDP500_SE_RGB_YUV_COEFFS + i * 4); + } + } + malidp_hw_setbits(hwdev, MALIDP_SE_MEMWRITE_EN, MALIDP500_SE_CONTROL); return 0; @@ -568,10 +607,6 @@ static int malidp550_rotmem_required(struct malidp_hw_device *hwdev, u16 w, u16 { u32 bytes_per_col; - /* raw RGB888 or BGR888 can't be rotated */ - if ((fmt == DRM_FORMAT_RGB888) || (fmt == DRM_FORMAT_BGR888)) - return -EINVAL; - switch (fmt) { /* 8 lines at 4 bytes per pixel */ case DRM_FORMAT_ARGB2101010: @@ -658,7 +693,8 @@ static long malidp550_se_calc_mclk(struct malidp_hw_device *hwdev, static int malidp550_enable_memwrite(struct malidp_hw_device *hwdev, dma_addr_t *addrs, s32 *pitches, - int num_planes, u16 w, u16 h, u32 fmt_id) + int num_planes, u16 w, u16 h, u32 fmt_id, + const s16 *rgb2yuv_coeffs) { u32 base = MALIDP550_SE_MEMWRITE_BASE; u32 de_base = malidp_get_block_base(hwdev, MALIDP_DE_BLOCK); @@ -689,6 +725,15 @@ static int malidp550_enable_memwrite(struct malidp_hw_device *hwdev, malidp_hw_setbits(hwdev, MALIDP550_SE_MEMWRITE_ONESHOT | MALIDP_SE_MEMWRITE_EN, MALIDP550_SE_CONTROL); + if (rgb2yuv_coeffs) { + int i; + + for (i = 0; i < MALIDP_COLORADJ_NUM_COEFFS; i++) { + malidp_hw_write(hwdev, rgb2yuv_coeffs[i], + MALIDP550_SE_RGB_YUV_COEFFS + i * 4); + } + } + return 0; } @@ -832,8 +877,8 @@ const struct malidp_hw malidp_device[MALIDP_MAX_DEVICES] = { .dc_base = MALIDP550_DC_BASE, .out_depth_base = MALIDP550_DE_OUTPUT_DEPTH, .features = MALIDP_REGMAP_HAS_CLEARIRQ, - .n_layers = ARRAY_SIZE(malidp550_layers), - .layers = malidp550_layers, + .n_layers = ARRAY_SIZE(malidp650_layers), + .layers = malidp650_layers, .de_irq_map = { .irq_mask = MALIDP_DE_IRQ_UNDERRUN | MALIDP650_DE_IRQ_DRIFT | diff --git a/drivers/gpu/drm/arm/malidp_hw.h b/drivers/gpu/drm/arm/malidp_hw.h index ad2e96915d44..40155e2ea9d9 100644 --- a/drivers/gpu/drm/arm/malidp_hw.h +++ b/drivers/gpu/drm/arm/malidp_hw.h @@ -36,6 +36,12 @@ enum { SE_MEMWRITE = BIT(5), }; +enum rotation_features { + ROTATE_NONE, /* does not support rotation at all */ + ROTATE_ANY, /* supports rotation on any buffers */ + ROTATE_COMPRESSED, /* supports rotation only on compressed buffers */ +}; + struct malidp_format_id { u32 format; /* DRM fourcc */ u8 layer; /* bitmask of layers supporting it */ @@ -62,6 +68,8 @@ struct malidp_layer { u16 ptr; /* address offset for the pointer register */ u16 stride_offset; /* offset to the first stride register. */ s16 yuv2rgb_offset; /* offset to the YUV->RGB matrix entries */ + u16 mmu_ctrl_offset; /* offset to the MMU control register */ + enum rotation_features rot; /* type of rotation supported */ }; enum malidp_scaling_coeff_set { @@ -191,7 +199,8 @@ struct malidp_hw { * @param fmt_id - internal format ID of output buffer */ int (*enable_memwrite)(struct malidp_hw_device *hwdev, dma_addr_t *addrs, - s32 *pitches, int num_planes, u16 w, u16 h, u32 fmt_id); + s32 *pitches, int num_planes, u16 w, u16 h, u32 fmt_id, + const s16 *rgb2yuv_coeffs); /* * Disable the writing to memory of the next frame's content. @@ -379,4 +388,9 @@ static inline void malidp_se_set_enh_coeffs(struct malidp_hw_device *hwdev) #define MALIDP_GAMMA_LUT_SIZE 4096 +#define AFBC_MOD_VALID_BITS (AFBC_FORMAT_MOD_BLOCK_SIZE_MASK | \ + AFBC_FORMAT_MOD_YTR | AFBC_FORMAT_MOD_SPLIT | \ + AFBC_FORMAT_MOD_SPARSE | AFBC_FORMAT_MOD_CBR | \ + AFBC_FORMAT_MOD_TILED | AFBC_FORMAT_MOD_SC) + #endif /* __MALIDP_HW_H__ */ diff --git a/drivers/gpu/drm/arm/malidp_mw.c b/drivers/gpu/drm/arm/malidp_mw.c index ba6ae66387c9..91472e5e0c8b 100644 --- a/drivers/gpu/drm/arm/malidp_mw.c +++ b/drivers/gpu/drm/arm/malidp_mw.c @@ -26,6 +26,8 @@ struct malidp_mw_connector_state { s32 pitches[2]; u8 format; u8 n_planes; + bool rgb2yuv_initialized; + const s16 *rgb2yuv_coeffs; }; static int malidp_mw_connector_get_modes(struct drm_connector *connector) @@ -84,7 +86,7 @@ static void malidp_mw_connector_destroy(struct drm_connector *connector) static struct drm_connector_state * malidp_mw_connector_duplicate_state(struct drm_connector *connector) { - struct malidp_mw_connector_state *mw_state; + struct malidp_mw_connector_state *mw_state, *mw_current_state; if (WARN_ON(!connector->state)) return NULL; @@ -93,7 +95,10 @@ malidp_mw_connector_duplicate_state(struct drm_connector *connector) if (!mw_state) return NULL; - /* No need to preserve any of our driver-local data */ + mw_current_state = to_mw_state(connector->state); + mw_state->rgb2yuv_coeffs = mw_current_state->rgb2yuv_coeffs; + mw_state->rgb2yuv_initialized = mw_current_state->rgb2yuv_initialized; + __drm_atomic_helper_connector_duplicate_state(connector, &mw_state->base); return &mw_state->base; @@ -108,6 +113,13 @@ static const struct drm_connector_funcs malidp_mw_connector_funcs = { .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, }; +static const s16 rgb2yuv_coeffs_bt709_limited[MALIDP_COLORADJ_NUM_COEFFS] = { + 47, 157, 16, + -26, -87, 112, + 112, -102, -10, + 16, 128, 128 +}; + static int malidp_mw_encoder_atomic_check(struct drm_encoder *encoder, struct drm_crtc_state *crtc_state, @@ -157,6 +169,9 @@ malidp_mw_encoder_atomic_check(struct drm_encoder *encoder, } mw_state->n_planes = n_planes; + if (fb->format->is_yuv) + mw_state->rgb2yuv_coeffs = rgb2yuv_coeffs_bt709_limited; + return 0; } @@ -239,10 +254,12 @@ void malidp_mw_atomic_commit(struct drm_device *drm, drm_writeback_queue_job(mw_conn, conn_state->writeback_job); conn_state->writeback_job = NULL; - hwdev->hw->enable_memwrite(hwdev, mw_state->addrs, mw_state->pitches, mw_state->n_planes, - fb->width, fb->height, mw_state->format); + fb->width, fb->height, mw_state->format, + !mw_state->rgb2yuv_initialized ? + mw_state->rgb2yuv_coeffs : NULL); + mw_state->rgb2yuv_initialized = !!mw_state->rgb2yuv_coeffs; } else { DRM_DEV_DEBUG_DRIVER(drm->dev, "Disable memwrite\n"); hwdev->hw->disable_memwrite(hwdev); diff --git a/drivers/gpu/drm/arm/malidp_planes.c b/drivers/gpu/drm/arm/malidp_planes.c index 49c37f6dd63e..837a24d56675 100644 --- a/drivers/gpu/drm/arm/malidp_planes.c +++ b/drivers/gpu/drm/arm/malidp_planes.c @@ -10,11 +10,14 @@ * ARM Mali DP plane manipulation routines. */ +#include <linux/iommu.h> + #include <drm/drmP.h> #include <drm/drm_atomic.h> #include <drm/drm_atomic_helper.h> #include <drm/drm_fb_cma_helper.h> #include <drm/drm_gem_cma_helper.h> +#include <drm/drm_gem_framebuffer_helper.h> #include <drm/drm_plane_helper.h> #include <drm/drm_print.h> @@ -36,6 +39,7 @@ #define LAYER_COMP_MASK (0x3 << 12) #define LAYER_COMP_PIXEL (0x3 << 12) #define LAYER_COMP_PLANE (0x2 << 12) +#define LAYER_PMUL_ENABLE (0x1 << 14) #define LAYER_ALPHA_OFFSET (16) #define LAYER_ALPHA_MASK (0xff) #define LAYER_ALPHA(x) (((x) & LAYER_ALPHA_MASK) << LAYER_ALPHA_OFFSET) @@ -56,6 +60,13 @@ */ #define MALIDP_ALPHA_LUT 0xffaa5500 +/* page sizes the MMU prefetcher can support */ +#define MALIDP_MMU_PREFETCH_PARTIAL_PGSIZES (SZ_4K | SZ_64K) +#define MALIDP_MMU_PREFETCH_FULL_PGSIZES (SZ_1M | SZ_2M) + +/* readahead for partial-frame prefetch */ +#define MALIDP_MMU_PREFETCH_READAHEAD 8 + static void malidp_de_plane_destroy(struct drm_plane *plane) { struct malidp_plane *mp = to_malidp_plane(plane); @@ -100,6 +111,9 @@ drm_plane_state *malidp_duplicate_plane_state(struct drm_plane *plane) state->format = m_state->format; state->n_planes = m_state->n_planes; + state->mmu_prefetch_mode = m_state->mmu_prefetch_mode; + state->mmu_prefetch_pgsize = m_state->mmu_prefetch_pgsize; + return &state->base; } @@ -112,6 +126,12 @@ static void malidp_destroy_plane_state(struct drm_plane *plane, kfree(m_state); } +static const char * const prefetch_mode_names[] = { + [MALIDP_PREFETCH_MODE_NONE] = "MMU_PREFETCH_NONE", + [MALIDP_PREFETCH_MODE_PARTIAL] = "MMU_PREFETCH_PARTIAL", + [MALIDP_PREFETCH_MODE_FULL] = "MMU_PREFETCH_FULL", +}; + static void malidp_plane_atomic_print_state(struct drm_printer *p, const struct drm_plane_state *state) { @@ -120,6 +140,9 @@ static void malidp_plane_atomic_print_state(struct drm_printer *p, drm_printf(p, "\trotmem_size=%u\n", ms->rotmem_size); drm_printf(p, "\tformat_id=%u\n", ms->format); drm_printf(p, "\tn_planes=%u\n", ms->n_planes); + drm_printf(p, "\tmmu_prefetch_mode=%s\n", + prefetch_mode_names[ms->mmu_prefetch_mode]); + drm_printf(p, "\tmmu_prefetch_pgsize=%d\n", ms->mmu_prefetch_pgsize); } static const struct drm_plane_funcs malidp_de_plane_funcs = { @@ -173,6 +196,199 @@ static int malidp_se_check_scaling(struct malidp_plane *mp, return 0; } +static u32 malidp_get_pgsize_bitmap(struct malidp_plane *mp) +{ + u32 pgsize_bitmap = 0; + + if (iommu_present(&platform_bus_type)) { + struct iommu_domain *mmu_dom = + iommu_get_domain_for_dev(mp->base.dev->dev); + + if (mmu_dom) + pgsize_bitmap = mmu_dom->pgsize_bitmap; + } + + return pgsize_bitmap; +} + +/* + * Check if the framebuffer is entirely made up of pages at least pgsize in + * size. Only a heuristic: assumes that each scatterlist entry has been aligned + * to the largest page size smaller than its length and that the MMU maps to + * the largest page size possible. + */ +static bool malidp_check_pages_threshold(struct malidp_plane_state *ms, + u32 pgsize) +{ + int i; + + for (i = 0; i < ms->n_planes; i++) { + struct drm_gem_object *obj; + struct drm_gem_cma_object *cma_obj; + struct sg_table *sgt; + struct scatterlist *sgl; + + obj = drm_gem_fb_get_obj(ms->base.fb, i); + cma_obj = to_drm_gem_cma_obj(obj); + + if (cma_obj->sgt) + sgt = cma_obj->sgt; + else + sgt = obj->dev->driver->gem_prime_get_sg_table(obj); + + if (!sgt) + return false; + + sgl = sgt->sgl; + + while (sgl) { + if (sgl->length < pgsize) { + if (!cma_obj->sgt) + kfree(sgt); + return false; + } + + sgl = sg_next(sgl); + } + if (!cma_obj->sgt) + kfree(sgt); + } + + return true; +} + +/* + * Check if it is possible to enable partial-frame MMU prefetch given the + * current format, AFBC state and rotation. + */ +static bool malidp_partial_prefetch_supported(u32 format, u64 modifier, + unsigned int rotation) +{ + bool afbc, sparse; + + /* rotation and horizontal flip not supported for partial prefetch */ + if (rotation & (DRM_MODE_ROTATE_90 | DRM_MODE_ROTATE_180 | + DRM_MODE_ROTATE_270 | DRM_MODE_REFLECT_X)) + return false; + + afbc = modifier & DRM_FORMAT_MOD_ARM_AFBC(0); + sparse = modifier & AFBC_FORMAT_MOD_SPARSE; + + switch (format) { + case DRM_FORMAT_ARGB2101010: + case DRM_FORMAT_RGBA1010102: + case DRM_FORMAT_BGRA1010102: + case DRM_FORMAT_ARGB8888: + case DRM_FORMAT_RGBA8888: + case DRM_FORMAT_BGRA8888: + case DRM_FORMAT_XRGB8888: + case DRM_FORMAT_XBGR8888: + case DRM_FORMAT_RGBX8888: + case DRM_FORMAT_BGRX8888: + case DRM_FORMAT_RGB888: + case DRM_FORMAT_RGBA5551: + case DRM_FORMAT_RGB565: + /* always supported */ + return true; + + case DRM_FORMAT_ABGR2101010: + case DRM_FORMAT_ABGR8888: + case DRM_FORMAT_ABGR1555: + case DRM_FORMAT_BGR565: + /* supported, but if AFBC then must be sparse mode */ + return (!afbc) || (afbc && sparse); + + case DRM_FORMAT_BGR888: + /* supported, but not for AFBC */ + return !afbc; + + case DRM_FORMAT_YUYV: + case DRM_FORMAT_UYVY: + case DRM_FORMAT_NV12: + case DRM_FORMAT_YUV420: + /* not supported */ + return false; + + default: + return false; + } +} + +/* + * Select the preferred MMU prefetch mode. Full-frame prefetch is preferred as + * long as the framebuffer is all large pages. Otherwise partial-frame prefetch + * is selected as long as it is supported for the current format. The selected + * page size for prefetch is returned in pgsize_bitmap. + */ +static enum mmu_prefetch_mode malidp_mmu_prefetch_select_mode + (struct malidp_plane_state *ms, u32 *pgsize_bitmap) +{ + u32 pgsizes; + + /* get the full-frame prefetch page size(s) supported by the MMU */ + pgsizes = *pgsize_bitmap & MALIDP_MMU_PREFETCH_FULL_PGSIZES; + + while (pgsizes) { + u32 largest_pgsize = 1 << __fls(pgsizes); + + if (malidp_check_pages_threshold(ms, largest_pgsize)) { + *pgsize_bitmap = largest_pgsize; + return MALIDP_PREFETCH_MODE_FULL; + } + + pgsizes -= largest_pgsize; + } + + /* get the partial-frame prefetch page size(s) supported by the MMU */ + pgsizes = *pgsize_bitmap & MALIDP_MMU_PREFETCH_PARTIAL_PGSIZES; + + if (malidp_partial_prefetch_supported(ms->base.fb->format->format, + ms->base.fb->modifier, + ms->base.rotation)) { + /* partial prefetch using the smallest page size */ + *pgsize_bitmap = 1 << __ffs(pgsizes); + return MALIDP_PREFETCH_MODE_PARTIAL; + } + *pgsize_bitmap = 0; + return MALIDP_PREFETCH_MODE_NONE; +} + +static u32 malidp_calc_mmu_control_value(enum mmu_prefetch_mode mode, + u8 readahead, u8 n_planes, u32 pgsize) +{ + u32 mmu_ctrl = 0; + + if (mode != MALIDP_PREFETCH_MODE_NONE) { + mmu_ctrl |= MALIDP_MMU_CTRL_EN; + + if (mode == MALIDP_PREFETCH_MODE_PARTIAL) { + mmu_ctrl |= MALIDP_MMU_CTRL_MODE; + mmu_ctrl |= MALIDP_MMU_CTRL_PP_NUM_REQ(readahead); + } + + if (pgsize == SZ_64K || pgsize == SZ_2M) { + int i; + + for (i = 0; i < n_planes; i++) + mmu_ctrl |= MALIDP_MMU_CTRL_PX_PS(i); + } + } + + return mmu_ctrl; +} + +static void malidp_de_prefetch_settings(struct malidp_plane *mp, + struct malidp_plane_state *ms) +{ + if (!mp->layer->mmu_ctrl_offset) + return; + + /* get the page sizes supported by the MMU */ + ms->mmu_prefetch_pgsize = malidp_get_pgsize_bitmap(mp); + ms->mmu_prefetch_mode = + malidp_mmu_prefetch_select_mode(ms, &ms->mmu_prefetch_pgsize); +} + static int malidp_de_plane_check(struct drm_plane *plane, struct drm_plane_state *state) { @@ -180,6 +396,7 @@ static int malidp_de_plane_check(struct drm_plane *plane, struct malidp_plane_state *ms = to_malidp_plane_state(state); bool rotated = state->rotation & MALIDP_ROTATED_MASK; struct drm_framebuffer *fb; + u16 pixel_alpha = state->pixel_blend_mode; int i, ret; if (!state->crtc || !state->fb) @@ -223,11 +440,20 @@ static int malidp_de_plane_check(struct drm_plane *plane, if (ret) return ret; - /* packed RGB888 / BGR888 can't be rotated or flipped */ - if (state->rotation != DRM_MODE_ROTATE_0 && - (fb->format->format == DRM_FORMAT_RGB888 || - fb->format->format == DRM_FORMAT_BGR888)) - return -EINVAL; + /* validate the rotation constraints for each layer */ + if (state->rotation != DRM_MODE_ROTATE_0) { + if (mp->layer->rot == ROTATE_NONE) + return -EINVAL; + if ((mp->layer->rot == ROTATE_COMPRESSED) && !(fb->modifier)) + return -EINVAL; + /* + * packed RGB888 / BGR888 can't be rotated or flipped + * unless they are stored in a compressed way + */ + if ((fb->format->format == DRM_FORMAT_RGB888 || + fb->format->format == DRM_FORMAT_BGR888) && !(fb->modifier)) + return -EINVAL; + } ms->rotmem_size = 0; if (state->rotation & MALIDP_ROTATED_MASK) { @@ -242,6 +468,14 @@ static int malidp_de_plane_check(struct drm_plane *plane, ms->rotmem_size = val; } + /* HW can't support plane + pixel blending */ + if ((state->alpha != DRM_BLEND_ALPHA_OPAQUE) && + (pixel_alpha != DRM_MODE_BLEND_PIXEL_NONE) && + fb->format->has_alpha) + return -EINVAL; + + malidp_de_prefetch_settings(mp, ms); + return 0; } @@ -318,22 +552,42 @@ static void malidp_de_set_color_encoding(struct malidp_plane *plane, } } +static void malidp_de_set_mmu_control(struct malidp_plane *mp, + struct malidp_plane_state *ms) +{ + u32 mmu_ctrl; + + /* check hardware supports MMU prefetch */ + if (!mp->layer->mmu_ctrl_offset) + return; + + mmu_ctrl = malidp_calc_mmu_control_value(ms->mmu_prefetch_mode, + MALIDP_MMU_PREFETCH_READAHEAD, + ms->n_planes, + ms->mmu_prefetch_pgsize); + + malidp_hw_write(mp->hwdev, mmu_ctrl, + mp->layer->base + mp->layer->mmu_ctrl_offset); +} + static void malidp_de_plane_update(struct drm_plane *plane, struct drm_plane_state *old_state) { struct malidp_plane *mp; struct malidp_plane_state *ms = to_malidp_plane_state(plane->state); + struct drm_plane_state *state = plane->state; + u16 pixel_alpha = state->pixel_blend_mode; + u8 plane_alpha = state->alpha >> 8; u32 src_w, src_h, dest_w, dest_h, val; int i; - bool format_has_alpha = plane->state->fb->format->has_alpha; mp = to_malidp_plane(plane); /* convert src values from Q16 fixed point to integer */ - src_w = plane->state->src_w >> 16; - src_h = plane->state->src_h >> 16; - dest_w = plane->state->crtc_w; - dest_h = plane->state->crtc_h; + src_w = state->src_w >> 16; + src_h = state->src_h >> 16; + dest_w = state->crtc_w; + dest_h = state->crtc_h; val = malidp_hw_read(mp->hwdev, mp->layer->base); val = (val & ~LAYER_FORMAT_MASK) | ms->format; @@ -342,14 +596,17 @@ static void malidp_de_plane_update(struct drm_plane *plane, for (i = 0; i < ms->n_planes; i++) { /* calculate the offset for the layer's plane registers */ u16 ptr = mp->layer->ptr + (i << 4); - dma_addr_t fb_addr = drm_fb_cma_get_gem_addr(plane->state->fb, - plane->state, i); + dma_addr_t fb_addr = drm_fb_cma_get_gem_addr(state->fb, + state, i); malidp_hw_write(mp->hwdev, lower_32_bits(fb_addr), ptr); malidp_hw_write(mp->hwdev, upper_32_bits(fb_addr), ptr + 4); } + + malidp_de_set_mmu_control(mp, ms); + malidp_de_set_plane_pitches(mp, ms->n_planes, - plane->state->fb->pitches); + state->fb->pitches); if ((plane->state->color_encoding != old_state->color_encoding) || (plane->state->color_range != old_state->color_range)) @@ -362,52 +619,56 @@ static void malidp_de_plane_update(struct drm_plane *plane, malidp_hw_write(mp->hwdev, LAYER_H_VAL(dest_w) | LAYER_V_VAL(dest_h), mp->layer->base + MALIDP_LAYER_COMP_SIZE); - malidp_hw_write(mp->hwdev, LAYER_H_VAL(plane->state->crtc_x) | - LAYER_V_VAL(plane->state->crtc_y), + malidp_hw_write(mp->hwdev, LAYER_H_VAL(state->crtc_x) | + LAYER_V_VAL(state->crtc_y), mp->layer->base + MALIDP_LAYER_OFFSET); - if (mp->layer->id == DE_SMART) + if (mp->layer->id == DE_SMART) { + /* + * Enable the first rectangle in the SMART layer to be + * able to use it as a drm plane. + */ + malidp_hw_write(mp->hwdev, 1, + mp->layer->base + MALIDP550_LS_ENABLE); malidp_hw_write(mp->hwdev, LAYER_H_VAL(src_w) | LAYER_V_VAL(src_h), mp->layer->base + MALIDP550_LS_R1_IN_SIZE); + } /* first clear the rotation bits */ val = malidp_hw_read(mp->hwdev, mp->layer->base + MALIDP_LAYER_CONTROL); val &= ~LAYER_ROT_MASK; /* setup the rotation and axis flip bits */ - if (plane->state->rotation & DRM_MODE_ROTATE_MASK) + if (state->rotation & DRM_MODE_ROTATE_MASK) val |= ilog2(plane->state->rotation & DRM_MODE_ROTATE_MASK) << LAYER_ROT_OFFSET; - if (plane->state->rotation & DRM_MODE_REFLECT_X) + if (state->rotation & DRM_MODE_REFLECT_X) val |= LAYER_H_FLIP; - if (plane->state->rotation & DRM_MODE_REFLECT_Y) + if (state->rotation & DRM_MODE_REFLECT_Y) val |= LAYER_V_FLIP; - val &= ~LAYER_COMP_MASK; - if (format_has_alpha) { - - /* - * always enable pixel alpha blending until we have a way - * to change blend modes - */ - val |= LAYER_COMP_PIXEL; - } else { + val &= ~(LAYER_COMP_MASK | LAYER_PMUL_ENABLE | LAYER_ALPHA(0xff)); - /* - * do not enable pixel alpha blending as the color channel - * does not have any alpha information - */ + if (state->alpha != DRM_BLEND_ALPHA_OPAQUE) { val |= LAYER_COMP_PLANE; - - /* Set layer alpha coefficient to 0xff ie fully opaque */ - val |= LAYER_ALPHA(0xff); + } else if (state->fb->format->has_alpha) { + /* We only care about blend mode if the format has alpha */ + switch (pixel_alpha) { + case DRM_MODE_BLEND_PREMULTI: + val |= LAYER_COMP_PIXEL | LAYER_PMUL_ENABLE; + break; + case DRM_MODE_BLEND_COVERAGE: + val |= LAYER_COMP_PIXEL; + break; + } } + val |= LAYER_ALPHA(plane_alpha); val &= ~LAYER_FLOWCFG(LAYER_FLOWCFG_MASK); - if (plane->state->crtc) { + if (state->crtc) { struct malidp_crtc_state *m = - to_malidp_crtc_state(plane->state->crtc->state); + to_malidp_crtc_state(state->crtc->state); if (m->scaler_config.scale_enable && m->scaler_config.plane_src_id == mp->layer->id) @@ -446,6 +707,9 @@ int malidp_de_planes_init(struct drm_device *drm) unsigned long crtcs = 1 << drm->mode_config.num_crtc; unsigned long flags = DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_90 | DRM_MODE_ROTATE_180 | DRM_MODE_ROTATE_270 | DRM_MODE_REFLECT_X | DRM_MODE_REFLECT_Y; + unsigned int blend_caps = BIT(DRM_MODE_BLEND_PIXEL_NONE) | + BIT(DRM_MODE_BLEND_PREMULTI) | + BIT(DRM_MODE_BLEND_COVERAGE); u32 *formats; int ret, i, j, n; @@ -483,13 +747,10 @@ int malidp_de_planes_init(struct drm_device *drm) plane->hwdev = malidp->dev; plane->layer = &map->layers[i]; + drm_plane_create_alpha_property(&plane->base); + drm_plane_create_blend_mode_property(&plane->base, blend_caps); + if (id == DE_SMART) { - /* - * Enable the first rectangle in the SMART layer to be - * able to use it as a drm plane. - */ - malidp_hw_write(malidp->dev, 1, - plane->layer->base + MALIDP550_LS_ENABLE); /* Skip the features which the SMART layer doesn't have. */ continue; } diff --git a/drivers/gpu/drm/arm/malidp_regs.h b/drivers/gpu/drm/arm/malidp_regs.h index 3579d36b2a71..7ce3e141464d 100644 --- a/drivers/gpu/drm/arm/malidp_regs.h +++ b/drivers/gpu/drm/arm/malidp_regs.h @@ -205,6 +205,7 @@ #define MALIDP500_SE_BASE 0x00c00 #define MALIDP500_SE_CONTROL 0x00c0c #define MALIDP500_SE_MEMWRITE_OUT_SIZE 0x00c2c +#define MALIDP500_SE_RGB_YUV_COEFFS 0x00C74 #define MALIDP500_SE_MEMWRITE_BASE 0x00e00 #define MALIDP500_DC_IRQ_BASE 0x00f00 #define MALIDP500_CONFIG_VALID 0x00f00 @@ -238,6 +239,7 @@ #define MALIDP550_SE_CONTROL 0x08010 #define MALIDP550_SE_MEMWRITE_ONESHOT (1 << 7) #define MALIDP550_SE_MEMWRITE_OUT_SIZE 0x08030 +#define MALIDP550_SE_RGB_YUV_COEFFS 0x08078 #define MALIDP550_SE_MEMWRITE_BASE 0x08100 #define MALIDP550_DC_BASE 0x0c000 #define MALIDP550_DC_CONTROL 0x0c010 @@ -245,6 +247,17 @@ #define MALIDP550_CONFIG_VALID 0x0c014 #define MALIDP550_CONFIG_ID 0x0ffd4 +/* register offsets specific to DP650 */ +#define MALIDP650_DE_LV_MMU_CTRL 0x000D0 +#define MALIDP650_DE_LG_MMU_CTRL 0x00048 +#define MALIDP650_DE_LS_MMU_CTRL 0x00078 + +/* bit masks to set the MMU control register */ +#define MALIDP_MMU_CTRL_EN (1 << 0) +#define MALIDP_MMU_CTRL_MODE (1 << 4) +#define MALIDP_MMU_CTRL_PX_PS(x) (1 << (8 + (x))) +#define MALIDP_MMU_CTRL_PP_NUM_REQ(x) (((x) & 0x7f) << 12) + /* * Starting with DP550 the register map blocks has been standardised to the * following layout: diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c index a504a5e05676..9b111e846847 100644 --- a/drivers/gpu/drm/drm_fb_helper.c +++ b/drivers/gpu/drm/drm_fb_helper.c @@ -56,6 +56,25 @@ MODULE_PARM_DESC(drm_fbdev_overalloc, "Overallocation of the fbdev buffer (%) [default=" __MODULE_STRING(CONFIG_DRM_FBDEV_OVERALLOC) "]"); +/* + * In order to keep user-space compatibility, we want in certain use-cases + * to keep leaking the fbdev physical address to the user-space program + * handling the fbdev buffer. + * This is a bad habit essentially kept into closed source opengl driver + * that should really be moved into open-source upstream projects instead + * of using legacy physical addresses in user space to communicate with + * other out-of-tree kernel modules. + * + * This module_param *should* be removed as soon as possible and be + * considered as a broken and legacy behaviour from a modern fbdev device. + */ +#if IS_ENABLED(CONFIG_DRM_FBDEV_LEAK_PHYS_SMEM) +static bool drm_leak_fbdev_smem = false; +module_param_unsafe(drm_leak_fbdev_smem, bool, 0600); +MODULE_PARM_DESC(fbdev_emulation, + "Allow unsafe leaking fbdev physical smem address [default=false]"); +#endif + static LIST_HEAD(kernel_fb_helper_list); static DEFINE_MUTEX(kernel_fb_helper_lock); @@ -2670,8 +2689,12 @@ __drm_fb_helper_initial_config_and_unlock(struct drm_fb_helper *fb_helper, info = fb_helper->fbdev; info->var.pixclock = 0; - /* don't leak any physical addresses to userspace */ - info->flags |= FBINFO_HIDE_SMEM_START; + /* Shamelessly allow physical address leaking to userspace */ +#if IS_ENABLED(CONFIG_DRM_FBDEV_LEAK_PHYS_SMEM) + if (!drm_leak_fbdev_smem) +#endif + /* don't leak any physical addresses to userspace */ + info->flags |= FBINFO_HIDE_SMEM_START; /* Need to drop locks to avoid recursive deadlock in * register_framebuffer. This is ok because the only thing left to do is @@ -3081,6 +3104,12 @@ int drm_fb_helper_generic_probe(struct drm_fb_helper *fb_helper, fbi->screen_size = fb->height * fb->pitches[0]; fbi->fix.smem_len = fbi->screen_size; fbi->screen_buffer = buffer->vaddr; + /* Shamelessly leak the physical address to user-space */ +#if IS_ENABLED(CONFIG_DRM_FBDEV_LEAK_PHYS_SMEM) + if (drm_leak_fbdev_smem && fbi->fix.smem_start == 0) + fbi->fix.smem_start = + page_to_phys(virt_to_page(fbi->screen_buffer)); +#endif strcpy(fbi->fix.id, "DRM emulated"); drm_fb_helper_fill_fix(fbi, fb->pitches[0], fb->format->depth); diff --git a/drivers/gpu/drm/drm_panel.c b/drivers/gpu/drm/drm_panel.c index 0db486d10d1c..c33f95e08e1b 100644 --- a/drivers/gpu/drm/drm_panel.c +++ b/drivers/gpu/drm/drm_panel.c @@ -24,7 +24,6 @@ #include <linux/err.h> #include <linux/module.h> -#include <drm/drm_device.h> #include <drm/drm_crtc.h> #include <drm/drm_panel.h> @@ -105,13 +104,6 @@ int drm_panel_attach(struct drm_panel *panel, struct drm_connector *connector) if (panel->connector) return -EBUSY; - panel->link = device_link_add(connector->dev->dev, panel->dev, 0); - if (!panel->link) { - dev_err(panel->dev, "failed to link panel to %s\n", - dev_name(connector->dev->dev)); - return -EINVAL; - } - panel->connector = connector; panel->drm = connector->dev; @@ -133,8 +125,6 @@ EXPORT_SYMBOL(drm_panel_attach); */ int drm_panel_detach(struct drm_panel *panel) { - device_link_del(panel->link); - panel->connector = NULL; panel->drm = NULL; diff --git a/drivers/gpu/drm/drm_panel_orientation_quirks.c b/drivers/gpu/drm/drm_panel_orientation_quirks.c index 280a41d33081..ab4e70e63f6e 100644 --- a/drivers/gpu/drm/drm_panel_orientation_quirks.c +++ b/drivers/gpu/drm/drm_panel_orientation_quirks.c @@ -30,6 +30,12 @@ struct drm_dmi_panel_orientation_data { int orientation; }; +static const struct drm_dmi_panel_orientation_data acer_s1003 = { + .width = 800, + .height = 1280, + .orientation = DRM_MODE_PANEL_ORIENTATION_RIGHT_UP, +}; + static const struct drm_dmi_panel_orientation_data asus_t100ha = { .width = 800, .height = 1280, @@ -75,7 +81,13 @@ static const struct drm_dmi_panel_orientation_data lcd800x1280_rightside_up = { }; static const struct dmi_system_id orientation_data[] = { - { /* Asus T100HA */ + { /* Acer One 10 (S1003) */ + .matches = { + DMI_EXACT_MATCH(DMI_SYS_VENDOR, "Acer"), + DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "One S1003"), + }, + .driver_data = (void *)&acer_s1003, + }, { /* Asus T100HA */ .matches = { DMI_EXACT_MATCH(DMI_SYS_VENDOR, "ASUSTeK COMPUTER INC."), DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "T100HAN"), diff --git a/drivers/gpu/drm/drm_syncobj.c b/drivers/gpu/drm/drm_syncobj.c index b7eaa603f368..c8252cd4c02d 100644 --- a/drivers/gpu/drm/drm_syncobj.c +++ b/drivers/gpu/drm/drm_syncobj.c @@ -157,6 +157,8 @@ static void drm_syncobj_fence_get_or_add_callback(struct drm_syncobj *syncobj, { u64 pt_value = 0; + WARN_ON(*fence); + if (syncobj->type == DRM_SYNCOBJ_TYPE_BINARY) { /*BINARY syncobj always wait on last pt */ pt_value = syncobj->signal_point; @@ -935,6 +937,9 @@ static signed long drm_syncobj_array_wait_timeout(struct drm_syncobj **syncobjs, if (flags & DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT) { for (i = 0; i < count; ++i) { + if (entries[i].fence) + continue; + drm_syncobj_fence_get_or_add_callback(syncobjs[i], &entries[i].fence, &entries[i].syncobj_cb, diff --git a/drivers/gpu/drm/etnaviv/etnaviv_drv.c b/drivers/gpu/drm/etnaviv/etnaviv_drv.c index 9b2720b41571..83c1f46670bf 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_drv.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_drv.c @@ -592,8 +592,6 @@ static int etnaviv_pdev_probe(struct platform_device *pdev) struct device *dev = &pdev->dev; struct component_match *match = NULL; - dma_set_coherent_mask(&pdev->dev, DMA_BIT_MASK(32)); - if (!dev->platform_data) { struct device_node *core_node; @@ -655,13 +653,30 @@ static int __init etnaviv_init(void) for_each_compatible_node(np, NULL, "vivante,gc") { if (!of_device_is_available(np)) continue; - pdev = platform_device_register_simple("etnaviv", -1, - NULL, 0); - if (IS_ERR(pdev)) { - ret = PTR_ERR(pdev); + + pdev = platform_device_alloc("etnaviv", -1); + if (!pdev) { + ret = -ENOMEM; + of_node_put(np); + goto unregister_platform_driver; + } + pdev->dev.coherent_dma_mask = DMA_BIT_MASK(40); + pdev->dev.dma_mask = &pdev->dev.coherent_dma_mask; + + /* + * Apply the same DMA configuration to the virtual etnaviv + * device as the GPU we found. This assumes that all Vivante + * GPUs in the system share the same DMA constraints. + */ + of_dma_configure(&pdev->dev, np, true); + + ret = platform_device_add(pdev); + if (ret) { + platform_device_put(pdev); of_node_put(np); goto unregister_platform_driver; } + etnaviv_drm = pdev; of_node_put(np); break; diff --git a/drivers/gpu/drm/etnaviv/etnaviv_sched.c b/drivers/gpu/drm/etnaviv/etnaviv_sched.c index 69e9b431bf1f..e7c3ed6c9a2e 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_sched.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_sched.c @@ -105,7 +105,7 @@ static void etnaviv_sched_timedout_job(struct drm_sched_job *sched_job) change = dma_addr - gpu->hangcheck_dma_addr; if (change < 0 || change > 16) { gpu->hangcheck_dma_addr = dma_addr; - schedule_delayed_work(&sched_job->work_tdr, + schedule_delayed_work(&sched_job->sched->work_tdr, sched_job->sched->timeout); return; } diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c index b599f74692e5..6f76baf4550a 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_drv.c +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c @@ -149,37 +149,15 @@ static struct drm_driver exynos_drm_driver = { static int exynos_drm_suspend(struct device *dev) { struct drm_device *drm_dev = dev_get_drvdata(dev); - struct exynos_drm_private *private; - - if (!drm_dev) - return 0; - - private = drm_dev->dev_private; - - drm_kms_helper_poll_disable(drm_dev); - exynos_drm_fbdev_suspend(drm_dev); - private->suspend_state = drm_atomic_helper_suspend(drm_dev); - if (IS_ERR(private->suspend_state)) { - exynos_drm_fbdev_resume(drm_dev); - drm_kms_helper_poll_enable(drm_dev); - return PTR_ERR(private->suspend_state); - } - return 0; + return drm_mode_config_helper_suspend(drm_dev); } static void exynos_drm_resume(struct device *dev) { struct drm_device *drm_dev = dev_get_drvdata(dev); - struct exynos_drm_private *private; - - if (!drm_dev) - return; - private = drm_dev->dev_private; - drm_atomic_helper_resume(drm_dev, private->suspend_state); - exynos_drm_fbdev_resume(drm_dev); - drm_kms_helper_poll_enable(drm_dev); + drm_mode_config_helper_resume(drm_dev); } static const struct dev_pm_ops exynos_drm_pm_ops = { diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h index c737c4bd2c19..ec9604f1272b 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_drv.h +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h @@ -92,6 +92,8 @@ struct exynos_drm_plane { #define EXYNOS_DRM_PLANE_CAP_SCALE (1 << 1) #define EXYNOS_DRM_PLANE_CAP_ZPOS (1 << 2) #define EXYNOS_DRM_PLANE_CAP_TILE (1 << 3) +#define EXYNOS_DRM_PLANE_CAP_PIX_BLEND (1 << 4) +#define EXYNOS_DRM_PLANE_CAP_WIN_BLEND (1 << 5) /* * Exynos DRM plane configuration structure. @@ -195,7 +197,6 @@ struct drm_exynos_file_private { */ struct exynos_drm_private { struct drm_fb_helper *fb_helper; - struct drm_atomic_state *suspend_state; struct device *g2d_dev; struct device *dma_dev; diff --git a/drivers/gpu/drm/exynos/exynos_drm_dsi.c b/drivers/gpu/drm/exynos/exynos_drm_dsi.c index 781b82c2c579..07af7758066d 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_dsi.c +++ b/drivers/gpu/drm/exynos/exynos_drm_dsi.c @@ -255,6 +255,7 @@ struct exynos_dsi { struct mipi_dsi_host dsi_host; struct drm_connector connector; struct drm_panel *panel; + struct drm_bridge *out_bridge; struct device *dev; void __iomem *reg_base; @@ -279,7 +280,7 @@ struct exynos_dsi { struct list_head transfer_list; const struct exynos_dsi_driver_data *driver_data; - struct device_node *bridge_node; + struct device_node *in_bridge_node; }; #define host_to_dsi(host) container_of(host, struct exynos_dsi, dsi_host) @@ -1382,29 +1383,37 @@ static void exynos_dsi_enable(struct drm_encoder *encoder) return; pm_runtime_get_sync(dsi->dev); - dsi->state |= DSIM_STATE_ENABLED; - ret = drm_panel_prepare(dsi->panel); - if (ret < 0) { - dsi->state &= ~DSIM_STATE_ENABLED; - pm_runtime_put_sync(dsi->dev); - return; + if (dsi->panel) { + ret = drm_panel_prepare(dsi->panel); + if (ret < 0) + goto err_put_sync; + } else { + drm_bridge_pre_enable(dsi->out_bridge); } exynos_dsi_set_display_mode(dsi); exynos_dsi_set_display_enable(dsi, true); - ret = drm_panel_enable(dsi->panel); - if (ret < 0) { - dsi->state &= ~DSIM_STATE_ENABLED; - exynos_dsi_set_display_enable(dsi, false); - drm_panel_unprepare(dsi->panel); - pm_runtime_put_sync(dsi->dev); - return; + if (dsi->panel) { + ret = drm_panel_enable(dsi->panel); + if (ret < 0) + goto err_display_disable; + } else { + drm_bridge_enable(dsi->out_bridge); } dsi->state |= DSIM_STATE_VIDOUT_AVAILABLE; + return; + +err_display_disable: + exynos_dsi_set_display_enable(dsi, false); + drm_panel_unprepare(dsi->panel); + +err_put_sync: + dsi->state &= ~DSIM_STATE_ENABLED; + pm_runtime_put(dsi->dev); } static void exynos_dsi_disable(struct drm_encoder *encoder) @@ -1417,11 +1426,11 @@ static void exynos_dsi_disable(struct drm_encoder *encoder) dsi->state &= ~DSIM_STATE_VIDOUT_AVAILABLE; drm_panel_disable(dsi->panel); + drm_bridge_disable(dsi->out_bridge); exynos_dsi_set_display_enable(dsi, false); drm_panel_unprepare(dsi->panel); - + drm_bridge_post_disable(dsi->out_bridge); dsi->state &= ~DSIM_STATE_ENABLED; - pm_runtime_put_sync(dsi->dev); } @@ -1499,7 +1508,30 @@ static int exynos_dsi_host_attach(struct mipi_dsi_host *host, struct mipi_dsi_device *device) { struct exynos_dsi *dsi = host_to_dsi(host); - struct drm_device *drm = dsi->connector.dev; + struct drm_encoder *encoder = &dsi->encoder; + struct drm_device *drm = encoder->dev; + struct drm_bridge *out_bridge; + + out_bridge = of_drm_find_bridge(device->dev.of_node); + if (out_bridge) { + drm_bridge_attach(encoder, out_bridge, NULL); + dsi->out_bridge = out_bridge; + encoder->bridge = NULL; + } else { + int ret = exynos_dsi_create_connector(encoder); + + if (ret) { + DRM_ERROR("failed to create connector ret = %d\n", ret); + drm_encoder_cleanup(encoder); + return ret; + } + + dsi->panel = of_drm_find_panel(device->dev.of_node); + if (dsi->panel) { + drm_panel_attach(dsi->panel, &dsi->connector); + dsi->connector.status = connector_status_connected; + } + } /* * This is a temporary solution and should be made by more generic way. @@ -1518,14 +1550,6 @@ static int exynos_dsi_host_attach(struct mipi_dsi_host *host, dsi->lanes = device->lanes; dsi->format = device->format; dsi->mode_flags = device->mode_flags; - dsi->panel = of_drm_find_panel(device->dev.of_node); - if (IS_ERR(dsi->panel)) - dsi->panel = NULL; - - if (dsi->panel) { - drm_panel_attach(dsi->panel, &dsi->connector); - dsi->connector.status = connector_status_connected; - } exynos_drm_crtc_get_by_type(drm, EXYNOS_DISPLAY_TYPE_LCD)->i80_mode = !(dsi->mode_flags & MIPI_DSI_MODE_VIDEO); @@ -1541,19 +1565,21 @@ static int exynos_dsi_host_detach(struct mipi_dsi_host *host, struct mipi_dsi_device *device) { struct exynos_dsi *dsi = host_to_dsi(host); - struct drm_device *drm = dsi->connector.dev; - - mutex_lock(&drm->mode_config.mutex); + struct drm_device *drm = dsi->encoder.dev; if (dsi->panel) { + mutex_lock(&drm->mode_config.mutex); exynos_dsi_disable(&dsi->encoder); drm_panel_detach(dsi->panel); dsi->panel = NULL; dsi->connector.status = connector_status_disconnected; + mutex_unlock(&drm->mode_config.mutex); + } else { + if (dsi->out_bridge->funcs->detach) + dsi->out_bridge->funcs->detach(dsi->out_bridge); + dsi->out_bridge = NULL; } - mutex_unlock(&drm->mode_config.mutex); - if (drm->mode_config.poll_enabled) drm_kms_helper_hotplug_event(drm); @@ -1634,7 +1660,7 @@ static int exynos_dsi_parse_dt(struct exynos_dsi *dsi) if (ret < 0) return ret; - dsi->bridge_node = of_graph_get_remote_node(node, DSI_PORT_IN, 0); + dsi->in_bridge_node = of_graph_get_remote_node(node, DSI_PORT_IN, 0); return 0; } @@ -1645,7 +1671,7 @@ static int exynos_dsi_bind(struct device *dev, struct device *master, struct drm_encoder *encoder = dev_get_drvdata(dev); struct exynos_dsi *dsi = encoder_to_dsi(encoder); struct drm_device *drm_dev = data; - struct drm_bridge *bridge; + struct drm_bridge *in_bridge; int ret; drm_encoder_init(drm_dev, encoder, &exynos_dsi_encoder_funcs, @@ -1657,17 +1683,10 @@ static int exynos_dsi_bind(struct device *dev, struct device *master, if (ret < 0) return ret; - ret = exynos_dsi_create_connector(encoder); - if (ret) { - DRM_ERROR("failed to create connector ret = %d\n", ret); - drm_encoder_cleanup(encoder); - return ret; - } - - if (dsi->bridge_node) { - bridge = of_drm_find_bridge(dsi->bridge_node); - if (bridge) - drm_bridge_attach(encoder, bridge, NULL); + if (dsi->in_bridge_node) { + in_bridge = of_drm_find_bridge(dsi->in_bridge_node); + if (in_bridge) + drm_bridge_attach(encoder, in_bridge, NULL); } return mipi_dsi_host_register(&dsi->dsi_host); @@ -1786,7 +1805,7 @@ static int exynos_dsi_remove(struct platform_device *pdev) { struct exynos_dsi *dsi = platform_get_drvdata(pdev); - of_node_put(dsi->bridge_node); + of_node_put(dsi->in_bridge_node); pm_runtime_disable(&pdev->dev); diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c index 132dd52d0ac7..918dd2c82209 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c +++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c @@ -270,20 +270,3 @@ void exynos_drm_fbdev_fini(struct drm_device *dev) private->fb_helper = NULL; } -void exynos_drm_fbdev_suspend(struct drm_device *dev) -{ - struct exynos_drm_private *private = dev->dev_private; - - console_lock(); - drm_fb_helper_set_suspend(private->fb_helper, 1); - console_unlock(); -} - -void exynos_drm_fbdev_resume(struct drm_device *dev) -{ - struct exynos_drm_private *private = dev->dev_private; - - console_lock(); - drm_fb_helper_set_suspend(private->fb_helper, 0); - console_unlock(); -} diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.h b/drivers/gpu/drm/exynos/exynos_drm_fbdev.h index b33847223a85..6840b6aadbc0 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.h +++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.h @@ -19,8 +19,6 @@ int exynos_drm_fbdev_init(struct drm_device *dev); void exynos_drm_fbdev_fini(struct drm_device *dev); -void exynos_drm_fbdev_suspend(struct drm_device *drm); -void exynos_drm_fbdev_resume(struct drm_device *drm); #else @@ -39,14 +37,6 @@ static inline void exynos_drm_fbdev_restore_mode(struct drm_device *dev) #define exynos_drm_output_poll_changed (NULL) -static inline void exynos_drm_fbdev_suspend(struct drm_device *drm) -{ -} - -static inline void exynos_drm_fbdev_resume(struct drm_device *drm) -{ -} - #endif #endif diff --git a/drivers/gpu/drm/exynos/exynos_drm_gsc.c b/drivers/gpu/drm/exynos/exynos_drm_gsc.c index 7ba414b52faa..ce15d46bfce8 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_gsc.c +++ b/drivers/gpu/drm/exynos/exynos_drm_gsc.c @@ -448,7 +448,7 @@ static void gsc_handle_irq(struct gsc_context *ctx, bool enable, } -static void gsc_src_set_fmt(struct gsc_context *ctx, u32 fmt) +static void gsc_src_set_fmt(struct gsc_context *ctx, u32 fmt, bool tiled) { u32 cfg; @@ -514,6 +514,9 @@ static void gsc_src_set_fmt(struct gsc_context *ctx, u32 fmt) break; } + if (tiled) + cfg |= (GSC_IN_TILE_C_16x8 | GSC_IN_TILE_MODE); + gsc_write(cfg, GSC_IN_CON); } @@ -632,7 +635,7 @@ static void gsc_src_set_addr(struct gsc_context *ctx, u32 buf_id, gsc_src_set_buf_seq(ctx, buf_id, true); } -static void gsc_dst_set_fmt(struct gsc_context *ctx, u32 fmt) +static void gsc_dst_set_fmt(struct gsc_context *ctx, u32 fmt, bool tiled) { u32 cfg; @@ -698,6 +701,9 @@ static void gsc_dst_set_fmt(struct gsc_context *ctx, u32 fmt) break; } + if (tiled) + cfg |= (GSC_IN_TILE_C_16x8 | GSC_OUT_TILE_MODE); + gsc_write(cfg, GSC_OUT_CON); } @@ -1122,11 +1128,11 @@ static int gsc_commit(struct exynos_drm_ipp *ipp, return ret; } - gsc_src_set_fmt(ctx, task->src.buf.fourcc); + gsc_src_set_fmt(ctx, task->src.buf.fourcc, task->src.buf.modifier); gsc_src_set_transf(ctx, task->transform.rotation); gsc_src_set_size(ctx, &task->src); gsc_src_set_addr(ctx, 0, &task->src); - gsc_dst_set_fmt(ctx, task->dst.buf.fourcc); + gsc_dst_set_fmt(ctx, task->dst.buf.fourcc, task->dst.buf.modifier); gsc_dst_set_size(ctx, &task->dst); gsc_dst_set_addr(ctx, 0, &task->dst); gsc_set_prescaler(ctx, &ctx->sc, &task->src.rect, &task->dst.rect); @@ -1200,6 +1206,10 @@ static const unsigned int gsc_formats[] = { DRM_FORMAT_YUV420, DRM_FORMAT_YVU420, DRM_FORMAT_YUV422, }; +static const unsigned int gsc_tiled_formats[] = { + DRM_FORMAT_NV12, DRM_FORMAT_NV21, +}; + static int gsc_probe(struct platform_device *pdev) { struct device *dev = &pdev->dev; @@ -1207,23 +1217,24 @@ static int gsc_probe(struct platform_device *pdev) struct exynos_drm_ipp_formats *formats; struct gsc_context *ctx; struct resource *res; - int ret, i; + int num_formats, ret, i, j; ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL); if (!ctx) return -ENOMEM; - formats = devm_kcalloc(dev, - ARRAY_SIZE(gsc_formats), sizeof(*formats), - GFP_KERNEL); - if (!formats) - return -ENOMEM; - driver_data = (struct gsc_driverdata *)of_device_get_match_data(dev); ctx->dev = dev; ctx->num_clocks = driver_data->num_clocks; ctx->clk_names = driver_data->clk_names; + /* construct formats/limits array */ + num_formats = ARRAY_SIZE(gsc_formats) + ARRAY_SIZE(gsc_tiled_formats); + formats = devm_kcalloc(dev, num_formats, sizeof(*formats), GFP_KERNEL); + if (!formats) + return -ENOMEM; + + /* linear formats */ for (i = 0; i < ARRAY_SIZE(gsc_formats); i++) { formats[i].fourcc = gsc_formats[i]; formats[i].type = DRM_EXYNOS_IPP_FORMAT_SOURCE | @@ -1231,8 +1242,19 @@ static int gsc_probe(struct platform_device *pdev) formats[i].limits = driver_data->limits; formats[i].num_limits = driver_data->num_limits; } + + /* tiled formats */ + for (j = i, i = 0; i < ARRAY_SIZE(gsc_tiled_formats); j++, i++) { + formats[j].fourcc = gsc_tiled_formats[i]; + formats[j].modifier = DRM_FORMAT_MOD_SAMSUNG_16_16_TILE; + formats[j].type = DRM_EXYNOS_IPP_FORMAT_SOURCE | + DRM_EXYNOS_IPP_FORMAT_DESTINATION; + formats[j].limits = driver_data->limits; + formats[j].num_limits = driver_data->num_limits; + } + ctx->formats = formats; - ctx->num_formats = ARRAY_SIZE(gsc_formats); + ctx->num_formats = num_formats; /* clock control */ for (i = 0; i < ctx->num_clocks; i++) { diff --git a/drivers/gpu/drm/exynos/exynos_drm_plane.c b/drivers/gpu/drm/exynos/exynos_drm_plane.c index dba29aec59b4..df0508e0e49e 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_plane.c +++ b/drivers/gpu/drm/exynos/exynos_drm_plane.c @@ -131,16 +131,14 @@ static void exynos_drm_plane_reset(struct drm_plane *plane) if (plane->state) { exynos_state = to_exynos_plane_state(plane->state); - if (exynos_state->base.fb) - drm_framebuffer_put(exynos_state->base.fb); + __drm_atomic_helper_plane_destroy_state(plane->state); kfree(exynos_state); plane->state = NULL; } exynos_state = kzalloc(sizeof(*exynos_state), GFP_KERNEL); if (exynos_state) { - plane->state = &exynos_state->base; - plane->state->plane = plane; + __drm_atomic_helper_plane_reset(plane, &exynos_state->base); plane->state->zpos = exynos_plane->config->zpos; } } @@ -300,6 +298,10 @@ int exynos_plane_init(struct drm_device *dev, const struct exynos_drm_plane_config *config) { int err; + unsigned int supported_modes = BIT(DRM_MODE_BLEND_PIXEL_NONE) | + BIT(DRM_MODE_BLEND_PREMULTI) | + BIT(DRM_MODE_BLEND_COVERAGE); + struct drm_plane *plane = &exynos_plane->base; err = drm_universal_plane_init(dev, &exynos_plane->base, 1 << dev->mode_config.num_crtc, @@ -320,5 +322,11 @@ int exynos_plane_init(struct drm_device *dev, exynos_plane_attach_zpos_property(&exynos_plane->base, config->zpos, !(config->capabilities & EXYNOS_DRM_PLANE_CAP_ZPOS)); + if (config->capabilities & EXYNOS_DRM_PLANE_CAP_PIX_BLEND) + drm_plane_create_blend_mode_property(plane, supported_modes); + + if (config->capabilities & EXYNOS_DRM_PLANE_CAP_WIN_BLEND) + drm_plane_create_alpha_property(plane); + return 0; } diff --git a/drivers/gpu/drm/exynos/exynos_drm_scaler.c b/drivers/gpu/drm/exynos/exynos_drm_scaler.c index 0ddb6eec7b11..cd66774e817d 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_scaler.c +++ b/drivers/gpu/drm/exynos/exynos_drm_scaler.c @@ -49,56 +49,46 @@ struct scaler_context { const struct scaler_data *scaler_data; }; -static u32 scaler_get_format(u32 drm_fmt) +struct scaler_format { + u32 drm_fmt; + u32 internal_fmt; + u32 chroma_tile_w; + u32 chroma_tile_h; +}; + +static const struct scaler_format scaler_formats[] = { + { DRM_FORMAT_NV12, SCALER_YUV420_2P_UV, 8, 8 }, + { DRM_FORMAT_NV21, SCALER_YUV420_2P_VU, 8, 8 }, + { DRM_FORMAT_YUV420, SCALER_YUV420_3P, 8, 8 }, + { DRM_FORMAT_YUYV, SCALER_YUV422_1P_YUYV, 16, 16 }, + { DRM_FORMAT_UYVY, SCALER_YUV422_1P_UYVY, 16, 16 }, + { DRM_FORMAT_YVYU, SCALER_YUV422_1P_YVYU, 16, 16 }, + { DRM_FORMAT_NV16, SCALER_YUV422_2P_UV, 8, 16 }, + { DRM_FORMAT_NV61, SCALER_YUV422_2P_VU, 8, 16 }, + { DRM_FORMAT_YUV422, SCALER_YUV422_3P, 8, 16 }, + { DRM_FORMAT_NV24, SCALER_YUV444_2P_UV, 16, 16 }, + { DRM_FORMAT_NV42, SCALER_YUV444_2P_VU, 16, 16 }, + { DRM_FORMAT_YUV444, SCALER_YUV444_3P, 16, 16 }, + { DRM_FORMAT_RGB565, SCALER_RGB_565, 0, 0 }, + { DRM_FORMAT_XRGB1555, SCALER_ARGB1555, 0, 0 }, + { DRM_FORMAT_ARGB1555, SCALER_ARGB1555, 0, 0 }, + { DRM_FORMAT_XRGB4444, SCALER_ARGB4444, 0, 0 }, + { DRM_FORMAT_ARGB4444, SCALER_ARGB4444, 0, 0 }, + { DRM_FORMAT_XRGB8888, SCALER_ARGB8888, 0, 0 }, + { DRM_FORMAT_ARGB8888, SCALER_ARGB8888, 0, 0 }, + { DRM_FORMAT_RGBX8888, SCALER_RGBA8888, 0, 0 }, + { DRM_FORMAT_RGBA8888, SCALER_RGBA8888, 0, 0 }, +}; + +static const struct scaler_format *scaler_get_format(u32 drm_fmt) { - switch (drm_fmt) { - case DRM_FORMAT_NV12: - return SCALER_YUV420_2P_UV; - case DRM_FORMAT_NV21: - return SCALER_YUV420_2P_VU; - case DRM_FORMAT_YUV420: - return SCALER_YUV420_3P; - case DRM_FORMAT_YUYV: - return SCALER_YUV422_1P_YUYV; - case DRM_FORMAT_UYVY: - return SCALER_YUV422_1P_UYVY; - case DRM_FORMAT_YVYU: - return SCALER_YUV422_1P_YVYU; - case DRM_FORMAT_NV16: - return SCALER_YUV422_2P_UV; - case DRM_FORMAT_NV61: - return SCALER_YUV422_2P_VU; - case DRM_FORMAT_YUV422: - return SCALER_YUV422_3P; - case DRM_FORMAT_NV24: - return SCALER_YUV444_2P_UV; - case DRM_FORMAT_NV42: - return SCALER_YUV444_2P_VU; - case DRM_FORMAT_YUV444: - return SCALER_YUV444_3P; - case DRM_FORMAT_RGB565: - return SCALER_RGB_565; - case DRM_FORMAT_XRGB1555: - return SCALER_ARGB1555; - case DRM_FORMAT_ARGB1555: - return SCALER_ARGB1555; - case DRM_FORMAT_XRGB4444: - return SCALER_ARGB4444; - case DRM_FORMAT_ARGB4444: - return SCALER_ARGB4444; - case DRM_FORMAT_XRGB8888: - return SCALER_ARGB8888; - case DRM_FORMAT_ARGB8888: - return SCALER_ARGB8888; - case DRM_FORMAT_RGBX8888: - return SCALER_RGBA8888; - case DRM_FORMAT_RGBA8888: - return SCALER_RGBA8888; - default: - break; - } + int i; - return 0; + for (i = 0; i < ARRAY_SIZE(scaler_formats); i++) + if (scaler_formats[i].drm_fmt == drm_fmt) + return &scaler_formats[i]; + + return NULL; } static inline int scaler_reset(struct scaler_context *scaler) @@ -152,11 +142,11 @@ static inline void scaler_enable_int(struct scaler_context *scaler) } static inline void scaler_set_src_fmt(struct scaler_context *scaler, - u32 src_fmt) + u32 src_fmt, u32 tile) { u32 val; - val = SCALER_SRC_CFG_SET_COLOR_FORMAT(src_fmt); + val = SCALER_SRC_CFG_SET_COLOR_FORMAT(src_fmt) | (tile << 10); scaler_write(val, SCALER_SRC_CFG); } @@ -188,15 +178,20 @@ static inline void scaler_set_src_span(struct scaler_context *scaler, scaler_write(val, SCALER_SRC_SPAN); } -static inline void scaler_set_src_luma_pos(struct scaler_context *scaler, - struct drm_exynos_ipp_task_rect *src_pos) +static inline void scaler_set_src_luma_chroma_pos(struct scaler_context *scaler, + struct drm_exynos_ipp_task_rect *src_pos, + const struct scaler_format *fmt) { u32 val; val = SCALER_SRC_Y_POS_SET_YH_POS(src_pos->x << 2); val |= SCALER_SRC_Y_POS_SET_YV_POS(src_pos->y << 2); scaler_write(val, SCALER_SRC_Y_POS); - scaler_write(val, SCALER_SRC_C_POS); /* ATTENTION! */ + val = SCALER_SRC_C_POS_SET_CH_POS( + (src_pos->x * fmt->chroma_tile_w / 16) << 2); + val |= SCALER_SRC_C_POS_SET_CV_POS( + (src_pos->y * fmt->chroma_tile_h / 16) << 2); + scaler_write(val, SCALER_SRC_C_POS); } static inline void scaler_set_src_wh(struct scaler_context *scaler, @@ -366,11 +361,12 @@ static int scaler_commit(struct exynos_drm_ipp *ipp, struct scaler_context *scaler = container_of(ipp, struct scaler_context, ipp); - u32 src_fmt = scaler_get_format(task->src.buf.fourcc); struct drm_exynos_ipp_task_rect *src_pos = &task->src.rect; - - u32 dst_fmt = scaler_get_format(task->dst.buf.fourcc); struct drm_exynos_ipp_task_rect *dst_pos = &task->dst.rect; + const struct scaler_format *src_fmt, *dst_fmt; + + src_fmt = scaler_get_format(task->src.buf.fourcc); + dst_fmt = scaler_get_format(task->dst.buf.fourcc); pm_runtime_get_sync(scaler->dev); if (scaler_reset(scaler)) { @@ -380,13 +376,14 @@ static int scaler_commit(struct exynos_drm_ipp *ipp, scaler->task = task; - scaler_set_src_fmt(scaler, src_fmt); + scaler_set_src_fmt( + scaler, src_fmt->internal_fmt, task->src.buf.modifier != 0); scaler_set_src_base(scaler, &task->src); scaler_set_src_span(scaler, &task->src); - scaler_set_src_luma_pos(scaler, src_pos); + scaler_set_src_luma_chroma_pos(scaler, src_pos, src_fmt); scaler_set_src_wh(scaler, src_pos); - scaler_set_dst_fmt(scaler, dst_fmt); + scaler_set_dst_fmt(scaler, dst_fmt->internal_fmt); scaler_set_dst_base(scaler, &task->dst); scaler_set_dst_span(scaler, &task->dst); scaler_set_dst_luma_pos(scaler, dst_pos); @@ -617,6 +614,16 @@ static const struct drm_exynos_ipp_limit scaler_5420_one_pixel_limits[] = { .v = { 65536 * 1 / 4, 65536 * 16 }) }, }; +static const struct drm_exynos_ipp_limit scaler_5420_tile_limits[] = { + { IPP_SIZE_LIMIT(BUFFER, .h = { 16, SZ_8K }, .v = { 16, SZ_8K })}, + { IPP_SIZE_LIMIT(AREA, .h.align = 16, .v.align = 16) }, + { IPP_SCALE_LIMIT(.h = {1, 1}, .v = {1, 1})}, + { } +}; + +#define IPP_SRCDST_TILE_FORMAT(f, l) \ + IPP_SRCDST_MFORMAT(f, DRM_FORMAT_MOD_SAMSUNG_16_16_TILE, (l)) + static const struct exynos_drm_ipp_formats exynos5420_formats[] = { /* SCALER_YUV420_2P_UV */ { IPP_SRCDST_FORMAT(NV21, scaler_5420_two_pixel_hv_limits) }, @@ -680,6 +687,18 @@ static const struct exynos_drm_ipp_formats exynos5420_formats[] = { /* SCALER_RGBA8888 */ { IPP_SRCDST_FORMAT(RGBA8888, scaler_5420_one_pixel_limits) }, + + /* SCALER_YUV420_2P_UV TILE */ + { IPP_SRCDST_TILE_FORMAT(NV21, scaler_5420_tile_limits) }, + + /* SCALER_YUV420_2P_VU TILE */ + { IPP_SRCDST_TILE_FORMAT(NV12, scaler_5420_tile_limits) }, + + /* SCALER_YUV420_3P TILE */ + { IPP_SRCDST_TILE_FORMAT(YUV420, scaler_5420_tile_limits) }, + + /* SCALER_YUV422_1P_YUYV TILE */ + { IPP_SRCDST_TILE_FORMAT(YUYV, scaler_5420_tile_limits) }, }; static const struct scaler_data exynos5420_data = { diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c index ffbf4a950f69..e3a4ecbc503b 100644 --- a/drivers/gpu/drm/exynos/exynos_mixer.c +++ b/drivers/gpu/drm/exynos/exynos_mixer.c @@ -131,14 +131,18 @@ static const struct exynos_drm_plane_config plane_configs[MIXER_WIN_NR] = { .pixel_formats = mixer_formats, .num_pixel_formats = ARRAY_SIZE(mixer_formats), .capabilities = EXYNOS_DRM_PLANE_CAP_DOUBLE | - EXYNOS_DRM_PLANE_CAP_ZPOS, + EXYNOS_DRM_PLANE_CAP_ZPOS | + EXYNOS_DRM_PLANE_CAP_PIX_BLEND | + EXYNOS_DRM_PLANE_CAP_WIN_BLEND, }, { .zpos = 1, .type = DRM_PLANE_TYPE_CURSOR, .pixel_formats = mixer_formats, .num_pixel_formats = ARRAY_SIZE(mixer_formats), .capabilities = EXYNOS_DRM_PLANE_CAP_DOUBLE | - EXYNOS_DRM_PLANE_CAP_ZPOS, + EXYNOS_DRM_PLANE_CAP_ZPOS | + EXYNOS_DRM_PLANE_CAP_PIX_BLEND | + EXYNOS_DRM_PLANE_CAP_WIN_BLEND, }, { .zpos = 2, .type = DRM_PLANE_TYPE_OVERLAY, @@ -146,7 +150,8 @@ static const struct exynos_drm_plane_config plane_configs[MIXER_WIN_NR] = { .num_pixel_formats = ARRAY_SIZE(vp_formats), .capabilities = EXYNOS_DRM_PLANE_CAP_SCALE | EXYNOS_DRM_PLANE_CAP_ZPOS | - EXYNOS_DRM_PLANE_CAP_TILE, + EXYNOS_DRM_PLANE_CAP_TILE | + EXYNOS_DRM_PLANE_CAP_WIN_BLEND, }, }; @@ -309,31 +314,42 @@ static void vp_default_filter(struct mixer_context *ctx) } static void mixer_cfg_gfx_blend(struct mixer_context *ctx, unsigned int win, - bool alpha) + unsigned int pixel_alpha, unsigned int alpha) { + u32 win_alpha = alpha >> 8; u32 val; val = MXR_GRP_CFG_COLOR_KEY_DISABLE; /* no blank key */ - if (alpha) { - /* blending based on pixel alpha */ + switch (pixel_alpha) { + case DRM_MODE_BLEND_PIXEL_NONE: + break; + case DRM_MODE_BLEND_COVERAGE: + val |= MXR_GRP_CFG_PIXEL_BLEND_EN; + break; + case DRM_MODE_BLEND_PREMULTI: + default: val |= MXR_GRP_CFG_BLEND_PRE_MUL; val |= MXR_GRP_CFG_PIXEL_BLEND_EN; + break; + } + + if (alpha != DRM_BLEND_ALPHA_OPAQUE) { + val |= MXR_GRP_CFG_WIN_BLEND_EN; + val |= win_alpha; } mixer_reg_writemask(ctx, MXR_GRAPHIC_CFG(win), val, MXR_GRP_CFG_MISC_MASK); } -static void mixer_cfg_vp_blend(struct mixer_context *ctx) +static void mixer_cfg_vp_blend(struct mixer_context *ctx, unsigned int alpha) { - u32 val; + u32 win_alpha = alpha >> 8; + u32 val = 0; - /* - * No blending at the moment since the NV12/NV21 pixelformats don't - * have an alpha channel. However the mixer supports a global alpha - * value for a layer. Once this functionality is exposed, we can - * support blending of the video layer through this. - */ - val = 0; + if (alpha != DRM_BLEND_ALPHA_OPAQUE) { + val |= MXR_VID_CFG_BLEND_EN; + val |= win_alpha; + } mixer_reg_write(ctx, MXR_VIDEO_CFG, val); } @@ -529,7 +545,7 @@ static void vp_video_buffer(struct mixer_context *ctx, vp_reg_write(ctx, VP_BOT_C_PTR, chroma_addr[1]); mixer_cfg_layer(ctx, plane->index, priority, true); - mixer_cfg_vp_blend(ctx); + mixer_cfg_vp_blend(ctx, state->base.alpha); spin_unlock_irqrestore(&ctx->reg_slock, flags); @@ -553,10 +569,16 @@ static void mixer_graph_buffer(struct mixer_context *ctx, unsigned int win = plane->index; unsigned int x_ratio = 0, y_ratio = 0; unsigned int dst_x_offset, dst_y_offset; + unsigned int pixel_alpha; dma_addr_t dma_addr; unsigned int fmt; u32 val; + if (fb->format->has_alpha) + pixel_alpha = state->base.pixel_blend_mode; + else + pixel_alpha = DRM_MODE_BLEND_PIXEL_NONE; + switch (fb->format->format) { case DRM_FORMAT_XRGB4444: case DRM_FORMAT_ARGB4444: @@ -616,7 +638,7 @@ static void mixer_graph_buffer(struct mixer_context *ctx, mixer_reg_write(ctx, MXR_GRAPHIC_BASE(win), dma_addr); mixer_cfg_layer(ctx, win, priority, true); - mixer_cfg_gfx_blend(ctx, win, fb->format->has_alpha); + mixer_cfg_gfx_blend(ctx, win, pixel_alpha, state->base.alpha); /* layer update mandatory for mixer 16.0.33.0 */ if (ctx->mxr_ver == MXR_VER_16_0_33_0 || diff --git a/drivers/gpu/drm/exynos/regs-mixer.h b/drivers/gpu/drm/exynos/regs-mixer.h index 189cfa2470a8..d2b8194a07bf 100644 --- a/drivers/gpu/drm/exynos/regs-mixer.h +++ b/drivers/gpu/drm/exynos/regs-mixer.h @@ -109,12 +109,15 @@ #define MXR_CFG_SCAN_HD (1 << 0) #define MXR_CFG_SCAN_MASK 0x47 +/* bits for MXR_VIDEO_CFG */ +#define MXR_VID_CFG_BLEND_EN (1 << 16) + /* bits for MXR_GRAPHICn_CFG */ #define MXR_GRP_CFG_COLOR_KEY_DISABLE (1 << 21) #define MXR_GRP_CFG_BLEND_PRE_MUL (1 << 20) #define MXR_GRP_CFG_WIN_BLEND_EN (1 << 17) #define MXR_GRP_CFG_PIXEL_BLEND_EN (1 << 16) -#define MXR_GRP_CFG_MISC_MASK ((3 << 16) | (3 << 20)) +#define MXR_GRP_CFG_MISC_MASK ((3 << 16) | (3 << 20) | 0xff) #define MXR_GRP_CFG_FORMAT_VAL(x) MXR_MASK_VAL(x, 11, 8) #define MXR_GRP_CFG_FORMAT_MASK MXR_GRP_CFG_FORMAT_VAL(~0) #define MXR_GRP_CFG_ALPHA_VAL(x) MXR_MASK_VAL(x, 7, 0) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 1f7051e97afb..4f3ac0a12889 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -4117,6 +4117,17 @@ i915_ring_test_irq_set(void *data, u64 val) { struct drm_i915_private *i915 = data; + /* GuC keeps the user interrupt permanently enabled for submission */ + if (USES_GUC_SUBMISSION(i915)) + return -ENODEV; + + /* + * From icl, we can no longer individually mask interrupt generation + * from each engine. + */ + if (INTEL_GEN(i915) >= 11) + return -ENODEV; + val &= INTEL_INFO(i915)->ring_mask; DRM_DEBUG_DRIVER("Masking interrupts on rings 0x%08llx\n", val); @@ -4178,7 +4189,7 @@ i915_drop_caches_set(void *data, u64 val) I915_WAIT_LOCKED, MAX_SCHEDULE_TIMEOUT); - if (val & DROP_RESET_SEQNO) { + if (ret == 0 && val & DROP_RESET_SEQNO) { intel_runtime_pm_get(i915); ret = i915_gem_set_global_seqno(&i915->drm, 1); intel_runtime_pm_put(i915); diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 2ddf8538cb47..44e2c0f5ec50 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -1063,6 +1063,300 @@ static void intel_sanitize_options(struct drm_i915_private *dev_priv) intel_gvt_sanitize_options(dev_priv); } +static enum dram_rank skl_get_dimm_rank(u8 size, u32 rank) +{ + if (size == 0) + return I915_DRAM_RANK_INVALID; + if (rank == SKL_DRAM_RANK_SINGLE) + return I915_DRAM_RANK_SINGLE; + else if (rank == SKL_DRAM_RANK_DUAL) + return I915_DRAM_RANK_DUAL; + + return I915_DRAM_RANK_INVALID; +} + +static bool +skl_is_16gb_dimm(enum dram_rank rank, u8 size, u8 width) +{ + if (rank == I915_DRAM_RANK_SINGLE && width == 8 && size == 16) + return true; + else if (rank == I915_DRAM_RANK_DUAL && width == 8 && size == 32) + return true; + else if (rank == SKL_DRAM_RANK_SINGLE && width == 16 && size == 8) + return true; + else if (rank == SKL_DRAM_RANK_DUAL && width == 16 && size == 16) + return true; + + return false; +} + +static int +skl_dram_get_channel_info(struct dram_channel_info *ch, u32 val) +{ + u32 tmp_l, tmp_s; + u32 s_val = val >> SKL_DRAM_S_SHIFT; + + if (!val) + return -EINVAL; + + tmp_l = val & SKL_DRAM_SIZE_MASK; + tmp_s = s_val & SKL_DRAM_SIZE_MASK; + + if (tmp_l == 0 && tmp_s == 0) + return -EINVAL; + + ch->l_info.size = tmp_l; + ch->s_info.size = tmp_s; + + tmp_l = (val & SKL_DRAM_WIDTH_MASK) >> SKL_DRAM_WIDTH_SHIFT; + tmp_s = (s_val & SKL_DRAM_WIDTH_MASK) >> SKL_DRAM_WIDTH_SHIFT; + ch->l_info.width = (1 << tmp_l) * 8; + ch->s_info.width = (1 << tmp_s) * 8; + + tmp_l = val & SKL_DRAM_RANK_MASK; + tmp_s = s_val & SKL_DRAM_RANK_MASK; + ch->l_info.rank = skl_get_dimm_rank(ch->l_info.size, tmp_l); + ch->s_info.rank = skl_get_dimm_rank(ch->s_info.size, tmp_s); + + if (ch->l_info.rank == I915_DRAM_RANK_DUAL || + ch->s_info.rank == I915_DRAM_RANK_DUAL) + ch->rank = I915_DRAM_RANK_DUAL; + else if (ch->l_info.rank == I915_DRAM_RANK_SINGLE && + ch->s_info.rank == I915_DRAM_RANK_SINGLE) + ch->rank = I915_DRAM_RANK_DUAL; + else + ch->rank = I915_DRAM_RANK_SINGLE; + + ch->is_16gb_dimm = skl_is_16gb_dimm(ch->l_info.rank, ch->l_info.size, + ch->l_info.width) || + skl_is_16gb_dimm(ch->s_info.rank, ch->s_info.size, + ch->s_info.width); + + DRM_DEBUG_KMS("(size:width:rank) L(%dGB:X%d:%s) S(%dGB:X%d:%s)\n", + ch->l_info.size, ch->l_info.width, + ch->l_info.rank ? "dual" : "single", + ch->s_info.size, ch->s_info.width, + ch->s_info.rank ? "dual" : "single"); + + return 0; +} + +static bool +intel_is_dram_symmetric(u32 val_ch0, u32 val_ch1, + struct dram_channel_info *ch0) +{ + return (val_ch0 == val_ch1 && + (ch0->s_info.size == 0 || + (ch0->l_info.size == ch0->s_info.size && + ch0->l_info.width == ch0->s_info.width && + ch0->l_info.rank == ch0->s_info.rank))); +} + +static int +skl_dram_get_channels_info(struct drm_i915_private *dev_priv) +{ + struct dram_info *dram_info = &dev_priv->dram_info; + struct dram_channel_info ch0, ch1; + u32 val_ch0, val_ch1; + int ret; + + val_ch0 = I915_READ(SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN); + ret = skl_dram_get_channel_info(&ch0, val_ch0); + if (ret == 0) + dram_info->num_channels++; + + val_ch1 = I915_READ(SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN); + ret = skl_dram_get_channel_info(&ch1, val_ch1); + if (ret == 0) + dram_info->num_channels++; + + if (dram_info->num_channels == 0) { + DRM_INFO("Number of memory channels is zero\n"); + return -EINVAL; + } + + dram_info->valid_dimm = true; + + /* + * If any of the channel is single rank channel, worst case output + * will be same as if single rank memory, so consider single rank + * memory. + */ + if (ch0.rank == I915_DRAM_RANK_SINGLE || + ch1.rank == I915_DRAM_RANK_SINGLE) + dram_info->rank = I915_DRAM_RANK_SINGLE; + else + dram_info->rank = max(ch0.rank, ch1.rank); + + if (dram_info->rank == I915_DRAM_RANK_INVALID) { + DRM_INFO("couldn't get memory rank information\n"); + return -EINVAL; + } + + if (ch0.is_16gb_dimm || ch1.is_16gb_dimm) + dram_info->is_16gb_dimm = true; + + dev_priv->dram_info.symmetric_memory = intel_is_dram_symmetric(val_ch0, + val_ch1, + &ch0); + + DRM_DEBUG_KMS("memory configuration is %sSymmetric memory\n", + dev_priv->dram_info.symmetric_memory ? "" : "not "); + return 0; +} + +static int +skl_get_dram_info(struct drm_i915_private *dev_priv) +{ + struct dram_info *dram_info = &dev_priv->dram_info; + u32 mem_freq_khz, val; + int ret; + + ret = skl_dram_get_channels_info(dev_priv); + if (ret) + return ret; + + val = I915_READ(SKL_MC_BIOS_DATA_0_0_0_MCHBAR_PCU); + mem_freq_khz = DIV_ROUND_UP((val & SKL_REQ_DATA_MASK) * + SKL_MEMORY_FREQ_MULTIPLIER_HZ, 1000); + + dram_info->bandwidth_kbps = dram_info->num_channels * + mem_freq_khz * 8; + + if (dram_info->bandwidth_kbps == 0) { + DRM_INFO("Couldn't get system memory bandwidth\n"); + return -EINVAL; + } + + dram_info->valid = true; + return 0; +} + +static int +bxt_get_dram_info(struct drm_i915_private *dev_priv) +{ + struct dram_info *dram_info = &dev_priv->dram_info; + u32 dram_channels; + u32 mem_freq_khz, val; + u8 num_active_channels; + int i; + + val = I915_READ(BXT_P_CR_MC_BIOS_REQ_0_0_0); + mem_freq_khz = DIV_ROUND_UP((val & BXT_REQ_DATA_MASK) * + BXT_MEMORY_FREQ_MULTIPLIER_HZ, 1000); + + dram_channels = val & BXT_DRAM_CHANNEL_ACTIVE_MASK; + num_active_channels = hweight32(dram_channels); + + /* Each active bit represents 4-byte channel */ + dram_info->bandwidth_kbps = (mem_freq_khz * num_active_channels * 4); + + if (dram_info->bandwidth_kbps == 0) { + DRM_INFO("Couldn't get system memory bandwidth\n"); + return -EINVAL; + } + + /* + * Now read each DUNIT8/9/10/11 to check the rank of each dimms. + */ + for (i = BXT_D_CR_DRP0_DUNIT_START; i <= BXT_D_CR_DRP0_DUNIT_END; i++) { + u8 size, width; + enum dram_rank rank; + u32 tmp; + + val = I915_READ(BXT_D_CR_DRP0_DUNIT(i)); + if (val == 0xFFFFFFFF) + continue; + + dram_info->num_channels++; + tmp = val & BXT_DRAM_RANK_MASK; + + if (tmp == BXT_DRAM_RANK_SINGLE) + rank = I915_DRAM_RANK_SINGLE; + else if (tmp == BXT_DRAM_RANK_DUAL) + rank = I915_DRAM_RANK_DUAL; + else + rank = I915_DRAM_RANK_INVALID; + + tmp = val & BXT_DRAM_SIZE_MASK; + if (tmp == BXT_DRAM_SIZE_4GB) + size = 4; + else if (tmp == BXT_DRAM_SIZE_6GB) + size = 6; + else if (tmp == BXT_DRAM_SIZE_8GB) + size = 8; + else if (tmp == BXT_DRAM_SIZE_12GB) + size = 12; + else if (tmp == BXT_DRAM_SIZE_16GB) + size = 16; + else + size = 0; + + tmp = (val & BXT_DRAM_WIDTH_MASK) >> BXT_DRAM_WIDTH_SHIFT; + width = (1 << tmp) * 8; + DRM_DEBUG_KMS("dram size:%dGB width:X%d rank:%s\n", size, + width, rank == I915_DRAM_RANK_SINGLE ? "single" : + rank == I915_DRAM_RANK_DUAL ? "dual" : "unknown"); + + /* + * If any of the channel is single rank channel, + * worst case output will be same as if single rank + * memory, so consider single rank memory. + */ + if (dram_info->rank == I915_DRAM_RANK_INVALID) + dram_info->rank = rank; + else if (rank == I915_DRAM_RANK_SINGLE) + dram_info->rank = I915_DRAM_RANK_SINGLE; + } + + if (dram_info->rank == I915_DRAM_RANK_INVALID) { + DRM_INFO("couldn't get memory rank information\n"); + return -EINVAL; + } + + dram_info->valid_dimm = true; + dram_info->valid = true; + return 0; +} + +static void +intel_get_dram_info(struct drm_i915_private *dev_priv) +{ + struct dram_info *dram_info = &dev_priv->dram_info; + char bandwidth_str[32]; + int ret; + + dram_info->valid = false; + dram_info->valid_dimm = false; + dram_info->is_16gb_dimm = false; + dram_info->rank = I915_DRAM_RANK_INVALID; + dram_info->bandwidth_kbps = 0; + dram_info->num_channels = 0; + + if (INTEL_GEN(dev_priv) < 9 || IS_GEMINILAKE(dev_priv)) + return; + + /* Need to calculate bandwidth only for Gen9 */ + if (IS_BROXTON(dev_priv)) + ret = bxt_get_dram_info(dev_priv); + else if (INTEL_GEN(dev_priv) == 9) + ret = skl_get_dram_info(dev_priv); + else + ret = skl_dram_get_channels_info(dev_priv); + if (ret) + return; + + if (dram_info->bandwidth_kbps) + sprintf(bandwidth_str, "%d KBps", dram_info->bandwidth_kbps); + else + sprintf(bandwidth_str, "unknown"); + DRM_DEBUG_KMS("DRAM bandwidth:%s, total-channels: %u\n", + bandwidth_str, dram_info->num_channels); + DRM_DEBUG_KMS("DRAM rank: %s rank 16GB-dimm:%s\n", + (dram_info->rank == I915_DRAM_RANK_DUAL) ? + "dual" : "single", yesno(dram_info->is_16gb_dimm)); +} + /** * i915_driver_init_hw - setup state requiring device access * @dev_priv: device private @@ -1180,6 +1474,12 @@ static int i915_driver_init_hw(struct drm_i915_private *dev_priv) goto err_msi; intel_opregion_setup(dev_priv); + /* + * Fill the dram structure to get the system raw bandwidth and + * dram info. This will be used for memory latency calculation. + */ + intel_get_dram_info(dev_priv); + return 0; diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 7ea442033a57..8624b4bdc242 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -87,8 +87,8 @@ #define DRIVER_NAME "i915" #define DRIVER_DESC "Intel Graphics" -#define DRIVER_DATE "20180906" -#define DRIVER_TIMESTAMP 1536242083 +#define DRIVER_DATE "20180921" +#define DRIVER_TIMESTAMP 1537521997 /* Use I915_STATE_WARN(x) and I915_STATE_WARN_ON() (rather than WARN() and * WARN_ON()) for hw state sanity checks to check for unexpected conditions @@ -1946,6 +1946,20 @@ struct drm_i915_private { bool distrust_bios_wm; } wm; + struct dram_info { + bool valid; + bool valid_dimm; + bool is_16gb_dimm; + u8 num_channels; + enum dram_rank { + I915_DRAM_RANK_INVALID = 0, + I915_DRAM_RANK_SINGLE, + I915_DRAM_RANK_DUAL + } rank; + u32 bandwidth_kbps; + bool symmetric_memory; + } dram_info; + struct i915_runtime_pm runtime_pm; struct { @@ -2159,6 +2173,15 @@ struct drm_i915_private { */ }; +struct dram_channel_info { + struct info { + u8 size, width; + enum dram_rank rank; + } l_info, s_info; + enum dram_rank rank; + bool is_16gb_dimm; +}; + static inline struct drm_i915_private *to_i915(const struct drm_device *dev) { return container_of(dev, struct drm_i915_private, drm); @@ -2284,7 +2307,7 @@ static inline struct scatterlist *__sg_next(struct scatterlist *sg) #define for_each_sgt_dma(__dmap, __iter, __sgt) \ for ((__iter) = __sgt_iter((__sgt)->sgl, true); \ ((__dmap) = (__iter).dma + (__iter).curr); \ - (((__iter).curr += PAGE_SIZE) >= (__iter).max) ? \ + (((__iter).curr += I915_GTT_PAGE_SIZE) >= (__iter).max) ? \ (__iter) = __sgt_iter(__sg_next((__iter).sgp), true), 0 : 0) /** @@ -3074,6 +3097,12 @@ enum i915_map_type { I915_MAP_FORCE_WC = I915_MAP_WC | I915_MAP_OVERRIDE, }; +static inline enum i915_map_type +i915_coherent_map_type(struct drm_i915_private *i915) +{ + return HAS_LLC(i915) ? I915_MAP_WB : I915_MAP_WC; +} + /** * i915_gem_object_pin_map - return a contiguous mapping of the entire object * @obj: the object to map into kernel address space @@ -3311,7 +3340,7 @@ int i915_gem_stolen_insert_node_in_range(struct drm_i915_private *dev_priv, void i915_gem_stolen_remove_node(struct drm_i915_private *dev_priv, struct drm_mm_node *node); int i915_gem_init_stolen(struct drm_i915_private *dev_priv); -void i915_gem_cleanup_stolen(struct drm_device *dev); +void i915_gem_cleanup_stolen(struct drm_i915_private *dev_priv); struct drm_i915_gem_object * i915_gem_object_create_stolen(struct drm_i915_private *dev_priv, resource_size_t size); diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 89834ce19acd..aa3969d52773 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -1127,11 +1127,7 @@ i915_gem_shmem_pread(struct drm_i915_gem_object *obj, offset = offset_in_page(args->offset); for (idx = args->offset >> PAGE_SHIFT; remain; idx++) { struct page *page = i915_gem_object_get_page(obj, idx); - int length; - - length = remain; - if (offset + length > PAGE_SIZE) - length = PAGE_SIZE - offset; + unsigned int length = min_t(u64, remain, PAGE_SIZE - offset); ret = shmem_pread(page, offset, length, user_data, page_to_phys(page) & obj_do_bit17_swizzling, @@ -1575,11 +1571,7 @@ i915_gem_shmem_pwrite(struct drm_i915_gem_object *obj, offset = offset_in_page(args->offset); for (idx = args->offset >> PAGE_SHIFT; remain; idx++) { struct page *page = i915_gem_object_get_page(obj, idx); - int length; - - length = remain; - if (offset + length > PAGE_SIZE) - length = PAGE_SIZE - offset; + unsigned int length = min_t(u64, remain, PAGE_SIZE - offset); ret = shmem_pwrite(page, offset, length, user_data, page_to_phys(page) & obj_do_bit17_swizzling, @@ -2506,7 +2498,9 @@ static bool i915_sg_trim(struct sg_table *orig_st) new_sg = new_st.sgl; for_each_sg(orig_st->sgl, sg, orig_st->nents, i) { sg_set_page(new_sg, sg_page(sg), sg->length, 0); - /* called before being DMA mapped, no need to copy sg->dma_* */ + sg_dma_address(new_sg) = sg_dma_address(sg); + sg_dma_len(new_sg) = sg_dma_len(sg); + new_sg = sg_next(new_sg); } GEM_BUG_ON(new_sg); /* Should walk exactly nents and hit the end */ @@ -3438,6 +3432,9 @@ bool i915_gem_unset_wedged(struct drm_i915_private *i915) i915_retire_requests(i915); GEM_BUG_ON(i915->gt.active_requests); + if (!intel_gpu_reset(i915, ALL_ENGINES)) + intel_engines_sanitize(i915); + /* * Undo nop_submit_request. We prevent all new i915 requests from * being queued (by disallowing execbuf whilst wedged) so having @@ -5414,8 +5411,19 @@ static int __intel_engines_record_defaults(struct drm_i915_private *i915) assert_kernel_context_is_current(i915); + /* + * Immediately park the GPU so that we enable powersaving and + * treat it as idle. The next time we issue a request, we will + * unpark and start using the engine->pinned_default_state, otherwise + * it is in limbo and an early reset may fail. + */ + __i915_gem_park(i915); + for_each_engine(engine, i915, id) { struct i915_vma *state; + void *vaddr; + + GEM_BUG_ON(to_intel_context(ctx, engine)->pin_count); state = to_intel_context(ctx, engine)->state; if (!state) @@ -5438,6 +5446,16 @@ static int __intel_engines_record_defaults(struct drm_i915_private *i915) goto err_active; engine->default_state = i915_gem_object_get(state->obj); + + /* Check we can acquire the image of the context state */ + vaddr = i915_gem_object_pin_map(engine->default_state, + I915_MAP_FORCE_WB); + if (IS_ERR(vaddr)) { + err = PTR_ERR(vaddr); + goto err_active; + } + + i915_gem_object_unpin_map(engine->default_state); } if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)) { diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index 747b8170a15a..f772593b99ab 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -862,7 +862,7 @@ int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data, ret = -EINVAL; break; case I915_CONTEXT_PARAM_NO_ZEROMAP: - args->value = ctx->flags & CONTEXT_NO_ZEROMAP; + args->value = test_bit(UCONTEXT_NO_ZEROMAP, &ctx->user_flags); break; case I915_CONTEXT_PARAM_GTT_SIZE: if (ctx->ppgtt) @@ -896,27 +896,23 @@ int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data, struct drm_i915_file_private *file_priv = file->driver_priv; struct drm_i915_gem_context_param *args = data; struct i915_gem_context *ctx; - int ret; + int ret = 0; ctx = i915_gem_context_lookup(file_priv, args->ctx_id); if (!ctx) return -ENOENT; - ret = i915_mutex_lock_interruptible(dev); - if (ret) - goto out; - switch (args->param) { case I915_CONTEXT_PARAM_BAN_PERIOD: ret = -EINVAL; break; case I915_CONTEXT_PARAM_NO_ZEROMAP: - if (args->size) { + if (args->size) ret = -EINVAL; - } else { - ctx->flags &= ~CONTEXT_NO_ZEROMAP; - ctx->flags |= args->value ? CONTEXT_NO_ZEROMAP : 0; - } + else if (args->value) + set_bit(UCONTEXT_NO_ZEROMAP, &ctx->user_flags); + else + clear_bit(UCONTEXT_NO_ZEROMAP, &ctx->user_flags); break; case I915_CONTEXT_PARAM_NO_ERROR_CAPTURE: if (args->size) @@ -960,9 +956,7 @@ int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data, ret = -EINVAL; break; } - mutex_unlock(&dev->struct_mutex); -out: i915_gem_context_put(ctx); return ret; } diff --git a/drivers/gpu/drm/i915/i915_gem_context.h b/drivers/gpu/drm/i915/i915_gem_context.h index e09673ca731d..08165f6a0a84 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.h +++ b/drivers/gpu/drm/i915/i915_gem_context.h @@ -117,15 +117,20 @@ struct i915_gem_context { struct rcu_head rcu; /** + * @user_flags: small set of booleans controlled by the user + */ + unsigned long user_flags; +#define UCONTEXT_NO_ZEROMAP 0 +#define UCONTEXT_NO_ERROR_CAPTURE 1 +#define UCONTEXT_BANNABLE 2 + + /** * @flags: small set of booleans */ unsigned long flags; -#define CONTEXT_NO_ZEROMAP BIT(0) -#define CONTEXT_NO_ERROR_CAPTURE 1 -#define CONTEXT_CLOSED 2 -#define CONTEXT_BANNABLE 3 -#define CONTEXT_BANNED 4 -#define CONTEXT_FORCE_SINGLE_SUBMISSION 5 +#define CONTEXT_BANNED 0 +#define CONTEXT_CLOSED 1 +#define CONTEXT_FORCE_SINGLE_SUBMISSION 2 /** * @hw_id: - unique identifier for the context @@ -209,37 +214,37 @@ static inline bool i915_gem_context_is_closed(const struct i915_gem_context *ctx static inline void i915_gem_context_set_closed(struct i915_gem_context *ctx) { GEM_BUG_ON(i915_gem_context_is_closed(ctx)); - __set_bit(CONTEXT_CLOSED, &ctx->flags); + set_bit(CONTEXT_CLOSED, &ctx->flags); } static inline bool i915_gem_context_no_error_capture(const struct i915_gem_context *ctx) { - return test_bit(CONTEXT_NO_ERROR_CAPTURE, &ctx->flags); + return test_bit(UCONTEXT_NO_ERROR_CAPTURE, &ctx->user_flags); } static inline void i915_gem_context_set_no_error_capture(struct i915_gem_context *ctx) { - __set_bit(CONTEXT_NO_ERROR_CAPTURE, &ctx->flags); + set_bit(UCONTEXT_NO_ERROR_CAPTURE, &ctx->user_flags); } static inline void i915_gem_context_clear_no_error_capture(struct i915_gem_context *ctx) { - __clear_bit(CONTEXT_NO_ERROR_CAPTURE, &ctx->flags); + clear_bit(UCONTEXT_NO_ERROR_CAPTURE, &ctx->user_flags); } static inline bool i915_gem_context_is_bannable(const struct i915_gem_context *ctx) { - return test_bit(CONTEXT_BANNABLE, &ctx->flags); + return test_bit(UCONTEXT_BANNABLE, &ctx->user_flags); } static inline void i915_gem_context_set_bannable(struct i915_gem_context *ctx) { - __set_bit(CONTEXT_BANNABLE, &ctx->flags); + set_bit(UCONTEXT_BANNABLE, &ctx->user_flags); } static inline void i915_gem_context_clear_bannable(struct i915_gem_context *ctx) { - __clear_bit(CONTEXT_BANNABLE, &ctx->flags); + clear_bit(UCONTEXT_BANNABLE, &ctx->user_flags); } static inline bool i915_gem_context_is_banned(const struct i915_gem_context *ctx) @@ -249,7 +254,7 @@ static inline bool i915_gem_context_is_banned(const struct i915_gem_context *ctx static inline void i915_gem_context_set_banned(struct i915_gem_context *ctx) { - __set_bit(CONTEXT_BANNED, &ctx->flags); + set_bit(CONTEXT_BANNED, &ctx->flags); } static inline bool i915_gem_context_force_single_submission(const struct i915_gem_context *ctx) diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 6d99651c6c4b..f90a09b83370 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -693,9 +693,14 @@ static int eb_reserve(struct i915_execbuffer *eb) eb_unreserve_vma(vma, &eb->flags[i]); if (flags & EXEC_OBJECT_PINNED) + /* Pinned must have their slot */ list_add(&vma->exec_link, &eb->unbound); else if (flags & __EXEC_OBJECT_NEEDS_MAP) + /* Map require the lowest 256MiB (aperture) */ list_add_tail(&vma->exec_link, &eb->unbound); + else if (!(flags & EXEC_OBJECT_SUPPORTS_48B_ADDRESS)) + /* Prioritise 4GiB region for restricted bo */ + list_add(&vma->exec_link, &last); else list_add_tail(&vma->exec_link, &last); } @@ -743,7 +748,7 @@ static int eb_select_context(struct i915_execbuffer *eb) } eb->context_flags = 0; - if (ctx->flags & CONTEXT_NO_ZEROMAP) + if (test_bit(UCONTEXT_NO_ZEROMAP, &ctx->user_flags)) eb->context_flags |= __EXEC_OBJECT_NEEDS_BIAS; return 0; diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index eb0e446d6482..56c7f8637311 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -1050,7 +1050,7 @@ gen8_ppgtt_insert_pte_entries(struct i915_hw_ppgtt *ppgtt, do { vaddr[idx->pte] = pte_encode | iter->dma; - iter->dma += PAGE_SIZE; + iter->dma += I915_GTT_PAGE_SIZE; if (iter->dma >= iter->max) { iter->sg = __sg_next(iter->sg); if (!iter->sg) { @@ -1144,7 +1144,7 @@ static void gen8_ppgtt_insert_huge_entries(struct i915_vma *vma, vma->page_sizes.sg & I915_GTT_PAGE_SIZE_64K && IS_ALIGNED(iter->dma, I915_GTT_PAGE_SIZE_64K) && (IS_ALIGNED(rem, I915_GTT_PAGE_SIZE_64K) || - rem >= (max - index) << PAGE_SHIFT)) + rem >= (max - index) * I915_GTT_PAGE_SIZE)) maybe_64K = true; vaddr = kmap_atomic_px(pt); @@ -1169,7 +1169,7 @@ static void gen8_ppgtt_insert_huge_entries(struct i915_vma *vma, if (maybe_64K && index < max && !(IS_ALIGNED(iter->dma, I915_GTT_PAGE_SIZE_64K) && (IS_ALIGNED(rem, I915_GTT_PAGE_SIZE_64K) || - rem >= (max - index) << PAGE_SHIFT))) + rem >= (max - index) * I915_GTT_PAGE_SIZE))) maybe_64K = false; if (unlikely(!IS_ALIGNED(iter->dma, page_size))) @@ -1759,7 +1759,7 @@ static void gen6_dump_ppgtt(struct i915_hw_ppgtt *base, struct seq_file *m) seq_printf(m, "\t\t(%03d, %04d) %08lx: ", pde, pte, - (pde * GEN6_PTES + pte) * PAGE_SIZE); + (pde * GEN6_PTES + pte) * I915_GTT_PAGE_SIZE); for (i = 0; i < 4; i++) { if (vaddr[pte + i] != scratch_pte) seq_printf(m, " %08x", vaddr[pte + i]); @@ -1842,10 +1842,10 @@ static void gen6_ppgtt_clear_range(struct i915_address_space *vm, u64 start, u64 length) { struct gen6_hw_ppgtt *ppgtt = to_gen6_ppgtt(i915_vm_to_ppgtt(vm)); - unsigned int first_entry = start >> PAGE_SHIFT; + unsigned int first_entry = start / I915_GTT_PAGE_SIZE; unsigned int pde = first_entry / GEN6_PTES; unsigned int pte = first_entry % GEN6_PTES; - unsigned int num_entries = length >> PAGE_SHIFT; + unsigned int num_entries = length / I915_GTT_PAGE_SIZE; const gen6_pte_t scratch_pte = ppgtt->scratch_pte; while (num_entries) { @@ -1886,7 +1886,7 @@ static void gen6_ppgtt_insert_entries(struct i915_address_space *vm, u32 flags) { struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm); - unsigned first_entry = vma->node.start >> PAGE_SHIFT; + unsigned first_entry = vma->node.start / I915_GTT_PAGE_SIZE; unsigned act_pt = first_entry / GEN6_PTES; unsigned act_pte = first_entry % GEN6_PTES; const u32 pte_encode = vm->pte_encode(0, cache_level, flags); @@ -1899,7 +1899,7 @@ static void gen6_ppgtt_insert_entries(struct i915_address_space *vm, do { vaddr[act_pte] = pte_encode | GEN6_PTE_ADDR_ENCODE(iter.dma); - iter.dma += PAGE_SIZE; + iter.dma += I915_GTT_PAGE_SIZE; if (iter.dma == iter.max) { iter.sg = __sg_next(iter.sg); if (!iter.sg) @@ -2037,7 +2037,7 @@ static int pd_vma_bind(struct i915_vma *vma, { struct i915_ggtt *ggtt = i915_vm_to_ggtt(vma->vm); struct gen6_hw_ppgtt *ppgtt = vma->private; - u32 ggtt_offset = i915_ggtt_offset(vma) / PAGE_SIZE; + u32 ggtt_offset = i915_ggtt_offset(vma) / I915_GTT_PAGE_SIZE; struct i915_page_table *pt; unsigned int pde; @@ -2163,7 +2163,7 @@ static struct i915_hw_ppgtt *gen6_ppgtt_create(struct drm_i915_private *i915) ppgtt->base.vm.i915 = i915; ppgtt->base.vm.dma = &i915->drm.pdev->dev; - ppgtt->base.vm.total = I915_PDES * GEN6_PTES * PAGE_SIZE; + ppgtt->base.vm.total = I915_PDES * GEN6_PTES * I915_GTT_PAGE_SIZE; i915_address_space_init(&ppgtt->base.vm, i915); @@ -2456,7 +2456,7 @@ static void gen8_ggtt_insert_page(struct i915_address_space *vm, { struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); gen8_pte_t __iomem *pte = - (gen8_pte_t __iomem *)ggtt->gsm + (offset >> PAGE_SHIFT); + (gen8_pte_t __iomem *)ggtt->gsm + offset / I915_GTT_PAGE_SIZE; gen8_set_pte(pte, gen8_pte_encode(addr, level, 0)); @@ -2480,7 +2480,7 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm, */ gtt_entries = (gen8_pte_t __iomem *)ggtt->gsm; - gtt_entries += vma->node.start >> PAGE_SHIFT; + gtt_entries += vma->node.start / I915_GTT_PAGE_SIZE; for_each_sgt_dma(addr, sgt_iter, vma->pages) gen8_set_pte(gtt_entries++, pte_encode | addr); @@ -2499,7 +2499,7 @@ static void gen6_ggtt_insert_page(struct i915_address_space *vm, { struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); gen6_pte_t __iomem *pte = - (gen6_pte_t __iomem *)ggtt->gsm + (offset >> PAGE_SHIFT); + (gen6_pte_t __iomem *)ggtt->gsm + offset / I915_GTT_PAGE_SIZE; iowrite32(vm->pte_encode(addr, level, flags), pte); @@ -2519,7 +2519,7 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm, { struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); gen6_pte_t __iomem *entries = (gen6_pte_t __iomem *)ggtt->gsm; - unsigned int i = vma->node.start >> PAGE_SHIFT; + unsigned int i = vma->node.start / I915_GTT_PAGE_SIZE; struct sgt_iter iter; dma_addr_t addr; for_each_sgt_dma(addr, iter, vma->pages) @@ -2541,8 +2541,8 @@ static void gen8_ggtt_clear_range(struct i915_address_space *vm, u64 start, u64 length) { struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); - unsigned first_entry = start >> PAGE_SHIFT; - unsigned num_entries = length >> PAGE_SHIFT; + unsigned first_entry = start / I915_GTT_PAGE_SIZE; + unsigned num_entries = length / I915_GTT_PAGE_SIZE; const gen8_pte_t scratch_pte = gen8_pte_encode(vm->scratch_page.daddr, I915_CACHE_LLC, 0); gen8_pte_t __iomem *gtt_base = @@ -2657,8 +2657,8 @@ static void gen6_ggtt_clear_range(struct i915_address_space *vm, u64 start, u64 length) { struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); - unsigned first_entry = start >> PAGE_SHIFT; - unsigned num_entries = length >> PAGE_SHIFT; + unsigned first_entry = start / I915_GTT_PAGE_SIZE; + unsigned num_entries = length / I915_GTT_PAGE_SIZE; gen6_pte_t scratch_pte, __iomem *gtt_base = (gen6_pte_t __iomem *)ggtt->gsm + first_entry; const int max_entries = ggtt_total_entries(ggtt) - first_entry; @@ -3005,7 +3005,7 @@ void i915_ggtt_cleanup_hw(struct drm_i915_private *dev_priv) arch_phys_wc_del(ggtt->mtrr); io_mapping_fini(&ggtt->iomap); - i915_gem_cleanup_stolen(&dev_priv->drm); + i915_gem_cleanup_stolen(dev_priv); } static unsigned int gen6_get_total_gtt_size(u16 snb_gmch_ctl) @@ -3023,7 +3023,7 @@ static unsigned int gen8_get_total_gtt_size(u16 bdw_gmch_ctl) bdw_gmch_ctl = 1 << bdw_gmch_ctl; #ifdef CONFIG_X86_32 - /* Limit 32b platforms to a 2GB GGTT: 4 << 20 / pte size * PAGE_SIZE */ + /* Limit 32b platforms to a 2GB GGTT: 4 << 20 / pte size * I915_GTT_PAGE_SIZE */ if (bdw_gmch_ctl > 4) bdw_gmch_ctl = 4; #endif @@ -3398,7 +3398,7 @@ static int gen8_gmch_probe(struct i915_ggtt *ggtt) else size = gen8_get_total_gtt_size(snb_gmch_ctl); - ggtt->vm.total = (size / sizeof(gen8_pte_t)) << PAGE_SHIFT; + ggtt->vm.total = (size / sizeof(gen8_pte_t)) * I915_GTT_PAGE_SIZE; ggtt->vm.cleanup = gen6_gmch_remove; ggtt->vm.insert_page = gen8_ggtt_insert_page; ggtt->vm.clear_range = nop_clear_range; @@ -3456,7 +3456,7 @@ static int gen6_gmch_probe(struct i915_ggtt *ggtt) pci_read_config_word(pdev, SNB_GMCH_CTRL, &snb_gmch_ctl); size = gen6_get_total_gtt_size(snb_gmch_ctl); - ggtt->vm.total = (size / sizeof(gen6_pte_t)) << PAGE_SHIFT; + ggtt->vm.total = (size / sizeof(gen6_pte_t)) * I915_GTT_PAGE_SIZE; ggtt->vm.clear_range = gen6_ggtt_clear_range; ggtt->vm.insert_page = gen6_ggtt_insert_page; @@ -3727,9 +3727,9 @@ rotate_pages(const dma_addr_t *in, unsigned int offset, * the entries so the sg list can be happily traversed. * The only thing we need are DMA addresses. */ - sg_set_page(sg, NULL, PAGE_SIZE, 0); + sg_set_page(sg, NULL, I915_GTT_PAGE_SIZE, 0); sg_dma_address(sg) = in[offset + src_idx]; - sg_dma_len(sg) = PAGE_SIZE; + sg_dma_len(sg) = I915_GTT_PAGE_SIZE; sg = sg_next(sg); src_idx -= stride; } @@ -3742,7 +3742,7 @@ static noinline struct sg_table * intel_rotate_pages(struct intel_rotation_info *rot_info, struct drm_i915_gem_object *obj) { - const unsigned long n_pages = obj->base.size / PAGE_SIZE; + const unsigned long n_pages = obj->base.size / I915_GTT_PAGE_SIZE; unsigned int size = intel_rotation_info_size(rot_info); struct sgt_iter sgt_iter; dma_addr_t dma_addr; diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c index 53440bf87650..f29a7ff7c362 100644 --- a/drivers/gpu/drm/i915/i915_gem_stolen.c +++ b/drivers/gpu/drm/i915/i915_gem_stolen.c @@ -167,10 +167,8 @@ static int i915_adjust_stolen(struct drm_i915_private *dev_priv, return 0; } -void i915_gem_cleanup_stolen(struct drm_device *dev) +void i915_gem_cleanup_stolen(struct drm_i915_private *dev_priv) { - struct drm_i915_private *dev_priv = to_i915(dev); - if (!drm_mm_initialized(&dev_priv->mm.stolen)) return; diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index f7f2aa71d8d9..2835cacd0d08 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -1365,15 +1365,20 @@ static void request_record_user_bo(struct i915_request *request, { struct i915_capture_list *c; struct drm_i915_error_object **bo; - long count; + long count, max; - count = 0; + max = 0; for (c = request->capture_list; c; c = c->next) - count++; + max++; + if (!max) + return; - bo = NULL; - if (count) - bo = kcalloc(count, sizeof(*bo), GFP_ATOMIC); + bo = kmalloc_array(max, sizeof(*bo), GFP_ATOMIC); + if (!bo) { + /* If we can't capture everything, try to capture something. */ + max = min_t(long, max, PAGE_SIZE / sizeof(*bo)); + bo = kmalloc_array(max, sizeof(*bo), GFP_ATOMIC); + } if (!bo) return; @@ -1382,7 +1387,8 @@ static void request_record_user_bo(struct i915_request *request, bo[count] = i915_error_object_create(request->i915, c->vma); if (!bo[count]) break; - count++; + if (++count == max) + break; } ee->user_bo = bo; diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index ccb20230df2c..664b96bb65a3 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -1680,107 +1680,6 @@ static void gen8_update_reg_state_unlocked(struct i915_gem_context *ctx, } /* - * Same as gen8_update_reg_state_unlocked only through the batchbuffer. This - * is only used by the kernel context. - */ -static int gen8_emit_oa_config(struct i915_request *rq, - const struct i915_oa_config *oa_config) -{ - struct drm_i915_private *dev_priv = rq->i915; - /* The MMIO offsets for Flex EU registers aren't contiguous */ - u32 flex_mmio[] = { - i915_mmio_reg_offset(EU_PERF_CNTL0), - i915_mmio_reg_offset(EU_PERF_CNTL1), - i915_mmio_reg_offset(EU_PERF_CNTL2), - i915_mmio_reg_offset(EU_PERF_CNTL3), - i915_mmio_reg_offset(EU_PERF_CNTL4), - i915_mmio_reg_offset(EU_PERF_CNTL5), - i915_mmio_reg_offset(EU_PERF_CNTL6), - }; - u32 *cs; - int i; - - cs = intel_ring_begin(rq, ARRAY_SIZE(flex_mmio) * 2 + 4); - if (IS_ERR(cs)) - return PTR_ERR(cs); - - *cs++ = MI_LOAD_REGISTER_IMM(ARRAY_SIZE(flex_mmio) + 1); - - *cs++ = i915_mmio_reg_offset(GEN8_OACTXCONTROL); - *cs++ = (dev_priv->perf.oa.period_exponent << GEN8_OA_TIMER_PERIOD_SHIFT) | - (dev_priv->perf.oa.periodic ? GEN8_OA_TIMER_ENABLE : 0) | - GEN8_OA_COUNTER_RESUME; - - for (i = 0; i < ARRAY_SIZE(flex_mmio); i++) { - u32 mmio = flex_mmio[i]; - - /* - * This arbitrary default will select the 'EU FPU0 Pipeline - * Active' event. In the future it's anticipated that there - * will be an explicit 'No Event' we can select, but not - * yet... - */ - u32 value = 0; - - if (oa_config) { - u32 j; - - for (j = 0; j < oa_config->flex_regs_len; j++) { - if (i915_mmio_reg_offset(oa_config->flex_regs[j].addr) == mmio) { - value = oa_config->flex_regs[j].value; - break; - } - } - } - - *cs++ = mmio; - *cs++ = value; - } - - *cs++ = MI_NOOP; - intel_ring_advance(rq, cs); - - return 0; -} - -static int gen8_switch_to_updated_kernel_context(struct drm_i915_private *dev_priv, - const struct i915_oa_config *oa_config) -{ - struct intel_engine_cs *engine = dev_priv->engine[RCS]; - struct i915_timeline *timeline; - struct i915_request *rq; - int ret; - - lockdep_assert_held(&dev_priv->drm.struct_mutex); - - i915_retire_requests(dev_priv); - - rq = i915_request_alloc(engine, dev_priv->kernel_context); - if (IS_ERR(rq)) - return PTR_ERR(rq); - - ret = gen8_emit_oa_config(rq, oa_config); - if (ret) { - i915_request_add(rq); - return ret; - } - - /* Queue this switch after all other activity */ - list_for_each_entry(timeline, &dev_priv->gt.timelines, link) { - struct i915_request *prev; - - prev = i915_gem_active_raw(&timeline->last_request, - &dev_priv->drm.struct_mutex); - if (prev) - i915_request_await_dma_fence(rq, &prev->fence); - } - - i915_request_add(rq); - - return 0; -} - -/* * Manages updating the per-context aspects of the OA stream * configuration across all contexts. * @@ -1808,17 +1707,13 @@ static int gen8_configure_all_contexts(struct drm_i915_private *dev_priv, const struct i915_oa_config *oa_config) { struct intel_engine_cs *engine = dev_priv->engine[RCS]; + unsigned int map_type = i915_coherent_map_type(dev_priv); struct i915_gem_context *ctx; + struct i915_request *rq; int ret; - unsigned int wait_flags = I915_WAIT_LOCKED; lockdep_assert_held(&dev_priv->drm.struct_mutex); - /* Switch away from any user context. */ - ret = gen8_switch_to_updated_kernel_context(dev_priv, oa_config); - if (ret) - return ret; - /* * The OA register config is setup through the context image. This image * might be written to by the GPU on context switch (in particular on @@ -1833,7 +1728,7 @@ static int gen8_configure_all_contexts(struct drm_i915_private *dev_priv, * the GPU from any submitted work. */ ret = i915_gem_wait_for_idle(dev_priv, - wait_flags, + I915_WAIT_LOCKED, MAX_SCHEDULE_TIMEOUT); if (ret) return ret; @@ -1847,7 +1742,7 @@ static int gen8_configure_all_contexts(struct drm_i915_private *dev_priv, if (!ce->state) continue; - regs = i915_gem_object_pin_map(ce->state->obj, I915_MAP_WB); + regs = i915_gem_object_pin_map(ce->state->obj, map_type); if (IS_ERR(regs)) return PTR_ERR(regs); @@ -1859,7 +1754,17 @@ static int gen8_configure_all_contexts(struct drm_i915_private *dev_priv, i915_gem_object_unpin_map(ce->state->obj); } - return ret; + /* + * Apply the configuration by doing one context restore of the edited + * context image. + */ + rq = i915_request_alloc(engine, dev_priv->kernel_context); + if (IS_ERR(rq)) + return PTR_ERR(rq); + + i915_request_add(rq); + + return 0; } static int gen8_enable_metric_set(struct drm_i915_private *dev_priv, diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 09bc8e730ee1..7c491ea3d052 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -9339,6 +9339,9 @@ enum skl_power_gate { #define DPCLKA_CFGCR0_ICL _MMIO(0x164280) #define DPCLKA_CFGCR0_DDI_CLK_OFF(port) (1 << ((port) == PORT_F ? 23 : \ (port) + 10)) +#define ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(port) (1 << ((port) + 10)) +#define ICL_DPCLKA_CFGCR0_TC_CLK_OFF(tc_port) (1 << ((tc_port) == PORT_TC4 ? \ + 21 : (tc_port) + 12)) #define DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(port) ((port) == PORT_F ? 21 : \ (port) * 2) #define DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(port) (3 << DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(port)) @@ -9583,6 +9586,54 @@ enum skl_power_gate { #define DC_STATE_DEBUG_MASK_CORES (1 << 0) #define DC_STATE_DEBUG_MASK_MEMORY_UP (1 << 1) +#define BXT_P_CR_MC_BIOS_REQ_0_0_0 _MMIO(MCHBAR_MIRROR_BASE_SNB + 0x7114) +#define BXT_REQ_DATA_MASK 0x3F +#define BXT_DRAM_CHANNEL_ACTIVE_SHIFT 12 +#define BXT_DRAM_CHANNEL_ACTIVE_MASK (0xF << 12) +#define BXT_MEMORY_FREQ_MULTIPLIER_HZ 133333333 + +#define BXT_D_CR_DRP0_DUNIT8 0x1000 +#define BXT_D_CR_DRP0_DUNIT9 0x1200 +#define BXT_D_CR_DRP0_DUNIT_START 8 +#define BXT_D_CR_DRP0_DUNIT_END 11 +#define BXT_D_CR_DRP0_DUNIT(x) _MMIO(MCHBAR_MIRROR_BASE_SNB + \ + _PICK_EVEN((x) - 8, BXT_D_CR_DRP0_DUNIT8,\ + BXT_D_CR_DRP0_DUNIT9)) +#define BXT_DRAM_RANK_MASK 0x3 +#define BXT_DRAM_RANK_SINGLE 0x1 +#define BXT_DRAM_RANK_DUAL 0x3 +#define BXT_DRAM_WIDTH_MASK (0x3 << 4) +#define BXT_DRAM_WIDTH_SHIFT 4 +#define BXT_DRAM_WIDTH_X8 (0x0 << 4) +#define BXT_DRAM_WIDTH_X16 (0x1 << 4) +#define BXT_DRAM_WIDTH_X32 (0x2 << 4) +#define BXT_DRAM_WIDTH_X64 (0x3 << 4) +#define BXT_DRAM_SIZE_MASK (0x7 << 6) +#define BXT_DRAM_SIZE_SHIFT 6 +#define BXT_DRAM_SIZE_4GB (0x0 << 6) +#define BXT_DRAM_SIZE_6GB (0x1 << 6) +#define BXT_DRAM_SIZE_8GB (0x2 << 6) +#define BXT_DRAM_SIZE_12GB (0x3 << 6) +#define BXT_DRAM_SIZE_16GB (0x4 << 6) + +#define SKL_MEMORY_FREQ_MULTIPLIER_HZ 266666666 +#define SKL_MC_BIOS_DATA_0_0_0_MCHBAR_PCU _MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5E04) +#define SKL_REQ_DATA_MASK (0xF << 0) + +#define SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN _MMIO(MCHBAR_MIRROR_BASE_SNB + 0x500C) +#define SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN _MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5010) +#define SKL_DRAM_S_SHIFT 16 +#define SKL_DRAM_SIZE_MASK 0x3F +#define SKL_DRAM_WIDTH_MASK (0x3 << 8) +#define SKL_DRAM_WIDTH_SHIFT 8 +#define SKL_DRAM_WIDTH_X8 (0x0 << 8) +#define SKL_DRAM_WIDTH_X16 (0x1 << 8) +#define SKL_DRAM_WIDTH_X32 (0x2 << 8) +#define SKL_DRAM_RANK_MASK (0x1 << 10) +#define SKL_DRAM_RANK_SHIFT 10 +#define SKL_DRAM_RANK_SINGLE (0x0 << 10) +#define SKL_DRAM_RANK_DUAL (0x1 << 10) + /* Please see hsw_read_dcomp() and hsw_write_dcomp() before using this register, * since on HSW we can't write to it using I915_WRITE. */ #define D_COMP_HSW _MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5F0C) @@ -10231,6 +10282,12 @@ enum skl_power_gate { #define PREPARE_COUNT_SHIFT 0 #define PREPARE_COUNT_MASK (0x3f << 0) +#define _ICL_DSI_T_INIT_MASTER_0 0x6b088 +#define _ICL_DSI_T_INIT_MASTER_1 0x6b888 +#define ICL_DSI_T_INIT_MASTER(port) _MMIO_PORT(port, \ + _ICL_DSI_T_INIT_MASTER_0,\ + _ICL_DSI_T_INIT_MASTER_1) + /* bits 31:0 */ #define _MIPIA_DBI_BW_CTRL (dev_priv->mipi_mmio_base + 0xb084) #define _MIPIC_DBI_BW_CTRL (dev_priv->mipi_mmio_base + 0xb884) diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index 09ed48833b54..a492385b2089 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -732,13 +732,13 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) rq = kmem_cache_alloc(i915->requests, GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_NOWARN); if (unlikely(!rq)) { + i915_retire_requests(i915); + /* Ratelimit ourselves to prevent oom from malicious clients */ - ret = i915_gem_wait_for_idle(i915, - I915_WAIT_LOCKED | - I915_WAIT_INTERRUPTIBLE, - MAX_SCHEDULE_TIMEOUT); - if (ret) - goto err_unreserve; + rq = i915_gem_active_raw(&ce->ring->timeline->last_request, + &i915->drm.struct_mutex); + if (rq) + cond_synchronize_rcu(rq->rcustate); /* * We've forced the client to stall and catch up with whatever @@ -758,6 +758,8 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) } } + rq->rcustate = get_state_synchronize_rcu(); + INIT_LIST_HEAD(&rq->active_list); rq->i915 = i915; rq->engine = engine; diff --git a/drivers/gpu/drm/i915/i915_request.h b/drivers/gpu/drm/i915/i915_request.h index 9898301ab7ef..7fa94b024968 100644 --- a/drivers/gpu/drm/i915/i915_request.h +++ b/drivers/gpu/drm/i915/i915_request.h @@ -101,6 +101,14 @@ struct i915_request { struct intel_signal_node signaling; /* + * The rcu epoch of when this request was allocated. Used to judiciously + * apply backpressure on future allocations to ensure that under + * mempressure there is sufficient RCU ticks for us to reclaim our + * RCU protected slabs. + */ + unsigned long rcustate; + + /* * Fences for the various phases in the request's lifetime. * * The submit fence is used to await upon all of the request's diff --git a/drivers/gpu/drm/i915/i915_sw_fence.c b/drivers/gpu/drm/i915/i915_sw_fence.c index 1de5173e53a2..6dbeed079ae5 100644 --- a/drivers/gpu/drm/i915/i915_sw_fence.c +++ b/drivers/gpu/drm/i915/i915_sw_fence.c @@ -24,13 +24,13 @@ enum { DEBUG_FENCE_NOTIFY, }; -#ifdef CONFIG_DRM_I915_SW_FENCE_DEBUG_OBJECTS - static void *i915_sw_fence_debug_hint(void *addr) { return (void *)(((struct i915_sw_fence *)addr)->flags & I915_SW_FENCE_MASK); } +#ifdef CONFIG_DRM_I915_SW_FENCE_DEBUG_OBJECTS + static struct debug_obj_descr i915_sw_fence_debug_descr = { .name = "i915_sw_fence", .debug_hint = i915_sw_fence_debug_hint, @@ -393,10 +393,11 @@ static void timer_i915_sw_fence_wake(struct timer_list *t) if (!fence) return; - pr_warn("asynchronous wait on fence %s:%s:%x timed out\n", - cb->dma->ops->get_driver_name(cb->dma), - cb->dma->ops->get_timeline_name(cb->dma), - cb->dma->seqno); + pr_notice("Asynchronous wait on fence %s:%s:%x timed out (hint:%pS)\n", + cb->dma->ops->get_driver_name(cb->dma), + cb->dma->ops->get_timeline_name(cb->dma), + cb->dma->seqno, + i915_sw_fence_debug_hint(fence)); i915_sw_fence_complete(fence); } diff --git a/drivers/gpu/drm/i915/intel_atomic_plane.c b/drivers/gpu/drm/i915/intel_atomic_plane.c index fa7df5fe154b..aabebe0d2e9b 100644 --- a/drivers/gpu/drm/i915/intel_atomic_plane.c +++ b/drivers/gpu/drm/i915/intel_atomic_plane.c @@ -113,71 +113,18 @@ int intel_plane_atomic_check_with_state(const struct intel_crtc_state *old_crtc_ struct intel_plane_state *intel_state) { struct drm_plane *plane = intel_state->base.plane; - struct drm_i915_private *dev_priv = to_i915(plane->dev); struct drm_plane_state *state = &intel_state->base; struct intel_plane *intel_plane = to_intel_plane(plane); - const struct drm_display_mode *adjusted_mode = - &crtc_state->base.adjusted_mode; int ret; if (!intel_state->base.crtc && !old_plane_state->base.crtc) return 0; - if (state->fb && drm_rotation_90_or_270(state->rotation)) { - struct drm_format_name_buf format_name; - - if (state->fb->modifier != I915_FORMAT_MOD_Y_TILED && - state->fb->modifier != I915_FORMAT_MOD_Yf_TILED) { - DRM_DEBUG_KMS("Y/Yf tiling required for 90/270!\n"); - return -EINVAL; - } - - /* - * 90/270 is not allowed with RGB64 16:16:16:16, - * RGB 16-bit 5:6:5, and Indexed 8-bit. - * TBD: Add RGB64 case once its added in supported format list. - */ - switch (state->fb->format->format) { - case DRM_FORMAT_C8: - case DRM_FORMAT_RGB565: - DRM_DEBUG_KMS("Unsupported pixel format %s for 90/270!\n", - drm_get_format_name(state->fb->format->format, - &format_name)); - return -EINVAL; - - default: - break; - } - } - - /* CHV ignores the mirror bit when the rotate bit is set :( */ - if (IS_CHERRYVIEW(dev_priv) && - state->rotation & DRM_MODE_ROTATE_180 && - state->rotation & DRM_MODE_REFLECT_X) { - DRM_DEBUG_KMS("Cannot rotate and reflect at the same time\n"); - return -EINVAL; - } - intel_state->base.visible = false; ret = intel_plane->check_plane(crtc_state, intel_state); if (ret) return ret; - /* - * Y-tiling is not supported in IF-ID Interlace mode in - * GEN9 and above. - */ - if (state->fb && INTEL_GEN(dev_priv) >= 9 && crtc_state->base.enable && - adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) { - if (state->fb->modifier == I915_FORMAT_MOD_Y_TILED || - state->fb->modifier == I915_FORMAT_MOD_Yf_TILED || - state->fb->modifier == I915_FORMAT_MOD_Y_TILED_CCS || - state->fb->modifier == I915_FORMAT_MOD_Yf_TILED_CCS) { - DRM_DEBUG_KMS("Y/Yf tiling not supported in IF-ID mode\n"); - return -EINVAL; - } - } - /* FIXME pre-g4x don't work like this */ if (state->visible) crtc_state->active_planes |= BIT(intel_plane->id); diff --git a/drivers/gpu/drm/i915/intel_csr.c b/drivers/gpu/drm/i915/intel_csr.c index 14cf4c367e36..d48186e9ddad 100644 --- a/drivers/gpu/drm/i915/intel_csr.c +++ b/drivers/gpu/drm/i915/intel_csr.c @@ -34,6 +34,10 @@ * low-power state and comes back to normal. */ +#define I915_CSR_ICL "i915/icl_dmc_ver1_07.bin" +MODULE_FIRMWARE(I915_CSR_ICL); +#define ICL_CSR_VERSION_REQUIRED CSR_VERSION(1, 7) + #define I915_CSR_GLK "i915/glk_dmc_ver1_04.bin" MODULE_FIRMWARE(I915_CSR_GLK); #define GLK_CSR_VERSION_REQUIRED CSR_VERSION(1, 4) @@ -304,6 +308,8 @@ static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv, if (csr->fw_path == i915_modparams.dmc_firmware_path) { /* Bypass version check for firmware override. */ required_version = csr->version; + } else if (IS_ICELAKE(dev_priv)) { + required_version = ICL_CSR_VERSION_REQUIRED; } else if (IS_CANNONLAKE(dev_priv)) { required_version = CNL_CSR_VERSION_REQUIRED; } else if (IS_GEMINILAKE(dev_priv)) { @@ -471,6 +477,8 @@ void intel_csr_ucode_init(struct drm_i915_private *dev_priv) if (i915_modparams.dmc_firmware_path) csr->fw_path = i915_modparams.dmc_firmware_path; + else if (IS_ICELAKE(dev_priv)) + csr->fw_path = I915_CSR_ICL; else if (IS_CANNONLAKE(dev_priv)) csr->fw_path = I915_CSR_CNL; else if (IS_GEMINILAKE(dev_priv)) diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index cd01a09c5e0f..5186cd7075f9 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -916,7 +916,7 @@ static int intel_ddi_hdmi_level(struct drm_i915_private *dev_priv, enum port por level = dev_priv->vbt.ddi_port_info[port].hdmi_level_shift; if (IS_ICELAKE(dev_priv)) { - if (port == PORT_A || port == PORT_B) + if (intel_port_is_combophy(dev_priv, port)) icl_get_combo_buf_trans(dev_priv, port, INTEL_OUTPUT_HDMI, &n_entries); else @@ -1535,7 +1535,7 @@ static void icl_ddi_clock_get(struct intel_encoder *encoder, uint32_t pll_id; pll_id = intel_get_shared_dpll_id(dev_priv, pipe_config->shared_dpll); - if (port == PORT_A || port == PORT_B) { + if (intel_port_is_combophy(dev_priv, port)) { if (intel_crtc_has_type(pipe_config, INTEL_OUTPUT_HDMI)) link_clock = cnl_calc_wrpll_link(dev_priv, pll_id); else @@ -2077,7 +2077,7 @@ out: static inline enum intel_display_power_domain intel_ddi_main_link_aux_domain(struct intel_dp *intel_dp) { - /* CNL HW requires corresponding AUX IOs to be powered up for PSR with + /* CNL+ HW requires corresponding AUX IOs to be powered up for PSR with * DC states enabled at the same time, while for driver initiated AUX * transfers we need the same AUX IOs to be powered but with DC states * disabled. Accordingly use the AUX power domain here which leaves DC @@ -2235,7 +2235,7 @@ u8 intel_ddi_dp_voltage_max(struct intel_encoder *encoder) int n_entries; if (IS_ICELAKE(dev_priv)) { - if (port == PORT_A || port == PORT_B) + if (intel_port_is_combophy(dev_priv, port)) icl_get_combo_buf_trans(dev_priv, port, encoder->type, &n_entries); else @@ -2669,9 +2669,10 @@ static void icl_ddi_vswing_sequence(struct intel_encoder *encoder, u32 level, enum intel_output_type type) { + struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); enum port port = encoder->port; - if (port == PORT_A || port == PORT_B) + if (intel_port_is_combophy(dev_priv, port)) icl_combo_phy_ddi_vswing_sequence(encoder, level, type); else icl_mg_phy_ddi_vswing_sequence(encoder, link_clock, level); @@ -2732,6 +2733,21 @@ uint32_t ddi_signal_levels(struct intel_dp *intel_dp) return DDI_BUF_TRANS_SELECT(level); } +static inline +uint32_t icl_dpclka_cfgcr0_clk_off(struct drm_i915_private *dev_priv, + enum port port) +{ + if (intel_port_is_combophy(dev_priv, port)) { + return ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(port); + } else if (intel_port_is_tc(dev_priv, port)) { + enum tc_port tc_port = intel_port_to_tc(dev_priv, port); + + return ICL_DPCLKA_CFGCR0_TC_CLK_OFF(tc_port); + } + + return 0; +} + void icl_map_plls_to_ports(struct drm_crtc *crtc, struct intel_crtc_state *crtc_state, struct drm_atomic_state *old_state) @@ -2755,16 +2771,16 @@ void icl_map_plls_to_ports(struct drm_crtc *crtc, mutex_lock(&dev_priv->dpll_lock); val = I915_READ(DPCLKA_CFGCR0_ICL); - WARN_ON((val & DPCLKA_CFGCR0_DDI_CLK_OFF(port)) == 0); + WARN_ON((val & icl_dpclka_cfgcr0_clk_off(dev_priv, port)) == 0); - if (port == PORT_A || port == PORT_B) { + if (intel_port_is_combophy(dev_priv, port)) { val &= ~DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(port); val |= DPCLKA_CFGCR0_DDI_CLK_SEL(pll->info->id, port); I915_WRITE(DPCLKA_CFGCR0_ICL, val); POSTING_READ(DPCLKA_CFGCR0_ICL); } - val &= ~DPCLKA_CFGCR0_DDI_CLK_OFF(port); + val &= ~icl_dpclka_cfgcr0_clk_off(dev_priv, port); I915_WRITE(DPCLKA_CFGCR0_ICL, val); mutex_unlock(&dev_priv->dpll_lock); @@ -2792,7 +2808,7 @@ void icl_unmap_plls_to_ports(struct drm_crtc *crtc, mutex_lock(&dev_priv->dpll_lock); I915_WRITE(DPCLKA_CFGCR0_ICL, I915_READ(DPCLKA_CFGCR0_ICL) | - DPCLKA_CFGCR0_DDI_CLK_OFF(port)); + icl_dpclka_cfgcr0_clk_off(dev_priv, port)); mutex_unlock(&dev_priv->dpll_lock); } } @@ -2810,7 +2826,7 @@ static void intel_ddi_clk_select(struct intel_encoder *encoder, mutex_lock(&dev_priv->dpll_lock); if (IS_ICELAKE(dev_priv)) { - if (port >= PORT_C) + if (!intel_port_is_combophy(dev_priv, port)) I915_WRITE(DDI_CLK_SEL(port), icl_pll_to_ddi_pll_sel(encoder, pll)); } else if (IS_CANNONLAKE(dev_priv)) { @@ -2852,7 +2868,7 @@ static void intel_ddi_clk_disable(struct intel_encoder *encoder) enum port port = encoder->port; if (IS_ICELAKE(dev_priv)) { - if (port >= PORT_C) + if (!intel_port_is_combophy(dev_priv, port)) I915_WRITE(DDI_CLK_SEL(port), DDI_CLK_SEL_NONE); } else if (IS_CANNONLAKE(dev_priv)) { I915_WRITE(DPCLKA_CFGCR0, I915_READ(DPCLKA_CFGCR0) | diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 5711cb701760..9741cc419e1b 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -1917,10 +1917,10 @@ static unsigned int intel_tile_size(const struct drm_i915_private *dev_priv) } static unsigned int -intel_tile_width_bytes(const struct drm_framebuffer *fb, int plane) +intel_tile_width_bytes(const struct drm_framebuffer *fb, int color_plane) { struct drm_i915_private *dev_priv = to_i915(fb->dev); - unsigned int cpp = fb->format->cpp[plane]; + unsigned int cpp = fb->format->cpp[color_plane]; switch (fb->modifier) { case DRM_FORMAT_MOD_LINEAR: @@ -1931,7 +1931,7 @@ intel_tile_width_bytes(const struct drm_framebuffer *fb, int plane) else return 512; case I915_FORMAT_MOD_Y_TILED_CCS: - if (plane == 1) + if (color_plane == 1) return 128; /* fall through */ case I915_FORMAT_MOD_Y_TILED: @@ -1940,7 +1940,7 @@ intel_tile_width_bytes(const struct drm_framebuffer *fb, int plane) else return 512; case I915_FORMAT_MOD_Yf_TILED_CCS: - if (plane == 1) + if (color_plane == 1) return 128; /* fall through */ case I915_FORMAT_MOD_Yf_TILED: @@ -1965,22 +1965,22 @@ intel_tile_width_bytes(const struct drm_framebuffer *fb, int plane) } static unsigned int -intel_tile_height(const struct drm_framebuffer *fb, int plane) +intel_tile_height(const struct drm_framebuffer *fb, int color_plane) { if (fb->modifier == DRM_FORMAT_MOD_LINEAR) return 1; else return intel_tile_size(to_i915(fb->dev)) / - intel_tile_width_bytes(fb, plane); + intel_tile_width_bytes(fb, color_plane); } /* Return the tile dimensions in pixel units */ -static void intel_tile_dims(const struct drm_framebuffer *fb, int plane, +static void intel_tile_dims(const struct drm_framebuffer *fb, int color_plane, unsigned int *tile_width, unsigned int *tile_height) { - unsigned int tile_width_bytes = intel_tile_width_bytes(fb, plane); - unsigned int cpp = fb->format->cpp[plane]; + unsigned int tile_width_bytes = intel_tile_width_bytes(fb, color_plane); + unsigned int cpp = fb->format->cpp[color_plane]; *tile_width = tile_width_bytes / cpp; *tile_height = intel_tile_size(to_i915(fb->dev)) / tile_width_bytes; @@ -1988,9 +1988,9 @@ static void intel_tile_dims(const struct drm_framebuffer *fb, int plane, unsigned int intel_fb_align_height(const struct drm_framebuffer *fb, - int plane, unsigned int height) + int color_plane, unsigned int height) { - unsigned int tile_height = intel_tile_height(fb, plane); + unsigned int tile_height = intel_tile_height(fb, color_plane); return ALIGN(height, tile_height); } @@ -2044,12 +2044,12 @@ static unsigned int intel_linear_alignment(const struct drm_i915_private *dev_pr } static unsigned int intel_surf_alignment(const struct drm_framebuffer *fb, - int plane) + int color_plane) { struct drm_i915_private *dev_priv = to_i915(fb->dev); /* AUX_DIST needs only 4K alignment */ - if (plane == 1) + if (color_plane == 1) return 4096; switch (fb->modifier) { @@ -2080,14 +2080,13 @@ static bool intel_plane_uses_fence(const struct intel_plane_state *plane_state) struct i915_vma * intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, - unsigned int rotation, + const struct i915_ggtt_view *view, bool uses_fence, unsigned long *out_flags) { struct drm_device *dev = fb->dev; struct drm_i915_private *dev_priv = to_i915(dev); struct drm_i915_gem_object *obj = intel_fb_obj(fb); - struct i915_ggtt_view view; struct i915_vma *vma; unsigned int pinctl; u32 alignment; @@ -2096,8 +2095,6 @@ intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, alignment = intel_surf_alignment(fb, 0); - intel_fill_fb_ggtt_view(&view, fb, rotation); - /* Note that the w/a also requires 64 PTE of padding following the * bo. We currently fill all unused PTE with the shadow page and so * we should always have valid PTE following the scanout preventing @@ -2130,7 +2127,7 @@ intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, pinctl |= PIN_MAPPABLE; vma = i915_gem_object_pin_to_display_plane(obj, - alignment, &view, pinctl); + alignment, view, pinctl); if (IS_ERR(vma)) goto err; @@ -2182,13 +2179,13 @@ void intel_unpin_fb_vma(struct i915_vma *vma, unsigned long flags) i915_vma_put(vma); } -static int intel_fb_pitch(const struct drm_framebuffer *fb, int plane, +static int intel_fb_pitch(const struct drm_framebuffer *fb, int color_plane, unsigned int rotation) { if (drm_rotation_90_or_270(rotation)) - return to_intel_framebuffer(fb)->rotated[plane].pitch; + return to_intel_framebuffer(fb)->rotated[color_plane].pitch; else - return fb->pitches[plane]; + return fb->pitches[color_plane]; } /* @@ -2199,11 +2196,11 @@ static int intel_fb_pitch(const struct drm_framebuffer *fb, int plane, */ u32 intel_fb_xy_to_linear(int x, int y, const struct intel_plane_state *state, - int plane) + int color_plane) { const struct drm_framebuffer *fb = state->base.fb; - unsigned int cpp = fb->format->cpp[plane]; - unsigned int pitch = fb->pitches[plane]; + unsigned int cpp = fb->format->cpp[color_plane]; + unsigned int pitch = state->color_plane[color_plane].stride; return y * pitch + x * cpp; } @@ -2215,28 +2212,28 @@ u32 intel_fb_xy_to_linear(int x, int y, */ void intel_add_fb_offsets(int *x, int *y, const struct intel_plane_state *state, - int plane) + int color_plane) { const struct intel_framebuffer *intel_fb = to_intel_framebuffer(state->base.fb); unsigned int rotation = state->base.rotation; if (drm_rotation_90_or_270(rotation)) { - *x += intel_fb->rotated[plane].x; - *y += intel_fb->rotated[plane].y; + *x += intel_fb->rotated[color_plane].x; + *y += intel_fb->rotated[color_plane].y; } else { - *x += intel_fb->normal[plane].x; - *y += intel_fb->normal[plane].y; + *x += intel_fb->normal[color_plane].x; + *y += intel_fb->normal[color_plane].y; } } -static u32 __intel_adjust_tile_offset(int *x, int *y, - unsigned int tile_width, - unsigned int tile_height, - unsigned int tile_size, - unsigned int pitch_tiles, - u32 old_offset, - u32 new_offset) +static u32 intel_adjust_tile_offset(int *x, int *y, + unsigned int tile_width, + unsigned int tile_height, + unsigned int tile_size, + unsigned int pitch_tiles, + u32 old_offset, + u32 new_offset) { unsigned int pitch_pixels = pitch_tiles * tile_width; unsigned int tiles; @@ -2257,14 +2254,15 @@ static u32 __intel_adjust_tile_offset(int *x, int *y, return new_offset; } -static u32 _intel_adjust_tile_offset(int *x, int *y, - const struct drm_framebuffer *fb, int plane, - unsigned int rotation, - u32 old_offset, u32 new_offset) +static u32 intel_adjust_aligned_offset(int *x, int *y, + const struct drm_framebuffer *fb, + int color_plane, + unsigned int rotation, + unsigned int pitch, + u32 old_offset, u32 new_offset) { - const struct drm_i915_private *dev_priv = to_i915(fb->dev); - unsigned int cpp = fb->format->cpp[plane]; - unsigned int pitch = intel_fb_pitch(fb, plane, rotation); + struct drm_i915_private *dev_priv = to_i915(fb->dev); + unsigned int cpp = fb->format->cpp[color_plane]; WARN_ON(new_offset > old_offset); @@ -2273,7 +2271,7 @@ static u32 _intel_adjust_tile_offset(int *x, int *y, unsigned int pitch_tiles; tile_size = intel_tile_size(dev_priv); - intel_tile_dims(fb, plane, &tile_width, &tile_height); + intel_tile_dims(fb, color_plane, &tile_width, &tile_height); if (drm_rotation_90_or_270(rotation)) { pitch_tiles = pitch / tile_height; @@ -2282,9 +2280,9 @@ static u32 _intel_adjust_tile_offset(int *x, int *y, pitch_tiles = pitch / (tile_width * cpp); } - __intel_adjust_tile_offset(x, y, tile_width, tile_height, - tile_size, pitch_tiles, - old_offset, new_offset); + intel_adjust_tile_offset(x, y, tile_width, tile_height, + tile_size, pitch_tiles, + old_offset, new_offset); } else { old_offset += *y * pitch + *x * cpp; @@ -2299,17 +2297,19 @@ static u32 _intel_adjust_tile_offset(int *x, int *y, * Adjust the tile offset by moving the difference into * the x/y offsets. */ -static u32 intel_adjust_tile_offset(int *x, int *y, - const struct intel_plane_state *state, int plane, - u32 old_offset, u32 new_offset) -{ - return _intel_adjust_tile_offset(x, y, state->base.fb, plane, - state->base.rotation, - old_offset, new_offset); +static u32 intel_plane_adjust_aligned_offset(int *x, int *y, + const struct intel_plane_state *state, + int color_plane, + u32 old_offset, u32 new_offset) +{ + return intel_adjust_aligned_offset(x, y, state->base.fb, color_plane, + state->base.rotation, + state->color_plane[color_plane].stride, + old_offset, new_offset); } /* - * Computes the linear offset to the base tile and adjusts + * Computes the aligned offset to the base tile and adjusts * x, y. bytes per pixel is assumed to be a power-of-two. * * In the 90/270 rotated case, x and y are assumed @@ -2322,15 +2322,16 @@ static u32 intel_adjust_tile_offset(int *x, int *y, * used. This is why the user has to pass in the pitch since it * is specified in the rotated orientation. */ -static u32 _intel_compute_tile_offset(const struct drm_i915_private *dev_priv, - int *x, int *y, - const struct drm_framebuffer *fb, int plane, - unsigned int pitch, - unsigned int rotation, - u32 alignment) +static u32 intel_compute_aligned_offset(struct drm_i915_private *dev_priv, + int *x, int *y, + const struct drm_framebuffer *fb, + int color_plane, + unsigned int pitch, + unsigned int rotation, + u32 alignment) { uint64_t fb_modifier = fb->modifier; - unsigned int cpp = fb->format->cpp[plane]; + unsigned int cpp = fb->format->cpp[color_plane]; u32 offset, offset_aligned; if (alignment) @@ -2341,7 +2342,7 @@ static u32 _intel_compute_tile_offset(const struct drm_i915_private *dev_priv, unsigned int tile_rows, tiles, pitch_tiles; tile_size = intel_tile_size(dev_priv); - intel_tile_dims(fb, plane, &tile_width, &tile_height); + intel_tile_dims(fb, color_plane, &tile_width, &tile_height); if (drm_rotation_90_or_270(rotation)) { pitch_tiles = pitch / tile_height; @@ -2359,9 +2360,9 @@ static u32 _intel_compute_tile_offset(const struct drm_i915_private *dev_priv, offset = (tile_rows * pitch_tiles + tiles) * tile_size; offset_aligned = offset & ~alignment; - __intel_adjust_tile_offset(x, y, tile_width, tile_height, - tile_size, pitch_tiles, - offset, offset_aligned); + intel_adjust_tile_offset(x, y, tile_width, tile_height, + tile_size, pitch_tiles, + offset, offset_aligned); } else { offset = *y * pitch + *x * cpp; offset_aligned = offset & ~alignment; @@ -2373,42 +2374,44 @@ static u32 _intel_compute_tile_offset(const struct drm_i915_private *dev_priv, return offset_aligned; } -u32 intel_compute_tile_offset(int *x, int *y, - const struct intel_plane_state *state, - int plane) +static u32 intel_plane_compute_aligned_offset(int *x, int *y, + const struct intel_plane_state *state, + int color_plane) { struct intel_plane *intel_plane = to_intel_plane(state->base.plane); struct drm_i915_private *dev_priv = to_i915(intel_plane->base.dev); const struct drm_framebuffer *fb = state->base.fb; unsigned int rotation = state->base.rotation; - int pitch = intel_fb_pitch(fb, plane, rotation); + int pitch = state->color_plane[color_plane].stride; u32 alignment; if (intel_plane->id == PLANE_CURSOR) alignment = intel_cursor_alignment(dev_priv); else - alignment = intel_surf_alignment(fb, plane); + alignment = intel_surf_alignment(fb, color_plane); - return _intel_compute_tile_offset(dev_priv, x, y, fb, plane, pitch, - rotation, alignment); + return intel_compute_aligned_offset(dev_priv, x, y, fb, color_plane, + pitch, rotation, alignment); } /* Convert the fb->offset[] into x/y offsets */ static int intel_fb_offset_to_xy(int *x, int *y, - const struct drm_framebuffer *fb, int plane) + const struct drm_framebuffer *fb, + int color_plane) { struct drm_i915_private *dev_priv = to_i915(fb->dev); if (fb->modifier != DRM_FORMAT_MOD_LINEAR && - fb->offsets[plane] % intel_tile_size(dev_priv)) + fb->offsets[color_plane] % intel_tile_size(dev_priv)) return -EINVAL; *x = 0; *y = 0; - _intel_adjust_tile_offset(x, y, - fb, plane, DRM_MODE_ROTATE_0, - fb->offsets[plane], 0); + intel_adjust_aligned_offset(x, y, + fb, color_plane, DRM_MODE_ROTATE_0, + fb->pitches[color_plane], + fb->offsets[color_plane], 0); return 0; } @@ -2565,9 +2568,10 @@ intel_fill_fb_info(struct drm_i915_private *dev_priv, intel_fb->normal[i].x = x; intel_fb->normal[i].y = y; - offset = _intel_compute_tile_offset(dev_priv, &x, &y, - fb, i, fb->pitches[i], - DRM_MODE_ROTATE_0, tile_size); + offset = intel_compute_aligned_offset(dev_priv, &x, &y, fb, i, + fb->pitches[i], + DRM_MODE_ROTATE_0, + tile_size); offset /= tile_size; if (fb->modifier != DRM_FORMAT_MOD_LINEAR) { @@ -2614,10 +2618,10 @@ intel_fill_fb_info(struct drm_i915_private *dev_priv, * We only keep the x/y offsets, so push all of the * gtt offset into the x/y offsets. */ - __intel_adjust_tile_offset(&x, &y, - tile_width, tile_height, - tile_size, pitch_tiles, - gtt_offset_rotated * tile_size, 0); + intel_adjust_tile_offset(&x, &y, + tile_width, tile_height, + tile_size, pitch_tiles, + gtt_offset_rotated * tile_size, 0); gtt_offset_rotated += rot_info->plane[i].width * rot_info->plane[i].height; @@ -2636,9 +2640,9 @@ intel_fill_fb_info(struct drm_i915_private *dev_priv, max_size = max(max_size, offset + size); } - if (max_size * tile_size > obj->base.size) { - DRM_DEBUG_KMS("fb too big for bo (need %u bytes, have %zu bytes)\n", - max_size * tile_size, obj->base.size); + if (mul_u32_u32(max_size, tile_size) > obj->base.size) { + DRM_DEBUG_KMS("fb too big for bo (need %llu bytes, have %zu bytes)\n", + mul_u32_u32(max_size, tile_size), obj->base.size); return -EINVAL; } @@ -2718,6 +2722,17 @@ intel_alloc_initial_plane_obj(struct intel_crtc *crtc, if (size_aligned * 2 > dev_priv->stolen_usable_size) return false; + switch (fb->modifier) { + case DRM_FORMAT_MOD_LINEAR: + case I915_FORMAT_MOD_X_TILED: + case I915_FORMAT_MOD_Y_TILED: + break; + default: + DRM_DEBUG_DRIVER("Unsupported modifier for initial FB: 0x%llx\n", + fb->modifier); + return false; + } + mutex_lock(&dev->struct_mutex); obj = i915_gem_object_create_stolen_for_preallocated(dev_priv, base_aligned, @@ -2727,8 +2742,17 @@ intel_alloc_initial_plane_obj(struct intel_crtc *crtc, if (!obj) return false; - if (plane_config->tiling == I915_TILING_X) - obj->tiling_and_stride = fb->pitches[0] | I915_TILING_X; + switch (plane_config->tiling) { + case I915_TILING_NONE: + break; + case I915_TILING_X: + case I915_TILING_Y: + obj->tiling_and_stride = fb->pitches[0] | plane_config->tiling; + break; + default: + MISSING_CASE(plane_config->tiling); + return false; + } mode_cmd.pixel_format = fb->format->format; mode_cmd.width = fb->width; @@ -2760,20 +2784,33 @@ intel_set_plane_visible(struct intel_crtc_state *crtc_state, plane_state->base.visible = visible; - /* FIXME pre-g4x don't work like this */ - if (visible) { + if (visible) crtc_state->base.plane_mask |= drm_plane_mask(&plane->base); - crtc_state->active_planes |= BIT(plane->id); - } else { + else crtc_state->base.plane_mask &= ~drm_plane_mask(&plane->base); - crtc_state->active_planes &= ~BIT(plane->id); - } DRM_DEBUG_KMS("%s active planes 0x%x\n", crtc_state->base.crtc->name, crtc_state->active_planes); } +static void fixup_active_planes(struct intel_crtc_state *crtc_state) +{ + struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); + struct drm_plane *plane; + + /* + * Active_planes aliases if multiple "primary" or cursor planes + * have been used on the same (or wrong) pipe. plane_mask uses + * unique ids, hence we can use that to reconstruct active_planes. + */ + crtc_state->active_planes = 0; + + drm_for_each_plane_mask(plane, &dev_priv->drm, + crtc_state->base.plane_mask) + crtc_state->active_planes |= BIT(to_intel_plane(plane)->id); +} + static void intel_plane_disable_noatomic(struct intel_crtc *crtc, struct intel_plane *plane) { @@ -2783,6 +2820,7 @@ static void intel_plane_disable_noatomic(struct intel_crtc *crtc, to_intel_plane_state(plane->base.state); intel_set_plane_visible(crtc_state, plane_state, false); + fixup_active_planes(crtc_state); if (plane->id == PLANE_PRIMARY) intel_pre_disable_primary_noatomic(&crtc->base); @@ -2801,7 +2839,6 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, struct drm_i915_gem_object *obj; struct drm_plane *primary = intel_crtc->base.primary; struct drm_plane_state *plane_state = primary->state; - struct drm_crtc_state *crtc_state = intel_crtc->base.state; struct intel_plane *intel_plane = to_intel_plane(primary); struct intel_plane_state *intel_state = to_intel_plane_state(plane_state); @@ -2853,10 +2890,15 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, return; valid_fb: + intel_fill_fb_ggtt_view(&intel_state->view, fb, + intel_state->base.rotation); + intel_state->color_plane[0].stride = + intel_fb_pitch(fb, 0, intel_state->base.rotation); + mutex_lock(&dev->struct_mutex); intel_state->vma = intel_pin_and_fence_fb_obj(fb, - primary->state->rotation, + &intel_state->view, intel_plane_uses_fence(intel_state), &intel_state->flags); mutex_unlock(&dev->struct_mutex); @@ -2891,18 +2933,15 @@ valid_fb: plane_state->fb = fb; plane_state->crtc = &intel_crtc->base; - intel_set_plane_visible(to_intel_crtc_state(crtc_state), - to_intel_plane_state(plane_state), - true); - atomic_or(to_intel_plane(primary)->frontbuffer_bit, &obj->frontbuffer_bits); } -static int skl_max_plane_width(const struct drm_framebuffer *fb, int plane, +static int skl_max_plane_width(const struct drm_framebuffer *fb, + int color_plane, unsigned int rotation) { - int cpp = fb->format->cpp[plane]; + int cpp = fb->format->cpp[color_plane]; switch (fb->modifier) { case DRM_FORMAT_MOD_LINEAR: @@ -2950,9 +2989,9 @@ static bool skl_check_main_ccs_coordinates(struct intel_plane_state *plane_state const struct drm_framebuffer *fb = plane_state->base.fb; int hsub = fb->format->hsub; int vsub = fb->format->vsub; - int aux_x = plane_state->aux.x; - int aux_y = plane_state->aux.y; - u32 aux_offset = plane_state->aux.offset; + int aux_x = plane_state->color_plane[1].x; + int aux_y = plane_state->color_plane[1].y; + u32 aux_offset = plane_state->color_plane[1].offset; u32 alignment = intel_surf_alignment(fb, 1); while (aux_offset >= main_offset && aux_y <= main_y) { @@ -2966,8 +3005,8 @@ static bool skl_check_main_ccs_coordinates(struct intel_plane_state *plane_state x = aux_x / hsub; y = aux_y / vsub; - aux_offset = intel_adjust_tile_offset(&x, &y, plane_state, 1, - aux_offset, aux_offset - alignment); + aux_offset = intel_plane_adjust_aligned_offset(&x, &y, plane_state, 1, + aux_offset, aux_offset - alignment); aux_x = x * hsub + aux_x % hsub; aux_y = y * vsub + aux_y % vsub; } @@ -2975,30 +3014,24 @@ static bool skl_check_main_ccs_coordinates(struct intel_plane_state *plane_state if (aux_x != main_x || aux_y != main_y) return false; - plane_state->aux.offset = aux_offset; - plane_state->aux.x = aux_x; - plane_state->aux.y = aux_y; + plane_state->color_plane[1].offset = aux_offset; + plane_state->color_plane[1].x = aux_x; + plane_state->color_plane[1].y = aux_y; return true; } -static int skl_check_main_surface(const struct intel_crtc_state *crtc_state, - struct intel_plane_state *plane_state) +static int skl_check_main_surface(struct intel_plane_state *plane_state) { - struct drm_i915_private *dev_priv = - to_i915(plane_state->base.plane->dev); const struct drm_framebuffer *fb = plane_state->base.fb; unsigned int rotation = plane_state->base.rotation; int x = plane_state->base.src.x1 >> 16; int y = plane_state->base.src.y1 >> 16; int w = drm_rect_width(&plane_state->base.src) >> 16; int h = drm_rect_height(&plane_state->base.src) >> 16; - int dst_x = plane_state->base.dst.x1; - int dst_w = drm_rect_width(&plane_state->base.dst); - int pipe_src_w = crtc_state->pipe_src_w; int max_width = skl_max_plane_width(fb, 0, rotation); int max_height = 4096; - u32 alignment, offset, aux_offset = plane_state->aux.offset; + u32 alignment, offset, aux_offset = plane_state->color_plane[1].offset; if (w > max_width || h > max_height) { DRM_DEBUG_KMS("requested Y/RGB source size %dx%d too big (limit %dx%d)\n", @@ -3006,26 +3039,8 @@ static int skl_check_main_surface(const struct intel_crtc_state *crtc_state, return -EINVAL; } - /* - * Display WA #1175: cnl,glk - * Planes other than the cursor may cause FIFO underflow and display - * corruption if starting less than 4 pixels from the right edge of - * the screen. - * Besides the above WA fix the similar problem, where planes other - * than the cursor ending less than 4 pixels from the left edge of the - * screen may cause FIFO underflow and display corruption. - */ - if ((IS_GEMINILAKE(dev_priv) || IS_CANNONLAKE(dev_priv)) && - (dst_x + dst_w < 4 || dst_x > pipe_src_w - 4)) { - DRM_DEBUG_KMS("requested plane X %s position %d invalid (valid range %d-%d)\n", - dst_x + dst_w < 4 ? "end" : "start", - dst_x + dst_w < 4 ? dst_x + dst_w : dst_x, - 4, pipe_src_w - 4); - return -ERANGE; - } - intel_add_fb_offsets(&x, &y, plane_state, 0); - offset = intel_compute_tile_offset(&x, &y, plane_state, 0); + offset = intel_plane_compute_aligned_offset(&x, &y, plane_state, 0); alignment = intel_surf_alignment(fb, 0); /* @@ -3034,8 +3049,8 @@ static int skl_check_main_surface(const struct intel_crtc_state *crtc_state, * sure that is what we will get. */ if (offset > aux_offset) - offset = intel_adjust_tile_offset(&x, &y, plane_state, 0, - offset, aux_offset & ~(alignment - 1)); + offset = intel_plane_adjust_aligned_offset(&x, &y, plane_state, 0, + offset, aux_offset & ~(alignment - 1)); /* * When using an X-tiled surface, the plane blows up @@ -3046,14 +3061,14 @@ static int skl_check_main_surface(const struct intel_crtc_state *crtc_state, if (fb->modifier == I915_FORMAT_MOD_X_TILED) { int cpp = fb->format->cpp[0]; - while ((x + w) * cpp > fb->pitches[0]) { + while ((x + w) * cpp > plane_state->color_plane[0].stride) { if (offset == 0) { DRM_DEBUG_KMS("Unable to find suitable display surface offset due to X-tiling\n"); return -EINVAL; } - offset = intel_adjust_tile_offset(&x, &y, plane_state, 0, - offset, offset - alignment); + offset = intel_plane_adjust_aligned_offset(&x, &y, plane_state, 0, + offset, offset - alignment); } } @@ -3066,26 +3081,25 @@ static int skl_check_main_surface(const struct intel_crtc_state *crtc_state, if (offset == 0) break; - offset = intel_adjust_tile_offset(&x, &y, plane_state, 0, - offset, offset - alignment); + offset = intel_plane_adjust_aligned_offset(&x, &y, plane_state, 0, + offset, offset - alignment); } - if (x != plane_state->aux.x || y != plane_state->aux.y) { + if (x != plane_state->color_plane[1].x || y != plane_state->color_plane[1].y) { DRM_DEBUG_KMS("Unable to find suitable display surface offset due to CCS\n"); return -EINVAL; } } - plane_state->main.offset = offset; - plane_state->main.x = x; - plane_state->main.y = y; + plane_state->color_plane[0].offset = offset; + plane_state->color_plane[0].x = x; + plane_state->color_plane[0].y = y; return 0; } static int -skl_check_nv12_surface(const struct intel_crtc_state *crtc_state, - struct intel_plane_state *plane_state) +skl_check_nv12_surface(struct intel_plane_state *plane_state) { /* Display WA #1106 */ if (plane_state->base.rotation != @@ -3119,7 +3133,7 @@ static int skl_check_nv12_aux_surface(struct intel_plane_state *plane_state) u32 offset; intel_add_fb_offsets(&x, &y, plane_state, 1); - offset = intel_compute_tile_offset(&x, &y, plane_state, 1); + offset = intel_plane_compute_aligned_offset(&x, &y, plane_state, 1); /* FIXME not quite sure how/if these apply to the chroma plane */ if (w > max_width || h > max_height) { @@ -3128,9 +3142,9 @@ static int skl_check_nv12_aux_surface(struct intel_plane_state *plane_state) return -EINVAL; } - plane_state->aux.offset = offset; - plane_state->aux.x = x; - plane_state->aux.y = y; + plane_state->color_plane[1].offset = offset; + plane_state->color_plane[1].x = x; + plane_state->color_plane[1].y = y; return 0; } @@ -3146,34 +3160,29 @@ static int skl_check_ccs_aux_surface(struct intel_plane_state *plane_state) int y = src_y / vsub; u32 offset; - if (plane_state->base.rotation & ~(DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_180)) { - DRM_DEBUG_KMS("RC support only with 0/180 degree rotation %x\n", - plane_state->base.rotation); - return -EINVAL; - } - intel_add_fb_offsets(&x, &y, plane_state, 1); - offset = intel_compute_tile_offset(&x, &y, plane_state, 1); + offset = intel_plane_compute_aligned_offset(&x, &y, plane_state, 1); - plane_state->aux.offset = offset; - plane_state->aux.x = x * hsub + src_x % hsub; - plane_state->aux.y = y * vsub + src_y % vsub; + plane_state->color_plane[1].offset = offset; + plane_state->color_plane[1].x = x * hsub + src_x % hsub; + plane_state->color_plane[1].y = y * vsub + src_y % vsub; return 0; } -int skl_check_plane_surface(const struct intel_crtc_state *crtc_state, - struct intel_plane_state *plane_state) +int skl_check_plane_surface(struct intel_plane_state *plane_state) { const struct drm_framebuffer *fb = plane_state->base.fb; unsigned int rotation = plane_state->base.rotation; int ret; - if (rotation & DRM_MODE_REFLECT_X && - fb->modifier == DRM_FORMAT_MOD_LINEAR) { - DRM_DEBUG_KMS("horizontal flip is not supported with linear surface formats\n"); - return -EINVAL; - } + intel_fill_fb_ggtt_view(&plane_state->view, fb, rotation); + plane_state->color_plane[0].stride = intel_fb_pitch(fb, 0, rotation); + plane_state->color_plane[1].stride = intel_fb_pitch(fb, 1, rotation); + + ret = intel_plane_check_stride(plane_state); + if (ret) + return ret; if (!plane_state->base.visible) return 0; @@ -3189,7 +3198,7 @@ int skl_check_plane_surface(const struct intel_crtc_state *crtc_state, * the main surface setup depends on it. */ if (fb->format->format == DRM_FORMAT_NV12) { - ret = skl_check_nv12_surface(crtc_state, plane_state); + ret = skl_check_nv12_surface(plane_state); if (ret) return ret; ret = skl_check_nv12_aux_surface(plane_state); @@ -3200,18 +3209,45 @@ int skl_check_plane_surface(const struct intel_crtc_state *crtc_state, if (ret) return ret; } else { - plane_state->aux.offset = ~0xfff; - plane_state->aux.x = 0; - plane_state->aux.y = 0; + plane_state->color_plane[1].offset = ~0xfff; + plane_state->color_plane[1].x = 0; + plane_state->color_plane[1].y = 0; } - ret = skl_check_main_surface(crtc_state, plane_state); + ret = skl_check_main_surface(plane_state); if (ret) return ret; return 0; } +unsigned int +i9xx_plane_max_stride(struct intel_plane *plane, + u32 pixel_format, u64 modifier, + unsigned int rotation) +{ + struct drm_i915_private *dev_priv = to_i915(plane->base.dev); + + if (!HAS_GMCH_DISPLAY(dev_priv)) { + return 32*1024; + } else if (INTEL_GEN(dev_priv) >= 4) { + if (modifier == I915_FORMAT_MOD_X_TILED) + return 16*1024; + else + return 32*1024; + } else if (INTEL_GEN(dev_priv) >= 3) { + if (modifier == I915_FORMAT_MOD_X_TILED) + return 8*1024; + else + return 16*1024; + } else { + if (plane->i9xx_plane == PLANE_C) + return 4*1024; + else + return 8*1024; + } +} + static u32 i9xx_plane_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { @@ -3278,21 +3314,30 @@ int i9xx_check_plane_surface(struct intel_plane_state *plane_state) { struct drm_i915_private *dev_priv = to_i915(plane_state->base.plane->dev); + const struct drm_framebuffer *fb = plane_state->base.fb; + unsigned int rotation = plane_state->base.rotation; int src_x = plane_state->base.src.x1 >> 16; int src_y = plane_state->base.src.y1 >> 16; u32 offset; + int ret; + + intel_fill_fb_ggtt_view(&plane_state->view, fb, rotation); + plane_state->color_plane[0].stride = intel_fb_pitch(fb, 0, rotation); + + ret = intel_plane_check_stride(plane_state); + if (ret) + return ret; intel_add_fb_offsets(&src_x, &src_y, plane_state, 0); if (INTEL_GEN(dev_priv) >= 4) - offset = intel_compute_tile_offset(&src_x, &src_y, - plane_state, 0); + offset = intel_plane_compute_aligned_offset(&src_x, &src_y, + plane_state, 0); else offset = 0; /* HSW/BDW do this automagically in hardware */ if (!IS_HASWELL(dev_priv) && !IS_BROADWELL(dev_priv)) { - unsigned int rotation = plane_state->base.rotation; int src_w = drm_rect_width(&plane_state->base.src) >> 16; int src_h = drm_rect_height(&plane_state->base.src) >> 16; @@ -3304,9 +3349,43 @@ int i9xx_check_plane_surface(struct intel_plane_state *plane_state) } } - plane_state->main.offset = offset; - plane_state->main.x = src_x; - plane_state->main.y = src_y; + plane_state->color_plane[0].offset = offset; + plane_state->color_plane[0].x = src_x; + plane_state->color_plane[0].y = src_y; + + return 0; +} + +static int +i9xx_plane_check(struct intel_crtc_state *crtc_state, + struct intel_plane_state *plane_state) +{ + int ret; + + ret = chv_plane_check_rotation(plane_state); + if (ret) + return ret; + + ret = drm_atomic_helper_check_plane_state(&plane_state->base, + &crtc_state->base, + DRM_PLANE_HELPER_NO_SCALING, + DRM_PLANE_HELPER_NO_SCALING, + false, true); + if (ret) + return ret; + + if (!plane_state->base.visible) + return 0; + + ret = intel_plane_check_src_coordinates(plane_state); + if (ret) + return ret; + + ret = i9xx_check_plane_surface(plane_state); + if (ret) + return ret; + + plane_state->ctl = i9xx_plane_ctl(crtc_state, plane_state); return 0; } @@ -3316,20 +3395,19 @@ static void i9xx_update_plane(struct intel_plane *plane, const struct intel_plane_state *plane_state) { struct drm_i915_private *dev_priv = to_i915(plane->base.dev); - const struct drm_framebuffer *fb = plane_state->base.fb; enum i9xx_plane_id i9xx_plane = plane->i9xx_plane; u32 linear_offset; u32 dspcntr = plane_state->ctl; i915_reg_t reg = DSPCNTR(i9xx_plane); - int x = plane_state->main.x; - int y = plane_state->main.y; + int x = plane_state->color_plane[0].x; + int y = plane_state->color_plane[0].y; unsigned long irqflags; u32 dspaddr_offset; linear_offset = intel_fb_xy_to_linear(x, y, plane_state, 0); if (INTEL_GEN(dev_priv) >= 4) - dspaddr_offset = plane_state->main.offset; + dspaddr_offset = plane_state->color_plane[0].offset; else dspaddr_offset = linear_offset; @@ -3353,7 +3431,7 @@ static void i9xx_update_plane(struct intel_plane *plane, I915_WRITE_FW(reg, dspcntr); - I915_WRITE_FW(DSPSTRIDE(i9xx_plane), fb->pitches[0]); + I915_WRITE_FW(DSPSTRIDE(i9xx_plane), plane_state->color_plane[0].stride); if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) { I915_WRITE_FW(DSPSURF(i9xx_plane), intel_plane_ggtt_offset(plane_state) + @@ -3428,12 +3506,12 @@ static bool i9xx_plane_get_hw_state(struct intel_plane *plane, } static u32 -intel_fb_stride_alignment(const struct drm_framebuffer *fb, int plane) +intel_fb_stride_alignment(const struct drm_framebuffer *fb, int color_plane) { if (fb->modifier == DRM_FORMAT_MOD_LINEAR) return 64; else - return intel_tile_width_bytes(fb, plane); + return intel_tile_width_bytes(fb, color_plane); } static void skl_detach_scaler(struct intel_crtc *intel_crtc, int id) @@ -3463,24 +3541,24 @@ static void skl_detach_scalers(struct intel_crtc *intel_crtc) } } -u32 skl_plane_stride(const struct drm_framebuffer *fb, int plane, - unsigned int rotation) +u32 skl_plane_stride(const struct intel_plane_state *plane_state, + int color_plane) { - u32 stride; + const struct drm_framebuffer *fb = plane_state->base.fb; + unsigned int rotation = plane_state->base.rotation; + u32 stride = plane_state->color_plane[color_plane].stride; - if (plane >= fb->format->num_planes) + if (color_plane >= fb->format->num_planes) return 0; - stride = intel_fb_pitch(fb, plane, rotation); - /* * The stride is either expressed as a multiple of 64 bytes chunks for * linear buffers or in number of tiles for tiled buffers. */ if (drm_rotation_90_or_270(rotation)) - stride /= intel_tile_height(fb, plane); + stride /= intel_tile_height(fb, color_plane); else - stride /= intel_fb_stride_alignment(fb, plane); + stride /= intel_fb_stride_alignment(fb, color_plane); return stride; } @@ -5883,6 +5961,17 @@ static void i9xx_pfit_enable(struct intel_crtc *crtc) I915_WRITE(BCLRPAT(crtc->pipe), 0); } +bool intel_port_is_combophy(struct drm_i915_private *dev_priv, enum port port) +{ + if (port == PORT_NONE) + return false; + + if (IS_ICELAKE(dev_priv)) + return port <= PORT_B; + + return false; +} + bool intel_port_is_tc(struct drm_i915_private *dev_priv, enum port port) { if (IS_ICELAKE(dev_priv)) @@ -6014,6 +6103,8 @@ static void valleyview_crtc_enable(struct intel_crtc_state *pipe_config, i9xx_set_pipeconf(intel_crtc); + intel_color_set_csc(&pipe_config->base); + intel_crtc->active = true; intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true); @@ -6113,8 +6204,8 @@ static void i9xx_pfit_disable(struct intel_crtc *crtc) assert_pipe_disabled(dev_priv, crtc->pipe); - DRM_DEBUG_DRIVER("disabling pfit, current: 0x%08x\n", - I915_READ(PFIT_CONTROL)); + DRM_DEBUG_KMS("disabling pfit, current: 0x%08x\n", + I915_READ(PFIT_CONTROL)); I915_WRITE(PFIT_CONTROL, 0); } @@ -8634,8 +8725,8 @@ static int ironlake_crtc_compute_clock(struct intel_crtc *crtc, ironlake_compute_dpll(crtc, crtc_state, NULL); if (!intel_get_shared_dpll(crtc, crtc_state, NULL)) { - DRM_DEBUG_DRIVER("failed to find PLL for pipe %c\n", - pipe_name(crtc->pipe)); + DRM_DEBUG_KMS("failed to find PLL for pipe %c\n", + pipe_name(crtc->pipe)); return -EINVAL; } @@ -8805,6 +8896,7 @@ skylake_get_initial_plane_config(struct intel_crtc *crtc, fb->modifier = I915_FORMAT_MOD_X_TILED; break; case PLANE_CTL_TILED_Y: + plane_config->tiling = I915_TILING_Y; if (val & PLANE_CTL_RENDER_DECOMPRESSION_ENABLE) fb->modifier = I915_FORMAT_MOD_Y_TILED_CCS; else @@ -9202,8 +9294,8 @@ static int haswell_crtc_compute_clock(struct intel_crtc *crtc, intel_get_crtc_new_encoder(state, crtc_state); if (!intel_get_shared_dpll(crtc, crtc_state, encoder)) { - DRM_DEBUG_DRIVER("failed to find PLL for pipe %c\n", - pipe_name(crtc->pipe)); + DRM_DEBUG_KMS("failed to find PLL for pipe %c\n", + pipe_name(crtc->pipe)); return -EINVAL; } } @@ -9592,7 +9684,7 @@ static u32 intel_cursor_base(const struct intel_plane_state *plane_state) else base = intel_plane_ggtt_offset(plane_state); - base += plane_state->main.offset; + base += plane_state->color_plane[0].offset; /* ILK+ do this automagically */ if (HAS_GMCH_DISPLAY(dev_priv) && @@ -9635,55 +9727,86 @@ static bool intel_cursor_size_ok(const struct intel_plane_state *plane_state) height > 0 && height <= config->cursor_height; } -static int intel_check_cursor(struct intel_crtc_state *crtc_state, - struct intel_plane_state *plane_state) +static int intel_cursor_check_surface(struct intel_plane_state *plane_state) { const struct drm_framebuffer *fb = plane_state->base.fb; + unsigned int rotation = plane_state->base.rotation; int src_x, src_y; u32 offset; int ret; - ret = drm_atomic_helper_check_plane_state(&plane_state->base, - &crtc_state->base, - DRM_PLANE_HELPER_NO_SCALING, - DRM_PLANE_HELPER_NO_SCALING, - true, true); + intel_fill_fb_ggtt_view(&plane_state->view, fb, rotation); + plane_state->color_plane[0].stride = intel_fb_pitch(fb, 0, rotation); + + ret = intel_plane_check_stride(plane_state); if (ret) return ret; - if (!fb) - return 0; - - if (fb->modifier != DRM_FORMAT_MOD_LINEAR) { - DRM_DEBUG_KMS("cursor cannot be tiled\n"); - return -EINVAL; - } - src_x = plane_state->base.src_x >> 16; src_y = plane_state->base.src_y >> 16; intel_add_fb_offsets(&src_x, &src_y, plane_state, 0); - offset = intel_compute_tile_offset(&src_x, &src_y, plane_state, 0); + offset = intel_plane_compute_aligned_offset(&src_x, &src_y, + plane_state, 0); if (src_x != 0 || src_y != 0) { DRM_DEBUG_KMS("Arbitrary cursor panning not supported\n"); return -EINVAL; } - plane_state->main.offset = offset; + plane_state->color_plane[0].offset = offset; return 0; } -static u32 i845_cursor_ctl(const struct intel_crtc_state *crtc_state, - const struct intel_plane_state *plane_state) +static int intel_check_cursor(struct intel_crtc_state *crtc_state, + struct intel_plane_state *plane_state) { const struct drm_framebuffer *fb = plane_state->base.fb; + int ret; + + if (fb && fb->modifier != DRM_FORMAT_MOD_LINEAR) { + DRM_DEBUG_KMS("cursor cannot be tiled\n"); + return -EINVAL; + } + + ret = drm_atomic_helper_check_plane_state(&plane_state->base, + &crtc_state->base, + DRM_PLANE_HELPER_NO_SCALING, + DRM_PLANE_HELPER_NO_SCALING, + true, true); + if (ret) + return ret; + + if (!plane_state->base.visible) + return 0; + + ret = intel_plane_check_src_coordinates(plane_state); + if (ret) + return ret; + + ret = intel_cursor_check_surface(plane_state); + if (ret) + return ret; + + return 0; +} + +static unsigned int +i845_cursor_max_stride(struct intel_plane *plane, + u32 pixel_format, u64 modifier, + unsigned int rotation) +{ + return 2048; +} +static u32 i845_cursor_ctl(const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state) +{ return CURSOR_ENABLE | CURSOR_GAMMA_ENABLE | CURSOR_FORMAT_ARGB | - CURSOR_STRIDE(fb->pitches[0]); + CURSOR_STRIDE(plane_state->color_plane[0].stride); } static bool i845_cursor_size_ok(const struct intel_plane_state *plane_state) @@ -9719,6 +9842,9 @@ static int i845_check_cursor(struct intel_crtc_state *crtc_state, return -EINVAL; } + WARN_ON(plane_state->base.visible && + plane_state->color_plane[0].stride != fb->pitches[0]); + switch (fb->pitches[0]) { case 256: case 512: @@ -9807,6 +9933,14 @@ static bool i845_cursor_get_hw_state(struct intel_plane *plane, return ret; } +static unsigned int +i9xx_cursor_max_stride(struct intel_plane *plane, + u32 pixel_format, u64 modifier, + unsigned int rotation) +{ + return plane->base.dev->mode_config.cursor_width * 4; +} + static u32 i9xx_cursor_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { @@ -9912,6 +10046,9 @@ static int i9xx_check_cursor(struct intel_crtc_state *crtc_state, return -EINVAL; } + WARN_ON(plane_state->base.visible && + plane_state->color_plane[0].stride != fb->pitches[0]); + if (fb->pitches[0] != plane_state->base.crtc_w * fb->format->cpp[0]) { DRM_DEBUG_KMS("Invalid cursor stride (%u) (cursor width %d)\n", fb->pitches[0], plane_state->base.crtc_w); @@ -12982,7 +13119,7 @@ static int intel_plane_pin_fb(struct intel_plane_state *plane_state) } vma = intel_pin_and_fence_fb_obj(fb, - plane_state->base.rotation, + &plane_state->view, intel_plane_uses_fence(plane_state), &plane_state->flags); if (IS_ERR(vma)) @@ -13160,19 +13297,17 @@ intel_cleanup_plane_fb(struct drm_plane *plane, } int -skl_max_scale(struct intel_crtc *intel_crtc, - struct intel_crtc_state *crtc_state, - uint32_t pixel_format) +skl_max_scale(const struct intel_crtc_state *crtc_state, + u32 pixel_format) { - struct drm_i915_private *dev_priv; + struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); int max_scale, mult; int crtc_clock, max_dotclk, tmpclk1, tmpclk2; - if (!intel_crtc || !crtc_state->base.enable) + if (!crtc_state->base.enable) return DRM_PLANE_HELPER_NO_SCALING; - dev_priv = to_i915(intel_crtc->base.dev); - crtc_clock = crtc_state->base.adjusted_mode.crtc_clock; max_dotclk = to_intel_atomic_state(crtc_state->base.state)->cdclk.logical.cdclk; @@ -13196,61 +13331,6 @@ skl_max_scale(struct intel_crtc *intel_crtc, return max_scale; } -static int -intel_check_primary_plane(struct intel_crtc_state *crtc_state, - struct intel_plane_state *state) -{ - struct intel_plane *plane = to_intel_plane(state->base.plane); - struct drm_i915_private *dev_priv = to_i915(plane->base.dev); - struct drm_crtc *crtc = state->base.crtc; - int min_scale = DRM_PLANE_HELPER_NO_SCALING; - int max_scale = DRM_PLANE_HELPER_NO_SCALING; - bool can_position = false; - int ret; - uint32_t pixel_format = 0; - - if (INTEL_GEN(dev_priv) >= 9) { - /* use scaler when colorkey is not required */ - if (!state->ckey.flags) { - min_scale = 1; - if (state->base.fb) - pixel_format = state->base.fb->format->format; - max_scale = skl_max_scale(to_intel_crtc(crtc), - crtc_state, pixel_format); - } - can_position = true; - } - - ret = drm_atomic_helper_check_plane_state(&state->base, - &crtc_state->base, - min_scale, max_scale, - can_position, true); - if (ret) - return ret; - - if (!state->base.fb) - return 0; - - if (INTEL_GEN(dev_priv) >= 9) { - ret = skl_check_plane_surface(crtc_state, state); - if (ret) - return ret; - - state->ctl = skl_plane_ctl(crtc_state, state); - } else { - ret = i9xx_check_plane_surface(state); - if (ret) - return ret; - - state->ctl = i9xx_plane_ctl(crtc_state, state); - } - - if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv)) - state->color_ctl = glk_plane_color_ctl(crtc_state, state); - - return 0; -} - static void intel_begin_crtc_commit(struct drm_crtc *crtc, struct drm_crtc_state *old_crtc_state) { @@ -13672,12 +13752,8 @@ intel_primary_plane_create(struct drm_i915_private *dev_priv, enum pipe pipe) primary->base.state = &state->base; - primary->can_scale = false; - primary->max_downscale = 1; - if (INTEL_GEN(dev_priv) >= 9) { - primary->can_scale = true; + if (INTEL_GEN(dev_priv) >= 9) state->scaler_id = -1; - } primary->pipe = pipe; /* * On gen2/3 only plane A can do FBC, but the panel fitter and LVDS @@ -13704,8 +13780,6 @@ intel_primary_plane_create(struct drm_i915_private *dev_priv, enum pipe pipe) fbc->possible_framebuffer_bits |= primary->frontbuffer_bit; } - primary->check_plane = intel_check_primary_plane; - if (INTEL_GEN(dev_priv) >= 9) { primary->has_ccs = skl_plane_has_ccs(dev_priv, pipe, PLANE_PRIMARY); @@ -13723,9 +13797,11 @@ intel_primary_plane_create(struct drm_i915_private *dev_priv, enum pipe pipe) else modifiers = skl_format_modifiers_noccs; + primary->max_stride = skl_plane_max_stride; primary->update_plane = skl_update_plane; primary->disable_plane = skl_disable_plane; primary->get_hw_state = skl_plane_get_hw_state; + primary->check_plane = skl_plane_check; plane_funcs = &skl_plane_funcs; } else if (INTEL_GEN(dev_priv) >= 4) { @@ -13733,9 +13809,11 @@ intel_primary_plane_create(struct drm_i915_private *dev_priv, enum pipe pipe) num_formats = ARRAY_SIZE(i965_primary_formats); modifiers = i9xx_format_modifiers; + primary->max_stride = i9xx_plane_max_stride; primary->update_plane = i9xx_update_plane; primary->disable_plane = i9xx_disable_plane; primary->get_hw_state = i9xx_plane_get_hw_state; + primary->check_plane = i9xx_plane_check; plane_funcs = &i965_plane_funcs; } else { @@ -13743,9 +13821,11 @@ intel_primary_plane_create(struct drm_i915_private *dev_priv, enum pipe pipe) num_formats = ARRAY_SIZE(i8xx_primary_formats); modifiers = i9xx_format_modifiers; + primary->max_stride = i9xx_plane_max_stride; primary->update_plane = i9xx_update_plane; primary->disable_plane = i9xx_disable_plane; primary->get_hw_state = i9xx_plane_get_hw_state; + primary->check_plane = i9xx_plane_check; plane_funcs = &i8xx_plane_funcs; } @@ -13842,19 +13922,19 @@ intel_cursor_plane_create(struct drm_i915_private *dev_priv, cursor->base.state = &state->base; - cursor->can_scale = false; - cursor->max_downscale = 1; cursor->pipe = pipe; cursor->i9xx_plane = (enum i9xx_plane_id) pipe; cursor->id = PLANE_CURSOR; cursor->frontbuffer_bit = INTEL_FRONTBUFFER(pipe, cursor->id); if (IS_I845G(dev_priv) || IS_I865G(dev_priv)) { + cursor->max_stride = i845_cursor_max_stride; cursor->update_plane = i845_update_cursor; cursor->disable_plane = i845_disable_cursor; cursor->get_hw_state = i845_cursor_get_hw_state; cursor->check_plane = i845_check_cursor; } else { + cursor->max_stride = i9xx_cursor_max_stride; cursor->update_plane = i9xx_update_cursor; cursor->disable_plane = i9xx_disable_cursor; cursor->get_hw_state = i9xx_cursor_get_hw_state; @@ -14380,31 +14460,18 @@ static u32 intel_fb_pitch_limit(struct drm_i915_private *dev_priv, uint64_t fb_modifier, uint32_t pixel_format) { - u32 gen = INTEL_GEN(dev_priv); + struct intel_crtc *crtc; + struct intel_plane *plane; - if (gen >= 9) { - int cpp = drm_format_plane_cpp(pixel_format, 0); + /* + * We assume the primary plane for pipe A has + * the highest stride limits of them all. + */ + crtc = intel_get_crtc_for_pipe(dev_priv, PIPE_A); + plane = to_intel_plane(crtc->base.primary); - /* "The stride in bytes must not exceed the of the size of 8K - * pixels and 32K bytes." - */ - return min(8192 * cpp, 32768); - } else if (gen >= 5 && !HAS_GMCH_DISPLAY(dev_priv)) { - return 32*1024; - } else if (gen >= 4) { - if (fb_modifier == I915_FORMAT_MOD_X_TILED) - return 16*1024; - else - return 32*1024; - } else if (gen >= 3) { - if (fb_modifier == I915_FORMAT_MOD_X_TILED) - return 8*1024; - else - return 16*1024; - } else { - /* XXX DSPC is limited to 4k tiled */ - return 8*1024; - } + return plane->max_stride(plane, pixel_format, fb_modifier, + DRM_MODE_ROTATE_0); } static int intel_framebuffer_init(struct intel_framebuffer *intel_fb, @@ -15424,17 +15491,6 @@ void i830_disable_pipe(struct drm_i915_private *dev_priv, enum pipe pipe) POSTING_READ(DPLL(pipe)); } -static bool intel_plane_mapping_ok(struct intel_crtc *crtc, - struct intel_plane *plane) -{ - enum pipe pipe; - - if (!plane->get_hw_state(plane, &pipe)) - return true; - - return pipe == crtc->pipe; -} - static void intel_sanitize_plane_mapping(struct drm_i915_private *dev_priv) { @@ -15446,13 +15502,20 @@ intel_sanitize_plane_mapping(struct drm_i915_private *dev_priv) for_each_intel_crtc(&dev_priv->drm, crtc) { struct intel_plane *plane = to_intel_plane(crtc->base.primary); + struct intel_crtc *plane_crtc; + enum pipe pipe; - if (intel_plane_mapping_ok(crtc, plane)) + if (!plane->get_hw_state(plane, &pipe)) + continue; + + if (pipe == crtc->pipe) continue; DRM_DEBUG_KMS("%s attached to the wrong pipe, disabling plane\n", plane->base.name); - intel_plane_disable_noatomic(crtc, plane); + + plane_crtc = intel_get_crtc_for_pipe(dev_priv, pipe); + intel_plane_disable_noatomic(plane_crtc, plane); } } @@ -15500,13 +15563,9 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc, I915_READ(reg) & ~PIPECONF_FRAME_START_DELAY_MASK); } - /* restore vblank interrupts to correct state */ - drm_crtc_vblank_reset(&crtc->base); if (crtc->active) { struct intel_plane *plane; - drm_crtc_vblank_on(&crtc->base); - /* Disable everything but the primary plane */ for_each_intel_plane_on_crtc(dev, crtc, plane) { const struct intel_plane_state *plane_state = @@ -15624,23 +15683,32 @@ void i915_redisable_vga(struct drm_i915_private *dev_priv) } /* FIXME read out full plane state for all planes */ -static void readout_plane_state(struct intel_crtc *crtc) +static void readout_plane_state(struct drm_i915_private *dev_priv) { - struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); - struct intel_crtc_state *crtc_state = - to_intel_crtc_state(crtc->base.state); struct intel_plane *plane; + struct intel_crtc *crtc; - for_each_intel_plane_on_crtc(&dev_priv->drm, crtc, plane) { + for_each_intel_plane(&dev_priv->drm, plane) { struct intel_plane_state *plane_state = to_intel_plane_state(plane->base.state); - enum pipe pipe; + struct intel_crtc_state *crtc_state; + enum pipe pipe = PIPE_A; bool visible; visible = plane->get_hw_state(plane, &pipe); + crtc = intel_get_crtc_for_pipe(dev_priv, pipe); + crtc_state = to_intel_crtc_state(crtc->base.state); + intel_set_plane_visible(crtc_state, plane_state, visible); } + + for_each_intel_crtc(&dev_priv->drm, crtc) { + struct intel_crtc_state *crtc_state = + to_intel_crtc_state(crtc->base.state); + + fixup_active_planes(crtc_state); + } } static void intel_modeset_readout_hw_state(struct drm_device *dev) @@ -15672,13 +15740,13 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) if (crtc_state->base.active) dev_priv->active_crtcs |= 1 << crtc->pipe; - readout_plane_state(crtc); - DRM_DEBUG_KMS("[CRTC:%d:%s] hw state readout: %s\n", crtc->base.base.id, crtc->base.name, enableddisabled(crtc_state->base.active)); } + readout_plane_state(dev_priv); + for (i = 0; i < dev_priv->num_shared_dpll; i++) { struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i]; @@ -15848,7 +15916,6 @@ intel_modeset_setup_hw_state(struct drm_device *dev, struct drm_modeset_acquire_ctx *ctx) { struct drm_i915_private *dev_priv = to_i915(dev); - enum pipe pipe; struct intel_crtc *crtc; struct intel_encoder *encoder; int i; @@ -15861,15 +15928,23 @@ intel_modeset_setup_hw_state(struct drm_device *dev, /* HW state is read out, now we need to sanitize this mess. */ get_encoder_power_domains(dev_priv); - intel_sanitize_plane_mapping(dev_priv); + /* + * intel_sanitize_plane_mapping() may need to do vblank + * waits, so we need vblank interrupts restored beforehand. + */ + for_each_intel_crtc(&dev_priv->drm, crtc) { + drm_crtc_vblank_reset(&crtc->base); - for_each_intel_encoder(dev, encoder) { - intel_sanitize_encoder(encoder); + if (crtc->active) + drm_crtc_vblank_on(&crtc->base); } - for_each_pipe(dev_priv, pipe) { - crtc = intel_get_crtc_for_pipe(dev_priv, pipe); + intel_sanitize_plane_mapping(dev_priv); + + for_each_intel_encoder(dev, encoder) + intel_sanitize_encoder(encoder); + for_each_intel_crtc(&dev_priv->drm, crtc) { intel_sanitize_crtc(crtc, ctx); intel_dump_pipe_config(crtc, crtc->config, "[setup_hw_state]"); diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 6b4c19123f2a..3fae4dab295f 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -557,6 +557,22 @@ static bool intel_dp_link_params_valid(struct intel_dp *intel_dp, int link_rate, return true; } +static bool intel_dp_can_link_train_fallback_for_edp(struct intel_dp *intel_dp, + int link_rate, + uint8_t lane_count) +{ + const struct drm_display_mode *fixed_mode = + intel_dp->attached_connector->panel.fixed_mode; + int mode_rate, max_rate; + + mode_rate = intel_dp_link_required(fixed_mode->clock, 18); + max_rate = intel_dp_max_data_rate(link_rate, lane_count); + if (mode_rate > max_rate) + return false; + + return true; +} + int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp, int link_rate, uint8_t lane_count) { @@ -566,9 +582,23 @@ int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp, intel_dp->num_common_rates, link_rate); if (index > 0) { + if (intel_dp_is_edp(intel_dp) && + !intel_dp_can_link_train_fallback_for_edp(intel_dp, + intel_dp->common_rates[index - 1], + lane_count)) { + DRM_DEBUG_KMS("Retrying Link training for eDP with same parameters\n"); + return 0; + } intel_dp->max_link_rate = intel_dp->common_rates[index - 1]; intel_dp->max_link_lane_count = lane_count; } else if (lane_count > 1) { + if (intel_dp_is_edp(intel_dp) && + !intel_dp_can_link_train_fallback_for_edp(intel_dp, + intel_dp_max_common_rate(intel_dp), + lane_count >> 1)) { + DRM_DEBUG_KMS("Retrying Link training for eDP with same parameters\n"); + return 0; + } intel_dp->max_link_rate = intel_dp_max_common_rate(intel_dp); intel_dp->max_link_lane_count = lane_count >> 1; } else { @@ -3704,7 +3734,7 @@ intel_dp_set_signal_levels(struct intel_dp *intel_dp) uint32_t signal_levels, mask = 0; uint8_t train_set = intel_dp->train_set[0]; - if (IS_GEN9_LP(dev_priv) || IS_CANNONLAKE(dev_priv)) { + if (IS_GEN9_LP(dev_priv) || INTEL_GEN(dev_priv) >= 10) { signal_levels = bxt_signal_levels(intel_dp); } else if (HAS_DDI(dev_priv)) { signal_levels = ddi_signal_levels(intel_dp); diff --git a/drivers/gpu/drm/i915/intel_dp_link_training.c b/drivers/gpu/drm/i915/intel_dp_link_training.c index a9f40985a621..30be0e39bd5f 100644 --- a/drivers/gpu/drm/i915/intel_dp_link_training.c +++ b/drivers/gpu/drm/i915/intel_dp_link_training.c @@ -367,22 +367,14 @@ intel_dp_start_link_train(struct intel_dp *intel_dp) return; failure_handling: - /* Dont fallback and prune modes if its eDP */ - if (!intel_dp_is_edp(intel_dp)) { - DRM_DEBUG_KMS("[CONNECTOR:%d:%s] Link Training failed at link rate = %d, lane count = %d", - intel_connector->base.base.id, - intel_connector->base.name, - intel_dp->link_rate, intel_dp->lane_count); - if (!intel_dp_get_link_train_fallback_values(intel_dp, - intel_dp->link_rate, - intel_dp->lane_count)) - /* Schedule a Hotplug Uevent to userspace to start modeset */ - schedule_work(&intel_connector->modeset_retry_work); - } else { - DRM_ERROR("[CONNECTOR:%d:%s] Link Training failed at link rate = %d, lane count = %d", - intel_connector->base.base.id, - intel_connector->base.name, - intel_dp->link_rate, intel_dp->lane_count); - } + DRM_DEBUG_KMS("[CONNECTOR:%d:%s] Link Training failed at link rate = %d, lane count = %d", + intel_connector->base.base.id, + intel_connector->base.name, + intel_dp->link_rate, intel_dp->lane_count); + if (!intel_dp_get_link_train_fallback_values(intel_dp, + intel_dp->link_rate, + intel_dp->lane_count)) + /* Schedule a Hotplug Uevent to userspace to start modeset */ + schedule_work(&intel_connector->modeset_retry_work); return; } diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c index 43db2e9ac575..7f155b4f1a7d 100644 --- a/drivers/gpu/drm/i915/intel_dp_mst.c +++ b/drivers/gpu/drm/i915/intel_dp_mst.c @@ -38,11 +38,11 @@ static bool intel_dp_mst_compute_config(struct intel_encoder *encoder, struct intel_dp_mst_encoder *intel_mst = enc_to_mst(&encoder->base); struct intel_digital_port *intel_dig_port = intel_mst->primary; struct intel_dp *intel_dp = &intel_dig_port->dp; - struct intel_connector *connector = - to_intel_connector(conn_state->connector); + struct drm_connector *connector = conn_state->connector; + void *port = to_intel_connector(connector)->port; struct drm_atomic_state *state = pipe_config->base.state; int bpp; - int lane_count, slots; + int lane_count, slots = 0; const struct drm_display_mode *adjusted_mode = &pipe_config->base.adjusted_mode; int mst_pbn; bool constant_n = drm_dp_has_quirk(&intel_dp->desc, @@ -70,17 +70,23 @@ static bool intel_dp_mst_compute_config(struct intel_encoder *encoder, pipe_config->port_clock = intel_dp_max_link_rate(intel_dp); - if (drm_dp_mst_port_has_audio(&intel_dp->mst_mgr, connector->port)) + if (drm_dp_mst_port_has_audio(&intel_dp->mst_mgr, port)) pipe_config->has_audio = true; mst_pbn = drm_dp_calc_pbn_mode(adjusted_mode->crtc_clock, bpp); pipe_config->pbn = mst_pbn; - slots = drm_dp_atomic_find_vcpi_slots(state, &intel_dp->mst_mgr, - connector->port, mst_pbn); - if (slots < 0) { - DRM_DEBUG_KMS("failed finding vcpi slots:%d\n", slots); - return false; + /* Zombie connectors can't have VCPI slots */ + if (READ_ONCE(connector->registered)) { + slots = drm_dp_atomic_find_vcpi_slots(state, + &intel_dp->mst_mgr, + port, + mst_pbn); + if (slots < 0) { + DRM_DEBUG_KMS("failed finding vcpi slots:%d\n", + slots); + return false; + } } intel_link_compute_m_n(bpp, lane_count, @@ -307,9 +313,8 @@ static int intel_dp_mst_get_ddc_modes(struct drm_connector *connector) struct edid *edid; int ret; - if (!intel_dp) { + if (!READ_ONCE(connector->registered)) return intel_connector_update_modes(connector, NULL); - } edid = drm_dp_mst_get_edid(connector, &intel_dp->mst_mgr, intel_connector->port); ret = intel_connector_update_modes(connector, edid); @@ -324,9 +329,10 @@ intel_dp_mst_detect(struct drm_connector *connector, bool force) struct intel_connector *intel_connector = to_intel_connector(connector); struct intel_dp *intel_dp = intel_connector->mst_port; - if (!intel_dp) + if (!READ_ONCE(connector->registered)) return connector_status_disconnected; - return drm_dp_mst_detect_port(connector, &intel_dp->mst_mgr, intel_connector->port); + return drm_dp_mst_detect_port(connector, &intel_dp->mst_mgr, + intel_connector->port); } static void @@ -366,7 +372,7 @@ intel_dp_mst_mode_valid(struct drm_connector *connector, int bpp = 24; /* MST uses fixed bpp */ int max_rate, mode_rate, max_lanes, max_link_clock; - if (!intel_dp) + if (!READ_ONCE(connector->registered)) return MODE_ERROR; if (mode->flags & DRM_MODE_FLAG_DBLSCAN) @@ -398,8 +404,6 @@ static struct drm_encoder *intel_mst_atomic_best_encoder(struct drm_connector *c struct intel_dp *intel_dp = intel_connector->mst_port; struct intel_crtc *crtc = to_intel_crtc(state->crtc); - if (!intel_dp) - return NULL; return &intel_dp->mst_encoders[crtc->pipe]->base.base; } @@ -499,7 +503,6 @@ static void intel_dp_register_mst_connector(struct drm_connector *connector) static void intel_dp_destroy_mst_connector(struct drm_dp_mst_topology_mgr *mgr, struct drm_connector *connector) { - struct intel_connector *intel_connector = to_intel_connector(connector); struct drm_i915_private *dev_priv = to_i915(connector->dev); DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n", connector->base.id, connector->name); @@ -508,10 +511,6 @@ static void intel_dp_destroy_mst_connector(struct drm_dp_mst_topology_mgr *mgr, if (dev_priv->fbdev) drm_fb_helper_remove_one_connector(&dev_priv->fbdev->helper, connector); - /* prevent race with the check in ->detect */ - drm_modeset_lock(&connector->dev->mode_config.connection_mutex, NULL); - intel_connector->mst_port = NULL; - drm_modeset_unlock(&connector->dev->mode_config.connection_mutex); drm_connector_put(connector); } diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index f5731215210a..f8dc84b2d2d3 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -497,18 +497,21 @@ struct intel_atomic_state { struct intel_plane_state { struct drm_plane_state base; + struct i915_ggtt_view view; struct i915_vma *vma; unsigned long flags; #define PLANE_HAS_FENCE BIT(0) struct { u32 offset; + /* + * Plane stride in: + * bytes for 0/180 degree rotation + * pixels for 90/270 degree rotation + */ + u32 stride; int x, y; - } main; - struct { - u32 offset; - int x, y; - } aux; + } color_plane[2]; /* plane control register */ u32 ctl; @@ -950,10 +953,8 @@ struct intel_plane { enum i9xx_plane_id i9xx_plane; enum plane_id id; enum pipe pipe; - bool can_scale; bool has_fbc; bool has_ccs; - int max_downscale; uint32_t frontbuffer_bit; struct { @@ -966,6 +967,9 @@ struct intel_plane { * the intel_plane_state structure and accessed via plane_state. */ + unsigned int (*max_stride)(struct intel_plane *plane, + u32 pixel_format, u64 modifier, + unsigned int rotation); void (*update_plane)(struct intel_plane *plane, const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state); @@ -1442,7 +1446,7 @@ void icl_unmap_plls_to_ports(struct drm_crtc *crtc, struct drm_atomic_state *old_state); unsigned int intel_fb_align_height(const struct drm_framebuffer *fb, - int plane, unsigned int height); + int color_plane, unsigned int height); /* intel_audio.c */ void intel_init_audio_hooks(struct drm_i915_private *dev_priv); @@ -1513,6 +1517,7 @@ void intel_connector_attach_encoder(struct intel_connector *connector, struct intel_encoder *encoder); struct drm_display_mode * intel_encoder_current_mode(struct intel_encoder *encoder); +bool intel_port_is_combophy(struct drm_i915_private *dev_priv, enum port port); bool intel_port_is_tc(struct drm_i915_private *dev_priv, enum port port); enum tc_port intel_port_to_tc(struct drm_i915_private *dev_priv, enum port port); @@ -1565,7 +1570,7 @@ void intel_release_load_detect_pipe(struct drm_connector *connector, struct drm_modeset_acquire_ctx *ctx); struct i915_vma * intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, - unsigned int rotation, + const struct i915_ggtt_view *view, bool uses_fence, unsigned long *out_flags); void intel_unpin_fb_vma(struct i915_vma *vma, unsigned long flags); @@ -1614,8 +1619,6 @@ void assert_fdi_rx_pll(struct drm_i915_private *dev_priv, void assert_pipe(struct drm_i915_private *dev_priv, enum pipe pipe, bool state); #define assert_pipe_enabled(d, p) assert_pipe(d, p, true) #define assert_pipe_disabled(d, p) assert_pipe(d, p, false) -u32 intel_compute_tile_offset(int *x, int *y, - const struct intel_plane_state *state, int plane); void intel_prepare_reset(struct drm_i915_private *dev_priv); void intel_finish_reset(struct drm_i915_private *dev_priv); void hsw_enable_pc8(struct drm_i915_private *dev_priv); @@ -1645,8 +1648,8 @@ void intel_crtc_arm_fifo_underrun(struct intel_crtc *crtc, u16 skl_scaler_calc_phase(int sub, bool chroma_center); int skl_update_scaler_crtc(struct intel_crtc_state *crtc_state); -int skl_max_scale(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state, - uint32_t pixel_format); +int skl_max_scale(const struct intel_crtc_state *crtc_state, + u32 pixel_format); static inline u32 intel_plane_ggtt_offset(const struct intel_plane_state *state) { @@ -1658,12 +1661,14 @@ u32 glk_plane_color_ctl(const struct intel_crtc_state *crtc_state, u32 skl_plane_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state); u32 glk_color_ctl(const struct intel_plane_state *plane_state); -u32 skl_plane_stride(const struct drm_framebuffer *fb, int plane, - unsigned int rotation); -int skl_check_plane_surface(const struct intel_crtc_state *crtc_state, - struct intel_plane_state *plane_state); +u32 skl_plane_stride(const struct intel_plane_state *plane_state, + int plane); +int skl_check_plane_surface(struct intel_plane_state *plane_state); int i9xx_check_plane_surface(struct intel_plane_state *plane_state); int skl_format_to_fourcc(int format, bool rgb_order, bool alpha); +unsigned int i9xx_plane_max_stride(struct intel_plane *plane, + u32 pixel_format, u64 modifier, + unsigned int rotation); /* intel_csr.c */ void intel_csr_ucode_init(struct drm_i915_private *); @@ -2131,6 +2136,14 @@ bool skl_plane_has_ccs(struct drm_i915_private *dev_priv, enum pipe pipe, enum plane_id plane_id); bool skl_plane_has_planar(struct drm_i915_private *dev_priv, enum pipe pipe, enum plane_id plane_id); +unsigned int skl_plane_max_stride(struct intel_plane *plane, + u32 pixel_format, u64 modifier, + unsigned int rotation); +int skl_plane_check(struct intel_crtc_state *crtc_state, + struct intel_plane_state *plane_state); +int intel_plane_check_stride(const struct intel_plane_state *plane_state); +int intel_plane_check_src_coordinates(struct intel_plane_state *plane_state); +int chv_plane_check_rotation(const struct intel_plane_state *plane_state); /* intel_tv.c */ void intel_tv_init(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 10cd051ba29e..217ed3ee1cab 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -990,6 +990,9 @@ bool intel_engine_is_idle(struct intel_engine_cs *engine) } local_bh_enable(); + /* Otherwise flush the tasklet if it was on another cpu */ + tasklet_unlock_wait(t); + if (READ_ONCE(engine->execlists.active)) return false; } diff --git a/drivers/gpu/drm/i915/intel_fbc.c b/drivers/gpu/drm/i915/intel_fbc.c index 01d1d2088f04..74d425c700ef 100644 --- a/drivers/gpu/drm/i915/intel_fbc.c +++ b/drivers/gpu/drm/i915/intel_fbc.c @@ -670,8 +670,8 @@ static void intel_fbc_update_state_cache(struct intel_crtc *crtc, cache->plane.src_w = drm_rect_width(&plane_state->base.src) >> 16; cache->plane.src_h = drm_rect_height(&plane_state->base.src) >> 16; cache->plane.visible = plane_state->base.visible; - cache->plane.adjusted_x = plane_state->main.x; - cache->plane.adjusted_y = plane_state->main.y; + cache->plane.adjusted_x = plane_state->color_plane[0].x; + cache->plane.adjusted_y = plane_state->color_plane[0].y; cache->plane.y = plane_state->base.src.y1 >> 16; if (!cache->plane.visible) diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c index fb2f9fce34cd..f99332972b7a 100644 --- a/drivers/gpu/drm/i915/intel_fbdev.c +++ b/drivers/gpu/drm/i915/intel_fbdev.c @@ -175,6 +175,9 @@ static int intelfb_create(struct drm_fb_helper *helper, struct drm_i915_private *dev_priv = to_i915(dev); struct pci_dev *pdev = dev_priv->drm.pdev; struct i915_ggtt *ggtt = &dev_priv->ggtt; + const struct i915_ggtt_view view = { + .type = I915_GGTT_VIEW_NORMAL, + }; struct fb_info *info; struct drm_framebuffer *fb; struct i915_vma *vma; @@ -214,8 +217,7 @@ static int intelfb_create(struct drm_fb_helper *helper, * BIOS is suitable for own access. */ vma = intel_pin_and_fence_fb_obj(&ifbdev->fb->base, - DRM_MODE_ROTATE_0, - false, &flags); + &view, false, &flags); if (IS_ERR(vma)) { ret = PTR_ERR(vma); goto out_unlock; diff --git a/drivers/gpu/drm/i915/intel_guc_submission.c b/drivers/gpu/drm/i915/intel_guc_submission.c index 07b9d313b019..a81f04d46e87 100644 --- a/drivers/gpu/drm/i915/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/intel_guc_submission.c @@ -557,16 +557,36 @@ static void inject_preempt_context(struct work_struct *work) preempt_work[engine->id]); struct intel_guc_client *client = guc->preempt_client; struct guc_stage_desc *stage_desc = __get_stage_desc(client); - u32 ctx_desc = lower_32_bits(to_intel_context(client->owner, - engine)->lrc_desc); + struct intel_context *ce = to_intel_context(client->owner, engine); u32 data[7]; - /* - * The ring contains commands to write GUC_PREEMPT_FINISHED into HWSP. - * See guc_fill_preempt_context(). - */ + if (!ce->ring->emit) { /* recreate upon load/resume */ + u32 addr = intel_hws_preempt_done_address(engine); + u32 *cs; + + cs = ce->ring->vaddr; + if (engine->id == RCS) { + cs = gen8_emit_ggtt_write_rcs(cs, + GUC_PREEMPT_FINISHED, + addr); + } else { + cs = gen8_emit_ggtt_write(cs, + GUC_PREEMPT_FINISHED, + addr); + *cs++ = MI_NOOP; + *cs++ = MI_NOOP; + } + *cs++ = MI_USER_INTERRUPT; + *cs++ = MI_NOOP; + + ce->ring->emit = GUC_PREEMPT_BREADCRUMB_BYTES; + GEM_BUG_ON((void *)cs - ce->ring->vaddr != ce->ring->emit); + + flush_ggtt_writes(ce->ring->vma); + } + spin_lock_irq(&client->wq_lock); - guc_wq_item_append(client, engine->guc_id, ctx_desc, + guc_wq_item_append(client, engine->guc_id, lower_32_bits(ce->lrc_desc), GUC_PREEMPT_BREADCRUMB_BYTES / sizeof(u64), 0); spin_unlock_irq(&client->wq_lock); @@ -1044,50 +1064,6 @@ static inline bool ctx_save_restore_disabled(struct intel_context *ce) #undef SR_DISABLED } -static void guc_fill_preempt_context(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct intel_guc_client *client = guc->preempt_client; - struct intel_engine_cs *engine; - enum intel_engine_id id; - - for_each_engine(engine, dev_priv, id) { - struct intel_context *ce = - to_intel_context(client->owner, engine); - u32 addr = intel_hws_preempt_done_address(engine); - u32 *cs; - - GEM_BUG_ON(!ce->pin_count); - - /* - * We rely on this context image *not* being saved after - * preemption. This ensures that the RING_HEAD / RING_TAIL - * remain pointing at initial values forever. - */ - GEM_BUG_ON(!ctx_save_restore_disabled(ce)); - - cs = ce->ring->vaddr; - if (id == RCS) { - cs = gen8_emit_ggtt_write_rcs(cs, - GUC_PREEMPT_FINISHED, - addr); - } else { - cs = gen8_emit_ggtt_write(cs, - GUC_PREEMPT_FINISHED, - addr); - *cs++ = MI_NOOP; - *cs++ = MI_NOOP; - } - *cs++ = MI_USER_INTERRUPT; - *cs++ = MI_NOOP; - - GEM_BUG_ON((void *)cs - ce->ring->vaddr != - GUC_PREEMPT_BREADCRUMB_BYTES); - - flush_ggtt_writes(ce->ring->vma); - } -} - static int guc_clients_create(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); @@ -1118,8 +1094,6 @@ static int guc_clients_create(struct intel_guc *guc) return PTR_ERR(client); } guc->preempt_client = client; - - guc_fill_preempt_context(guc); } return 0; diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 9b1f0e5211a0..43957bb37a42 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -1294,7 +1294,7 @@ static int __context_pin(struct i915_gem_context *ctx, struct i915_vma *vma) * on an active context (which by nature is already on the GPU). */ if (!(vma->flags & I915_VMA_GLOBAL_BIND)) { - err = i915_gem_object_set_to_gtt_domain(vma->obj, true); + err = i915_gem_object_set_to_wc_domain(vma->obj, true); if (err) return err; } @@ -1322,7 +1322,9 @@ __execlists_context_pin(struct intel_engine_cs *engine, if (ret) goto err; - vaddr = i915_gem_object_pin_map(ce->state->obj, I915_MAP_WB); + vaddr = i915_gem_object_pin_map(ce->state->obj, + i915_coherent_map_type(ctx->i915) | + I915_MAP_OVERRIDE); if (IS_ERR(vaddr)) { ret = PTR_ERR(vaddr); goto unpin_vma; @@ -1338,11 +1340,13 @@ __execlists_context_pin(struct intel_engine_cs *engine, intel_lr_context_descriptor_update(ctx, engine, ce); + GEM_BUG_ON(!intel_ring_offset_valid(ce->ring, ce->ring->head)); + ce->lrc_reg_state = vaddr + LRC_STATE_PN * PAGE_SIZE; ce->lrc_reg_state[CTX_RING_BUFFER_START+1] = i915_ggtt_offset(ce->ring->vma); - GEM_BUG_ON(!intel_ring_offset_valid(ce->ring, ce->ring->head)); - ce->lrc_reg_state[CTX_RING_HEAD+1] = ce->ring->head; + ce->lrc_reg_state[CTX_RING_HEAD + 1] = ce->ring->head; + ce->lrc_reg_state[CTX_RING_TAIL + 1] = ce->ring->tail; ce->state->obj->pin_global++; i915_gem_context_get(ctx); @@ -2392,7 +2396,7 @@ static int logical_ring_init(struct intel_engine_cs *engine) ret = intel_engine_init_common(engine); if (ret) - goto error; + return ret; if (HAS_LOGICAL_RING_ELSQ(i915)) { execlists->submit_reg = i915->regs + @@ -2434,10 +2438,6 @@ static int logical_ring_init(struct intel_engine_cs *engine) reset_csb_pointers(execlists); return 0; - -error: - intel_logical_ring_cleanup(engine); - return ret; } int logical_render_ring_init(struct intel_engine_cs *engine) @@ -2460,10 +2460,14 @@ int logical_render_ring_init(struct intel_engine_cs *engine) engine->emit_breadcrumb = gen8_emit_breadcrumb_rcs; engine->emit_breadcrumb_sz = gen8_emit_breadcrumb_rcs_sz; - ret = intel_engine_create_scratch(engine, PAGE_SIZE); + ret = logical_ring_init(engine); if (ret) return ret; + ret = intel_engine_create_scratch(engine, PAGE_SIZE); + if (ret) + goto err_cleanup_common; + ret = intel_init_workaround_bb(engine); if (ret) { /* @@ -2475,7 +2479,11 @@ int logical_render_ring_init(struct intel_engine_cs *engine) ret); } - return logical_ring_init(engine); + return 0; + +err_cleanup_common: + intel_engine_cleanup_common(engine); + return ret; } int logical_xcs_ring_init(struct intel_engine_cs *engine) @@ -2841,13 +2849,14 @@ error_deref_obj: return ret; } -void intel_lr_context_resume(struct drm_i915_private *dev_priv) +void intel_lr_context_resume(struct drm_i915_private *i915) { struct intel_engine_cs *engine; struct i915_gem_context *ctx; enum intel_engine_id id; - /* Because we emit WA_TAIL_DWORDS there may be a disparity + /* + * Because we emit WA_TAIL_DWORDS there may be a disparity * between our bookkeeping in ce->ring->head and ce->ring->tail and * that stored in context. As we only write new commands from * ce->ring->tail onwards, everything before that is junk. If the GPU @@ -2857,28 +2866,22 @@ void intel_lr_context_resume(struct drm_i915_private *dev_priv) * So to avoid that we reset the context images upon resume. For * simplicity, we just zero everything out. */ - list_for_each_entry(ctx, &dev_priv->contexts.list, link) { - for_each_engine(engine, dev_priv, id) { + list_for_each_entry(ctx, &i915->contexts.list, link) { + for_each_engine(engine, i915, id) { struct intel_context *ce = to_intel_context(ctx, engine); - u32 *reg; if (!ce->state) continue; - reg = i915_gem_object_pin_map(ce->state->obj, - I915_MAP_WB); - if (WARN_ON(IS_ERR(reg))) - continue; - - reg += LRC_STATE_PN * PAGE_SIZE / sizeof(*reg); - reg[CTX_RING_HEAD+1] = 0; - reg[CTX_RING_TAIL+1] = 0; + intel_ring_reset(ce->ring, 0); - ce->state->obj->mm.dirty = true; - i915_gem_object_unpin_map(ce->state->obj); + if (ce->pin_count) { /* otherwise done in context_pin */ + u32 *regs = ce->lrc_reg_state; - intel_ring_reset(ce->ring, 0); + regs[CTX_RING_HEAD + 1] = ce->ring->head; + regs[CTX_RING_TAIL + 1] = ce->ring->tail; + } } } } diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c index 443dfaefd7a6..72eb7e48e8bc 100644 --- a/drivers/gpu/drm/i915/intel_overlay.c +++ b/drivers/gpu/drm/i915/intel_overlay.c @@ -487,23 +487,6 @@ void intel_overlay_reset(struct drm_i915_private *dev_priv) overlay->active = false; } -struct put_image_params { - int format; - short dst_x; - short dst_y; - short dst_w; - short dst_h; - short src_w; - short src_scan_h; - short src_scan_w; - short src_h; - short stride_Y; - short stride_UV; - int offset_Y; - int offset_U; - int offset_V; -}; - static int packed_depth_bytes(u32 format) { switch (format & I915_OVERLAY_DEPTH_MASK) { @@ -618,25 +601,25 @@ static void update_polyphase_filter(struct overlay_registers __iomem *regs) static bool update_scaling_factors(struct intel_overlay *overlay, struct overlay_registers __iomem *regs, - struct put_image_params *params) + struct drm_intel_overlay_put_image *params) { /* fixed point with a 12 bit shift */ u32 xscale, yscale, xscale_UV, yscale_UV; #define FP_SHIFT 12 #define FRACT_MASK 0xfff bool scale_changed = false; - int uv_hscale = uv_hsubsampling(params->format); - int uv_vscale = uv_vsubsampling(params->format); + int uv_hscale = uv_hsubsampling(params->flags); + int uv_vscale = uv_vsubsampling(params->flags); - if (params->dst_w > 1) - xscale = ((params->src_scan_w - 1) << FP_SHIFT) - /(params->dst_w); + if (params->dst_width > 1) + xscale = ((params->src_scan_width - 1) << FP_SHIFT) / + params->dst_width; else xscale = 1 << FP_SHIFT; - if (params->dst_h > 1) - yscale = ((params->src_scan_h - 1) << FP_SHIFT) - /(params->dst_h); + if (params->dst_height > 1) + yscale = ((params->src_scan_height - 1) << FP_SHIFT) / + params->dst_height; else yscale = 1 << FP_SHIFT; @@ -713,12 +696,12 @@ static void update_colorkey(struct intel_overlay *overlay, iowrite32(flags, ®s->DCLRKM); } -static u32 overlay_cmd_reg(struct put_image_params *params) +static u32 overlay_cmd_reg(struct drm_intel_overlay_put_image *params) { u32 cmd = OCMD_ENABLE | OCMD_BUF_TYPE_FRAME | OCMD_BUFFER0; - if (params->format & I915_OVERLAY_YUV_PLANAR) { - switch (params->format & I915_OVERLAY_DEPTH_MASK) { + if (params->flags & I915_OVERLAY_YUV_PLANAR) { + switch (params->flags & I915_OVERLAY_DEPTH_MASK) { case I915_OVERLAY_YUV422: cmd |= OCMD_YUV_422_PLANAR; break; @@ -731,7 +714,7 @@ static u32 overlay_cmd_reg(struct put_image_params *params) break; } } else { /* YUV packed */ - switch (params->format & I915_OVERLAY_DEPTH_MASK) { + switch (params->flags & I915_OVERLAY_DEPTH_MASK) { case I915_OVERLAY_YUV422: cmd |= OCMD_YUV_422_PACKED; break; @@ -740,7 +723,7 @@ static u32 overlay_cmd_reg(struct put_image_params *params) break; } - switch (params->format & I915_OVERLAY_SWAP_MASK) { + switch (params->flags & I915_OVERLAY_SWAP_MASK) { case I915_OVERLAY_NO_SWAP: break; case I915_OVERLAY_UV_SWAP: @@ -760,7 +743,7 @@ static u32 overlay_cmd_reg(struct put_image_params *params) static int intel_overlay_do_put_image(struct intel_overlay *overlay, struct drm_i915_gem_object *new_bo, - struct put_image_params *params) + struct drm_intel_overlay_put_image *params) { struct overlay_registers __iomem *regs = overlay->regs; struct drm_i915_private *dev_priv = overlay->i915; @@ -806,35 +789,40 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay, goto out_unpin; } - iowrite32((params->dst_y << 16) | params->dst_x, ®s->DWINPOS); - iowrite32((params->dst_h << 16) | params->dst_w, ®s->DWINSZ); + iowrite32(params->dst_y << 16 | params->dst_x, ®s->DWINPOS); + iowrite32(params->dst_height << 16 | params->dst_width, ®s->DWINSZ); - if (params->format & I915_OVERLAY_YUV_PACKED) - tmp_width = packed_width_bytes(params->format, params->src_w); + if (params->flags & I915_OVERLAY_YUV_PACKED) + tmp_width = packed_width_bytes(params->flags, + params->src_width); else - tmp_width = params->src_w; + tmp_width = params->src_width; - swidth = params->src_w; + swidth = params->src_width; swidthsw = calc_swidthsw(dev_priv, params->offset_Y, tmp_width); - sheight = params->src_h; + sheight = params->src_height; iowrite32(i915_ggtt_offset(vma) + params->offset_Y, ®s->OBUF_0Y); ostride = params->stride_Y; - if (params->format & I915_OVERLAY_YUV_PLANAR) { - int uv_hscale = uv_hsubsampling(params->format); - int uv_vscale = uv_vsubsampling(params->format); + if (params->flags & I915_OVERLAY_YUV_PLANAR) { + int uv_hscale = uv_hsubsampling(params->flags); + int uv_vscale = uv_vsubsampling(params->flags); u32 tmp_U, tmp_V; - swidth |= (params->src_w/uv_hscale) << 16; + + swidth |= (params->src_width / uv_hscale) << 16; + sheight |= (params->src_height / uv_vscale) << 16; + tmp_U = calc_swidthsw(dev_priv, params->offset_U, - params->src_w/uv_hscale); + params->src_width / uv_hscale); tmp_V = calc_swidthsw(dev_priv, params->offset_V, - params->src_w/uv_hscale); - swidthsw |= max_t(u32, tmp_U, tmp_V) << 16; - sheight |= (params->src_h/uv_vscale) << 16; + params->src_width / uv_hscale); + swidthsw |= max(tmp_U, tmp_V) << 16; + iowrite32(i915_ggtt_offset(vma) + params->offset_U, ®s->OBUF_0U); iowrite32(i915_ggtt_offset(vma) + params->offset_V, ®s->OBUF_0V); + ostride |= params->stride_UV << 16; } @@ -938,15 +926,16 @@ static int check_overlay_dst(struct intel_overlay *overlay, return -EINVAL; } -static int check_overlay_scaling(struct put_image_params *rec) +static int check_overlay_scaling(struct drm_intel_overlay_put_image *rec) { u32 tmp; /* downscaling limit is 8.0 */ - tmp = ((rec->src_scan_h << 16) / rec->dst_h) >> 16; + tmp = ((rec->src_scan_height << 16) / rec->dst_height) >> 16; if (tmp > 7) return -EINVAL; - tmp = ((rec->src_scan_w << 16) / rec->dst_w) >> 16; + + tmp = ((rec->src_scan_width << 16) / rec->dst_width) >> 16; if (tmp > 7) return -EINVAL; @@ -1067,13 +1056,12 @@ static int check_overlay_src(struct drm_i915_private *dev_priv, int intel_overlay_put_image_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { - struct drm_intel_overlay_put_image *put_image_rec = data; + struct drm_intel_overlay_put_image *params = data; struct drm_i915_private *dev_priv = to_i915(dev); struct intel_overlay *overlay; struct drm_crtc *drmmode_crtc; struct intel_crtc *crtc; struct drm_i915_gem_object *new_bo; - struct put_image_params *params; int ret; overlay = dev_priv->overlay; @@ -1082,7 +1070,7 @@ int intel_overlay_put_image_ioctl(struct drm_device *dev, void *data, return -ENODEV; } - if (!(put_image_rec->flags & I915_OVERLAY_ENABLE)) { + if (!(params->flags & I915_OVERLAY_ENABLE)) { drm_modeset_lock_all(dev); mutex_lock(&dev->struct_mutex); @@ -1094,22 +1082,14 @@ int intel_overlay_put_image_ioctl(struct drm_device *dev, void *data, return ret; } - params = kmalloc(sizeof(*params), GFP_KERNEL); - if (!params) - return -ENOMEM; - - drmmode_crtc = drm_crtc_find(dev, file_priv, put_image_rec->crtc_id); - if (!drmmode_crtc) { - ret = -ENOENT; - goto out_free; - } + drmmode_crtc = drm_crtc_find(dev, file_priv, params->crtc_id); + if (!drmmode_crtc) + return -ENOENT; crtc = to_intel_crtc(drmmode_crtc); - new_bo = i915_gem_object_lookup(file_priv, put_image_rec->bo_handle); - if (!new_bo) { - ret = -ENOENT; - goto out_free; - } + new_bo = i915_gem_object_lookup(file_priv, params->bo_handle); + if (!new_bo) + return -ENOENT; drm_modeset_lock_all(dev); mutex_lock(&dev->struct_mutex); @@ -1145,42 +1125,27 @@ int intel_overlay_put_image_ioctl(struct drm_device *dev, void *data, overlay->pfit_active = false; } - ret = check_overlay_dst(overlay, put_image_rec); + ret = check_overlay_dst(overlay, params); if (ret != 0) goto out_unlock; if (overlay->pfit_active) { - params->dst_y = ((((u32)put_image_rec->dst_y) << 12) / + params->dst_y = (((u32)params->dst_y << 12) / overlay->pfit_vscale_ratio); /* shifting right rounds downwards, so add 1 */ - params->dst_h = ((((u32)put_image_rec->dst_height) << 12) / + params->dst_height = (((u32)params->dst_height << 12) / overlay->pfit_vscale_ratio) + 1; - } else { - params->dst_y = put_image_rec->dst_y; - params->dst_h = put_image_rec->dst_height; } - params->dst_x = put_image_rec->dst_x; - params->dst_w = put_image_rec->dst_width; - - params->src_w = put_image_rec->src_width; - params->src_h = put_image_rec->src_height; - params->src_scan_w = put_image_rec->src_scan_width; - params->src_scan_h = put_image_rec->src_scan_height; - if (params->src_scan_h > params->src_h || - params->src_scan_w > params->src_w) { + + if (params->src_scan_height > params->src_height || + params->src_scan_width > params->src_width) { ret = -EINVAL; goto out_unlock; } - ret = check_overlay_src(dev_priv, put_image_rec, new_bo); + ret = check_overlay_src(dev_priv, params, new_bo); if (ret != 0) goto out_unlock; - params->format = put_image_rec->flags & ~I915_OVERLAY_FLAGS_MASK; - params->stride_Y = put_image_rec->stride_Y; - params->stride_UV = put_image_rec->stride_UV; - params->offset_Y = put_image_rec->offset_Y; - params->offset_U = put_image_rec->offset_U; - params->offset_V = put_image_rec->offset_V; /* Check scaling after src size to prevent a divide-by-zero. */ ret = check_overlay_scaling(params); @@ -1195,16 +1160,12 @@ int intel_overlay_put_image_ioctl(struct drm_device *dev, void *data, drm_modeset_unlock_all(dev); i915_gem_object_put(new_bo); - kfree(params); - return 0; out_unlock: mutex_unlock(&dev->struct_mutex); drm_modeset_unlock_all(dev); i915_gem_object_put(new_bo); -out_free: - kfree(params); return ret; } diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index d99e5fabe93c..1db9b8328275 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -2875,6 +2875,16 @@ static void intel_read_wm_latency(struct drm_i915_private *dev_priv, } } + /* + * WA Level-0 adjustment for 16GB DIMMs: SKL+ + * If we could not get dimm info enable this WA to prevent from + * any underrun. If not able to get Dimm info assume 16GB dimm + * to avoid any underrun. + */ + if (!dev_priv->dram_info.valid_dimm || + dev_priv->dram_info.is_16gb_dimm) + wm[0] += 1; + } else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) { uint64_t sskpd = I915_READ64(MCH_SSKPD); @@ -6108,10 +6118,13 @@ void intel_enable_ipc(struct drm_i915_private *dev_priv) u32 val; /* Display WA #0477 WaDisableIPC: skl */ - if (IS_SKYLAKE(dev_priv)) { + if (IS_SKYLAKE(dev_priv)) + dev_priv->ipc_enabled = false; + + /* Display WA #1141: SKL:all KBL:all CFL */ + if ((IS_KABYLAKE(dev_priv) || IS_COFFEELAKE(dev_priv)) && + !dev_priv->dram_info.symmetric_memory) dev_priv->ipc_enabled = false; - return; - } val = I915_READ(DISP_ARB_CTL2); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 472939f5c18f..d0ef50bf930a 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1677,9 +1677,26 @@ static int switch_context(struct i915_request *rq) GEM_BUG_ON(HAS_EXECLISTS(rq->i915)); if (ppgtt) { - ret = load_pd_dir(rq, ppgtt); - if (ret) - goto err; + int loops; + + /* + * Baytail takes a little more convincing that it really needs + * to reload the PD between contexts. It is not just a little + * longer, as adding more stalls after the load_pd_dir (i.e. + * adding a long loop around flush_pd_dir) is not as effective + * as reloading the PD umpteen times. 32 is derived from + * experimentation (gem_exec_parallel/fds) and has no good + * explanation. + */ + loops = 1; + if (engine->id == BCS && IS_VALLEYVIEW(engine->i915)) + loops = 32; + + do { + ret = load_pd_dir(rq, ppgtt); + if (ret) + goto err; + } while (--loops); if (intel_engine_flag(engine) & ppgtt->pd_dirty_rings) { unwind_mm = intel_engine_flag(engine); diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c index 480dadb1047b..0fdabce647ab 100644 --- a/drivers/gpu/drm/i915/intel_runtime_pm.c +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c @@ -1996,6 +1996,7 @@ void intel_display_power_put(struct drm_i915_private *dev_priv, BIT_ULL(POWER_DOMAIN_PORT_DDI_F_IO)) #define ICL_AUX_A_IO_POWER_DOMAINS ( \ + BIT_ULL(POWER_DOMAIN_AUX_IO_A) | \ BIT_ULL(POWER_DOMAIN_AUX_A)) #define ICL_AUX_B_IO_POWER_DOMAINS ( \ BIT_ULL(POWER_DOMAIN_AUX_B)) @@ -3563,6 +3564,9 @@ static void icl_display_core_init(struct drm_i915_private *dev_priv, /* 7. Setup MBUS. */ icl_mbus_init(dev_priv); + + if (resume && dev_priv->csr.dmc_payload) + intel_csr_load_program(dev_priv); } static void icl_display_core_uninit(struct drm_i915_private *dev_priv) diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c index 812fe7b06f87..701372e512a8 100644 --- a/drivers/gpu/drm/i915/intel_sdvo.c +++ b/drivers/gpu/drm/i915/intel_sdvo.c @@ -99,32 +99,13 @@ struct intel_sdvo { */ uint16_t hotplug_active; - /** - * This is set if we're going to treat the device as TV-out. - * - * While we have these nice friendly flags for output types that ought - * to decide this for us, the S-Video output on our HDMI+S-Video card - * shows up as RGB1 (VGA). - */ - bool is_tv; - enum port port; - /** - * This is set if we treat the device as HDMI, instead of DVI. - */ - bool is_hdmi; bool has_hdmi_monitor; bool has_hdmi_audio; bool rgb_quant_range_selectable; /** - * This is set if we detect output of sdvo device as LVDS and - * have a valid fixed mode to use with the panel. - */ - bool is_lvds; - - /** * This is sdvo fixed pannel mode pointer */ struct drm_display_mode *sdvo_lvds_fixed_mode; @@ -172,6 +153,11 @@ struct intel_sdvo_connector { /* this is to get the range of margin.*/ u32 max_hscan, max_vscan; + + /** + * This is set if we treat the device as HDMI, instead of DVI. + */ + bool is_hdmi; }; struct intel_sdvo_connector_state { @@ -766,6 +752,7 @@ static bool intel_sdvo_get_input_timing(struct intel_sdvo *intel_sdvo, static bool intel_sdvo_create_preferred_input_timing(struct intel_sdvo *intel_sdvo, + struct intel_sdvo_connector *intel_sdvo_connector, uint16_t clock, uint16_t width, uint16_t height) @@ -778,7 +765,7 @@ intel_sdvo_create_preferred_input_timing(struct intel_sdvo *intel_sdvo, args.height = height; args.interlace = 0; - if (intel_sdvo->is_lvds && + if (IS_LVDS(intel_sdvo_connector) && (intel_sdvo->sdvo_lvds_fixed_mode->hdisplay != width || intel_sdvo->sdvo_lvds_fixed_mode->vdisplay != height)) args.scaled = 1; @@ -1067,6 +1054,7 @@ intel_sdvo_set_output_timings_from_mode(struct intel_sdvo *intel_sdvo, */ static bool intel_sdvo_get_preferred_input_mode(struct intel_sdvo *intel_sdvo, + struct intel_sdvo_connector *intel_sdvo_connector, const struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) { @@ -1077,6 +1065,7 @@ intel_sdvo_get_preferred_input_mode(struct intel_sdvo *intel_sdvo, return false; if (!intel_sdvo_create_preferred_input_timing(intel_sdvo, + intel_sdvo_connector, mode->clock / 10, mode->hdisplay, mode->vdisplay)) @@ -1127,6 +1116,8 @@ static bool intel_sdvo_compute_config(struct intel_encoder *encoder, struct intel_sdvo *intel_sdvo = to_sdvo(encoder); struct intel_sdvo_connector_state *intel_sdvo_state = to_intel_sdvo_connector_state(conn_state); + struct intel_sdvo_connector *intel_sdvo_connector = + to_intel_sdvo_connector(conn_state->connector); struct drm_display_mode *adjusted_mode = &pipe_config->base.adjusted_mode; struct drm_display_mode *mode = &pipe_config->base.mode; @@ -1142,20 +1133,22 @@ static bool intel_sdvo_compute_config(struct intel_encoder *encoder, * timings, even though this isn't really the right place in * the sequence to do it. Oh well. */ - if (intel_sdvo->is_tv) { + if (IS_TV(intel_sdvo_connector)) { if (!intel_sdvo_set_output_timings_from_mode(intel_sdvo, mode)) return false; (void) intel_sdvo_get_preferred_input_mode(intel_sdvo, + intel_sdvo_connector, mode, adjusted_mode); pipe_config->sdvo_tv_clock = true; - } else if (intel_sdvo->is_lvds) { + } else if (IS_LVDS(intel_sdvo_connector)) { if (!intel_sdvo_set_output_timings_from_mode(intel_sdvo, intel_sdvo->sdvo_lvds_fixed_mode)) return false; (void) intel_sdvo_get_preferred_input_mode(intel_sdvo, + intel_sdvo_connector, mode, adjusted_mode); } @@ -1194,11 +1187,11 @@ static bool intel_sdvo_compute_config(struct intel_encoder *encoder, } /* Clock computation needs to happen after pixel multiplier. */ - if (intel_sdvo->is_tv) + if (IS_TV(intel_sdvo_connector)) i9xx_adjust_sdvo_tv_clock(pipe_config); /* Set user selected PAR to incoming mode's member */ - if (intel_sdvo->is_hdmi) + if (intel_sdvo_connector->is_hdmi) adjusted_mode->picture_aspect_ratio = conn_state->picture_aspect_ratio; return true; @@ -1275,6 +1268,8 @@ static void intel_sdvo_pre_enable(struct intel_encoder *intel_encoder, const struct drm_display_mode *adjusted_mode = &crtc_state->base.adjusted_mode; const struct intel_sdvo_connector_state *sdvo_state = to_intel_sdvo_connector_state(conn_state); + const struct intel_sdvo_connector *intel_sdvo_connector = + to_intel_sdvo_connector(conn_state->connector); const struct drm_display_mode *mode = &crtc_state->base.mode; struct intel_sdvo *intel_sdvo = to_sdvo(intel_encoder); u32 sdvox; @@ -1304,7 +1299,7 @@ static void intel_sdvo_pre_enable(struct intel_encoder *intel_encoder, return; /* lvds has a special fixed output timing. */ - if (intel_sdvo->is_lvds) + if (IS_LVDS(intel_sdvo_connector)) intel_sdvo_get_dtd_from_mode(&output_dtd, intel_sdvo->sdvo_lvds_fixed_mode); else @@ -1325,13 +1320,13 @@ static void intel_sdvo_pre_enable(struct intel_encoder *intel_encoder, } else intel_sdvo_set_encode(intel_sdvo, SDVO_ENCODE_DVI); - if (intel_sdvo->is_tv && + if (IS_TV(intel_sdvo_connector) && !intel_sdvo_set_tv_format(intel_sdvo, conn_state)) return; intel_sdvo_get_dtd_from_mode(&input_dtd, adjusted_mode); - if (intel_sdvo->is_tv || intel_sdvo->is_lvds) + if (IS_TV(intel_sdvo_connector) || IS_LVDS(intel_sdvo_connector)) input_dtd.part2.sdvo_flags = intel_sdvo->dtd_sdvo_flags; if (!intel_sdvo_set_input_timing(intel_sdvo, &input_dtd)) DRM_INFO("Setting input timings on %s failed\n", @@ -1630,6 +1625,8 @@ intel_sdvo_mode_valid(struct drm_connector *connector, struct drm_display_mode *mode) { struct intel_sdvo *intel_sdvo = intel_attached_sdvo(connector); + struct intel_sdvo_connector *intel_sdvo_connector = + to_intel_sdvo_connector(connector); int max_dotclk = to_i915(connector->dev)->max_dotclk_freq; if (mode->flags & DRM_MODE_FLAG_DBLSCAN) @@ -1644,7 +1641,7 @@ intel_sdvo_mode_valid(struct drm_connector *connector, if (mode->clock > max_dotclk) return MODE_CLOCK_HIGH; - if (intel_sdvo->is_lvds) { + if (IS_LVDS(intel_sdvo_connector)) { if (mode->hdisplay > intel_sdvo->sdvo_lvds_fixed_mode->hdisplay) return MODE_PANEL; @@ -1759,6 +1756,8 @@ static enum drm_connector_status intel_sdvo_tmds_sink_detect(struct drm_connector *connector) { struct intel_sdvo *intel_sdvo = intel_attached_sdvo(connector); + struct intel_sdvo_connector *intel_sdvo_connector = + to_intel_sdvo_connector(connector); enum drm_connector_status status; struct edid *edid; @@ -1797,7 +1796,7 @@ intel_sdvo_tmds_sink_detect(struct drm_connector *connector) /* DDC bus is shared, match EDID to connector type */ if (edid->input & DRM_EDID_INPUT_DIGITAL) { status = connector_status_connected; - if (intel_sdvo->is_hdmi) { + if (intel_sdvo_connector->is_hdmi) { intel_sdvo->has_hdmi_monitor = drm_detect_hdmi_monitor(edid); intel_sdvo->has_hdmi_audio = drm_detect_monitor_audio(edid); intel_sdvo->rgb_quant_range_selectable = @@ -1875,17 +1874,6 @@ intel_sdvo_detect(struct drm_connector *connector, bool force) ret = connector_status_connected; } - /* May update encoder flag for like clock for SDVO TV, etc.*/ - if (ret == connector_status_connected) { - intel_sdvo->is_tv = false; - intel_sdvo->is_lvds = false; - - if (response & SDVO_TV_MASK) - intel_sdvo->is_tv = true; - if (response & SDVO_LVDS_MASK) - intel_sdvo->is_lvds = intel_sdvo->sdvo_lvds_fixed_mode != NULL; - } - return ret; } @@ -2054,16 +2042,6 @@ static void intel_sdvo_get_lvds_modes(struct drm_connector *connector) * arranged in priority order. */ intel_ddc_get_modes(connector, &intel_sdvo->ddc); - - list_for_each_entry(newmode, &connector->probed_modes, head) { - if (newmode->type & DRM_MODE_TYPE_PREFERRED) { - intel_sdvo->sdvo_lvds_fixed_mode = - drm_mode_duplicate(connector->dev, newmode); - - intel_sdvo->is_lvds = true; - break; - } - } } static int intel_sdvo_get_modes(struct drm_connector *connector) @@ -2555,7 +2533,7 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device) if (INTEL_GEN(dev_priv) >= 4 && intel_sdvo_is_hdmi_connector(intel_sdvo, device)) { connector->connector_type = DRM_MODE_CONNECTOR_HDMIA; - intel_sdvo->is_hdmi = true; + intel_sdvo_connector->is_hdmi = true; } if (intel_sdvo_connector_init(intel_sdvo_connector, intel_sdvo) < 0) { @@ -2563,7 +2541,7 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device) return false; } - if (intel_sdvo->is_hdmi) + if (intel_sdvo_connector->is_hdmi) intel_sdvo_add_hdmi_properties(intel_sdvo, intel_sdvo_connector); return true; @@ -2591,8 +2569,6 @@ intel_sdvo_tv_init(struct intel_sdvo *intel_sdvo, int type) intel_sdvo->controlled_output |= type; intel_sdvo_connector->output_flag = type; - intel_sdvo->is_tv = true; - if (intel_sdvo_connector_init(intel_sdvo_connector, intel_sdvo) < 0) { kfree(intel_sdvo_connector); return false; @@ -2654,6 +2630,7 @@ intel_sdvo_lvds_init(struct intel_sdvo *intel_sdvo, int device) struct drm_connector *connector; struct intel_connector *intel_connector; struct intel_sdvo_connector *intel_sdvo_connector; + struct drm_display_mode *mode; DRM_DEBUG_KMS("initialising LVDS device %d\n", device); @@ -2682,6 +2659,19 @@ intel_sdvo_lvds_init(struct intel_sdvo *intel_sdvo, int device) if (!intel_sdvo_create_enhance_property(intel_sdvo, intel_sdvo_connector)) goto err; + intel_sdvo_get_lvds_modes(connector); + + list_for_each_entry(mode, &connector->probed_modes, head) { + if (mode->type & DRM_MODE_TYPE_PREFERRED) { + intel_sdvo->sdvo_lvds_fixed_mode = + drm_mode_duplicate(connector->dev, mode); + break; + } + } + + if (!intel_sdvo->sdvo_lvds_fixed_mode) + goto err; + return true; err: @@ -2692,9 +2682,6 @@ err: static bool intel_sdvo_output_setup(struct intel_sdvo *intel_sdvo, uint16_t flags) { - intel_sdvo->is_tv = false; - intel_sdvo->is_lvds = false; - /* SDVO requires XXX1 function may not exist unless it has XXX0 function.*/ if (flags & SDVO_OUTPUT_TMDS0) diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c index 9600ccfc5b76..5fd2f7bf3927 100644 --- a/drivers/gpu/drm/i915/intel_sprite.c +++ b/drivers/gpu/drm/i915/intel_sprite.c @@ -230,6 +230,78 @@ void intel_pipe_update_end(struct intel_crtc_state *new_crtc_state) #endif } +int intel_plane_check_stride(const struct intel_plane_state *plane_state) +{ + struct intel_plane *plane = to_intel_plane(plane_state->base.plane); + const struct drm_framebuffer *fb = plane_state->base.fb; + unsigned int rotation = plane_state->base.rotation; + u32 stride, max_stride; + + /* FIXME other color planes? */ + stride = plane_state->color_plane[0].stride; + max_stride = plane->max_stride(plane, fb->format->format, + fb->modifier, rotation); + + if (stride > max_stride) { + DRM_DEBUG_KMS("[FB:%d] stride (%d) exceeds [PLANE:%d:%s] max stride (%d)\n", + fb->base.id, stride, + plane->base.base.id, plane->base.name, max_stride); + return -EINVAL; + } + + return 0; +} + +int intel_plane_check_src_coordinates(struct intel_plane_state *plane_state) +{ + const struct drm_framebuffer *fb = plane_state->base.fb; + struct drm_rect *src = &plane_state->base.src; + u32 src_x, src_y, src_w, src_h; + + /* + * Hardware doesn't handle subpixel coordinates. + * Adjust to (macro)pixel boundary, but be careful not to + * increase the source viewport size, because that could + * push the downscaling factor out of bounds. + */ + src_x = src->x1 >> 16; + src_w = drm_rect_width(src) >> 16; + src_y = src->y1 >> 16; + src_h = drm_rect_height(src) >> 16; + + src->x1 = src_x << 16; + src->x2 = (src_x + src_w) << 16; + src->y1 = src_y << 16; + src->y2 = (src_y + src_h) << 16; + + if (fb->format->is_yuv && + fb->format->format != DRM_FORMAT_NV12 && + (src_x & 1 || src_w & 1)) { + DRM_DEBUG_KMS("src x/w (%u, %u) must be a multiple of 2 for YUV planes\n", + src_x, src_w); + return -EINVAL; + } + + return 0; +} + +unsigned int +skl_plane_max_stride(struct intel_plane *plane, + u32 pixel_format, u64 modifier, + unsigned int rotation) +{ + int cpp = drm_format_plane_cpp(pixel_format, 0); + + /* + * "The stride in bytes must not exceed the + * of the size of 8K pixels and 32K bytes." + */ + if (drm_rotation_90_or_270(rotation)) + return min(8192, 32768 / cpp); + else + return min(8192 * cpp, 32768); +} + void skl_update_plane(struct intel_plane *plane, const struct intel_crtc_state *crtc_state, @@ -241,16 +313,15 @@ skl_update_plane(struct intel_plane *plane, enum pipe pipe = plane->pipe; u32 plane_ctl = plane_state->ctl; const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; - u32 surf_addr = plane_state->main.offset; - unsigned int rotation = plane_state->base.rotation; - u32 stride = skl_plane_stride(fb, 0, rotation); - u32 aux_stride = skl_plane_stride(fb, 1, rotation); + u32 surf_addr = plane_state->color_plane[0].offset; + u32 stride = skl_plane_stride(plane_state, 0); + u32 aux_stride = skl_plane_stride(plane_state, 1); int crtc_x = plane_state->base.dst.x1; int crtc_y = plane_state->base.dst.y1; uint32_t crtc_w = drm_rect_width(&plane_state->base.dst); uint32_t crtc_h = drm_rect_height(&plane_state->base.dst); - uint32_t x = plane_state->main.x; - uint32_t y = plane_state->main.y; + uint32_t x = plane_state->color_plane[0].x; + uint32_t y = plane_state->color_plane[0].y; uint32_t src_w = drm_rect_width(&plane_state->base.src) >> 16; uint32_t src_h = drm_rect_height(&plane_state->base.src) >> 16; unsigned long irqflags; @@ -277,9 +348,10 @@ skl_update_plane(struct intel_plane *plane, I915_WRITE_FW(PLANE_STRIDE(pipe, plane_id), stride); I915_WRITE_FW(PLANE_SIZE(pipe, plane_id), (src_h << 16) | src_w); I915_WRITE_FW(PLANE_AUX_DIST(pipe, plane_id), - (plane_state->aux.offset - surf_addr) | aux_stride); + (plane_state->color_plane[1].offset - surf_addr) | aux_stride); I915_WRITE_FW(PLANE_AUX_OFFSET(pipe, plane_id), - (plane_state->aux.y << 16) | plane_state->aux.x); + (plane_state->color_plane[1].y << 16) | + plane_state->color_plane[1].x); /* program plane scaler */ if (plane_state->scaler_id >= 0) { @@ -545,15 +617,15 @@ vlv_update_plane(struct intel_plane *plane, enum pipe pipe = plane->pipe; enum plane_id plane_id = plane->id; u32 sprctl = plane_state->ctl; - u32 sprsurf_offset = plane_state->main.offset; + u32 sprsurf_offset = plane_state->color_plane[0].offset; u32 linear_offset; const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; int crtc_x = plane_state->base.dst.x1; int crtc_y = plane_state->base.dst.y1; uint32_t crtc_w = drm_rect_width(&plane_state->base.dst); uint32_t crtc_h = drm_rect_height(&plane_state->base.dst); - uint32_t x = plane_state->main.x; - uint32_t y = plane_state->main.y; + uint32_t x = plane_state->color_plane[0].x; + uint32_t y = plane_state->color_plane[0].y; unsigned long irqflags; /* Sizes are 0 based */ @@ -574,7 +646,8 @@ vlv_update_plane(struct intel_plane *plane, I915_WRITE_FW(SPKEYMAXVAL(pipe, plane_id), key->max_value); I915_WRITE_FW(SPKEYMSK(pipe, plane_id), key->channel_mask); } - I915_WRITE_FW(SPSTRIDE(pipe, plane_id), fb->pitches[0]); + I915_WRITE_FW(SPSTRIDE(pipe, plane_id), + plane_state->color_plane[0].stride); I915_WRITE_FW(SPPOS(pipe, plane_id), (crtc_y << 16) | crtc_x); if (fb->modifier == I915_FORMAT_MOD_X_TILED) @@ -704,15 +777,15 @@ ivb_update_plane(struct intel_plane *plane, const struct drm_framebuffer *fb = plane_state->base.fb; enum pipe pipe = plane->pipe; u32 sprctl = plane_state->ctl, sprscale = 0; - u32 sprsurf_offset = plane_state->main.offset; + u32 sprsurf_offset = plane_state->color_plane[0].offset; u32 linear_offset; const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; int crtc_x = plane_state->base.dst.x1; int crtc_y = plane_state->base.dst.y1; uint32_t crtc_w = drm_rect_width(&plane_state->base.dst); uint32_t crtc_h = drm_rect_height(&plane_state->base.dst); - uint32_t x = plane_state->main.x; - uint32_t y = plane_state->main.y; + uint32_t x = plane_state->color_plane[0].x; + uint32_t y = plane_state->color_plane[0].y; uint32_t src_w = drm_rect_width(&plane_state->base.src) >> 16; uint32_t src_h = drm_rect_height(&plane_state->base.src) >> 16; unsigned long irqflags; @@ -736,7 +809,7 @@ ivb_update_plane(struct intel_plane *plane, I915_WRITE_FW(SPRKEYMSK(pipe), key->channel_mask); } - I915_WRITE_FW(SPRSTRIDE(pipe), fb->pitches[0]); + I915_WRITE_FW(SPRSTRIDE(pipe), plane_state->color_plane[0].stride); I915_WRITE_FW(SPRPOS(pipe), (crtc_y << 16) | crtc_x); /* HSW consolidates SPRTILEOFF and SPRLINOFF into a single SPROFFSET @@ -749,7 +822,7 @@ ivb_update_plane(struct intel_plane *plane, I915_WRITE_FW(SPRLINOFF(pipe), linear_offset); I915_WRITE_FW(SPRSIZE(pipe), (crtc_h << 16) | crtc_w); - if (plane->can_scale) + if (IS_IVYBRIDGE(dev_priv)) I915_WRITE_FW(SPRSCALE(pipe), sprscale); I915_WRITE_FW(SPRCTL(pipe), sprctl); I915_WRITE_FW(SPRSURF(pipe), @@ -770,7 +843,7 @@ ivb_disable_plane(struct intel_plane *plane, struct intel_crtc *crtc) I915_WRITE_FW(SPRCTL(pipe), 0); /* Can't leave the scaler enabled... */ - if (plane->can_scale) + if (IS_IVYBRIDGE(dev_priv)) I915_WRITE_FW(SPRSCALE(pipe), 0); I915_WRITE_FW(SPRSURF(pipe), 0); @@ -800,6 +873,14 @@ ivb_plane_get_hw_state(struct intel_plane *plane, return ret; } +static unsigned int +g4x_sprite_max_stride(struct intel_plane *plane, + u32 pixel_format, u64 modifier, + unsigned int rotation) +{ + return 16384; +} + static u32 g4x_sprite_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { @@ -868,15 +949,15 @@ g4x_update_plane(struct intel_plane *plane, const struct drm_framebuffer *fb = plane_state->base.fb; enum pipe pipe = plane->pipe; u32 dvscntr = plane_state->ctl, dvsscale = 0; - u32 dvssurf_offset = plane_state->main.offset; + u32 dvssurf_offset = plane_state->color_plane[0].offset; u32 linear_offset; const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; int crtc_x = plane_state->base.dst.x1; int crtc_y = plane_state->base.dst.y1; uint32_t crtc_w = drm_rect_width(&plane_state->base.dst); uint32_t crtc_h = drm_rect_height(&plane_state->base.dst); - uint32_t x = plane_state->main.x; - uint32_t y = plane_state->main.y; + uint32_t x = plane_state->color_plane[0].x; + uint32_t y = plane_state->color_plane[0].y; uint32_t src_w = drm_rect_width(&plane_state->base.src) >> 16; uint32_t src_h = drm_rect_height(&plane_state->base.src) >> 16; unsigned long irqflags; @@ -900,7 +981,7 @@ g4x_update_plane(struct intel_plane *plane, I915_WRITE_FW(DVSKEYMSK(pipe), key->channel_mask); } - I915_WRITE_FW(DVSSTRIDE(pipe), fb->pitches[0]); + I915_WRITE_FW(DVSSTRIDE(pipe), plane_state->color_plane[0].stride); I915_WRITE_FW(DVSPOS(pipe), (crtc_y << 16) | crtc_x); if (fb->modifier == I915_FORMAT_MOD_X_TILED) @@ -959,144 +1040,309 @@ g4x_plane_get_hw_state(struct intel_plane *plane, } static int -intel_check_sprite_plane(struct intel_crtc_state *crtc_state, - struct intel_plane_state *state) +g4x_sprite_check_scaling(struct intel_crtc_state *crtc_state, + struct intel_plane_state *plane_state) { - struct intel_plane *plane = to_intel_plane(state->base.plane); - struct drm_i915_private *dev_priv = to_i915(plane->base.dev); - struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); - struct drm_framebuffer *fb = state->base.fb; - int max_stride = INTEL_GEN(dev_priv) >= 9 ? 32768 : 16384; - int max_scale, min_scale; - bool can_scale; - int ret; - uint32_t pixel_format = 0; - - if (!fb) { - state->base.visible = false; + const struct drm_framebuffer *fb = plane_state->base.fb; + const struct drm_rect *src = &plane_state->base.src; + const struct drm_rect *dst = &plane_state->base.dst; + int src_x, src_y, src_w, src_h, crtc_w, crtc_h; + const struct drm_display_mode *adjusted_mode = + &crtc_state->base.adjusted_mode; + unsigned int cpp = fb->format->cpp[0]; + unsigned int width_bytes; + int min_width, min_height; + + crtc_w = drm_rect_width(dst); + crtc_h = drm_rect_height(dst); + + src_x = src->x1 >> 16; + src_y = src->y1 >> 16; + src_w = drm_rect_width(src) >> 16; + src_h = drm_rect_height(src) >> 16; + + if (src_w == crtc_w && src_h == crtc_h) return 0; + + min_width = 3; + + if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) { + if (src_h & 1) { + DRM_DEBUG_KMS("Source height must be even with interlaced modes\n"); + return -EINVAL; + } + min_height = 6; + } else { + min_height = 3; } - /* Don't modify another pipe's plane */ - if (plane->pipe != crtc->pipe) { - DRM_DEBUG_KMS("Wrong plane <-> crtc mapping\n"); + width_bytes = ((src_x * cpp) & 63) + src_w * cpp; + + if (src_w < min_width || src_h < min_height || + src_w > 2048 || src_h > 2048) { + DRM_DEBUG_KMS("Source dimensions (%dx%d) exceed hardware limits (%dx%d - %dx%d)\n", + src_w, src_h, min_width, min_height, 2048, 2048); return -EINVAL; } - /* FIXME check all gen limits */ - if (fb->width < 3 || fb->height < 3 || fb->pitches[0] > max_stride) { - DRM_DEBUG_KMS("Unsuitable framebuffer for plane\n"); + if (width_bytes > 4096) { + DRM_DEBUG_KMS("Fetch width (%d) exceeds hardware max with scaling (%u)\n", + width_bytes, 4096); return -EINVAL; } - /* setup can_scale, min_scale, max_scale */ - if (INTEL_GEN(dev_priv) >= 9) { - if (state->base.fb) - pixel_format = state->base.fb->format->format; - /* use scaler when colorkey is not required */ - if (!state->ckey.flags) { - can_scale = 1; - min_scale = 1; - max_scale = - skl_max_scale(crtc, crtc_state, pixel_format); - } else { - can_scale = 0; - min_scale = DRM_PLANE_HELPER_NO_SCALING; - max_scale = DRM_PLANE_HELPER_NO_SCALING; - } + if (width_bytes > 4096 || fb->pitches[0] > 4096) { + DRM_DEBUG_KMS("Stride (%u) exceeds hardware max with scaling (%u)\n", + fb->pitches[0], 4096); + return -EINVAL; + } + + return 0; +} + +static int +g4x_sprite_check(struct intel_crtc_state *crtc_state, + struct intel_plane_state *plane_state) +{ + struct intel_plane *plane = to_intel_plane(plane_state->base.plane); + struct drm_i915_private *dev_priv = to_i915(plane->base.dev); + int max_scale, min_scale; + int ret; + + if (INTEL_GEN(dev_priv) < 7) { + min_scale = 1; + max_scale = 16 << 16; + } else if (IS_IVYBRIDGE(dev_priv)) { + min_scale = 1; + max_scale = 2 << 16; } else { - can_scale = plane->can_scale; - max_scale = plane->max_downscale << 16; - min_scale = plane->can_scale ? 1 : (1 << 16); + min_scale = DRM_PLANE_HELPER_NO_SCALING; + max_scale = DRM_PLANE_HELPER_NO_SCALING; } - ret = drm_atomic_helper_check_plane_state(&state->base, + ret = drm_atomic_helper_check_plane_state(&plane_state->base, &crtc_state->base, min_scale, max_scale, true, true); if (ret) return ret; - if (state->base.visible) { - struct drm_rect *src = &state->base.src; - struct drm_rect *dst = &state->base.dst; - unsigned int crtc_w = drm_rect_width(dst); - unsigned int crtc_h = drm_rect_height(dst); - uint32_t src_x, src_y, src_w, src_h; + if (!plane_state->base.visible) + return 0; + + ret = intel_plane_check_src_coordinates(plane_state); + if (ret) + return ret; + + ret = g4x_sprite_check_scaling(crtc_state, plane_state); + if (ret) + return ret; + + ret = i9xx_check_plane_surface(plane_state); + if (ret) + return ret; + + if (INTEL_GEN(dev_priv) >= 7) + plane_state->ctl = ivb_sprite_ctl(crtc_state, plane_state); + else + plane_state->ctl = g4x_sprite_ctl(crtc_state, plane_state); + + return 0; +} + +int chv_plane_check_rotation(const struct intel_plane_state *plane_state) +{ + struct intel_plane *plane = to_intel_plane(plane_state->base.plane); + struct drm_i915_private *dev_priv = to_i915(plane->base.dev); + unsigned int rotation = plane_state->base.rotation; + + /* CHV ignores the mirror bit when the rotate bit is set :( */ + if (IS_CHERRYVIEW(dev_priv) && + rotation & DRM_MODE_ROTATE_180 && + rotation & DRM_MODE_REFLECT_X) { + DRM_DEBUG_KMS("Cannot rotate and reflect at the same time\n"); + return -EINVAL; + } + + return 0; +} + +static int +vlv_sprite_check(struct intel_crtc_state *crtc_state, + struct intel_plane_state *plane_state) +{ + int ret; + + ret = chv_plane_check_rotation(plane_state); + if (ret) + return ret; + + ret = drm_atomic_helper_check_plane_state(&plane_state->base, + &crtc_state->base, + DRM_PLANE_HELPER_NO_SCALING, + DRM_PLANE_HELPER_NO_SCALING, + true, true); + if (ret) + return ret; + + if (!plane_state->base.visible) + return 0; + + ret = intel_plane_check_src_coordinates(plane_state); + if (ret) + return ret; + + ret = i9xx_check_plane_surface(plane_state); + if (ret) + return ret; + + plane_state->ctl = vlv_sprite_ctl(crtc_state, plane_state); + + return 0; +} + +static int skl_plane_check_fb(const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state) +{ + const struct drm_framebuffer *fb = plane_state->base.fb; + unsigned int rotation = plane_state->base.rotation; + struct drm_format_name_buf format_name; + + if (!fb) + return 0; + + if (rotation & ~(DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_180) && + is_ccs_modifier(fb->modifier)) { + DRM_DEBUG_KMS("RC support only with 0/180 degree rotation (%x)\n", + rotation); + return -EINVAL; + } + + if (rotation & DRM_MODE_REFLECT_X && + fb->modifier == DRM_FORMAT_MOD_LINEAR) { + DRM_DEBUG_KMS("horizontal flip is not supported with linear surface formats\n"); + return -EINVAL; + } + + if (drm_rotation_90_or_270(rotation)) { + if (fb->modifier != I915_FORMAT_MOD_Y_TILED && + fb->modifier != I915_FORMAT_MOD_Yf_TILED) { + DRM_DEBUG_KMS("Y/Yf tiling required for 90/270!\n"); + return -EINVAL; + } /* - * Hardware doesn't handle subpixel coordinates. - * Adjust to (macro)pixel boundary, but be careful not to - * increase the source viewport size, because that could - * push the downscaling factor out of bounds. + * 90/270 is not allowed with RGB64 16:16:16:16, + * RGB 16-bit 5:6:5, and Indexed 8-bit. + * TBD: Add RGB64 case once its added in supported format list. */ - src_x = src->x1 >> 16; - src_w = drm_rect_width(src) >> 16; - src_y = src->y1 >> 16; - src_h = drm_rect_height(src) >> 16; - - src->x1 = src_x << 16; - src->x2 = (src_x + src_w) << 16; - src->y1 = src_y << 16; - src->y2 = (src_y + src_h) << 16; - - if (fb->format->is_yuv && - fb->format->format != DRM_FORMAT_NV12 && - (src_x % 2 || src_w % 2)) { - DRM_DEBUG_KMS("src x/w (%u, %u) must be a multiple of 2 for YUV planes\n", - src_x, src_w); + switch (fb->format->format) { + case DRM_FORMAT_C8: + case DRM_FORMAT_RGB565: + DRM_DEBUG_KMS("Unsupported pixel format %s for 90/270!\n", + drm_get_format_name(fb->format->format, + &format_name)); return -EINVAL; + default: + break; } + } - /* Check size restrictions when scaling */ - if (src_w != crtc_w || src_h != crtc_h) { - unsigned int width_bytes; - int cpp = fb->format->cpp[0]; + /* Y-tiling is not supported in IF-ID Interlace mode */ + if (crtc_state->base.enable && + crtc_state->base.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE && + (fb->modifier == I915_FORMAT_MOD_Y_TILED || + fb->modifier == I915_FORMAT_MOD_Yf_TILED || + fb->modifier == I915_FORMAT_MOD_Y_TILED_CCS || + fb->modifier == I915_FORMAT_MOD_Yf_TILED_CCS)) { + DRM_DEBUG_KMS("Y/Yf tiling not supported in IF-ID mode\n"); + return -EINVAL; + } - WARN_ON(!can_scale); + return 0; +} - width_bytes = ((src_x * cpp) & 63) + src_w * cpp; +static int skl_plane_check_dst_coordinates(const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state) +{ + struct drm_i915_private *dev_priv = + to_i915(plane_state->base.plane->dev); + int crtc_x = plane_state->base.dst.x1; + int crtc_w = drm_rect_width(&plane_state->base.dst); + int pipe_src_w = crtc_state->pipe_src_w; - /* FIXME interlacing min height is 6 */ - if (INTEL_GEN(dev_priv) < 9 && ( - src_w < 3 || src_h < 3 || - src_w > 2048 || src_h > 2048 || - crtc_w < 3 || crtc_h < 3 || - width_bytes > 4096 || fb->pitches[0] > 4096)) { - DRM_DEBUG_KMS("Source dimensions exceed hardware limits\n"); - return -EINVAL; - } - } + /* + * Display WA #1175: cnl,glk + * Planes other than the cursor may cause FIFO underflow and display + * corruption if starting less than 4 pixels from the right edge of + * the screen. + * Besides the above WA fix the similar problem, where planes other + * than the cursor ending less than 4 pixels from the left edge of the + * screen may cause FIFO underflow and display corruption. + */ + if ((IS_GEMINILAKE(dev_priv) || IS_CANNONLAKE(dev_priv)) && + (crtc_x + crtc_w < 4 || crtc_x > pipe_src_w - 4)) { + DRM_DEBUG_KMS("requested plane X %s position %d invalid (valid range %d-%d)\n", + crtc_x + crtc_w < 4 ? "end" : "start", + crtc_x + crtc_w < 4 ? crtc_x + crtc_w : crtc_x, + 4, pipe_src_w - 4); + return -ERANGE; } - if (INTEL_GEN(dev_priv) >= 9) { - ret = skl_check_plane_surface(crtc_state, state); - if (ret) - return ret; + return 0; +} - state->ctl = skl_plane_ctl(crtc_state, state); - } else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { - ret = i9xx_check_plane_surface(state); - if (ret) - return ret; +int skl_plane_check(struct intel_crtc_state *crtc_state, + struct intel_plane_state *plane_state) +{ + struct intel_plane *plane = to_intel_plane(plane_state->base.plane); + struct drm_i915_private *dev_priv = to_i915(plane->base.dev); + int max_scale, min_scale; + int ret; - state->ctl = vlv_sprite_ctl(crtc_state, state); - } else if (INTEL_GEN(dev_priv) >= 7) { - ret = i9xx_check_plane_surface(state); - if (ret) - return ret; + ret = skl_plane_check_fb(crtc_state, plane_state); + if (ret) + return ret; - state->ctl = ivb_sprite_ctl(crtc_state, state); - } else { - ret = i9xx_check_plane_surface(state); - if (ret) - return ret; + /* use scaler when colorkey is not required */ + if (!plane_state->ckey.flags) { + const struct drm_framebuffer *fb = plane_state->base.fb; - state->ctl = g4x_sprite_ctl(crtc_state, state); + min_scale = 1; + max_scale = skl_max_scale(crtc_state, + fb ? fb->format->format : 0); + } else { + min_scale = DRM_PLANE_HELPER_NO_SCALING; + max_scale = DRM_PLANE_HELPER_NO_SCALING; } + ret = drm_atomic_helper_check_plane_state(&plane_state->base, + &crtc_state->base, + min_scale, max_scale, + true, true); + if (ret) + return ret; + + if (!plane_state->base.visible) + return 0; + + ret = skl_plane_check_dst_coordinates(crtc_state, plane_state); + if (ret) + return ret; + + ret = intel_plane_check_src_coordinates(plane_state); + if (ret) + return ret; + + ret = skl_check_plane_surface(plane_state); + if (ret) + return ret; + + plane_state->ctl = skl_plane_ctl(crtc_state, plane_state); + if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv)) - state->color_ctl = glk_plane_color_ctl(crtc_state, state); + plane_state->color_ctl = glk_plane_color_ctl(crtc_state, + plane_state); return 0; } @@ -1523,15 +1769,16 @@ intel_sprite_plane_create(struct drm_i915_private *dev_priv, intel_plane->base.state = &state->base; if (INTEL_GEN(dev_priv) >= 9) { - intel_plane->can_scale = true; state->scaler_id = -1; intel_plane->has_ccs = skl_plane_has_ccs(dev_priv, pipe, PLANE_SPRITE0 + plane); + intel_plane->max_stride = skl_plane_max_stride; intel_plane->update_plane = skl_update_plane; intel_plane->disable_plane = skl_disable_plane; intel_plane->get_hw_state = skl_plane_get_hw_state; + intel_plane->check_plane = skl_plane_check; if (skl_plane_has_planar(dev_priv, pipe, PLANE_SPRITE0 + plane)) { @@ -1549,12 +1796,11 @@ intel_sprite_plane_create(struct drm_i915_private *dev_priv, plane_funcs = &skl_plane_funcs; } else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { - intel_plane->can_scale = false; - intel_plane->max_downscale = 1; - + intel_plane->max_stride = i9xx_plane_max_stride; intel_plane->update_plane = vlv_update_plane; intel_plane->disable_plane = vlv_disable_plane; intel_plane->get_hw_state = vlv_plane_get_hw_state; + intel_plane->check_plane = vlv_sprite_check; plane_formats = vlv_plane_formats; num_plane_formats = ARRAY_SIZE(vlv_plane_formats); @@ -1562,17 +1808,11 @@ intel_sprite_plane_create(struct drm_i915_private *dev_priv, plane_funcs = &vlv_sprite_funcs; } else if (INTEL_GEN(dev_priv) >= 7) { - if (IS_IVYBRIDGE(dev_priv)) { - intel_plane->can_scale = true; - intel_plane->max_downscale = 2; - } else { - intel_plane->can_scale = false; - intel_plane->max_downscale = 1; - } - + intel_plane->max_stride = g4x_sprite_max_stride; intel_plane->update_plane = ivb_update_plane; intel_plane->disable_plane = ivb_disable_plane; intel_plane->get_hw_state = ivb_plane_get_hw_state; + intel_plane->check_plane = g4x_sprite_check; plane_formats = snb_plane_formats; num_plane_formats = ARRAY_SIZE(snb_plane_formats); @@ -1580,12 +1820,11 @@ intel_sprite_plane_create(struct drm_i915_private *dev_priv, plane_funcs = &snb_sprite_funcs; } else { - intel_plane->can_scale = true; - intel_plane->max_downscale = 16; - + intel_plane->max_stride = g4x_sprite_max_stride; intel_plane->update_plane = g4x_update_plane; intel_plane->disable_plane = g4x_disable_plane; intel_plane->get_hw_state = g4x_plane_get_hw_state; + intel_plane->check_plane = g4x_sprite_check; modifiers = i9xx_plane_format_modifiers; if (IS_GEN6(dev_priv)) { @@ -1618,7 +1857,6 @@ intel_sprite_plane_create(struct drm_i915_private *dev_priv, intel_plane->i9xx_plane = plane; intel_plane->id = PLANE_SPRITE0 + plane; intel_plane->frontbuffer_bit = INTEL_FRONTBUFFER(pipe, intel_plane->id); - intel_plane->check_plane = intel_check_sprite_plane; possible_crtcs = (1 << pipe); diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 7c95697e1a35..b1b3e81b6e24 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -401,6 +401,10 @@ int intel_uc_init_hw(struct drm_i915_private *i915) ret = intel_guc_submission_enable(guc); if (ret) goto err_communication; + } else if (INTEL_GEN(i915) < 11) { + ret = intel_guc_sample_forcewake(guc); + if (ret) + goto err_communication; } dev_info(i915->drm.dev, "GuC firmware version %u.%u\n", diff --git a/drivers/gpu/drm/i915/selftests/huge_pages.c b/drivers/gpu/drm/i915/selftests/huge_pages.c index e272127783fe..8d03f64eabd7 100644 --- a/drivers/gpu/drm/i915/selftests/huge_pages.c +++ b/drivers/gpu/drm/i915/selftests/huge_pages.c @@ -235,6 +235,8 @@ static int fake_get_huge_pages(struct drm_i915_gem_object *obj) sg = sg_next(sg); } while (1); + i915_sg_trim(st); + obj->mm.madv = I915_MADV_DONTNEED; __i915_gem_object_set_pages(obj, st, sg_page_sizes); diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_coherency.c b/drivers/gpu/drm/i915/selftests/i915_gem_coherency.c index 4e6a221063ac..f7392c1ffe75 100644 --- a/drivers/gpu/drm/i915/selftests/i915_gem_coherency.c +++ b/drivers/gpu/drm/i915/selftests/i915_gem_coherency.c @@ -298,6 +298,7 @@ static int igt_gem_coherency(void *arg) values = offsets + ncachelines; mutex_lock(&i915->drm.struct_mutex); + intel_runtime_pm_get(i915); for (over = igt_coherency_mode; over->name; over++) { if (!over->set) continue; @@ -375,6 +376,7 @@ static int igt_gem_coherency(void *arg) } } unlock: + intel_runtime_pm_put(i915); mutex_unlock(&i915->drm.struct_mutex); kfree(offsets); return err; diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_context.c b/drivers/gpu/drm/i915/selftests/i915_gem_context.c index 1c92560d35da..76df25aa90c9 100644 --- a/drivers/gpu/drm/i915/selftests/i915_gem_context.c +++ b/drivers/gpu/drm/i915/selftests/i915_gem_context.c @@ -22,6 +22,8 @@ * */ +#include <linux/prime_numbers.h> + #include "../i915_selftest.h" #include "i915_random.h" #include "igt_flush_test.h" @@ -32,6 +34,200 @@ #define DW_PER_PAGE (PAGE_SIZE / sizeof(u32)) +struct live_test { + struct drm_i915_private *i915; + const char *func; + const char *name; + + unsigned int reset_count; +}; + +static int begin_live_test(struct live_test *t, + struct drm_i915_private *i915, + const char *func, + const char *name) +{ + int err; + + t->i915 = i915; + t->func = func; + t->name = name; + + err = i915_gem_wait_for_idle(i915, + I915_WAIT_LOCKED, + MAX_SCHEDULE_TIMEOUT); + if (err) { + pr_err("%s(%s): failed to idle before, with err=%d!", + func, name, err); + return err; + } + + i915->gpu_error.missed_irq_rings = 0; + t->reset_count = i915_reset_count(&i915->gpu_error); + + return 0; +} + +static int end_live_test(struct live_test *t) +{ + struct drm_i915_private *i915 = t->i915; + + if (igt_flush_test(i915, I915_WAIT_LOCKED)) + return -EIO; + + if (t->reset_count != i915_reset_count(&i915->gpu_error)) { + pr_err("%s(%s): GPU was reset %d times!\n", + t->func, t->name, + i915_reset_count(&i915->gpu_error) - t->reset_count); + return -EIO; + } + + if (i915->gpu_error.missed_irq_rings) { + pr_err("%s(%s): Missed interrupts on engines %lx\n", + t->func, t->name, i915->gpu_error.missed_irq_rings); + return -EIO; + } + + return 0; +} + +static int live_nop_switch(void *arg) +{ + const unsigned int nctx = 1024; + struct drm_i915_private *i915 = arg; + struct intel_engine_cs *engine; + struct i915_gem_context **ctx; + enum intel_engine_id id; + struct drm_file *file; + struct live_test t; + unsigned long n; + int err = -ENODEV; + + /* + * Create as many contexts as we can feasibly get away with + * and check we can switch between them rapidly. + * + * Serves as very simple stress test for submission and HW switching + * between contexts. + */ + + if (!DRIVER_CAPS(i915)->has_logical_contexts) + return 0; + + file = mock_file(i915); + if (IS_ERR(file)) + return PTR_ERR(file); + + mutex_lock(&i915->drm.struct_mutex); + intel_runtime_pm_get(i915); + + ctx = kcalloc(nctx, sizeof(*ctx), GFP_KERNEL); + if (!ctx) { + err = -ENOMEM; + goto out_unlock; + } + + for (n = 0; n < nctx; n++) { + ctx[n] = i915_gem_create_context(i915, file->driver_priv); + if (IS_ERR(ctx[n])) { + err = PTR_ERR(ctx[n]); + goto out_unlock; + } + } + + for_each_engine(engine, i915, id) { + struct i915_request *rq; + unsigned long end_time, prime; + ktime_t times[2] = {}; + + times[0] = ktime_get_raw(); + for (n = 0; n < nctx; n++) { + rq = i915_request_alloc(engine, ctx[n]); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + goto out_unlock; + } + i915_request_add(rq); + } + if (i915_request_wait(rq, + I915_WAIT_LOCKED, + HZ / 5) < 0) { + pr_err("Failed to populated %d contexts\n", nctx); + i915_gem_set_wedged(i915); + err = -EIO; + goto out_unlock; + } + + times[1] = ktime_get_raw(); + + pr_info("Populated %d contexts on %s in %lluns\n", + nctx, engine->name, ktime_to_ns(times[1] - times[0])); + + err = begin_live_test(&t, i915, __func__, engine->name); + if (err) + goto out_unlock; + + end_time = jiffies + i915_selftest.timeout_jiffies; + for_each_prime_number_from(prime, 2, 8192) { + times[1] = ktime_get_raw(); + + for (n = 0; n < prime; n++) { + rq = i915_request_alloc(engine, ctx[n % nctx]); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + goto out_unlock; + } + + /* + * This space is left intentionally blank. + * + * We do not actually want to perform any + * action with this request, we just want + * to measure the latency in allocation + * and submission of our breadcrumbs - + * ensuring that the bare request is sufficient + * for the system to work (i.e. proper HEAD + * tracking of the rings, interrupt handling, + * etc). It also gives us the lowest bounds + * for latency. + */ + + i915_request_add(rq); + } + if (i915_request_wait(rq, + I915_WAIT_LOCKED, + HZ / 5) < 0) { + pr_err("Switching between %ld contexts timed out\n", + prime); + i915_gem_set_wedged(i915); + break; + } + + times[1] = ktime_sub(ktime_get_raw(), times[1]); + if (prime == 2) + times[0] = times[1]; + + if (__igt_timeout(end_time, NULL)) + break; + } + + err = end_live_test(&t); + if (err) + goto out_unlock; + + pr_info("Switch latencies on %s: 1 = %lluns, %lu = %lluns\n", + engine->name, + ktime_to_ns(times[0]), + prime - 1, div64_u64(ktime_to_ns(times[1]), prime - 1)); + } + +out_unlock: + intel_runtime_pm_put(i915); + mutex_unlock(&i915->drm.struct_mutex); + mock_file_free(i915, file); + return err; +} + static struct i915_vma * gpu_fill_dw(struct i915_vma *vma, u64 offset, unsigned long count, u32 value) { @@ -195,6 +391,7 @@ err_request: i915_request_add(rq); err_batch: i915_vma_unpin(batch); + i915_vma_put(batch); err_vma: i915_vma_unpin(vma); return err; @@ -636,6 +833,8 @@ static int igt_switch_to_kernel_context(void *arg) */ mutex_lock(&i915->drm.struct_mutex); + intel_runtime_pm_get(i915); + ctx = kernel_context(i915); if (IS_ERR(ctx)) { mutex_unlock(&i915->drm.struct_mutex); @@ -658,6 +857,8 @@ out_unlock: GEM_TRACE_DUMP_ON(err); if (igt_flush_test(i915, I915_WAIT_LOCKED)) err = -EIO; + + intel_runtime_pm_put(i915); mutex_unlock(&i915->drm.struct_mutex); kernel_context_close(ctx); @@ -713,6 +914,7 @@ int i915_gem_context_live_selftests(struct drm_i915_private *dev_priv) { static const struct i915_subtest tests[] = { SUBTEST(igt_switch_to_kernel_context), + SUBTEST(live_nop_switch), SUBTEST(igt_ctx_exec), SUBTEST(igt_ctx_readonly), }; diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_object.c b/drivers/gpu/drm/i915/selftests/i915_gem_object.c index 6d3516d5bff9..c3999dd2021e 100644 --- a/drivers/gpu/drm/i915/selftests/i915_gem_object.c +++ b/drivers/gpu/drm/i915/selftests/i915_gem_object.c @@ -501,6 +501,8 @@ static bool assert_mmap_offset(struct drm_i915_private *i915, static void disable_retire_worker(struct drm_i915_private *i915) { + i915_gem_shrinker_unregister(i915); + mutex_lock(&i915->drm.struct_mutex); if (!i915->gt.active_requests++) { intel_runtime_pm_get(i915); @@ -613,6 +615,7 @@ out_park: else queue_delayed_work(i915->wq, &i915->gt.idle_work, 0); mutex_unlock(&i915->drm.struct_mutex); + i915_gem_shrinker_register(i915); return err; err_obj: i915_gem_object_put(obj); diff --git a/drivers/gpu/drm/i915/selftests/i915_request.c b/drivers/gpu/drm/i915/selftests/i915_request.c index c4aac6141e04..07e557815308 100644 --- a/drivers/gpu/drm/i915/selftests/i915_request.c +++ b/drivers/gpu/drm/i915/selftests/i915_request.c @@ -342,6 +342,7 @@ static int live_nop_request(void *arg) */ mutex_lock(&i915->drm.struct_mutex); + intel_runtime_pm_get(i915); for_each_engine(engine, i915, id) { struct i915_request *request = NULL; @@ -402,6 +403,7 @@ static int live_nop_request(void *arg) } out_unlock: + intel_runtime_pm_put(i915); mutex_unlock(&i915->drm.struct_mutex); return err; } @@ -487,6 +489,7 @@ static int live_empty_request(void *arg) */ mutex_lock(&i915->drm.struct_mutex); + intel_runtime_pm_get(i915); batch = empty_batch(i915); if (IS_ERR(batch)) { @@ -550,6 +553,7 @@ out_batch: i915_vma_unpin(batch); i915_vma_put(batch); out_unlock: + intel_runtime_pm_put(i915); mutex_unlock(&i915->drm.struct_mutex); return err; } @@ -644,6 +648,7 @@ static int live_all_engines(void *arg) */ mutex_lock(&i915->drm.struct_mutex); + intel_runtime_pm_get(i915); err = begin_live_test(&t, i915, __func__, ""); if (err) @@ -726,6 +731,7 @@ out_request: i915_vma_unpin(batch); i915_vma_put(batch); out_unlock: + intel_runtime_pm_put(i915); mutex_unlock(&i915->drm.struct_mutex); return err; } @@ -747,6 +753,7 @@ static int live_sequential_engines(void *arg) */ mutex_lock(&i915->drm.struct_mutex); + intel_runtime_pm_get(i915); err = begin_live_test(&t, i915, __func__, ""); if (err) @@ -853,6 +860,7 @@ out_request: i915_request_put(request[id]); } out_unlock: + intel_runtime_pm_put(i915); mutex_unlock(&i915->drm.struct_mutex); return err; } diff --git a/drivers/gpu/drm/i915/selftests/intel_guc.c b/drivers/gpu/drm/i915/selftests/intel_guc.c index 90ba88c972cf..0c0ab82b6228 100644 --- a/drivers/gpu/drm/i915/selftests/intel_guc.c +++ b/drivers/gpu/drm/i915/selftests/intel_guc.c @@ -142,6 +142,7 @@ static int igt_guc_clients(void *args) GEM_BUG_ON(!HAS_GUC(dev_priv)); mutex_lock(&dev_priv->drm.struct_mutex); + intel_runtime_pm_get(dev_priv); guc = &dev_priv->guc; if (!guc) { @@ -269,6 +270,7 @@ out: guc_clients_create(guc); guc_clients_doorbell_init(guc); unlock: + intel_runtime_pm_put(dev_priv); mutex_unlock(&dev_priv->drm.struct_mutex); return err; } @@ -287,6 +289,7 @@ static int igt_guc_doorbells(void *arg) GEM_BUG_ON(!HAS_GUC(dev_priv)); mutex_lock(&dev_priv->drm.struct_mutex); + intel_runtime_pm_get(dev_priv); guc = &dev_priv->guc; if (!guc) { @@ -379,6 +382,7 @@ out: guc_client_free(clients[i]); } unlock: + intel_runtime_pm_put(dev_priv); mutex_unlock(&dev_priv->drm.struct_mutex); return err; } diff --git a/drivers/gpu/drm/i915/selftests/intel_lrc.c b/drivers/gpu/drm/i915/selftests/intel_lrc.c index 582566faef09..1aea7a8f2224 100644 --- a/drivers/gpu/drm/i915/selftests/intel_lrc.c +++ b/drivers/gpu/drm/i915/selftests/intel_lrc.c @@ -221,6 +221,7 @@ static int live_sanitycheck(void *arg) return 0; mutex_lock(&i915->drm.struct_mutex); + intel_runtime_pm_get(i915); if (spinner_init(&spin, i915)) goto err_unlock; @@ -261,6 +262,7 @@ err_spin: spinner_fini(&spin); err_unlock: igt_flush_test(i915, I915_WAIT_LOCKED); + intel_runtime_pm_put(i915); mutex_unlock(&i915->drm.struct_mutex); return err; } @@ -278,6 +280,7 @@ static int live_preempt(void *arg) return 0; mutex_lock(&i915->drm.struct_mutex); + intel_runtime_pm_get(i915); if (spinner_init(&spin_hi, i915)) goto err_unlock; @@ -350,6 +353,7 @@ err_spin_hi: spinner_fini(&spin_hi); err_unlock: igt_flush_test(i915, I915_WAIT_LOCKED); + intel_runtime_pm_put(i915); mutex_unlock(&i915->drm.struct_mutex); return err; } @@ -368,6 +372,7 @@ static int live_late_preempt(void *arg) return 0; mutex_lock(&i915->drm.struct_mutex); + intel_runtime_pm_get(i915); if (spinner_init(&spin_hi, i915)) goto err_unlock; @@ -440,6 +445,7 @@ err_spin_hi: spinner_fini(&spin_hi); err_unlock: igt_flush_test(i915, I915_WAIT_LOCKED); + intel_runtime_pm_put(i915); mutex_unlock(&i915->drm.struct_mutex); return err; @@ -467,6 +473,7 @@ static int live_preempt_hang(void *arg) return 0; mutex_lock(&i915->drm.struct_mutex); + intel_runtime_pm_get(i915); if (spinner_init(&spin_hi, i915)) goto err_unlock; @@ -561,6 +568,7 @@ err_spin_hi: spinner_fini(&spin_hi); err_unlock: igt_flush_test(i915, I915_WAIT_LOCKED); + intel_runtime_pm_put(i915); mutex_unlock(&i915->drm.struct_mutex); return err; } diff --git a/drivers/gpu/drm/i915/selftests/intel_workarounds.c b/drivers/gpu/drm/i915/selftests/intel_workarounds.c index 0d39b3bf0c0d..d1a0923d2f38 100644 --- a/drivers/gpu/drm/i915/selftests/intel_workarounds.c +++ b/drivers/gpu/drm/i915/selftests/intel_workarounds.c @@ -44,7 +44,9 @@ read_nonprivs(struct i915_gem_context *ctx, struct intel_engine_cs *engine) if (err) goto err_obj; + intel_runtime_pm_get(engine->i915); rq = i915_request_alloc(engine, ctx); + intel_runtime_pm_put(engine->i915); if (IS_ERR(rq)) { err = PTR_ERR(rq); goto err_pin; @@ -175,7 +177,10 @@ static int switch_to_scratch_context(struct intel_engine_cs *engine) if (IS_ERR(ctx)) return PTR_ERR(ctx); + intel_runtime_pm_get(engine->i915); rq = i915_request_alloc(engine, ctx); + intel_runtime_pm_put(engine->i915); + kernel_context_close(ctx); if (IS_ERR(rq)) return PTR_ERR(rq); diff --git a/drivers/gpu/drm/mediatek/Makefile b/drivers/gpu/drm/mediatek/Makefile index ce83c396a742..82ae49c64221 100644 --- a/drivers/gpu/drm/mediatek/Makefile +++ b/drivers/gpu/drm/mediatek/Makefile @@ -1,4 +1,5 @@ # SPDX-License-Identifier: GPL-2.0 + mediatek-drm-y := mtk_disp_color.o \ mtk_disp_ovl.o \ mtk_disp_rdma.o \ @@ -18,6 +19,8 @@ obj-$(CONFIG_DRM_MEDIATEK) += mediatek-drm.o mediatek-drm-hdmi-objs := mtk_cec.o \ mtk_hdmi.o \ mtk_hdmi_ddc.o \ - mtk_mt8173_hdmi_phy.o + mtk_mt2701_hdmi_phy.o \ + mtk_mt8173_hdmi_phy.o \ + mtk_hdmi_phy.o obj-$(CONFIG_DRM_MEDIATEK_HDMI) += mediatek-drm-hdmi.o diff --git a/drivers/gpu/drm/mediatek/mtk_dpi.c b/drivers/gpu/drm/mediatek/mtk_dpi.c index 6c0ea39d5739..62a9d47df948 100644 --- a/drivers/gpu/drm/mediatek/mtk_dpi.c +++ b/drivers/gpu/drm/mediatek/mtk_dpi.c @@ -14,10 +14,12 @@ #include <drm/drmP.h> #include <drm/drm_crtc.h> #include <drm/drm_crtc_helper.h> +#include <drm/drm_of.h> #include <linux/kernel.h> #include <linux/component.h> #include <linux/platform_device.h> #include <linux/of.h> +#include <linux/of_device.h> #include <linux/of_graph.h> #include <linux/interrupt.h> #include <linux/types.h> @@ -72,12 +74,12 @@ struct mtk_dpi { struct clk *tvd_clk; int irq; struct drm_display_mode mode; + const struct mtk_dpi_conf *conf; enum mtk_dpi_out_color_format color_format; enum mtk_dpi_out_yc_map yc_map; enum mtk_dpi_out_bit_num bit_num; enum mtk_dpi_out_channel_swap channel_swap; - bool power_sta; - u8 power_ctl; + int refcount; }; static inline struct mtk_dpi *mtk_dpi_from_encoder(struct drm_encoder *e) @@ -90,11 +92,6 @@ enum mtk_dpi_polarity { MTK_DPI_POLARITY_FALLING, }; -enum mtk_dpi_power_ctl { - DPI_POWER_START = BIT(0), - DPI_POWER_ENABLE = BIT(1), -}; - struct mtk_dpi_polarities { enum mtk_dpi_polarity de_pol; enum mtk_dpi_polarity ck_pol; @@ -116,6 +113,12 @@ struct mtk_dpi_yc_limit { u16 c_bottom; }; +struct mtk_dpi_conf { + unsigned int (*cal_factor)(int clock); + u32 reg_h_fre_con; + bool edge_sel_en; +}; + static void mtk_dpi_mask(struct mtk_dpi *dpi, u32 offset, u32 val, u32 mask) { u32 tmp = readl(dpi->regs + offset) & ~mask; @@ -341,7 +344,13 @@ static void mtk_dpi_config_swap_input(struct mtk_dpi *dpi, bool enable) static void mtk_dpi_config_2n_h_fre(struct mtk_dpi *dpi) { - mtk_dpi_mask(dpi, DPI_H_FRE_CON, H_FRE_2N, H_FRE_2N); + mtk_dpi_mask(dpi, dpi->conf->reg_h_fre_con, H_FRE_2N, H_FRE_2N); +} + +static void mtk_dpi_config_disable_edge(struct mtk_dpi *dpi) +{ + if (dpi->conf->edge_sel_en) + mtk_dpi_mask(dpi, dpi->conf->reg_h_fre_con, 0, EDGE_SEL_EN); } static void mtk_dpi_config_color_format(struct mtk_dpi *dpi, @@ -367,40 +376,30 @@ static void mtk_dpi_config_color_format(struct mtk_dpi *dpi, } } -static void mtk_dpi_power_off(struct mtk_dpi *dpi, enum mtk_dpi_power_ctl pctl) +static void mtk_dpi_power_off(struct mtk_dpi *dpi) { - dpi->power_ctl &= ~pctl; - - if ((dpi->power_ctl & DPI_POWER_START) || - (dpi->power_ctl & DPI_POWER_ENABLE)) + if (WARN_ON(dpi->refcount == 0)) return; - if (!dpi->power_sta) + if (--dpi->refcount != 0) return; mtk_dpi_disable(dpi); clk_disable_unprepare(dpi->pixel_clk); clk_disable_unprepare(dpi->engine_clk); - dpi->power_sta = false; } -static int mtk_dpi_power_on(struct mtk_dpi *dpi, enum mtk_dpi_power_ctl pctl) +static int mtk_dpi_power_on(struct mtk_dpi *dpi) { int ret; - dpi->power_ctl |= pctl; - - if (!(dpi->power_ctl & DPI_POWER_START) && - !(dpi->power_ctl & DPI_POWER_ENABLE)) - return 0; - - if (dpi->power_sta) + if (++dpi->refcount != 1) return 0; ret = clk_prepare_enable(dpi->engine_clk); if (ret) { dev_err(dpi->dev, "Failed to enable engine clock: %d\n", ret); - goto err_eng; + goto err_refcount; } ret = clk_prepare_enable(dpi->pixel_clk); @@ -410,13 +409,12 @@ static int mtk_dpi_power_on(struct mtk_dpi *dpi, enum mtk_dpi_power_ctl pctl) } mtk_dpi_enable(dpi); - dpi->power_sta = true; return 0; err_pixel: clk_disable_unprepare(dpi->engine_clk); -err_eng: - dpi->power_ctl &= ~pctl; +err_refcount: + dpi->refcount--; return ret; } @@ -435,15 +433,7 @@ static int mtk_dpi_set_display_mode(struct mtk_dpi *dpi, unsigned int factor; /* let pll_rate can fix the valid range of tvdpll (1G~2GHz) */ - - if (mode->clock <= 27000) - factor = 3 << 4; - else if (mode->clock <= 84000) - factor = 3 << 3; - else if (mode->clock <= 167000) - factor = 3 << 2; - else - factor = 3 << 1; + factor = dpi->conf->cal_factor(mode->clock); drm_display_mode_to_videomode(mode, &vm); pll_rate = vm.pixelclock * factor; @@ -518,6 +508,7 @@ static int mtk_dpi_set_display_mode(struct mtk_dpi *dpi, mtk_dpi_config_yc_map(dpi, dpi->yc_map); mtk_dpi_config_color_format(dpi, dpi->color_format); mtk_dpi_config_2n_h_fre(dpi); + mtk_dpi_config_disable_edge(dpi); mtk_dpi_sw_reset(dpi, false); return 0; @@ -552,14 +543,14 @@ static void mtk_dpi_encoder_disable(struct drm_encoder *encoder) { struct mtk_dpi *dpi = mtk_dpi_from_encoder(encoder); - mtk_dpi_power_off(dpi, DPI_POWER_ENABLE); + mtk_dpi_power_off(dpi); } static void mtk_dpi_encoder_enable(struct drm_encoder *encoder) { struct mtk_dpi *dpi = mtk_dpi_from_encoder(encoder); - mtk_dpi_power_on(dpi, DPI_POWER_ENABLE); + mtk_dpi_power_on(dpi); mtk_dpi_set_display_mode(dpi, &dpi->mode); } @@ -582,14 +573,14 @@ static void mtk_dpi_start(struct mtk_ddp_comp *comp) { struct mtk_dpi *dpi = container_of(comp, struct mtk_dpi, ddp_comp); - mtk_dpi_power_on(dpi, DPI_POWER_START); + mtk_dpi_power_on(dpi); } static void mtk_dpi_stop(struct mtk_ddp_comp *comp) { struct mtk_dpi *dpi = container_of(comp, struct mtk_dpi, ddp_comp); - mtk_dpi_power_off(dpi, DPI_POWER_START); + mtk_dpi_power_off(dpi); } static const struct mtk_ddp_comp_funcs mtk_dpi_funcs = { @@ -656,12 +647,46 @@ static const struct component_ops mtk_dpi_component_ops = { .unbind = mtk_dpi_unbind, }; +static unsigned int mt8173_calculate_factor(int clock) +{ + if (clock <= 27000) + return 3 << 4; + else if (clock <= 84000) + return 3 << 3; + else if (clock <= 167000) + return 3 << 2; + else + return 3 << 1; +} + +static unsigned int mt2701_calculate_factor(int clock) +{ + if (clock <= 64000) + return 16; + else if (clock <= 128000) + return 8; + else if (clock <= 256000) + return 4; + else + return 2; +} + +static const struct mtk_dpi_conf mt8173_conf = { + .cal_factor = mt8173_calculate_factor, + .reg_h_fre_con = 0xe0, +}; + +static const struct mtk_dpi_conf mt2701_conf = { + .cal_factor = mt2701_calculate_factor, + .reg_h_fre_con = 0xb0, + .edge_sel_en = true, +}; + static int mtk_dpi_probe(struct platform_device *pdev) { struct device *dev = &pdev->dev; struct mtk_dpi *dpi; struct resource *mem; - struct device_node *bridge_node; int comp_id; int ret; @@ -670,6 +695,7 @@ static int mtk_dpi_probe(struct platform_device *pdev) return -ENOMEM; dpi->dev = dev; + dpi->conf = (struct mtk_dpi_conf *)of_device_get_match_data(dev); mem = platform_get_resource(pdev, IORESOURCE_MEM, 0); dpi->regs = devm_ioremap_resource(dev, mem); @@ -706,16 +732,12 @@ static int mtk_dpi_probe(struct platform_device *pdev) return -EINVAL; } - bridge_node = of_graph_get_remote_node(dev->of_node, 0, 0); - if (!bridge_node) - return -ENODEV; - - dev_info(dev, "Found bridge node: %pOF\n", bridge_node); + ret = drm_of_find_panel_or_bridge(dev->of_node, 0, 0, + NULL, &dpi->bridge); + if (ret) + return ret; - dpi->bridge = of_drm_find_bridge(bridge_node); - of_node_put(bridge_node); - if (!dpi->bridge) - return -EPROBE_DEFER; + dev_info(dev, "Found bridge node: %pOF\n", dpi->bridge->of_node); comp_id = mtk_ddp_comp_get_id(dev->of_node, MTK_DPI); if (comp_id < 0) { @@ -749,8 +771,13 @@ static int mtk_dpi_remove(struct platform_device *pdev) } static const struct of_device_id mtk_dpi_of_ids[] = { - { .compatible = "mediatek,mt8173-dpi", }, - {} + { .compatible = "mediatek,mt2701-dpi", + .data = &mt2701_conf, + }, + { .compatible = "mediatek,mt8173-dpi", + .data = &mt8173_conf, + }, + { }, }; struct platform_driver mtk_dpi_driver = { diff --git a/drivers/gpu/drm/mediatek/mtk_dpi_regs.h b/drivers/gpu/drm/mediatek/mtk_dpi_regs.h index 4b6ad4751a31..d9db8c4cacd7 100644 --- a/drivers/gpu/drm/mediatek/mtk_dpi_regs.h +++ b/drivers/gpu/drm/mediatek/mtk_dpi_regs.h @@ -223,6 +223,6 @@ #define ESAV_CODE2 (0xFFF << 0) #define ESAV_CODE3_MSB BIT(16) -#define DPI_H_FRE_CON 0xE0 +#define EDGE_SEL_EN BIT(5) #define H_FRE_2N BIT(25) #endif /* __MTK_DPI_REGS_H */ diff --git a/drivers/gpu/drm/mediatek/mtk_drm_ddp.c b/drivers/gpu/drm/mediatek/mtk_drm_ddp.c index 546b3e3b300b..579ce28d801d 100644 --- a/drivers/gpu/drm/mediatek/mtk_drm_ddp.c +++ b/drivers/gpu/drm/mediatek/mtk_drm_ddp.c @@ -39,6 +39,7 @@ #define DISP_REG_CONFIG_DISP_OVL_MOUT_EN 0x030 #define DISP_REG_CONFIG_OUT_SEL 0x04c #define DISP_REG_CONFIG_DSI_SEL 0x050 +#define DISP_REG_CONFIG_DPI_SEL 0x064 #define DISP_REG_MUTEX_EN(n) (0x20 + 0x20 * (n)) #define DISP_REG_MUTEX(n) (0x24 + 0x20 * (n)) @@ -136,7 +137,10 @@ #define OVL_MOUT_EN_RDMA 0x1 #define BLS_TO_DSI_RDMA1_TO_DPI1 0x8 +#define BLS_TO_DPI_RDMA1_TO_DSI 0x2 #define DSI_SEL_IN_BLS 0x0 +#define DPI_SEL_IN_BLS 0x0 +#define DSI_SEL_IN_RDMA 0x1 struct mtk_disp_mutex { int id; @@ -339,9 +343,17 @@ static void mtk_ddp_sout_sel(void __iomem *config_regs, enum mtk_ddp_comp_id cur, enum mtk_ddp_comp_id next) { - if (cur == DDP_COMPONENT_BLS && next == DDP_COMPONENT_DSI0) + if (cur == DDP_COMPONENT_BLS && next == DDP_COMPONENT_DSI0) { writel_relaxed(BLS_TO_DSI_RDMA1_TO_DPI1, config_regs + DISP_REG_CONFIG_OUT_SEL); + } else if (cur == DDP_COMPONENT_BLS && next == DDP_COMPONENT_DPI0) { + writel_relaxed(BLS_TO_DPI_RDMA1_TO_DSI, + config_regs + DISP_REG_CONFIG_OUT_SEL); + writel_relaxed(DSI_SEL_IN_RDMA, + config_regs + DISP_REG_CONFIG_DSI_SEL); + writel_relaxed(DPI_SEL_IN_BLS, + config_regs + DISP_REG_CONFIG_DPI_SEL); + } } void mtk_ddp_add_comp_to_path(void __iomem *config_regs, diff --git a/drivers/gpu/drm/mediatek/mtk_drm_ddp_comp.c b/drivers/gpu/drm/mediatek/mtk_drm_ddp_comp.c index ff974d82a4a6..54ca794db3e9 100644 --- a/drivers/gpu/drm/mediatek/mtk_drm_ddp_comp.c +++ b/drivers/gpu/drm/mediatek/mtk_drm_ddp_comp.c @@ -294,7 +294,7 @@ int mtk_ddp_comp_init(struct device *dev, struct device_node *node, comp->irq = of_irq_get(node, 0); comp->clk = of_clk_get(node, 0); if (IS_ERR(comp->clk)) - comp->clk = NULL; + return PTR_ERR(comp->clk); /* Only DMA capable components need the LARB property */ comp->larb_dev = NULL; diff --git a/drivers/gpu/drm/mediatek/mtk_drm_drv.c b/drivers/gpu/drm/mediatek/mtk_drm_drv.c index 47ec604289b7..6422e99952fe 100644 --- a/drivers/gpu/drm/mediatek/mtk_drm_drv.c +++ b/drivers/gpu/drm/mediatek/mtk_drm_drv.c @@ -424,6 +424,8 @@ static const struct of_device_id mtk_ddp_comp_dt_ids[] = { .data = (void *)MTK_DSI }, { .compatible = "mediatek,mt8173-dsi", .data = (void *)MTK_DSI }, + { .compatible = "mediatek,mt2701-dpi", + .data = (void *)MTK_DPI }, { .compatible = "mediatek,mt8173-dpi", .data = (void *)MTK_DPI }, { .compatible = "mediatek,mt2701-disp-mutex", diff --git a/drivers/gpu/drm/mediatek/mtk_hdmi.c b/drivers/gpu/drm/mediatek/mtk_hdmi.c index 2d45d1dd9554..11e3644da79a 100644 --- a/drivers/gpu/drm/mediatek/mtk_hdmi.c +++ b/drivers/gpu/drm/mediatek/mtk_hdmi.c @@ -233,6 +233,7 @@ static void mtk_hdmi_hw_vid_black(struct mtk_hdmi *hdmi, bool black) static void mtk_hdmi_hw_make_reg_writable(struct mtk_hdmi *hdmi, bool enable) { struct arm_smccc_res res; + struct mtk_hdmi_phy *hdmi_phy = phy_get_drvdata(hdmi->phy); /* * MT8173 HDMI hardware has an output control bit to enable/disable HDMI @@ -240,8 +241,13 @@ static void mtk_hdmi_hw_make_reg_writable(struct mtk_hdmi *hdmi, bool enable) * The ARM trusted firmware provides an API for the HDMI driver to set * this control bit to enable HDMI output in supervisor mode. */ - arm_smccc_smc(MTK_SIP_SET_AUTHORIZED_SECURE_REG, 0x14000904, 0x80000000, - 0, 0, 0, 0, 0, &res); + if (hdmi_phy->conf && hdmi_phy->conf->tz_disabled) + regmap_update_bits(hdmi->sys_regmap, + hdmi->sys_offset + HDMI_SYS_CFG20, + 0x80008005, enable ? 0x80000005 : 0x8000); + else + arm_smccc_smc(MTK_SIP_SET_AUTHORIZED_SECURE_REG, 0x14000904, + 0x80000000, 0, 0, 0, 0, 0, &res); regmap_update_bits(hdmi->sys_regmap, hdmi->sys_offset + HDMI_SYS_CFG20, HDMI_PCLK_FREE_RUN, enable ? HDMI_PCLK_FREE_RUN : 0); @@ -1575,6 +1581,11 @@ static int mtk_hdmi_audio_hw_params(struct device *dev, void *data, hdmi_params.aud_i2s_fmt = HDMI_I2S_MODE_I2S_24BIT; hdmi_params.aud_mclk = HDMI_AUD_MCLK_128FS; break; + case HDMI_SPDIF: + hdmi_params.aud_codec = HDMI_AUDIO_CODING_TYPE_PCM; + hdmi_params.aud_sampe_size = HDMI_AUDIO_SAMPLE_SIZE_16; + hdmi_params.aud_input_type = HDMI_AUD_INPUT_SPDIF; + break; default: dev_err(hdmi->dev, "%s: Invalid DAI format %d\n", __func__, daifmt->fmt); diff --git a/drivers/gpu/drm/mediatek/mtk_hdmi.h b/drivers/gpu/drm/mediatek/mtk_hdmi.h index 6371b3de1ff6..3e9fb8d19802 100644 --- a/drivers/gpu/drm/mediatek/mtk_hdmi.h +++ b/drivers/gpu/drm/mediatek/mtk_hdmi.h @@ -13,11 +13,11 @@ */ #ifndef _MTK_HDMI_CTRL_H #define _MTK_HDMI_CTRL_H +#include "mtk_hdmi_phy.h" struct platform_driver; extern struct platform_driver mtk_cec_driver; extern struct platform_driver mtk_hdmi_ddc_driver; -extern struct platform_driver mtk_hdmi_phy_driver; #endif /* _MTK_HDMI_CTRL_H */ diff --git a/drivers/gpu/drm/mediatek/mtk_hdmi_phy.c b/drivers/gpu/drm/mediatek/mtk_hdmi_phy.c new file mode 100644 index 000000000000..4ef9c57ffd44 --- /dev/null +++ b/drivers/gpu/drm/mediatek/mtk_hdmi_phy.c @@ -0,0 +1,235 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2018 MediaTek Inc. + * Author: Jie Qiu <jie.qiu@mediatek.com> + */ + +#include "mtk_hdmi_phy.h" + +static int mtk_hdmi_phy_power_on(struct phy *phy); +static int mtk_hdmi_phy_power_off(struct phy *phy); + +static const struct phy_ops mtk_hdmi_phy_dev_ops = { + .power_on = mtk_hdmi_phy_power_on, + .power_off = mtk_hdmi_phy_power_off, + .owner = THIS_MODULE, +}; + +long mtk_hdmi_pll_round_rate(struct clk_hw *hw, unsigned long rate, + unsigned long *parent_rate) +{ + struct mtk_hdmi_phy *hdmi_phy = to_mtk_hdmi_phy(hw); + + hdmi_phy->pll_rate = rate; + if (rate <= 74250000) + *parent_rate = rate; + else + *parent_rate = rate / 2; + + return rate; +} + +unsigned long mtk_hdmi_pll_recalc_rate(struct clk_hw *hw, + unsigned long parent_rate) +{ + struct mtk_hdmi_phy *hdmi_phy = to_mtk_hdmi_phy(hw); + + return hdmi_phy->pll_rate; +} + +void mtk_hdmi_phy_clear_bits(struct mtk_hdmi_phy *hdmi_phy, u32 offset, + u32 bits) +{ + void __iomem *reg = hdmi_phy->regs + offset; + u32 tmp; + + tmp = readl(reg); + tmp &= ~bits; + writel(tmp, reg); +} + +void mtk_hdmi_phy_set_bits(struct mtk_hdmi_phy *hdmi_phy, u32 offset, + u32 bits) +{ + void __iomem *reg = hdmi_phy->regs + offset; + u32 tmp; + + tmp = readl(reg); + tmp |= bits; + writel(tmp, reg); +} + +void mtk_hdmi_phy_mask(struct mtk_hdmi_phy *hdmi_phy, u32 offset, + u32 val, u32 mask) +{ + void __iomem *reg = hdmi_phy->regs + offset; + u32 tmp; + + tmp = readl(reg); + tmp = (tmp & ~mask) | (val & mask); + writel(tmp, reg); +} + +inline struct mtk_hdmi_phy *to_mtk_hdmi_phy(struct clk_hw *hw) +{ + return container_of(hw, struct mtk_hdmi_phy, pll_hw); +} + +static int mtk_hdmi_phy_power_on(struct phy *phy) +{ + struct mtk_hdmi_phy *hdmi_phy = phy_get_drvdata(phy); + int ret; + + ret = clk_prepare_enable(hdmi_phy->pll); + if (ret < 0) + return ret; + + hdmi_phy->conf->hdmi_phy_enable_tmds(hdmi_phy); + return 0; +} + +static int mtk_hdmi_phy_power_off(struct phy *phy) +{ + struct mtk_hdmi_phy *hdmi_phy = phy_get_drvdata(phy); + + hdmi_phy->conf->hdmi_phy_disable_tmds(hdmi_phy); + clk_disable_unprepare(hdmi_phy->pll); + + return 0; +} + +static const struct phy_ops * +mtk_hdmi_phy_dev_get_ops(const struct mtk_hdmi_phy *hdmi_phy) +{ + if (hdmi_phy && hdmi_phy->conf && + hdmi_phy->conf->hdmi_phy_enable_tmds && + hdmi_phy->conf->hdmi_phy_disable_tmds) + return &mtk_hdmi_phy_dev_ops; + + dev_err(hdmi_phy->dev, "Failed to get dev ops of phy\n"); + return NULL; +} + +static void mtk_hdmi_phy_clk_get_ops(struct mtk_hdmi_phy *hdmi_phy, + const struct clk_ops **ops) +{ + if (hdmi_phy && hdmi_phy->conf && hdmi_phy->conf->hdmi_phy_clk_ops) + *ops = hdmi_phy->conf->hdmi_phy_clk_ops; + else + dev_err(hdmi_phy->dev, "Failed to get clk ops of phy\n"); +} + +static int mtk_hdmi_phy_probe(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct mtk_hdmi_phy *hdmi_phy; + struct resource *mem; + struct clk *ref_clk; + const char *ref_clk_name; + struct clk_init_data clk_init = { + .num_parents = 1, + .parent_names = (const char * const *)&ref_clk_name, + .flags = CLK_SET_RATE_PARENT | CLK_SET_RATE_GATE, + }; + + struct phy *phy; + struct phy_provider *phy_provider; + int ret; + + hdmi_phy = devm_kzalloc(dev, sizeof(*hdmi_phy), GFP_KERNEL); + if (!hdmi_phy) + return -ENOMEM; + + mem = platform_get_resource(pdev, IORESOURCE_MEM, 0); + hdmi_phy->regs = devm_ioremap_resource(dev, mem); + if (IS_ERR(hdmi_phy->regs)) { + ret = PTR_ERR(hdmi_phy->regs); + dev_err(dev, "Failed to get memory resource: %d\n", ret); + return ret; + } + + ref_clk = devm_clk_get(dev, "pll_ref"); + if (IS_ERR(ref_clk)) { + ret = PTR_ERR(ref_clk); + dev_err(&pdev->dev, "Failed to get PLL reference clock: %d\n", + ret); + return ret; + } + ref_clk_name = __clk_get_name(ref_clk); + + ret = of_property_read_string(dev->of_node, "clock-output-names", + &clk_init.name); + if (ret < 0) { + dev_err(dev, "Failed to read clock-output-names: %d\n", ret); + return ret; + } + + hdmi_phy->dev = dev; + hdmi_phy->conf = + (struct mtk_hdmi_phy_conf *)of_device_get_match_data(dev); + mtk_hdmi_phy_clk_get_ops(hdmi_phy, &clk_init.ops); + hdmi_phy->pll_hw.init = &clk_init; + hdmi_phy->pll = devm_clk_register(dev, &hdmi_phy->pll_hw); + if (IS_ERR(hdmi_phy->pll)) { + ret = PTR_ERR(hdmi_phy->pll); + dev_err(dev, "Failed to register PLL: %d\n", ret); + return ret; + } + + ret = of_property_read_u32(dev->of_node, "mediatek,ibias", + &hdmi_phy->ibias); + if (ret < 0) { + dev_err(&pdev->dev, "Failed to get ibias: %d\n", ret); + return ret; + } + + ret = of_property_read_u32(dev->of_node, "mediatek,ibias_up", + &hdmi_phy->ibias_up); + if (ret < 0) { + dev_err(&pdev->dev, "Failed to get ibias up: %d\n", ret); + return ret; + } + + dev_info(dev, "Using default TX DRV impedance: 4.2k/36\n"); + hdmi_phy->drv_imp_clk = 0x30; + hdmi_phy->drv_imp_d2 = 0x30; + hdmi_phy->drv_imp_d1 = 0x30; + hdmi_phy->drv_imp_d0 = 0x30; + + phy = devm_phy_create(dev, NULL, mtk_hdmi_phy_dev_get_ops(hdmi_phy)); + if (IS_ERR(phy)) { + dev_err(dev, "Failed to create HDMI PHY\n"); + return PTR_ERR(phy); + } + phy_set_drvdata(phy, hdmi_phy); + + phy_provider = devm_of_phy_provider_register(dev, of_phy_simple_xlate); + if (IS_ERR(phy_provider)) { + dev_err(dev, "Failed to register HDMI PHY\n"); + return PTR_ERR(phy_provider); + } + + return of_clk_add_provider(dev->of_node, of_clk_src_simple_get, + hdmi_phy->pll); +} + +static const struct of_device_id mtk_hdmi_phy_match[] = { + { .compatible = "mediatek,mt2701-hdmi-phy", + .data = &mtk_hdmi_phy_2701_conf, + }, + { .compatible = "mediatek,mt8173-hdmi-phy", + .data = &mtk_hdmi_phy_8173_conf, + }, + {}, +}; + +struct platform_driver mtk_hdmi_phy_driver = { + .probe = mtk_hdmi_phy_probe, + .driver = { + .name = "mediatek-hdmi-phy", + .of_match_table = mtk_hdmi_phy_match, + }, +}; + +MODULE_DESCRIPTION("MediaTek HDMI PHY Driver"); +MODULE_LICENSE("GPL v2"); diff --git a/drivers/gpu/drm/mediatek/mtk_hdmi_phy.h b/drivers/gpu/drm/mediatek/mtk_hdmi_phy.h new file mode 100644 index 000000000000..f39b1fc66612 --- /dev/null +++ b/drivers/gpu/drm/mediatek/mtk_hdmi_phy.h @@ -0,0 +1,60 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2018 MediaTek Inc. + * Author: Chunhui Dai <chunhui.dai@mediatek.com> + */ + +#ifndef _MTK_HDMI_PHY_H +#define _MTK_HDMI_PHY_H +#include <linux/clk.h> +#include <linux/clk-provider.h> +#include <linux/delay.h> +#include <linux/io.h> +#include <linux/mfd/syscon.h> +#include <linux/module.h> +#include <linux/of_device.h> +#include <linux/phy/phy.h> +#include <linux/platform_device.h> +#include <linux/types.h> + +struct mtk_hdmi_phy; + +struct mtk_hdmi_phy_conf { + bool tz_disabled; + const struct clk_ops *hdmi_phy_clk_ops; + void (*hdmi_phy_enable_tmds)(struct mtk_hdmi_phy *hdmi_phy); + void (*hdmi_phy_disable_tmds)(struct mtk_hdmi_phy *hdmi_phy); +}; + +struct mtk_hdmi_phy { + void __iomem *regs; + struct device *dev; + struct mtk_hdmi_phy_conf *conf; + struct clk *pll; + struct clk_hw pll_hw; + unsigned long pll_rate; + unsigned char drv_imp_clk; + unsigned char drv_imp_d2; + unsigned char drv_imp_d1; + unsigned char drv_imp_d0; + unsigned int ibias; + unsigned int ibias_up; +}; + +void mtk_hdmi_phy_clear_bits(struct mtk_hdmi_phy *hdmi_phy, u32 offset, + u32 bits); +void mtk_hdmi_phy_set_bits(struct mtk_hdmi_phy *hdmi_phy, u32 offset, + u32 bits); +void mtk_hdmi_phy_mask(struct mtk_hdmi_phy *hdmi_phy, u32 offset, + u32 val, u32 mask); +struct mtk_hdmi_phy *to_mtk_hdmi_phy(struct clk_hw *hw); +long mtk_hdmi_pll_round_rate(struct clk_hw *hw, unsigned long rate, + unsigned long *parent_rate); +unsigned long mtk_hdmi_pll_recalc_rate(struct clk_hw *hw, + unsigned long parent_rate); + +extern struct platform_driver mtk_hdmi_phy_driver; +extern struct mtk_hdmi_phy_conf mtk_hdmi_phy_8173_conf; +extern struct mtk_hdmi_phy_conf mtk_hdmi_phy_2701_conf; + +#endif /* _MTK_HDMI_PHY_H */ diff --git a/drivers/gpu/drm/mediatek/mtk_mt2701_hdmi_phy.c b/drivers/gpu/drm/mediatek/mtk_mt2701_hdmi_phy.c new file mode 100644 index 000000000000..fcc42dc6ea7f --- /dev/null +++ b/drivers/gpu/drm/mediatek/mtk_mt2701_hdmi_phy.c @@ -0,0 +1,212 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2018 MediaTek Inc. + * Author: Chunhui Dai <chunhui.dai@mediatek.com> + */ + +#include "mtk_hdmi_phy.h" + +#define HDMI_CON0 0x00 +#define RG_HDMITX_DRV_IBIAS 0 +#define RG_HDMITX_DRV_IBIAS_MASK (0x3f << 0) +#define RG_HDMITX_EN_SER 12 +#define RG_HDMITX_EN_SER_MASK (0x0f << 12) +#define RG_HDMITX_EN_SLDO 16 +#define RG_HDMITX_EN_SLDO_MASK (0x0f << 16) +#define RG_HDMITX_EN_PRED 20 +#define RG_HDMITX_EN_PRED_MASK (0x0f << 20) +#define RG_HDMITX_EN_IMP 24 +#define RG_HDMITX_EN_IMP_MASK (0x0f << 24) +#define RG_HDMITX_EN_DRV 28 +#define RG_HDMITX_EN_DRV_MASK (0x0f << 28) + +#define HDMI_CON1 0x04 +#define RG_HDMITX_PRED_IBIAS 18 +#define RG_HDMITX_PRED_IBIAS_MASK (0x0f << 18) +#define RG_HDMITX_PRED_IMP (0x01 << 22) +#define RG_HDMITX_DRV_IMP 26 +#define RG_HDMITX_DRV_IMP_MASK (0x3f << 26) + +#define HDMI_CON2 0x08 +#define RG_HDMITX_EN_TX_CKLDO (0x01 << 0) +#define RG_HDMITX_EN_TX_POSDIV (0x01 << 1) +#define RG_HDMITX_TX_POSDIV 3 +#define RG_HDMITX_TX_POSDIV_MASK (0x03 << 3) +#define RG_HDMITX_EN_MBIAS (0x01 << 6) +#define RG_HDMITX_MBIAS_LPF_EN (0x01 << 7) + +#define HDMI_CON4 0x10 +#define RG_HDMITX_RESERVE_MASK (0xffffffff << 0) + +#define HDMI_CON6 0x18 +#define RG_HTPLL_BR 0 +#define RG_HTPLL_BR_MASK (0x03 << 0) +#define RG_HTPLL_BC 2 +#define RG_HTPLL_BC_MASK (0x03 << 2) +#define RG_HTPLL_BP 4 +#define RG_HTPLL_BP_MASK (0x0f << 4) +#define RG_HTPLL_IR 8 +#define RG_HTPLL_IR_MASK (0x0f << 8) +#define RG_HTPLL_IC 12 +#define RG_HTPLL_IC_MASK (0x0f << 12) +#define RG_HTPLL_POSDIV 16 +#define RG_HTPLL_POSDIV_MASK (0x03 << 16) +#define RG_HTPLL_PREDIV 18 +#define RG_HTPLL_PREDIV_MASK (0x03 << 18) +#define RG_HTPLL_FBKSEL 20 +#define RG_HTPLL_FBKSEL_MASK (0x03 << 20) +#define RG_HTPLL_RLH_EN (0x01 << 22) +#define RG_HTPLL_FBKDIV 24 +#define RG_HTPLL_FBKDIV_MASK (0x7f << 24) +#define RG_HTPLL_EN (0x01 << 31) + +#define HDMI_CON7 0x1c +#define RG_HTPLL_AUTOK_EN (0x01 << 23) +#define RG_HTPLL_DIVEN 28 +#define RG_HTPLL_DIVEN_MASK (0x07 << 28) + +static int mtk_hdmi_pll_prepare(struct clk_hw *hw) +{ + struct mtk_hdmi_phy *hdmi_phy = to_mtk_hdmi_phy(hw); + + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON7, RG_HTPLL_AUTOK_EN); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_RLH_EN); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_POSDIV_MASK); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_EN_MBIAS); + usleep_range(80, 100); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_EN); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_EN_TX_CKLDO); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_SLDO_MASK); + usleep_range(80, 100); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_MBIAS_LPF_EN); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_EN_TX_POSDIV); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_SER_MASK); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_PRED_MASK); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_DRV_MASK); + usleep_range(80, 100); + return 0; +} + +static void mtk_hdmi_pll_unprepare(struct clk_hw *hw) +{ + struct mtk_hdmi_phy *hdmi_phy = to_mtk_hdmi_phy(hw); + + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_DRV_MASK); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_PRED_MASK); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_SER_MASK); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_EN_TX_POSDIV); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_MBIAS_LPF_EN); + usleep_range(80, 100); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_SLDO_MASK); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_EN_TX_CKLDO); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_EN); + usleep_range(80, 100); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_EN_MBIAS); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_POSDIV_MASK); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_RLH_EN); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON7, RG_HTPLL_AUTOK_EN); + usleep_range(80, 100); +} + +static int mtk_hdmi_pll_set_rate(struct clk_hw *hw, unsigned long rate, + unsigned long parent_rate) +{ + struct mtk_hdmi_phy *hdmi_phy = to_mtk_hdmi_phy(hw); + u32 pos_div; + + if (rate <= 64000000) + pos_div = 3; + else if (rate <= 12800000) + pos_div = 1; + else + pos_div = 1; + + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_PREDIV_MASK); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_POSDIV_MASK); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON6, (0x1 << RG_HTPLL_IC), + RG_HTPLL_IC_MASK); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON6, (0x1 << RG_HTPLL_IR), + RG_HTPLL_IR_MASK); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON2, (pos_div << RG_HDMITX_TX_POSDIV), + RG_HDMITX_TX_POSDIV_MASK); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON6, (1 << RG_HTPLL_FBKSEL), + RG_HTPLL_FBKSEL_MASK); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON6, (19 << RG_HTPLL_FBKDIV), + RG_HTPLL_FBKDIV_MASK); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON7, (0x2 << RG_HTPLL_DIVEN), + RG_HTPLL_DIVEN_MASK); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON6, (0xc << RG_HTPLL_BP), + RG_HTPLL_BP_MASK); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON6, (0x2 << RG_HTPLL_BC), + RG_HTPLL_BC_MASK); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON6, (0x1 << RG_HTPLL_BR), + RG_HTPLL_BR_MASK); + + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON1, RG_HDMITX_PRED_IMP); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON1, (0x3 << RG_HDMITX_PRED_IBIAS), + RG_HDMITX_PRED_IBIAS_MASK); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_IMP_MASK); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON1, (0x28 << RG_HDMITX_DRV_IMP), + RG_HDMITX_DRV_IMP_MASK); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON4, 0x28, RG_HDMITX_RESERVE_MASK); + mtk_hdmi_phy_mask(hdmi_phy, HDMI_CON0, (0xa << RG_HDMITX_DRV_IBIAS), + RG_HDMITX_DRV_IBIAS_MASK); + return 0; +} + +static const struct clk_ops mtk_hdmi_phy_pll_ops = { + .prepare = mtk_hdmi_pll_prepare, + .unprepare = mtk_hdmi_pll_unprepare, + .set_rate = mtk_hdmi_pll_set_rate, + .round_rate = mtk_hdmi_pll_round_rate, + .recalc_rate = mtk_hdmi_pll_recalc_rate, +}; + +static void mtk_hdmi_phy_enable_tmds(struct mtk_hdmi_phy *hdmi_phy) +{ + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON7, RG_HTPLL_AUTOK_EN); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_RLH_EN); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_POSDIV_MASK); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_EN_MBIAS); + usleep_range(80, 100); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_EN); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_EN_TX_CKLDO); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_SLDO_MASK); + usleep_range(80, 100); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_MBIAS_LPF_EN); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_EN_TX_POSDIV); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_SER_MASK); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_PRED_MASK); + mtk_hdmi_phy_set_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_DRV_MASK); + usleep_range(80, 100); +} + +static void mtk_hdmi_phy_disable_tmds(struct mtk_hdmi_phy *hdmi_phy) +{ + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_DRV_MASK); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_PRED_MASK); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_SER_MASK); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_EN_TX_POSDIV); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_MBIAS_LPF_EN); + usleep_range(80, 100); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON0, RG_HDMITX_EN_SLDO_MASK); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_EN_TX_CKLDO); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_EN); + usleep_range(80, 100); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON2, RG_HDMITX_EN_MBIAS); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_POSDIV_MASK); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON6, RG_HTPLL_RLH_EN); + mtk_hdmi_phy_clear_bits(hdmi_phy, HDMI_CON7, RG_HTPLL_AUTOK_EN); + usleep_range(80, 100); +} + +struct mtk_hdmi_phy_conf mtk_hdmi_phy_2701_conf = { + .tz_disabled = true, + .hdmi_phy_clk_ops = &mtk_hdmi_phy_pll_ops, + .hdmi_phy_enable_tmds = mtk_hdmi_phy_enable_tmds, + .hdmi_phy_disable_tmds = mtk_hdmi_phy_disable_tmds, +}; + +MODULE_AUTHOR("Chunhui Dai <chunhui.dai@mediatek.com>"); +MODULE_DESCRIPTION("MediaTek HDMI PHY Driver"); +MODULE_LICENSE("GPL v2"); diff --git a/drivers/gpu/drm/mediatek/mtk_mt8173_hdmi_phy.c b/drivers/gpu/drm/mediatek/mtk_mt8173_hdmi_phy.c index 51cb9cfb6646..ed5916b27658 100644 --- a/drivers/gpu/drm/mediatek/mtk_mt8173_hdmi_phy.c +++ b/drivers/gpu/drm/mediatek/mtk_mt8173_hdmi_phy.c @@ -12,15 +12,7 @@ * GNU General Public License for more details. */ -#include <linux/clk.h> -#include <linux/clk-provider.h> -#include <linux/delay.h> -#include <linux/io.h> -#include <linux/mfd/syscon.h> -#include <linux/module.h> -#include <linux/phy/phy.h> -#include <linux/platform_device.h> -#include <linux/types.h> +#include "mtk_hdmi_phy.h" #define HDMI_CON0 0x00 #define RG_HDMITX_PLL_EN BIT(31) @@ -123,20 +115,6 @@ #define RGS_HDMITX_5T1_EDG (0xf << 4) #define RGS_HDMITX_PLUG_TST BIT(0) -struct mtk_hdmi_phy { - void __iomem *regs; - struct device *dev; - struct clk *pll; - struct clk_hw pll_hw; - unsigned long pll_rate; - u8 drv_imp_clk; - u8 drv_imp_d2; - u8 drv_imp_d1; - u8 drv_imp_d0; - u32 ibias; - u32 ibias_up; -}; - static const u8 PREDIV[3][4] = { {0x0, 0x0, 0x0, 0x0}, /* 27Mhz */ {0x1, 0x1, 0x1, 0x1}, /* 74Mhz */ @@ -185,44 +163,6 @@ static const u8 HTPLLBR[3][4] = { {0x1, 0x2, 0x2, 0x1} /* 148Mhz */ }; -static void mtk_hdmi_phy_clear_bits(struct mtk_hdmi_phy *hdmi_phy, u32 offset, - u32 bits) -{ - void __iomem *reg = hdmi_phy->regs + offset; - u32 tmp; - - tmp = readl(reg); - tmp &= ~bits; - writel(tmp, reg); -} - -static void mtk_hdmi_phy_set_bits(struct mtk_hdmi_phy *hdmi_phy, u32 offset, - u32 bits) -{ - void __iomem *reg = hdmi_phy->regs + offset; - u32 tmp; - - tmp = readl(reg); - tmp |= bits; - writel(tmp, reg); -} - -static void mtk_hdmi_phy_mask(struct mtk_hdmi_phy *hdmi_phy, u32 offset, - u32 val, u32 mask) -{ - void __iomem *reg = hdmi_phy->regs + offset; - u32 tmp; - - tmp = readl(reg); - tmp = (tmp & ~mask) | (val & mask); - writel(tmp, reg); -} - -static inline struct mtk_hdmi_phy *to_mtk_hdmi_phy(struct clk_hw *hw) -{ - return container_of(hw, struct mtk_hdmi_phy, pll_hw); -} - static int mtk_hdmi_pll_prepare(struct clk_hw *hw) { struct mtk_hdmi_phy *hdmi_phy = to_mtk_hdmi_phy(hw); @@ -345,29 +285,7 @@ static int mtk_hdmi_pll_set_rate(struct clk_hw *hw, unsigned long rate, return 0; } -static long mtk_hdmi_pll_round_rate(struct clk_hw *hw, unsigned long rate, - unsigned long *parent_rate) -{ - struct mtk_hdmi_phy *hdmi_phy = to_mtk_hdmi_phy(hw); - - hdmi_phy->pll_rate = rate; - if (rate <= 74250000) - *parent_rate = rate; - else - *parent_rate = rate / 2; - - return rate; -} - -static unsigned long mtk_hdmi_pll_recalc_rate(struct clk_hw *hw, - unsigned long parent_rate) -{ - struct mtk_hdmi_phy *hdmi_phy = to_mtk_hdmi_phy(hw); - - return hdmi_phy->pll_rate; -} - -static const struct clk_ops mtk_hdmi_pll_ops = { +static const struct clk_ops mtk_hdmi_phy_pll_ops = { .prepare = mtk_hdmi_pll_prepare, .unprepare = mtk_hdmi_pll_unprepare, .set_rate = mtk_hdmi_pll_set_rate, @@ -390,142 +308,10 @@ static void mtk_hdmi_phy_disable_tmds(struct mtk_hdmi_phy *hdmi_phy) RG_HDMITX_SER_EN); } -static int mtk_hdmi_phy_power_on(struct phy *phy) -{ - struct mtk_hdmi_phy *hdmi_phy = phy_get_drvdata(phy); - int ret; - - ret = clk_prepare_enable(hdmi_phy->pll); - if (ret < 0) - return ret; - - mtk_hdmi_phy_enable_tmds(hdmi_phy); - - return 0; -} - -static int mtk_hdmi_phy_power_off(struct phy *phy) -{ - struct mtk_hdmi_phy *hdmi_phy = phy_get_drvdata(phy); - - mtk_hdmi_phy_disable_tmds(hdmi_phy); - clk_disable_unprepare(hdmi_phy->pll); - - return 0; -} - -static const struct phy_ops mtk_hdmi_phy_ops = { - .power_on = mtk_hdmi_phy_power_on, - .power_off = mtk_hdmi_phy_power_off, - .owner = THIS_MODULE, -}; - -static int mtk_hdmi_phy_probe(struct platform_device *pdev) -{ - struct device *dev = &pdev->dev; - struct mtk_hdmi_phy *hdmi_phy; - struct resource *mem; - struct clk *ref_clk; - const char *ref_clk_name; - struct clk_init_data clk_init = { - .ops = &mtk_hdmi_pll_ops, - .num_parents = 1, - .parent_names = (const char * const *)&ref_clk_name, - .flags = CLK_SET_RATE_PARENT | CLK_SET_RATE_GATE, - }; - struct phy *phy; - struct phy_provider *phy_provider; - int ret; - - hdmi_phy = devm_kzalloc(dev, sizeof(*hdmi_phy), GFP_KERNEL); - if (!hdmi_phy) - return -ENOMEM; - - mem = platform_get_resource(pdev, IORESOURCE_MEM, 0); - hdmi_phy->regs = devm_ioremap_resource(dev, mem); - if (IS_ERR(hdmi_phy->regs)) { - ret = PTR_ERR(hdmi_phy->regs); - dev_err(dev, "Failed to get memory resource: %d\n", ret); - return ret; - } - - ref_clk = devm_clk_get(dev, "pll_ref"); - if (IS_ERR(ref_clk)) { - ret = PTR_ERR(ref_clk); - dev_err(&pdev->dev, "Failed to get PLL reference clock: %d\n", - ret); - return ret; - } - ref_clk_name = __clk_get_name(ref_clk); - - ret = of_property_read_string(dev->of_node, "clock-output-names", - &clk_init.name); - if (ret < 0) { - dev_err(dev, "Failed to read clock-output-names: %d\n", ret); - return ret; - } - - hdmi_phy->pll_hw.init = &clk_init; - hdmi_phy->pll = devm_clk_register(dev, &hdmi_phy->pll_hw); - if (IS_ERR(hdmi_phy->pll)) { - ret = PTR_ERR(hdmi_phy->pll); - dev_err(dev, "Failed to register PLL: %d\n", ret); - return ret; - } - - ret = of_property_read_u32(dev->of_node, "mediatek,ibias", - &hdmi_phy->ibias); - if (ret < 0) { - dev_err(&pdev->dev, "Failed to get ibias: %d\n", ret); - return ret; - } - - ret = of_property_read_u32(dev->of_node, "mediatek,ibias_up", - &hdmi_phy->ibias_up); - if (ret < 0) { - dev_err(&pdev->dev, "Failed to get ibias up: %d\n", ret); - return ret; - } - - dev_info(dev, "Using default TX DRV impedance: 4.2k/36\n"); - hdmi_phy->drv_imp_clk = 0x30; - hdmi_phy->drv_imp_d2 = 0x30; - hdmi_phy->drv_imp_d1 = 0x30; - hdmi_phy->drv_imp_d0 = 0x30; - - phy = devm_phy_create(dev, NULL, &mtk_hdmi_phy_ops); - if (IS_ERR(phy)) { - dev_err(dev, "Failed to create HDMI PHY\n"); - return PTR_ERR(phy); - } - phy_set_drvdata(phy, hdmi_phy); - - phy_provider = devm_of_phy_provider_register(dev, of_phy_simple_xlate); - if (IS_ERR(phy_provider)) - return PTR_ERR(phy_provider); - - hdmi_phy->dev = dev; - return of_clk_add_provider(dev->of_node, of_clk_src_simple_get, - hdmi_phy->pll); -} - -static int mtk_hdmi_phy_remove(struct platform_device *pdev) -{ - return 0; -} - -static const struct of_device_id mtk_hdmi_phy_match[] = { - { .compatible = "mediatek,mt8173-hdmi-phy", }, - {}, -}; - -struct platform_driver mtk_hdmi_phy_driver = { - .probe = mtk_hdmi_phy_probe, - .remove = mtk_hdmi_phy_remove, - .driver = { - .name = "mediatek-hdmi-phy", - .of_match_table = mtk_hdmi_phy_match, - }, +struct mtk_hdmi_phy_conf mtk_hdmi_phy_8173_conf = { + .hdmi_phy_clk_ops = &mtk_hdmi_phy_pll_ops, + .hdmi_phy_enable_tmds = mtk_hdmi_phy_enable_tmds, + .hdmi_phy_disable_tmds = mtk_hdmi_phy_disable_tmds, }; MODULE_AUTHOR("Jie Qiu <jie.qiu@mediatek.com>"); diff --git a/drivers/gpu/drm/msm/Makefile b/drivers/gpu/drm/msm/Makefile index 261fa79d456d..19ab521d4c3a 100644 --- a/drivers/gpu/drm/msm/Makefile +++ b/drivers/gpu/drm/msm/Makefile @@ -58,7 +58,6 @@ msm-y := \ disp/dpu1/dpu_formats.o \ disp/dpu1/dpu_hw_blk.o \ disp/dpu1/dpu_hw_catalog.o \ - disp/dpu1/dpu_hw_cdm.o \ disp/dpu1/dpu_hw_ctl.o \ disp/dpu1/dpu_hw_interrupts.o \ disp/dpu1/dpu_hw_intf.o \ diff --git a/drivers/gpu/drm/msm/adreno/a2xx.xml.h b/drivers/gpu/drm/msm/adreno/a2xx.xml.h index 4bff0a740c7d..12b0ba270b5e 100644 --- a/drivers/gpu/drm/msm/adreno/a2xx.xml.h +++ b/drivers/gpu/drm/msm/adreno/a2xx.xml.h @@ -12,12 +12,12 @@ The rules-ng-ng source files this header was generated from are: - /home/robclark/src/envytools/rnndb/freedreno_copyright.xml ( 1572 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a2xx.xml ( 36805 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/adreno_common.xml ( 13634 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42393 bytes, from 2018-08-06 18:45:45) +- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42585 bytes, from 2018-10-04 19:06:37) - /home/robclark/src/envytools/rnndb/adreno/a3xx.xml ( 83840 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a4xx.xml ( 112086 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 101627 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-07-03 19:37:13) +- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-10-04 19:06:37) +- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 139581 bytes, from 2018-10-04 19:06:42) +- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-09-14 13:03:07) - /home/robclark/src/envytools/rnndb/adreno/ocmem.xml ( 1773 bytes, from 2018-07-03 19:37:13) Copyright (C) 2013-2018 by the following authors: diff --git a/drivers/gpu/drm/msm/adreno/a3xx.xml.h b/drivers/gpu/drm/msm/adreno/a3xx.xml.h index 645a19aef399..a89f7bb8b5cc 100644 --- a/drivers/gpu/drm/msm/adreno/a3xx.xml.h +++ b/drivers/gpu/drm/msm/adreno/a3xx.xml.h @@ -12,12 +12,12 @@ The rules-ng-ng source files this header was generated from are: - /home/robclark/src/envytools/rnndb/freedreno_copyright.xml ( 1572 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a2xx.xml ( 36805 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/adreno_common.xml ( 13634 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42393 bytes, from 2018-08-06 18:45:45) +- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42585 bytes, from 2018-10-04 19:06:37) - /home/robclark/src/envytools/rnndb/adreno/a3xx.xml ( 83840 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a4xx.xml ( 112086 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 101627 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-07-03 19:37:13) +- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-10-04 19:06:37) +- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 139581 bytes, from 2018-10-04 19:06:42) +- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-09-14 13:03:07) - /home/robclark/src/envytools/rnndb/adreno/ocmem.xml ( 1773 bytes, from 2018-07-03 19:37:13) Copyright (C) 2013-2018 by the following authors: diff --git a/drivers/gpu/drm/msm/adreno/a4xx.xml.h b/drivers/gpu/drm/msm/adreno/a4xx.xml.h index 19565e87aa7b..858690f52854 100644 --- a/drivers/gpu/drm/msm/adreno/a4xx.xml.h +++ b/drivers/gpu/drm/msm/adreno/a4xx.xml.h @@ -12,12 +12,12 @@ The rules-ng-ng source files this header was generated from are: - /home/robclark/src/envytools/rnndb/freedreno_copyright.xml ( 1572 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a2xx.xml ( 36805 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/adreno_common.xml ( 13634 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42393 bytes, from 2018-08-06 18:45:45) +- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42585 bytes, from 2018-10-04 19:06:37) - /home/robclark/src/envytools/rnndb/adreno/a3xx.xml ( 83840 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a4xx.xml ( 112086 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 101627 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-07-03 19:37:13) +- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-10-04 19:06:37) +- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 139581 bytes, from 2018-10-04 19:06:42) +- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-09-14 13:03:07) - /home/robclark/src/envytools/rnndb/adreno/ocmem.xml ( 1773 bytes, from 2018-07-03 19:37:13) Copyright (C) 2013-2018 by the following authors: diff --git a/drivers/gpu/drm/msm/adreno/a5xx.xml.h b/drivers/gpu/drm/msm/adreno/a5xx.xml.h index 182d37ff3794..b4944cc0e62f 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx.xml.h +++ b/drivers/gpu/drm/msm/adreno/a5xx.xml.h @@ -12,12 +12,12 @@ The rules-ng-ng source files this header was generated from are: - /home/robclark/src/envytools/rnndb/freedreno_copyright.xml ( 1572 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a2xx.xml ( 36805 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/adreno_common.xml ( 13634 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42393 bytes, from 2018-08-06 18:45:45) +- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42585 bytes, from 2018-10-04 19:06:37) - /home/robclark/src/envytools/rnndb/adreno/a3xx.xml ( 83840 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a4xx.xml ( 112086 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 101627 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-07-03 19:37:13) +- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-10-04 19:06:37) +- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 139581 bytes, from 2018-10-04 19:06:42) +- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-09-14 13:03:07) - /home/robclark/src/envytools/rnndb/adreno/ocmem.xml ( 1773 bytes, from 2018-07-03 19:37:13) Copyright (C) 2013-2018 by the following authors: diff --git a/drivers/gpu/drm/msm/adreno/a5xx_debugfs.c b/drivers/gpu/drm/msm/adreno/a5xx_debugfs.c index 059ec7d394d0..d2127b1c4ece 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_debugfs.c +++ b/drivers/gpu/drm/msm/adreno/a5xx_debugfs.c @@ -132,14 +132,14 @@ reset_set(void *data, u64 val) if (a5xx_gpu->pm4_bo) { if (a5xx_gpu->pm4_iova) msm_gem_put_iova(a5xx_gpu->pm4_bo, gpu->aspace); - drm_gem_object_unreference(a5xx_gpu->pm4_bo); + drm_gem_object_put(a5xx_gpu->pm4_bo); a5xx_gpu->pm4_bo = NULL; } if (a5xx_gpu->pfp_bo) { if (a5xx_gpu->pfp_iova) msm_gem_put_iova(a5xx_gpu->pfp_bo, gpu->aspace); - drm_gem_object_unreference(a5xx_gpu->pfp_bo); + drm_gem_object_put(a5xx_gpu->pfp_bo); a5xx_gpu->pfp_bo = NULL; } diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c index 29b3d9d7d54b..8edd80bb0428 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c @@ -1231,7 +1231,7 @@ static void a5xx_crashdumper_free(struct msm_gpu *gpu, msm_gem_put_iova(dumper->bo, gpu->aspace); msm_gem_put_vaddr(dumper->bo); - drm_gem_object_unreference(dumper->bo); + drm_gem_object_put(dumper->bo); } static int a5xx_crashdumper_run(struct msm_gpu *gpu, @@ -1433,12 +1433,22 @@ static struct msm_ringbuffer *a5xx_active_ring(struct msm_gpu *gpu) return a5xx_gpu->cur_ring; } -static int a5xx_gpu_busy(struct msm_gpu *gpu, uint64_t *value) +static unsigned long a5xx_gpu_busy(struct msm_gpu *gpu) { - *value = gpu_read64(gpu, REG_A5XX_RBBM_PERFCTR_RBBM_0_LO, - REG_A5XX_RBBM_PERFCTR_RBBM_0_HI); + u64 busy_cycles, busy_time; - return 0; + busy_cycles = gpu_read64(gpu, REG_A5XX_RBBM_PERFCTR_RBBM_0_LO, + REG_A5XX_RBBM_PERFCTR_RBBM_0_HI); + + busy_time = busy_cycles - gpu->devfreq.busy_cycles; + do_div(busy_time, clk_get_rate(gpu->core_clk) / 1000000); + + gpu->devfreq.busy_cycles = busy_cycles; + + if (WARN_ON(busy_time > ~0LU)) + return ~0LU; + + return (unsigned long)busy_time; } static const struct adreno_gpu_funcs funcs = { diff --git a/drivers/gpu/drm/msm/adreno/a5xx_power.c b/drivers/gpu/drm/msm/adreno/a5xx_power.c index e9c0e56dbec0..7a41e1c147e4 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_power.c +++ b/drivers/gpu/drm/msm/adreno/a5xx_power.c @@ -323,7 +323,7 @@ err: if (a5xx_gpu->gpmu_iova) msm_gem_put_iova(a5xx_gpu->gpmu_bo, gpu->aspace); if (a5xx_gpu->gpmu_bo) - drm_gem_object_unreference(a5xx_gpu->gpmu_bo); + drm_gem_object_put(a5xx_gpu->gpmu_bo); a5xx_gpu->gpmu_bo = NULL; a5xx_gpu->gpmu_iova = 0; diff --git a/drivers/gpu/drm/msm/adreno/a5xx_preempt.c b/drivers/gpu/drm/msm/adreno/a5xx_preempt.c index 970c7963ae29..4c357ead1be6 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_preempt.c +++ b/drivers/gpu/drm/msm/adreno/a5xx_preempt.c @@ -208,6 +208,13 @@ void a5xx_preempt_hw_init(struct msm_gpu *gpu) struct a5xx_gpu *a5xx_gpu = to_a5xx_gpu(adreno_gpu); int i; + /* Always come up on rb 0 */ + a5xx_gpu->cur_ring = gpu->rb[0]; + + /* No preemption if we only have one ring */ + if (gpu->nr_rings == 1) + return; + for (i = 0; i < gpu->nr_rings; i++) { a5xx_gpu->preempt[i]->wptr = 0; a5xx_gpu->preempt[i]->rptr = 0; @@ -220,9 +227,6 @@ void a5xx_preempt_hw_init(struct msm_gpu *gpu) /* Reset the preemption state */ set_preempt_state(a5xx_gpu, PREEMPT_NONE); - - /* Always come up on rb 0 */ - a5xx_gpu->cur_ring = gpu->rb[0]; } static int preempt_init_ring(struct a5xx_gpu *a5xx_gpu, @@ -272,7 +276,7 @@ void a5xx_preempt_fini(struct msm_gpu *gpu) if (a5xx_gpu->preempt_iova[i]) msm_gem_put_iova(a5xx_gpu->preempt_bo[i], gpu->aspace); - drm_gem_object_unreference(a5xx_gpu->preempt_bo[i]); + drm_gem_object_put(a5xx_gpu->preempt_bo[i]); a5xx_gpu->preempt_bo[i] = NULL; } } diff --git a/drivers/gpu/drm/msm/adreno/a6xx.xml.h b/drivers/gpu/drm/msm/adreno/a6xx.xml.h index 87eab51f7000..a6f7c40454a6 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx.xml.h +++ b/drivers/gpu/drm/msm/adreno/a6xx.xml.h @@ -12,12 +12,12 @@ The rules-ng-ng source files this header was generated from are: - /home/robclark/src/envytools/rnndb/freedreno_copyright.xml ( 1572 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a2xx.xml ( 36805 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/adreno_common.xml ( 13634 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42393 bytes, from 2018-08-06 18:45:45) +- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42585 bytes, from 2018-10-04 19:06:37) - /home/robclark/src/envytools/rnndb/adreno/a3xx.xml ( 83840 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a4xx.xml ( 112086 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 101627 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-07-03 19:37:13) +- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-10-04 19:06:37) +- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 139581 bytes, from 2018-10-04 19:06:42) +- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-09-14 13:03:07) - /home/robclark/src/envytools/rnndb/adreno/ocmem.xml ( 1773 bytes, from 2018-07-03 19:37:13) Copyright (C) 2013-2018 by the following authors: @@ -268,8 +268,687 @@ enum a6xx_depth_format { DEPTH6_32 = 4, }; +enum a6xx_shader_id { + A6XX_TP0_TMO_DATA = 9, + A6XX_TP0_SMO_DATA = 10, + A6XX_TP0_MIPMAP_BASE_DATA = 11, + A6XX_TP1_TMO_DATA = 25, + A6XX_TP1_SMO_DATA = 26, + A6XX_TP1_MIPMAP_BASE_DATA = 27, + A6XX_SP_INST_DATA = 41, + A6XX_SP_LB_0_DATA = 42, + A6XX_SP_LB_1_DATA = 43, + A6XX_SP_LB_2_DATA = 44, + A6XX_SP_LB_3_DATA = 45, + A6XX_SP_LB_4_DATA = 46, + A6XX_SP_LB_5_DATA = 47, + A6XX_SP_CB_BINDLESS_DATA = 48, + A6XX_SP_CB_LEGACY_DATA = 49, + A6XX_SP_UAV_DATA = 50, + A6XX_SP_INST_TAG = 51, + A6XX_SP_CB_BINDLESS_TAG = 52, + A6XX_SP_TMO_UMO_TAG = 53, + A6XX_SP_SMO_TAG = 54, + A6XX_SP_STATE_DATA = 55, + A6XX_HLSQ_CHUNK_CVS_RAM = 73, + A6XX_HLSQ_CHUNK_CPS_RAM = 74, + A6XX_HLSQ_CHUNK_CVS_RAM_TAG = 75, + A6XX_HLSQ_CHUNK_CPS_RAM_TAG = 76, + A6XX_HLSQ_ICB_CVS_CB_BASE_TAG = 77, + A6XX_HLSQ_ICB_CPS_CB_BASE_TAG = 78, + A6XX_HLSQ_CVS_MISC_RAM = 80, + A6XX_HLSQ_CPS_MISC_RAM = 81, + A6XX_HLSQ_INST_RAM = 82, + A6XX_HLSQ_GFX_CVS_CONST_RAM = 83, + A6XX_HLSQ_GFX_CPS_CONST_RAM = 84, + A6XX_HLSQ_CVS_MISC_RAM_TAG = 85, + A6XX_HLSQ_CPS_MISC_RAM_TAG = 86, + A6XX_HLSQ_INST_RAM_TAG = 87, + A6XX_HLSQ_GFX_CVS_CONST_RAM_TAG = 88, + A6XX_HLSQ_GFX_CPS_CONST_RAM_TAG = 89, + A6XX_HLSQ_PWR_REST_RAM = 90, + A6XX_HLSQ_PWR_REST_TAG = 91, + A6XX_HLSQ_DATAPATH_META = 96, + A6XX_HLSQ_FRONTEND_META = 97, + A6XX_HLSQ_INDIRECT_META = 98, + A6XX_HLSQ_BACKEND_META = 99, +}; + +enum a6xx_debugbus_id { + A6XX_DBGBUS_CP = 1, + A6XX_DBGBUS_RBBM = 2, + A6XX_DBGBUS_VBIF = 3, + A6XX_DBGBUS_HLSQ = 4, + A6XX_DBGBUS_UCHE = 5, + A6XX_DBGBUS_DPM = 6, + A6XX_DBGBUS_TESS = 7, + A6XX_DBGBUS_PC = 8, + A6XX_DBGBUS_VFDP = 9, + A6XX_DBGBUS_VPC = 10, + A6XX_DBGBUS_TSE = 11, + A6XX_DBGBUS_RAS = 12, + A6XX_DBGBUS_VSC = 13, + A6XX_DBGBUS_COM = 14, + A6XX_DBGBUS_LRZ = 16, + A6XX_DBGBUS_A2D = 17, + A6XX_DBGBUS_CCUFCHE = 18, + A6XX_DBGBUS_GMU_CX = 19, + A6XX_DBGBUS_RBP = 20, + A6XX_DBGBUS_DCS = 21, + A6XX_DBGBUS_DBGC = 22, + A6XX_DBGBUS_CX = 23, + A6XX_DBGBUS_GMU_GX = 24, + A6XX_DBGBUS_TPFCHE = 25, + A6XX_DBGBUS_GBIF_GX = 26, + A6XX_DBGBUS_GPC = 29, + A6XX_DBGBUS_LARC = 30, + A6XX_DBGBUS_HLSQ_SPTP = 31, + A6XX_DBGBUS_RB_0 = 32, + A6XX_DBGBUS_RB_1 = 33, + A6XX_DBGBUS_UCHE_WRAPPER = 36, + A6XX_DBGBUS_CCU_0 = 40, + A6XX_DBGBUS_CCU_1 = 41, + A6XX_DBGBUS_VFD_0 = 56, + A6XX_DBGBUS_VFD_1 = 57, + A6XX_DBGBUS_VFD_2 = 58, + A6XX_DBGBUS_VFD_3 = 59, + A6XX_DBGBUS_SP_0 = 64, + A6XX_DBGBUS_SP_1 = 65, + A6XX_DBGBUS_TPL1_0 = 72, + A6XX_DBGBUS_TPL1_1 = 73, + A6XX_DBGBUS_TPL1_2 = 74, + A6XX_DBGBUS_TPL1_3 = 75, +}; + enum a6xx_cp_perfcounter_select { PERF_CP_ALWAYS_COUNT = 0, + PERF_CP_BUSY_GFX_CORE_IDLE = 1, + PERF_CP_BUSY_CYCLES = 2, + PERF_CP_NUM_PREEMPTIONS = 3, + PERF_CP_PREEMPTION_REACTION_DELAY = 4, + PERF_CP_PREEMPTION_SWITCH_OUT_TIME = 5, + PERF_CP_PREEMPTION_SWITCH_IN_TIME = 6, + PERF_CP_DEAD_DRAWS_IN_BIN_RENDER = 7, + PERF_CP_PREDICATED_DRAWS_KILLED = 8, + PERF_CP_MODE_SWITCH = 9, + PERF_CP_ZPASS_DONE = 10, + PERF_CP_CONTEXT_DONE = 11, + PERF_CP_CACHE_FLUSH = 12, + PERF_CP_LONG_PREEMPTIONS = 13, + PERF_CP_SQE_I_CACHE_STARVE = 14, + PERF_CP_SQE_IDLE = 15, + PERF_CP_SQE_PM4_STARVE_RB_IB = 16, + PERF_CP_SQE_PM4_STARVE_SDS = 17, + PERF_CP_SQE_MRB_STARVE = 18, + PERF_CP_SQE_RRB_STARVE = 19, + PERF_CP_SQE_VSD_STARVE = 20, + PERF_CP_VSD_DECODE_STARVE = 21, + PERF_CP_SQE_PIPE_OUT_STALL = 22, + PERF_CP_SQE_SYNC_STALL = 23, + PERF_CP_SQE_PM4_WFI_STALL = 24, + PERF_CP_SQE_SYS_WFI_STALL = 25, + PERF_CP_SQE_T4_EXEC = 26, + PERF_CP_SQE_LOAD_STATE_EXEC = 27, + PERF_CP_SQE_SAVE_SDS_STATE = 28, + PERF_CP_SQE_DRAW_EXEC = 29, + PERF_CP_SQE_CTXT_REG_BUNCH_EXEC = 30, + PERF_CP_SQE_EXEC_PROFILED = 31, + PERF_CP_MEMORY_POOL_EMPTY = 32, + PERF_CP_MEMORY_POOL_SYNC_STALL = 33, + PERF_CP_MEMORY_POOL_ABOVE_THRESH = 34, + PERF_CP_AHB_WR_STALL_PRE_DRAWS = 35, + PERF_CP_AHB_STALL_SQE_GMU = 36, + PERF_CP_AHB_STALL_SQE_WR_OTHER = 37, + PERF_CP_AHB_STALL_SQE_RD_OTHER = 38, + PERF_CP_CLUSTER0_EMPTY = 39, + PERF_CP_CLUSTER1_EMPTY = 40, + PERF_CP_CLUSTER2_EMPTY = 41, + PERF_CP_CLUSTER3_EMPTY = 42, + PERF_CP_CLUSTER4_EMPTY = 43, + PERF_CP_CLUSTER5_EMPTY = 44, + PERF_CP_PM4_DATA = 45, + PERF_CP_PM4_HEADERS = 46, + PERF_CP_VBIF_READ_BEATS = 47, + PERF_CP_VBIF_WRITE_BEATS = 48, + PERF_CP_SQE_INSTR_COUNTER = 49, +}; + +enum a6xx_rbbm_perfcounter_select { + PERF_RBBM_ALWAYS_COUNT = 0, + PERF_RBBM_ALWAYS_ON = 1, + PERF_RBBM_TSE_BUSY = 2, + PERF_RBBM_RAS_BUSY = 3, + PERF_RBBM_PC_DCALL_BUSY = 4, + PERF_RBBM_PC_VSD_BUSY = 5, + PERF_RBBM_STATUS_MASKED = 6, + PERF_RBBM_COM_BUSY = 7, + PERF_RBBM_DCOM_BUSY = 8, + PERF_RBBM_VBIF_BUSY = 9, + PERF_RBBM_VSC_BUSY = 10, + PERF_RBBM_TESS_BUSY = 11, + PERF_RBBM_UCHE_BUSY = 12, + PERF_RBBM_HLSQ_BUSY = 13, +}; + +enum a6xx_pc_perfcounter_select { + PERF_PC_BUSY_CYCLES = 0, + PERF_PC_WORKING_CYCLES = 1, + PERF_PC_STALL_CYCLES_VFD = 2, + PERF_PC_STALL_CYCLES_TSE = 3, + PERF_PC_STALL_CYCLES_VPC = 4, + PERF_PC_STALL_CYCLES_UCHE = 5, + PERF_PC_STALL_CYCLES_TESS = 6, + PERF_PC_STALL_CYCLES_TSE_ONLY = 7, + PERF_PC_STALL_CYCLES_VPC_ONLY = 8, + PERF_PC_PASS1_TF_STALL_CYCLES = 9, + PERF_PC_STARVE_CYCLES_FOR_INDEX = 10, + PERF_PC_STARVE_CYCLES_FOR_TESS_FACTOR = 11, + PERF_PC_STARVE_CYCLES_FOR_VIZ_STREAM = 12, + PERF_PC_STARVE_CYCLES_FOR_POSITION = 13, + PERF_PC_STARVE_CYCLES_DI = 14, + PERF_PC_VIS_STREAMS_LOADED = 15, + PERF_PC_INSTANCES = 16, + PERF_PC_VPC_PRIMITIVES = 17, + PERF_PC_DEAD_PRIM = 18, + PERF_PC_LIVE_PRIM = 19, + PERF_PC_VERTEX_HITS = 20, + PERF_PC_IA_VERTICES = 21, + PERF_PC_IA_PRIMITIVES = 22, + PERF_PC_GS_PRIMITIVES = 23, + PERF_PC_HS_INVOCATIONS = 24, + PERF_PC_DS_INVOCATIONS = 25, + PERF_PC_VS_INVOCATIONS = 26, + PERF_PC_GS_INVOCATIONS = 27, + PERF_PC_DS_PRIMITIVES = 28, + PERF_PC_VPC_POS_DATA_TRANSACTION = 29, + PERF_PC_3D_DRAWCALLS = 30, + PERF_PC_2D_DRAWCALLS = 31, + PERF_PC_NON_DRAWCALL_GLOBAL_EVENTS = 32, + PERF_TESS_BUSY_CYCLES = 33, + PERF_TESS_WORKING_CYCLES = 34, + PERF_TESS_STALL_CYCLES_PC = 35, + PERF_TESS_STARVE_CYCLES_PC = 36, + PERF_PC_TSE_TRANSACTION = 37, + PERF_PC_TSE_VERTEX = 38, + PERF_PC_TESS_PC_UV_TRANS = 39, + PERF_PC_TESS_PC_UV_PATCHES = 40, + PERF_PC_TESS_FACTOR_TRANS = 41, +}; + +enum a6xx_vfd_perfcounter_select { + PERF_VFD_BUSY_CYCLES = 0, + PERF_VFD_STALL_CYCLES_UCHE = 1, + PERF_VFD_STALL_CYCLES_VPC_ALLOC = 2, + PERF_VFD_STALL_CYCLES_SP_INFO = 3, + PERF_VFD_STALL_CYCLES_SP_ATTR = 4, + PERF_VFD_STARVE_CYCLES_UCHE = 5, + PERF_VFD_RBUFFER_FULL = 6, + PERF_VFD_ATTR_INFO_FIFO_FULL = 7, + PERF_VFD_DECODED_ATTRIBUTE_BYTES = 8, + PERF_VFD_NUM_ATTRIBUTES = 9, + PERF_VFD_UPPER_SHADER_FIBERS = 10, + PERF_VFD_LOWER_SHADER_FIBERS = 11, + PERF_VFD_MODE_0_FIBERS = 12, + PERF_VFD_MODE_1_FIBERS = 13, + PERF_VFD_MODE_2_FIBERS = 14, + PERF_VFD_MODE_3_FIBERS = 15, + PERF_VFD_MODE_4_FIBERS = 16, + PERF_VFD_TOTAL_VERTICES = 17, + PERF_VFDP_STALL_CYCLES_VFD = 18, + PERF_VFDP_STALL_CYCLES_VFD_INDEX = 19, + PERF_VFDP_STALL_CYCLES_VFD_PROG = 20, + PERF_VFDP_STARVE_CYCLES_PC = 21, + PERF_VFDP_VS_STAGE_WAVES = 22, +}; + +enum a6xx_hslq_perfcounter_select { + PERF_HLSQ_BUSY_CYCLES = 0, + PERF_HLSQ_STALL_CYCLES_UCHE = 1, + PERF_HLSQ_STALL_CYCLES_SP_STATE = 2, + PERF_HLSQ_STALL_CYCLES_SP_FS_STAGE = 3, + PERF_HLSQ_UCHE_LATENCY_CYCLES = 4, + PERF_HLSQ_UCHE_LATENCY_COUNT = 5, + PERF_HLSQ_FS_STAGE_1X_WAVES = 6, + PERF_HLSQ_FS_STAGE_2X_WAVES = 7, + PERF_HLSQ_QUADS = 8, + PERF_HLSQ_CS_INVOCATIONS = 9, + PERF_HLSQ_COMPUTE_DRAWCALLS = 10, + PERF_HLSQ_FS_DATA_WAIT_PROGRAMMING = 11, + PERF_HLSQ_DUAL_FS_PROG_ACTIVE = 12, + PERF_HLSQ_DUAL_VS_PROG_ACTIVE = 13, + PERF_HLSQ_FS_BATCH_COUNT_ZERO = 14, + PERF_HLSQ_VS_BATCH_COUNT_ZERO = 15, + PERF_HLSQ_WAVE_PENDING_NO_QUAD = 16, + PERF_HLSQ_WAVE_PENDING_NO_PRIM_BASE = 17, + PERF_HLSQ_STALL_CYCLES_VPC = 18, + PERF_HLSQ_PIXELS = 19, + PERF_HLSQ_DRAW_MODE_SWITCH_VSFS_SYNC = 20, +}; + +enum a6xx_vpc_perfcounter_select { + PERF_VPC_BUSY_CYCLES = 0, + PERF_VPC_WORKING_CYCLES = 1, + PERF_VPC_STALL_CYCLES_UCHE = 2, + PERF_VPC_STALL_CYCLES_VFD_WACK = 3, + PERF_VPC_STALL_CYCLES_HLSQ_PRIM_ALLOC = 4, + PERF_VPC_STALL_CYCLES_PC = 5, + PERF_VPC_STALL_CYCLES_SP_LM = 6, + PERF_VPC_STARVE_CYCLES_SP = 7, + PERF_VPC_STARVE_CYCLES_LRZ = 8, + PERF_VPC_PC_PRIMITIVES = 9, + PERF_VPC_SP_COMPONENTS = 10, + PERF_VPC_STALL_CYCLES_VPCRAM_POS = 11, + PERF_VPC_LRZ_ASSIGN_PRIMITIVES = 12, + PERF_VPC_RB_VISIBLE_PRIMITIVES = 13, + PERF_VPC_LM_TRANSACTION = 14, + PERF_VPC_STREAMOUT_TRANSACTION = 15, + PERF_VPC_VS_BUSY_CYCLES = 16, + PERF_VPC_PS_BUSY_CYCLES = 17, + PERF_VPC_VS_WORKING_CYCLES = 18, + PERF_VPC_PS_WORKING_CYCLES = 19, + PERF_VPC_STARVE_CYCLES_RB = 20, + PERF_VPC_NUM_VPCRAM_READ_POS = 21, + PERF_VPC_WIT_FULL_CYCLES = 22, + PERF_VPC_VPCRAM_FULL_CYCLES = 23, + PERF_VPC_LM_FULL_WAIT_FOR_INTP_END = 24, + PERF_VPC_NUM_VPCRAM_WRITE = 25, + PERF_VPC_NUM_VPCRAM_READ_SO = 26, + PERF_VPC_NUM_ATTR_REQ_LM = 27, +}; + +enum a6xx_tse_perfcounter_select { + PERF_TSE_BUSY_CYCLES = 0, + PERF_TSE_CLIPPING_CYCLES = 1, + PERF_TSE_STALL_CYCLES_RAS = 2, + PERF_TSE_STALL_CYCLES_LRZ_BARYPLANE = 3, + PERF_TSE_STALL_CYCLES_LRZ_ZPLANE = 4, + PERF_TSE_STARVE_CYCLES_PC = 5, + PERF_TSE_INPUT_PRIM = 6, + PERF_TSE_INPUT_NULL_PRIM = 7, + PERF_TSE_TRIVAL_REJ_PRIM = 8, + PERF_TSE_CLIPPED_PRIM = 9, + PERF_TSE_ZERO_AREA_PRIM = 10, + PERF_TSE_FACENESS_CULLED_PRIM = 11, + PERF_TSE_ZERO_PIXEL_PRIM = 12, + PERF_TSE_OUTPUT_NULL_PRIM = 13, + PERF_TSE_OUTPUT_VISIBLE_PRIM = 14, + PERF_TSE_CINVOCATION = 15, + PERF_TSE_CPRIMITIVES = 16, + PERF_TSE_2D_INPUT_PRIM = 17, + PERF_TSE_2D_ALIVE_CYCLES = 18, + PERF_TSE_CLIP_PLANES = 19, +}; + +enum a6xx_ras_perfcounter_select { + PERF_RAS_BUSY_CYCLES = 0, + PERF_RAS_SUPERTILE_ACTIVE_CYCLES = 1, + PERF_RAS_STALL_CYCLES_LRZ = 2, + PERF_RAS_STARVE_CYCLES_TSE = 3, + PERF_RAS_SUPER_TILES = 4, + PERF_RAS_8X4_TILES = 5, + PERF_RAS_MASKGEN_ACTIVE = 6, + PERF_RAS_FULLY_COVERED_SUPER_TILES = 7, + PERF_RAS_FULLY_COVERED_8X4_TILES = 8, + PERF_RAS_PRIM_KILLED_INVISILBE = 9, + PERF_RAS_SUPERTILE_GEN_ACTIVE_CYCLES = 10, + PERF_RAS_LRZ_INTF_WORKING_CYCLES = 11, + PERF_RAS_BLOCKS = 12, +}; + +enum a6xx_uche_perfcounter_select { + PERF_UCHE_BUSY_CYCLES = 0, + PERF_UCHE_STALL_CYCLES_ARBITER = 1, + PERF_UCHE_VBIF_LATENCY_CYCLES = 2, + PERF_UCHE_VBIF_LATENCY_SAMPLES = 3, + PERF_UCHE_VBIF_READ_BEATS_TP = 4, + PERF_UCHE_VBIF_READ_BEATS_VFD = 5, + PERF_UCHE_VBIF_READ_BEATS_HLSQ = 6, + PERF_UCHE_VBIF_READ_BEATS_LRZ = 7, + PERF_UCHE_VBIF_READ_BEATS_SP = 8, + PERF_UCHE_READ_REQUESTS_TP = 9, + PERF_UCHE_READ_REQUESTS_VFD = 10, + PERF_UCHE_READ_REQUESTS_HLSQ = 11, + PERF_UCHE_READ_REQUESTS_LRZ = 12, + PERF_UCHE_READ_REQUESTS_SP = 13, + PERF_UCHE_WRITE_REQUESTS_LRZ = 14, + PERF_UCHE_WRITE_REQUESTS_SP = 15, + PERF_UCHE_WRITE_REQUESTS_VPC = 16, + PERF_UCHE_WRITE_REQUESTS_VSC = 17, + PERF_UCHE_EVICTS = 18, + PERF_UCHE_BANK_REQ0 = 19, + PERF_UCHE_BANK_REQ1 = 20, + PERF_UCHE_BANK_REQ2 = 21, + PERF_UCHE_BANK_REQ3 = 22, + PERF_UCHE_BANK_REQ4 = 23, + PERF_UCHE_BANK_REQ5 = 24, + PERF_UCHE_BANK_REQ6 = 25, + PERF_UCHE_BANK_REQ7 = 26, + PERF_UCHE_VBIF_READ_BEATS_CH0 = 27, + PERF_UCHE_VBIF_READ_BEATS_CH1 = 28, + PERF_UCHE_GMEM_READ_BEATS = 29, + PERF_UCHE_TPH_REF_FULL = 30, + PERF_UCHE_TPH_VICTIM_FULL = 31, + PERF_UCHE_TPH_EXT_FULL = 32, + PERF_UCHE_VBIF_STALL_WRITE_DATA = 33, + PERF_UCHE_DCMP_LATENCY_SAMPLES = 34, + PERF_UCHE_DCMP_LATENCY_CYCLES = 35, + PERF_UCHE_VBIF_READ_BEATS_PC = 36, + PERF_UCHE_READ_REQUESTS_PC = 37, + PERF_UCHE_RAM_READ_REQ = 38, + PERF_UCHE_RAM_WRITE_REQ = 39, +}; + +enum a6xx_tp_perfcounter_select { + PERF_TP_BUSY_CYCLES = 0, + PERF_TP_STALL_CYCLES_UCHE = 1, + PERF_TP_LATENCY_CYCLES = 2, + PERF_TP_LATENCY_TRANS = 3, + PERF_TP_FLAG_CACHE_REQUEST_SAMPLES = 4, + PERF_TP_FLAG_CACHE_REQUEST_LATENCY = 5, + PERF_TP_L1_CACHELINE_REQUESTS = 6, + PERF_TP_L1_CACHELINE_MISSES = 7, + PERF_TP_SP_TP_TRANS = 8, + PERF_TP_TP_SP_TRANS = 9, + PERF_TP_OUTPUT_PIXELS = 10, + PERF_TP_FILTER_WORKLOAD_16BIT = 11, + PERF_TP_FILTER_WORKLOAD_32BIT = 12, + PERF_TP_QUADS_RECEIVED = 13, + PERF_TP_QUADS_OFFSET = 14, + PERF_TP_QUADS_SHADOW = 15, + PERF_TP_QUADS_ARRAY = 16, + PERF_TP_QUADS_GRADIENT = 17, + PERF_TP_QUADS_1D = 18, + PERF_TP_QUADS_2D = 19, + PERF_TP_QUADS_BUFFER = 20, + PERF_TP_QUADS_3D = 21, + PERF_TP_QUADS_CUBE = 22, + PERF_TP_DIVERGENT_QUADS_RECEIVED = 23, + PERF_TP_PRT_NON_RESIDENT_EVENTS = 24, + PERF_TP_OUTPUT_PIXELS_POINT = 25, + PERF_TP_OUTPUT_PIXELS_BILINEAR = 26, + PERF_TP_OUTPUT_PIXELS_MIP = 27, + PERF_TP_OUTPUT_PIXELS_ANISO = 28, + PERF_TP_OUTPUT_PIXELS_ZERO_LOD = 29, + PERF_TP_FLAG_CACHE_REQUESTS = 30, + PERF_TP_FLAG_CACHE_MISSES = 31, + PERF_TP_L1_5_L2_REQUESTS = 32, + PERF_TP_2D_OUTPUT_PIXELS = 33, + PERF_TP_2D_OUTPUT_PIXELS_POINT = 34, + PERF_TP_2D_OUTPUT_PIXELS_BILINEAR = 35, + PERF_TP_2D_FILTER_WORKLOAD_16BIT = 36, + PERF_TP_2D_FILTER_WORKLOAD_32BIT = 37, + PERF_TP_TPA2TPC_TRANS = 38, + PERF_TP_L1_MISSES_ASTC_1TILE = 39, + PERF_TP_L1_MISSES_ASTC_2TILE = 40, + PERF_TP_L1_MISSES_ASTC_4TILE = 41, + PERF_TP_L1_5_L2_COMPRESS_REQS = 42, + PERF_TP_L1_5_L2_COMPRESS_MISS = 43, + PERF_TP_L1_BANK_CONFLICT = 44, + PERF_TP_L1_5_MISS_LATENCY_CYCLES = 45, + PERF_TP_L1_5_MISS_LATENCY_TRANS = 46, + PERF_TP_QUADS_CONSTANT_MULTIPLIED = 47, + PERF_TP_FRONTEND_WORKING_CYCLES = 48, + PERF_TP_L1_TAG_WORKING_CYCLES = 49, + PERF_TP_L1_DATA_WRITE_WORKING_CYCLES = 50, + PERF_TP_PRE_L1_DECOM_WORKING_CYCLES = 51, + PERF_TP_BACKEND_WORKING_CYCLES = 52, + PERF_TP_FLAG_CACHE_WORKING_CYCLES = 53, + PERF_TP_L1_5_CACHE_WORKING_CYCLES = 54, + PERF_TP_STARVE_CYCLES_SP = 55, + PERF_TP_STARVE_CYCLES_UCHE = 56, +}; + +enum a6xx_sp_perfcounter_select { + PERF_SP_BUSY_CYCLES = 0, + PERF_SP_ALU_WORKING_CYCLES = 1, + PERF_SP_EFU_WORKING_CYCLES = 2, + PERF_SP_STALL_CYCLES_VPC = 3, + PERF_SP_STALL_CYCLES_TP = 4, + PERF_SP_STALL_CYCLES_UCHE = 5, + PERF_SP_STALL_CYCLES_RB = 6, + PERF_SP_NON_EXECUTION_CYCLES = 7, + PERF_SP_WAVE_CONTEXTS = 8, + PERF_SP_WAVE_CONTEXT_CYCLES = 9, + PERF_SP_FS_STAGE_WAVE_CYCLES = 10, + PERF_SP_FS_STAGE_WAVE_SAMPLES = 11, + PERF_SP_VS_STAGE_WAVE_CYCLES = 12, + PERF_SP_VS_STAGE_WAVE_SAMPLES = 13, + PERF_SP_FS_STAGE_DURATION_CYCLES = 14, + PERF_SP_VS_STAGE_DURATION_CYCLES = 15, + PERF_SP_WAVE_CTRL_CYCLES = 16, + PERF_SP_WAVE_LOAD_CYCLES = 17, + PERF_SP_WAVE_EMIT_CYCLES = 18, + PERF_SP_WAVE_NOP_CYCLES = 19, + PERF_SP_WAVE_WAIT_CYCLES = 20, + PERF_SP_WAVE_FETCH_CYCLES = 21, + PERF_SP_WAVE_IDLE_CYCLES = 22, + PERF_SP_WAVE_END_CYCLES = 23, + PERF_SP_WAVE_LONG_SYNC_CYCLES = 24, + PERF_SP_WAVE_SHORT_SYNC_CYCLES = 25, + PERF_SP_WAVE_JOIN_CYCLES = 26, + PERF_SP_LM_LOAD_INSTRUCTIONS = 27, + PERF_SP_LM_STORE_INSTRUCTIONS = 28, + PERF_SP_LM_ATOMICS = 29, + PERF_SP_GM_LOAD_INSTRUCTIONS = 30, + PERF_SP_GM_STORE_INSTRUCTIONS = 31, + PERF_SP_GM_ATOMICS = 32, + PERF_SP_VS_STAGE_TEX_INSTRUCTIONS = 33, + PERF_SP_VS_STAGE_EFU_INSTRUCTIONS = 34, + PERF_SP_VS_STAGE_FULL_ALU_INSTRUCTIONS = 35, + PERF_SP_VS_STAGE_HALF_ALU_INSTRUCTIONS = 36, + PERF_SP_FS_STAGE_TEX_INSTRUCTIONS = 37, + PERF_SP_FS_STAGE_CFLOW_INSTRUCTIONS = 38, + PERF_SP_FS_STAGE_EFU_INSTRUCTIONS = 39, + PERF_SP_FS_STAGE_FULL_ALU_INSTRUCTIONS = 40, + PERF_SP_FS_STAGE_HALF_ALU_INSTRUCTIONS = 41, + PERF_SP_FS_STAGE_BARY_INSTRUCTIONS = 42, + PERF_SP_VS_INSTRUCTIONS = 43, + PERF_SP_FS_INSTRUCTIONS = 44, + PERF_SP_ADDR_LOCK_COUNT = 45, + PERF_SP_UCHE_READ_TRANS = 46, + PERF_SP_UCHE_WRITE_TRANS = 47, + PERF_SP_EXPORT_VPC_TRANS = 48, + PERF_SP_EXPORT_RB_TRANS = 49, + PERF_SP_PIXELS_KILLED = 50, + PERF_SP_ICL1_REQUESTS = 51, + PERF_SP_ICL1_MISSES = 52, + PERF_SP_HS_INSTRUCTIONS = 53, + PERF_SP_DS_INSTRUCTIONS = 54, + PERF_SP_GS_INSTRUCTIONS = 55, + PERF_SP_CS_INSTRUCTIONS = 56, + PERF_SP_GPR_READ = 57, + PERF_SP_GPR_WRITE = 58, + PERF_SP_FS_STAGE_HALF_EFU_INSTRUCTIONS = 59, + PERF_SP_VS_STAGE_HALF_EFU_INSTRUCTIONS = 60, + PERF_SP_LM_BANK_CONFLICTS = 61, + PERF_SP_TEX_CONTROL_WORKING_CYCLES = 62, + PERF_SP_LOAD_CONTROL_WORKING_CYCLES = 63, + PERF_SP_FLOW_CONTROL_WORKING_CYCLES = 64, + PERF_SP_LM_WORKING_CYCLES = 65, + PERF_SP_DISPATCHER_WORKING_CYCLES = 66, + PERF_SP_SEQUENCER_WORKING_CYCLES = 67, + PERF_SP_LOW_EFFICIENCY_STARVED_BY_TP = 68, + PERF_SP_STARVE_CYCLES_HLSQ = 69, + PERF_SP_NON_EXECUTION_LS_CYCLES = 70, + PERF_SP_WORKING_EU = 71, + PERF_SP_ANY_EU_WORKING = 72, + PERF_SP_WORKING_EU_FS_STAGE = 73, + PERF_SP_ANY_EU_WORKING_FS_STAGE = 74, + PERF_SP_WORKING_EU_VS_STAGE = 75, + PERF_SP_ANY_EU_WORKING_VS_STAGE = 76, + PERF_SP_WORKING_EU_CS_STAGE = 77, + PERF_SP_ANY_EU_WORKING_CS_STAGE = 78, + PERF_SP_GPR_READ_PREFETCH = 79, + PERF_SP_GPR_READ_CONFLICT = 80, + PERF_SP_GPR_WRITE_CONFLICT = 81, + PERF_SP_GM_LOAD_LATENCY_CYCLES = 82, + PERF_SP_GM_LOAD_LATENCY_SAMPLES = 83, + PERF_SP_EXECUTABLE_WAVES = 84, +}; + +enum a6xx_rb_perfcounter_select { + PERF_RB_BUSY_CYCLES = 0, + PERF_RB_STALL_CYCLES_HLSQ = 1, + PERF_RB_STALL_CYCLES_FIFO0_FULL = 2, + PERF_RB_STALL_CYCLES_FIFO1_FULL = 3, + PERF_RB_STALL_CYCLES_FIFO2_FULL = 4, + PERF_RB_STARVE_CYCLES_SP = 5, + PERF_RB_STARVE_CYCLES_LRZ_TILE = 6, + PERF_RB_STARVE_CYCLES_CCU = 7, + PERF_RB_STARVE_CYCLES_Z_PLANE = 8, + PERF_RB_STARVE_CYCLES_BARY_PLANE = 9, + PERF_RB_Z_WORKLOAD = 10, + PERF_RB_HLSQ_ACTIVE = 11, + PERF_RB_Z_READ = 12, + PERF_RB_Z_WRITE = 13, + PERF_RB_C_READ = 14, + PERF_RB_C_WRITE = 15, + PERF_RB_TOTAL_PASS = 16, + PERF_RB_Z_PASS = 17, + PERF_RB_Z_FAIL = 18, + PERF_RB_S_FAIL = 19, + PERF_RB_BLENDED_FXP_COMPONENTS = 20, + PERF_RB_BLENDED_FP16_COMPONENTS = 21, + PERF_RB_PS_INVOCATIONS = 22, + PERF_RB_2D_ALIVE_CYCLES = 23, + PERF_RB_2D_STALL_CYCLES_A2D = 24, + PERF_RB_2D_STARVE_CYCLES_SRC = 25, + PERF_RB_2D_STARVE_CYCLES_SP = 26, + PERF_RB_2D_STARVE_CYCLES_DST = 27, + PERF_RB_2D_VALID_PIXELS = 28, + PERF_RB_3D_PIXELS = 29, + PERF_RB_BLENDER_WORKING_CYCLES = 30, + PERF_RB_ZPROC_WORKING_CYCLES = 31, + PERF_RB_CPROC_WORKING_CYCLES = 32, + PERF_RB_SAMPLER_WORKING_CYCLES = 33, + PERF_RB_STALL_CYCLES_CCU_COLOR_READ = 34, + PERF_RB_STALL_CYCLES_CCU_COLOR_WRITE = 35, + PERF_RB_STALL_CYCLES_CCU_DEPTH_READ = 36, + PERF_RB_STALL_CYCLES_CCU_DEPTH_WRITE = 37, + PERF_RB_STALL_CYCLES_VPC = 38, + PERF_RB_2D_INPUT_TRANS = 39, + PERF_RB_2D_OUTPUT_RB_DST_TRANS = 40, + PERF_RB_2D_OUTPUT_RB_SRC_TRANS = 41, + PERF_RB_BLENDED_FP32_COMPONENTS = 42, + PERF_RB_COLOR_PIX_TILES = 43, + PERF_RB_STALL_CYCLES_CCU = 44, + PERF_RB_EARLY_Z_ARB3_GRANT = 45, + PERF_RB_LATE_Z_ARB3_GRANT = 46, + PERF_RB_EARLY_Z_SKIP_GRANT = 47, +}; + +enum a6xx_vsc_perfcounter_select { + PERF_VSC_BUSY_CYCLES = 0, + PERF_VSC_WORKING_CYCLES = 1, + PERF_VSC_STALL_CYCLES_UCHE = 2, + PERF_VSC_EOT_NUM = 3, + PERF_VSC_INPUT_TILES = 4, +}; + +enum a6xx_ccu_perfcounter_select { + PERF_CCU_BUSY_CYCLES = 0, + PERF_CCU_STALL_CYCLES_RB_DEPTH_RETURN = 1, + PERF_CCU_STALL_CYCLES_RB_COLOR_RETURN = 2, + PERF_CCU_STARVE_CYCLES_FLAG_RETURN = 3, + PERF_CCU_DEPTH_BLOCKS = 4, + PERF_CCU_COLOR_BLOCKS = 5, + PERF_CCU_DEPTH_BLOCK_HIT = 6, + PERF_CCU_COLOR_BLOCK_HIT = 7, + PERF_CCU_PARTIAL_BLOCK_READ = 8, + PERF_CCU_GMEM_READ = 9, + PERF_CCU_GMEM_WRITE = 10, + PERF_CCU_DEPTH_READ_FLAG0_COUNT = 11, + PERF_CCU_DEPTH_READ_FLAG1_COUNT = 12, + PERF_CCU_DEPTH_READ_FLAG2_COUNT = 13, + PERF_CCU_DEPTH_READ_FLAG3_COUNT = 14, + PERF_CCU_DEPTH_READ_FLAG4_COUNT = 15, + PERF_CCU_DEPTH_READ_FLAG5_COUNT = 16, + PERF_CCU_DEPTH_READ_FLAG6_COUNT = 17, + PERF_CCU_DEPTH_READ_FLAG8_COUNT = 18, + PERF_CCU_COLOR_READ_FLAG0_COUNT = 19, + PERF_CCU_COLOR_READ_FLAG1_COUNT = 20, + PERF_CCU_COLOR_READ_FLAG2_COUNT = 21, + PERF_CCU_COLOR_READ_FLAG3_COUNT = 22, + PERF_CCU_COLOR_READ_FLAG4_COUNT = 23, + PERF_CCU_COLOR_READ_FLAG5_COUNT = 24, + PERF_CCU_COLOR_READ_FLAG6_COUNT = 25, + PERF_CCU_COLOR_READ_FLAG8_COUNT = 26, + PERF_CCU_2D_RD_REQ = 27, + PERF_CCU_2D_WR_REQ = 28, +}; + +enum a6xx_lrz_perfcounter_select { + PERF_LRZ_BUSY_CYCLES = 0, + PERF_LRZ_STARVE_CYCLES_RAS = 1, + PERF_LRZ_STALL_CYCLES_RB = 2, + PERF_LRZ_STALL_CYCLES_VSC = 3, + PERF_LRZ_STALL_CYCLES_VPC = 4, + PERF_LRZ_STALL_CYCLES_FLAG_PREFETCH = 5, + PERF_LRZ_STALL_CYCLES_UCHE = 6, + PERF_LRZ_LRZ_READ = 7, + PERF_LRZ_LRZ_WRITE = 8, + PERF_LRZ_READ_LATENCY = 9, + PERF_LRZ_MERGE_CACHE_UPDATING = 10, + PERF_LRZ_PRIM_KILLED_BY_MASKGEN = 11, + PERF_LRZ_PRIM_KILLED_BY_LRZ = 12, + PERF_LRZ_VISIBLE_PRIM_AFTER_LRZ = 13, + PERF_LRZ_FULL_8X8_TILES = 14, + PERF_LRZ_PARTIAL_8X8_TILES = 15, + PERF_LRZ_TILE_KILLED = 16, + PERF_LRZ_TOTAL_PIXEL = 17, + PERF_LRZ_VISIBLE_PIXEL_AFTER_LRZ = 18, + PERF_LRZ_FULLY_COVERED_TILES = 19, + PERF_LRZ_PARTIAL_COVERED_TILES = 20, + PERF_LRZ_FEEDBACK_ACCEPT = 21, + PERF_LRZ_FEEDBACK_DISCARD = 22, + PERF_LRZ_FEEDBACK_STALL = 23, + PERF_LRZ_STALL_CYCLES_RB_ZPLANE = 24, + PERF_LRZ_STALL_CYCLES_RB_BPLANE = 25, + PERF_LRZ_STALL_CYCLES_VC = 26, + PERF_LRZ_RAS_MASK_TRANS = 27, +}; + +enum a6xx_cmp_perfcounter_select { + PERF_CMPDECMP_STALL_CYCLES_ARB = 0, + PERF_CMPDECMP_VBIF_LATENCY_CYCLES = 1, + PERF_CMPDECMP_VBIF_LATENCY_SAMPLES = 2, + PERF_CMPDECMP_VBIF_READ_DATA_CCU = 3, + PERF_CMPDECMP_VBIF_WRITE_DATA_CCU = 4, + PERF_CMPDECMP_VBIF_READ_REQUEST = 5, + PERF_CMPDECMP_VBIF_WRITE_REQUEST = 6, + PERF_CMPDECMP_VBIF_READ_DATA = 7, + PERF_CMPDECMP_VBIF_WRITE_DATA = 8, + PERF_CMPDECMP_FLAG_FETCH_CYCLES = 9, + PERF_CMPDECMP_FLAG_FETCH_SAMPLES = 10, + PERF_CMPDECMP_DEPTH_WRITE_FLAG1_COUNT = 11, + PERF_CMPDECMP_DEPTH_WRITE_FLAG2_COUNT = 12, + PERF_CMPDECMP_DEPTH_WRITE_FLAG3_COUNT = 13, + PERF_CMPDECMP_DEPTH_WRITE_FLAG4_COUNT = 14, + PERF_CMPDECMP_DEPTH_WRITE_FLAG5_COUNT = 15, + PERF_CMPDECMP_DEPTH_WRITE_FLAG6_COUNT = 16, + PERF_CMPDECMP_DEPTH_WRITE_FLAG8_COUNT = 17, + PERF_CMPDECMP_COLOR_WRITE_FLAG1_COUNT = 18, + PERF_CMPDECMP_COLOR_WRITE_FLAG2_COUNT = 19, + PERF_CMPDECMP_COLOR_WRITE_FLAG3_COUNT = 20, + PERF_CMPDECMP_COLOR_WRITE_FLAG4_COUNT = 21, + PERF_CMPDECMP_COLOR_WRITE_FLAG5_COUNT = 22, + PERF_CMPDECMP_COLOR_WRITE_FLAG6_COUNT = 23, + PERF_CMPDECMP_COLOR_WRITE_FLAG8_COUNT = 24, + PERF_CMPDECMP_2D_STALL_CYCLES_VBIF_REQ = 25, + PERF_CMPDECMP_2D_STALL_CYCLES_VBIF_WR = 26, + PERF_CMPDECMP_2D_STALL_CYCLES_VBIF_RETURN = 27, + PERF_CMPDECMP_2D_RD_DATA = 28, + PERF_CMPDECMP_2D_WR_DATA = 29, + PERF_CMPDECMP_VBIF_READ_DATA_UCHE_CH0 = 30, + PERF_CMPDECMP_VBIF_READ_DATA_UCHE_CH1 = 31, + PERF_CMPDECMP_2D_OUTPUT_TRANS = 32, + PERF_CMPDECMP_VBIF_WRITE_DATA_UCHE = 33, + PERF_CMPDECMP_DEPTH_WRITE_FLAG0_COUNT = 34, + PERF_CMPDECMP_COLOR_WRITE_FLAG0_COUNT = 35, + PERF_CMPDECMP_COLOR_WRITE_FLAGALPHA_COUNT = 36, + PERF_CMPDECMP_2D_BUSY_CYCLES = 37, + PERF_CMPDECMP_2D_REORDER_STARVE_CYCLES = 38, + PERF_CMPDECMP_2D_PIXELS = 39, }; enum a6xx_tex_filter { @@ -1765,12 +2444,39 @@ static inline uint32_t A6XX_UCHE_CLIENT_PF_PERFSEL(uint32_t val) #define REG_A6XX_VBIF_VERSION 0x00003000 +#define REG_A6XX_VBIF_CLKON 0x00003001 +#define A6XX_VBIF_CLKON_FORCE_ON_TESTBUS 0x00000002 + #define REG_A6XX_VBIF_GATE_OFF_WRREQ_EN 0x0000302a #define REG_A6XX_VBIF_XIN_HALT_CTRL0 0x00003080 #define REG_A6XX_VBIF_XIN_HALT_CTRL1 0x00003081 +#define REG_A6XX_VBIF_TEST_BUS_OUT_CTRL 0x00003084 + +#define REG_A6XX_VBIF_TEST_BUS1_CTRL0 0x00003085 + +#define REG_A6XX_VBIF_TEST_BUS1_CTRL1 0x00003086 +#define A6XX_VBIF_TEST_BUS1_CTRL1_DATA_SEL__MASK 0x0000000f +#define A6XX_VBIF_TEST_BUS1_CTRL1_DATA_SEL__SHIFT 0 +static inline uint32_t A6XX_VBIF_TEST_BUS1_CTRL1_DATA_SEL(uint32_t val) +{ + return ((val) << A6XX_VBIF_TEST_BUS1_CTRL1_DATA_SEL__SHIFT) & A6XX_VBIF_TEST_BUS1_CTRL1_DATA_SEL__MASK; +} + +#define REG_A6XX_VBIF_TEST_BUS2_CTRL0 0x00003087 + +#define REG_A6XX_VBIF_TEST_BUS2_CTRL1 0x00003088 +#define A6XX_VBIF_TEST_BUS2_CTRL1_DATA_SEL__MASK 0x000001ff +#define A6XX_VBIF_TEST_BUS2_CTRL1_DATA_SEL__SHIFT 0 +static inline uint32_t A6XX_VBIF_TEST_BUS2_CTRL1_DATA_SEL(uint32_t val) +{ + return ((val) << A6XX_VBIF_TEST_BUS2_CTRL1_DATA_SEL__SHIFT) & A6XX_VBIF_TEST_BUS2_CTRL1_DATA_SEL__MASK; +} + +#define REG_A6XX_VBIF_TEST_BUS_OUT 0x0000308c + #define REG_A6XX_VBIF_PERF_CNT_SEL0 0x000030d0 #define REG_A6XX_VBIF_PERF_CNT_SEL1 0x000030d1 @@ -1813,313 +2519,79 @@ static inline uint32_t A6XX_UCHE_CLIENT_PF_PERFSEL(uint32_t val) #define REG_A6XX_VBIF_PERF_PWR_CNT_HIGH2 0x0000311a -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_SEL_A 0x00018400 - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_SEL_B 0x00018401 - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_SEL_C 0x00018402 - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_SEL_D 0x00018403 -#define A6XX_CX_DBGC_CFG_DBGBUS_SEL_D_PING_INDEX__MASK 0x000000ff -#define A6XX_CX_DBGC_CFG_DBGBUS_SEL_D_PING_INDEX__SHIFT 0 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_SEL_D_PING_INDEX(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_SEL_D_PING_INDEX__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_SEL_D_PING_INDEX__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_SEL_D_PING_BLK_SEL__MASK 0x0000ff00 -#define A6XX_CX_DBGC_CFG_DBGBUS_SEL_D_PING_BLK_SEL__SHIFT 8 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_SEL_D_PING_BLK_SEL(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_SEL_D_PING_BLK_SEL__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_SEL_D_PING_BLK_SEL__MASK; -} - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_CNTLT 0x00018404 -#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_TRACEEN__MASK 0x0000003f -#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_TRACEEN__SHIFT 0 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_TRACEEN(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_TRACEEN__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_TRACEEN__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_GRANU__MASK 0x00007000 -#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_GRANU__SHIFT 12 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_GRANU(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_GRANU__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_GRANU__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_SEGT__MASK 0xf0000000 -#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_SEGT__SHIFT 28 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_SEGT(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_SEGT__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_SEGT__MASK; -} - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_CNTLM 0x00018405 -#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLM_ENABLE__MASK 0x0f000000 -#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLM_ENABLE__SHIFT 24 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_CNTLM_ENABLE(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_CNTLM_ENABLE__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_CNTLM_ENABLE__MASK; -} - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_IVTL_0 0x00018408 - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_IVTL_1 0x00018409 - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_IVTL_2 0x0001840a - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_IVTL_3 0x0001840b - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_MASKL_0 0x0001840c - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_MASKL_1 0x0001840d - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_MASKL_2 0x0001840e - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_MASKL_3 0x0001840f - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0 0x00018410 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL0__MASK 0x0000000f -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL0__SHIFT 0 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL0(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL0__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL0__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL1__MASK 0x000000f0 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL1__SHIFT 4 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL1(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL1__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL1__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL2__MASK 0x00000f00 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL2__SHIFT 8 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL2(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL2__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL2__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL3__MASK 0x0000f000 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL3__SHIFT 12 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL3(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL3__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL3__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL4__MASK 0x000f0000 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL4__SHIFT 16 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL4(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL4__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL4__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL5__MASK 0x00f00000 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL5__SHIFT 20 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL5(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL5__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL5__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL6__MASK 0x0f000000 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL6__SHIFT 24 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL6(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL6__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL6__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL7__MASK 0xf0000000 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL7__SHIFT 28 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL7(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL7__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL7__MASK; -} - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1 0x00018411 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL8__MASK 0x0000000f -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL8__SHIFT 0 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL8(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL8__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL8__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL9__MASK 0x000000f0 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL9__SHIFT 4 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL9(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL9__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL9__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL10__MASK 0x00000f00 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL10__SHIFT 8 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL10(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL10__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL10__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL11__MASK 0x0000f000 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL11__SHIFT 12 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL11(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL11__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL11__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL12__MASK 0x000f0000 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL12__SHIFT 16 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL12(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL12__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL12__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL13__MASK 0x00f00000 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL13__SHIFT 20 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL13(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL13__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL13__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL14__MASK 0x0f000000 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL14__SHIFT 24 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL14(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL14__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL14__MASK; -} -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL15__MASK 0xf0000000 -#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL15__SHIFT 28 -static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL15(uint32_t val) -{ - return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL15__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL15__MASK; -} - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_TRACE_BUF1 0x0001842f - -#define REG_A6XX_CX_DBGC_CFG_DBGBUS_TRACE_BUF2 0x00018430 - -#define REG_A6XX_PDC_GPU_ENABLE_PDC 0x00021140 - -#define REG_A6XX_PDC_GPU_SEQ_START_ADDR 0x00021148 - -#define REG_A6XX_PDC_GPU_TCS0_CONTROL 0x00021540 - -#define REG_A6XX_PDC_GPU_TCS0_CMD_ENABLE_BANK 0x00021541 - -#define REG_A6XX_PDC_GPU_TCS0_CMD_WAIT_FOR_CMPL_BANK 0x00021542 - -#define REG_A6XX_PDC_GPU_TCS0_CMD0_MSGID 0x00021543 - -#define REG_A6XX_PDC_GPU_TCS0_CMD0_ADDR 0x00021544 - -#define REG_A6XX_PDC_GPU_TCS0_CMD0_DATA 0x00021545 - -#define REG_A6XX_PDC_GPU_TCS1_CONTROL 0x00021572 - -#define REG_A6XX_PDC_GPU_TCS1_CMD_ENABLE_BANK 0x00021573 - -#define REG_A6XX_PDC_GPU_TCS1_CMD_WAIT_FOR_CMPL_BANK 0x00021574 - -#define REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID 0x00021575 - -#define REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR 0x00021576 - -#define REG_A6XX_PDC_GPU_TCS1_CMD0_DATA 0x00021577 - -#define REG_A6XX_PDC_GPU_TCS2_CONTROL 0x000215a4 - -#define REG_A6XX_PDC_GPU_TCS2_CMD_ENABLE_BANK 0x000215a5 - -#define REG_A6XX_PDC_GPU_TCS2_CMD_WAIT_FOR_CMPL_BANK 0x000215a6 - -#define REG_A6XX_PDC_GPU_TCS2_CMD0_MSGID 0x000215a7 - -#define REG_A6XX_PDC_GPU_TCS2_CMD0_ADDR 0x000215a8 - -#define REG_A6XX_PDC_GPU_TCS2_CMD0_DATA 0x000215a9 - -#define REG_A6XX_PDC_GPU_TCS3_CONTROL 0x000215d6 - -#define REG_A6XX_PDC_GPU_TCS3_CMD_ENABLE_BANK 0x000215d7 - -#define REG_A6XX_PDC_GPU_TCS3_CMD_WAIT_FOR_CMPL_BANK 0x000215d8 - -#define REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID 0x000215d9 - -#define REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR 0x000215da - -#define REG_A6XX_PDC_GPU_TCS3_CMD0_DATA 0x000215db - -#define REG_A6XX_PDC_GPU_SEQ_MEM_0 0x000a0000 - -#define REG_A6XX_X1_WINDOW_OFFSET 0x000088d4 -#define A6XX_X1_WINDOW_OFFSET_WINDOW_OFFSET_DISABLE 0x80000000 -#define A6XX_X1_WINDOW_OFFSET_X__MASK 0x00007fff -#define A6XX_X1_WINDOW_OFFSET_X__SHIFT 0 -static inline uint32_t A6XX_X1_WINDOW_OFFSET_X(uint32_t val) -{ - return ((val) << A6XX_X1_WINDOW_OFFSET_X__SHIFT) & A6XX_X1_WINDOW_OFFSET_X__MASK; -} -#define A6XX_X1_WINDOW_OFFSET_Y__MASK 0x7fff0000 -#define A6XX_X1_WINDOW_OFFSET_Y__SHIFT 16 -static inline uint32_t A6XX_X1_WINDOW_OFFSET_Y(uint32_t val) -{ - return ((val) << A6XX_X1_WINDOW_OFFSET_Y__SHIFT) & A6XX_X1_WINDOW_OFFSET_Y__MASK; -} - -#define REG_A6XX_X2_WINDOW_OFFSET 0x0000b4d1 -#define A6XX_X2_WINDOW_OFFSET_WINDOW_OFFSET_DISABLE 0x80000000 -#define A6XX_X2_WINDOW_OFFSET_X__MASK 0x00007fff -#define A6XX_X2_WINDOW_OFFSET_X__SHIFT 0 -static inline uint32_t A6XX_X2_WINDOW_OFFSET_X(uint32_t val) +#define REG_A6XX_RB_WINDOW_OFFSET2 0x000088d4 +#define A6XX_RB_WINDOW_OFFSET2_WINDOW_OFFSET_DISABLE 0x80000000 +#define A6XX_RB_WINDOW_OFFSET2_X__MASK 0x00007fff +#define A6XX_RB_WINDOW_OFFSET2_X__SHIFT 0 +static inline uint32_t A6XX_RB_WINDOW_OFFSET2_X(uint32_t val) { - return ((val) << A6XX_X2_WINDOW_OFFSET_X__SHIFT) & A6XX_X2_WINDOW_OFFSET_X__MASK; + return ((val) << A6XX_RB_WINDOW_OFFSET2_X__SHIFT) & A6XX_RB_WINDOW_OFFSET2_X__MASK; } -#define A6XX_X2_WINDOW_OFFSET_Y__MASK 0x7fff0000 -#define A6XX_X2_WINDOW_OFFSET_Y__SHIFT 16 -static inline uint32_t A6XX_X2_WINDOW_OFFSET_Y(uint32_t val) +#define A6XX_RB_WINDOW_OFFSET2_Y__MASK 0x7fff0000 +#define A6XX_RB_WINDOW_OFFSET2_Y__SHIFT 16 +static inline uint32_t A6XX_RB_WINDOW_OFFSET2_Y(uint32_t val) { - return ((val) << A6XX_X2_WINDOW_OFFSET_Y__SHIFT) & A6XX_X2_WINDOW_OFFSET_Y__MASK; + return ((val) << A6XX_RB_WINDOW_OFFSET2_Y__SHIFT) & A6XX_RB_WINDOW_OFFSET2_Y__MASK; } -#define REG_A6XX_X3_WINDOW_OFFSET 0x0000b307 -#define A6XX_X3_WINDOW_OFFSET_WINDOW_OFFSET_DISABLE 0x80000000 -#define A6XX_X3_WINDOW_OFFSET_X__MASK 0x00007fff -#define A6XX_X3_WINDOW_OFFSET_X__SHIFT 0 -static inline uint32_t A6XX_X3_WINDOW_OFFSET_X(uint32_t val) +#define REG_A6XX_SP_WINDOW_OFFSET 0x0000b4d1 +#define A6XX_SP_WINDOW_OFFSET_WINDOW_OFFSET_DISABLE 0x80000000 +#define A6XX_SP_WINDOW_OFFSET_X__MASK 0x00007fff +#define A6XX_SP_WINDOW_OFFSET_X__SHIFT 0 +static inline uint32_t A6XX_SP_WINDOW_OFFSET_X(uint32_t val) { - return ((val) << A6XX_X3_WINDOW_OFFSET_X__SHIFT) & A6XX_X3_WINDOW_OFFSET_X__MASK; + return ((val) << A6XX_SP_WINDOW_OFFSET_X__SHIFT) & A6XX_SP_WINDOW_OFFSET_X__MASK; } -#define A6XX_X3_WINDOW_OFFSET_Y__MASK 0x7fff0000 -#define A6XX_X3_WINDOW_OFFSET_Y__SHIFT 16 -static inline uint32_t A6XX_X3_WINDOW_OFFSET_Y(uint32_t val) +#define A6XX_SP_WINDOW_OFFSET_Y__MASK 0x7fff0000 +#define A6XX_SP_WINDOW_OFFSET_Y__SHIFT 16 +static inline uint32_t A6XX_SP_WINDOW_OFFSET_Y(uint32_t val) { - return ((val) << A6XX_X3_WINDOW_OFFSET_Y__SHIFT) & A6XX_X3_WINDOW_OFFSET_Y__MASK; + return ((val) << A6XX_SP_WINDOW_OFFSET_Y__SHIFT) & A6XX_SP_WINDOW_OFFSET_Y__MASK; } -#define REG_A6XX_X1_BIN_SIZE 0x000080a1 -#define A6XX_X1_BIN_SIZE_WIDTH__MASK 0x000000ff -#define A6XX_X1_BIN_SIZE_WIDTH__SHIFT 0 -static inline uint32_t A6XX_X1_BIN_SIZE_WIDTH(uint32_t val) +#define REG_A6XX_SP_TP_WINDOW_OFFSET 0x0000b307 +#define A6XX_SP_TP_WINDOW_OFFSET_WINDOW_OFFSET_DISABLE 0x80000000 +#define A6XX_SP_TP_WINDOW_OFFSET_X__MASK 0x00007fff +#define A6XX_SP_TP_WINDOW_OFFSET_X__SHIFT 0 +static inline uint32_t A6XX_SP_TP_WINDOW_OFFSET_X(uint32_t val) { - return ((val >> 5) << A6XX_X1_BIN_SIZE_WIDTH__SHIFT) & A6XX_X1_BIN_SIZE_WIDTH__MASK; + return ((val) << A6XX_SP_TP_WINDOW_OFFSET_X__SHIFT) & A6XX_SP_TP_WINDOW_OFFSET_X__MASK; } -#define A6XX_X1_BIN_SIZE_HEIGHT__MASK 0x0001ff00 -#define A6XX_X1_BIN_SIZE_HEIGHT__SHIFT 8 -static inline uint32_t A6XX_X1_BIN_SIZE_HEIGHT(uint32_t val) +#define A6XX_SP_TP_WINDOW_OFFSET_Y__MASK 0x7fff0000 +#define A6XX_SP_TP_WINDOW_OFFSET_Y__SHIFT 16 +static inline uint32_t A6XX_SP_TP_WINDOW_OFFSET_Y(uint32_t val) { - return ((val >> 4) << A6XX_X1_BIN_SIZE_HEIGHT__SHIFT) & A6XX_X1_BIN_SIZE_HEIGHT__MASK; + return ((val) << A6XX_SP_TP_WINDOW_OFFSET_Y__SHIFT) & A6XX_SP_TP_WINDOW_OFFSET_Y__MASK; } -#define REG_A6XX_X2_BIN_SIZE 0x00008800 -#define A6XX_X2_BIN_SIZE_WIDTH__MASK 0x000000ff -#define A6XX_X2_BIN_SIZE_WIDTH__SHIFT 0 -static inline uint32_t A6XX_X2_BIN_SIZE_WIDTH(uint32_t val) +#define REG_A6XX_GRAS_BIN_CONTROL 0x000080a1 +#define A6XX_GRAS_BIN_CONTROL_BINW__MASK 0x000000ff +#define A6XX_GRAS_BIN_CONTROL_BINW__SHIFT 0 +static inline uint32_t A6XX_GRAS_BIN_CONTROL_BINW(uint32_t val) { - return ((val >> 5) << A6XX_X2_BIN_SIZE_WIDTH__SHIFT) & A6XX_X2_BIN_SIZE_WIDTH__MASK; + return ((val >> 5) << A6XX_GRAS_BIN_CONTROL_BINW__SHIFT) & A6XX_GRAS_BIN_CONTROL_BINW__MASK; } -#define A6XX_X2_BIN_SIZE_HEIGHT__MASK 0x0001ff00 -#define A6XX_X2_BIN_SIZE_HEIGHT__SHIFT 8 -static inline uint32_t A6XX_X2_BIN_SIZE_HEIGHT(uint32_t val) +#define A6XX_GRAS_BIN_CONTROL_BINH__MASK 0x0001ff00 +#define A6XX_GRAS_BIN_CONTROL_BINH__SHIFT 8 +static inline uint32_t A6XX_GRAS_BIN_CONTROL_BINH(uint32_t val) { - return ((val >> 4) << A6XX_X2_BIN_SIZE_HEIGHT__SHIFT) & A6XX_X2_BIN_SIZE_HEIGHT__MASK; + return ((val >> 4) << A6XX_GRAS_BIN_CONTROL_BINH__SHIFT) & A6XX_GRAS_BIN_CONTROL_BINH__MASK; } +#define A6XX_GRAS_BIN_CONTROL_BINNING_PASS 0x00040000 +#define A6XX_GRAS_BIN_CONTROL_USE_VIZ 0x00200000 -#define REG_A6XX_X3_BIN_SIZE 0x000088d3 -#define A6XX_X3_BIN_SIZE_WIDTH__MASK 0x000000ff -#define A6XX_X3_BIN_SIZE_WIDTH__SHIFT 0 -static inline uint32_t A6XX_X3_BIN_SIZE_WIDTH(uint32_t val) +#define REG_A6XX_RB_BIN_CONTROL2 0x000088d3 +#define A6XX_RB_BIN_CONTROL2_BINW__MASK 0x000000ff +#define A6XX_RB_BIN_CONTROL2_BINW__SHIFT 0 +static inline uint32_t A6XX_RB_BIN_CONTROL2_BINW(uint32_t val) { - return ((val >> 5) << A6XX_X3_BIN_SIZE_WIDTH__SHIFT) & A6XX_X3_BIN_SIZE_WIDTH__MASK; + return ((val >> 5) << A6XX_RB_BIN_CONTROL2_BINW__SHIFT) & A6XX_RB_BIN_CONTROL2_BINW__MASK; } -#define A6XX_X3_BIN_SIZE_HEIGHT__MASK 0x0001ff00 -#define A6XX_X3_BIN_SIZE_HEIGHT__SHIFT 8 -static inline uint32_t A6XX_X3_BIN_SIZE_HEIGHT(uint32_t val) +#define A6XX_RB_BIN_CONTROL2_BINH__MASK 0x0001ff00 +#define A6XX_RB_BIN_CONTROL2_BINH__SHIFT 8 +static inline uint32_t A6XX_RB_BIN_CONTROL2_BINH(uint32_t val) { - return ((val >> 4) << A6XX_X3_BIN_SIZE_HEIGHT__SHIFT) & A6XX_X3_BIN_SIZE_HEIGHT__MASK; + return ((val >> 4) << A6XX_RB_BIN_CONTROL2_BINH__SHIFT) & A6XX_RB_BIN_CONTROL2_BINH__MASK; } #define REG_A6XX_VSC_BIN_SIZE 0x00000c02 @@ -2182,11 +2654,19 @@ static inline uint32_t A6XX_VSC_PIPE_CONFIG_REG_H(uint32_t val) return ((val) << A6XX_VSC_PIPE_CONFIG_REG_H__SHIFT) & A6XX_VSC_PIPE_CONFIG_REG_H__MASK; } -#define REG_A6XX_VSC_XXX_ADDRESS_LO 0x00000c30 +#define REG_A6XX_VSC_PIPE_DATA2_ADDRESS_LO 0x00000c30 -#define REG_A6XX_VSC_XXX_ADDRESS_HI 0x00000c31 +#define REG_A6XX_VSC_PIPE_DATA2_ADDRESS_HI 0x00000c31 -#define REG_A6XX_VSC_XXX_PITCH 0x00000c32 +#define REG_A6XX_VSC_PIPE_DATA2_PITCH 0x00000c32 + +#define REG_A6XX_VSC_PIPE_DATA2_ARRAY_PITCH 0x00000c33 +#define A6XX_VSC_PIPE_DATA2_ARRAY_PITCH__MASK 0xffffffff +#define A6XX_VSC_PIPE_DATA2_ARRAY_PITCH__SHIFT 0 +static inline uint32_t A6XX_VSC_PIPE_DATA2_ARRAY_PITCH(uint32_t val) +{ + return ((val >> 4) << A6XX_VSC_PIPE_DATA2_ARRAY_PITCH__SHIFT) & A6XX_VSC_PIPE_DATA2_ARRAY_PITCH__MASK; +} #define REG_A6XX_VSC_PIPE_DATA_ADDRESS_LO 0x00000c34 @@ -2194,18 +2674,29 @@ static inline uint32_t A6XX_VSC_PIPE_CONFIG_REG_H(uint32_t val) #define REG_A6XX_VSC_PIPE_DATA_PITCH 0x00000c36 +#define REG_A6XX_VSC_PIPE_DATA_ARRAY_PITCH 0x00000c37 +#define A6XX_VSC_PIPE_DATA_ARRAY_PITCH__MASK 0xffffffff +#define A6XX_VSC_PIPE_DATA_ARRAY_PITCH__SHIFT 0 +static inline uint32_t A6XX_VSC_PIPE_DATA_ARRAY_PITCH(uint32_t val) +{ + return ((val >> 4) << A6XX_VSC_PIPE_DATA_ARRAY_PITCH__SHIFT) & A6XX_VSC_PIPE_DATA_ARRAY_PITCH__MASK; +} + static inline uint32_t REG_A6XX_VSC_SIZE(uint32_t i0) { return 0x00000c78 + 0x1*i0; } static inline uint32_t REG_A6XX_VSC_SIZE_REG(uint32_t i0) { return 0x00000c78 + 0x1*i0; } #define REG_A6XX_UCHE_UNKNOWN_0E12 0x00000e12 +#define REG_A6XX_GRAS_UNKNOWN_8000 0x00008000 + #define REG_A6XX_GRAS_UNKNOWN_8001 0x00008001 #define REG_A6XX_GRAS_UNKNOWN_8004 0x00008004 #define REG_A6XX_GRAS_CNTL 0x00008005 #define A6XX_GRAS_CNTL_VARYING 0x00000001 +#define A6XX_GRAS_CNTL_UNK3 0x00000008 #define A6XX_GRAS_CNTL_XCOORD 0x00000040 #define A6XX_GRAS_CNTL_YCOORD 0x00000080 #define A6XX_GRAS_CNTL_ZCOORD 0x00000100 @@ -2308,6 +2799,9 @@ static inline uint32_t A6XX_GRAS_SU_POINT_SIZE(float val) return ((((int32_t)(val * 16.0))) << A6XX_GRAS_SU_POINT_SIZE__SHIFT) & A6XX_GRAS_SU_POINT_SIZE__MASK; } +#define REG_A6XX_GRAS_SU_DEPTH_PLANE_CNTL 0x00008094 +#define A6XX_GRAS_SU_DEPTH_PLANE_CNTL_FRAG_WRITES_Z 0x00000001 + #define REG_A6XX_GRAS_SU_POLY_OFFSET_SCALE 0x00008095 #define A6XX_GRAS_SU_POLY_OFFSET_SCALE__MASK 0xffffffff #define A6XX_GRAS_SU_POLY_OFFSET_SCALE__SHIFT 0 @@ -2344,6 +2838,8 @@ static inline uint32_t A6XX_GRAS_SU_DEPTH_BUFFER_INFO_DEPTH_FORMAT(enum a6xx_dep #define REG_A6XX_GRAS_UNKNOWN_809B 0x0000809b +#define REG_A6XX_GRAS_UNKNOWN_80A0 0x000080a0 + #define REG_A6XX_GRAS_RAS_MSAA_CNTL 0x000080a2 #define A6XX_GRAS_RAS_MSAA_CNTL_SAMPLES__MASK 0x00000003 #define A6XX_GRAS_RAS_MSAA_CNTL_SAMPLES__SHIFT 0 @@ -2464,6 +2960,8 @@ static inline uint32_t A6XX_GRAS_SC_WINDOW_SCISSOR_BR_Y(uint32_t val) #define A6XX_GRAS_LRZ_CNTL_LRZ_WRITE 0x00000002 #define A6XX_GRAS_LRZ_CNTL_GREATER 0x00000004 +#define REG_A6XX_GRAS_UNKNOWN_8101 0x00008101 + #define REG_A6XX_GRAS_2D_BLIT_INFO 0x00008102 #define A6XX_GRAS_2D_BLIT_INFO_COLOR_FORMAT__MASK 0x000000ff #define A6XX_GRAS_2D_BLIT_INFO_COLOR_FORMAT__SHIFT 0 @@ -2494,6 +2992,10 @@ static inline uint32_t A6XX_GRAS_LRZ_BUFFER_PITCH_ARRAY_PITCH(uint32_t val) #define REG_A6XX_GRAS_LRZ_FAST_CLEAR_BUFFER_BASE_HI 0x00008107 +#define REG_A6XX_GRAS_UNKNOWN_8109 0x00008109 + +#define REG_A6XX_GRAS_UNKNOWN_8110 0x00008110 + #define REG_A6XX_GRAS_2D_BLIT_CNTL 0x00008400 #define REG_A6XX_GRAS_2D_SRC_TL_X 0x00008401 @@ -2590,6 +3092,33 @@ static inline uint32_t A6XX_GRAS_RESOLVE_CNTL_2_Y(uint32_t val) #define REG_A6XX_GRAS_UNKNOWN_8600 0x00008600 +#define REG_A6XX_RB_BIN_CONTROL 0x00008800 +#define A6XX_RB_BIN_CONTROL_BINW__MASK 0x000000ff +#define A6XX_RB_BIN_CONTROL_BINW__SHIFT 0 +static inline uint32_t A6XX_RB_BIN_CONTROL_BINW(uint32_t val) +{ + return ((val >> 5) << A6XX_RB_BIN_CONTROL_BINW__SHIFT) & A6XX_RB_BIN_CONTROL_BINW__MASK; +} +#define A6XX_RB_BIN_CONTROL_BINH__MASK 0x0001ff00 +#define A6XX_RB_BIN_CONTROL_BINH__SHIFT 8 +static inline uint32_t A6XX_RB_BIN_CONTROL_BINH(uint32_t val) +{ + return ((val >> 4) << A6XX_RB_BIN_CONTROL_BINH__SHIFT) & A6XX_RB_BIN_CONTROL_BINH__MASK; +} +#define A6XX_RB_BIN_CONTROL_BINNING_PASS 0x00040000 +#define A6XX_RB_BIN_CONTROL_USE_VIZ 0x00200000 + +#define REG_A6XX_RB_RENDER_CNTL 0x00008801 +#define A6XX_RB_RENDER_CNTL_UNK4 0x00000010 +#define A6XX_RB_RENDER_CNTL_BINNING 0x00000080 +#define A6XX_RB_RENDER_CNTL_FLAG_DEPTH 0x00004000 +#define A6XX_RB_RENDER_CNTL_FLAG_MRTS__MASK 0x00ff0000 +#define A6XX_RB_RENDER_CNTL_FLAG_MRTS__SHIFT 16 +static inline uint32_t A6XX_RB_RENDER_CNTL_FLAG_MRTS(uint32_t val) +{ + return ((val) << A6XX_RB_RENDER_CNTL_FLAG_MRTS__SHIFT) & A6XX_RB_RENDER_CNTL_FLAG_MRTS__MASK; +} + #define REG_A6XX_RB_RAS_MSAA_CNTL 0x00008802 #define A6XX_RB_RAS_MSAA_CNTL_SAMPLES__MASK 0x00000003 #define A6XX_RB_RAS_MSAA_CNTL_SAMPLES__SHIFT 0 @@ -2615,6 +3144,7 @@ static inline uint32_t A6XX_RB_DEST_MSAA_CNTL_SAMPLES(enum a3xx_msaa_samples val #define REG_A6XX_RB_RENDER_CONTROL0 0x00008809 #define A6XX_RB_RENDER_CONTROL0_VARYING 0x00000001 +#define A6XX_RB_RENDER_CONTROL0_UNK3 0x00000008 #define A6XX_RB_RENDER_CONTROL0_XCOORD 0x00000040 #define A6XX_RB_RENDER_CONTROL0_YCOORD 0x00000080 #define A6XX_RB_RENDER_CONTROL0_ZCOORD 0x00000100 @@ -2747,6 +3277,10 @@ static inline uint32_t A6XX_RB_DITHER_CNTL_DITHER_MODE_MRT7(enum adreno_rb_dithe #define A6XX_RB_SRGB_CNTL_SRGB_MRT6 0x00000040 #define A6XX_RB_SRGB_CNTL_SRGB_MRT7 0x00000080 +#define REG_A6XX_RB_UNKNOWN_8810 0x00008810 + +#define REG_A6XX_RB_UNKNOWN_8811 0x00008811 + #define REG_A6XX_RB_UNKNOWN_8818 0x00008818 #define REG_A6XX_RB_UNKNOWN_8819 0x00008819 @@ -2837,7 +3371,6 @@ static inline uint32_t A6XX_RB_MRT_BUF_INFO_COLOR_SWAP(enum a3xx_color_swap val) { return ((val) << A6XX_RB_MRT_BUF_INFO_COLOR_SWAP__SHIFT) & A6XX_RB_MRT_BUF_INFO_COLOR_SWAP__MASK; } -#define A6XX_RB_MRT_BUF_INFO_COLOR_SRGB 0x00008000 static inline uint32_t REG_A6XX_RB_MRT_PITCH(uint32_t i0) { return 0x00008823 + 0x8*i0; } #define A6XX_RB_MRT_PITCH__MASK 0xffffffff @@ -2923,6 +3456,9 @@ static inline uint32_t A6XX_RB_BLEND_CNTL_SAMPLE_MASK(uint32_t val) return ((val) << A6XX_RB_BLEND_CNTL_SAMPLE_MASK__SHIFT) & A6XX_RB_BLEND_CNTL_SAMPLE_MASK__MASK; } +#define REG_A6XX_RB_DEPTH_PLANE_CNTL 0x00008870 +#define A6XX_RB_DEPTH_PLANE_CNTL_FRAG_WRITES_Z 0x00000001 + #define REG_A6XX_RB_DEPTH_CNTL 0x00008871 #define A6XX_RB_DEPTH_CNTL_Z_ENABLE 0x00000001 #define A6XX_RB_DEPTH_CNTL_Z_WRITE_ENABLE 0x00000002 @@ -3053,6 +3589,12 @@ static inline uint32_t A6XX_RB_STENCILREF_REF(uint32_t val) { return ((val) << A6XX_RB_STENCILREF_REF__SHIFT) & A6XX_RB_STENCILREF_REF__MASK; } +#define A6XX_RB_STENCILREF_BFREF__MASK 0x0000ff00 +#define A6XX_RB_STENCILREF_BFREF__SHIFT 8 +static inline uint32_t A6XX_RB_STENCILREF_BFREF(uint32_t val) +{ + return ((val) << A6XX_RB_STENCILREF_BFREF__SHIFT) & A6XX_RB_STENCILREF_BFREF__MASK; +} #define REG_A6XX_RB_STENCILMASK 0x00008888 #define A6XX_RB_STENCILMASK_MASK__MASK 0x000000ff @@ -3061,6 +3603,12 @@ static inline uint32_t A6XX_RB_STENCILMASK_MASK(uint32_t val) { return ((val) << A6XX_RB_STENCILMASK_MASK__SHIFT) & A6XX_RB_STENCILMASK_MASK__MASK; } +#define A6XX_RB_STENCILMASK_BFMASK__MASK 0x0000ff00 +#define A6XX_RB_STENCILMASK_BFMASK__SHIFT 8 +static inline uint32_t A6XX_RB_STENCILMASK_BFMASK(uint32_t val) +{ + return ((val) << A6XX_RB_STENCILMASK_BFMASK__SHIFT) & A6XX_RB_STENCILMASK_BFMASK__MASK; +} #define REG_A6XX_RB_STENCILWRMASK 0x00008889 #define A6XX_RB_STENCILWRMASK_WRMASK__MASK 0x000000ff @@ -3069,6 +3617,12 @@ static inline uint32_t A6XX_RB_STENCILWRMASK_WRMASK(uint32_t val) { return ((val) << A6XX_RB_STENCILWRMASK_WRMASK__SHIFT) & A6XX_RB_STENCILWRMASK_WRMASK__MASK; } +#define A6XX_RB_STENCILWRMASK_BFWRMASK__MASK 0x0000ff00 +#define A6XX_RB_STENCILWRMASK_BFWRMASK__SHIFT 8 +static inline uint32_t A6XX_RB_STENCILWRMASK_BFWRMASK(uint32_t val) +{ + return ((val) << A6XX_RB_STENCILWRMASK_BFWRMASK__SHIFT) & A6XX_RB_STENCILWRMASK_BFWRMASK__MASK; +} #define REG_A6XX_RB_WINDOW_OFFSET 0x00008890 #define A6XX_RB_WINDOW_OFFSET_WINDOW_OFFSET_DISABLE 0x80000000 @@ -3177,14 +3731,14 @@ static inline uint32_t A6XX_RB_BLIT_DST_ARRAY_PITCH(uint32_t val) #define REG_A6XX_RB_BLIT_INFO 0x000088e3 #define A6XX_RB_BLIT_INFO_UNK0 0x00000001 -#define A6XX_RB_BLIT_INFO_FAST_CLEAR 0x00000002 +#define A6XX_RB_BLIT_INFO_GMEM 0x00000002 #define A6XX_RB_BLIT_INFO_INTEGER 0x00000004 -#define A6XX_RB_BLIT_INFO_UNK3 0x00000008 -#define A6XX_RB_BLIT_INFO_MASK__MASK 0x000000f0 -#define A6XX_RB_BLIT_INFO_MASK__SHIFT 4 -static inline uint32_t A6XX_RB_BLIT_INFO_MASK(uint32_t val) +#define A6XX_RB_BLIT_INFO_DEPTH 0x00000008 +#define A6XX_RB_BLIT_INFO_CLEAR_MASK__MASK 0x000000f0 +#define A6XX_RB_BLIT_INFO_CLEAR_MASK__SHIFT 4 +static inline uint32_t A6XX_RB_BLIT_INFO_CLEAR_MASK(uint32_t val) { - return ((val) << A6XX_RB_BLIT_INFO_MASK__SHIFT) & A6XX_RB_BLIT_INFO_MASK__MASK; + return ((val) << A6XX_RB_BLIT_INFO_CLEAR_MASK__SHIFT) & A6XX_RB_BLIT_INFO_CLEAR_MASK__MASK; } #define REG_A6XX_RB_UNKNOWN_88F0 0x000088f0 @@ -3274,12 +3828,16 @@ static inline uint32_t A6XX_RB_2D_DST_SIZE_PITCH(uint32_t val) #define REG_A6XX_RB_UNKNOWN_8E01 0x00008e01 +#define REG_A6XX_RB_UNKNOWN_8E04 0x00008e04 + #define REG_A6XX_RB_CCU_CNTL 0x00008e07 #define REG_A6XX_VPC_UNKNOWN_9101 0x00009101 #define REG_A6XX_VPC_GS_SIV_CNTL 0x00009104 +#define REG_A6XX_VPC_UNKNOWN_9107 0x00009107 + #define REG_A6XX_VPC_UNKNOWN_9108 0x00009108 static inline uint32_t REG_A6XX_VPC_VARYING_INTERP(uint32_t i0) { return 0x00009200 + 0x1*i0; } @@ -3385,6 +3943,9 @@ static inline uint32_t A6XX_VPC_CNTL_0_NUMNONPOSVAR(uint32_t val) #define A6XX_VPC_SO_BUF_CNTL_BUF3 0x00000200 #define A6XX_VPC_SO_BUF_CNTL_ENABLE 0x00008000 +#define REG_A6XX_VPC_SO_OVERRIDE 0x00009306 +#define A6XX_VPC_SO_OVERRIDE_SO_DISABLE 0x00000001 + #define REG_A6XX_VPC_UNKNOWN_9600 0x00009600 #define REG_A6XX_VPC_UNKNOWN_9602 0x00009602 @@ -3397,8 +3958,14 @@ static inline uint32_t A6XX_VPC_CNTL_0_NUMNONPOSVAR(uint32_t val) #define REG_A6XX_PC_UNKNOWN_9805 0x00009805 +#define REG_A6XX_PC_UNKNOWN_9806 0x00009806 + +#define REG_A6XX_PC_UNKNOWN_9980 0x00009980 + #define REG_A6XX_PC_UNKNOWN_9981 0x00009981 +#define REG_A6XX_PC_UNKNOWN_9990 0x00009990 + #define REG_A6XX_PC_PRIMITIVE_CNTL_0 0x00009b00 #define A6XX_PC_PRIMITIVE_CNTL_0_PRIMITIVE_RESTART 0x00000001 #define A6XX_PC_PRIMITIVE_CNTL_0_PROVOKING_VTX_LAST 0x00000002 @@ -3410,6 +3977,7 @@ static inline uint32_t A6XX_PC_PRIMITIVE_CNTL_1_STRIDE_IN_VPC(uint32_t val) { return ((val) << A6XX_PC_PRIMITIVE_CNTL_1_STRIDE_IN_VPC__SHIFT) & A6XX_PC_PRIMITIVE_CNTL_1_STRIDE_IN_VPC__MASK; } +#define A6XX_PC_PRIMITIVE_CNTL_1_PSIZE 0x00000100 #define REG_A6XX_PC_UNKNOWN_9B06 0x00009b06 @@ -3488,6 +4056,8 @@ static inline uint32_t A6XX_VFD_CONTROL_3_REGID_TESSY(uint32_t val) #define REG_A6XX_VFD_UNKNOWN_A008 0x0000a008 +#define REG_A6XX_VFD_UNKNOWN_A009 0x0000a009 + #define REG_A6XX_VFD_INDEX_OFFSET 0x0000a00e #define REG_A6XX_VFD_INSTANCE_START_OFFSET 0x0000a00f @@ -3640,6 +4210,8 @@ static inline uint32_t A6XX_SP_VS_CTRL_REG0_THREADSIZE(enum a3xx_threadsize val) #define A6XX_SP_VS_CTRL_REG0_PIXLODENABLE 0x04000000 #define A6XX_SP_VS_CTRL_REG0_MERGEDREGS 0x80000000 +#define REG_A6XX_SP_UNKNOWN_A81B 0x0000a81b + #define REG_A6XX_SP_VS_OBJ_START_LO 0x0000a81c #define REG_A6XX_SP_VS_OBJ_START_HI 0x0000a81d @@ -3884,6 +4456,8 @@ static inline uint32_t A6XX_SP_FS_CTRL_REG0_THREADSIZE(enum a3xx_threadsize val) #define A6XX_SP_FS_CTRL_REG0_PIXLODENABLE 0x04000000 #define A6XX_SP_FS_CTRL_REG0_MERGEDREGS 0x80000000 +#define REG_A6XX_SP_UNKNOWN_A982 0x0000a982 + #define REG_A6XX_SP_FS_OBJ_START_LO 0x0000a983 #define REG_A6XX_SP_FS_OBJ_START_HI 0x0000a984 @@ -3979,7 +4553,8 @@ static inline uint32_t A6XX_SP_FS_MRT_REG_COLOR_FORMAT(enum a6xx_color_fmt val) } #define A6XX_SP_FS_MRT_REG_COLOR_SINT 0x00000100 #define A6XX_SP_FS_MRT_REG_COLOR_UINT 0x00000200 -#define A6XX_SP_FS_MRT_REG_COLOR_SRGB 0x00000400 + +#define REG_A6XX_SP_UNKNOWN_A99E 0x0000a99e #define REG_A6XX_SP_FS_TEX_COUNT 0x0000a9a7 @@ -4066,14 +4641,20 @@ static inline uint32_t A6XX_SP_FS_CONFIG_NSAMP(uint32_t val) #define REG_A6XX_SP_FS_INSTRLEN 0x0000ab05 +#define REG_A6XX_SP_UNKNOWN_AB20 0x0000ab20 + #define REG_A6XX_SP_UNKNOWN_AE00 0x0000ae00 +#define REG_A6XX_SP_UNKNOWN_AE03 0x0000ae03 + #define REG_A6XX_SP_UNKNOWN_AE04 0x0000ae04 #define REG_A6XX_SP_UNKNOWN_AE0F 0x0000ae0f #define REG_A6XX_SP_UNKNOWN_B182 0x0000b182 +#define REG_A6XX_SP_UNKNOWN_B183 0x0000b183 + #define REG_A6XX_SP_TP_RAS_MSAA_CNTL 0x0000b300 #define A6XX_SP_TP_RAS_MSAA_CNTL_SAMPLES__MASK 0x00000003 #define A6XX_SP_TP_RAS_MSAA_CNTL_SAMPLES__SHIFT 0 @@ -4097,6 +4678,8 @@ static inline uint32_t A6XX_SP_TP_DEST_MSAA_CNTL_SAMPLES(enum a3xx_msaa_samples #define REG_A6XX_SP_TP_UNKNOWN_B304 0x0000b304 +#define REG_A6XX_SP_TP_UNKNOWN_B309 0x0000b309 + #define REG_A6XX_SP_PS_2D_SRC_INFO 0x0000b4c0 #define A6XX_SP_PS_2D_SRC_INFO_COLOR_FORMAT__MASK 0x000000ff #define A6XX_SP_PS_2D_SRC_INFO_COLOR_FORMAT__SHIFT 0 @@ -4162,6 +4745,8 @@ static inline uint32_t A6XX_HLSQ_GS_CNTL_CONSTLEN(uint32_t val) return ((val >> 2) << A6XX_HLSQ_GS_CNTL_CONSTLEN__SHIFT) & A6XX_HLSQ_GS_CNTL_CONSTLEN__MASK; } +#define REG_A6XX_HLSQ_UNKNOWN_B980 0x0000b980 + #define REG_A6XX_HLSQ_CONTROL_1_REG 0x0000b982 #define REG_A6XX_HLSQ_CONTROL_2_REG 0x0000b983 @@ -4537,11 +5122,11 @@ static inline uint32_t A6XX_TEX_CONST_7_FLAG_LO(uint32_t val) } #define REG_A6XX_TEX_CONST_8 0x00000008 -#define A6XX_TEX_CONST_8_BASE_HI__MASK 0x0001ffff -#define A6XX_TEX_CONST_8_BASE_HI__SHIFT 0 -static inline uint32_t A6XX_TEX_CONST_8_BASE_HI(uint32_t val) +#define A6XX_TEX_CONST_8_FLAG_HI__MASK 0x0001ffff +#define A6XX_TEX_CONST_8_FLAG_HI__SHIFT 0 +static inline uint32_t A6XX_TEX_CONST_8_FLAG_HI(uint32_t val) { - return ((val) << A6XX_TEX_CONST_8_BASE_HI__SHIFT) & A6XX_TEX_CONST_8_BASE_HI__MASK; + return ((val) << A6XX_TEX_CONST_8_FLAG_HI__SHIFT) & A6XX_TEX_CONST_8_FLAG_HI__MASK; } #define REG_A6XX_TEX_CONST_9 0x00000009 @@ -4558,5 +5143,227 @@ static inline uint32_t A6XX_TEX_CONST_8_BASE_HI(uint32_t val) #define REG_A6XX_TEX_CONST_15 0x0000000f +#define REG_A6XX_PDC_GPU_ENABLE_PDC 0x00001140 + +#define REG_A6XX_PDC_GPU_SEQ_START_ADDR 0x00001148 + +#define REG_A6XX_PDC_GPU_TCS0_CONTROL 0x00001540 + +#define REG_A6XX_PDC_GPU_TCS0_CMD_ENABLE_BANK 0x00001541 + +#define REG_A6XX_PDC_GPU_TCS0_CMD_WAIT_FOR_CMPL_BANK 0x00001542 + +#define REG_A6XX_PDC_GPU_TCS0_CMD0_MSGID 0x00001543 + +#define REG_A6XX_PDC_GPU_TCS0_CMD0_ADDR 0x00001544 + +#define REG_A6XX_PDC_GPU_TCS0_CMD0_DATA 0x00001545 + +#define REG_A6XX_PDC_GPU_TCS1_CONTROL 0x00001572 + +#define REG_A6XX_PDC_GPU_TCS1_CMD_ENABLE_BANK 0x00001573 + +#define REG_A6XX_PDC_GPU_TCS1_CMD_WAIT_FOR_CMPL_BANK 0x00001574 + +#define REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID 0x00001575 + +#define REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR 0x00001576 + +#define REG_A6XX_PDC_GPU_TCS1_CMD0_DATA 0x00001577 + +#define REG_A6XX_PDC_GPU_TCS2_CONTROL 0x000015a4 + +#define REG_A6XX_PDC_GPU_TCS2_CMD_ENABLE_BANK 0x000015a5 + +#define REG_A6XX_PDC_GPU_TCS2_CMD_WAIT_FOR_CMPL_BANK 0x000015a6 + +#define REG_A6XX_PDC_GPU_TCS2_CMD0_MSGID 0x000015a7 + +#define REG_A6XX_PDC_GPU_TCS2_CMD0_ADDR 0x000015a8 + +#define REG_A6XX_PDC_GPU_TCS2_CMD0_DATA 0x000015a9 + +#define REG_A6XX_PDC_GPU_TCS3_CONTROL 0x000015d6 + +#define REG_A6XX_PDC_GPU_TCS3_CMD_ENABLE_BANK 0x000015d7 + +#define REG_A6XX_PDC_GPU_TCS3_CMD_WAIT_FOR_CMPL_BANK 0x000015d8 + +#define REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID 0x000015d9 + +#define REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR 0x000015da + +#define REG_A6XX_PDC_GPU_TCS3_CMD0_DATA 0x000015db + +#define REG_A6XX_PDC_GPU_SEQ_MEM_0 0x00000000 + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_SEL_A 0x00000000 +#define A6XX_CX_DBGC_CFG_DBGBUS_SEL_A_PING_INDEX__MASK 0x000000ff +#define A6XX_CX_DBGC_CFG_DBGBUS_SEL_A_PING_INDEX__SHIFT 0 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_SEL_A_PING_INDEX(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_SEL_A_PING_INDEX__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_SEL_A_PING_INDEX__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_SEL_A_PING_BLK_SEL__MASK 0x0000ff00 +#define A6XX_CX_DBGC_CFG_DBGBUS_SEL_A_PING_BLK_SEL__SHIFT 8 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_SEL_A_PING_BLK_SEL(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_SEL_A_PING_BLK_SEL__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_SEL_A_PING_BLK_SEL__MASK; +} + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_SEL_B 0x00000001 + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_SEL_C 0x00000002 + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_SEL_D 0x00000003 + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_CNTLT 0x00000004 +#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_TRACEEN__MASK 0x0000003f +#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_TRACEEN__SHIFT 0 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_TRACEEN(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_TRACEEN__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_TRACEEN__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_GRANU__MASK 0x00007000 +#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_GRANU__SHIFT 12 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_GRANU(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_GRANU__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_GRANU__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_SEGT__MASK 0xf0000000 +#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_SEGT__SHIFT 28 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_SEGT(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_SEGT__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_CNTLT_SEGT__MASK; +} + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_CNTLM 0x00000005 +#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLM_ENABLE__MASK 0x0f000000 +#define A6XX_CX_DBGC_CFG_DBGBUS_CNTLM_ENABLE__SHIFT 24 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_CNTLM_ENABLE(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_CNTLM_ENABLE__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_CNTLM_ENABLE__MASK; +} + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_IVTL_0 0x00000008 + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_IVTL_1 0x00000009 + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_IVTL_2 0x0000000a + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_IVTL_3 0x0000000b + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_MASKL_0 0x0000000c + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_MASKL_1 0x0000000d + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_MASKL_2 0x0000000e + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_MASKL_3 0x0000000f + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0 0x00000010 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL0__MASK 0x0000000f +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL0__SHIFT 0 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL0(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL0__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL0__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL1__MASK 0x000000f0 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL1__SHIFT 4 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL1(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL1__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL1__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL2__MASK 0x00000f00 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL2__SHIFT 8 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL2(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL2__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL2__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL3__MASK 0x0000f000 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL3__SHIFT 12 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL3(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL3__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL3__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL4__MASK 0x000f0000 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL4__SHIFT 16 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL4(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL4__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL4__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL5__MASK 0x00f00000 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL5__SHIFT 20 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL5(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL5__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL5__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL6__MASK 0x0f000000 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL6__SHIFT 24 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL6(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL6__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL6__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL7__MASK 0xf0000000 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL7__SHIFT 28 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL7(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL7__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_0_BYTEL7__MASK; +} + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1 0x00000011 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL8__MASK 0x0000000f +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL8__SHIFT 0 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL8(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL8__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL8__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL9__MASK 0x000000f0 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL9__SHIFT 4 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL9(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL9__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL9__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL10__MASK 0x00000f00 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL10__SHIFT 8 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL10(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL10__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL10__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL11__MASK 0x0000f000 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL11__SHIFT 12 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL11(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL11__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL11__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL12__MASK 0x000f0000 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL12__SHIFT 16 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL12(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL12__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL12__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL13__MASK 0x00f00000 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL13__SHIFT 20 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL13(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL13__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL13__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL14__MASK 0x0f000000 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL14__SHIFT 24 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL14(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL14__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL14__MASK; +} +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL15__MASK 0xf0000000 +#define A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL15__SHIFT 28 +static inline uint32_t A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL15(uint32_t val) +{ + return ((val) << A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL15__SHIFT) & A6XX_CX_DBGC_CFG_DBGBUS_BYTEL_1_BYTEL15__MASK; +} + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_TRACE_BUF1 0x0000002f + +#define REG_A6XX_CX_DBGC_CFG_DBGBUS_TRACE_BUF2 0x00000030 + #endif /* A6XX_XML */ diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c index bbb8126ec5c5..d4e98e5876bc 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c @@ -2,7 +2,6 @@ /* Copyright (c) 2017-2018 The Linux Foundation. All rights reserved. */ #include <linux/clk.h> -#include <linux/iopoll.h> #include <linux/pm_opp.h> #include <soc/qcom/cmd-db.h> @@ -42,9 +41,6 @@ static irqreturn_t a6xx_hfi_irq(int irq, void *data) status = gmu_read(gmu, REG_A6XX_GMU_GMU2HOST_INTR_INFO); gmu_write(gmu, REG_A6XX_GMU_GMU2HOST_INTR_CLR, status); - if (status & A6XX_GMU_GMU2HOST_INTR_INFO_MSGQ) - tasklet_schedule(&gmu->hfi_tasklet); - if (status & A6XX_GMU_GMU2HOST_INTR_INFO_CM3_FAULT) { dev_err_ratelimited(gmu->dev, "GMU firmware fault\n"); @@ -65,12 +61,14 @@ static bool a6xx_gmu_gx_is_on(struct a6xx_gmu *gmu) A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_GX_HM_CLK_OFF)); } -static int a6xx_gmu_set_freq(struct a6xx_gmu *gmu, int index) +static void __a6xx_gmu_set_freq(struct a6xx_gmu *gmu, int index) { + int ret; + gmu_write(gmu, REG_A6XX_GMU_DCVS_ACK_OPTION, 0); gmu_write(gmu, REG_A6XX_GMU_DCVS_PERF_SETTING, - ((index << 24) & 0xff) | (3 & 0xf)); + ((3 & 0xf) << 28) | index); /* * Send an invalid index as a vote for the bus bandwidth and let the @@ -82,7 +80,37 @@ static int a6xx_gmu_set_freq(struct a6xx_gmu *gmu, int index) a6xx_gmu_set_oob(gmu, GMU_OOB_DCVS_SET); a6xx_gmu_clear_oob(gmu, GMU_OOB_DCVS_SET); - return gmu_read(gmu, REG_A6XX_GMU_DCVS_RETURN); + ret = gmu_read(gmu, REG_A6XX_GMU_DCVS_RETURN); + if (ret) + dev_err(gmu->dev, "GMU set GPU frequency error: %d\n", ret); + + gmu->freq = gmu->gpu_freqs[index]; +} + +void a6xx_gmu_set_freq(struct msm_gpu *gpu, unsigned long freq) +{ + struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); + struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu); + struct a6xx_gmu *gmu = &a6xx_gpu->gmu; + u32 perf_index = 0; + + if (freq == gmu->freq) + return; + + for (perf_index = 0; perf_index < gmu->nr_gpu_freqs - 1; perf_index++) + if (freq == gmu->gpu_freqs[perf_index]) + break; + + __a6xx_gmu_set_freq(gmu, perf_index); +} + +unsigned long a6xx_gmu_get_freq(struct msm_gpu *gpu) +{ + struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); + struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu); + struct a6xx_gmu *gmu = &a6xx_gpu->gmu; + + return gmu->freq; } static bool a6xx_gmu_check_idle_level(struct a6xx_gmu *gmu) @@ -135,9 +163,6 @@ static int a6xx_gmu_hfi_start(struct a6xx_gmu *gmu) u32 val; int ret; - gmu_rmw(gmu, REG_A6XX_GMU_GMU2HOST_INTR_MASK, - A6XX_GMU_GMU2HOST_INTR_INFO_MSGQ, 0); - gmu_write(gmu, REG_A6XX_GMU_HFI_CTRL_INIT, 1); ret = gmu_poll_timeout(gmu, REG_A6XX_GMU_HFI_CTRL_STATUS, val, @@ -348,8 +373,23 @@ static void a6xx_rpmh_stop(struct a6xx_gmu *gmu) gmu_write(gmu, REG_A6XX_GMU_RSCC_CONTROL_REQ, 0); } +static inline void pdc_write(void __iomem *ptr, u32 offset, u32 value) +{ + return msm_writel(value, ptr + (offset << 2)); +} + +static void __iomem *a6xx_gmu_get_mmio(struct platform_device *pdev, + const char *name); + static void a6xx_gmu_rpmh_init(struct a6xx_gmu *gmu) { + struct platform_device *pdev = to_platform_device(gmu->dev); + void __iomem *pdcptr = a6xx_gmu_get_mmio(pdev, "gmu_pdc"); + void __iomem *seqptr = a6xx_gmu_get_mmio(pdev, "gmu_pdc_seq"); + + if (!pdcptr || !seqptr) + goto err; + /* Disable SDE clock gating */ gmu_write(gmu, REG_A6XX_GPU_RSCC_RSC_STATUS0_DRV0, BIT(24)); @@ -374,44 +414,48 @@ static void a6xx_gmu_rpmh_init(struct a6xx_gmu *gmu) gmu_write(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 4, 0x0020e8a8); /* Load PDC sequencer uCode for power up and power down sequence */ - pdc_write(gmu, REG_A6XX_PDC_GPU_SEQ_MEM_0, 0xfebea1e1); - pdc_write(gmu, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 1, 0xa5a4a3a2); - pdc_write(gmu, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 2, 0x8382a6e0); - pdc_write(gmu, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 3, 0xbce3e284); - pdc_write(gmu, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 4, 0x002081fc); + pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0, 0xfebea1e1); + pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 1, 0xa5a4a3a2); + pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 2, 0x8382a6e0); + pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 3, 0xbce3e284); + pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 4, 0x002081fc); /* Set TCS commands used by PDC sequence for low power modes */ - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD_ENABLE_BANK, 7); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD_WAIT_FOR_CMPL_BANK, 0); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CONTROL, 0); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID, 0x10108); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR, 0x30010); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA, 1); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID + 4, 0x10108); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR + 4, 0x30000); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA + 4, 0x0); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID + 8, 0x10108); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR + 8, 0x30080); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA + 8, 0x0); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD_ENABLE_BANK, 7); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD_WAIT_FOR_CMPL_BANK, 0); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CONTROL, 0); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID, 0x10108); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR, 0x30010); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA, 2); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID + 4, 0x10108); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR + 4, 0x30000); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 4, 0x3); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID + 8, 0x10108); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR + 8, 0x30080); - pdc_write(gmu, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 8, 0x3); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD_ENABLE_BANK, 7); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD_WAIT_FOR_CMPL_BANK, 0); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CONTROL, 0); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID, 0x10108); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR, 0x30010); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA, 1); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID + 4, 0x10108); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR + 4, 0x30000); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA + 4, 0x0); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID + 8, 0x10108); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR + 8, 0x30080); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA + 8, 0x0); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD_ENABLE_BANK, 7); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD_WAIT_FOR_CMPL_BANK, 0); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CONTROL, 0); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID, 0x10108); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR, 0x30010); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA, 2); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID + 4, 0x10108); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR + 4, 0x30000); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 4, 0x3); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID + 8, 0x10108); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR + 8, 0x30080); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 8, 0x3); /* Setup GPU PDC */ - pdc_write(gmu, REG_A6XX_PDC_GPU_SEQ_START_ADDR, 0); - pdc_write(gmu, REG_A6XX_PDC_GPU_ENABLE_PDC, 0x80000001); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_SEQ_START_ADDR, 0); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_ENABLE_PDC, 0x80000001); /* ensure no writes happen before the uCode is fully written */ wmb(); + +err: + devm_iounmap(gmu->dev, pdcptr); + devm_iounmap(gmu->dev, seqptr); } /* @@ -547,8 +591,7 @@ static int a6xx_gmu_fw_start(struct a6xx_gmu *gmu, unsigned int state) } #define A6XX_HFI_IRQ_MASK \ - (A6XX_GMU_GMU2HOST_INTR_INFO_MSGQ | \ - A6XX_GMU_GMU2HOST_INTR_INFO_CM3_FAULT) + (A6XX_GMU_GMU2HOST_INTR_INFO_CM3_FAULT) #define A6XX_GMU_IRQ_MASK \ (A6XX_GMU_AO_HOST_INTERRUPT_STATUS_WDOG_BITE | \ @@ -626,7 +669,7 @@ int a6xx_gmu_reset(struct a6xx_gpu *a6xx_gpu) ret = a6xx_hfi_start(gmu, GMU_COLD_BOOT); /* Set the GPU back to the highest power frequency */ - a6xx_gmu_set_freq(gmu, gmu->nr_gpu_freqs - 1); + __a6xx_gmu_set_freq(gmu, gmu->nr_gpu_freqs - 1); out: if (ret) @@ -665,7 +708,7 @@ int a6xx_gmu_resume(struct a6xx_gpu *a6xx_gpu) ret = a6xx_hfi_start(gmu, status); /* Set the GPU to the highest power frequency */ - a6xx_gmu_set_freq(gmu, gmu->nr_gpu_freqs - 1); + __a6xx_gmu_set_freq(gmu, gmu->nr_gpu_freqs - 1); out: /* Make sure to turn off the boot OOB request on error */ @@ -1140,7 +1183,7 @@ int a6xx_gmu_probe(struct a6xx_gpu *a6xx_gpu, struct device_node *node) gmu->dev = &pdev->dev; - of_dma_configure(gmu->dev, node, false); + of_dma_configure(gmu->dev, node, true); /* Fow now, don't do anything fancy until we get our feet under us */ gmu->idle_level = GMU_IDLE_STATE_ACTIVE; @@ -1170,11 +1213,7 @@ int a6xx_gmu_probe(struct a6xx_gpu *a6xx_gpu, struct device_node *node) /* Map the GMU registers */ gmu->mmio = a6xx_gmu_get_mmio(pdev, "gmu"); - - /* Map the GPU power domain controller registers */ - gmu->pdc_mmio = a6xx_gmu_get_mmio(pdev, "gmu_pdc"); - - if (IS_ERR(gmu->mmio) || IS_ERR(gmu->pdc_mmio)) + if (IS_ERR(gmu->mmio)) goto err; /* Get the HFI and GMU interrupts */ @@ -1184,9 +1223,6 @@ int a6xx_gmu_probe(struct a6xx_gpu *a6xx_gpu, struct device_node *node) if (gmu->hfi_irq < 0 || gmu->gmu_irq < 0) goto err; - /* Set up a tasklet to handle GMU HFI responses */ - tasklet_init(&gmu->hfi_tasklet, a6xx_hfi_task, (unsigned long) gmu); - /* Get the power levels for the GMU and GPU */ a6xx_gmu_pwrlevels_probe(gmu); diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.h b/drivers/gpu/drm/msm/adreno/a6xx_gmu.h index d9a386c18799..35f765afae45 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.h +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.h @@ -4,6 +4,7 @@ #ifndef _A6XX_GMU_H_ #define _A6XX_GMU_H_ +#include <linux/iopoll.h> #include <linux/interrupt.h> #include "msm_drv.h" #include "a6xx_hfi.h" @@ -47,7 +48,6 @@ struct a6xx_gmu { struct device *dev; void * __iomem mmio; - void * __iomem pdc_mmio; int hfi_irq; int gmu_irq; @@ -74,6 +74,8 @@ struct a6xx_gmu { unsigned long gmu_freqs[4]; u32 cx_arc_votes[4]; + unsigned long freq; + struct a6xx_hfi_queue queues[2]; struct tasklet_struct hfi_tasklet; @@ -89,11 +91,6 @@ static inline void gmu_write(struct a6xx_gmu *gmu, u32 offset, u32 value) return msm_writel(value, gmu->mmio + (offset << 2)); } -static inline void pdc_write(struct a6xx_gmu *gmu, u32 offset, u32 value) -{ - return msm_writel(value, gmu->pdc_mmio + (offset << 2)); -} - static inline void gmu_rmw(struct a6xx_gmu *gmu, u32 reg, u32 mask, u32 or) { u32 val = gmu_read(gmu, reg); @@ -103,6 +100,16 @@ static inline void gmu_rmw(struct a6xx_gmu *gmu, u32 reg, u32 mask, u32 or) gmu_write(gmu, reg, val | or); } +static inline u64 gmu_read64(struct a6xx_gmu *gmu, u32 lo, u32 hi) +{ + u64 val; + + val = (u64) msm_readl(gmu->mmio + (lo << 2)); + val |= ((u64) msm_readl(gmu->mmio + (hi << 2)) << 32); + + return val; +} + #define gmu_poll_timeout(gmu, addr, val, cond, interval, timeout) \ readl_poll_timeout((gmu)->mmio + ((addr) << 2), val, cond, \ interval, timeout) @@ -157,6 +164,4 @@ void a6xx_hfi_init(struct a6xx_gmu *gmu); int a6xx_hfi_start(struct a6xx_gmu *gmu, int boot_state); void a6xx_hfi_stop(struct a6xx_gmu *gmu); -void a6xx_hfi_task(unsigned long data); - #endif diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.xml.h b/drivers/gpu/drm/msm/adreno/a6xx_gmu.xml.h index ef68098d2adc..db56f263ed77 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.xml.h +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.xml.h @@ -12,12 +12,12 @@ The rules-ng-ng source files this header was generated from are: - /home/robclark/src/envytools/rnndb/freedreno_copyright.xml ( 1572 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a2xx.xml ( 36805 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/adreno_common.xml ( 13634 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42393 bytes, from 2018-08-06 18:45:45) +- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42585 bytes, from 2018-10-04 19:06:37) - /home/robclark/src/envytools/rnndb/adreno/a3xx.xml ( 83840 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a4xx.xml ( 112086 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 101627 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-07-03 19:37:13) +- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-10-04 19:06:37) +- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 139581 bytes, from 2018-10-04 19:06:42) +- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-09-14 13:03:07) - /home/robclark/src/envytools/rnndb/adreno/ocmem.xml ( 1773 bytes, from 2018-07-03 19:37:13) Copyright (C) 2013-2018 by the following authors: @@ -167,8 +167,8 @@ static inline uint32_t A6XX_GMU_PWR_COL_INTER_FRAME_CTRL_MIN_PASS_LENGTH(uint32_ #define REG_A6XX_GMU_SPTPRAC_PWR_CLK_STATUS 0x000050d0 #define A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_SPTPRAC_GDSC_POWERING_OFF 0x00000001 #define A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_SPTPRAC_GDSC_POWERING_ON 0x00000002 -#define A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_SPTPRAC_GDSC_POWER_ON 0x00000004 -#define A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_SPTPRAC_GDSC_POWER_OFF 0x00000008 +#define A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_SPTPRAC_GDSC_POWER_OFF 0x00000004 +#define A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_SPTPRAC_GDSC_POWER_ON 0x00000008 #define A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_SP_CLOCK_OFF 0x00000010 #define A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_GMU_UP_POWER_STATE 0x00000020 #define A6XX_GMU_SPTPRAC_PWR_CLK_STATUS_GX_HM_GDSC_POWER_OFF 0x00000040 diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c index c629f742a1d1..631257c297fd 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c @@ -7,6 +7,8 @@ #include "a6xx_gpu.h" #include "a6xx_gmu.xml.h" +#include <linux/devfreq.h> + static inline bool _a6xx_check_idle(struct msm_gpu *gpu) { struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); @@ -438,10 +440,8 @@ static int a6xx_hw_init(struct msm_gpu *gpu) gpu_write(gpu, REG_A6XX_CP_PROTECT(22), A6XX_PROTECT_RW(0x900, 0x4d)); gpu_write(gpu, REG_A6XX_CP_PROTECT(23), A6XX_PROTECT_RW(0x98d, 0x76)); gpu_write(gpu, REG_A6XX_CP_PROTECT(24), - A6XX_PROTECT_RDONLY(0x8d0, 0x23)); - gpu_write(gpu, REG_A6XX_CP_PROTECT(25), A6XX_PROTECT_RDONLY(0x980, 0x4)); - gpu_write(gpu, REG_A6XX_CP_PROTECT(26), A6XX_PROTECT_RW(0xa630, 0x0)); + gpu_write(gpu, REG_A6XX_CP_PROTECT(25), A6XX_PROTECT_RW(0xa630, 0x0)); /* Enable interrupts */ gpu_write(gpu, REG_A6XX_RBBM_INT_0_MASK, A6XX_INT_MASK); @@ -682,6 +682,8 @@ static int a6xx_pm_resume(struct msm_gpu *gpu) gpu->needs_hw_init = true; + msm_gpu_resume_devfreq(gpu); + return ret; } @@ -690,6 +692,8 @@ static int a6xx_pm_suspend(struct msm_gpu *gpu) struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu); + devfreq_suspend_device(gpu->devfreq.devfreq); + /* * Make sure the GMU is idle before continuing (because some transitions * may use VBIF @@ -744,7 +748,7 @@ static void a6xx_destroy(struct msm_gpu *gpu) if (a6xx_gpu->sqe_bo) { if (a6xx_gpu->sqe_iova) msm_gem_put_iova(a6xx_gpu->sqe_bo, gpu->aspace); - drm_gem_object_unreference_unlocked(a6xx_gpu->sqe_bo); + drm_gem_object_put_unlocked(a6xx_gpu->sqe_bo); } a6xx_gmu_remove(a6xx_gpu); @@ -753,6 +757,27 @@ static void a6xx_destroy(struct msm_gpu *gpu) kfree(a6xx_gpu); } +static unsigned long a6xx_gpu_busy(struct msm_gpu *gpu) +{ + struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); + struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu); + u64 busy_cycles, busy_time; + + busy_cycles = gmu_read64(&a6xx_gpu->gmu, + REG_A6XX_GMU_CX_GMU_POWER_COUNTER_XOCLK_0_L, + REG_A6XX_GMU_CX_GMU_POWER_COUNTER_XOCLK_0_H); + + busy_time = (busy_cycles - gpu->devfreq.busy_cycles) * 10; + do_div(busy_time, 192); + + gpu->devfreq.busy_cycles = busy_cycles; + + if (WARN_ON(busy_time > ~0LU)) + return ~0LU; + + return (unsigned long)busy_time; +} + static const struct adreno_gpu_funcs funcs = { .base = { .get_param = adreno_get_param, @@ -768,6 +793,9 @@ static const struct adreno_gpu_funcs funcs = { #if defined(CONFIG_DEBUG_FS) || defined(CONFIG_DEV_COREDUMP) .show = a6xx_show, #endif + .gpu_busy = a6xx_gpu_busy, + .gpu_get_freq = a6xx_gmu_get_freq, + .gpu_set_freq = a6xx_gmu_set_freq, }, .get_timestamp = a6xx_get_timestamp, }; @@ -799,7 +827,7 @@ struct msm_gpu *a6xx_gpu_init(struct drm_device *dev) } /* Check if there is a GMU phandle and set it up */ - node = of_parse_phandle(pdev->dev.of_node, "gmu", 0); + node = of_parse_phandle(pdev->dev.of_node, "qcom,gmu", 0); /* FIXME: How do we gracefully handle this? */ BUG_ON(!node); diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.h b/drivers/gpu/drm/msm/adreno/a6xx_gpu.h index dd69e5b0e692..4127dcebc202 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.h +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.h @@ -56,5 +56,6 @@ void a6xx_gmu_clear_oob(struct a6xx_gmu *gmu, enum a6xx_gmu_oob_state state); int a6xx_gmu_probe(struct a6xx_gpu *a6xx_gpu, struct device_node *node); void a6xx_gmu_remove(struct a6xx_gpu *a6xx_gpu); - +void a6xx_gmu_set_freq(struct msm_gpu *gpu, unsigned long freq); +unsigned long a6xx_gmu_get_freq(struct msm_gpu *gpu); #endif /* __A6XX_GPU_H__ */ diff --git a/drivers/gpu/drm/msm/adreno/a6xx_hfi.c b/drivers/gpu/drm/msm/adreno/a6xx_hfi.c index f19ef4cb6ea4..6ff9baec2658 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_hfi.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_hfi.c @@ -79,83 +79,72 @@ static int a6xx_hfi_queue_write(struct a6xx_gmu *gmu, return 0; } -struct a6xx_hfi_response { - u32 id; - u32 seqnum; - struct list_head node; - struct completion complete; - - u32 error; - u32 payload[16]; -}; +static int a6xx_hfi_wait_for_ack(struct a6xx_gmu *gmu, u32 id, u32 seqnum, + u32 *payload, u32 payload_size) +{ + struct a6xx_hfi_queue *queue = &gmu->queues[HFI_RESPONSE_QUEUE]; + u32 val; + int ret; -/* - * Incoming HFI ack messages can come in out of order so we need to store all - * the pending messages on a list until they are handled. - */ -static spinlock_t hfi_ack_lock = __SPIN_LOCK_UNLOCKED(message_lock); -static LIST_HEAD(hfi_ack_list); + /* Wait for a response */ + ret = gmu_poll_timeout(gmu, REG_A6XX_GMU_GMU2HOST_INTR_INFO, val, + val & A6XX_GMU_GMU2HOST_INTR_INFO_MSGQ, 100, 5000); -static void a6xx_hfi_handle_ack(struct a6xx_gmu *gmu, - struct a6xx_hfi_msg_response *msg) -{ - struct a6xx_hfi_response *resp; - u32 id, seqnum; - - /* msg->ret_header contains the header of the message being acked */ - id = HFI_HEADER_ID(msg->ret_header); - seqnum = HFI_HEADER_SEQNUM(msg->ret_header); - - spin_lock(&hfi_ack_lock); - list_for_each_entry(resp, &hfi_ack_list, node) { - if (resp->id == id && resp->seqnum == seqnum) { - resp->error = msg->error; - memcpy(resp->payload, msg->payload, - sizeof(resp->payload)); - - complete(&resp->complete); - spin_unlock(&hfi_ack_lock); - return; - } + if (ret) { + dev_err(gmu->dev, + "Message %s id %d timed out waiting for response\n", + a6xx_hfi_msg_id[id], seqnum); + return -ETIMEDOUT; } - spin_unlock(&hfi_ack_lock); - dev_err(gmu->dev, "Nobody was waiting for HFI message %d\n", seqnum); -} + /* Clear the interrupt */ + gmu_write(gmu, REG_A6XX_GMU_GMU2HOST_INTR_CLR, + A6XX_GMU_GMU2HOST_INTR_INFO_MSGQ); -static void a6xx_hfi_handle_error(struct a6xx_gmu *gmu, - struct a6xx_hfi_msg_response *msg) -{ - struct a6xx_hfi_msg_error *error = (struct a6xx_hfi_msg_error *) msg; + for (;;) { + struct a6xx_hfi_msg_response resp; - dev_err(gmu->dev, "GMU firmware error %d\n", error->code); -} + /* Get the next packet */ + ret = a6xx_hfi_queue_read(queue, (u32 *) &resp, + sizeof(resp) >> 2); -void a6xx_hfi_task(unsigned long data) -{ - struct a6xx_gmu *gmu = (struct a6xx_gmu *) data; - struct a6xx_hfi_queue *queue = &gmu->queues[HFI_RESPONSE_QUEUE]; - struct a6xx_hfi_msg_response resp; + /* If the queue is empty our response never made it */ + if (!ret) { + dev_err(gmu->dev, + "The HFI response queue is unexpectedly empty\n"); - for (;;) { - u32 id; - int ret = a6xx_hfi_queue_read(queue, (u32 *) &resp, - sizeof(resp) >> 2); + return -ENOENT; + } + + if (HFI_HEADER_ID(resp.header) == HFI_F2H_MSG_ERROR) { + struct a6xx_hfi_msg_error *error = + (struct a6xx_hfi_msg_error *) &resp; - /* Returns the number of bytes copied or negative on error */ - if (ret <= 0) { - if (ret < 0) - dev_err(gmu->dev, - "Unable to read the HFI message queue\n"); - break; + dev_err(gmu->dev, "GMU firmware error %d\n", + error->code); + continue; + } + + if (seqnum != HFI_HEADER_SEQNUM(resp.ret_header)) { + dev_err(gmu->dev, + "Unexpected message id %d on the response queue\n", + HFI_HEADER_SEQNUM(resp.ret_header)); + continue; + } + + if (resp.error) { + dev_err(gmu->dev, + "Message %s id %d returned error %d\n", + a6xx_hfi_msg_id[id], seqnum, resp.error); + return -EINVAL; } - id = HFI_HEADER_ID(resp.header); + /* All is well, copy over the buffer */ + if (payload && payload_size) + memcpy(payload, resp.payload, + min_t(u32, payload_size, sizeof(resp.payload))); - if (id == HFI_F2H_MSG_ACK) - a6xx_hfi_handle_ack(gmu, &resp); - else if (id == HFI_F2H_MSG_ERROR) - a6xx_hfi_handle_error(gmu, &resp); + return 0; } } @@ -163,7 +152,6 @@ static int a6xx_hfi_send_msg(struct a6xx_gmu *gmu, int id, void *data, u32 size, u32 *payload, u32 payload_size) { struct a6xx_hfi_queue *queue = &gmu->queues[HFI_COMMAND_QUEUE]; - struct a6xx_hfi_response resp = { 0 }; int ret, dwords = size >> 2; u32 seqnum; @@ -173,53 +161,14 @@ static int a6xx_hfi_send_msg(struct a6xx_gmu *gmu, int id, *((u32 *) data) = (seqnum << 20) | (HFI_MSG_CMD << 16) | (dwords << 8) | id; - init_completion(&resp.complete); - resp.id = id; - resp.seqnum = seqnum; - - spin_lock_bh(&hfi_ack_lock); - list_add_tail(&resp.node, &hfi_ack_list); - spin_unlock_bh(&hfi_ack_lock); - ret = a6xx_hfi_queue_write(gmu, queue, data, dwords); if (ret) { dev_err(gmu->dev, "Unable to send message %s id %d\n", a6xx_hfi_msg_id[id], seqnum); - goto out; - } - - /* Wait up to 5 seconds for the response */ - ret = wait_for_completion_timeout(&resp.complete, - msecs_to_jiffies(5000)); - if (!ret) { - dev_err(gmu->dev, - "Message %s id %d timed out waiting for response\n", - a6xx_hfi_msg_id[id], seqnum); - ret = -ETIMEDOUT; - } else - ret = 0; - -out: - spin_lock_bh(&hfi_ack_lock); - list_del(&resp.node); - spin_unlock_bh(&hfi_ack_lock); - - if (ret) return ret; - - if (resp.error) { - dev_err(gmu->dev, "Message %s id %d returned error %d\n", - a6xx_hfi_msg_id[id], seqnum, resp.error); - return -EINVAL; } - if (payload && payload_size) { - int copy = min_t(u32, payload_size, sizeof(resp.payload)); - - memcpy(payload, resp.payload, copy); - } - - return 0; + return a6xx_hfi_wait_for_ack(gmu, id, seqnum, payload, payload_size); } static int a6xx_hfi_send_gmu_init(struct a6xx_gmu *gmu, int boot_state) diff --git a/drivers/gpu/drm/msm/adreno/adreno_common.xml.h b/drivers/gpu/drm/msm/adreno/adreno_common.xml.h index 5dace1350810..1318959d504d 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_common.xml.h +++ b/drivers/gpu/drm/msm/adreno/adreno_common.xml.h @@ -12,12 +12,12 @@ The rules-ng-ng source files this header was generated from are: - /home/robclark/src/envytools/rnndb/freedreno_copyright.xml ( 1572 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a2xx.xml ( 36805 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/adreno_common.xml ( 13634 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42393 bytes, from 2018-08-06 18:45:45) +- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42585 bytes, from 2018-10-04 19:06:37) - /home/robclark/src/envytools/rnndb/adreno/a3xx.xml ( 83840 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a4xx.xml ( 112086 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 101627 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-07-03 19:37:13) +- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-10-04 19:06:37) +- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 139581 bytes, from 2018-10-04 19:06:42) +- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-09-14 13:03:07) - /home/robclark/src/envytools/rnndb/adreno/ocmem.xml ( 1773 bytes, from 2018-07-03 19:37:13) Copyright (C) 2013-2018 by the following authors: diff --git a/drivers/gpu/drm/msm/adreno/adreno_device.c b/drivers/gpu/drm/msm/adreno/adreno_device.c index 7d3e9a129ac7..86abdb2b3a9c 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_device.c +++ b/drivers/gpu/drm/msm/adreno/adreno_device.c @@ -120,6 +120,7 @@ static const struct adreno_info gpulist[] = { [ADRENO_FW_GMU] = "a630_gmu.bin", }, .gmem = SZ_1M, + .inactive_period = DRM_MSM_INACTIVE_PERIOD, .init = a6xx_gpu_init, }, }; diff --git a/drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h b/drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h index 03a91e10b310..15eb03bed984 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h +++ b/drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h @@ -12,12 +12,12 @@ The rules-ng-ng source files this header was generated from are: - /home/robclark/src/envytools/rnndb/freedreno_copyright.xml ( 1572 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a2xx.xml ( 36805 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/adreno_common.xml ( 13634 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42393 bytes, from 2018-08-06 18:45:45) +- /home/robclark/src/envytools/rnndb/adreno/adreno_pm4.xml ( 42585 bytes, from 2018-10-04 19:06:37) - /home/robclark/src/envytools/rnndb/adreno/a3xx.xml ( 83840 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a4xx.xml ( 112086 bytes, from 2018-07-03 19:37:13) -- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 101627 bytes, from 2018-08-06 18:45:45) -- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-07-03 19:37:13) +- /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-10-04 19:06:37) +- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 139581 bytes, from 2018-10-04 19:06:42) +- /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-09-14 13:03:07) - /home/robclark/src/envytools/rnndb/adreno/ocmem.xml ( 1773 bytes, from 2018-07-03 19:37:13) Copyright (C) 2013-2018 by the following authors: @@ -237,7 +237,7 @@ enum adreno_pm4_type3_packets { CP_UNK_A6XX_14 = 20, CP_UNK_A6XX_36 = 54, CP_UNK_A6XX_55 = 85, - UNK_A6XX_6D = 109, + CP_REG_WRITE = 109, }; enum adreno_state_block { @@ -968,19 +968,19 @@ static inline uint32_t CP_SET_BIN_DATA5_4_BIN_SIZE_ADDRESS_HI(uint32_t val) } #define REG_CP_SET_BIN_DATA5_5 0x00000005 -#define CP_SET_BIN_DATA5_5_XXX_ADDRESS_LO__MASK 0xffffffff -#define CP_SET_BIN_DATA5_5_XXX_ADDRESS_LO__SHIFT 0 -static inline uint32_t CP_SET_BIN_DATA5_5_XXX_ADDRESS_LO(uint32_t val) +#define CP_SET_BIN_DATA5_5_BIN_DATA_ADDR2_LO__MASK 0xffffffff +#define CP_SET_BIN_DATA5_5_BIN_DATA_ADDR2_LO__SHIFT 0 +static inline uint32_t CP_SET_BIN_DATA5_5_BIN_DATA_ADDR2_LO(uint32_t val) { - return ((val) << CP_SET_BIN_DATA5_5_XXX_ADDRESS_LO__SHIFT) & CP_SET_BIN_DATA5_5_XXX_ADDRESS_LO__MASK; + return ((val) << CP_SET_BIN_DATA5_5_BIN_DATA_ADDR2_LO__SHIFT) & CP_SET_BIN_DATA5_5_BIN_DATA_ADDR2_LO__MASK; } #define REG_CP_SET_BIN_DATA5_6 0x00000006 -#define CP_SET_BIN_DATA5_6_XXX_ADDRESS_HI__MASK 0xffffffff -#define CP_SET_BIN_DATA5_6_XXX_ADDRESS_HI__SHIFT 0 -static inline uint32_t CP_SET_BIN_DATA5_6_XXX_ADDRESS_HI(uint32_t val) +#define CP_SET_BIN_DATA5_6_BIN_DATA_ADDR2_LO__MASK 0xffffffff +#define CP_SET_BIN_DATA5_6_BIN_DATA_ADDR2_LO__SHIFT 0 +static inline uint32_t CP_SET_BIN_DATA5_6_BIN_DATA_ADDR2_LO(uint32_t val) { - return ((val) << CP_SET_BIN_DATA5_6_XXX_ADDRESS_HI__SHIFT) & CP_SET_BIN_DATA5_6_XXX_ADDRESS_HI__MASK; + return ((val) << CP_SET_BIN_DATA5_6_BIN_DATA_ADDR2_LO__SHIFT) & CP_SET_BIN_DATA5_6_BIN_DATA_ADDR2_LO__MASK; } #define REG_CP_REG_TO_MEM_0 0x00000000 diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c index 80cbf75bc2ff..d4530d60767b 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c @@ -47,237 +47,17 @@ #define LEFT_MIXER 0 #define RIGHT_MIXER 1 -#define MISR_BUFF_SIZE 256 - -static inline struct dpu_kms *_dpu_crtc_get_kms(struct drm_crtc *crtc) -{ - struct msm_drm_private *priv; - - if (!crtc || !crtc->dev || !crtc->dev->dev_private) { - DPU_ERROR("invalid crtc\n"); - return NULL; - } - priv = crtc->dev->dev_private; - if (!priv || !priv->kms) { - DPU_ERROR("invalid kms\n"); - return NULL; - } - - return to_dpu_kms(priv->kms); -} - -static inline int _dpu_crtc_power_enable(struct dpu_crtc *dpu_crtc, bool enable) -{ - struct drm_crtc *crtc; - struct msm_drm_private *priv; - struct dpu_kms *dpu_kms; - - if (!dpu_crtc) { - DPU_ERROR("invalid dpu crtc\n"); - return -EINVAL; - } - - crtc = &dpu_crtc->base; - if (!crtc->dev || !crtc->dev->dev_private) { - DPU_ERROR("invalid drm device\n"); - return -EINVAL; - } - - priv = crtc->dev->dev_private; - if (!priv->kms) { - DPU_ERROR("invalid kms\n"); - return -EINVAL; - } - - dpu_kms = to_dpu_kms(priv->kms); - - if (enable) - pm_runtime_get_sync(&dpu_kms->pdev->dev); - else - pm_runtime_put_sync(&dpu_kms->pdev->dev); - - return 0; -} - -/** - * _dpu_crtc_rp_to_crtc - get crtc from resource pool object - * @rp: Pointer to resource pool - * return: Pointer to drm crtc if success; null otherwise - */ -static struct drm_crtc *_dpu_crtc_rp_to_crtc(struct dpu_crtc_respool *rp) -{ - if (!rp) - return NULL; - - return container_of(rp, struct dpu_crtc_state, rp)->base.crtc; -} - -/** - * _dpu_crtc_rp_reclaim - reclaim unused, or all if forced, resources in pool - * @rp: Pointer to resource pool - * @force: True to reclaim all resources; otherwise, reclaim only unused ones - * return: None - */ -static void _dpu_crtc_rp_reclaim(struct dpu_crtc_respool *rp, bool force) +static inline int _dpu_crtc_get_mixer_width(struct dpu_crtc_state *cstate, + struct drm_display_mode *mode) { - struct dpu_crtc_res *res, *next; - struct drm_crtc *crtc; - - crtc = _dpu_crtc_rp_to_crtc(rp); - if (!crtc) { - DPU_ERROR("invalid crtc\n"); - return; - } - - DPU_DEBUG("crtc%d.%u %s\n", crtc->base.id, rp->sequence_id, - force ? "destroy" : "free_unused"); - - list_for_each_entry_safe(res, next, &rp->res_list, list) { - if (!force && !(res->flags & DPU_CRTC_RES_FLAG_FREE)) - continue; - DPU_DEBUG("crtc%d.%u reclaim res:0x%x/0x%llx/%pK/%d\n", - crtc->base.id, rp->sequence_id, - res->type, res->tag, res->val, - atomic_read(&res->refcount)); - list_del(&res->list); - if (res->ops.put) - res->ops.put(res->val); - kfree(res); - } + return mode->hdisplay / cstate->num_mixers; } -/** - * _dpu_crtc_rp_free_unused - free unused resource in pool - * @rp: Pointer to resource pool - * return: none - */ -static void _dpu_crtc_rp_free_unused(struct dpu_crtc_respool *rp) -{ - mutex_lock(rp->rp_lock); - _dpu_crtc_rp_reclaim(rp, false); - mutex_unlock(rp->rp_lock); -} - -/** - * _dpu_crtc_rp_destroy - destroy resource pool - * @rp: Pointer to resource pool - * return: None - */ -static void _dpu_crtc_rp_destroy(struct dpu_crtc_respool *rp) -{ - mutex_lock(rp->rp_lock); - list_del_init(&rp->rp_list); - _dpu_crtc_rp_reclaim(rp, true); - mutex_unlock(rp->rp_lock); -} - -/** - * _dpu_crtc_hw_blk_get - get callback for hardware block - * @val: Resource handle - * @type: Resource type - * @tag: Search tag for given resource - * return: Resource handle - */ -static void *_dpu_crtc_hw_blk_get(void *val, u32 type, u64 tag) -{ - DPU_DEBUG("res:%d/0x%llx/%pK\n", type, tag, val); - return dpu_hw_blk_get(val, type, tag); -} - -/** - * _dpu_crtc_hw_blk_put - put callback for hardware block - * @val: Resource handle - * return: None - */ -static void _dpu_crtc_hw_blk_put(void *val) -{ - DPU_DEBUG("res://%pK\n", val); - dpu_hw_blk_put(val); -} - -/** - * _dpu_crtc_rp_duplicate - duplicate resource pool and reset reference count - * @rp: Pointer to original resource pool - * @dup_rp: Pointer to duplicated resource pool - * return: None - */ -static void _dpu_crtc_rp_duplicate(struct dpu_crtc_respool *rp, - struct dpu_crtc_respool *dup_rp) -{ - struct dpu_crtc_res *res, *dup_res; - struct drm_crtc *crtc; - - if (!rp || !dup_rp || !rp->rp_head) { - DPU_ERROR("invalid resource pool\n"); - return; - } - - crtc = _dpu_crtc_rp_to_crtc(rp); - if (!crtc) { - DPU_ERROR("invalid crtc\n"); - return; - } - - DPU_DEBUG("crtc%d.%u duplicate\n", crtc->base.id, rp->sequence_id); - - mutex_lock(rp->rp_lock); - dup_rp->sequence_id = rp->sequence_id + 1; - INIT_LIST_HEAD(&dup_rp->res_list); - dup_rp->ops = rp->ops; - list_for_each_entry(res, &rp->res_list, list) { - dup_res = kzalloc(sizeof(struct dpu_crtc_res), GFP_KERNEL); - if (!dup_res) { - mutex_unlock(rp->rp_lock); - return; - } - INIT_LIST_HEAD(&dup_res->list); - atomic_set(&dup_res->refcount, 0); - dup_res->type = res->type; - dup_res->tag = res->tag; - dup_res->val = res->val; - dup_res->ops = res->ops; - dup_res->flags = DPU_CRTC_RES_FLAG_FREE; - DPU_DEBUG("crtc%d.%u dup res:0x%x/0x%llx/%pK/%d\n", - crtc->base.id, dup_rp->sequence_id, - dup_res->type, dup_res->tag, dup_res->val, - atomic_read(&dup_res->refcount)); - list_add_tail(&dup_res->list, &dup_rp->res_list); - if (dup_res->ops.get) - dup_res->ops.get(dup_res->val, 0, -1); - } - - dup_rp->rp_lock = rp->rp_lock; - dup_rp->rp_head = rp->rp_head; - INIT_LIST_HEAD(&dup_rp->rp_list); - list_add_tail(&dup_rp->rp_list, rp->rp_head); - mutex_unlock(rp->rp_lock); -} - -/** - * _dpu_crtc_rp_reset - reset resource pool after allocation - * @rp: Pointer to original resource pool - * @rp_lock: Pointer to serialization resource pool lock - * @rp_head: Pointer to crtc resource pool head - * return: None - */ -static void _dpu_crtc_rp_reset(struct dpu_crtc_respool *rp, - struct mutex *rp_lock, struct list_head *rp_head) +static inline struct dpu_kms *_dpu_crtc_get_kms(struct drm_crtc *crtc) { - if (!rp || !rp_lock || !rp_head) { - DPU_ERROR("invalid resource pool\n"); - return; - } + struct msm_drm_private *priv = crtc->dev->dev_private; - mutex_lock(rp_lock); - rp->rp_lock = rp_lock; - rp->rp_head = rp_head; - INIT_LIST_HEAD(&rp->rp_list); - rp->sequence_id = 0; - INIT_LIST_HEAD(&rp->res_list); - rp->ops.get = _dpu_crtc_hw_blk_get; - rp->ops.put = _dpu_crtc_hw_blk_put; - list_add_tail(&rp->rp_list, rp->rp_head); - mutex_unlock(rp_lock); + return to_dpu_kms(priv->kms); } static void dpu_crtc_destroy(struct drm_crtc *crtc) @@ -297,14 +77,29 @@ static void dpu_crtc_destroy(struct drm_crtc *crtc) } static void _dpu_crtc_setup_blend_cfg(struct dpu_crtc_mixer *mixer, - struct dpu_plane_state *pstate) + struct dpu_plane_state *pstate, struct dpu_format *format) { struct dpu_hw_mixer *lm = mixer->hw_lm; + uint32_t blend_op; + struct drm_format_name_buf format_name; /* default to opaque blending */ - lm->ops.setup_blend_config(lm, pstate->stage, 0XFF, 0, - DPU_BLEND_FG_ALPHA_FG_CONST | - DPU_BLEND_BG_ALPHA_BG_CONST); + blend_op = DPU_BLEND_FG_ALPHA_FG_CONST | + DPU_BLEND_BG_ALPHA_BG_CONST; + + if (format->alpha_enable) { + /* coverage blending */ + blend_op = DPU_BLEND_FG_ALPHA_FG_PIXEL | + DPU_BLEND_BG_ALPHA_FG_PIXEL | + DPU_BLEND_BG_INV_ALPHA; + } + + lm->ops.setup_blend_config(lm, pstate->stage, + 0xFF, 0, blend_op); + + DPU_DEBUG("format:%s, alpha_en:%u blend_op:0x%x\n", + drm_get_format_name(format->base.pixel_format, &format_name), + format->alpha_enable, blend_op); } static void _dpu_crtc_program_lm_output_roi(struct drm_crtc *crtc) @@ -317,9 +112,9 @@ static void _dpu_crtc_program_lm_output_roi(struct drm_crtc *crtc) crtc_state = to_dpu_crtc_state(crtc->state); lm_horiz_position = 0; - for (lm_idx = 0; lm_idx < dpu_crtc->num_mixers; lm_idx++) { + for (lm_idx = 0; lm_idx < crtc_state->num_mixers; lm_idx++) { const struct drm_rect *lm_roi = &crtc_state->lm_bounds[lm_idx]; - struct dpu_hw_mixer *hw_lm = dpu_crtc->mixers[lm_idx].hw_lm; + struct dpu_hw_mixer *hw_lm = crtc_state->mixers[lm_idx].hw_lm; struct dpu_hw_mixer_cfg cfg; if (!lm_roi || !drm_rect_visible(lm_roi)) @@ -339,28 +134,17 @@ static void _dpu_crtc_blend_setup_mixer(struct drm_crtc *crtc, struct drm_plane *plane; struct drm_framebuffer *fb; struct drm_plane_state *state; - struct dpu_crtc_state *cstate; + struct dpu_crtc_state *cstate = to_dpu_crtc_state(crtc->state); struct dpu_plane_state *pstate = NULL; struct dpu_format *format; - struct dpu_hw_ctl *ctl; - struct dpu_hw_mixer *lm; - struct dpu_hw_stage_cfg *stage_cfg; + struct dpu_hw_ctl *ctl = mixer->lm_ctl; + struct dpu_hw_stage_cfg *stage_cfg = &dpu_crtc->stage_cfg; u32 flush_mask; uint32_t stage_idx, lm_idx; int zpos_cnt[DPU_STAGE_MAX + 1] = { 0 }; bool bg_alpha_enable = false; - if (!dpu_crtc || !mixer) { - DPU_ERROR("invalid dpu_crtc or mixer\n"); - return; - } - - ctl = mixer->hw_ctl; - lm = mixer->hw_lm; - stage_cfg = &dpu_crtc->stage_cfg; - cstate = to_dpu_crtc_state(crtc->state); - drm_atomic_crtc_for_each_plane(plane, crtc) { state = plane->state; if (!state) @@ -379,10 +163,6 @@ static void _dpu_crtc_blend_setup_mixer(struct drm_crtc *crtc, state->fb ? state->fb->base.id : -1); format = to_dpu_format(msm_framebuffer_format(pstate->base.fb)); - if (!format) { - DPU_ERROR("invalid format\n"); - return; - } if (pstate->stage == DPU_STAGE_BASE && format->alpha_enable) bg_alpha_enable = true; @@ -400,8 +180,9 @@ static void _dpu_crtc_blend_setup_mixer(struct drm_crtc *crtc, fb ? fb->modifier : 0); /* blend config update */ - for (lm_idx = 0; lm_idx < dpu_crtc->num_mixers; lm_idx++) { - _dpu_crtc_setup_blend_cfg(mixer + lm_idx, pstate); + for (lm_idx = 0; lm_idx < cstate->num_mixers; lm_idx++) { + _dpu_crtc_setup_blend_cfg(mixer + lm_idx, + pstate, format); mixer[lm_idx].flush_mask |= flush_mask; @@ -422,38 +203,25 @@ static void _dpu_crtc_blend_setup_mixer(struct drm_crtc *crtc, */ static void _dpu_crtc_blend_setup(struct drm_crtc *crtc) { - struct dpu_crtc *dpu_crtc; - struct dpu_crtc_state *dpu_crtc_state; - struct dpu_crtc_mixer *mixer; + struct dpu_crtc *dpu_crtc = to_dpu_crtc(crtc); + struct dpu_crtc_state *cstate = to_dpu_crtc_state(crtc->state); + struct dpu_crtc_mixer *mixer = cstate->mixers; struct dpu_hw_ctl *ctl; struct dpu_hw_mixer *lm; - int i; - if (!crtc) - return; - - dpu_crtc = to_dpu_crtc(crtc); - dpu_crtc_state = to_dpu_crtc_state(crtc->state); - mixer = dpu_crtc->mixers; - DPU_DEBUG("%s\n", dpu_crtc->name); - if (dpu_crtc->num_mixers > CRTC_DUAL_MIXERS) { - DPU_ERROR("invalid number mixers: %d\n", dpu_crtc->num_mixers); - return; - } - - for (i = 0; i < dpu_crtc->num_mixers; i++) { - if (!mixer[i].hw_lm || !mixer[i].hw_ctl) { + for (i = 0; i < cstate->num_mixers; i++) { + if (!mixer[i].hw_lm || !mixer[i].lm_ctl) { DPU_ERROR("invalid lm or ctl assigned to mixer\n"); return; } mixer[i].mixer_op_mode = 0; mixer[i].flush_mask = 0; - if (mixer[i].hw_ctl->ops.clear_all_blendstages) - mixer[i].hw_ctl->ops.clear_all_blendstages( - mixer[i].hw_ctl); + if (mixer[i].lm_ctl->ops.clear_all_blendstages) + mixer[i].lm_ctl->ops.clear_all_blendstages( + mixer[i].lm_ctl); } /* initialize stage cfg */ @@ -461,8 +229,8 @@ static void _dpu_crtc_blend_setup(struct drm_crtc *crtc) _dpu_crtc_blend_setup_mixer(crtc, dpu_crtc, mixer); - for (i = 0; i < dpu_crtc->num_mixers; i++) { - ctl = mixer[i].hw_ctl; + for (i = 0; i < cstate->num_mixers; i++) { + ctl = mixer[i].lm_ctl; lm = mixer[i].hw_lm; lm->ops.setup_alpha_out(lm, mixer[i].mixer_op_mode); @@ -543,34 +311,13 @@ static void dpu_crtc_vblank_cb(void *data) static void dpu_crtc_frame_event_work(struct kthread_work *work) { - struct msm_drm_private *priv; - struct dpu_crtc_frame_event *fevent; - struct drm_crtc *crtc; - struct dpu_crtc *dpu_crtc; - struct dpu_kms *dpu_kms; + struct dpu_crtc_frame_event *fevent = container_of(work, + struct dpu_crtc_frame_event, work); + struct drm_crtc *crtc = fevent->crtc; + struct dpu_crtc *dpu_crtc = to_dpu_crtc(crtc); unsigned long flags; bool frame_done = false; - if (!work) { - DPU_ERROR("invalid work handle\n"); - return; - } - - fevent = container_of(work, struct dpu_crtc_frame_event, work); - if (!fevent->crtc || !fevent->crtc->state) { - DPU_ERROR("invalid crtc\n"); - return; - } - - crtc = fevent->crtc; - dpu_crtc = to_dpu_crtc(crtc); - - dpu_kms = _dpu_crtc_get_kms(crtc); - if (!dpu_kms) { - DPU_ERROR("invalid kms handle\n"); - return; - } - priv = dpu_kms->dev->dev_private; DPU_ATRACE_BEGIN("crtc_frame_event"); DRM_DEBUG_KMS("crtc%d event:%u ts:%lld\n", crtc->base.id, fevent->event, @@ -636,11 +383,6 @@ static void dpu_crtc_frame_event_cb(void *data, u32 event) unsigned long flags; u32 crtc_id; - if (!crtc || !crtc->dev || !crtc->dev->dev_private) { - DPU_ERROR("invalid parameters\n"); - return; - } - /* Nothing to do on idle event */ if (event & DPU_ENCODER_FRAME_EVENT_IDLE) return; @@ -683,7 +425,7 @@ static void _dpu_crtc_setup_mixer_for_encoder( struct drm_crtc *crtc, struct drm_encoder *enc) { - struct dpu_crtc *dpu_crtc = to_dpu_crtc(crtc); + struct dpu_crtc_state *cstate = to_dpu_crtc_state(crtc->state); struct dpu_kms *dpu_kms = _dpu_crtc_get_kms(crtc); struct dpu_rm *rm = &dpu_kms->rm; struct dpu_crtc_mixer *mixer; @@ -695,8 +437,8 @@ static void _dpu_crtc_setup_mixer_for_encoder( dpu_rm_init_hw_iter(&ctl_iter, enc->base.id, DPU_HW_BLK_CTL); /* Set up all the mixers and ctls reserved by this encoder */ - for (i = dpu_crtc->num_mixers; i < ARRAY_SIZE(dpu_crtc->mixers); i++) { - mixer = &dpu_crtc->mixers[i]; + for (i = cstate->num_mixers; i < ARRAY_SIZE(cstate->mixers); i++) { + mixer = &cstate->mixers[i]; if (!dpu_rm_get_hw(rm, &lm_iter)) break; @@ -706,14 +448,14 @@ static void _dpu_crtc_setup_mixer_for_encoder( if (!dpu_rm_get_hw(rm, &ctl_iter)) { DPU_DEBUG("no ctl assigned to lm %d, using previous\n", mixer->hw_lm->idx - LM_0); - mixer->hw_ctl = last_valid_ctl; + mixer->lm_ctl = last_valid_ctl; } else { - mixer->hw_ctl = (struct dpu_hw_ctl *)ctl_iter.hw; - last_valid_ctl = mixer->hw_ctl; + mixer->lm_ctl = (struct dpu_hw_ctl *)ctl_iter.hw; + last_valid_ctl = mixer->lm_ctl; } /* Shouldn't happen, mixers are always >= ctls */ - if (!mixer->hw_ctl) { + if (!mixer->lm_ctl) { DPU_ERROR("no valid ctls found for lm %d\n", mixer->hw_lm->idx - LM_0); return; @@ -721,11 +463,11 @@ static void _dpu_crtc_setup_mixer_for_encoder( mixer->encoder = enc; - dpu_crtc->num_mixers++; + cstate->num_mixers++; DPU_DEBUG("setup mixer %d: lm %d\n", i, mixer->hw_lm->idx - LM_0); DPU_DEBUG("setup mixer %d: ctl %d\n", - i, mixer->hw_ctl->idx - CTL_0); + i, mixer->lm_ctl->idx - CTL_0); } } @@ -734,10 +476,6 @@ static void _dpu_crtc_setup_mixers(struct drm_crtc *crtc) struct dpu_crtc *dpu_crtc = to_dpu_crtc(crtc); struct drm_encoder *enc; - dpu_crtc->num_mixers = 0; - dpu_crtc->mixers_swapped = false; - memset(dpu_crtc->mixers, 0, sizeof(dpu_crtc->mixers)); - mutex_lock(&dpu_crtc->crtc_lock); /* Check for mixers on all encoders attached to this crtc */ list_for_each_entry(enc, &crtc->dev->mode_config.encoder_list, head) { @@ -753,24 +491,13 @@ static void _dpu_crtc_setup_mixers(struct drm_crtc *crtc) static void _dpu_crtc_setup_lm_bounds(struct drm_crtc *crtc, struct drm_crtc_state *state) { - struct dpu_crtc *dpu_crtc; - struct dpu_crtc_state *cstate; - struct drm_display_mode *adj_mode; - u32 crtc_split_width; + struct dpu_crtc *dpu_crtc = to_dpu_crtc(crtc); + struct dpu_crtc_state *cstate = to_dpu_crtc_state(state); + struct drm_display_mode *adj_mode = &state->adjusted_mode; + u32 crtc_split_width = _dpu_crtc_get_mixer_width(cstate, adj_mode); int i; - if (!crtc || !state) { - DPU_ERROR("invalid args\n"); - return; - } - - dpu_crtc = to_dpu_crtc(crtc); - cstate = to_dpu_crtc_state(state); - - adj_mode = &state->adjusted_mode; - crtc_split_width = dpu_crtc_get_mixer_width(dpu_crtc, cstate, adj_mode); - - for (i = 0; i < dpu_crtc->num_mixers; i++) { + for (i = 0; i < cstate->num_mixers; i++) { struct drm_rect *r = &cstate->lm_bounds[i]; r->x1 = crtc_split_width * i; r->y1 = 0; @@ -787,6 +514,7 @@ static void dpu_crtc_atomic_begin(struct drm_crtc *crtc, struct drm_crtc_state *old_state) { struct dpu_crtc *dpu_crtc; + struct dpu_crtc_state *cstate; struct drm_encoder *encoder; struct drm_device *dev; unsigned long flags; @@ -806,10 +534,11 @@ static void dpu_crtc_atomic_begin(struct drm_crtc *crtc, DPU_DEBUG("crtc%d\n", crtc->base.id); dpu_crtc = to_dpu_crtc(crtc); + cstate = to_dpu_crtc_state(crtc->state); dev = crtc->dev; smmu_state = &dpu_crtc->smmu_state; - if (!dpu_crtc->num_mixers) { + if (!cstate->num_mixers) { _dpu_crtc_setup_mixers(crtc); _dpu_crtc_setup_lm_bounds(crtc, crtc->state); } @@ -836,7 +565,7 @@ static void dpu_crtc_atomic_begin(struct drm_crtc *crtc, * it means we are trying to flush a CRTC whose state is disabled: * nothing else needs to be done. */ - if (unlikely(!dpu_crtc->num_mixers)) + if (unlikely(!cstate->num_mixers)) return; _dpu_crtc_blend_setup(crtc); @@ -861,11 +590,6 @@ static void dpu_crtc_atomic_flush(struct drm_crtc *crtc, unsigned long flags; struct dpu_crtc_state *cstate; - if (!crtc || !crtc->dev || !crtc->dev->dev_private) { - DPU_ERROR("invalid crtc\n"); - return; - } - if (!crtc->state->enable) { DPU_DEBUG("crtc%d -> enable %d, skip atomic_flush\n", crtc->base.id, crtc->state->enable); @@ -900,7 +624,7 @@ static void dpu_crtc_atomic_flush(struct drm_crtc *crtc, * it means we are trying to flush a CRTC whose state is disabled: * nothing else needs to be done. */ - if (unlikely(!dpu_crtc->num_mixers)) + if (unlikely(!cstate->num_mixers)) return; /* @@ -951,8 +675,6 @@ static void dpu_crtc_destroy_state(struct drm_crtc *crtc, DPU_DEBUG("crtc%d\n", crtc->base.id); - _dpu_crtc_rp_destroy(&cstate->rp); - __drm_atomic_helper_crtc_destroy_state(state); kfree(cstate); @@ -960,15 +682,9 @@ static void dpu_crtc_destroy_state(struct drm_crtc *crtc, static int _dpu_crtc_wait_for_frame_done(struct drm_crtc *crtc) { - struct dpu_crtc *dpu_crtc; + struct dpu_crtc *dpu_crtc = to_dpu_crtc(crtc); int ret, rc = 0; - if (!crtc) { - DPU_ERROR("invalid argument\n"); - return -EINVAL; - } - dpu_crtc = to_dpu_crtc(crtc); - if (!atomic_read(&dpu_crtc->frame_pending)) { DPU_DEBUG("no frames pending\n"); return 0; @@ -989,35 +705,18 @@ static int _dpu_crtc_wait_for_frame_done(struct drm_crtc *crtc) void dpu_crtc_commit_kickoff(struct drm_crtc *crtc) { struct drm_encoder *encoder; - struct drm_device *dev; - struct dpu_crtc *dpu_crtc; - struct msm_drm_private *priv; - struct dpu_kms *dpu_kms; - struct dpu_crtc_state *cstate; + struct drm_device *dev = crtc->dev; + struct dpu_crtc *dpu_crtc = to_dpu_crtc(crtc); + struct dpu_kms *dpu_kms = _dpu_crtc_get_kms(crtc); + struct dpu_crtc_state *cstate = to_dpu_crtc_state(crtc->state); int ret; - if (!crtc) { - DPU_ERROR("invalid argument\n"); - return; - } - dev = crtc->dev; - dpu_crtc = to_dpu_crtc(crtc); - dpu_kms = _dpu_crtc_get_kms(crtc); - - if (!dpu_kms || !dpu_kms->dev || !dpu_kms->dev->dev_private) { - DPU_ERROR("invalid argument\n"); - return; - } - - priv = dpu_kms->dev->dev_private; - cstate = to_dpu_crtc_state(crtc->state); - /* * If no mixers has been allocated in dpu_crtc_atomic_check(), * it means we are trying to start a CRTC whose state is disabled: * nothing else needs to be done. */ - if (unlikely(!dpu_crtc->num_mixers)) + if (unlikely(!cstate->num_mixers)) return; DPU_ATRACE_BEGIN("crtc_commit"); @@ -1072,33 +771,19 @@ end: * _dpu_crtc_vblank_enable_no_lock - update power resource and vblank request * @dpu_crtc: Pointer to dpu crtc structure * @enable: Whether to enable/disable vblanks - * - * @Return: error code */ -static int _dpu_crtc_vblank_enable_no_lock( +static void _dpu_crtc_vblank_enable_no_lock( struct dpu_crtc *dpu_crtc, bool enable) { - struct drm_device *dev; - struct drm_crtc *crtc; + struct drm_crtc *crtc = &dpu_crtc->base; + struct drm_device *dev = crtc->dev; struct drm_encoder *enc; - if (!dpu_crtc) { - DPU_ERROR("invalid crtc\n"); - return -EINVAL; - } - - crtc = &dpu_crtc->base; - dev = crtc->dev; - if (enable) { - int ret; - /* drop lock since power crtc cb may try to re-acquire lock */ mutex_unlock(&dpu_crtc->crtc_lock); - ret = _dpu_crtc_power_enable(dpu_crtc, true); + pm_runtime_get_sync(dev->dev); mutex_lock(&dpu_crtc->crtc_lock); - if (ret) - return ret; list_for_each_entry(enc, &dev->mode_config.encoder_list, head) { if (enc->crtc != crtc) @@ -1125,11 +810,9 @@ static int _dpu_crtc_vblank_enable_no_lock( /* drop lock since power crtc cb may try to re-acquire lock */ mutex_unlock(&dpu_crtc->crtc_lock); - _dpu_crtc_power_enable(dpu_crtc, false); + pm_runtime_put_sync(dev->dev); mutex_lock(&dpu_crtc->crtc_lock); } - - return 0; } /** @@ -1139,23 +822,7 @@ static int _dpu_crtc_vblank_enable_no_lock( */ static void _dpu_crtc_set_suspend(struct drm_crtc *crtc, bool enable) { - struct dpu_crtc *dpu_crtc; - struct msm_drm_private *priv; - struct dpu_kms *dpu_kms; - int ret = 0; - - if (!crtc || !crtc->dev || !crtc->dev->dev_private) { - DPU_ERROR("invalid crtc\n"); - return; - } - dpu_crtc = to_dpu_crtc(crtc); - priv = crtc->dev->dev_private; - - if (!priv->kms) { - DPU_ERROR("invalid crtc kms\n"); - return; - } - dpu_kms = to_dpu_kms(priv->kms); + struct dpu_crtc *dpu_crtc = to_dpu_crtc(crtc); DRM_DEBUG_KMS("crtc%d suspend = %d\n", crtc->base.id, enable); @@ -1170,10 +837,7 @@ static void _dpu_crtc_set_suspend(struct drm_crtc *crtc, bool enable) DPU_DEBUG("crtc%d suspend already set to %d, ignoring update\n", crtc->base.id, enable); else if (dpu_crtc->enabled && dpu_crtc->vblank_requested) { - ret = _dpu_crtc_vblank_enable_no_lock(dpu_crtc, !enable); - if (ret) - DPU_ERROR("%s vblank enable failed: %d\n", - dpu_crtc->name, ret); + _dpu_crtc_vblank_enable_no_lock(dpu_crtc, !enable); } dpu_crtc->suspend = enable; @@ -1206,8 +870,6 @@ static struct drm_crtc_state *dpu_crtc_duplicate_state(struct drm_crtc *crtc) /* duplicate base helper */ __drm_atomic_helper_crtc_duplicate_state(crtc, &cstate->base); - _dpu_crtc_rp_duplicate(&old_cstate->rp, &cstate->rp); - return &cstate->base; } @@ -1244,9 +906,6 @@ static void dpu_crtc_reset(struct drm_crtc *crtc) return; } - _dpu_crtc_rp_reset(&cstate->rp, &dpu_crtc->rp_lock, - &dpu_crtc->rp_head); - cstate->base.crtc = crtc; crtc->state = &cstate->base; } @@ -1254,62 +913,19 @@ static void dpu_crtc_reset(struct drm_crtc *crtc) static void dpu_crtc_handle_power_event(u32 event_type, void *arg) { struct drm_crtc *crtc = arg; - struct dpu_crtc *dpu_crtc; + struct dpu_crtc *dpu_crtc = to_dpu_crtc(crtc); struct drm_encoder *encoder; - struct dpu_crtc_mixer *m; - u32 i, misr_status; - - if (!crtc) { - DPU_ERROR("invalid crtc\n"); - return; - } - dpu_crtc = to_dpu_crtc(crtc); mutex_lock(&dpu_crtc->crtc_lock); trace_dpu_crtc_handle_power_event(DRMID(crtc), event_type); - switch (event_type) { - case DPU_POWER_EVENT_POST_ENABLE: - /* restore encoder; crtc will be programmed during commit */ - drm_for_each_encoder(encoder, crtc->dev) { - if (encoder->crtc != crtc) - continue; - - dpu_encoder_virt_restore(encoder); - } - - for (i = 0; i < dpu_crtc->num_mixers; ++i) { - m = &dpu_crtc->mixers[i]; - if (!m->hw_lm || !m->hw_lm->ops.setup_misr || - !dpu_crtc->misr_enable) - continue; - - m->hw_lm->ops.setup_misr(m->hw_lm, true, - dpu_crtc->misr_frame_count); - } - break; - case DPU_POWER_EVENT_PRE_DISABLE: - for (i = 0; i < dpu_crtc->num_mixers; ++i) { - m = &dpu_crtc->mixers[i]; - if (!m->hw_lm || !m->hw_lm->ops.collect_misr || - !dpu_crtc->misr_enable) - continue; + /* restore encoder; crtc will be programmed during commit */ + drm_for_each_encoder(encoder, crtc->dev) { + if (encoder->crtc != crtc) + continue; - misr_status = m->hw_lm->ops.collect_misr(m->hw_lm); - dpu_crtc->misr_data[i] = misr_status ? misr_status : - dpu_crtc->misr_data[i]; - } - break; - case DPU_POWER_EVENT_POST_DISABLE: - /** - * Nothing to do. All the planes on the CRTC will be - * programmed for every frame - */ - break; - default: - DPU_DEBUG("event:%d not handled\n", event_type); - break; + dpu_encoder_virt_restore(encoder); } mutex_unlock(&dpu_crtc->crtc_lock); @@ -1322,7 +938,6 @@ static void dpu_crtc_disable(struct drm_crtc *crtc) struct drm_display_mode *mode; struct drm_encoder *encoder; struct msm_drm_private *priv; - int ret; unsigned long flags; if (!crtc || !crtc->dev || !crtc->dev->dev_private || !crtc->state) { @@ -1353,10 +968,7 @@ static void dpu_crtc_disable(struct drm_crtc *crtc) trace_dpu_crtc_disable(DRMID(crtc), false, dpu_crtc); if (dpu_crtc->enabled && !dpu_crtc->suspend && dpu_crtc->vblank_requested) { - ret = _dpu_crtc_vblank_enable_no_lock(dpu_crtc, false); - if (ret) - DPU_ERROR("%s vblank enable failed: %d\n", - dpu_crtc->name, ret); + _dpu_crtc_vblank_enable_no_lock(dpu_crtc, false); } dpu_crtc->enabled = false; @@ -1379,9 +991,8 @@ static void dpu_crtc_disable(struct drm_crtc *crtc) dpu_power_handle_unregister_event(dpu_crtc->phandle, dpu_crtc->power_event); - memset(dpu_crtc->mixers, 0, sizeof(dpu_crtc->mixers)); - dpu_crtc->num_mixers = 0; - dpu_crtc->mixers_swapped = false; + memset(cstate->mixers, 0, sizeof(cstate->mixers)); + cstate->num_mixers = 0; /* disable clk & bw control until clk & bw properties are set */ cstate->bw_control = false; @@ -1403,7 +1014,6 @@ static void dpu_crtc_enable(struct drm_crtc *crtc, struct dpu_crtc *dpu_crtc; struct drm_encoder *encoder; struct msm_drm_private *priv; - int ret; if (!crtc || !crtc->dev || !crtc->dev->dev_private) { DPU_ERROR("invalid crtc\n"); @@ -1425,10 +1035,7 @@ static void dpu_crtc_enable(struct drm_crtc *crtc, trace_dpu_crtc_enable(DRMID(crtc), true, dpu_crtc); if (!dpu_crtc->enabled && !dpu_crtc->suspend && dpu_crtc->vblank_requested) { - ret = _dpu_crtc_vblank_enable_no_lock(dpu_crtc, true); - if (ret) - DPU_ERROR("%s vblank enable failed: %d\n", - dpu_crtc->name, ret); + _dpu_crtc_vblank_enable_no_lock(dpu_crtc, true); } dpu_crtc->enabled = true; @@ -1438,9 +1045,7 @@ static void dpu_crtc_enable(struct drm_crtc *crtc, drm_crtc_vblank_on(crtc); dpu_crtc->power_event = dpu_power_handle_register_event( - dpu_crtc->phandle, - DPU_POWER_EVENT_POST_ENABLE | DPU_POWER_EVENT_POST_DISABLE | - DPU_POWER_EVENT_PRE_DISABLE, + dpu_crtc->phandle, DPU_POWER_EVENT_ENABLE, dpu_crtc_handle_power_event, crtc, dpu_crtc->name); } @@ -1496,7 +1101,7 @@ static int dpu_crtc_atomic_check(struct drm_crtc *crtc, memset(pipe_staged, 0, sizeof(pipe_staged)); - mixer_width = dpu_crtc_get_mixer_width(dpu_crtc, cstate, mode); + mixer_width = _dpu_crtc_get_mixer_width(cstate, mode); _dpu_crtc_setup_lm_bounds(crtc, state); @@ -1535,8 +1140,7 @@ static int dpu_crtc_atomic_check(struct drm_crtc *crtc, cnt++; dst = drm_plane_state_dest(pstate); - if (!drm_rect_intersect(&clip, &dst) || - !drm_rect_equals(&clip, &dst)) { + if (!drm_rect_intersect(&clip, &dst)) { DPU_ERROR("invalid vertical/horizontal destination\n"); DPU_ERROR("display: " DRM_RECT_FMT " plane: " DRM_RECT_FMT "\n", DRM_RECT_ARG(&crtc_rect), @@ -1679,7 +1283,6 @@ static int dpu_crtc_atomic_check(struct drm_crtc *crtc, } end: - _dpu_crtc_rp_free_unused(&cstate->rp); kfree(pstates); return rc; } @@ -1687,7 +1290,6 @@ end: int dpu_crtc_vblank(struct drm_crtc *crtc, bool en) { struct dpu_crtc *dpu_crtc; - int ret; if (!crtc) { DPU_ERROR("invalid crtc\n"); @@ -1698,10 +1300,7 @@ int dpu_crtc_vblank(struct drm_crtc *crtc, bool en) mutex_lock(&dpu_crtc->crtc_lock); trace_dpu_crtc_vblank(DRMID(&dpu_crtc->base), en, dpu_crtc); if (dpu_crtc->enabled && !dpu_crtc->suspend) { - ret = _dpu_crtc_vblank_enable_no_lock(dpu_crtc, en); - if (ret) - DPU_ERROR("%s vblank enable failed: %d\n", - dpu_crtc->name, ret); + _dpu_crtc_vblank_enable_no_lock(dpu_crtc, en); } dpu_crtc->vblank_requested = en; mutex_unlock(&dpu_crtc->crtc_lock); @@ -1730,26 +1329,28 @@ static int _dpu_debugfs_status_show(struct seq_file *s, void *data) dpu_crtc = s->private; crtc = &dpu_crtc->base; + + drm_modeset_lock_all(crtc->dev); cstate = to_dpu_crtc_state(crtc->state); mutex_lock(&dpu_crtc->crtc_lock); mode = &crtc->state->adjusted_mode; - out_width = dpu_crtc_get_mixer_width(dpu_crtc, cstate, mode); + out_width = _dpu_crtc_get_mixer_width(cstate, mode); seq_printf(s, "crtc:%d width:%d height:%d\n", crtc->base.id, mode->hdisplay, mode->vdisplay); seq_puts(s, "\n"); - for (i = 0; i < dpu_crtc->num_mixers; ++i) { - m = &dpu_crtc->mixers[i]; + for (i = 0; i < cstate->num_mixers; ++i) { + m = &cstate->mixers[i]; if (!m->hw_lm) seq_printf(s, "\tmixer[%d] has no lm\n", i); - else if (!m->hw_ctl) + else if (!m->lm_ctl) seq_printf(s, "\tmixer[%d] has no ctl\n", i); else seq_printf(s, "\tmixer:%d ctl:%d width:%d height:%d\n", - m->hw_lm->idx - LM_0, m->hw_ctl->idx - CTL_0, + m->hw_lm->idx - LM_0, m->lm_ctl->idx - CTL_0, out_width, mode->vdisplay); } @@ -1822,6 +1423,7 @@ static int _dpu_debugfs_status_show(struct seq_file *s, void *data) seq_printf(s, "vblank_enable:%d\n", dpu_crtc->vblank_requested); mutex_unlock(&dpu_crtc->crtc_lock); + drm_modeset_unlock_all(crtc->dev); return 0; } @@ -1831,113 +1433,6 @@ static int _dpu_debugfs_status_open(struct inode *inode, struct file *file) return single_open(file, _dpu_debugfs_status_show, inode->i_private); } -static ssize_t _dpu_crtc_misr_setup(struct file *file, - const char __user *user_buf, size_t count, loff_t *ppos) -{ - struct dpu_crtc *dpu_crtc; - struct dpu_crtc_mixer *m; - int i = 0, rc; - char buf[MISR_BUFF_SIZE + 1]; - u32 frame_count, enable; - size_t buff_copy; - - if (!file || !file->private_data) - return -EINVAL; - - dpu_crtc = file->private_data; - buff_copy = min_t(size_t, count, MISR_BUFF_SIZE); - if (copy_from_user(buf, user_buf, buff_copy)) { - DPU_ERROR("buffer copy failed\n"); - return -EINVAL; - } - - buf[buff_copy] = 0; /* end of string */ - - if (sscanf(buf, "%u %u", &enable, &frame_count) != 2) - return -EINVAL; - - rc = _dpu_crtc_power_enable(dpu_crtc, true); - if (rc) - return rc; - - mutex_lock(&dpu_crtc->crtc_lock); - dpu_crtc->misr_enable = enable; - dpu_crtc->misr_frame_count = frame_count; - for (i = 0; i < dpu_crtc->num_mixers; ++i) { - dpu_crtc->misr_data[i] = 0; - m = &dpu_crtc->mixers[i]; - if (!m->hw_lm || !m->hw_lm->ops.setup_misr) - continue; - - m->hw_lm->ops.setup_misr(m->hw_lm, enable, frame_count); - } - mutex_unlock(&dpu_crtc->crtc_lock); - _dpu_crtc_power_enable(dpu_crtc, false); - - return count; -} - -static ssize_t _dpu_crtc_misr_read(struct file *file, - char __user *user_buff, size_t count, loff_t *ppos) -{ - struct dpu_crtc *dpu_crtc; - struct dpu_crtc_mixer *m; - int i = 0, rc; - u32 misr_status; - ssize_t len = 0; - char buf[MISR_BUFF_SIZE + 1] = {'\0'}; - - if (*ppos) - return 0; - - if (!file || !file->private_data) - return -EINVAL; - - dpu_crtc = file->private_data; - rc = _dpu_crtc_power_enable(dpu_crtc, true); - if (rc) - return rc; - - mutex_lock(&dpu_crtc->crtc_lock); - if (!dpu_crtc->misr_enable) { - len += snprintf(buf + len, MISR_BUFF_SIZE - len, - "disabled\n"); - goto buff_check; - } - - for (i = 0; i < dpu_crtc->num_mixers; ++i) { - m = &dpu_crtc->mixers[i]; - if (!m->hw_lm || !m->hw_lm->ops.collect_misr) - continue; - - misr_status = m->hw_lm->ops.collect_misr(m->hw_lm); - dpu_crtc->misr_data[i] = misr_status ? misr_status : - dpu_crtc->misr_data[i]; - len += snprintf(buf + len, MISR_BUFF_SIZE - len, "lm idx:%d\n", - m->hw_lm->idx - LM_0); - len += snprintf(buf + len, MISR_BUFF_SIZE - len, "0x%x\n", - dpu_crtc->misr_data[i]); - } - -buff_check: - if (count <= len) { - len = 0; - goto end; - } - - if (copy_to_user(user_buff, buf, len)) { - len = -EFAULT; - goto end; - } - - *ppos += len; /* increase offset */ - -end: - mutex_unlock(&dpu_crtc->crtc_lock); - _dpu_crtc_power_enable(dpu_crtc, false); - return len; -} - #define DEFINE_DPU_DEBUGFS_SEQ_FOPS(__prefix) \ static int __prefix ## _open(struct inode *inode, struct file *file) \ { \ @@ -1955,8 +1450,6 @@ static int dpu_crtc_debugfs_state_show(struct seq_file *s, void *v) { struct drm_crtc *crtc = (struct drm_crtc *) s->private; struct dpu_crtc *dpu_crtc = to_dpu_crtc(crtc); - struct dpu_crtc_res *res; - struct dpu_crtc_respool *rp; int i; seq_printf(s, "client type: %d\n", dpu_crtc_get_client_type(crtc)); @@ -1973,17 +1466,6 @@ static int dpu_crtc_debugfs_state_show(struct seq_file *s, void *v) dpu_crtc->cur_perf.max_per_pipe_ib[i]); } - mutex_lock(&dpu_crtc->rp_lock); - list_for_each_entry(rp, &dpu_crtc->rp_head, rp_list) { - seq_printf(s, "rp.%d: ", rp->sequence_id); - list_for_each_entry(res, &rp->res_list, list) - seq_printf(s, "0x%x/0x%llx/%pK/%d ", - res->type, res->tag, res->val, - atomic_read(&res->refcount)); - seq_puts(s, "\n"); - } - mutex_unlock(&dpu_crtc->rp_lock); - return 0; } DEFINE_DPU_DEBUGFS_SEQ_FOPS(dpu_crtc_debugfs_state); @@ -1999,19 +1481,12 @@ static int _dpu_crtc_init_debugfs(struct drm_crtc *crtc) .llseek = seq_lseek, .release = single_release, }; - static const struct file_operations debugfs_misr_fops = { - .open = simple_open, - .read = _dpu_crtc_misr_read, - .write = _dpu_crtc_misr_setup, - }; if (!crtc) return -EINVAL; dpu_crtc = to_dpu_crtc(crtc); dpu_kms = _dpu_crtc_get_kms(crtc); - if (!dpu_kms) - return -EINVAL; dpu_crtc->debugfs_root = debugfs_create_dir(dpu_crtc->name, crtc->dev->primary->debugfs_root); @@ -2026,8 +1501,6 @@ static int _dpu_crtc_init_debugfs(struct drm_crtc *crtc) dpu_crtc->debugfs_root, &dpu_crtc->base, &dpu_crtc_debugfs_state_fops); - debugfs_create_file("misr_data", 0600, dpu_crtc->debugfs_root, - dpu_crtc, &debugfs_misr_fops); return 0; } @@ -2082,7 +1555,8 @@ static const struct drm_crtc_helper_funcs dpu_crtc_helper_funcs = { }; /* initialize crtc */ -struct drm_crtc *dpu_crtc_init(struct drm_device *dev, struct drm_plane *plane) +struct drm_crtc *dpu_crtc_init(struct drm_device *dev, struct drm_plane *plane, + struct drm_plane *cursor) { struct drm_crtc *crtc = NULL; struct dpu_crtc *dpu_crtc = NULL; @@ -2104,9 +1578,6 @@ struct drm_crtc *dpu_crtc_init(struct drm_device *dev, struct drm_plane *plane) spin_lock_init(&dpu_crtc->spin_lock); atomic_set(&dpu_crtc->frame_pending, 0); - mutex_init(&dpu_crtc->rp_lock); - INIT_LIST_HEAD(&dpu_crtc->rp_head); - init_completion(&dpu_crtc->frame_done_comp); INIT_LIST_HEAD(&dpu_crtc->frame_event_list); @@ -2119,7 +1590,7 @@ struct drm_crtc *dpu_crtc_init(struct drm_device *dev, struct drm_plane *plane) dpu_crtc_frame_event_work); } - drm_crtc_init_with_planes(dev, crtc, plane, NULL, &dpu_crtc_funcs, + drm_crtc_init_with_planes(dev, crtc, plane, cursor, &dpu_crtc_funcs, NULL); drm_crtc_helper_add(crtc, &dpu_crtc_helper_funcs); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.h index e87109e608e9..3723b4830335 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.h @@ -83,14 +83,14 @@ struct dpu_crtc_smmu_state_data { /** * struct dpu_crtc_mixer: stores the map for each virtual pipeline in the CRTC * @hw_lm: LM HW Driver context - * @hw_ctl: CTL Path HW driver context + * @lm_ctl: CTL Path HW driver context * @encoder: Encoder attached to this lm & ctl * @mixer_op_mode: mixer blending operation mode * @flush_mask: mixer flush mask for ctl, mixer and pipe */ struct dpu_crtc_mixer { struct dpu_hw_mixer *hw_lm; - struct dpu_hw_ctl *hw_ctl; + struct dpu_hw_ctl *lm_ctl; struct drm_encoder *encoder; u32 mixer_op_mode; u32 flush_mask; @@ -121,11 +121,6 @@ struct dpu_crtc_frame_event { * struct dpu_crtc - virtualized CRTC data structure * @base : Base drm crtc structure * @name : ASCII description of this crtc - * @num_ctls : Number of ctl paths in use - * @num_mixers : Number of mixers in use - * @mixers_swapped: Whether the mixers have been swapped for left/right update - * especially in the case of DSC Merge. - * @mixers : List of active mixers * @event : Pointer to last received drm vblank event. If there is a * pending vblank event, this will be non-null. * @vsync_count : Running count of received vsync events @@ -156,27 +151,14 @@ struct dpu_crtc_frame_event { * @event_thread : Pointer to event handler thread * @event_worker : Event worker queue * @event_lock : Spinlock around event handling code - * @misr_enable : boolean entry indicates misr enable/disable status. - * @misr_frame_count : misr frame count provided by client - * @misr_data : store misr data before turning off the clocks. * @phandle: Pointer to power handler * @power_event : registered power event handle * @cur_perf : current performance committed to clock/bandwidth driver - * @rp_lock : serialization lock for resource pool - * @rp_head : list of active resource pool - * @scl3_cfg_lut : qseed3 lut config */ struct dpu_crtc { struct drm_crtc base; char name[DPU_CRTC_NAME_SIZE]; - /* HW Resources reserved for the crtc */ - u32 num_ctls; - u32 num_mixers; - bool mixers_swapped; - struct dpu_crtc_mixer mixers[CRTC_DUAL_MIXERS]; - struct dpu_hw_scaler3_lut_cfg *scl3_lut_cfg; - struct drm_pending_vblank_event *event; u32 vsync_count; @@ -206,77 +188,20 @@ struct dpu_crtc { /* for handling internal event thread */ spinlock_t event_lock; - bool misr_enable; - u32 misr_frame_count; - u32 misr_data[CRTC_DUAL_MIXERS]; struct dpu_power_handle *phandle; struct dpu_power_event *power_event; struct dpu_core_perf_params cur_perf; - struct mutex rp_lock; - struct list_head rp_head; - struct dpu_crtc_smmu_state_data smmu_state; }; #define to_dpu_crtc(x) container_of(x, struct dpu_crtc, base) /** - * struct dpu_crtc_res_ops - common operations for crtc resources - * @get: get given resource - * @put: put given resource - */ -struct dpu_crtc_res_ops { - void *(*get)(void *val, u32 type, u64 tag); - void (*put)(void *val); -}; - -#define DPU_CRTC_RES_FLAG_FREE BIT(0) - -/** - * struct dpu_crtc_res - definition of crtc resources - * @list: list of crtc resource - * @type: crtc resource type - * @tag: unique identifier per type - * @refcount: reference/usage count - * @ops: callback operations - * @val: resource handle associated with type/tag - * @flags: customization flags - */ -struct dpu_crtc_res { - struct list_head list; - u32 type; - u64 tag; - atomic_t refcount; - struct dpu_crtc_res_ops ops; - void *val; - u32 flags; -}; - -/** - * dpu_crtc_respool - crtc resource pool - * @rp_lock: pointer to serialization lock - * @rp_head: pointer to head of active resource pools of this crtc - * @rp_list: list of crtc resource pool - * @sequence_id: sequence identifier, incremented per state duplication - * @res_list: list of resource managed by this resource pool - * @ops: resource operations for parent resource pool - */ -struct dpu_crtc_respool { - struct mutex *rp_lock; - struct list_head *rp_head; - struct list_head rp_list; - u32 sequence_id; - struct list_head res_list; - struct dpu_crtc_res_ops ops; -}; - -/** * struct dpu_crtc_state - dpu container for atomic crtc state * @base: Base drm crtc state structure - * @is_ppsplit : Whether current topology requires PPSplit special handling * @bw_control : true if bw/clk controlled by core bw/clk properties * @bw_split_vote : true if bw controlled by llcc/dram bw properties * @lm_bounds : LM boundaries based on current mode full resolution, no ROI. @@ -285,41 +210,41 @@ struct dpu_crtc_respool { * @property_values: Current crtc property values * @input_fence_timeout_ns : Cached input fence timeout, in ns * @new_perf: new performance state being requested + * @num_mixers : Number of mixers in use + * @mixers : List of active mixers + * @num_ctls : Number of ctl paths in use + * @hw_ctls : List of active ctl paths */ struct dpu_crtc_state { struct drm_crtc_state base; bool bw_control; bool bw_split_vote; - - bool is_ppsplit; struct drm_rect lm_bounds[CRTC_DUAL_MIXERS]; uint64_t input_fence_timeout_ns; struct dpu_core_perf_params new_perf; - struct dpu_crtc_respool rp; + + /* HW Resources reserved for the crtc */ + u32 num_mixers; + struct dpu_crtc_mixer mixers[CRTC_DUAL_MIXERS]; + + u32 num_ctls; + struct dpu_hw_ctl *hw_ctls[CRTC_DUAL_MIXERS]; }; #define to_dpu_crtc_state(x) \ container_of(x, struct dpu_crtc_state, base) /** - * dpu_crtc_get_mixer_width - get the mixer width - * Mixer width will be same as panel width(/2 for split) + * dpu_crtc_state_is_stereo - Is crtc virtualized with two mixers? + * @cstate: Pointer to dpu crtc state + * @Return: true - has two mixers, false - has one mixer */ -static inline int dpu_crtc_get_mixer_width(struct dpu_crtc *dpu_crtc, - struct dpu_crtc_state *cstate, struct drm_display_mode *mode) +static inline bool dpu_crtc_state_is_stereo(struct dpu_crtc_state *cstate) { - u32 mixer_width; - - if (!dpu_crtc || !cstate || !mode) - return 0; - - mixer_width = (dpu_crtc->num_mixers == CRTC_DUAL_MIXERS ? - mode->hdisplay / CRTC_DUAL_MIXERS : mode->hdisplay); - - return mixer_width; + return cstate->num_mixers == CRTC_DUAL_MIXERS; } /** @@ -375,9 +300,11 @@ void dpu_crtc_complete_commit(struct drm_crtc *crtc, * dpu_crtc_init - create a new crtc object * @dev: dpu device * @plane: base plane + * @cursor: cursor plane * @Return: new crtc object or error */ -struct drm_crtc *dpu_crtc_init(struct drm_device *dev, struct drm_plane *plane); +struct drm_crtc *dpu_crtc_init(struct drm_device *dev, struct drm_plane *plane, + struct drm_plane *cursor); /** * dpu_crtc_register_custom_event - api for enabling/disabling crtc event diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c index 1b4de3486ef9..96cdf06e7da2 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c @@ -65,8 +65,6 @@ #define MAX_CHANNELS_PER_ENC 2 -#define MISR_BUFF_SIZE 256 - #define IDLE_SHORT_TIMEOUT 1 #define MAX_VDISPLAY_SPLIT 1080 @@ -161,8 +159,6 @@ enum dpu_enc_rc_states { * @frame_done_timer: watchdog timer for frame done event * @vsync_event_timer: vsync timer * @disp_info: local copy of msm_display_info struct - * @misr_enable: misr enable/disable status - * @misr_frame_count: misr frame count before start capturing the data * @idle_pc_supported: indicate if idle power collaps is supported * @rc_lock: resource control mutex lock to protect * virt encoder over various state changes @@ -179,11 +175,10 @@ struct dpu_encoder_virt { spinlock_t enc_spinlock; uint32_t bus_scaling_client; - uint32_t display_num_of_h_tiles; - unsigned int num_phys_encs; struct dpu_encoder_phys *phys_encs[MAX_PHYS_ENCODERS_PER_VIRTUAL]; struct dpu_encoder_phys *cur_master; + struct dpu_encoder_phys *cur_slave; struct dpu_hw_pingpong *hw_pp[MAX_CHANNELS_PER_ENC]; bool intfs_swapped; @@ -202,8 +197,6 @@ struct dpu_encoder_virt { struct timer_list vsync_event_timer; struct msm_display_info disp_info; - bool misr_enable; - u32 misr_frame_count; bool idle_pc_supported; struct mutex rc_lock; @@ -443,30 +436,22 @@ int dpu_encoder_helper_unregister_irq(struct dpu_encoder_phys *phys_enc, } void dpu_encoder_get_hw_resources(struct drm_encoder *drm_enc, - struct dpu_encoder_hw_resources *hw_res, - struct drm_connector_state *conn_state) + struct dpu_encoder_hw_resources *hw_res) { struct dpu_encoder_virt *dpu_enc = NULL; int i = 0; - if (!hw_res || !drm_enc || !conn_state) { - DPU_ERROR("invalid argument(s), drm_enc %d, res %d, state %d\n", - drm_enc != 0, hw_res != 0, conn_state != 0); - return; - } - dpu_enc = to_dpu_encoder_virt(drm_enc); DPU_DEBUG_ENC(dpu_enc, "\n"); /* Query resources used by phys encs, expected to be without overlap */ memset(hw_res, 0, sizeof(*hw_res)); - hw_res->display_num_of_h_tiles = dpu_enc->display_num_of_h_tiles; for (i = 0; i < dpu_enc->num_phys_encs; i++) { struct dpu_encoder_phys *phys = dpu_enc->phys_encs[i]; if (phys && phys->ops.get_hw_resources) - phys->ops.get_hw_resources(phys, hw_res, conn_state); + phys->ops.get_hw_resources(phys, hw_res); } } @@ -525,7 +510,7 @@ void dpu_encoder_helper_split_config( hw_mdptop = phys_enc->hw_mdptop; disp_info = &dpu_enc->disp_info; - if (disp_info->intf_type != DRM_MODE_CONNECTOR_DSI) + if (disp_info->intf_type != DRM_MODE_ENCODER_DSI) return; /** @@ -660,7 +645,7 @@ static int dpu_encoder_virt_atomic_check( if (drm_atomic_crtc_needs_modeset(crtc_state) && dpu_enc->mode_set_complete) { ret = dpu_rm_reserve(&dpu_kms->rm, drm_enc, crtc_state, - conn_state, topology, true); + topology, true); dpu_enc->mode_set_complete = false; } } @@ -1016,9 +1001,9 @@ static void dpu_encoder_virt_mode_set(struct drm_encoder *drm_enc, struct dpu_kms *dpu_kms; struct list_head *connector_list; struct drm_connector *conn = NULL, *conn_iter; - struct dpu_rm_hw_iter pp_iter; + struct dpu_rm_hw_iter pp_iter, ctl_iter; struct msm_display_topology topology; - enum dpu_rm_topology_name topology_name; + struct dpu_hw_ctl *hw_ctl[MAX_CHANNELS_PER_ENC] = { NULL }; int i = 0, ret; if (!drm_enc) { @@ -1051,7 +1036,7 @@ static void dpu_encoder_virt_mode_set(struct drm_encoder *drm_enc, /* Reserve dynamic resources now. Indicating non-AtomicTest phase */ ret = dpu_rm_reserve(&dpu_kms->rm, drm_enc, drm_enc->crtc->state, - conn->state, topology, false); + topology, false); if (ret) { DPU_ERROR_ENC(dpu_enc, "failed to reserve hw resources, %d\n", ret); @@ -1066,19 +1051,33 @@ static void dpu_encoder_virt_mode_set(struct drm_encoder *drm_enc, dpu_enc->hw_pp[i] = (struct dpu_hw_pingpong *) pp_iter.hw; } - topology_name = dpu_rm_get_topology_name(topology); + dpu_rm_init_hw_iter(&ctl_iter, drm_enc->base.id, DPU_HW_BLK_CTL); + for (i = 0; i < MAX_CHANNELS_PER_ENC; i++) { + if (!dpu_rm_get_hw(&dpu_kms->rm, &ctl_iter)) + break; + hw_ctl[i] = (struct dpu_hw_ctl *)ctl_iter.hw; + } + for (i = 0; i < dpu_enc->num_phys_encs; i++) { struct dpu_encoder_phys *phys = dpu_enc->phys_encs[i]; if (phys) { if (!dpu_enc->hw_pp[i]) { - DPU_ERROR_ENC(dpu_enc, - "invalid pingpong block for the encoder\n"); + DPU_ERROR_ENC(dpu_enc, "no pp block assigned" + "at idx: %d\n", i); return; } + + if (!hw_ctl[i]) { + DPU_ERROR_ENC(dpu_enc, "no ctl block assigned" + "at idx: %d\n", i); + return; + } + phys->hw_pp = dpu_enc->hw_pp[i]; + phys->hw_ctl = hw_ctl[i]; + phys->connector = conn->state->connector; - phys->topology_name = topology_name; if (phys->ops.mode_set) phys->ops.mode_set(phys, mode, adj_mode); } @@ -1111,12 +1110,6 @@ static void _dpu_encoder_virt_enable_helper(struct drm_encoder *drm_enc) return; } - if (dpu_enc->disp_info.intf_type == DRM_MODE_CONNECTOR_DisplayPort && - dpu_enc->cur_master->hw_mdptop && - dpu_enc->cur_master->hw_mdptop->ops.intf_audio_select) - dpu_enc->cur_master->hw_mdptop->ops.intf_audio_select( - dpu_enc->cur_master->hw_mdptop); - if (dpu_enc->cur_master->hw_mdptop && dpu_enc->cur_master->hw_mdptop->ops.reset_ubwc) dpu_enc->cur_master->hw_mdptop->ops.reset_ubwc( @@ -1153,7 +1146,7 @@ void dpu_encoder_virt_restore(struct drm_encoder *drm_enc) static void dpu_encoder_virt_enable(struct drm_encoder *drm_enc) { struct dpu_encoder_virt *dpu_enc = NULL; - int i, ret = 0; + int ret = 0; struct drm_display_mode *cur_mode = NULL; if (!drm_enc) { @@ -1166,21 +1159,12 @@ static void dpu_encoder_virt_enable(struct drm_encoder *drm_enc) trace_dpu_enc_enable(DRMID(drm_enc), cur_mode->hdisplay, cur_mode->vdisplay); - dpu_enc->cur_master = NULL; - for (i = 0; i < dpu_enc->num_phys_encs; i++) { - struct dpu_encoder_phys *phys = dpu_enc->phys_encs[i]; + /* always enable slave encoder before master */ + if (dpu_enc->cur_slave && dpu_enc->cur_slave->ops.enable) + dpu_enc->cur_slave->ops.enable(dpu_enc->cur_slave); - if (phys && phys->ops.is_master && phys->ops.is_master(phys)) { - DPU_DEBUG_ENC(dpu_enc, "master is now idx %d\n", i); - dpu_enc->cur_master = phys; - break; - } - } - - if (!dpu_enc->cur_master) { - DPU_ERROR("virt encoder has no master! num_phys %d\n", i); - return; - } + if (dpu_enc->cur_master && dpu_enc->cur_master->ops.enable) + dpu_enc->cur_master->ops.enable(dpu_enc->cur_master); ret = dpu_encoder_resource_control(drm_enc, DPU_ENC_RC_EVENT_KICKOFF); if (ret) { @@ -1189,26 +1173,6 @@ static void dpu_encoder_virt_enable(struct drm_encoder *drm_enc) return; } - for (i = 0; i < dpu_enc->num_phys_encs; i++) { - struct dpu_encoder_phys *phys = dpu_enc->phys_encs[i]; - - if (!phys) - continue; - - if (phys != dpu_enc->cur_master) { - if (phys->ops.enable) - phys->ops.enable(phys); - } - - if (dpu_enc->misr_enable && (dpu_enc->disp_info.capabilities & - MSM_DISPLAY_CAP_VID_MODE) && phys->ops.setup_misr) - phys->ops.setup_misr(phys, true, - dpu_enc->misr_frame_count); - } - - if (dpu_enc->cur_master->ops.enable) - dpu_enc->cur_master->ops.enable(dpu_enc->cur_master); - _dpu_encoder_virt_enable_helper(drm_enc); } @@ -1266,8 +1230,6 @@ static void dpu_encoder_virt_disable(struct drm_encoder *drm_enc) dpu_enc->phys_encs[i]->connector = NULL; } - dpu_enc->cur_master = NULL; - DPU_DEBUG_ENC(dpu_enc, "encoder disabled\n"); dpu_rm_release(&dpu_kms->rm, drm_enc); @@ -1397,9 +1359,9 @@ static void dpu_encoder_frame_done_callback( /* One of the physical encoders has become idle */ for (i = 0; i < dpu_enc->num_phys_encs; i++) { if (dpu_enc->phys_encs[i] == ready_phys) { - clear_bit(i, dpu_enc->frame_busy_mask); trace_dpu_enc_frame_done_cb(DRMID(drm_enc), i, dpu_enc->frame_busy_mask[0]); + clear_bit(i, dpu_enc->frame_busy_mask); } } @@ -1480,7 +1442,8 @@ static inline void _dpu_encoder_trigger_flush(struct drm_encoder *drm_enc, ret = ctl->ops.get_pending_flush(ctl); trace_dpu_enc_trigger_flush(DRMID(drm_enc), phys->intf_idx, - pending_kickoff_cnt, ctl->idx, ret); + pending_kickoff_cnt, ctl->idx, + extra_flush_bits, ret); } /** @@ -1879,7 +1842,7 @@ void dpu_encoder_kickoff(struct drm_encoder *drm_enc) phys->ops.handle_post_kickoff(phys); } - if (dpu_enc->disp_info.intf_type == DRM_MODE_CONNECTOR_DSI && + if (dpu_enc->disp_info.intf_type == DRM_MODE_ENCODER_DSI && !_dpu_encoder_wakeup_time(drm_enc, &wakeup_time)) { trace_dpu_enc_early_kickoff(DRMID(drm_enc), ktime_to_ms(wakeup_time)); @@ -1955,113 +1918,6 @@ static int _dpu_encoder_debugfs_status_open(struct inode *inode, return single_open(file, _dpu_encoder_status_show, inode->i_private); } -static ssize_t _dpu_encoder_misr_setup(struct file *file, - const char __user *user_buf, size_t count, loff_t *ppos) -{ - struct dpu_encoder_virt *dpu_enc; - int i = 0, rc; - char buf[MISR_BUFF_SIZE + 1]; - size_t buff_copy; - u32 frame_count, enable; - - if (!file || !file->private_data) - return -EINVAL; - - dpu_enc = file->private_data; - - buff_copy = min_t(size_t, count, MISR_BUFF_SIZE); - if (copy_from_user(buf, user_buf, buff_copy)) - return -EINVAL; - - buf[buff_copy] = 0; /* end of string */ - - if (sscanf(buf, "%u %u", &enable, &frame_count) != 2) - return -EINVAL; - - rc = _dpu_encoder_power_enable(dpu_enc, true); - if (rc) - return rc; - - mutex_lock(&dpu_enc->enc_lock); - dpu_enc->misr_enable = enable; - dpu_enc->misr_frame_count = frame_count; - for (i = 0; i < dpu_enc->num_phys_encs; i++) { - struct dpu_encoder_phys *phys = dpu_enc->phys_encs[i]; - - if (!phys || !phys->ops.setup_misr) - continue; - - phys->ops.setup_misr(phys, enable, frame_count); - } - mutex_unlock(&dpu_enc->enc_lock); - _dpu_encoder_power_enable(dpu_enc, false); - - return count; -} - -static ssize_t _dpu_encoder_misr_read(struct file *file, - char __user *user_buff, size_t count, loff_t *ppos) -{ - struct dpu_encoder_virt *dpu_enc; - int i = 0, len = 0; - char buf[MISR_BUFF_SIZE + 1] = {'\0'}; - int rc; - - if (*ppos) - return 0; - - if (!file || !file->private_data) - return -EINVAL; - - dpu_enc = file->private_data; - - rc = _dpu_encoder_power_enable(dpu_enc, true); - if (rc) - return rc; - - mutex_lock(&dpu_enc->enc_lock); - if (!dpu_enc->misr_enable) { - len += snprintf(buf + len, MISR_BUFF_SIZE - len, - "disabled\n"); - goto buff_check; - } else if (dpu_enc->disp_info.capabilities & - ~MSM_DISPLAY_CAP_VID_MODE) { - len += snprintf(buf + len, MISR_BUFF_SIZE - len, - "unsupported\n"); - goto buff_check; - } - - for (i = 0; i < dpu_enc->num_phys_encs; i++) { - struct dpu_encoder_phys *phys = dpu_enc->phys_encs[i]; - - if (!phys || !phys->ops.collect_misr) - continue; - - len += snprintf(buf + len, MISR_BUFF_SIZE - len, - "Intf idx:%d\n", phys->intf_idx - INTF_0); - len += snprintf(buf + len, MISR_BUFF_SIZE - len, "0x%x\n", - phys->ops.collect_misr(phys)); - } - -buff_check: - if (count <= len) { - len = 0; - goto end; - } - - if (copy_to_user(user_buff, buf, len)) { - len = -EFAULT; - goto end; - } - - *ppos += len; /* increase offset */ - -end: - mutex_unlock(&dpu_enc->enc_lock); - _dpu_encoder_power_enable(dpu_enc, false); - return len; -} - static int _dpu_encoder_init_debugfs(struct drm_encoder *drm_enc) { struct dpu_encoder_virt *dpu_enc; @@ -2076,12 +1932,6 @@ static int _dpu_encoder_init_debugfs(struct drm_encoder *drm_enc) .release = single_release, }; - static const struct file_operations debugfs_misr_fops = { - .open = simple_open, - .read = _dpu_encoder_misr_read, - .write = _dpu_encoder_misr_setup, - }; - char name[DPU_NAME_SIZE]; if (!drm_enc || !drm_enc->dev || !drm_enc->dev->dev_private) { @@ -2105,9 +1955,6 @@ static int _dpu_encoder_init_debugfs(struct drm_encoder *drm_enc) debugfs_create_file("status", 0600, dpu_enc->debugfs_root, dpu_enc, &debugfs_status_fops); - debugfs_create_file("misr_data", 0600, - dpu_enc->debugfs_root, dpu_enc, &debugfs_misr_fops); - for (i = 0; i < dpu_enc->num_phys_encs; i++) if (dpu_enc->phys_encs[i] && dpu_enc->phys_encs[i]->ops.late_register) @@ -2195,6 +2042,11 @@ static int dpu_encoder_virt_add_phys_encs( ++dpu_enc->num_phys_encs; } + if (params->split_role == ENC_ROLE_SLAVE) + dpu_enc->cur_slave = enc; + else + dpu_enc->cur_master = enc; + return 0; } @@ -2206,8 +2058,7 @@ static const struct dpu_encoder_virt_ops dpu_encoder_parent_ops = { static int dpu_encoder_setup_display(struct dpu_encoder_virt *dpu_enc, struct dpu_kms *dpu_kms, - struct msm_display_info *disp_info, - int *drm_enc_mode) + struct msm_display_info *disp_info) { int ret = 0; int i = 0; @@ -2220,6 +2071,8 @@ static int dpu_encoder_setup_display(struct dpu_encoder_virt *dpu_enc, return -EINVAL; } + dpu_enc->cur_master = NULL; + memset(&phys_params, 0, sizeof(phys_params)); phys_params.dpu_kms = dpu_kms; phys_params.parent = &dpu_enc->base; @@ -2228,24 +2081,17 @@ static int dpu_encoder_setup_display(struct dpu_encoder_virt *dpu_enc, DPU_DEBUG("\n"); - if (disp_info->intf_type == DRM_MODE_CONNECTOR_DSI) { - *drm_enc_mode = DRM_MODE_ENCODER_DSI; + switch (disp_info->intf_type) { + case DRM_MODE_ENCODER_DSI: intf_type = INTF_DSI; - } else if (disp_info->intf_type == DRM_MODE_CONNECTOR_HDMIA) { - *drm_enc_mode = DRM_MODE_ENCODER_TMDS; - intf_type = INTF_HDMI; - } else if (disp_info->intf_type == DRM_MODE_CONNECTOR_DisplayPort) { - *drm_enc_mode = DRM_MODE_ENCODER_TMDS; - intf_type = INTF_DP; - } else { + break; + default: DPU_ERROR_ENC(dpu_enc, "unsupported display interface type\n"); return -EINVAL; } WARN_ON(disp_info->num_of_h_tiles < 1); - dpu_enc->display_num_of_h_tiles = disp_info->num_of_h_tiles; - DPU_DEBUG("dsi_info->num_of_h_tiles %d\n", disp_info->num_of_h_tiles); if ((disp_info->capabilities & MSM_DISPLAY_CAP_CMD_MODE) || @@ -2358,25 +2204,22 @@ int dpu_encoder_setup(struct drm_device *dev, struct drm_encoder *enc, struct dpu_kms *dpu_kms = to_dpu_kms(priv->kms); struct drm_encoder *drm_enc = NULL; struct dpu_encoder_virt *dpu_enc = NULL; - int drm_enc_mode = DRM_MODE_ENCODER_NONE; int ret = 0; dpu_enc = to_dpu_encoder_virt(enc); mutex_init(&dpu_enc->enc_lock); - ret = dpu_encoder_setup_display(dpu_enc, dpu_kms, disp_info, - &drm_enc_mode); + ret = dpu_encoder_setup_display(dpu_enc, dpu_kms, disp_info); if (ret) goto fail; - dpu_enc->cur_master = NULL; spin_lock_init(&dpu_enc->enc_spinlock); atomic_set(&dpu_enc->frame_done_timeout, 0); timer_setup(&dpu_enc->frame_done_timer, dpu_encoder_frame_done_timeout, 0); - if (disp_info->intf_type == DRM_MODE_CONNECTOR_DSI) + if (disp_info->intf_type == DRM_MODE_ENCODER_DSI) timer_setup(&dpu_enc->vsync_event_timer, dpu_encoder_vsync_event_handler, 0); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h index 60f809fc7c13..9dbf38f446d9 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h @@ -32,15 +32,9 @@ /** * Encoder functions and data types * @intfs: Interfaces this encoder is using, INTF_MODE_NONE if unused - * @needs_cdm: Encoder requests a CDM based on pixel format conversion needs - * @display_num_of_h_tiles: Number of horizontal tiles in case of split - * interface - * @topology: Topology of the display */ struct dpu_encoder_hw_resources { enum dpu_intf_mode intfs[INTF_MAX]; - bool needs_cdm; - u32 display_num_of_h_tiles; }; /** @@ -56,11 +50,9 @@ struct dpu_encoder_kickoff_params { * dpu_encoder_get_hw_resources - Populate table of required hardware resources * @encoder: encoder pointer * @hw_res: resource table to populate with encoder required resources - * @conn_state: report hw reqs based on this proposed connector state */ void dpu_encoder_get_hw_resources(struct drm_encoder *encoder, - struct dpu_encoder_hw_resources *hw_res, - struct drm_connector_state *conn_state); + struct dpu_encoder_hw_resources *hw_res); /** * dpu_encoder_register_vblank_callback - provide callback to encoder that diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys.h index c7df8aad6613..964efcc757a4 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys.h @@ -22,8 +22,8 @@ #include "dpu_hw_pingpong.h" #include "dpu_hw_ctl.h" #include "dpu_hw_top.h" -#include "dpu_hw_cdm.h" #include "dpu_encoder.h" +#include "dpu_crtc.h" #define DPU_ENCODER_NAME_MAX 16 @@ -114,8 +114,6 @@ struct dpu_encoder_virt_ops { * @handle_post_kickoff: Do any work necessary post-kickoff work * @trigger_start: Process start event on physical encoder * @needs_single_flush: Whether encoder slaves need to be flushed - * @setup_misr: Sets up MISR, enable and disables based on sysfs - * @collect_misr: Collects MISR data on frame update * @hw_reset: Issue HW recovery such as CTL reset and clear * DPU_ENC_ERR_NEEDS_HW_RESET state * @irq_control: Handler to enable/disable all the encoder IRQs @@ -143,8 +141,7 @@ struct dpu_encoder_phys_ops { struct drm_connector_state *conn_state); void (*destroy)(struct dpu_encoder_phys *encoder); void (*get_hw_resources)(struct dpu_encoder_phys *encoder, - struct dpu_encoder_hw_resources *hw_res, - struct drm_connector_state *conn_state); + struct dpu_encoder_hw_resources *hw_res); int (*control_vblank_irq)(struct dpu_encoder_phys *enc, bool enable); int (*wait_for_commit_done)(struct dpu_encoder_phys *phys_enc); int (*wait_for_tx_complete)(struct dpu_encoder_phys *phys_enc); @@ -154,10 +151,6 @@ struct dpu_encoder_phys_ops { void (*handle_post_kickoff)(struct dpu_encoder_phys *phys_enc); void (*trigger_start)(struct dpu_encoder_phys *phys_enc); bool (*needs_single_flush)(struct dpu_encoder_phys *phys_enc); - - void (*setup_misr)(struct dpu_encoder_phys *phys_encs, - bool enable, u32 frame_count); - u32 (*collect_misr)(struct dpu_encoder_phys *phys_enc); void (*hw_reset)(struct dpu_encoder_phys *phys_enc); void (*irq_control)(struct dpu_encoder_phys *phys, bool enable); void (*prepare_idle_pc)(struct dpu_encoder_phys *phys_enc); @@ -210,8 +203,6 @@ struct dpu_encoder_irq { * @parent_ops: Callbacks exposed by the parent to the phys_enc * @hw_mdptop: Hardware interface to the top registers * @hw_ctl: Hardware interface to the ctl registers - * @hw_cdm: Hardware interface to the cdm registers - * @cdm_cfg: Chroma-down hardware configuration * @hw_pp: Hardware interface to the ping pong registers * @dpu_kms: Pointer to the dpu_kms top level * @cached_mode: DRM mode cached at mode_set time, acted on in enable @@ -219,7 +210,6 @@ struct dpu_encoder_irq { * @split_role: Role to play in a split-panel configuration * @intf_mode: Interface mode * @intf_idx: Interface index on dpu hardware - * @topology_name: topology selected for the display * @enc_spinlock: Virtual-Encoder-Wide Spin Lock for IRQ purposes * @enable_state: Enable state tracking * @vblank_refcount: Reference count of vblank request @@ -241,15 +231,12 @@ struct dpu_encoder_phys { const struct dpu_encoder_virt_ops *parent_ops; struct dpu_hw_mdp *hw_mdptop; struct dpu_hw_ctl *hw_ctl; - struct dpu_hw_cdm *hw_cdm; - struct dpu_hw_cdm_cfg cdm_cfg; struct dpu_hw_pingpong *hw_pp; struct dpu_kms *dpu_kms; struct drm_display_mode cached_mode; enum dpu_enc_split_role split_role; enum dpu_intf_mode intf_mode; enum dpu_intf intf_idx; - enum dpu_rm_topology_name topology_name; spinlock_t *enc_spinlock; enum dpu_enc_enable_state enable_state; atomic_t vblank_refcount; @@ -367,11 +354,15 @@ void dpu_encoder_helper_hw_reset(struct dpu_encoder_phys *phys_enc); static inline enum dpu_3d_blend_mode dpu_encoder_helper_get_3d_blend_mode( struct dpu_encoder_phys *phys_enc) { + struct dpu_crtc_state *dpu_cstate; + if (!phys_enc || phys_enc->enable_state == DPU_ENC_DISABLING) return BLEND_3D_NONE; + dpu_cstate = to_dpu_crtc_state(phys_enc->parent->crtc->state); + if (phys_enc->split_role == ENC_ROLE_SOLO && - phys_enc->topology_name == DPU_RM_TOPOLOGY_DUALPIPE_3DMERGE) + dpu_crtc_state_is_stereo(dpu_cstate)) return BLEND_3D_H_ROW_INT; return BLEND_3D_NONE; diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c index 3084675ed425..b2d7f0ded24c 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c @@ -196,9 +196,6 @@ static void dpu_encoder_phys_cmd_mode_set( { struct dpu_encoder_phys_cmd *cmd_enc = to_dpu_encoder_phys_cmd(phys_enc); - struct dpu_rm *rm = &phys_enc->dpu_kms->rm; - struct dpu_rm_hw_iter iter; - int i, instance; if (!phys_enc || !mode || !adj_mode) { DPU_ERROR("invalid args\n"); @@ -208,22 +205,6 @@ static void dpu_encoder_phys_cmd_mode_set( DPU_DEBUG_CMDENC(cmd_enc, "caching mode:\n"); drm_mode_debug_printmodeline(adj_mode); - instance = phys_enc->split_role == ENC_ROLE_SLAVE ? 1 : 0; - - /* Retrieve previously allocated HW Resources. Shouldn't fail */ - dpu_rm_init_hw_iter(&iter, phys_enc->parent->base.id, DPU_HW_BLK_CTL); - for (i = 0; i <= instance; i++) { - if (dpu_rm_get_hw(rm, &iter)) - phys_enc->hw_ctl = (struct dpu_hw_ctl *)iter.hw; - } - - if (IS_ERR_OR_NULL(phys_enc->hw_ctl)) { - DPU_ERROR_CMDENC(cmd_enc, "failed to init ctl: %ld\n", - PTR_ERR(phys_enc->hw_ctl)); - phys_enc->hw_ctl = NULL; - return; - } - _dpu_encoder_phys_cmd_setup_irq_hw_idx(phys_enc); } @@ -618,23 +599,8 @@ static void dpu_encoder_phys_cmd_destroy(struct dpu_encoder_phys *phys_enc) static void dpu_encoder_phys_cmd_get_hw_resources( struct dpu_encoder_phys *phys_enc, - struct dpu_encoder_hw_resources *hw_res, - struct drm_connector_state *conn_state) + struct dpu_encoder_hw_resources *hw_res) { - struct dpu_encoder_phys_cmd *cmd_enc = - to_dpu_encoder_phys_cmd(phys_enc); - - if (!phys_enc) { - DPU_ERROR("invalid encoder\n"); - return; - } - - if ((phys_enc->intf_idx - INTF_0) >= INTF_MAX) { - DPU_ERROR("invalid intf idx:%d\n", phys_enc->intf_idx); - return; - } - - DPU_DEBUG_CMDENC(cmd_enc, "\n"); hw_res->intfs[phys_enc->intf_idx - INTF_0] = INTF_MODE_CMD; } @@ -823,7 +789,6 @@ struct dpu_encoder_phys *dpu_encoder_phys_cmd_init( { struct dpu_encoder_phys *phys_enc = NULL; struct dpu_encoder_phys_cmd *cmd_enc = NULL; - struct dpu_hw_mdp *hw_mdp; struct dpu_encoder_irq *irq; int i, ret = 0; @@ -836,14 +801,7 @@ struct dpu_encoder_phys *dpu_encoder_phys_cmd_init( goto fail; } phys_enc = &cmd_enc->base; - - hw_mdp = dpu_rm_get_mdp(&p->dpu_kms->rm); - if (IS_ERR_OR_NULL(hw_mdp)) { - ret = PTR_ERR(hw_mdp); - DPU_ERROR("failed to get mdptop\n"); - goto fail_mdp_init; - } - phys_enc->hw_mdptop = hw_mdp; + phys_enc->hw_mdptop = p->dpu_kms->hw_mdp; phys_enc->intf_idx = p->intf_idx; dpu_encoder_phys_cmd_init_ops(&phys_enc->ops); @@ -898,8 +856,6 @@ struct dpu_encoder_phys *dpu_encoder_phys_cmd_init( return phys_enc; -fail_mdp_init: - kfree(cmd_enc); fail: return ERR_PTR(ret); } diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_vid.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_vid.c index 14fc7c2a6bb7..84de385a9f62 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_vid.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_vid.c @@ -355,13 +355,14 @@ static void dpu_encoder_phys_vid_underrun_irq(void *arg, int irq_idx) static bool _dpu_encoder_phys_is_dual_ctl(struct dpu_encoder_phys *phys_enc) { + struct dpu_crtc_state *dpu_cstate; + if (!phys_enc) return false; - if (phys_enc->topology_name == DPU_RM_TOPOLOGY_DUALPIPE) - return true; + dpu_cstate = to_dpu_crtc_state(phys_enc->parent->crtc->state); - return false; + return dpu_cstate->num_ctls > 1; } static bool dpu_encoder_phys_vid_needs_single_flush( @@ -395,9 +396,6 @@ static void dpu_encoder_phys_vid_mode_set( struct drm_display_mode *mode, struct drm_display_mode *adj_mode) { - struct dpu_rm *rm; - struct dpu_rm_hw_iter iter; - int i, instance; struct dpu_encoder_phys_vid *vid_enc; if (!phys_enc || !phys_enc->dpu_kms) { @@ -405,7 +403,6 @@ static void dpu_encoder_phys_vid_mode_set( return; } - rm = &phys_enc->dpu_kms->rm; vid_enc = to_dpu_encoder_phys_vid(phys_enc); if (adj_mode) { @@ -414,21 +411,6 @@ static void dpu_encoder_phys_vid_mode_set( DPU_DEBUG_VIDENC(vid_enc, "caching mode:\n"); } - instance = phys_enc->split_role == ENC_ROLE_SLAVE ? 1 : 0; - - /* Retrieve previously allocated HW Resources. Shouldn't fail */ - dpu_rm_init_hw_iter(&iter, phys_enc->parent->base.id, DPU_HW_BLK_CTL); - for (i = 0; i <= instance; i++) { - if (dpu_rm_get_hw(rm, &iter)) - phys_enc->hw_ctl = (struct dpu_hw_ctl *)iter.hw; - } - if (IS_ERR_OR_NULL(phys_enc->hw_ctl)) { - DPU_ERROR_VIDENC(vid_enc, "failed to init ctl, %ld\n", - PTR_ERR(phys_enc->hw_ctl)); - phys_enc->hw_ctl = NULL; - return; - } - _dpu_encoder_phys_vid_setup_irq_hw_idx(phys_enc); } @@ -481,7 +463,7 @@ static void dpu_encoder_phys_vid_enable(struct dpu_encoder_phys *phys_enc) { struct msm_drm_private *priv; struct dpu_encoder_phys_vid *vid_enc; - struct dpu_hw_intf *intf; + struct dpu_rm_hw_iter iter; struct dpu_hw_ctl *ctl; u32 flush_mask = 0; @@ -493,11 +475,20 @@ static void dpu_encoder_phys_vid_enable(struct dpu_encoder_phys *phys_enc) priv = phys_enc->parent->dev->dev_private; vid_enc = to_dpu_encoder_phys_vid(phys_enc); - intf = vid_enc->hw_intf; ctl = phys_enc->hw_ctl; - if (!vid_enc->hw_intf || !phys_enc->hw_ctl) { - DPU_ERROR("invalid hw_intf %d hw_ctl %d\n", - vid_enc->hw_intf != 0, phys_enc->hw_ctl != 0); + + dpu_rm_init_hw_iter(&iter, phys_enc->parent->base.id, DPU_HW_BLK_INTF); + while (dpu_rm_get_hw(&phys_enc->dpu_kms->rm, &iter)) { + struct dpu_hw_intf *hw_intf = (struct dpu_hw_intf *)iter.hw; + + if (hw_intf->idx == phys_enc->intf_idx) { + vid_enc->hw_intf = hw_intf; + break; + } + } + + if (!vid_enc->hw_intf) { + DPU_ERROR("hw_intf not assigned\n"); return; } @@ -519,7 +510,7 @@ static void dpu_encoder_phys_vid_enable(struct dpu_encoder_phys *phys_enc) !dpu_encoder_phys_vid_is_master(phys_enc)) goto skip_flush; - ctl->ops.get_bitmask_intf(ctl, &flush_mask, intf->idx); + ctl->ops.get_bitmask_intf(ctl, &flush_mask, vid_enc->hw_intf->idx); ctl->ops.update_pending_flush(ctl, flush_mask); skip_flush: @@ -547,25 +538,9 @@ static void dpu_encoder_phys_vid_destroy(struct dpu_encoder_phys *phys_enc) static void dpu_encoder_phys_vid_get_hw_resources( struct dpu_encoder_phys *phys_enc, - struct dpu_encoder_hw_resources *hw_res, - struct drm_connector_state *conn_state) + struct dpu_encoder_hw_resources *hw_res) { - struct dpu_encoder_phys_vid *vid_enc; - - if (!phys_enc || !hw_res) { - DPU_ERROR("invalid arg(s), enc %d hw_res %d conn_state %d\n", - phys_enc != 0, hw_res != 0, conn_state != 0); - return; - } - - vid_enc = to_dpu_encoder_phys_vid(phys_enc); - if (!vid_enc->hw_intf) { - DPU_ERROR("invalid arg(s), hw_intf\n"); - return; - } - - DPU_DEBUG_VIDENC(vid_enc, "\n"); - hw_res->intfs[vid_enc->hw_intf->idx - INTF_0] = INTF_MODE_VIDEO; + hw_res->intfs[phys_enc->intf_idx - INTF_0] = INTF_MODE_VIDEO; } static int _dpu_encoder_phys_vid_wait_for_vblank( @@ -756,32 +731,6 @@ static void dpu_encoder_phys_vid_irq_control(struct dpu_encoder_phys *phys_enc, } } -static void dpu_encoder_phys_vid_setup_misr(struct dpu_encoder_phys *phys_enc, - bool enable, u32 frame_count) -{ - struct dpu_encoder_phys_vid *vid_enc; - - if (!phys_enc) - return; - vid_enc = to_dpu_encoder_phys_vid(phys_enc); - - if (vid_enc->hw_intf && vid_enc->hw_intf->ops.setup_misr) - vid_enc->hw_intf->ops.setup_misr(vid_enc->hw_intf, - enable, frame_count); -} - -static u32 dpu_encoder_phys_vid_collect_misr(struct dpu_encoder_phys *phys_enc) -{ - struct dpu_encoder_phys_vid *vid_enc; - - if (!phys_enc) - return 0; - vid_enc = to_dpu_encoder_phys_vid(phys_enc); - - return vid_enc->hw_intf && vid_enc->hw_intf->ops.collect_misr ? - vid_enc->hw_intf->ops.collect_misr(vid_enc->hw_intf) : 0; -} - static int dpu_encoder_phys_vid_get_line_count( struct dpu_encoder_phys *phys_enc) { @@ -817,8 +766,6 @@ static void dpu_encoder_phys_vid_init_ops(struct dpu_encoder_phys_ops *ops) ops->prepare_for_kickoff = dpu_encoder_phys_vid_prepare_for_kickoff; ops->handle_post_kickoff = dpu_encoder_phys_vid_handle_post_kickoff; ops->needs_single_flush = dpu_encoder_phys_vid_needs_single_flush; - ops->setup_misr = dpu_encoder_phys_vid_setup_misr; - ops->collect_misr = dpu_encoder_phys_vid_collect_misr; ops->hw_reset = dpu_encoder_helper_hw_reset; ops->get_line_count = dpu_encoder_phys_vid_get_line_count; } @@ -828,8 +775,6 @@ struct dpu_encoder_phys *dpu_encoder_phys_vid_init( { struct dpu_encoder_phys *phys_enc = NULL; struct dpu_encoder_phys_vid *vid_enc = NULL; - struct dpu_rm_hw_iter iter; - struct dpu_hw_mdp *hw_mdp; struct dpu_encoder_irq *irq; int i, ret = 0; @@ -846,35 +791,9 @@ struct dpu_encoder_phys *dpu_encoder_phys_vid_init( phys_enc = &vid_enc->base; - hw_mdp = dpu_rm_get_mdp(&p->dpu_kms->rm); - if (IS_ERR_OR_NULL(hw_mdp)) { - ret = PTR_ERR(hw_mdp); - DPU_ERROR("failed to get mdptop\n"); - goto fail; - } - phys_enc->hw_mdptop = hw_mdp; + phys_enc->hw_mdptop = p->dpu_kms->hw_mdp; phys_enc->intf_idx = p->intf_idx; - /** - * hw_intf resource permanently assigned to this encoder - * Other resources allocated at atomic commit time by use case - */ - dpu_rm_init_hw_iter(&iter, 0, DPU_HW_BLK_INTF); - while (dpu_rm_get_hw(&p->dpu_kms->rm, &iter)) { - struct dpu_hw_intf *hw_intf = (struct dpu_hw_intf *)iter.hw; - - if (hw_intf->idx == p->intf_idx) { - vid_enc->hw_intf = hw_intf; - break; - } - } - - if (!vid_enc->hw_intf) { - ret = -EINVAL; - DPU_ERROR("failed to get hw_intf\n"); - goto fail; - } - DPU_DEBUG_VIDENC(vid_enc, "\n"); dpu_encoder_phys_vid_init_ops(&phys_enc->ops); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c index 44ee06398b1d..512ac0834d2b 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c @@ -29,6 +29,9 @@ BIT(DPU_SSPP_TS_PREFILL) | BIT(DPU_SSPP_TS_PREFILL_REC1) |\ BIT(DPU_SSPP_CDP) | BIT(DPU_SSPP_EXCL_RECT)) +#define DMA_CURSOR_SDM845_MASK \ + (DMA_SDM845_MASK | BIT(DPU_SSPP_CURSOR)) + #define MIXER_SDM845_MASK \ (BIT(DPU_MIXER_SOURCESPLIT) | BIT(DPU_DIM_LAYER)) @@ -71,7 +74,6 @@ static struct dpu_mdp_cfg sdm845_mdp[] = { .base = 0x0, .len = 0x45C, .features = 0, .highest_bank_bit = 0x2, - .has_dest_scaler = true, .clk_ctrls[DPU_CLK_CTRL_VIG0] = { .reg_off = 0x2AC, .bit_off = 0}, .clk_ctrls[DPU_CLK_CTRL_VIG1] = { @@ -174,45 +176,35 @@ static const struct dpu_sspp_sub_blks sdm845_dma_sblk_1 = _DMA_SBLK("9", 2); static const struct dpu_sspp_sub_blks sdm845_dma_sblk_2 = _DMA_SBLK("10", 3); static const struct dpu_sspp_sub_blks sdm845_dma_sblk_3 = _DMA_SBLK("11", 4); -#define SSPP_VIG_BLK(_name, _id, _base, _sblk, _xinid, _clkctrl) \ - { \ - .name = _name, .id = _id, \ - .base = _base, .len = 0x1c8, \ - .features = VIG_SDM845_MASK, \ - .sblk = &_sblk, \ - .xin_id = _xinid, \ - .type = SSPP_TYPE_VIG, \ - .clk_ctrl = _clkctrl \ - } - -#define SSPP_DMA_BLK(_name, _id, _base, _sblk, _xinid, _clkctrl) \ +#define SSPP_BLK(_name, _id, _base, _features, \ + _sblk, _xinid, _type, _clkctrl) \ { \ .name = _name, .id = _id, \ .base = _base, .len = 0x1c8, \ - .features = DMA_SDM845_MASK, \ + .features = _features, \ .sblk = &_sblk, \ .xin_id = _xinid, \ - .type = SSPP_TYPE_DMA, \ + .type = _type, \ .clk_ctrl = _clkctrl \ } static struct dpu_sspp_cfg sdm845_sspp[] = { - SSPP_VIG_BLK("sspp_0", SSPP_VIG0, 0x4000, - sdm845_vig_sblk_0, 0, DPU_CLK_CTRL_VIG0), - SSPP_VIG_BLK("sspp_1", SSPP_VIG1, 0x6000, - sdm845_vig_sblk_1, 4, DPU_CLK_CTRL_VIG1), - SSPP_VIG_BLK("sspp_2", SSPP_VIG2, 0x8000, - sdm845_vig_sblk_2, 8, DPU_CLK_CTRL_VIG2), - SSPP_VIG_BLK("sspp_3", SSPP_VIG3, 0xa000, - sdm845_vig_sblk_3, 12, DPU_CLK_CTRL_VIG3), - SSPP_DMA_BLK("sspp_8", SSPP_DMA0, 0x24000, - sdm845_dma_sblk_0, 1, DPU_CLK_CTRL_DMA0), - SSPP_DMA_BLK("sspp_9", SSPP_DMA1, 0x26000, - sdm845_dma_sblk_1, 5, DPU_CLK_CTRL_DMA1), - SSPP_DMA_BLK("sspp_10", SSPP_DMA2, 0x28000, - sdm845_dma_sblk_2, 9, DPU_CLK_CTRL_CURSOR0), - SSPP_DMA_BLK("sspp_11", SSPP_DMA3, 0x2a000, - sdm845_dma_sblk_3, 13, DPU_CLK_CTRL_CURSOR1), + SSPP_BLK("sspp_0", SSPP_VIG0, 0x4000, VIG_SDM845_MASK, + sdm845_vig_sblk_0, 0, SSPP_TYPE_VIG, DPU_CLK_CTRL_VIG0), + SSPP_BLK("sspp_1", SSPP_VIG1, 0x6000, VIG_SDM845_MASK, + sdm845_vig_sblk_1, 4, SSPP_TYPE_VIG, DPU_CLK_CTRL_VIG1), + SSPP_BLK("sspp_2", SSPP_VIG2, 0x8000, VIG_SDM845_MASK, + sdm845_vig_sblk_2, 8, SSPP_TYPE_VIG, DPU_CLK_CTRL_VIG2), + SSPP_BLK("sspp_3", SSPP_VIG3, 0xa000, VIG_SDM845_MASK, + sdm845_vig_sblk_3, 12, SSPP_TYPE_VIG, DPU_CLK_CTRL_VIG3), + SSPP_BLK("sspp_8", SSPP_DMA0, 0x24000, DMA_SDM845_MASK, + sdm845_dma_sblk_0, 1, SSPP_TYPE_DMA, DPU_CLK_CTRL_DMA0), + SSPP_BLK("sspp_9", SSPP_DMA1, 0x26000, DMA_SDM845_MASK, + sdm845_dma_sblk_1, 5, SSPP_TYPE_DMA, DPU_CLK_CTRL_DMA1), + SSPP_BLK("sspp_10", SSPP_DMA2, 0x28000, DMA_CURSOR_SDM845_MASK, + sdm845_dma_sblk_2, 9, SSPP_TYPE_DMA, DPU_CLK_CTRL_CURSOR0), + SSPP_BLK("sspp_11", SSPP_DMA3, 0x2a000, DMA_CURSOR_SDM845_MASK, + sdm845_dma_sblk_3, 13, SSPP_TYPE_DMA, DPU_CLK_CTRL_CURSOR1), }; /************************************************************* @@ -227,48 +219,23 @@ static const struct dpu_lm_sub_blks sdm845_lm_sblk = { }, }; -#define LM_BLK(_name, _id, _base, _ds, _pp, _lmpair) \ +#define LM_BLK(_name, _id, _base, _pp, _lmpair) \ { \ .name = _name, .id = _id, \ .base = _base, .len = 0x320, \ .features = MIXER_SDM845_MASK, \ .sblk = &sdm845_lm_sblk, \ - .ds = _ds, \ .pingpong = _pp, \ .lm_pair_mask = (1 << _lmpair) \ } static struct dpu_lm_cfg sdm845_lm[] = { - LM_BLK("lm_0", LM_0, 0x44000, DS_0, PINGPONG_0, LM_1), - LM_BLK("lm_1", LM_1, 0x45000, DS_1, PINGPONG_1, LM_0), - LM_BLK("lm_2", LM_2, 0x46000, DS_MAX, PINGPONG_2, LM_5), - LM_BLK("lm_3", LM_3, 0x0, DS_MAX, PINGPONG_MAX, 0), - LM_BLK("lm_4", LM_4, 0x0, DS_MAX, PINGPONG_MAX, 0), - LM_BLK("lm_5", LM_5, 0x49000, DS_MAX, PINGPONG_3, LM_2), -}; - -/************************************************************* - * DS sub blocks config - *************************************************************/ -static const struct dpu_ds_top_cfg sdm845_ds_top = { - .name = "ds_top_0", .id = DS_TOP, - .base = 0x60000, .len = 0xc, - .maxinputwidth = DEFAULT_DPU_LINE_WIDTH, - .maxoutputwidth = DEFAULT_DPU_OUTPUT_LINE_WIDTH, - .maxupscale = MAX_UPSCALE_RATIO, -}; - -#define DS_BLK(_name, _id, _base) \ - {\ - .name = _name, .id = _id, \ - .base = _base, .len = 0x800, \ - .features = DPU_SSPP_SCALER_QSEED3, \ - .top = &sdm845_ds_top \ - } - -static struct dpu_ds_cfg sdm845_ds[] = { - DS_BLK("ds_0", DS_0, 0x800), - DS_BLK("ds_1", DS_1, 0x1000), + LM_BLK("lm_0", LM_0, 0x44000, PINGPONG_0, LM_1), + LM_BLK("lm_1", LM_1, 0x45000, PINGPONG_1, LM_0), + LM_BLK("lm_2", LM_2, 0x46000, PINGPONG_2, LM_5), + LM_BLK("lm_3", LM_3, 0x0, PINGPONG_MAX, 0), + LM_BLK("lm_4", LM_4, 0x0, PINGPONG_MAX, 0), + LM_BLK("lm_5", LM_5, 0x49000, PINGPONG_3, LM_2), }; /************************************************************* @@ -328,18 +295,6 @@ static struct dpu_intf_cfg sdm845_intf[] = { }; /************************************************************* - * CDM sub blocks config - *************************************************************/ -static struct dpu_cdm_cfg sdm845_cdm[] = { - { - .name = "cdm_0", .id = CDM_0, - .base = 0x79200, .len = 0x224, - .features = 0, - .intf_connect = BIT(INTF_3), - }, -}; - -/************************************************************* * VBIF sub blocks config *************************************************************/ /* VBIF QOS remap */ @@ -461,12 +416,8 @@ static void sdm845_cfg_init(struct dpu_mdss_cfg *dpu_cfg) .sspp = sdm845_sspp, .mixer_count = ARRAY_SIZE(sdm845_lm), .mixer = sdm845_lm, - .ds_count = ARRAY_SIZE(sdm845_ds), - .ds = sdm845_ds, .pingpong_count = ARRAY_SIZE(sdm845_pp), .pingpong = sdm845_pp, - .cdm_count = ARRAY_SIZE(sdm845_cdm), - .cdm = sdm845_cdm, .intf_count = ARRAY_SIZE(sdm845_intf), .intf = sdm845_intf, .vbif_count = ARRAY_SIZE(sdm845_vbif), diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h index f0cb0d4fc80e..dc060e7358e4 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h @@ -428,7 +428,6 @@ struct dpu_clk_ctrl_reg { * @highest_bank_bit: UBWC parameter * @ubwc_static: ubwc static configuration * @ubwc_swizzle: ubwc default swizzle setting - * @has_dest_scaler: indicates support of destination scaler * @clk_ctrls clock control register definition */ struct dpu_mdp_cfg { @@ -436,7 +435,6 @@ struct dpu_mdp_cfg { u32 highest_bank_bit; u32 ubwc_static; u32 ubwc_swizzle; - bool has_dest_scaler; struct dpu_clk_ctrl_reg clk_ctrls[DPU_CLK_CTRL_MAX]; }; @@ -474,50 +472,16 @@ struct dpu_sspp_cfg { * @features bit mask identifying sub-blocks/features * @sblk: LM Sub-blocks information * @pingpong: ID of connected PingPong, PINGPONG_MAX if unsupported - * @ds: ID of connected DS, DS_MAX if unsupported * @lm_pair_mask: Bitmask of LMs that can be controlled by same CTL */ struct dpu_lm_cfg { DPU_HW_BLK_INFO; const struct dpu_lm_sub_blks *sblk; u32 pingpong; - u32 ds; unsigned long lm_pair_mask; }; /** - * struct dpu_ds_top_cfg - information of dest scaler top - * @id enum identifying this block - * @base register offset of this block - * @features bit mask identifying features - * @version hw version of dest scaler - * @maxinputwidth maximum input line width - * @maxoutputwidth maximum output line width - * @maxupscale maximum upscale ratio - */ -struct dpu_ds_top_cfg { - DPU_HW_BLK_INFO; - u32 version; - u32 maxinputwidth; - u32 maxoutputwidth; - u32 maxupscale; -}; - -/** - * struct dpu_ds_cfg - information of dest scaler blocks - * @id enum identifying this block - * @base register offset wrt DS top offset - * @features bit mask identifying features - * @version hw version of the qseed block - * @top DS top information - */ -struct dpu_ds_cfg { - DPU_HW_BLK_INFO; - u32 version; - const struct dpu_ds_top_cfg *top; -}; - -/** * struct dpu_pingpong_cfg - information of PING-PONG blocks * @id enum identifying this block * @base register offset of this block @@ -530,18 +494,6 @@ struct dpu_pingpong_cfg { }; /** - * struct dpu_cdm_cfg - information of chroma down blocks - * @id enum identifying this block - * @base register offset of this block - * @features bit mask identifying sub-blocks/features - * @intf_connect Bitmask of INTF IDs this CDM can connect to - */ -struct dpu_cdm_cfg { - DPU_HW_BLK_INFO; - unsigned long intf_connect; -}; - -/** * struct dpu_intf_cfg - information of timing engine blocks * @id enum identifying this block * @base register offset of this block @@ -728,15 +680,9 @@ struct dpu_mdss_cfg { u32 mixer_count; struct dpu_lm_cfg *mixer; - u32 ds_count; - struct dpu_ds_cfg *ds; - u32 pingpong_count; struct dpu_pingpong_cfg *pingpong; - u32 cdm_count; - struct dpu_cdm_cfg *cdm; - u32 intf_count; struct dpu_intf_cfg *intf; @@ -771,9 +717,7 @@ struct dpu_mdss_hw_cfg_handler { #define BLK_DMA(s) ((s)->dma) #define BLK_CURSOR(s) ((s)->cursor) #define BLK_MIXER(s) ((s)->mixer) -#define BLK_DS(s) ((s)->ds) #define BLK_PINGPONG(s) ((s)->pingpong) -#define BLK_CDM(s) ((s)->cdm) #define BLK_INTF(s) ((s)->intf) #define BLK_AD(s) ((s)->ad) diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_cdm.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_cdm.c deleted file mode 100644 index 554874ba0c3b..000000000000 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_cdm.c +++ /dev/null @@ -1,323 +0,0 @@ -/* Copyright (c) 2015-2018, The Linux Foundation. All rights reserved. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License version 2 and - * only version 2 as published by the Free Software Foundation. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - */ - -#include "dpu_hw_mdss.h" -#include "dpu_hwio.h" -#include "dpu_hw_catalog.h" -#include "dpu_hw_cdm.h" -#include "dpu_dbg.h" -#include "dpu_kms.h" - -#define CDM_CSC_10_OPMODE 0x000 -#define CDM_CSC_10_BASE 0x004 - -#define CDM_CDWN2_OP_MODE 0x100 -#define CDM_CDWN2_CLAMP_OUT 0x104 -#define CDM_CDWN2_PARAMS_3D_0 0x108 -#define CDM_CDWN2_PARAMS_3D_1 0x10C -#define CDM_CDWN2_COEFF_COSITE_H_0 0x110 -#define CDM_CDWN2_COEFF_COSITE_H_1 0x114 -#define CDM_CDWN2_COEFF_COSITE_H_2 0x118 -#define CDM_CDWN2_COEFF_OFFSITE_H_0 0x11C -#define CDM_CDWN2_COEFF_OFFSITE_H_1 0x120 -#define CDM_CDWN2_COEFF_OFFSITE_H_2 0x124 -#define CDM_CDWN2_COEFF_COSITE_V 0x128 -#define CDM_CDWN2_COEFF_OFFSITE_V 0x12C -#define CDM_CDWN2_OUT_SIZE 0x130 - -#define CDM_HDMI_PACK_OP_MODE 0x200 -#define CDM_CSC_10_MATRIX_COEFF_0 0x004 - -/** - * Horizontal coefficients for cosite chroma downscale - * s13 representation of coefficients - */ -static u32 cosite_h_coeff[] = {0x00000016, 0x000001cc, 0x0100009e}; - -/** - * Horizontal coefficients for offsite chroma downscale - */ -static u32 offsite_h_coeff[] = {0x000b0005, 0x01db01eb, 0x00e40046}; - -/** - * Vertical coefficients for cosite chroma downscale - */ -static u32 cosite_v_coeff[] = {0x00080004}; -/** - * Vertical coefficients for offsite chroma downscale - */ -static u32 offsite_v_coeff[] = {0x00060002}; - -/* Limited Range rgb2yuv coeff with clamp and bias values for CSC 10 module */ -static struct dpu_csc_cfg rgb2yuv_cfg = { - { - 0x0083, 0x0102, 0x0032, - 0x1fb5, 0x1f6c, 0x00e1, - 0x00e1, 0x1f45, 0x1fdc - }, - { 0x00, 0x00, 0x00 }, - { 0x0040, 0x0200, 0x0200 }, - { 0x000, 0x3ff, 0x000, 0x3ff, 0x000, 0x3ff }, - { 0x040, 0x3ac, 0x040, 0x3c0, 0x040, 0x3c0 }, -}; - -static struct dpu_cdm_cfg *_cdm_offset(enum dpu_cdm cdm, - struct dpu_mdss_cfg *m, - void __iomem *addr, - struct dpu_hw_blk_reg_map *b) -{ - int i; - - for (i = 0; i < m->cdm_count; i++) { - if (cdm == m->cdm[i].id) { - b->base_off = addr; - b->blk_off = m->cdm[i].base; - b->length = m->cdm[i].len; - b->hwversion = m->hwversion; - b->log_mask = DPU_DBG_MASK_CDM; - return &m->cdm[i]; - } - } - - return ERR_PTR(-EINVAL); -} - -static int dpu_hw_cdm_setup_csc_10bit(struct dpu_hw_cdm *ctx, - struct dpu_csc_cfg *data) -{ - dpu_hw_csc_setup(&ctx->hw, CDM_CSC_10_MATRIX_COEFF_0, data, true); - - return 0; -} - -static int dpu_hw_cdm_setup_cdwn(struct dpu_hw_cdm *ctx, - struct dpu_hw_cdm_cfg *cfg) -{ - struct dpu_hw_blk_reg_map *c = &ctx->hw; - u32 opmode = 0; - u32 out_size = 0; - - if (cfg->output_bit_depth == CDM_CDWN_OUTPUT_10BIT) - opmode &= ~BIT(7); - else - opmode |= BIT(7); - - /* ENABLE DWNS_H bit */ - opmode |= BIT(1); - - switch (cfg->h_cdwn_type) { - case CDM_CDWN_DISABLE: - /* CLEAR METHOD_H field */ - opmode &= ~(0x18); - /* CLEAR DWNS_H bit */ - opmode &= ~BIT(1); - break; - case CDM_CDWN_PIXEL_DROP: - /* Clear METHOD_H field (pixel drop is 0) */ - opmode &= ~(0x18); - break; - case CDM_CDWN_AVG: - /* Clear METHOD_H field (Average is 0x1) */ - opmode &= ~(0x18); - opmode |= (0x1 << 0x3); - break; - case CDM_CDWN_COSITE: - /* Clear METHOD_H field (Average is 0x2) */ - opmode &= ~(0x18); - opmode |= (0x2 << 0x3); - /* Co-site horizontal coefficients */ - DPU_REG_WRITE(c, CDM_CDWN2_COEFF_COSITE_H_0, - cosite_h_coeff[0]); - DPU_REG_WRITE(c, CDM_CDWN2_COEFF_COSITE_H_1, - cosite_h_coeff[1]); - DPU_REG_WRITE(c, CDM_CDWN2_COEFF_COSITE_H_2, - cosite_h_coeff[2]); - break; - case CDM_CDWN_OFFSITE: - /* Clear METHOD_H field (Average is 0x3) */ - opmode &= ~(0x18); - opmode |= (0x3 << 0x3); - - /* Off-site horizontal coefficients */ - DPU_REG_WRITE(c, CDM_CDWN2_COEFF_OFFSITE_H_0, - offsite_h_coeff[0]); - DPU_REG_WRITE(c, CDM_CDWN2_COEFF_OFFSITE_H_1, - offsite_h_coeff[1]); - DPU_REG_WRITE(c, CDM_CDWN2_COEFF_OFFSITE_H_2, - offsite_h_coeff[2]); - break; - default: - pr_err("%s invalid horz down sampling type\n", __func__); - return -EINVAL; - } - - /* ENABLE DWNS_V bit */ - opmode |= BIT(2); - - switch (cfg->v_cdwn_type) { - case CDM_CDWN_DISABLE: - /* CLEAR METHOD_V field */ - opmode &= ~(0x60); - /* CLEAR DWNS_V bit */ - opmode &= ~BIT(2); - break; - case CDM_CDWN_PIXEL_DROP: - /* Clear METHOD_V field (pixel drop is 0) */ - opmode &= ~(0x60); - break; - case CDM_CDWN_AVG: - /* Clear METHOD_V field (Average is 0x1) */ - opmode &= ~(0x60); - opmode |= (0x1 << 0x5); - break; - case CDM_CDWN_COSITE: - /* Clear METHOD_V field (Average is 0x2) */ - opmode &= ~(0x60); - opmode |= (0x2 << 0x5); - /* Co-site vertical coefficients */ - DPU_REG_WRITE(c, - CDM_CDWN2_COEFF_COSITE_V, - cosite_v_coeff[0]); - break; - case CDM_CDWN_OFFSITE: - /* Clear METHOD_V field (Average is 0x3) */ - opmode &= ~(0x60); - opmode |= (0x3 << 0x5); - - /* Off-site vertical coefficients */ - DPU_REG_WRITE(c, - CDM_CDWN2_COEFF_OFFSITE_V, - offsite_v_coeff[0]); - break; - default: - return -EINVAL; - } - - if (cfg->v_cdwn_type || cfg->h_cdwn_type) - opmode |= BIT(0); /* EN CDWN module */ - else - opmode &= ~BIT(0); - - out_size = (cfg->output_width & 0xFFFF) | - ((cfg->output_height & 0xFFFF) << 16); - DPU_REG_WRITE(c, CDM_CDWN2_OUT_SIZE, out_size); - DPU_REG_WRITE(c, CDM_CDWN2_OP_MODE, opmode); - DPU_REG_WRITE(c, CDM_CDWN2_CLAMP_OUT, - ((0x3FF << 16) | 0x0)); - - return 0; -} - -static int dpu_hw_cdm_enable(struct dpu_hw_cdm *ctx, - struct dpu_hw_cdm_cfg *cdm) -{ - struct dpu_hw_blk_reg_map *c = &ctx->hw; - const struct dpu_format *fmt = cdm->output_fmt; - struct cdm_output_cfg cdm_cfg = { 0 }; - u32 opmode = 0; - u32 csc = 0; - - if (!DPU_FORMAT_IS_YUV(fmt)) - return -EINVAL; - - if (cdm->output_type == CDM_CDWN_OUTPUT_HDMI) { - if (fmt->chroma_sample != DPU_CHROMA_H1V2) - return -EINVAL; /*unsupported format */ - opmode = BIT(0); - opmode |= (fmt->chroma_sample << 1); - cdm_cfg.intf_en = true; - } - - csc |= BIT(2); - csc &= ~BIT(1); - csc |= BIT(0); - - if (ctx->hw_mdp && ctx->hw_mdp->ops.setup_cdm_output) - ctx->hw_mdp->ops.setup_cdm_output(ctx->hw_mdp, &cdm_cfg); - - DPU_REG_WRITE(c, CDM_CSC_10_OPMODE, csc); - DPU_REG_WRITE(c, CDM_HDMI_PACK_OP_MODE, opmode); - return 0; -} - -static void dpu_hw_cdm_disable(struct dpu_hw_cdm *ctx) -{ - struct cdm_output_cfg cdm_cfg = { 0 }; - - if (ctx->hw_mdp && ctx->hw_mdp->ops.setup_cdm_output) - ctx->hw_mdp->ops.setup_cdm_output(ctx->hw_mdp, &cdm_cfg); -} - -static void _setup_cdm_ops(struct dpu_hw_cdm_ops *ops, - unsigned long features) -{ - ops->setup_csc_data = dpu_hw_cdm_setup_csc_10bit; - ops->setup_cdwn = dpu_hw_cdm_setup_cdwn; - ops->enable = dpu_hw_cdm_enable; - ops->disable = dpu_hw_cdm_disable; -} - -static struct dpu_hw_blk_ops dpu_hw_ops = { - .start = NULL, - .stop = NULL, -}; - -struct dpu_hw_cdm *dpu_hw_cdm_init(enum dpu_cdm idx, - void __iomem *addr, - struct dpu_mdss_cfg *m, - struct dpu_hw_mdp *hw_mdp) -{ - struct dpu_hw_cdm *c; - struct dpu_cdm_cfg *cfg; - int rc; - - c = kzalloc(sizeof(*c), GFP_KERNEL); - if (!c) - return ERR_PTR(-ENOMEM); - - cfg = _cdm_offset(idx, m, addr, &c->hw); - if (IS_ERR_OR_NULL(cfg)) { - kfree(c); - return ERR_PTR(-EINVAL); - } - - c->idx = idx; - c->caps = cfg; - _setup_cdm_ops(&c->ops, c->caps->features); - c->hw_mdp = hw_mdp; - - rc = dpu_hw_blk_init(&c->base, DPU_HW_BLK_CDM, idx, &dpu_hw_ops); - if (rc) { - DPU_ERROR("failed to init hw blk %d\n", rc); - goto blk_init_error; - } - - /* - * Perform any default initialization for the chroma down module - * @setup default csc coefficients - */ - dpu_hw_cdm_setup_csc_10bit(c, &rgb2yuv_cfg); - - return c; - -blk_init_error: - kzfree(c); - - return ERR_PTR(rc); -} - -void dpu_hw_cdm_destroy(struct dpu_hw_cdm *cdm) -{ - if (cdm) - dpu_hw_blk_destroy(&cdm->base); - kfree(cdm); -} diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_cdm.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_cdm.h deleted file mode 100644 index 5cceb1ecb8e0..000000000000 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_cdm.h +++ /dev/null @@ -1,139 +0,0 @@ -/* Copyright (c) 2015-2018, The Linux Foundation. All rights reserved. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License version 2 and - * only version 2 as published by the Free Software Foundation. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - */ - -#ifndef _DPU_HW_CDM_H -#define _DPU_HW_CDM_H - -#include "dpu_hw_mdss.h" -#include "dpu_hw_top.h" -#include "dpu_hw_blk.h" - -struct dpu_hw_cdm; - -struct dpu_hw_cdm_cfg { - u32 output_width; - u32 output_height; - u32 output_bit_depth; - u32 h_cdwn_type; - u32 v_cdwn_type; - const struct dpu_format *output_fmt; - u32 output_type; - int flags; -}; - -enum dpu_hw_cdwn_type { - CDM_CDWN_DISABLE, - CDM_CDWN_PIXEL_DROP, - CDM_CDWN_AVG, - CDM_CDWN_COSITE, - CDM_CDWN_OFFSITE, -}; - -enum dpu_hw_cdwn_output_type { - CDM_CDWN_OUTPUT_HDMI, - CDM_CDWN_OUTPUT_WB, -}; - -enum dpu_hw_cdwn_output_bit_depth { - CDM_CDWN_OUTPUT_8BIT, - CDM_CDWN_OUTPUT_10BIT, -}; - -/** - * struct dpu_hw_cdm_ops : Interface to the chroma down Hw driver functions - * Assumption is these functions will be called after - * clocks are enabled - * @setup_csc: Programs the csc matrix - * @setup_cdwn: Sets up the chroma down sub module - * @enable: Enables the output to interface and programs the - * output packer - * @disable: Puts the cdm in bypass mode - */ -struct dpu_hw_cdm_ops { - /** - * Programs the CSC matrix for conversion from RGB space to YUV space, - * it is optional to call this function as this matrix is automatically - * set during initialization, user should call this if it wants - * to program a different matrix than default matrix. - * @cdm: Pointer to the chroma down context structure - * @data Pointer to CSC configuration data - * return: 0 if success; error code otherwise - */ - int (*setup_csc_data)(struct dpu_hw_cdm *cdm, - struct dpu_csc_cfg *data); - - /** - * Programs the Chroma downsample part. - * @cdm Pointer to chroma down context - */ - int (*setup_cdwn)(struct dpu_hw_cdm *cdm, - struct dpu_hw_cdm_cfg *cfg); - - /** - * Enable the CDM module - * @cdm Pointer to chroma down context - */ - int (*enable)(struct dpu_hw_cdm *cdm, - struct dpu_hw_cdm_cfg *cfg); - - /** - * Disable the CDM module - * @cdm Pointer to chroma down context - */ - void (*disable)(struct dpu_hw_cdm *cdm); -}; - -struct dpu_hw_cdm { - struct dpu_hw_blk base; - struct dpu_hw_blk_reg_map hw; - - /* chroma down */ - const struct dpu_cdm_cfg *caps; - enum dpu_cdm idx; - - /* mdp top hw driver */ - struct dpu_hw_mdp *hw_mdp; - - /* ops */ - struct dpu_hw_cdm_ops ops; -}; - -/** - * dpu_hw_cdm - convert base object dpu_hw_base to container - * @hw: Pointer to base hardware block - * return: Pointer to hardware block container - */ -static inline struct dpu_hw_cdm *to_dpu_hw_cdm(struct dpu_hw_blk *hw) -{ - return container_of(hw, struct dpu_hw_cdm, base); -} - -/** - * dpu_hw_cdm_init - initializes the cdm hw driver object. - * should be called once before accessing every cdm. - * @idx: cdm index for which driver object is required - * @addr: mapped register io address of MDP - * @m : pointer to mdss catalog data - * @hw_mdp: pointer to mdp top hw driver object - */ -struct dpu_hw_cdm *dpu_hw_cdm_init(enum dpu_cdm idx, - void __iomem *addr, - struct dpu_mdss_cfg *m, - struct dpu_hw_mdp *hw_mdp); - -/** - * dpu_hw_cdm_destroy - destroys CDM driver context - * @cdm: pointer to CDM driver context - */ -void dpu_hw_cdm_destroy(struct dpu_hw_cdm *cdm); - -#endif /*_DPU_HW_CDM_H */ diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.c index 06be7cf7ce50..eec1051f2afc 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.c @@ -224,19 +224,6 @@ static inline int dpu_hw_ctl_get_bitmask_intf(struct dpu_hw_ctl *ctx, return 0; } -static inline int dpu_hw_ctl_get_bitmask_cdm(struct dpu_hw_ctl *ctx, - u32 *flushbits, enum dpu_cdm cdm) -{ - switch (cdm) { - case CDM_0: - *flushbits |= BIT(26); - break; - default: - return -EINVAL; - } - return 0; -} - static u32 dpu_hw_ctl_poll_reset_status(struct dpu_hw_ctl *ctx, u32 timeout_us) { struct dpu_hw_blk_reg_map *c = &ctx->hw; @@ -310,7 +297,7 @@ static void dpu_hw_ctl_setup_blendstage(struct dpu_hw_ctl *ctx, u32 mixercfg = 0, mixercfg_ext = 0, mix, ext; u32 mixercfg_ext2 = 0, mixercfg_ext3 = 0; int i, j; - u8 stages; + int stages; int pipes_per_stage; stages = _mixer_stages(ctx->mixer_hw_caps, ctx->mixer_count, lm); @@ -485,7 +472,6 @@ static void _setup_ctl_ops(struct dpu_hw_ctl_ops *ops, ops->get_bitmask_sspp = dpu_hw_ctl_get_bitmask_sspp; ops->get_bitmask_mixer = dpu_hw_ctl_get_bitmask_mixer; ops->get_bitmask_intf = dpu_hw_ctl_get_bitmask_intf; - ops->get_bitmask_cdm = dpu_hw_ctl_get_bitmask_cdm; }; static struct dpu_hw_blk_ops dpu_hw_ops = { diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.h index c66a71f8b839..6f313faca43e 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_ctl.h @@ -142,10 +142,6 @@ struct dpu_hw_ctl_ops { u32 *flushbits, enum dpu_intf blk); - int (*get_bitmask_cdm)(struct dpu_hw_ctl *ctx, - u32 *flushbits, - enum dpu_cdm blk); - /** * Set all blend stages to disabled * @ctx : ctl path ctx pointer diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_intf.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_intf.c index d280df5613c9..9c6bba0ac7c3 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_intf.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_intf.c @@ -65,9 +65,6 @@ #define INTF_FRAME_COUNT 0x0AC #define INTF_LINE_COUNT 0x0B0 -#define INTF_MISR_CTRL 0x180 -#define INTF_MISR_SIGNATURE 0x184 - static struct dpu_intf_cfg *_intf_offset(enum dpu_intf intf, struct dpu_mdss_cfg *m, void __iomem *addr, @@ -246,30 +243,6 @@ static void dpu_hw_intf_get_status( } } -static void dpu_hw_intf_setup_misr(struct dpu_hw_intf *intf, - bool enable, u32 frame_count) -{ - struct dpu_hw_blk_reg_map *c = &intf->hw; - u32 config = 0; - - DPU_REG_WRITE(c, INTF_MISR_CTRL, MISR_CTRL_STATUS_CLEAR); - /* clear misr data */ - wmb(); - - if (enable) - config = (frame_count & MISR_FRAME_COUNT_MASK) | - MISR_CTRL_ENABLE | INTF_MISR_CTRL_FREE_RUN_MASK; - - DPU_REG_WRITE(c, INTF_MISR_CTRL, config); -} - -static u32 dpu_hw_intf_collect_misr(struct dpu_hw_intf *intf) -{ - struct dpu_hw_blk_reg_map *c = &intf->hw; - - return DPU_REG_READ(c, INTF_MISR_SIGNATURE); -} - static u32 dpu_hw_intf_get_line_count(struct dpu_hw_intf *intf) { struct dpu_hw_blk_reg_map *c; @@ -289,8 +262,6 @@ static void _setup_intf_ops(struct dpu_hw_intf_ops *ops, ops->setup_prg_fetch = dpu_hw_intf_setup_prg_fetch; ops->get_status = dpu_hw_intf_get_status; ops->enable_timing = dpu_hw_intf_enable_timing_engine; - ops->setup_misr = dpu_hw_intf_setup_misr; - ops->collect_misr = dpu_hw_intf_collect_misr; ops->get_line_count = dpu_hw_intf_get_line_count; } diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_intf.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_intf.h index a79d735da68d..3b77df460dea 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_intf.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_intf.h @@ -59,8 +59,6 @@ struct intf_status { * @ setup_prog_fetch : enables/disables the programmable fetch logic * @ enable_timing: enable/disable timing engine * @ get_status: returns if timing engine is enabled or not - * @ setup_misr: enables/disables MISR in HW register - * @ collect_misr: reads and stores MISR data from HW register * @ get_line_count: reads current vertical line counter */ struct dpu_hw_intf_ops { @@ -77,11 +75,6 @@ struct dpu_hw_intf_ops { void (*get_status)(struct dpu_hw_intf *intf, struct intf_status *status); - void (*setup_misr)(struct dpu_hw_intf *intf, - bool enable, u32 frame_count); - - u32 (*collect_misr)(struct dpu_hw_intf *intf); - u32 (*get_line_count)(struct dpu_hw_intf *intf); }; diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_lm.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_lm.c index 4ab72b0f07a5..acb8dc8acaa5 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_lm.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_lm.c @@ -34,9 +34,6 @@ #define LM_BLEND0_FG_ALPHA 0x04 #define LM_BLEND0_BG_ALPHA 0x08 -#define LM_MISR_CTRL 0x310 -#define LM_MISR_SIGNATURE 0x314 - static struct dpu_lm_cfg *_lm_offset(enum dpu_lm mixer, struct dpu_mdss_cfg *m, void __iomem *addr, @@ -171,30 +168,6 @@ static void dpu_hw_lm_gc(struct dpu_hw_mixer *mixer, { } -static void dpu_hw_lm_setup_misr(struct dpu_hw_mixer *ctx, - bool enable, u32 frame_count) -{ - struct dpu_hw_blk_reg_map *c = &ctx->hw; - u32 config = 0; - - DPU_REG_WRITE(c, LM_MISR_CTRL, MISR_CTRL_STATUS_CLEAR); - /* clear misr data */ - wmb(); - - if (enable) - config = (frame_count & MISR_FRAME_COUNT_MASK) | - MISR_CTRL_ENABLE | INTF_MISR_CTRL_FREE_RUN_MASK; - - DPU_REG_WRITE(c, LM_MISR_CTRL, config); -} - -static u32 dpu_hw_lm_collect_misr(struct dpu_hw_mixer *ctx) -{ - struct dpu_hw_blk_reg_map *c = &ctx->hw; - - return DPU_REG_READ(c, LM_MISR_SIGNATURE); -} - static void _setup_mixer_ops(struct dpu_mdss_cfg *m, struct dpu_hw_lm_ops *ops, unsigned long features) @@ -207,8 +180,6 @@ static void _setup_mixer_ops(struct dpu_mdss_cfg *m, ops->setup_alpha_out = dpu_hw_lm_setup_color3; ops->setup_border_color = dpu_hw_lm_setup_border_color; ops->setup_gc = dpu_hw_lm_gc; - ops->setup_misr = dpu_hw_lm_setup_misr; - ops->collect_misr = dpu_hw_lm_collect_misr; }; static struct dpu_hw_blk_ops dpu_hw_ops = { diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_lm.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_lm.h index e29e5dab31bf..5b036aca8340 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_lm.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_lm.h @@ -66,13 +66,6 @@ struct dpu_hw_lm_ops { */ void (*setup_gc)(struct dpu_hw_mixer *mixer, void *cfg); - - /* setup_misr: enables/disables MISR in HW register */ - void (*setup_misr)(struct dpu_hw_mixer *ctx, - bool enable, u32 frame_count); - - /* collect_misr: reads and stores MISR data from HW register */ - u32 (*collect_misr)(struct dpu_hw_mixer *ctx); }; struct dpu_hw_mixer { diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h index 35e6bf930924..68c54d2c9677 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h @@ -100,7 +100,6 @@ enum dpu_hw_blk_type { DPU_HW_BLK_SSPP, DPU_HW_BLK_LM, DPU_HW_BLK_CTL, - DPU_HW_BLK_CDM, DPU_HW_BLK_PINGPONG, DPU_HW_BLK_INTF, DPU_HW_BLK_WB, @@ -173,13 +172,6 @@ enum dpu_dspp { DSPP_MAX }; -enum dpu_ds { - DS_TOP, - DS_0, - DS_1, - DS_MAX -}; - enum dpu_ctl { CTL_0 = 1, CTL_1, @@ -189,12 +181,6 @@ enum dpu_ctl { CTL_MAX }; -enum dpu_cdm { - CDM_0 = 1, - CDM_1, - CDM_MAX -}; - enum dpu_pingpong { PINGPONG_0 = 1, PINGPONG_1, @@ -246,12 +232,6 @@ enum dpu_wb { WB_MAX }; -enum dpu_ad { - AD_0 = 0x1, - AD_1, - AD_MAX -}; - enum dpu_cwb { CWB_0 = 0x1, CWB_1, @@ -451,15 +431,14 @@ struct dpu_mdss_color { * Define bit masks for h/w logging. */ #define DPU_DBG_MASK_NONE (1 << 0) -#define DPU_DBG_MASK_CDM (1 << 1) -#define DPU_DBG_MASK_INTF (1 << 2) -#define DPU_DBG_MASK_LM (1 << 3) -#define DPU_DBG_MASK_CTL (1 << 4) -#define DPU_DBG_MASK_PINGPONG (1 << 5) -#define DPU_DBG_MASK_SSPP (1 << 6) -#define DPU_DBG_MASK_WB (1 << 7) -#define DPU_DBG_MASK_TOP (1 << 8) -#define DPU_DBG_MASK_VBIF (1 << 9) -#define DPU_DBG_MASK_ROT (1 << 10) +#define DPU_DBG_MASK_INTF (1 << 1) +#define DPU_DBG_MASK_LM (1 << 2) +#define DPU_DBG_MASK_CTL (1 << 3) +#define DPU_DBG_MASK_PINGPONG (1 << 4) +#define DPU_DBG_MASK_SSPP (1 << 5) +#define DPU_DBG_MASK_WB (1 << 6) +#define DPU_DBG_MASK_TOP (1 << 7) +#define DPU_DBG_MASK_VBIF (1 << 8) +#define DPU_DBG_MASK_ROT (1 << 9) #endif /* _DPU_HW_MDSS_H */ diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.c index db2798e862fc..b8781256e21b 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.c @@ -98,23 +98,6 @@ static void dpu_hw_setup_split_pipe(struct dpu_hw_mdp *mdp, DPU_REG_WRITE(c, SPLIT_DISPLAY_EN, cfg->en & 0x1); } -static void dpu_hw_setup_cdm_output(struct dpu_hw_mdp *mdp, - struct cdm_output_cfg *cfg) -{ - struct dpu_hw_blk_reg_map *c; - u32 out_ctl = 0; - - if (!mdp || !cfg) - return; - - c = &mdp->hw; - - if (cfg->intf_en) - out_ctl |= BIT(19); - - DPU_REG_WRITE(c, MDP_OUT_CTL_0, out_ctl); -} - static bool dpu_hw_setup_clk_force_ctrl(struct dpu_hw_mdp *mdp, enum dpu_clk_ctrl_type clk_ctrl, bool enable) { @@ -307,7 +290,6 @@ static void _setup_mdp_ops(struct dpu_hw_mdp_ops *ops, unsigned long cap) { ops->setup_split_pipe = dpu_hw_setup_split_pipe; - ops->setup_cdm_output = dpu_hw_setup_cdm_output; ops->setup_clk_force_ctrl = dpu_hw_setup_clk_force_ctrl; ops->get_danger_status = dpu_hw_get_danger_status; ops->setup_vsync_source = dpu_hw_setup_vsync_source; diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.h index 899925aaa6d7..192e338f20bb 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.h @@ -52,14 +52,6 @@ struct split_pipe_cfg { }; /** - * struct cdm_output_cfg: output configuration for cdm - * @intf_en : enable/disable interface output - */ -struct cdm_output_cfg { - bool intf_en; -}; - -/** * struct dpu_danger_safe_status: danger and safe status signals * @mdp: top level status * @sspp: source pipe status @@ -89,7 +81,6 @@ struct dpu_vsync_source_cfg { * Assumption is these functions will be called after clocks are enabled. * @setup_split_pipe : Programs the pipe control registers * @setup_pp_split : Programs the pp split control registers - * @setup_cdm_output : programs cdm control * @setup_traffic_shaper : programs traffic shaper control */ struct dpu_hw_mdp_ops { @@ -102,14 +93,6 @@ struct dpu_hw_mdp_ops { struct split_pipe_cfg *p); /** - * setup_cdm_output() : Setup selection control of the cdm data path - * @mdp : mdp top context driver - * @cfg : cdm output configuration - */ - void (*setup_cdm_output)(struct dpu_hw_mdp *mdp, - struct cdm_output_cfg *cfg); - - /** * setup_traffic_shaper() : Setup traffic shaper control * @mdp : mdp top context driver * @cfg : traffic shaper configuration diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.c index 4cabae480a7b..cb5c0170374b 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.c @@ -50,9 +50,6 @@ static u32 dpu_hw_util_log_mask = DPU_DBG_MASK_NONE; #define QSEED3_CLK_CTRL0 0x54 #define QSEED3_CLK_CTRL1 0x58 #define QSEED3_CLK_STATUS 0x5C -#define QSEED3_MISR_CTRL 0x70 -#define QSEED3_MISR_SIGNATURE_0 0x74 -#define QSEED3_MISR_SIGNATURE_1 0x78 #define QSEED3_PHASE_INIT_Y_H 0x90 #define QSEED3_PHASE_INIT_Y_V 0x94 #define QSEED3_PHASE_INIT_UV_H 0x98 diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.h index 1240f505ca53..321fc64ddd0e 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_util.h @@ -148,16 +148,6 @@ struct dpu_hw_scaler3_cfg { struct dpu_hw_scaler3_de_cfg de; }; -struct dpu_hw_scaler3_lut_cfg { - bool is_configured; - u32 *dir_lut; - size_t dir_len; - u32 *cir_lut; - size_t cir_len; - u32 *sep_lut; - size_t sep_len; -}; - /** * struct dpu_drm_pix_ext_v1 - version 1 of pixel ext structure * @num_ext_pxls_lr: Number of total horizontal pixels @@ -325,12 +315,6 @@ int dpu_reg_read(struct dpu_hw_blk_reg_map *c, u32 reg_off); #define DPU_REG_WRITE(c, off, val) dpu_reg_write(c, off, val, #off) #define DPU_REG_READ(c, off) dpu_reg_read(c, off) -#define MISR_FRAME_COUNT_MASK 0xFF -#define MISR_CTRL_ENABLE BIT(8) -#define MISR_CTRL_STATUS BIT(9) -#define MISR_CTRL_STATUS_CLEAR BIT(10) -#define INTF_MISR_CTRL_FREE_RUN_MASK BIT(31) - void *dpu_hw_util_get_dir(void); void dpu_hw_setup_scaler3(struct dpu_hw_blk_reg_map *c, diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c index 7dd6bd2d6d37..0a683e65a9f3 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c @@ -450,7 +450,7 @@ static void _dpu_kms_initialize_dsi(struct drm_device *dev, int i, rc; /*TODO: Support two independent DSI connectors */ - encoder = dpu_encoder_init(dev, DRM_MODE_CONNECTOR_DSI); + encoder = dpu_encoder_init(dev, DRM_MODE_ENCODER_DSI); if (IS_ERR_OR_NULL(encoder)) { DPU_ERROR("encoder init failed for dsi display\n"); return; @@ -531,12 +531,13 @@ static int _dpu_kms_drm_obj_init(struct dpu_kms *dpu_kms) { struct drm_device *dev; struct drm_plane *primary_planes[MAX_PLANES], *plane; + struct drm_plane *cursor_planes[MAX_PLANES] = { NULL }; struct drm_crtc *crtc; struct msm_drm_private *priv; struct dpu_mdss_cfg *catalog; - int primary_planes_idx = 0, i, ret; + int primary_planes_idx = 0, cursor_planes_idx = 0, i, ret; int max_crtc_count; if (!dpu_kms || !dpu_kms->dev || !dpu_kms->dev->dev) { @@ -556,16 +557,24 @@ static int _dpu_kms_drm_obj_init(struct dpu_kms *dpu_kms) max_crtc_count = min(catalog->mixer_count, priv->num_encoders); - /* Create the planes */ + /* Create the planes, keeping track of one primary/cursor per crtc */ for (i = 0; i < catalog->sspp_count; i++) { - bool primary = true; - - if (catalog->sspp[i].features & BIT(DPU_SSPP_CURSOR) - || primary_planes_idx >= max_crtc_count) - primary = false; - - plane = dpu_plane_init(dev, catalog->sspp[i].id, primary, - (1UL << max_crtc_count) - 1, 0); + enum drm_plane_type type; + + if ((catalog->sspp[i].features & BIT(DPU_SSPP_CURSOR)) + && cursor_planes_idx < max_crtc_count) + type = DRM_PLANE_TYPE_CURSOR; + else if (primary_planes_idx < max_crtc_count) + type = DRM_PLANE_TYPE_PRIMARY; + else + type = DRM_PLANE_TYPE_OVERLAY; + + DPU_DEBUG("Create plane type %d with features %lx (cur %lx)\n", + type, catalog->sspp[i].features, + catalog->sspp[i].features & BIT(DPU_SSPP_CURSOR)); + + plane = dpu_plane_init(dev, catalog->sspp[i].id, type, + (1UL << max_crtc_count) - 1, 0); if (IS_ERR(plane)) { DPU_ERROR("dpu_plane_init failed\n"); ret = PTR_ERR(plane); @@ -573,7 +582,9 @@ static int _dpu_kms_drm_obj_init(struct dpu_kms *dpu_kms) } priv->planes[priv->num_planes++] = plane; - if (primary) + if (type == DRM_PLANE_TYPE_CURSOR) + cursor_planes[cursor_planes_idx++] = plane; + else if (type == DRM_PLANE_TYPE_PRIMARY) primary_planes[primary_planes_idx++] = plane; } @@ -581,7 +592,7 @@ static int _dpu_kms_drm_obj_init(struct dpu_kms *dpu_kms) /* Create one CRTC per encoder */ for (i = 0; i < max_crtc_count; i++) { - crtc = dpu_crtc_init(dev, primary_planes[i]); + crtc = dpu_crtc_init(dev, primary_planes[i], cursor_planes[i]); if (IS_ERR(crtc)) { ret = PTR_ERR(crtc); goto fail; @@ -956,8 +967,7 @@ static void dpu_kms_handle_power_event(u32 event_type, void *usr) if (!dpu_kms) return; - if (event_type == DPU_POWER_EVENT_POST_ENABLE) - dpu_vbif_init_memtypes(dpu_kms); + dpu_vbif_init_memtypes(dpu_kms); } static int dpu_kms_hw_init(struct msm_kms *kms) @@ -1144,10 +1154,9 @@ static int dpu_kms_hw_init(struct msm_kms *kms) /* * Handle (re)initializations during power enable */ - dpu_kms_handle_power_event(DPU_POWER_EVENT_POST_ENABLE, dpu_kms); + dpu_kms_handle_power_event(DPU_POWER_EVENT_ENABLE, dpu_kms); dpu_kms->power_event = dpu_power_handle_register_event( - &dpu_kms->phandle, - DPU_POWER_EVENT_POST_ENABLE, + &dpu_kms->phandle, DPU_POWER_EVENT_ENABLE, dpu_kms_handle_power_event, dpu_kms, "kms"); pm_runtime_put_sync(&dpu_kms->pdev->dev); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_mdss.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_mdss.c index 9e533b86682c..2235ef8129f4 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_mdss.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_mdss.c @@ -158,6 +158,8 @@ static void dpu_mdss_destroy(struct drm_device *dev) _dpu_mdss_irq_domain_fini(dpu_mdss); + free_irq(platform_get_irq(pdev, 0), dpu_mdss); + msm_dss_put_clk(mp->clk_config, mp->num_clk); devm_kfree(&pdev->dev, mp->clk_config); @@ -215,7 +217,7 @@ int dpu_mdss_init(struct drm_device *dev) if (ret) goto irq_domain_error; - ret = devm_request_irq(dev->dev, platform_get_irq(pdev, 0), + ret = request_irq(platform_get_irq(pdev, 0), dpu_mdss_irq, 0, "dpu_mdss_isr", dpu_mdss); if (ret) { DPU_ERROR("failed to init irq: %d\n", ret); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c index ec959f847d5f..d77a8cb15404 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c @@ -125,26 +125,11 @@ struct dpu_plane { static struct dpu_kms *_dpu_plane_get_kms(struct drm_plane *plane) { - struct msm_drm_private *priv; + struct msm_drm_private *priv = plane->dev->dev_private; - if (!plane || !plane->dev) - return NULL; - priv = plane->dev->dev_private; - if (!priv) - return NULL; return to_dpu_kms(priv->kms); } -static bool dpu_plane_enabled(struct drm_plane_state *state) -{ - return state && state->fb && state->crtc; -} - -static bool dpu_plane_sspp_enabled(struct drm_plane_state *state) -{ - return state && state->crtc; -} - /** * _dpu_plane_calc_fill_level - calculate fill level of the given source format * @plane: Pointer to drm plane @@ -160,7 +145,7 @@ static inline int _dpu_plane_calc_fill_level(struct drm_plane *plane, u32 fixed_buff_size; u32 total_fl; - if (!plane || !fmt || !plane->state || !src_width || !fmt->bpp) { + if (!fmt || !plane->state || !src_width || !fmt->bpp) { DPU_ERROR("invalid arguments\n"); return 0; } @@ -170,7 +155,7 @@ static inline int _dpu_plane_calc_fill_level(struct drm_plane *plane, fixed_buff_size = pdpu->pipe_sblk->common->pixel_ram_size; list_for_each_entry(tmp, &pdpu->mplane_list, mplane_list) { - if (!dpu_plane_enabled(tmp->base.state)) + if (!tmp->base.state->visible) continue; DPU_DEBUG("plane%d/%d src_width:%d/%d\n", pdpu->base.base.id, tmp->base.base.id, @@ -241,26 +226,11 @@ static u64 _dpu_plane_get_qos_lut(const struct dpu_qos_lut_tbl *tbl, static void _dpu_plane_set_qos_lut(struct drm_plane *plane, struct drm_framebuffer *fb) { - struct dpu_plane *pdpu; + struct dpu_plane *pdpu = to_dpu_plane(plane); const struct dpu_format *fmt = NULL; u64 qos_lut; u32 total_fl = 0, lut_usage; - if (!plane || !fb) { - DPU_ERROR("invalid arguments plane %d fb %d\n", - plane != 0, fb != 0); - return; - } - - pdpu = to_dpu_plane(plane); - - if (!pdpu->pipe_hw || !pdpu->pipe_sblk || !pdpu->catalog) { - DPU_ERROR("invalid arguments\n"); - return; - } else if (!pdpu->pipe_hw->ops.setup_creq_lut) { - return; - } - if (!pdpu->is_rt_pipe) { lut_usage = DPU_QOS_LUT_USAGE_NRT; } else { @@ -302,24 +272,10 @@ static void _dpu_plane_set_qos_lut(struct drm_plane *plane, static void _dpu_plane_set_danger_lut(struct drm_plane *plane, struct drm_framebuffer *fb) { - struct dpu_plane *pdpu; + struct dpu_plane *pdpu = to_dpu_plane(plane); const struct dpu_format *fmt = NULL; u32 danger_lut, safe_lut; - if (!plane || !fb) { - DPU_ERROR("invalid arguments\n"); - return; - } - - pdpu = to_dpu_plane(plane); - - if (!pdpu->pipe_hw || !pdpu->pipe_sblk || !pdpu->catalog) { - DPU_ERROR("invalid arguments\n"); - return; - } else if (!pdpu->pipe_hw->ops.setup_danger_safe_lut) { - return; - } - if (!pdpu->is_rt_pipe) { danger_lut = pdpu->catalog->perf.danger_lut_tbl [DPU_QOS_LUT_USAGE_NRT]; @@ -373,21 +329,7 @@ static void _dpu_plane_set_danger_lut(struct drm_plane *plane, static void _dpu_plane_set_qos_ctrl(struct drm_plane *plane, bool enable, u32 flags) { - struct dpu_plane *pdpu; - - if (!plane) { - DPU_ERROR("invalid arguments\n"); - return; - } - - pdpu = to_dpu_plane(plane); - - if (!pdpu->pipe_hw || !pdpu->pipe_sblk) { - DPU_ERROR("invalid arguments\n"); - return; - } else if (!pdpu->pipe_hw->ops.setup_qos_ctrl) { - return; - } + struct dpu_plane *pdpu = to_dpu_plane(plane); if (flags & DPU_PLANE_QOS_VBLANK_CTRL) { pdpu->pipe_qos_cfg.creq_vblank = pdpu->pipe_sblk->creq_vblank; @@ -423,35 +365,17 @@ static void _dpu_plane_set_qos_ctrl(struct drm_plane *plane, &pdpu->pipe_qos_cfg); } -int dpu_plane_danger_signal_ctrl(struct drm_plane *plane, bool enable) +static void dpu_plane_danger_signal_ctrl(struct drm_plane *plane, bool enable) { - struct dpu_plane *pdpu; - struct msm_drm_private *priv; - struct dpu_kms *dpu_kms; - - if (!plane || !plane->dev) { - DPU_ERROR("invalid arguments\n"); - return -EINVAL; - } - - priv = plane->dev->dev_private; - if (!priv || !priv->kms) { - DPU_ERROR("invalid KMS reference\n"); - return -EINVAL; - } - - dpu_kms = to_dpu_kms(priv->kms); - pdpu = to_dpu_plane(plane); + struct dpu_plane *pdpu = to_dpu_plane(plane); + struct dpu_kms *dpu_kms = _dpu_plane_get_kms(plane); if (!pdpu->is_rt_pipe) - goto end; + return; pm_runtime_get_sync(&dpu_kms->pdev->dev); _dpu_plane_set_qos_ctrl(plane, enable, DPU_PLANE_QOS_PANIC_CTRL); pm_runtime_put_sync(&dpu_kms->pdev->dev); - -end: - return 0; } /** @@ -462,29 +386,9 @@ end: static void _dpu_plane_set_ot_limit(struct drm_plane *plane, struct drm_crtc *crtc) { - struct dpu_plane *pdpu; + struct dpu_plane *pdpu = to_dpu_plane(plane); struct dpu_vbif_set_ot_params ot_params; - struct msm_drm_private *priv; - struct dpu_kms *dpu_kms; - - if (!plane || !plane->dev || !crtc) { - DPU_ERROR("invalid arguments plane %d crtc %d\n", - plane != 0, crtc != 0); - return; - } - - priv = plane->dev->dev_private; - if (!priv || !priv->kms) { - DPU_ERROR("invalid KMS reference\n"); - return; - } - - dpu_kms = to_dpu_kms(priv->kms); - pdpu = to_dpu_plane(plane); - if (!pdpu->pipe_hw) { - DPU_ERROR("invalid pipe reference\n"); - return; - } + struct dpu_kms *dpu_kms = _dpu_plane_get_kms(plane); memset(&ot_params, 0, sizeof(ot_params)); ot_params.xin_id = pdpu->pipe_hw->cap->xin_id; @@ -506,28 +410,9 @@ static void _dpu_plane_set_ot_limit(struct drm_plane *plane, */ static void _dpu_plane_set_qos_remap(struct drm_plane *plane) { - struct dpu_plane *pdpu; + struct dpu_plane *pdpu = to_dpu_plane(plane); struct dpu_vbif_set_qos_params qos_params; - struct msm_drm_private *priv; - struct dpu_kms *dpu_kms; - - if (!plane || !plane->dev) { - DPU_ERROR("invalid arguments\n"); - return; - } - - priv = plane->dev->dev_private; - if (!priv || !priv->kms) { - DPU_ERROR("invalid KMS reference\n"); - return; - } - - dpu_kms = to_dpu_kms(priv->kms); - pdpu = to_dpu_plane(plane); - if (!pdpu->pipe_hw) { - DPU_ERROR("invalid pipe reference\n"); - return; - } + struct dpu_kms *dpu_kms = _dpu_plane_get_kms(plane); memset(&qos_params, 0, sizeof(qos_params)); qos_params.vbif_idx = VBIF_RT; @@ -548,27 +433,12 @@ static void _dpu_plane_set_qos_remap(struct drm_plane *plane) /** * _dpu_plane_get_aspace: gets the address space */ -static int _dpu_plane_get_aspace( - struct dpu_plane *pdpu, - struct dpu_plane_state *pstate, - struct msm_gem_address_space **aspace) +static inline struct msm_gem_address_space *_dpu_plane_get_aspace( + struct dpu_plane *pdpu) { - struct dpu_kms *kms; + struct dpu_kms *kms = _dpu_plane_get_kms(&pdpu->base); - if (!pdpu || !pstate || !aspace) { - DPU_ERROR("invalid parameters\n"); - return -EINVAL; - } - - kms = _dpu_plane_get_kms(&pdpu->base); - if (!kms) { - DPU_ERROR("invalid kms\n"); - return -EINVAL; - } - - *aspace = kms->base.aspace; - - return 0; + return kms->base.aspace; } static inline void _dpu_plane_set_scanout(struct drm_plane *plane, @@ -576,29 +446,10 @@ static inline void _dpu_plane_set_scanout(struct drm_plane *plane, struct dpu_hw_pipe_cfg *pipe_cfg, struct drm_framebuffer *fb) { - struct dpu_plane *pdpu; - struct msm_gem_address_space *aspace = NULL; + struct dpu_plane *pdpu = to_dpu_plane(plane); + struct msm_gem_address_space *aspace = _dpu_plane_get_aspace(pdpu); int ret; - if (!plane || !pstate || !pipe_cfg || !fb) { - DPU_ERROR( - "invalid arg(s), plane %d state %d cfg %d fb %d\n", - plane != 0, pstate != 0, pipe_cfg != 0, fb != 0); - return; - } - - pdpu = to_dpu_plane(plane); - if (!pdpu->pipe_hw) { - DPU_ERROR_PLANE(pdpu, "invalid pipe_hw\n"); - return; - } - - ret = _dpu_plane_get_aspace(pdpu, pstate, &aspace); - if (ret) { - DPU_ERROR_PLANE(pdpu, "Failed to get aspace %d\n", ret); - return; - } - ret = dpu_format_populate_layout(aspace, fb, &pipe_cfg->layout); if (ret == -EAGAIN) DPU_DEBUG_PLANE(pdpu, "not updating same src addrs\n"); @@ -622,15 +473,6 @@ static void _dpu_plane_setup_scaler3(struct dpu_plane *pdpu, { uint32_t i; - if (!pdpu || !pstate || !scale_cfg || !fmt || !chroma_subsmpl_h || - !chroma_subsmpl_v) { - DPU_ERROR( - "pdpu %d pstate %d scale_cfg %d fmt %d smp_h %d smp_v %d\n", - !!pdpu, !!pstate, !!scale_cfg, !!fmt, chroma_subsmpl_h, - chroma_subsmpl_v); - return; - } - memset(scale_cfg, 0, sizeof(*scale_cfg)); memset(&pstate->pixel_ext, 0, sizeof(struct dpu_hw_pixel_ext)); @@ -734,17 +576,8 @@ static void _dpu_plane_setup_scaler(struct dpu_plane *pdpu, struct dpu_plane_state *pstate, const struct dpu_format *fmt, bool color_fill) { - struct dpu_hw_pixel_ext *pe; uint32_t chroma_subsmpl_h, chroma_subsmpl_v; - if (!pdpu || !fmt || !pstate) { - DPU_ERROR("invalid arg(s), plane %d fmt %d state %d\n", - pdpu != 0, fmt != 0, pstate != 0); - return; - } - - pe = &pstate->pixel_ext; - /* don't chroma subsample if decimating */ chroma_subsmpl_h = drm_format_horz_chroma_subsampling(fmt->base.pixel_format); @@ -772,21 +605,8 @@ static int _dpu_plane_color_fill(struct dpu_plane *pdpu, uint32_t color, uint32_t alpha) { const struct dpu_format *fmt; - const struct drm_plane *plane; - struct dpu_plane_state *pstate; - - if (!pdpu || !pdpu->base.state) { - DPU_ERROR("invalid plane\n"); - return -EINVAL; - } - - if (!pdpu->pipe_hw) { - DPU_ERROR_PLANE(pdpu, "invalid plane h/w pointer\n"); - return -EINVAL; - } - - plane = &pdpu->base; - pstate = to_dpu_plane_state(plane->state); + const struct drm_plane *plane = &pdpu->base; + struct dpu_plane_state *pstate = to_dpu_plane_state(plane->state); DPU_DEBUG_PLANE(pdpu, "\n"); @@ -837,12 +657,7 @@ static int _dpu_plane_color_fill(struct dpu_plane *pdpu, void dpu_plane_clear_multirect(const struct drm_plane_state *drm_state) { - struct dpu_plane_state *pstate; - - if (!drm_state) - return; - - pstate = to_dpu_plane_state(drm_state); + struct dpu_plane_state *pstate = to_dpu_plane_state(drm_state); pstate->multirect_index = DPU_SSPP_RECT_SOLO; pstate->multirect_mode = DPU_SSPP_MULTIRECT_NONE; @@ -973,15 +788,6 @@ done: void dpu_plane_get_ctl_flush(struct drm_plane *plane, struct dpu_hw_ctl *ctl, u32 *flush_sspp) { - struct dpu_plane_state *pstate; - - if (!plane || !flush_sspp) { - DPU_ERROR("invalid parameters\n"); - return; - } - - pstate = to_dpu_plane_state(plane->state); - *flush_sspp = ctl->ops.get_bitmask_sspp(ctl, dpu_plane_pipe(plane)); } @@ -995,7 +801,7 @@ static int dpu_plane_prepare_fb(struct drm_plane *plane, struct drm_gem_object *obj; struct msm_gem_object *msm_obj; struct dma_fence *fence; - struct msm_gem_address_space *aspace; + struct msm_gem_address_space *aspace = _dpu_plane_get_aspace(pdpu); int ret; if (!new_state->fb) @@ -1003,12 +809,6 @@ static int dpu_plane_prepare_fb(struct drm_plane *plane, DPU_DEBUG_PLANE(pdpu, "FB[%u]\n", fb->base.id); - ret = _dpu_plane_get_aspace(pdpu, pstate, &aspace); - if (ret) { - DPU_ERROR_PLANE(pdpu, "Failed to get aspace\n"); - return ret; - } - /* cache aspace */ pstate->aspace = aspace; @@ -1078,33 +878,30 @@ static bool dpu_plane_validate_src(struct drm_rect *src, drm_rect_equals(fb_rect, src); } -static int dpu_plane_sspp_atomic_check(struct drm_plane *plane, - struct drm_plane_state *state) +static int dpu_plane_atomic_check(struct drm_plane *plane, + struct drm_plane_state *state) { - int ret = 0; - struct dpu_plane *pdpu; - struct dpu_plane_state *pstate; + int ret = 0, min_scale; + struct dpu_plane *pdpu = to_dpu_plane(plane); + const struct drm_crtc_state *crtc_state = NULL; const struct dpu_format *fmt; struct drm_rect src, dst, fb_rect = { 0 }; - uint32_t max_upscale = 1, max_downscale = 1; uint32_t min_src_size, max_linewidth; - int hscale = 1, vscale = 1; - if (!plane || !state) { - DPU_ERROR("invalid arg(s), plane %d state %d\n", - plane != 0, state != 0); - ret = -EINVAL; - goto exit; - } - - pdpu = to_dpu_plane(plane); - pstate = to_dpu_plane_state(state); + if (state->crtc) + crtc_state = drm_atomic_get_new_crtc_state(state->state, + state->crtc); - if (!pdpu->pipe_sblk) { - DPU_ERROR_PLANE(pdpu, "invalid catalog\n"); - ret = -EINVAL; - goto exit; + min_scale = FRAC_16_16(1, pdpu->pipe_sblk->maxdwnscale); + ret = drm_atomic_helper_check_plane_state(state, crtc_state, min_scale, + pdpu->pipe_sblk->maxupscale << 16, + true, true); + if (ret) { + DPU_ERROR_PLANE(pdpu, "Check plane state failed (%d)\n", ret); + return ret; } + if (!state->visible) + return 0; src.x1 = state->src_x >> 16; src.y1 = state->src_y >> 16; @@ -1118,25 +915,6 @@ static int dpu_plane_sspp_atomic_check(struct drm_plane *plane, max_linewidth = pdpu->pipe_sblk->common->maxlinewidth; - if (pdpu->features & DPU_SSPP_SCALER) { - max_downscale = pdpu->pipe_sblk->maxdwnscale; - max_upscale = pdpu->pipe_sblk->maxupscale; - } - if (drm_rect_width(&src) < drm_rect_width(&dst)) - hscale = drm_rect_calc_hscale(&src, &dst, 1, max_upscale); - else - hscale = drm_rect_calc_hscale(&dst, &src, 1, max_downscale); - if (drm_rect_height(&src) < drm_rect_height(&dst)) - vscale = drm_rect_calc_vscale(&src, &dst, 1, max_upscale); - else - vscale = drm_rect_calc_vscale(&dst, &src, 1, max_downscale); - - DPU_DEBUG_PLANE(pdpu, "check %d -> %d\n", - dpu_plane_enabled(plane->state), dpu_plane_enabled(state)); - - if (!dpu_plane_enabled(state)) - goto exit; - fmt = to_dpu_format(msm_framebuffer_format(state->fb)); min_src_size = DPU_FORMAT_IS_YUV(fmt) ? 2 : 1; @@ -1147,13 +925,13 @@ static int dpu_plane_sspp_atomic_check(struct drm_plane *plane, | BIT(DPU_SSPP_CSC_10BIT))))) { DPU_ERROR_PLANE(pdpu, "plane doesn't have scaler/csc for yuv\n"); - ret = -EINVAL; + return -EINVAL; /* check src bounds */ } else if (!dpu_plane_validate_src(&src, &fb_rect, min_src_size)) { DPU_ERROR_PLANE(pdpu, "invalid source " DRM_RECT_FMT "\n", DRM_RECT_ARG(&src)); - ret = -E2BIG; + return -E2BIG; /* valid yuv image */ } else if (DPU_FORMAT_IS_YUV(fmt) && @@ -1162,41 +940,22 @@ static int dpu_plane_sspp_atomic_check(struct drm_plane *plane, drm_rect_height(&src) & 0x1)) { DPU_ERROR_PLANE(pdpu, "invalid yuv source " DRM_RECT_FMT "\n", DRM_RECT_ARG(&src)); - ret = -EINVAL; + return -EINVAL; /* min dst support */ } else if (drm_rect_width(&dst) < 0x1 || drm_rect_height(&dst) < 0x1) { DPU_ERROR_PLANE(pdpu, "invalid dest rect " DRM_RECT_FMT "\n", DRM_RECT_ARG(&dst)); - ret = -EINVAL; + return -EINVAL; /* check decimated source width */ } else if (drm_rect_width(&src) > max_linewidth) { DPU_ERROR_PLANE(pdpu, "invalid src " DRM_RECT_FMT " line:%u\n", DRM_RECT_ARG(&src), max_linewidth); - ret = -E2BIG; - - /* check scaler capability */ - } else if (hscale < 0 || vscale < 0) { - DPU_ERROR_PLANE(pdpu, "invalid scaling requested src=" - DRM_RECT_FMT " dst=" DRM_RECT_FMT "\n", - DRM_RECT_ARG(&src), DRM_RECT_ARG(&dst)); - ret = -E2BIG; + return -E2BIG; } -exit: - return ret; -} - -static int dpu_plane_atomic_check(struct drm_plane *plane, - struct drm_plane_state *state) -{ - if (!state->fb) - return 0; - - DPU_DEBUG_PLANE(to_dpu_plane(plane), "\n"); - - return dpu_plane_sspp_atomic_check(plane, state); + return 0; } void dpu_plane_flush(struct drm_plane *plane) @@ -1245,46 +1004,16 @@ void dpu_plane_set_error(struct drm_plane *plane, bool error) pdpu->is_error = error; } -static int dpu_plane_sspp_atomic_update(struct drm_plane *plane, - struct drm_plane_state *old_state) +static void dpu_plane_sspp_atomic_update(struct drm_plane *plane) { - uint32_t nplanes, src_flags; - struct dpu_plane *pdpu; - struct drm_plane_state *state; - struct dpu_plane_state *pstate; - struct dpu_plane_state *old_pstate; - const struct dpu_format *fmt; - struct drm_crtc *crtc; - struct drm_framebuffer *fb; - struct drm_rect src, dst; - - if (!plane) { - DPU_ERROR("invalid plane\n"); - return -EINVAL; - } else if (!plane->state) { - DPU_ERROR("invalid plane state\n"); - return -EINVAL; - } else if (!old_state) { - DPU_ERROR("invalid old state\n"); - return -EINVAL; - } - - pdpu = to_dpu_plane(plane); - state = plane->state; - - pstate = to_dpu_plane_state(state); - - old_pstate = to_dpu_plane_state(old_state); - - crtc = state->crtc; - fb = state->fb; - if (!crtc || !fb) { - DPU_ERROR_PLANE(pdpu, "invalid crtc %d or fb %d\n", - crtc != 0, fb != 0); - return -EINVAL; - } - fmt = to_dpu_format(msm_framebuffer_format(fb)); - nplanes = fmt->num_planes; + uint32_t src_flags; + struct dpu_plane *pdpu = to_dpu_plane(plane); + struct drm_plane_state *state = plane->state; + struct dpu_plane_state *pstate = to_dpu_plane_state(state); + struct drm_crtc *crtc = state->crtc; + struct drm_framebuffer *fb = state->fb; + const struct dpu_format *fmt = + to_dpu_format(msm_framebuffer_format(fb)); memset(&(pdpu->pipe_cfg), 0, sizeof(struct dpu_hw_pipe_cfg)); @@ -1295,28 +1024,27 @@ static int dpu_plane_sspp_atomic_update(struct drm_plane *plane, pdpu->is_rt_pipe = (dpu_crtc_get_client_type(crtc) != NRT_CLIENT); _dpu_plane_set_qos_ctrl(plane, false, DPU_PLANE_QOS_PANIC_CTRL); - src.x1 = state->src_x >> 16; - src.y1 = state->src_y >> 16; - src.x2 = src.x1 + (state->src_w >> 16); - src.y2 = src.y1 + (state->src_h >> 16); + DPU_DEBUG_PLANE(pdpu, "FB[%u] " DRM_RECT_FP_FMT "->crtc%u " DRM_RECT_FMT + ", %4.4s ubwc %d\n", fb->base.id, DRM_RECT_FP_ARG(&state->src), + crtc->base.id, DRM_RECT_ARG(&state->dst), + (char *)&fmt->base.pixel_format, DPU_FORMAT_IS_UBWC(fmt)); - dst = drm_plane_state_dest(state); + pdpu->pipe_cfg.src_rect = state->src; - DPU_DEBUG_PLANE(pdpu, "FB[%u] " DRM_RECT_FMT "->crtc%u " DRM_RECT_FMT - ", %4.4s ubwc %d\n", fb->base.id, DRM_RECT_ARG(&src), - crtc->base.id, DRM_RECT_ARG(&dst), - (char *)&fmt->base.pixel_format, - DPU_FORMAT_IS_UBWC(fmt)); + /* state->src is 16.16, src_rect is not */ + pdpu->pipe_cfg.src_rect.x1 >>= 16; + pdpu->pipe_cfg.src_rect.x2 >>= 16; + pdpu->pipe_cfg.src_rect.y1 >>= 16; + pdpu->pipe_cfg.src_rect.y2 >>= 16; - pdpu->pipe_cfg.src_rect = src; - pdpu->pipe_cfg.dst_rect = dst; + pdpu->pipe_cfg.dst_rect = state->dst; _dpu_plane_setup_scaler(pdpu, pstate, fmt, false); /* override for color fill */ if (pdpu->color_fill & DPU_PLANE_COLOR_FILL_FLAG) { /* skip remaining processing on color fill */ - return 0; + return; } if (pdpu->pipe_hw->ops.setup_rects) { @@ -1387,30 +1115,13 @@ static int dpu_plane_sspp_atomic_update(struct drm_plane *plane, } _dpu_plane_set_qos_remap(plane); - return 0; } -static void _dpu_plane_atomic_disable(struct drm_plane *plane, - struct drm_plane_state *old_state) +static void _dpu_plane_atomic_disable(struct drm_plane *plane) { - struct dpu_plane *pdpu; - struct drm_plane_state *state; - struct dpu_plane_state *pstate; - - if (!plane) { - DPU_ERROR("invalid plane\n"); - return; - } else if (!plane->state) { - DPU_ERROR("invalid plane state\n"); - return; - } else if (!old_state) { - DPU_ERROR("invalid old state\n"); - return; - } - - pdpu = to_dpu_plane(plane); - state = plane->state; - pstate = to_dpu_plane_state(state); + struct dpu_plane *pdpu = to_dpu_plane(plane); + struct drm_plane_state *state = plane->state; + struct dpu_plane_state *pstate = to_dpu_plane_state(state); trace_dpu_plane_disable(DRMID(plane), is_dpu_plane_virtual(plane), pstate->multirect_mode); @@ -1426,31 +1137,17 @@ static void _dpu_plane_atomic_disable(struct drm_plane *plane, static void dpu_plane_atomic_update(struct drm_plane *plane, struct drm_plane_state *old_state) { - struct dpu_plane *pdpu; - struct drm_plane_state *state; - - if (!plane) { - DPU_ERROR("invalid plane\n"); - return; - } else if (!plane->state) { - DPU_ERROR("invalid plane state\n"); - return; - } + struct dpu_plane *pdpu = to_dpu_plane(plane); + struct drm_plane_state *state = plane->state; - pdpu = to_dpu_plane(plane); pdpu->is_error = false; - state = plane->state; DPU_DEBUG_PLANE(pdpu, "\n"); - if (!dpu_plane_sspp_enabled(state)) { - _dpu_plane_atomic_disable(plane, old_state); + if (!state->visible) { + _dpu_plane_atomic_disable(plane); } else { - int ret; - - ret = dpu_plane_sspp_atomic_update(plane, old_state); - /* atomic_check should have ensured that this doesn't fail */ - WARN_ON(ret < 0); + dpu_plane_sspp_atomic_update(plane); } } @@ -1485,8 +1182,7 @@ static void dpu_plane_destroy(struct drm_plane *plane) /* this will destroy the states as well */ drm_plane_cleanup(plane); - if (pdpu->pipe_hw) - dpu_hw_sspp_destroy(pdpu->pipe_hw); + dpu_hw_sspp_destroy(pdpu->pipe_hw); kfree(pdpu); } @@ -1505,9 +1201,7 @@ static void dpu_plane_destroy_state(struct drm_plane *plane, pstate = to_dpu_plane_state(state); - /* remove ref count for frame buffers */ - if (state->fb) - drm_framebuffer_put(state->fb); + __drm_atomic_helper_plane_destroy_state(state); kfree(pstate); } @@ -1827,40 +1521,17 @@ bool is_dpu_plane_virtual(struct drm_plane *plane) /* initialize plane */ struct drm_plane *dpu_plane_init(struct drm_device *dev, - uint32_t pipe, bool primary_plane, + uint32_t pipe, enum drm_plane_type type, unsigned long possible_crtcs, u32 master_plane_id) { struct drm_plane *plane = NULL, *master_plane = NULL; const struct dpu_format_extended *format_list; struct dpu_plane *pdpu; - struct msm_drm_private *priv; - struct dpu_kms *kms; - enum drm_plane_type type; + struct msm_drm_private *priv = dev->dev_private; + struct dpu_kms *kms = to_dpu_kms(priv->kms); int zpos_max = DPU_ZPOS_MAX; int ret = -EINVAL; - if (!dev) { - DPU_ERROR("[%u]device is NULL\n", pipe); - goto exit; - } - - priv = dev->dev_private; - if (!priv) { - DPU_ERROR("[%u]private data is NULL\n", pipe); - goto exit; - } - - if (!priv->kms) { - DPU_ERROR("[%u]invalid KMS reference\n", pipe); - goto exit; - } - kms = to_dpu_kms(priv->kms); - - if (!kms->catalog) { - DPU_ERROR("[%u]invalid catalog reference\n", pipe); - goto exit; - } - /* create and zero local structure */ pdpu = kzalloc(sizeof(*pdpu), GFP_KERNEL); if (!pdpu) { @@ -1916,12 +1587,6 @@ struct drm_plane *dpu_plane_init(struct drm_device *dev, goto clean_sspp; } - if (pdpu->features & BIT(DPU_SSPP_CURSOR)) - type = DRM_PLANE_TYPE_CURSOR; - else if (primary_plane) - type = DRM_PLANE_TYPE_PRIMARY; - else - type = DRM_PLANE_TYPE_OVERLAY; ret = drm_universal_plane_init(dev, plane, 0xff, &dpu_plane_funcs, pdpu->formats, pdpu->nformats, NULL, type, NULL); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.h index f6fe6ddc7a3a..7fed0b627708 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.h @@ -122,7 +122,7 @@ void dpu_plane_set_error(struct drm_plane *plane, bool error); * dpu_plane_init - create new dpu plane for the given pipe * @dev: Pointer to DRM device * @pipe: dpu hardware pipe identifier - * @primary_plane: true if this pipe is primary plane for crtc + * @type: Plane type - PRIMARY/OVERLAY/CURSOR * @possible_crtcs: bitmask of crtc that can be attached to the given pipe * @master_plane_id: primary plane id of a multirect pipe. 0 value passed for * a regular plane initialization. A non-zero primary plane @@ -130,7 +130,7 @@ void dpu_plane_set_error(struct drm_plane *plane, bool error); * */ struct drm_plane *dpu_plane_init(struct drm_device *dev, - uint32_t pipe, bool primary_plane, + uint32_t pipe, enum drm_plane_type type, unsigned long possible_crtcs, u32 master_plane_id); /** diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_power_handle.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_power_handle.c index a75eebca2f37..fc14116789f2 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_power_handle.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_power_handle.c @@ -145,6 +145,7 @@ int dpu_power_resource_enable(struct dpu_power_handle *phandle, bool changed = false; u32 max_usecase_ndx = VOTE_INDEX_DISABLE, prev_usecase_ndx; struct dpu_power_client *client; + u32 event_type; if (!phandle || !pclient) { pr_err("invalid input argument\n"); @@ -181,19 +182,9 @@ int dpu_power_resource_enable(struct dpu_power_handle *phandle, if (!changed) goto end; - if (enable) { - dpu_power_event_trigger_locked(phandle, - DPU_POWER_EVENT_PRE_ENABLE); - dpu_power_event_trigger_locked(phandle, - DPU_POWER_EVENT_POST_ENABLE); - - } else { - dpu_power_event_trigger_locked(phandle, - DPU_POWER_EVENT_PRE_DISABLE); - dpu_power_event_trigger_locked(phandle, - DPU_POWER_EVENT_POST_DISABLE); - } + event_type = enable ? DPU_POWER_EVENT_ENABLE : DPU_POWER_EVENT_DISABLE; + dpu_power_event_trigger_locked(phandle, event_type); end: mutex_unlock(&phandle->phandle_lock); return 0; diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_power_handle.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_power_handle.h index 344f74464eca..a65b7a297f21 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_power_handle.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_power_handle.h @@ -23,17 +23,9 @@ #include "dpu_io_util.h" -/* event will be triggered before power handler disable */ -#define DPU_POWER_EVENT_PRE_DISABLE 0x1 - -/* event will be triggered after power handler disable */ -#define DPU_POWER_EVENT_POST_DISABLE 0x2 - -/* event will be triggered before power handler enable */ -#define DPU_POWER_EVENT_PRE_ENABLE 0x4 - -/* event will be triggered after power handler enable */ -#define DPU_POWER_EVENT_POST_ENABLE 0x8 +/* events will be triggered on power handler enable/disable */ +#define DPU_POWER_EVENT_DISABLE BIT(0) +#define DPU_POWER_EVENT_ENABLE BIT(1) /** * mdss_bus_vote_type: register bus vote type diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c index 13c0a36d4ef9..bdb117709674 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c @@ -16,7 +16,6 @@ #include "dpu_kms.h" #include "dpu_hw_lm.h" #include "dpu_hw_ctl.h" -#include "dpu_hw_cdm.h" #include "dpu_hw_pingpong.h" #include "dpu_hw_intf.h" #include "dpu_encoder.h" @@ -25,38 +24,13 @@ #define RESERVED_BY_OTHER(h, r) \ ((h)->rsvp && ((h)->rsvp->enc_id != (r)->enc_id)) -#define RM_RQ_LOCK(r) ((r)->top_ctrl & BIT(DPU_RM_TOPCTL_RESERVE_LOCK)) -#define RM_RQ_CLEAR(r) ((r)->top_ctrl & BIT(DPU_RM_TOPCTL_RESERVE_CLEAR)) -#define RM_RQ_DS(r) ((r)->top_ctrl & BIT(DPU_RM_TOPCTL_DS)) -#define RM_IS_TOPOLOGY_MATCH(t, r) ((t).num_lm == (r).num_lm && \ - (t).num_comp_enc == (r).num_enc && \ - (t).num_intf == (r).num_intf) - -struct dpu_rm_topology_def { - enum dpu_rm_topology_name top_name; - int num_lm; - int num_comp_enc; - int num_intf; - int num_ctl; - int needs_split_display; -}; - -static const struct dpu_rm_topology_def g_top_table[] = { - { DPU_RM_TOPOLOGY_NONE, 0, 0, 0, 0, false }, - { DPU_RM_TOPOLOGY_SINGLEPIPE, 1, 0, 1, 1, false }, - { DPU_RM_TOPOLOGY_DUALPIPE, 2, 0, 2, 2, true }, - { DPU_RM_TOPOLOGY_DUALPIPE_3DMERGE, 2, 0, 1, 1, false }, -}; - /** * struct dpu_rm_requirements - Reservation requirements parameter bundle - * @top_ctrl: topology control preference from kernel client - * @top: selected topology for the display + * @topology: selected topology for the display * @hw_res: Hardware resources required as reported by the encoders */ struct dpu_rm_requirements { - uint64_t top_ctrl; - const struct dpu_rm_topology_def *topology; + struct msm_display_topology topology; struct dpu_encoder_hw_resources hw_res; }; @@ -72,13 +46,11 @@ struct dpu_rm_requirements { * @enc_id: Reservations are tracked by Encoder DRM object ID. * CRTCs may be connected to multiple Encoders. * An encoder or connector id identifies the display path. - * @topology DRM<->HW topology use case */ struct dpu_rm_rsvp { struct list_head list; uint32_t seq; uint32_t enc_id; - enum dpu_rm_topology_name topology; }; /** @@ -122,8 +94,8 @@ static void _dpu_rm_print_rsvps( DPU_DEBUG("%d\n", stage); list_for_each_entry(rsvp, &rm->rsvps, list) { - DRM_DEBUG_KMS("%d rsvp[s%ue%u] topology %d\n", stage, rsvp->seq, - rsvp->enc_id, rsvp->topology); + DRM_DEBUG_KMS("%d rsvp[s%ue%u]\n", stage, rsvp->seq, + rsvp->enc_id); } for (type = 0; type < DPU_HW_BLK_MAX; type++) { @@ -146,18 +118,6 @@ struct dpu_hw_mdp *dpu_rm_get_mdp(struct dpu_rm *rm) return rm->hw_mdp; } -enum dpu_rm_topology_name -dpu_rm_get_topology_name(struct msm_display_topology topology) -{ - int i; - - for (i = 0; i < DPU_RM_TOPOLOGY_MAX; i++) - if (RM_IS_TOPOLOGY_MATCH(g_top_table[i], topology)) - return g_top_table[i].top_name; - - return DPU_RM_TOPOLOGY_NONE; -} - void dpu_rm_init_hw_iter( struct dpu_rm_hw_iter *iter, uint32_t enc_id, @@ -229,9 +189,6 @@ static void _dpu_rm_hw_destroy(enum dpu_hw_blk_type type, void *hw) case DPU_HW_BLK_CTL: dpu_hw_ctl_destroy(hw); break; - case DPU_HW_BLK_CDM: - dpu_hw_cdm_destroy(hw); - break; case DPU_HW_BLK_PINGPONG: dpu_hw_pingpong_destroy(hw); break; @@ -305,9 +262,6 @@ static int _dpu_rm_hw_blk_create( case DPU_HW_BLK_CTL: hw = dpu_hw_ctl_init(id, mmio, cat); break; - case DPU_HW_BLK_CDM: - hw = dpu_hw_cdm_init(id, mmio, cat, hw_mdp); - break; case DPU_HW_BLK_PINGPONG: hw = dpu_hw_pingpong_init(id, mmio, cat); break; @@ -438,15 +392,6 @@ int dpu_rm_init(struct dpu_rm *rm, } } - for (i = 0; i < cat->cdm_count; i++) { - rc = _dpu_rm_hw_blk_create(rm, cat, mmio, DPU_HW_BLK_CDM, - cat->cdm[i].id, &cat->cdm[i]); - if (rc) { - DPU_ERROR("failed: cdm hw not available\n"); - goto fail; - } - } - return 0; fail: @@ -455,6 +400,11 @@ fail: return rc; } +static bool _dpu_rm_needs_split_display(const struct msm_display_topology *top) +{ + return top->num_intf > 1; +} + /** * _dpu_rm_check_lm_and_get_connected_blks - check if proposed layer mixer meets * proposed use case requirements, incl. hardwired dependent blocks like @@ -538,14 +488,14 @@ static int _dpu_rm_reserve_lms( int lm_count = 0; int i, rc = 0; - if (!reqs->topology->num_lm) { - DPU_ERROR("invalid number of lm: %d\n", reqs->topology->num_lm); + if (!reqs->topology.num_lm) { + DPU_ERROR("invalid number of lm: %d\n", reqs->topology.num_lm); return -EINVAL; } /* Find a primary mixer */ dpu_rm_init_hw_iter(&iter_i, 0, DPU_HW_BLK_LM); - while (lm_count != reqs->topology->num_lm && + while (lm_count != reqs->topology.num_lm && _dpu_rm_get_hw_locked(rm, &iter_i)) { memset(&lm, 0, sizeof(lm)); memset(&pp, 0, sizeof(pp)); @@ -563,7 +513,7 @@ static int _dpu_rm_reserve_lms( /* Valid primary mixer found, find matching peers */ dpu_rm_init_hw_iter(&iter_j, 0, DPU_HW_BLK_LM); - while (lm_count != reqs->topology->num_lm && + while (lm_count != reqs->topology.num_lm && _dpu_rm_get_hw_locked(rm, &iter_j)) { if (iter_i.blk == iter_j.blk) continue; @@ -578,7 +528,7 @@ static int _dpu_rm_reserve_lms( } } - if (lm_count != reqs->topology->num_lm) { + if (lm_count != reqs->topology.num_lm) { DPU_DEBUG("unable to find appropriate mixers\n"); return -ENAVAIL; } @@ -600,14 +550,20 @@ static int _dpu_rm_reserve_lms( static int _dpu_rm_reserve_ctls( struct dpu_rm *rm, struct dpu_rm_rsvp *rsvp, - const struct dpu_rm_topology_def *top) + const struct msm_display_topology *top) { struct dpu_rm_hw_blk *ctls[MAX_BLOCKS]; struct dpu_rm_hw_iter iter; - int i = 0; + int i = 0, num_ctls = 0; + bool needs_split_display = false; memset(&ctls, 0, sizeof(ctls)); + /* each hw_intf needs its own hw_ctrl to program its control path */ + num_ctls = top->num_intf; + + needs_split_display = _dpu_rm_needs_split_display(top); + dpu_rm_init_hw_iter(&iter, 0, DPU_HW_BLK_CTL); while (_dpu_rm_get_hw_locked(rm, &iter)) { const struct dpu_hw_ctl *ctl = to_dpu_hw_ctl(iter.blk->hw); @@ -621,20 +577,20 @@ static int _dpu_rm_reserve_ctls( DPU_DEBUG("ctl %d caps 0x%lX\n", iter.blk->id, features); - if (top->needs_split_display != has_split_display) + if (needs_split_display != has_split_display) continue; ctls[i] = iter.blk; DPU_DEBUG("ctl %d match\n", iter.blk->id); - if (++i == top->num_ctl) + if (++i == num_ctls) break; } - if (i != top->num_ctl) + if (i != num_ctls) return -ENAVAIL; - for (i = 0; i < ARRAY_SIZE(ctls) && i < top->num_ctl; i++) { + for (i = 0; i < ARRAY_SIZE(ctls) && i < num_ctls; i++) { ctls[i]->rsvp_nxt = rsvp; trace_dpu_rm_reserve_ctls(ctls[i]->id, ctls[i]->type, rsvp->enc_id); @@ -643,55 +599,11 @@ static int _dpu_rm_reserve_ctls( return 0; } -static int _dpu_rm_reserve_cdm( - struct dpu_rm *rm, - struct dpu_rm_rsvp *rsvp, - uint32_t id, - enum dpu_hw_blk_type type) -{ - struct dpu_rm_hw_iter iter; - - DRM_DEBUG_KMS("type %d id %d\n", type, id); - - dpu_rm_init_hw_iter(&iter, 0, DPU_HW_BLK_CDM); - while (_dpu_rm_get_hw_locked(rm, &iter)) { - const struct dpu_hw_cdm *cdm = to_dpu_hw_cdm(iter.blk->hw); - const struct dpu_cdm_cfg *caps = cdm->caps; - bool match = false; - - if (RESERVED_BY_OTHER(iter.blk, rsvp)) - continue; - - if (type == DPU_HW_BLK_INTF && id != INTF_MAX) - match = test_bit(id, &caps->intf_connect); - - DRM_DEBUG_KMS("iter: type:%d id:%d enc:%d cdm:%lu match:%d\n", - iter.blk->type, iter.blk->id, rsvp->enc_id, - caps->intf_connect, match); - - if (!match) - continue; - - trace_dpu_rm_reserve_cdm(iter.blk->id, iter.blk->type, - rsvp->enc_id); - iter.blk->rsvp_nxt = rsvp; - break; - } - - if (!iter.hw) { - DPU_ERROR("couldn't reserve cdm for type %d id %d\n", type, id); - return -ENAVAIL; - } - - return 0; -} - static int _dpu_rm_reserve_intf( struct dpu_rm *rm, struct dpu_rm_rsvp *rsvp, uint32_t id, - enum dpu_hw_blk_type type, - bool needs_cdm) + enum dpu_hw_blk_type type) { struct dpu_rm_hw_iter iter; int ret = 0; @@ -719,9 +631,6 @@ static int _dpu_rm_reserve_intf( return -EINVAL; } - if (needs_cdm) - ret = _dpu_rm_reserve_cdm(rm, rsvp, id, type); - return ret; } @@ -738,7 +647,7 @@ static int _dpu_rm_reserve_intf_related_hw( continue; id = i + INTF_0; ret = _dpu_rm_reserve_intf(rm, rsvp, id, - DPU_HW_BLK_INTF, hw_res->needs_cdm); + DPU_HW_BLK_INTF); if (ret) return ret; } @@ -750,17 +659,14 @@ static int _dpu_rm_make_next_rsvp( struct dpu_rm *rm, struct drm_encoder *enc, struct drm_crtc_state *crtc_state, - struct drm_connector_state *conn_state, struct dpu_rm_rsvp *rsvp, struct dpu_rm_requirements *reqs) { int ret; - struct dpu_rm_topology_def topology; /* Create reservation info, tag reserved blocks with it as we go */ rsvp->seq = ++rm->rsvp_next_seq; rsvp->enc_id = enc->base.id; - rsvp->topology = reqs->topology->top_name; list_add_tail(&rsvp->list, &rm->rsvps); ret = _dpu_rm_reserve_lms(rm, rsvp, reqs); @@ -769,23 +675,12 @@ static int _dpu_rm_make_next_rsvp( return ret; } - /* - * Do assignment preferring to give away low-resource CTLs first: - * - Check mixers without Split Display - * - Only then allow to grab from CTLs with split display capability - */ - _dpu_rm_reserve_ctls(rm, rsvp, reqs->topology); - if (ret && !reqs->topology->needs_split_display) { - memcpy(&topology, reqs->topology, sizeof(topology)); - topology.needs_split_display = true; - _dpu_rm_reserve_ctls(rm, rsvp, &topology); - } + ret = _dpu_rm_reserve_ctls(rm, rsvp, &reqs->topology); if (ret) { DPU_ERROR("unable to find appropriate CTL\n"); return ret; } - /* Assign INTFs and blks whose usage is tied to them: CTL & CDM */ ret = _dpu_rm_reserve_intf_related_hw(rm, rsvp, &reqs->hw_res); if (ret) return ret; @@ -797,44 +692,16 @@ static int _dpu_rm_populate_requirements( struct dpu_rm *rm, struct drm_encoder *enc, struct drm_crtc_state *crtc_state, - struct drm_connector_state *conn_state, struct dpu_rm_requirements *reqs, struct msm_display_topology req_topology) { - int i; + dpu_encoder_get_hw_resources(enc, &reqs->hw_res); - memset(reqs, 0, sizeof(*reqs)); + reqs->topology = req_topology; - dpu_encoder_get_hw_resources(enc, &reqs->hw_res, conn_state); - - for (i = 0; i < DPU_RM_TOPOLOGY_MAX; i++) { - if (RM_IS_TOPOLOGY_MATCH(g_top_table[i], - req_topology)) { - reqs->topology = &g_top_table[i]; - break; - } - } - - if (!reqs->topology) { - DPU_ERROR("invalid topology for the display\n"); - return -EINVAL; - } - - /** - * Set the requirement based on caps if not set from user space - * This will ensure to select LM tied with DS blocks - * Currently, DS blocks are tied with LM 0 and LM 1 (primary display) - */ - if (!RM_RQ_DS(reqs) && rm->hw_mdp->caps->has_dest_scaler && - conn_state->connector->connector_type == DRM_MODE_CONNECTOR_DSI) - reqs->top_ctrl |= BIT(DPU_RM_TOPCTL_DS); - - DRM_DEBUG_KMS("top_ctrl: 0x%llX num_h_tiles: %d\n", reqs->top_ctrl, - reqs->hw_res.display_num_of_h_tiles); - DRM_DEBUG_KMS("num_lm: %d num_ctl: %d topology: %d split_display: %d\n", - reqs->topology->num_lm, reqs->topology->num_ctl, - reqs->topology->top_name, - reqs->topology->needs_split_display); + DRM_DEBUG_KMS("num_lm: %d num_enc: %d num_intf: %d\n", + reqs->topology.num_lm, reqs->topology.num_enc, + reqs->topology.num_intf); return 0; } @@ -860,29 +727,12 @@ static struct dpu_rm_rsvp *_dpu_rm_get_rsvp( return NULL; } -static struct drm_connector *_dpu_rm_get_connector( - struct drm_encoder *enc) -{ - struct drm_connector *conn = NULL; - struct list_head *connector_list = - &enc->dev->mode_config.connector_list; - - list_for_each_entry(conn, connector_list, head) - if (conn->encoder == enc) - return conn; - - return NULL; -} - /** * _dpu_rm_release_rsvp - release resources and release a reservation * @rm: KMS handle * @rsvp: RSVP pointer to release and release resources for */ -static void _dpu_rm_release_rsvp( - struct dpu_rm *rm, - struct dpu_rm_rsvp *rsvp, - struct drm_connector *conn) +static void _dpu_rm_release_rsvp(struct dpu_rm *rm, struct dpu_rm_rsvp *rsvp) { struct dpu_rm_rsvp *rsvp_c, *rsvp_n; struct dpu_rm_hw_blk *blk; @@ -923,7 +773,6 @@ static void _dpu_rm_release_rsvp( void dpu_rm_release(struct dpu_rm *rm, struct drm_encoder *enc) { struct dpu_rm_rsvp *rsvp; - struct drm_connector *conn; if (!rm || !enc) { DPU_ERROR("invalid params\n"); @@ -938,25 +787,15 @@ void dpu_rm_release(struct dpu_rm *rm, struct drm_encoder *enc) goto end; } - conn = _dpu_rm_get_connector(enc); - if (!conn) { - DPU_ERROR("failed to get connector for enc %d\n", enc->base.id); - goto end; - } - - _dpu_rm_release_rsvp(rm, rsvp, conn); + _dpu_rm_release_rsvp(rm, rsvp); end: mutex_unlock(&rm->rm_lock); } -static int _dpu_rm_commit_rsvp( - struct dpu_rm *rm, - struct dpu_rm_rsvp *rsvp, - struct drm_connector_state *conn_state) +static void _dpu_rm_commit_rsvp(struct dpu_rm *rm, struct dpu_rm_rsvp *rsvp) { struct dpu_rm_hw_blk *blk; enum dpu_hw_blk_type type; - int ret = 0; /* Swap next rsvp to be the active */ for (type = 0; type < DPU_HW_BLK_MAX; type++) { @@ -967,19 +806,12 @@ static int _dpu_rm_commit_rsvp( } } } - - if (!ret) - DRM_DEBUG_KMS("rsrv enc %d topology %d\n", rsvp->enc_id, - rsvp->topology); - - return ret; } int dpu_rm_reserve( struct dpu_rm *rm, struct drm_encoder *enc, struct drm_crtc_state *crtc_state, - struct drm_connector_state *conn_state, struct msm_display_topology topology, bool test_only) { @@ -987,25 +819,19 @@ int dpu_rm_reserve( struct dpu_rm_requirements reqs; int ret; - if (!rm || !enc || !crtc_state || !conn_state) { - DPU_ERROR("invalid arguments\n"); - return -EINVAL; - } - /* Check if this is just a page-flip */ if (!drm_atomic_crtc_needs_modeset(crtc_state)) return 0; - DRM_DEBUG_KMS("reserving hw for conn %d enc %d crtc %d test_only %d\n", - conn_state->connector->base.id, enc->base.id, - crtc_state->crtc->base.id, test_only); + DRM_DEBUG_KMS("reserving hw for enc %d crtc %d test_only %d\n", + enc->base.id, crtc_state->crtc->base.id, test_only); mutex_lock(&rm->rm_lock); _dpu_rm_print_rsvps(rm, DPU_RM_STAGE_BEGIN); - ret = _dpu_rm_populate_requirements(rm, enc, crtc_state, - conn_state, &reqs, topology); + ret = _dpu_rm_populate_requirements(rm, enc, crtc_state, &reqs, + topology); if (ret) { DPU_ERROR("failed to populate hw requirements\n"); goto end; @@ -1030,28 +856,15 @@ int dpu_rm_reserve( rsvp_cur = _dpu_rm_get_rsvp(rm, enc); - /* - * User can request that we clear out any reservation during the - * atomic_check phase by using this CLEAR bit - */ - if (rsvp_cur && test_only && RM_RQ_CLEAR(&reqs)) { - DPU_DEBUG("test_only & CLEAR: clear rsvp[s%de%d]\n", - rsvp_cur->seq, rsvp_cur->enc_id); - _dpu_rm_release_rsvp(rm, rsvp_cur, conn_state->connector); - rsvp_cur = NULL; - _dpu_rm_print_rsvps(rm, DPU_RM_STAGE_AFTER_CLEAR); - } - /* Check the proposed reservation, store it in hw's "next" field */ - ret = _dpu_rm_make_next_rsvp(rm, enc, crtc_state, conn_state, - rsvp_nxt, &reqs); + ret = _dpu_rm_make_next_rsvp(rm, enc, crtc_state, rsvp_nxt, &reqs); _dpu_rm_print_rsvps(rm, DPU_RM_STAGE_AFTER_RSVPNEXT); if (ret) { DPU_ERROR("failed to reserve hw resources: %d\n", ret); - _dpu_rm_release_rsvp(rm, rsvp_nxt, conn_state->connector); - } else if (test_only && !RM_RQ_LOCK(&reqs)) { + _dpu_rm_release_rsvp(rm, rsvp_nxt); + } else if (test_only) { /* * Normally, if test_only, test the reservation and then undo * However, if the user requests LOCK, then keep the reservation @@ -1059,15 +872,11 @@ int dpu_rm_reserve( */ DPU_DEBUG("test_only: discard test rsvp[s%de%d]\n", rsvp_nxt->seq, rsvp_nxt->enc_id); - _dpu_rm_release_rsvp(rm, rsvp_nxt, conn_state->connector); + _dpu_rm_release_rsvp(rm, rsvp_nxt); } else { - if (test_only && RM_RQ_LOCK(&reqs)) - DPU_DEBUG("test_only & LOCK: lock rsvp[s%de%d]\n", - rsvp_nxt->seq, rsvp_nxt->enc_id); - - _dpu_rm_release_rsvp(rm, rsvp_cur, conn_state->connector); + _dpu_rm_release_rsvp(rm, rsvp_cur); - ret = _dpu_rm_commit_rsvp(rm, rsvp_nxt, conn_state); + _dpu_rm_commit_rsvp(rm, rsvp_nxt); } _dpu_rm_print_rsvps(rm, DPU_RM_STAGE_FINAL); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.h index ffd1841a6067..b8273bd23801 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.h @@ -21,39 +21,6 @@ #include "dpu_hw_top.h" /** - * enum dpu_rm_topology_name - HW resource use case in use by connector - * @DPU_RM_TOPOLOGY_NONE: No topology in use currently - * @DPU_RM_TOPOLOGY_SINGLEPIPE: 1 LM, 1 PP, 1 INTF/WB - * @DPU_RM_TOPOLOGY_DUALPIPE: 2 LM, 2 PP, 2 INTF/WB - * @DPU_RM_TOPOLOGY_DUALPIPE_3DMERGE: 2 LM, 2 PP, 3DMux, 1 INTF/WB - */ -enum dpu_rm_topology_name { - DPU_RM_TOPOLOGY_NONE = 0, - DPU_RM_TOPOLOGY_SINGLEPIPE, - DPU_RM_TOPOLOGY_DUALPIPE, - DPU_RM_TOPOLOGY_DUALPIPE_3DMERGE, - DPU_RM_TOPOLOGY_MAX, -}; - -/** - * enum dpu_rm_topology_control - HW resource use case in use by connector - * @DPU_RM_TOPCTL_RESERVE_LOCK: If set, in AtomicTest phase, after a successful - * test, reserve the resources for this display. - * Normal behavior would not impact the reservation - * list during the AtomicTest phase. - * @DPU_RM_TOPCTL_RESERVE_CLEAR: If set, in AtomicTest phase, before testing, - * release any reservation held by this display. - * Normal behavior would not impact the - * reservation list during the AtomicTest phase. - * @DPU_RM_TOPCTL_DS : Require layer mixers with DS capabilities - */ -enum dpu_rm_topology_control { - DPU_RM_TOPCTL_RESERVE_LOCK, - DPU_RM_TOPCTL_RESERVE_CLEAR, - DPU_RM_TOPCTL_DS, -}; - -/** * struct dpu_rm - DPU dynamic hardware resource manager * @dev: device handle for event logging purposes * @rsvps: list of hardware reservations by each crtc->encoder->connector @@ -125,7 +92,6 @@ int dpu_rm_destroy(struct dpu_rm *rm); * @rm: DPU Resource Manager handle * @drm_enc: DRM Encoder handle * @crtc_state: Proposed Atomic DRM CRTC State handle - * @conn_state: Proposed Atomic DRM Connector State handle * @topology: Pointer to topology info for the display * @test_only: Atomic-Test phase, discard results (unless property overrides) * @Return: 0 on Success otherwise -ERROR @@ -133,7 +99,6 @@ int dpu_rm_destroy(struct dpu_rm *rm); int dpu_rm_reserve(struct dpu_rm *rm, struct drm_encoder *drm_enc, struct drm_crtc_state *crtc_state, - struct drm_connector_state *conn_state, struct msm_display_topology topology, bool test_only); @@ -187,13 +152,4 @@ bool dpu_rm_get_hw(struct dpu_rm *rm, struct dpu_rm_hw_iter *iter); */ int dpu_rm_check_property_topctl(uint64_t val); -/** - * dpu_rm_get_topology_name - returns the name of the the given topology - * definition - * @topology: topology definition - * @Return: name of the topology - */ -enum dpu_rm_topology_name -dpu_rm_get_topology_name(struct msm_display_topology topology); - #endif /* __DPU_RM_H__ */ diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_trace.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_trace.h index ae0ca5076238..e12c4cefb742 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_trace.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_trace.h @@ -468,14 +468,16 @@ TRACE_EVENT(dpu_enc_frame_done_cb, TRACE_EVENT(dpu_enc_trigger_flush, TP_PROTO(uint32_t drm_id, enum dpu_intf intf_idx, - int pending_kickoff_cnt, int ctl_idx, u32 pending_flush_ret), + int pending_kickoff_cnt, int ctl_idx, u32 extra_flush_bits, + u32 pending_flush_ret), TP_ARGS(drm_id, intf_idx, pending_kickoff_cnt, ctl_idx, - pending_flush_ret), + extra_flush_bits, pending_flush_ret), TP_STRUCT__entry( __field( uint32_t, drm_id ) __field( enum dpu_intf, intf_idx ) __field( int, pending_kickoff_cnt ) __field( int, ctl_idx ) + __field( u32, extra_flush_bits ) __field( u32, pending_flush_ret ) ), TP_fast_assign( @@ -483,12 +485,14 @@ TRACE_EVENT(dpu_enc_trigger_flush, __entry->intf_idx = intf_idx; __entry->pending_kickoff_cnt = pending_kickoff_cnt; __entry->ctl_idx = ctl_idx; + __entry->extra_flush_bits = extra_flush_bits; __entry->pending_flush_ret = pending_flush_ret; ), TP_printk("id=%u, intf_idx=%d, pending_kickoff_cnt=%d ctl_idx=%d " - "pending_flush_ret=%u", __entry->drm_id, - __entry->intf_idx, __entry->pending_kickoff_cnt, - __entry->ctl_idx, __entry->pending_flush_ret) + "extra_flush_bits=0x%x pending_flush_ret=0x%x", + __entry->drm_id, __entry->intf_idx, + __entry->pending_kickoff_cnt, __entry->ctl_idx, + __entry->extra_flush_bits, __entry->pending_flush_ret) ); DECLARE_EVENT_CLASS(dpu_enc_ktime_template, @@ -682,37 +686,41 @@ TRACE_EVENT(dpu_crtc_setup_mixer, TP_STRUCT__entry( __field( uint32_t, crtc_id ) __field( uint32_t, plane_id ) - __field( struct drm_plane_state*,state ) - __field( struct dpu_plane_state*,pstate ) + __field( uint32_t, fb_id ) + __field_struct( struct drm_rect, src_rect ) + __field_struct( struct drm_rect, dst_rect ) __field( uint32_t, stage_idx ) + __field( enum dpu_stage, stage ) __field( enum dpu_sspp, sspp ) + __field( uint32_t, multirect_idx ) + __field( uint32_t, multirect_mode ) __field( uint32_t, pixel_format ) __field( uint64_t, modifier ) ), TP_fast_assign( __entry->crtc_id = crtc_id; __entry->plane_id = plane_id; - __entry->state = state; - __entry->pstate = pstate; + __entry->fb_id = state ? state->fb->base.id : 0; + __entry->src_rect = drm_plane_state_src(state); + __entry->dst_rect = drm_plane_state_dest(state); __entry->stage_idx = stage_idx; + __entry->stage = pstate->stage; __entry->sspp = sspp; + __entry->multirect_idx = pstate->multirect_index; + __entry->multirect_mode = pstate->multirect_mode; __entry->pixel_format = pixel_format; __entry->modifier = modifier; ), - TP_printk("crtc_id:%u plane_id:%u fb_id:%u src:{%ux%u+%ux%u} " - "dst:{%ux%u+%ux%u} stage_idx:%u stage:%d, sspp:%d " + TP_printk("crtc_id:%u plane_id:%u fb_id:%u src:" DRM_RECT_FP_FMT + " dst:" DRM_RECT_FMT " stage_idx:%u stage:%d, sspp:%d " "multirect_index:%d multirect_mode:%u pix_format:%u " "modifier:%llu", - __entry->crtc_id, __entry->plane_id, - __entry->state->fb ? __entry->state->fb->base.id : -1, - __entry->state->src_w >> 16, __entry->state->src_h >> 16, - __entry->state->src_x >> 16, __entry->state->src_y >> 16, - __entry->state->crtc_w, __entry->state->crtc_h, - __entry->state->crtc_x, __entry->state->crtc_y, - __entry->stage_idx, __entry->pstate->stage, __entry->sspp, - __entry->pstate->multirect_index, - __entry->pstate->multirect_mode, __entry->pixel_format, - __entry->modifier) + __entry->crtc_id, __entry->plane_id, __entry->fb_id, + DRM_RECT_FP_ARG(&__entry->src_rect), + DRM_RECT_ARG(&__entry->dst_rect), + __entry->stage_idx, __entry->stage, __entry->sspp, + __entry->multirect_idx, __entry->multirect_mode, + __entry->pixel_format, __entry->modifier) ); TRACE_EVENT(dpu_crtc_setup_lm_bounds, @@ -721,15 +729,15 @@ TRACE_EVENT(dpu_crtc_setup_lm_bounds, TP_STRUCT__entry( __field( uint32_t, drm_id ) __field( int, mixer ) - __field( struct drm_rect *, bounds ) + __field_struct( struct drm_rect, bounds ) ), TP_fast_assign( __entry->drm_id = drm_id; __entry->mixer = mixer; - __entry->bounds = bounds; + __entry->bounds = *bounds; ), TP_printk("id:%u mixer:%d bounds:" DRM_RECT_FMT, __entry->drm_id, - __entry->mixer, DRM_RECT_ARG(__entry->bounds)) + __entry->mixer, DRM_RECT_ARG(&__entry->bounds)) ); TRACE_EVENT(dpu_crtc_vblank_enable, @@ -740,21 +748,25 @@ TRACE_EVENT(dpu_crtc_vblank_enable, __field( uint32_t, drm_id ) __field( uint32_t, enc_id ) __field( bool, enable ) - __field( struct dpu_crtc *, crtc ) + __field( bool, enabled ) + __field( bool, suspend ) + __field( bool, vblank_requested ) ), TP_fast_assign( __entry->drm_id = drm_id; __entry->enc_id = enc_id; __entry->enable = enable; - __entry->crtc = crtc; + __entry->enabled = crtc->enabled; + __entry->suspend = crtc->suspend; + __entry->vblank_requested = crtc->vblank_requested; ), TP_printk("id:%u encoder:%u enable:%s state{enabled:%s suspend:%s " "vblank_req:%s}", __entry->drm_id, __entry->enc_id, __entry->enable ? "true" : "false", - __entry->crtc->enabled ? "true" : "false", - __entry->crtc->suspend ? "true" : "false", - __entry->crtc->vblank_requested ? "true" : "false") + __entry->enabled ? "true" : "false", + __entry->suspend ? "true" : "false", + __entry->vblank_requested ? "true" : "false") ); DECLARE_EVENT_CLASS(dpu_crtc_enable_template, @@ -763,18 +775,22 @@ DECLARE_EVENT_CLASS(dpu_crtc_enable_template, TP_STRUCT__entry( __field( uint32_t, drm_id ) __field( bool, enable ) - __field( struct dpu_crtc *, crtc ) + __field( bool, enabled ) + __field( bool, suspend ) + __field( bool, vblank_requested ) ), TP_fast_assign( __entry->drm_id = drm_id; __entry->enable = enable; - __entry->crtc = crtc; + __entry->enabled = crtc->enabled; + __entry->suspend = crtc->suspend; + __entry->vblank_requested = crtc->vblank_requested; ), TP_printk("id:%u enable:%s state{enabled:%s suspend:%s vblank_req:%s}", __entry->drm_id, __entry->enable ? "true" : "false", - __entry->crtc->enabled ? "true" : "false", - __entry->crtc->suspend ? "true" : "false", - __entry->crtc->vblank_requested ? "true" : "false") + __entry->enabled ? "true" : "false", + __entry->suspend ? "true" : "false", + __entry->vblank_requested ? "true" : "false") ); DEFINE_EVENT(dpu_crtc_enable_template, dpu_crtc_set_suspend, TP_PROTO(uint32_t drm_id, bool enable, struct dpu_crtc *crtc), @@ -814,24 +830,24 @@ TRACE_EVENT(dpu_plane_set_scanout, TP_ARGS(index, layout, multirect_index), TP_STRUCT__entry( __field( enum dpu_sspp, index ) - __field( struct dpu_hw_fmt_layout*, layout ) + __field_struct( struct dpu_hw_fmt_layout, layout ) __field( enum dpu_sspp_multirect_index, multirect_index) ), TP_fast_assign( __entry->index = index; - __entry->layout = layout; + __entry->layout = *layout; __entry->multirect_index = multirect_index; ), TP_printk("index:%d layout:{%ux%u @ [%u/%u, %u/%u, %u/%u, %u/%u]} " - "multirect_index:%d", __entry->index, __entry->layout->width, - __entry->layout->height, __entry->layout->plane_addr[0], - __entry->layout->plane_size[0], - __entry->layout->plane_addr[1], - __entry->layout->plane_size[1], - __entry->layout->plane_addr[2], - __entry->layout->plane_size[2], - __entry->layout->plane_addr[3], - __entry->layout->plane_size[3], __entry->multirect_index) + "multirect_index:%d", __entry->index, __entry->layout.width, + __entry->layout.height, __entry->layout.plane_addr[0], + __entry->layout.plane_size[0], + __entry->layout.plane_addr[1], + __entry->layout.plane_size[1], + __entry->layout.plane_addr[2], + __entry->layout.plane_size[2], + __entry->layout.plane_addr[3], + __entry->layout.plane_size[3], __entry->multirect_index) ); TRACE_EVENT(dpu_plane_disable, @@ -868,10 +884,6 @@ DECLARE_EVENT_CLASS(dpu_rm_iter_template, TP_printk("id:%d type:%d enc_id:%u", __entry->id, __entry->type, __entry->enc_id) ); -DEFINE_EVENT(dpu_rm_iter_template, dpu_rm_reserve_cdm, - TP_PROTO(uint32_t id, enum dpu_hw_blk_type type, uint32_t enc_id), - TP_ARGS(id, type, enc_id) -); DEFINE_EVENT(dpu_rm_iter_template, dpu_rm_reserve_intf, TP_PROTO(uint32_t id, enum dpu_hw_blk_type type, uint32_t enc_id), TP_ARGS(id, type, enc_id) @@ -979,16 +991,16 @@ TRACE_EVENT(dpu_core_perf_update_clk, TP_PROTO(struct drm_device *dev, bool stop_req, u64 clk_rate), TP_ARGS(dev, stop_req, clk_rate), TP_STRUCT__entry( - __field( struct drm_device *, dev ) + __string( dev_name, dev->unique ) __field( bool, stop_req ) __field( u64, clk_rate ) ), TP_fast_assign( - __entry->dev = dev; + __assign_str(dev_name, dev->unique); __entry->stop_req = stop_req; __entry->clk_rate = clk_rate; ), - TP_printk("dev:%s stop_req:%s clk_rate:%llu", __entry->dev->unique, + TP_printk("dev:%s stop_req:%s clk_rate:%llu", __get_str(dev_name), __entry->stop_req ? "true" : "false", __entry->clk_rate) ); diff --git a/drivers/gpu/drm/msm/disp/mdp5/mdp5_plane.c b/drivers/gpu/drm/msm/disp/mdp5/mdp5_plane.c index d5e4f0de321a..310459541e48 100644 --- a/drivers/gpu/drm/msm/disp/mdp5/mdp5_plane.c +++ b/drivers/gpu/drm/msm/disp/mdp5/mdp5_plane.c @@ -184,7 +184,7 @@ static void mdp5_plane_reset(struct drm_plane *plane) struct mdp5_plane_state *mdp5_state; if (plane->state && plane->state->fb) - drm_framebuffer_unreference(plane->state->fb); + drm_framebuffer_put(plane->state->fb); kfree(to_mdp5_plane_state(plane->state)); mdp5_state = kzalloc(sizeof(*mdp5_state), GFP_KERNEL); @@ -227,7 +227,7 @@ static void mdp5_plane_destroy_state(struct drm_plane *plane, struct mdp5_plane_state *pstate = to_mdp5_plane_state(state); if (state->fb) - drm_framebuffer_unreference(state->fb); + drm_framebuffer_put(state->fb); kfree(pstate); } @@ -258,7 +258,6 @@ static void mdp5_plane_cleanup_fb(struct drm_plane *plane, msm_framebuffer_cleanup(fb, kms->aspace); } -#define FRAC_16_16(mult, div) (((mult) << 16) / (div)) static int mdp5_plane_atomic_check_with_state(struct drm_crtc_state *crtc_state, struct drm_plane_state *state) { diff --git a/drivers/gpu/drm/msm/dsi/dsi.c b/drivers/gpu/drm/msm/dsi/dsi.c index ff8164cc6738..a9768f823290 100644 --- a/drivers/gpu/drm/msm/dsi/dsi.c +++ b/drivers/gpu/drm/msm/dsi/dsi.c @@ -83,6 +83,7 @@ static struct msm_dsi *dsi_init(struct platform_device *pdev) return ERR_PTR(-ENOMEM); DBG("dsi probed=%p", msm_dsi); + msm_dsi->id = -1; msm_dsi->pdev = pdev; platform_set_drvdata(pdev, msm_dsi); @@ -117,8 +118,13 @@ static int dsi_bind(struct device *dev, struct device *master, void *data) DBG(""); msm_dsi = dsi_init(pdev); - if (IS_ERR(msm_dsi)) - return PTR_ERR(msm_dsi); + if (IS_ERR(msm_dsi)) { + /* Don't fail the bind if the dsi port is not connected */ + if (PTR_ERR(msm_dsi) == -ENODEV) + return 0; + else + return PTR_ERR(msm_dsi); + } priv->dsi[msm_dsi->id] = msm_dsi; diff --git a/drivers/gpu/drm/msm/dsi/dsi_host.c b/drivers/gpu/drm/msm/dsi/dsi_host.c index 96fb5f635314..9c6c523eacdc 100644 --- a/drivers/gpu/drm/msm/dsi/dsi_host.c +++ b/drivers/gpu/drm/msm/dsi/dsi_host.c @@ -1750,6 +1750,7 @@ static int dsi_host_parse_dt(struct msm_dsi_host *msm_host) if (ret) { dev_err(dev, "%s: invalid lane configuration %d\n", __func__, ret); + ret = -EINVAL; goto err; } @@ -1757,6 +1758,7 @@ static int dsi_host_parse_dt(struct msm_dsi_host *msm_host) device_node = of_graph_get_remote_node(np, 1, 0); if (!device_node) { dev_dbg(dev, "%s: no valid device\n", __func__); + ret = -ENODEV; goto err; } diff --git a/drivers/gpu/drm/msm/dsi/dsi_manager.c b/drivers/gpu/drm/msm/dsi/dsi_manager.c index 5224010d90e4..80aa6344185e 100644 --- a/drivers/gpu/drm/msm/dsi/dsi_manager.c +++ b/drivers/gpu/drm/msm/dsi/dsi_manager.c @@ -839,6 +839,8 @@ void msm_dsi_manager_unregister(struct msm_dsi *msm_dsi) if (msm_dsi->host) msm_dsi_host_unregister(msm_dsi->host); - msm_dsim->dsi[msm_dsi->id] = NULL; + + if (msm_dsi->id >= 0) + msm_dsim->dsi[msm_dsi->id] = NULL; } diff --git a/drivers/gpu/drm/msm/msm_drv.c b/drivers/gpu/drm/msm/msm_drv.c index 69dbdba183fe..5e758d95751a 100644 --- a/drivers/gpu/drm/msm/msm_drv.c +++ b/drivers/gpu/drm/msm/msm_drv.c @@ -338,7 +338,7 @@ static int msm_drm_uninit(struct device *dev) mdss->funcs->destroy(ddev); ddev->dev_private = NULL; - drm_dev_unref(ddev); + drm_dev_put(ddev); kfree(priv); @@ -453,7 +453,7 @@ static int msm_drm_init(struct device *dev, struct drm_driver *drv) priv = kzalloc(sizeof(*priv), GFP_KERNEL); if (!priv) { ret = -ENOMEM; - goto err_unref_drm_dev; + goto err_put_drm_dev; } ddev->dev_private = priv; @@ -654,8 +654,8 @@ err_destroy_mdss: mdss->funcs->destroy(ddev); err_free_priv: kfree(priv); -err_unref_drm_dev: - drm_dev_unref(ddev); +err_put_drm_dev: + drm_dev_put(ddev); return ret; } diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h index 8e510d5c758a..9d11f321f5a9 100644 --- a/drivers/gpu/drm/msm/msm_drv.h +++ b/drivers/gpu/drm/msm/msm_drv.h @@ -62,6 +62,8 @@ struct msm_gem_vma; #define MAX_BRIDGES 8 #define MAX_CONNECTORS 8 +#define FRAC_16_16(mult, div) (((mult) << 16) / (div)) + struct msm_file_private { rwlock_t queuelock; struct list_head submitqueues; diff --git a/drivers/gpu/drm/msm/msm_gem_submit.c b/drivers/gpu/drm/msm/msm_gem_submit.c index 7bd83e0afa97..7a7923e6220d 100644 --- a/drivers/gpu/drm/msm/msm_gem_submit.c +++ b/drivers/gpu/drm/msm/msm_gem_submit.c @@ -144,7 +144,7 @@ static int submit_lookup_objects(struct msm_gem_submit *submit, goto out_unlock; } - drm_gem_object_reference(obj); + drm_gem_object_get(obj); submit->bos[i].obj = msm_obj; @@ -396,7 +396,7 @@ static void submit_cleanup(struct msm_gem_submit *submit) struct msm_gem_object *msm_obj = submit->bos[i].obj; submit_unlock_unpin_bo(submit, i, false); list_del_init(&msm_obj->submit_entry); - drm_gem_object_unreference(&msm_obj->base); + drm_gem_object_put(&msm_obj->base); } ww_acquire_fini(&submit->ticket); diff --git a/drivers/gpu/drm/msm/msm_gpu.c b/drivers/gpu/drm/msm/msm_gpu.c index 5e808cfec345..11aac8337066 100644 --- a/drivers/gpu/drm/msm/msm_gpu.c +++ b/drivers/gpu/drm/msm/msm_gpu.c @@ -41,7 +41,11 @@ static int msm_devfreq_target(struct device *dev, unsigned long *freq, if (IS_ERR(opp)) return PTR_ERR(opp); - clk_set_rate(gpu->core_clk, *freq); + if (gpu->funcs->gpu_set_freq) + gpu->funcs->gpu_set_freq(gpu, (u64)*freq); + else + clk_set_rate(gpu->core_clk, *freq); + dev_pm_opp_put(opp); return 0; @@ -51,16 +55,14 @@ static int msm_devfreq_get_dev_status(struct device *dev, struct devfreq_dev_status *status) { struct msm_gpu *gpu = platform_get_drvdata(to_platform_device(dev)); - u64 cycles; - u32 freq = ((u32) status->current_frequency) / 1000000; ktime_t time; - status->current_frequency = (unsigned long) clk_get_rate(gpu->core_clk); - gpu->funcs->gpu_busy(gpu, &cycles); - - status->busy_time = ((u32) (cycles - gpu->devfreq.busy_cycles)) / freq; + if (gpu->funcs->gpu_get_freq) + status->current_frequency = gpu->funcs->gpu_get_freq(gpu); + else + status->current_frequency = clk_get_rate(gpu->core_clk); - gpu->devfreq.busy_cycles = cycles; + status->busy_time = gpu->funcs->gpu_busy(gpu); time = ktime_get(); status->total_time = ktime_us_delta(time, gpu->devfreq.time); @@ -73,7 +75,10 @@ static int msm_devfreq_get_cur_freq(struct device *dev, unsigned long *freq) { struct msm_gpu *gpu = platform_get_drvdata(to_platform_device(dev)); - *freq = (unsigned long) clk_get_rate(gpu->core_clk); + if (gpu->funcs->gpu_get_freq) + *freq = gpu->funcs->gpu_get_freq(gpu); + else + *freq = clk_get_rate(gpu->core_clk); return 0; } @@ -88,7 +93,7 @@ static struct devfreq_dev_profile msm_devfreq_profile = { static void msm_devfreq_init(struct msm_gpu *gpu) { /* We need target support to do devfreq */ - if (!gpu->funcs->gpu_busy || !gpu->core_clk) + if (!gpu->funcs->gpu_busy) return; msm_devfreq_profile.initial_freq = gpu->fast_rate; @@ -105,6 +110,8 @@ static void msm_devfreq_init(struct msm_gpu *gpu) dev_err(&gpu->pdev->dev, "Couldn't initialize GPU devfreq\n"); gpu->devfreq.devfreq = NULL; } + + devfreq_suspend_device(gpu->devfreq.devfreq); } static int enable_pwrrail(struct msm_gpu *gpu) @@ -184,6 +191,14 @@ static int disable_axi(struct msm_gpu *gpu) return 0; } +void msm_gpu_resume_devfreq(struct msm_gpu *gpu) +{ + gpu->devfreq.busy_cycles = 0; + gpu->devfreq.time = ktime_get(); + + devfreq_resume_device(gpu->devfreq.devfreq); +} + int msm_gpu_pm_resume(struct msm_gpu *gpu) { int ret; @@ -202,12 +217,7 @@ int msm_gpu_pm_resume(struct msm_gpu *gpu) if (ret) return ret; - if (gpu->devfreq.devfreq) { - gpu->devfreq.busy_cycles = 0; - gpu->devfreq.time = ktime_get(); - - devfreq_resume_device(gpu->devfreq.devfreq); - } + msm_gpu_resume_devfreq(gpu); gpu->needs_hw_init = true; @@ -220,8 +230,7 @@ int msm_gpu_pm_suspend(struct msm_gpu *gpu) DBG("%s", gpu->name); - if (gpu->devfreq.devfreq) - devfreq_suspend_device(gpu->devfreq.devfreq); + devfreq_suspend_device(gpu->devfreq.devfreq); ret = disable_axi(gpu); if (ret) @@ -367,8 +376,8 @@ static void msm_gpu_crashstate_capture(struct msm_gpu *gpu, msm_gpu_devcoredump_read, msm_gpu_devcoredump_free); } #else -static void msm_gpu_crashstate_capture(struct msm_gpu *gpu, char *comm, - char *cmd) +static void msm_gpu_crashstate_capture(struct msm_gpu *gpu, + struct msm_gem_submit *submit, char *comm, char *cmd) { } #endif diff --git a/drivers/gpu/drm/msm/msm_gpu.h b/drivers/gpu/drm/msm/msm_gpu.h index 9122ee6e55e4..f82bac086666 100644 --- a/drivers/gpu/drm/msm/msm_gpu.h +++ b/drivers/gpu/drm/msm/msm_gpu.h @@ -70,9 +70,11 @@ struct msm_gpu_funcs { /* for generation specific debugfs: */ int (*debugfs_init)(struct msm_gpu *gpu, struct drm_minor *minor); #endif - int (*gpu_busy)(struct msm_gpu *gpu, uint64_t *value); + unsigned long (*gpu_busy)(struct msm_gpu *gpu); struct msm_gpu_state *(*gpu_state_get)(struct msm_gpu *gpu); int (*gpu_state_put)(struct msm_gpu_state *state); + unsigned long (*gpu_get_freq)(struct msm_gpu *gpu); + void (*gpu_set_freq)(struct msm_gpu *gpu, unsigned long freq); }; struct msm_gpu { @@ -264,6 +266,7 @@ static inline void gpu_write64(struct msm_gpu *gpu, u32 lo, u32 hi, u64 val) int msm_gpu_pm_suspend(struct msm_gpu *gpu); int msm_gpu_pm_resume(struct msm_gpu *gpu); +void msm_gpu_resume_devfreq(struct msm_gpu *gpu); int msm_gpu_hw_init(struct msm_gpu *gpu); diff --git a/drivers/gpu/drm/msm/msm_rd.c b/drivers/gpu/drm/msm/msm_rd.c index 3aa8a8576abe..cca933458439 100644 --- a/drivers/gpu/drm/msm/msm_rd.c +++ b/drivers/gpu/drm/msm/msm_rd.c @@ -366,7 +366,7 @@ void msm_rd_dump_submit(struct msm_rd_state *rd, struct msm_gem_submit *submit, va_list args; va_start(args, fmt); - n = vsnprintf(msg, sizeof(msg), fmt, args); + n = vscnprintf(msg, sizeof(msg), fmt, args); va_end(args); rd_write_section(rd, RD_CMD, msg, ALIGN(n, 4)); @@ -375,11 +375,11 @@ void msm_rd_dump_submit(struct msm_rd_state *rd, struct msm_gem_submit *submit, rcu_read_lock(); task = pid_task(submit->pid, PIDTYPE_PID); if (task) { - n = snprintf(msg, sizeof(msg), "%.*s/%d: fence=%u", + n = scnprintf(msg, sizeof(msg), "%.*s/%d: fence=%u", TASK_COMM_LEN, task->comm, pid_nr(submit->pid), submit->seqno); } else { - n = snprintf(msg, sizeof(msg), "???/%d: fence=%u", + n = scnprintf(msg, sizeof(msg), "???/%d: fence=%u", pid_nr(submit->pid), submit->seqno); } rcu_read_unlock(); diff --git a/drivers/gpu/drm/nouveau/dispnv50/disp.c b/drivers/gpu/drm/nouveau/dispnv50/disp.c index 1dbd1dcdcf15..5f163a025e89 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/disp.c +++ b/drivers/gpu/drm/nouveau/dispnv50/disp.c @@ -36,6 +36,7 @@ #include <drm/drm_dp_helper.h> #include <drm/drm_fb_helper.h> #include <drm/drm_plane_helper.h> +#include <drm/drm_scdc_helper.h> #include <drm/drm_edid.h> #include <nvif/class.h> @@ -531,6 +532,7 @@ nv50_hdmi_disable(struct drm_encoder *encoder, struct nouveau_crtc *nv_crtc) static void nv50_hdmi_enable(struct drm_encoder *encoder, struct drm_display_mode *mode) { + struct nouveau_drm *drm = nouveau_drm(encoder->dev); struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc); struct nv50_disp *disp = nv50_disp(encoder->dev); @@ -548,9 +550,12 @@ nv50_hdmi_enable(struct drm_encoder *encoder, struct drm_display_mode *mode) .pwr.rekey = 56, /* binary driver, and tegra, constant */ }; struct nouveau_connector *nv_connector; + struct drm_hdmi_info *hdmi; u32 max_ac_packet; union hdmi_infoframe avi_frame; union hdmi_infoframe vendor_frame; + bool scdc_supported, high_tmds_clock_ratio = false, scrambling = false; + u8 config; int ret; int size; @@ -558,8 +563,11 @@ nv50_hdmi_enable(struct drm_encoder *encoder, struct drm_display_mode *mode) if (!drm_detect_hdmi_monitor(nv_connector->edid)) return; + hdmi = &nv_connector->base.display_info.hdmi; + scdc_supported = hdmi->scdc.supported; + ret = drm_hdmi_avi_infoframe_from_display_mode(&avi_frame.avi, mode, - false); + scdc_supported); if (!ret) { /* We have an AVI InfoFrame, populate it to the display */ args.pwr.avi_infoframe_length @@ -582,12 +590,42 @@ nv50_hdmi_enable(struct drm_encoder *encoder, struct drm_display_mode *mode) max_ac_packet -= 18; /* constant from tegra */ args.pwr.max_ac_packet = max_ac_packet / 32; + if (hdmi->scdc.scrambling.supported) { + high_tmds_clock_ratio = mode->clock > 340000; + scrambling = high_tmds_clock_ratio || + hdmi->scdc.scrambling.low_rates; + } + + args.pwr.scdc = + NV50_DISP_SOR_HDMI_PWR_V0_SCDC_SCRAMBLE * scrambling | + NV50_DISP_SOR_HDMI_PWR_V0_SCDC_DIV_BY_4 * high_tmds_clock_ratio; + size = sizeof(args.base) + sizeof(args.pwr) + args.pwr.avi_infoframe_length + args.pwr.vendor_infoframe_length; nvif_mthd(&disp->disp->object, 0, &args, size); + nv50_audio_enable(encoder, mode); + + /* If SCDC is supported by the downstream monitor, update + * divider / scrambling settings to what we programmed above. + */ + if (!hdmi->scdc.scrambling.supported) + return; + + ret = drm_scdc_readb(nv_encoder->i2c, SCDC_TMDS_CONFIG, &config); + if (ret < 0) { + NV_ERROR(drm, "Failure to read SCDC_TMDS_CONFIG: %d\n", ret); + return; + } + config &= ~(SCDC_TMDS_BIT_CLOCK_RATIO_BY_40 | SCDC_SCRAMBLING_ENABLE); + config |= SCDC_TMDS_BIT_CLOCK_RATIO_BY_40 * high_tmds_clock_ratio; + config |= SCDC_SCRAMBLING_ENABLE * scrambling; + ret = drm_scdc_writeb(nv_encoder->i2c, SCDC_TMDS_CONFIG, config); + if (ret < 0) + NV_ERROR(drm, "Failure to write SCDC_TMDS_CONFIG = 0x%02x: %d\n", + config, ret); } /****************************************************************************** diff --git a/drivers/gpu/drm/nouveau/include/nvif/cl5070.h b/drivers/gpu/drm/nouveau/include/nvif/cl5070.h index 7cdf53615d7b..bced81987269 100644 --- a/drivers/gpu/drm/nouveau/include/nvif/cl5070.h +++ b/drivers/gpu/drm/nouveau/include/nvif/cl5070.h @@ -69,7 +69,10 @@ struct nv50_disp_sor_hdmi_pwr_v0 { __u8 rekey; __u8 avi_infoframe_length; __u8 vendor_infoframe_length; - __u8 pad06[2]; +#define NV50_DISP_SOR_HDMI_PWR_V0_SCDC_SCRAMBLE (1 << 0) +#define NV50_DISP_SOR_HDMI_PWR_V0_SCDC_DIV_BY_4 (1 << 1) + __u8 scdc; + __u8 pad07[1]; }; struct nv50_disp_sor_lvds_script_v0 { diff --git a/drivers/gpu/drm/nouveau/nouveau_backlight.c b/drivers/gpu/drm/nouveau/nouveau_backlight.c index 408b955e5c39..5f5be6368aed 100644 --- a/drivers/gpu/drm/nouveau/nouveau_backlight.c +++ b/drivers/gpu/drm/nouveau/nouveau_backlight.c @@ -37,18 +37,19 @@ #include "nouveau_drv.h" #include "nouveau_reg.h" #include "nouveau_encoder.h" +#include "nouveau_connector.h" static struct ida bl_ida; #define BL_NAME_SIZE 15 // 12 for name + 2 for digits + 1 for '\0' -struct backlight_connector { - struct list_head head; +struct nouveau_backlight { + struct backlight_device *dev; int id; }; static bool -nouveau_get_backlight_name(char backlight_name[BL_NAME_SIZE], struct backlight_connector - *connector) +nouveau_get_backlight_name(char backlight_name[BL_NAME_SIZE], + struct nouveau_backlight *bl) { const int nb = ida_simple_get(&bl_ida, 0, 0, GFP_KERNEL); if (nb < 0 || nb >= 100) @@ -57,17 +58,18 @@ nouveau_get_backlight_name(char backlight_name[BL_NAME_SIZE], struct backlight_c snprintf(backlight_name, BL_NAME_SIZE, "nv_backlight%d", nb); else snprintf(backlight_name, BL_NAME_SIZE, "nv_backlight"); - connector->id = nb; + bl->id = nb; return true; } static int nv40_get_intensity(struct backlight_device *bd) { - struct nouveau_drm *drm = bl_get_data(bd); + struct nouveau_encoder *nv_encoder = bl_get_data(bd); + struct nouveau_drm *drm = nouveau_drm(nv_encoder->base.base.dev); struct nvif_object *device = &drm->client.device.object; int val = (nvif_rd32(device, NV40_PMC_BACKLIGHT) & - NV40_PMC_BACKLIGHT_MASK) >> 16; + NV40_PMC_BACKLIGHT_MASK) >> 16; return val; } @@ -75,13 +77,14 @@ nv40_get_intensity(struct backlight_device *bd) static int nv40_set_intensity(struct backlight_device *bd) { - struct nouveau_drm *drm = bl_get_data(bd); + struct nouveau_encoder *nv_encoder = bl_get_data(bd); + struct nouveau_drm *drm = nouveau_drm(nv_encoder->base.base.dev); struct nvif_object *device = &drm->client.device.object; int val = bd->props.brightness; int reg = nvif_rd32(device, NV40_PMC_BACKLIGHT); nvif_wr32(device, NV40_PMC_BACKLIGHT, - (val << 16) | (reg & ~NV40_PMC_BACKLIGHT_MASK)); + (val << 16) | (reg & ~NV40_PMC_BACKLIGHT_MASK)); return 0; } @@ -93,38 +96,19 @@ static const struct backlight_ops nv40_bl_ops = { }; static int -nv40_backlight_init(struct drm_connector *connector) +nv40_backlight_init(struct nouveau_encoder *encoder, + struct backlight_properties *props, + const struct backlight_ops **ops) { - struct nouveau_drm *drm = nouveau_drm(connector->dev); + struct nouveau_drm *drm = nouveau_drm(encoder->base.base.dev); struct nvif_object *device = &drm->client.device.object; - struct backlight_properties props; - struct backlight_device *bd; - struct backlight_connector bl_connector; - char backlight_name[BL_NAME_SIZE]; if (!(nvif_rd32(device, NV40_PMC_BACKLIGHT) & NV40_PMC_BACKLIGHT_MASK)) - return 0; - - memset(&props, 0, sizeof(struct backlight_properties)); - props.type = BACKLIGHT_RAW; - props.max_brightness = 31; - if (!nouveau_get_backlight_name(backlight_name, &bl_connector)) { - NV_ERROR(drm, "Failed to retrieve a unique name for the backlight interface\n"); - return 0; - } - bd = backlight_device_register(backlight_name , connector->kdev, drm, - &nv40_bl_ops, &props); - - if (IS_ERR(bd)) { - if (bl_connector.id > 0) - ida_simple_remove(&bl_ida, bl_connector.id); - return PTR_ERR(bd); - } - list_add(&bl_connector.head, &drm->bl_connectors); - drm->backlight = bd; - bd->props.brightness = nv40_get_intensity(bd); - backlight_update_status(bd); + return -ENODEV; + props->type = BACKLIGHT_RAW; + props->max_brightness = 31; + *ops = &nv40_bl_ops; return 0; } @@ -154,7 +138,7 @@ nv50_set_intensity(struct backlight_device *bd) u32 val = (bd->props.brightness * div) / 100; nvif_wr32(device, NV50_PDISP_SOR_PWM_CTL(or), - NV50_PDISP_SOR_PWM_CTL_NEW | val); + NV50_PDISP_SOR_PWM_CTL_NEW | val); return 0; } @@ -194,9 +178,10 @@ nva3_set_intensity(struct backlight_device *bd) div = nvif_rd32(device, NV50_PDISP_SOR_PWM_DIV(or)); val = (bd->props.brightness * div) / 100; if (div) { - nvif_wr32(device, NV50_PDISP_SOR_PWM_CTL(or), val | - NV50_PDISP_SOR_PWM_CTL_NEW | - NVA3_PDISP_SOR_PWM_CTL_UNK); + nvif_wr32(device, NV50_PDISP_SOR_PWM_CTL(or), + val | + NV50_PDISP_SOR_PWM_CTL_NEW | + NVA3_PDISP_SOR_PWM_CTL_UNK); return 0; } @@ -210,110 +195,119 @@ static const struct backlight_ops nva3_bl_ops = { }; static int -nv50_backlight_init(struct drm_connector *connector) +nv50_backlight_init(struct nouveau_encoder *nv_encoder, + struct backlight_properties *props, + const struct backlight_ops **ops) { - struct nouveau_drm *drm = nouveau_drm(connector->dev); + struct nouveau_drm *drm = nouveau_drm(nv_encoder->base.base.dev); struct nvif_object *device = &drm->client.device.object; - struct nouveau_encoder *nv_encoder; - struct backlight_properties props; - struct backlight_device *bd; - const struct backlight_ops *ops; - struct backlight_connector bl_connector; - char backlight_name[BL_NAME_SIZE]; - - nv_encoder = find_encoder(connector, DCB_OUTPUT_LVDS); - if (!nv_encoder) { - nv_encoder = find_encoder(connector, DCB_OUTPUT_DP); - if (!nv_encoder) - return -ENODEV; - } if (!nvif_rd32(device, NV50_PDISP_SOR_PWM_CTL(ffs(nv_encoder->dcb->or) - 1))) - return 0; + return -ENODEV; if (drm->client.device.info.chipset <= 0xa0 || drm->client.device.info.chipset == 0xaa || drm->client.device.info.chipset == 0xac) - ops = &nv50_bl_ops; + *ops = &nv50_bl_ops; else - ops = &nva3_bl_ops; - - memset(&props, 0, sizeof(struct backlight_properties)); - props.type = BACKLIGHT_RAW; - props.max_brightness = 100; - if (!nouveau_get_backlight_name(backlight_name, &bl_connector)) { - NV_ERROR(drm, "Failed to retrieve a unique name for the backlight interface\n"); - return 0; - } - bd = backlight_device_register(backlight_name , connector->kdev, - nv_encoder, ops, &props); + *ops = &nva3_bl_ops; - if (IS_ERR(bd)) { - if (bl_connector.id > 0) - ida_simple_remove(&bl_ida, bl_connector.id); - return PTR_ERR(bd); - } + props->type = BACKLIGHT_RAW; + props->max_brightness = 100; - list_add(&bl_connector.head, &drm->bl_connectors); - drm->backlight = bd; - bd->props.brightness = bd->ops->get_brightness(bd); - backlight_update_status(bd); return 0; } int -nouveau_backlight_init(struct drm_device *dev) +nouveau_backlight_init(struct drm_connector *connector) { - struct nouveau_drm *drm = nouveau_drm(dev); + struct nouveau_drm *drm = nouveau_drm(connector->dev); + struct nouveau_backlight *bl; + struct nouveau_encoder *nv_encoder = NULL; struct nvif_device *device = &drm->client.device; - struct drm_connector *connector; - struct drm_connector_list_iter conn_iter; - - INIT_LIST_HEAD(&drm->bl_connectors); + char backlight_name[BL_NAME_SIZE]; + struct backlight_properties props = {0}; + const struct backlight_ops *ops; + int ret; if (apple_gmux_present()) { - NV_INFO(drm, "Apple GMUX detected: not registering Nouveau backlight interface\n"); + NV_INFO_ONCE(drm, "Apple GMUX detected: not registering Nouveau backlight interface\n"); return 0; } - drm_connector_list_iter_begin(dev, &conn_iter); - drm_for_each_connector_iter(connector, &conn_iter) { - if (connector->connector_type != DRM_MODE_CONNECTOR_LVDS && - connector->connector_type != DRM_MODE_CONNECTOR_eDP) - continue; - - switch (device->info.family) { - case NV_DEVICE_INFO_V0_CURIE: - return nv40_backlight_init(connector); - case NV_DEVICE_INFO_V0_TESLA: - case NV_DEVICE_INFO_V0_FERMI: - case NV_DEVICE_INFO_V0_KEPLER: - case NV_DEVICE_INFO_V0_MAXWELL: - return nv50_backlight_init(connector); - default: - break; - } + if (connector->connector_type == DRM_MODE_CONNECTOR_LVDS) + nv_encoder = find_encoder(connector, DCB_OUTPUT_LVDS); + else if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) + nv_encoder = find_encoder(connector, DCB_OUTPUT_DP); + else + return 0; + + if (!nv_encoder) + return 0; + + switch (device->info.family) { + case NV_DEVICE_INFO_V0_CURIE: + ret = nv40_backlight_init(nv_encoder, &props, &ops); + break; + case NV_DEVICE_INFO_V0_TESLA: + case NV_DEVICE_INFO_V0_FERMI: + case NV_DEVICE_INFO_V0_KEPLER: + case NV_DEVICE_INFO_V0_MAXWELL: + ret = nv50_backlight_init(nv_encoder, &props, &ops); + break; + default: + return 0; } - drm_connector_list_iter_end(&conn_iter); + + if (ret == -ENODEV) + return 0; + else if (ret) + return ret; + + bl = kzalloc(sizeof(*bl), GFP_KERNEL); + if (!bl) + return -ENOMEM; + + if (!nouveau_get_backlight_name(backlight_name, bl)) { + NV_ERROR(drm, "Failed to retrieve a unique name for the backlight interface\n"); + goto fail_alloc; + } + + bl->dev = backlight_device_register(backlight_name, connector->kdev, + nv_encoder, ops, &props); + if (IS_ERR(bl->dev)) { + if (bl->id >= 0) + ida_simple_remove(&bl_ida, bl->id); + ret = PTR_ERR(bl->dev); + goto fail_alloc; + } + + nouveau_connector(connector)->backlight = bl; + bl->dev->props.brightness = bl->dev->ops->get_brightness(bl->dev); + backlight_update_status(bl->dev); return 0; + +fail_alloc: + kfree(bl); + return ret; } void -nouveau_backlight_exit(struct drm_device *dev) +nouveau_backlight_fini(struct drm_connector *connector) { - struct nouveau_drm *drm = nouveau_drm(dev); - struct backlight_connector *connector; + struct nouveau_connector *nv_conn = nouveau_connector(connector); + struct nouveau_backlight *bl = nv_conn->backlight; - list_for_each_entry(connector, &drm->bl_connectors, head) { - if (connector->id >= 0) - ida_simple_remove(&bl_ida, connector->id); - } + if (!bl) + return; - if (drm->backlight) { - backlight_device_unregister(drm->backlight); - drm->backlight = NULL; - } + if (bl->id >= 0) + ida_simple_remove(&bl_ida, bl->id); + + backlight_device_unregister(bl->dev); + nv_conn->backlight = NULL; + kfree(bl); } void diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c index dde8724aa8f5..fd80661dff92 100644 --- a/drivers/gpu/drm/nouveau/nouveau_connector.c +++ b/drivers/gpu/drm/nouveau/nouveau_connector.c @@ -886,6 +886,22 @@ nouveau_connector_detect_depth(struct drm_connector *connector) } static int +nouveau_connector_late_register(struct drm_connector *connector) +{ + int ret; + + ret = nouveau_backlight_init(connector); + + return ret; +} + +static void +nouveau_connector_early_unregister(struct drm_connector *connector) +{ + nouveau_backlight_fini(connector); +} + +static int nouveau_connector_get_modes(struct drm_connector *connector) { struct drm_device *dev = connector->dev; @@ -953,18 +969,33 @@ nouveau_connector_get_modes(struct drm_connector *connector) } static unsigned -get_tmds_link_bandwidth(struct drm_connector *connector, bool hdmi) +get_tmds_link_bandwidth(struct drm_connector *connector) { struct nouveau_connector *nv_connector = nouveau_connector(connector); + struct nouveau_encoder *nv_encoder = nv_connector->detected_encoder; struct nouveau_drm *drm = nouveau_drm(connector->dev); struct dcb_output *dcb = nv_connector->detected_encoder->dcb; + struct drm_display_info *info = NULL; + const unsigned duallink_scale = + nouveau_duallink && nv_encoder->dcb->duallink_possible ? 2 : 1; + + if (drm_detect_hdmi_monitor(nv_connector->edid)) + info = &nv_connector->base.display_info; - if (hdmi) { + if (info) { if (nouveau_hdmimhz > 0) return nouveau_hdmimhz * 1000; /* Note: these limits are conservative, some Fermi's * can do 297 MHz. Unclear how this can be determined. */ + if (drm->client.device.info.chipset >= 0x120) { + const int max_tmds_clock = + info->hdmi.scdc.scrambling.supported ? + 594000 : 340000; + return info->max_tmds_clock ? + min(info->max_tmds_clock, max_tmds_clock) : + max_tmds_clock; + } if (drm->client.device.info.family >= NV_DEVICE_INFO_V0_KEPLER) return 297000; if (drm->client.device.info.family >= NV_DEVICE_INFO_V0_FERMI) @@ -972,13 +1003,13 @@ get_tmds_link_bandwidth(struct drm_connector *connector, bool hdmi) } if (dcb->location != DCB_LOC_ON_CHIP || drm->client.device.info.chipset >= 0x46) - return 165000; + return 165000 * duallink_scale; else if (drm->client.device.info.chipset >= 0x40) - return 155000; + return 155000 * duallink_scale; else if (drm->client.device.info.chipset >= 0x18) - return 135000; + return 135000 * duallink_scale; else - return 112000; + return 112000 * duallink_scale; } static enum drm_mode_status @@ -990,7 +1021,6 @@ nouveau_connector_mode_valid(struct drm_connector *connector, struct drm_encoder *encoder = to_drm_encoder(nv_encoder); unsigned min_clock = 25000, max_clock = min_clock; unsigned clock = mode->clock; - bool hdmi; switch (nv_encoder->dcb->type) { case DCB_OUTPUT_LVDS: @@ -1003,11 +1033,7 @@ nouveau_connector_mode_valid(struct drm_connector *connector, max_clock = 400000; break; case DCB_OUTPUT_TMDS: - hdmi = drm_detect_hdmi_monitor(nv_connector->edid); - max_clock = get_tmds_link_bandwidth(connector, hdmi); - if (!hdmi && nouveau_duallink && - nv_encoder->dcb->duallink_possible) - max_clock *= 2; + max_clock = get_tmds_link_bandwidth(connector); break; case DCB_OUTPUT_ANALOG: max_clock = nv_encoder->dcb->crtconf.maxfreq; @@ -1069,6 +1095,8 @@ nouveau_connector_funcs = { .atomic_destroy_state = nouveau_conn_atomic_destroy_state, .atomic_set_property = nouveau_conn_atomic_set_property, .atomic_get_property = nouveau_conn_atomic_get_property, + .late_register = nouveau_connector_late_register, + .early_unregister = nouveau_connector_early_unregister, }; static const struct drm_connector_funcs @@ -1084,6 +1112,8 @@ nouveau_connector_funcs_lvds = { .atomic_destroy_state = nouveau_conn_atomic_destroy_state, .atomic_set_property = nouveau_conn_atomic_set_property, .atomic_get_property = nouveau_conn_atomic_get_property, + .late_register = nouveau_connector_late_register, + .early_unregister = nouveau_connector_early_unregister, }; static int diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.h b/drivers/gpu/drm/nouveau/nouveau_connector.h index 0acc07555bcd..f57ef35b1e5e 100644 --- a/drivers/gpu/drm/nouveau/nouveau_connector.h +++ b/drivers/gpu/drm/nouveau/nouveau_connector.h @@ -39,6 +39,10 @@ struct nvkm_i2c_port; +#ifdef CONFIG_DRM_NOUVEAU_BACKLIGHT +struct nouveau_backlight; +#endif + struct nouveau_connector { struct drm_connector base; enum dcb_connector_type type; @@ -55,6 +59,9 @@ struct nouveau_connector { struct nouveau_encoder *detected_encoder; struct edid *edid; struct drm_display_mode *native_mode; +#ifdef CONFIG_DRM_NOUVEAU_BACKLIGHT + struct nouveau_backlight *backlight; +#endif }; static inline struct nouveau_connector *nouveau_connector( @@ -181,4 +188,30 @@ int nouveau_conn_atomic_get_property(struct drm_connector *, const struct drm_connector_state *, struct drm_property *, u64 *); struct drm_display_mode *nouveau_conn_native_mode(struct drm_connector *); + +#ifdef CONFIG_DRM_NOUVEAU_BACKLIGHT +extern int nouveau_backlight_init(struct drm_connector *); +extern void nouveau_backlight_fini(struct drm_connector *); +extern void nouveau_backlight_ctor(void); +extern void nouveau_backlight_dtor(void); +#else +static inline int +nouveau_backlight_init(struct drm_connector *connector) +{ + return 0; +} + +static inline void +nouveau_backlight_fini(struct drm_connector *connector) { +} + +static inline void +nouveau_backlight_ctor(void) { +} + +static inline void +nouveau_backlight_dtor(void) { +} +#endif + #endif /* __NOUVEAU_CONNECTOR_H__ */ diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c index 540c0cbbfcee..f326ffd86766 100644 --- a/drivers/gpu/drm/nouveau/nouveau_display.c +++ b/drivers/gpu/drm/nouveau/nouveau_display.c @@ -582,7 +582,6 @@ nouveau_display_create(struct drm_device *dev) goto vblank_err; } - nouveau_backlight_init(dev); INIT_WORK(&drm->hpd_work, nouveau_display_hpd_work); #ifdef CONFIG_ACPI drm->acpi_nb.notifier_call = nouveau_display_acpi_ntfy; @@ -607,7 +606,6 @@ nouveau_display_destroy(struct drm_device *dev) #ifdef CONFIG_ACPI unregister_acpi_notifier(&nouveau_drm(dev)->acpi_nb); #endif - nouveau_backlight_exit(dev); nouveau_display_vblank_fini(dev); drm_kms_helper_poll_fini(dev); diff --git a/drivers/gpu/drm/nouveau/nouveau_display.h b/drivers/gpu/drm/nouveau/nouveau_display.h index ff92b54ce448..eb77e41c2d4e 100644 --- a/drivers/gpu/drm/nouveau/nouveau_display.h +++ b/drivers/gpu/drm/nouveau/nouveau_display.h @@ -85,31 +85,6 @@ int nouveau_display_dumb_map_offset(struct drm_file *, struct drm_device *, void nouveau_hdmi_mode_set(struct drm_encoder *, struct drm_display_mode *); -#ifdef CONFIG_DRM_NOUVEAU_BACKLIGHT -extern int nouveau_backlight_init(struct drm_device *); -extern void nouveau_backlight_exit(struct drm_device *); -extern void nouveau_backlight_ctor(void); -extern void nouveau_backlight_dtor(void); -#else -static inline int -nouveau_backlight_init(struct drm_device *dev) -{ - return 0; -} - -static inline void -nouveau_backlight_exit(struct drm_device *dev) { -} - -static inline void -nouveau_backlight_ctor(void) { -} - -static inline void -nouveau_backlight_dtor(void) { -} -#endif - struct drm_framebuffer * nouveau_user_framebuffer_create(struct drm_device *, struct drm_file *, const struct drm_mode_fb_cmd2 *); diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c index 74d2283f2c28..2b2baf6e0e0d 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drm.c +++ b/drivers/gpu/drm/nouveau/nouveau_drm.c @@ -458,75 +458,8 @@ nouveau_accel_init(struct nouveau_drm *drm) nouveau_bo_move_init(drm); } -static int nouveau_drm_probe(struct pci_dev *pdev, - const struct pci_device_id *pent) -{ - struct nvkm_device *device; - struct apertures_struct *aper; - bool boot = false; - int ret; - - if (vga_switcheroo_client_probe_defer(pdev)) - return -EPROBE_DEFER; - - /* We need to check that the chipset is supported before booting - * fbdev off the hardware, as there's no way to put it back. - */ - ret = nvkm_device_pci_new(pdev, NULL, "error", true, false, 0, &device); - if (ret) - return ret; - - nvkm_device_del(&device); - - /* Remove conflicting drivers (vesafb, efifb etc). */ - aper = alloc_apertures(3); - if (!aper) - return -ENOMEM; - - aper->ranges[0].base = pci_resource_start(pdev, 1); - aper->ranges[0].size = pci_resource_len(pdev, 1); - aper->count = 1; - - if (pci_resource_len(pdev, 2)) { - aper->ranges[aper->count].base = pci_resource_start(pdev, 2); - aper->ranges[aper->count].size = pci_resource_len(pdev, 2); - aper->count++; - } - - if (pci_resource_len(pdev, 3)) { - aper->ranges[aper->count].base = pci_resource_start(pdev, 3); - aper->ranges[aper->count].size = pci_resource_len(pdev, 3); - aper->count++; - } - -#ifdef CONFIG_X86 - boot = pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW; -#endif - if (nouveau_modeset != 2) - drm_fb_helper_remove_conflicting_framebuffers(aper, "nouveaufb", boot); - kfree(aper); - - ret = nvkm_device_pci_new(pdev, nouveau_config, nouveau_debug, - true, true, ~0ULL, &device); - if (ret) - return ret; - - pci_set_master(pdev); - - if (nouveau_atomic) - driver_pci.driver_features |= DRIVER_ATOMIC; - - ret = drm_get_pci_dev(pdev, pent, &driver_pci); - if (ret) { - nvkm_device_del(&device); - return ret; - } - - return 0; -} - static int -nouveau_drm_load(struct drm_device *dev, unsigned long flags) +nouveau_drm_device_init(struct drm_device *dev) { struct nouveau_drm *drm; int ret; @@ -538,11 +471,11 @@ nouveau_drm_load(struct drm_device *dev, unsigned long flags) ret = nouveau_cli_init(drm, "DRM-master", &drm->master); if (ret) - return ret; + goto fail_alloc; ret = nouveau_cli_init(drm, "DRM", &drm->client); if (ret) - return ret; + goto fail_master; dev->irq_enabled = true; @@ -605,13 +538,15 @@ fail_bios: fail_ttm: nouveau_vga_fini(drm); nouveau_cli_fini(&drm->client); +fail_master: nouveau_cli_fini(&drm->master); +fail_alloc: kfree(drm); return ret; } static void -nouveau_drm_unload(struct drm_device *dev) +nouveau_drm_device_fini(struct drm_device *dev) { struct nouveau_drm *drm = nouveau_drm(dev); @@ -640,18 +575,116 @@ nouveau_drm_unload(struct drm_device *dev) kfree(drm); } +static int nouveau_drm_probe(struct pci_dev *pdev, + const struct pci_device_id *pent) +{ + struct nvkm_device *device; + struct drm_device *drm_dev; + struct apertures_struct *aper; + bool boot = false; + int ret; + + if (vga_switcheroo_client_probe_defer(pdev)) + return -EPROBE_DEFER; + + /* We need to check that the chipset is supported before booting + * fbdev off the hardware, as there's no way to put it back. + */ + ret = nvkm_device_pci_new(pdev, NULL, "error", true, false, 0, &device); + if (ret) + return ret; + + nvkm_device_del(&device); + + /* Remove conflicting drivers (vesafb, efifb etc). */ + aper = alloc_apertures(3); + if (!aper) + return -ENOMEM; + + aper->ranges[0].base = pci_resource_start(pdev, 1); + aper->ranges[0].size = pci_resource_len(pdev, 1); + aper->count = 1; + + if (pci_resource_len(pdev, 2)) { + aper->ranges[aper->count].base = pci_resource_start(pdev, 2); + aper->ranges[aper->count].size = pci_resource_len(pdev, 2); + aper->count++; + } + + if (pci_resource_len(pdev, 3)) { + aper->ranges[aper->count].base = pci_resource_start(pdev, 3); + aper->ranges[aper->count].size = pci_resource_len(pdev, 3); + aper->count++; + } + +#ifdef CONFIG_X86 + boot = pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW; +#endif + if (nouveau_modeset != 2) + drm_fb_helper_remove_conflicting_framebuffers(aper, "nouveaufb", boot); + kfree(aper); + + ret = nvkm_device_pci_new(pdev, nouveau_config, nouveau_debug, + true, true, ~0ULL, &device); + if (ret) + return ret; + + pci_set_master(pdev); + + if (nouveau_atomic) + driver_pci.driver_features |= DRIVER_ATOMIC; + + drm_dev = drm_dev_alloc(&driver_pci, &pdev->dev); + if (IS_ERR(drm_dev)) { + ret = PTR_ERR(drm_dev); + goto fail_nvkm; + } + + ret = pci_enable_device(pdev); + if (ret) + goto fail_drm; + + drm_dev->pdev = pdev; + pci_set_drvdata(pdev, drm_dev); + + ret = nouveau_drm_device_init(drm_dev); + if (ret) + goto fail_pci; + + ret = drm_dev_register(drm_dev, pent->driver_data); + if (ret) + goto fail_drm_dev_init; + + return 0; + +fail_drm_dev_init: + nouveau_drm_device_fini(drm_dev); +fail_pci: + pci_disable_device(pdev); +fail_drm: + drm_dev_put(drm_dev); +fail_nvkm: + nvkm_device_del(&device); + return ret; +} + void nouveau_drm_device_remove(struct drm_device *dev) { + struct pci_dev *pdev = dev->pdev; struct nouveau_drm *drm = nouveau_drm(dev); struct nvkm_client *client; struct nvkm_device *device; + drm_dev_unregister(dev); + dev->irq_enabled = false; client = nvxx_client(&drm->client.base); device = nvkm_device_find(client->device); - drm_put_dev(dev); + nouveau_drm_device_fini(dev); + pci_disable_device(pdev); + drm_dev_put(dev); nvkm_device_del(&device); } @@ -1018,8 +1051,6 @@ driver_stub = { DRIVER_GEM | DRIVER_MODESET | DRIVER_PRIME | DRIVER_RENDER | DRIVER_KMS_LEGACY_CONTEXT, - .load = nouveau_drm_load, - .unload = nouveau_drm_unload, .open = nouveau_drm_open, .postclose = nouveau_drm_postclose, .lastclose = nouveau_vga_lastclose, diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h index 6e1acaec3400..0b2191fa96f7 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drv.h +++ b/drivers/gpu/drm/nouveau/nouveau_drv.h @@ -194,8 +194,6 @@ struct nouveau_drm { /* modesetting */ struct nvbios vbios; struct nouveau_display *display; - struct backlight_device *backlight; - struct list_head bl_connectors; struct work_struct hpd_work; struct work_struct fbcon_work; int fbcon_new_state; @@ -244,10 +242,12 @@ void nouveau_drm_device_remove(struct drm_device *dev); struct nouveau_cli *_cli = (c); \ dev_##l(_cli->drm->dev->dev, "%s: "f, _cli->name, ##a); \ } while(0) + #define NV_FATAL(drm,f,a...) NV_PRINTK(crit, &(drm)->client, f, ##a) #define NV_ERROR(drm,f,a...) NV_PRINTK(err, &(drm)->client, f, ##a) #define NV_WARN(drm,f,a...) NV_PRINTK(warn, &(drm)->client, f, ##a) #define NV_INFO(drm,f,a...) NV_PRINTK(info, &(drm)->client, f, ##a) + #define NV_DEBUG(drm,f,a...) do { \ if (unlikely(drm_debug & DRM_UT_DRIVER)) \ NV_PRINTK(info, &(drm)->client, f, ##a); \ @@ -257,6 +257,12 @@ void nouveau_drm_device_remove(struct drm_device *dev); NV_PRINTK(info, &(drm)->client, f, ##a); \ } while(0) +#define NV_PRINTK_ONCE(l,c,f,a...) NV_PRINTK(l##_once,c,f, ##a) + +#define NV_ERROR_ONCE(drm,f,a...) NV_PRINTK_ONCE(err, &(drm)->client, f, ##a) +#define NV_WARN_ONCE(drm,f,a...) NV_PRINTK_ONCE(warn, &(drm)->client, f, ##a) +#define NV_INFO_ONCE(drm,f,a...) NV_PRINTK_ONCE(info, &(drm)->client, f, ##a) + extern int nouveau_modeset; #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild index 3d485dbf310a..8089ac9a12e2 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild @@ -50,6 +50,7 @@ nvkm-y += nvkm/engine/disp/hdmig84.o nvkm-y += nvkm/engine/disp/hdmigt215.o nvkm-y += nvkm/engine/disp/hdmigf119.o nvkm-y += nvkm/engine/disp/hdmigk104.o +nvkm-y += nvkm/engine/disp/hdmigm200.o nvkm-y += nvkm/engine/disp/hdmigv100.o nvkm-y += nvkm/engine/disp/conn.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/hdmigm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/hdmigm200.c new file mode 100644 index 000000000000..9b16a08eb4d9 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/hdmigm200.c @@ -0,0 +1,36 @@ +/* + * Copyright 2018 Ilia Mirkin + * + * 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: Ilia Mirkin + */ +#include "hdmi.h" + +void +gm200_hdmi_scdc(struct nvkm_ior *ior, int head, u8 scdc) +{ + struct nvkm_device *device = ior->disp->engine.subdev.device; + const u32 hoff = head * 0x800; + const u32 ctrl = scdc & 0x3; + + nvkm_mask(device, 0x61c5bc + hoff, 0x00000003, ctrl); + + ior->tmds.high_speed = !!(scdc & 0x2); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h index 19911211a12a..0f0c86c32ec3 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h @@ -41,6 +41,11 @@ struct nvkm_ior { u8 nr; u8 bw; } dp; + + /* Armed TMDS state. */ + struct { + bool high_speed; + } tmds; }; struct nvkm_ior_func { @@ -61,6 +66,7 @@ struct nvkm_ior_func { void (*ctrl)(struct nvkm_ior *, int head, bool enable, u8 max_ac_packet, u8 rekey, u8 *avi, u8 avi_size, u8 *vendor, u8 vendor_size); + void (*scdc)(struct nvkm_ior *, int head, u8 scdc); } hdmi; struct { @@ -144,6 +150,8 @@ void gf119_hdmi_ctrl(struct nvkm_ior *, int, bool, u8, u8, u8 *, u8 , u8 *, u8); void gk104_hdmi_ctrl(struct nvkm_ior *, int, bool, u8, u8, u8 *, u8 , u8 *, u8); void gv100_hdmi_ctrl(struct nvkm_ior *, int, bool, u8, u8, u8 *, u8 , u8 *, u8); +void gm200_hdmi_scdc(struct nvkm_ior *, int, u8); + void gt215_hda_hpd(struct nvkm_ior *, int, bool); void gt215_hda_eld(struct nvkm_ior *, u8 *, u8); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.c index 3aa5a2879239..5f758948d6e1 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.c @@ -176,9 +176,10 @@ nv50_disp_root_mthd_(struct nvkm_object *object, u32 mthd, void *data, u32 size) nvif_ioctl(object, "disp sor hdmi ctrl size %d\n", size); if (!(ret = nvif_unpack(ret, &data, &size, args->v0, 0, 0, true))) { nvif_ioctl(object, "disp sor hdmi ctrl vers %d state %d " - "max_ac_packet %d rekey %d\n", + "max_ac_packet %d rekey %d scdc %d\n", args->v0.version, args->v0.state, - args->v0.max_ac_packet, args->v0.rekey); + args->v0.max_ac_packet, args->v0.rekey, + args->v0.scdc); if (args->v0.max_ac_packet > 0x1f || args->v0.rekey > 0x7f) return -EINVAL; if ((args->v0.avi_infoframe_length @@ -202,6 +203,11 @@ nv50_disp_root_mthd_(struct nvkm_object *object, u32 mthd, void *data, u32 size) args->v0.max_ac_packet, args->v0.rekey, avi, avi_size, vendor, vendor_size); + + if (outp->ior->func->hdmi.scdc) + outp->ior->func->hdmi.scdc( + outp->ior, hidx, args->v0.scdc); + return 0; } break; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c index e6e6dfbb1283..456a5a143522 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c @@ -120,13 +120,16 @@ void gf119_sor_clock(struct nvkm_ior *sor) { struct nvkm_device *device = sor->disp->engine.subdev.device; - const int div = sor->asy.link == 3; const u32 soff = nv50_ior_base(sor); + u32 div1 = sor->asy.link == 3; + u32 div2 = sor->asy.link == 3; if (sor->asy.proto == TMDS) { - /* NFI why, but this sets DP_LINK_BW_2_7 when using TMDS. */ - nvkm_mask(device, 0x612300 + soff, 0x007c0000, 0x0a << 18); + const u32 speed = sor->tmds.high_speed ? 0x14 : 0x0a; + nvkm_mask(device, 0x612300 + soff, 0x007c0000, speed << 18); + if (sor->tmds.high_speed) + div2 = 1; } - nvkm_mask(device, 0x612300 + soff, 0x00000707, (div << 8) | div); + nvkm_mask(device, 0x612300 + soff, 0x00000707, (div2 << 8) | div1); } void diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c index d892bdf04034..384f82652bec 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c @@ -99,6 +99,7 @@ gm200_sor = { .clock = gf119_sor_clock, .hdmi = { .ctrl = gk104_hdmi_ctrl, + .scdc = gm200_hdmi_scdc, }, .dp = { .lanes = { 0, 1, 2, 3 }, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgv100.c index 040db8a338de..8ba881a729ee 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgv100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgv100.c @@ -88,6 +88,7 @@ gv100_sor = { .clock = gf119_sor_clock, .hdmi = { .ctrl = gv100_hdmi_ctrl, + .scdc = gm200_hdmi_scdc, }, .dp = { .lanes = { 0, 1, 2, 3 }, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/acr_r352.c b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/acr_r352.c index d02e183717dc..5c14d6ac855d 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/acr_r352.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/acr_r352.c @@ -801,6 +801,7 @@ acr_r352_load(struct nvkm_acr *_acr, struct nvkm_falcon *falcon, bl = acr->hsbl_unload_blob; } else { nvkm_error(_acr->subdev, "invalid secure boot blob!\n"); + kfree(bl_desc); return -EINVAL; } diff --git a/drivers/gpu/drm/omapdrm/dss/dispc.c b/drivers/gpu/drm/omapdrm/dss/dispc.c index e61a9592a650..ba82d916719c 100644 --- a/drivers/gpu/drm/omapdrm/dss/dispc.c +++ b/drivers/gpu/drm/omapdrm/dss/dispc.c @@ -1140,18 +1140,6 @@ static void dispc_ovl_set_color_mode(struct dispc_device *dispc, REG_FLD_MOD(dispc, DISPC_OVL_ATTRIBUTES(plane), m, 4, 1); } -static bool format_is_yuv(u32 fourcc) -{ - switch (fourcc) { - case DRM_FORMAT_YUYV: - case DRM_FORMAT_UYVY: - case DRM_FORMAT_NV12: - return true; - default: - return false; - } -} - static void dispc_ovl_configure_burst_type(struct dispc_device *dispc, enum omap_plane_id plane, enum omap_dss_rotation_type rotation) @@ -1910,11 +1898,14 @@ static void dispc_ovl_set_scaling_uv(struct dispc_device *dispc, int scale_x = out_width != orig_width; int scale_y = out_height != orig_height; bool chroma_upscale = plane != OMAP_DSS_WB; + const struct drm_format_info *info; + + info = drm_format_info(fourcc); if (!dispc_has_feature(dispc, FEAT_HANDLE_UV_SEPARATE)) return; - if (!format_is_yuv(fourcc)) { + if (!info->is_yuv) { /* reset chroma resampling for RGB formats */ if (plane != OMAP_DSS_WB) REG_FLD_MOD(dispc, DISPC_OVL_ATTRIBUTES2(plane), @@ -2624,7 +2615,7 @@ static int dispc_ovl_setup_common(struct dispc_device *dispc, unsigned int offset0, offset1; s32 row_inc; s32 pix_inc; - u16 frame_width, frame_height; + u16 frame_width; unsigned int field_offset = 0; u16 in_height = height; u16 in_width = width; @@ -2632,6 +2623,9 @@ static int dispc_ovl_setup_common(struct dispc_device *dispc, bool ilace = !!(vm->flags & DISPLAY_FLAGS_INTERLACED); unsigned long pclk = dispc_plane_pclk_rate(dispc, plane); unsigned long lclk = dispc_plane_lclk_rate(dispc, plane); + const struct drm_format_info *info; + + info = drm_format_info(fourcc); /* when setting up WB, dispc_plane_pclk_rate() returns 0 */ if (plane == OMAP_DSS_WB) @@ -2640,7 +2634,7 @@ static int dispc_ovl_setup_common(struct dispc_device *dispc, if (paddr == 0 && rotation_type != OMAP_DSS_ROT_TILER) return -EINVAL; - if (format_is_yuv(fourcc) && (in_width & 1)) { + if (info->is_yuv && (in_width & 1)) { DSSERR("input width %d is not even for YUV format\n", in_width); return -EINVAL; } @@ -2680,7 +2674,7 @@ static int dispc_ovl_setup_common(struct dispc_device *dispc, DSSDBG("predecimation %d x %x, new input size %d x %d\n", x_predecim, y_predecim, in_width, in_height); - if (format_is_yuv(fourcc) && (in_width & 1)) { + if (info->is_yuv && (in_width & 1)) { DSSDBG("predecimated input width is not even for YUV format\n"); DSSDBG("adjusting input width %d -> %d\n", in_width, in_width & ~1); @@ -2688,7 +2682,7 @@ static int dispc_ovl_setup_common(struct dispc_device *dispc, in_width &= ~1; } - if (format_is_yuv(fourcc)) + if (info->is_yuv) cconv = 1; if (ilace && !fieldmode) { @@ -2714,13 +2708,10 @@ static int dispc_ovl_setup_common(struct dispc_device *dispc, row_inc = 0; pix_inc = 0; - if (plane == OMAP_DSS_WB) { + if (plane == OMAP_DSS_WB) frame_width = out_width; - frame_height = out_height; - } else { + else frame_width = in_width; - frame_height = height; - } calc_offset(screen_width, frame_width, fourcc, fieldmode, field_offset, diff --git a/drivers/gpu/drm/omapdrm/dss/dss.c b/drivers/gpu/drm/omapdrm/dss/dss.c index 19fc4dfc429e..1aaf260aa9b8 100644 --- a/drivers/gpu/drm/omapdrm/dss/dss.c +++ b/drivers/gpu/drm/omapdrm/dss/dss.c @@ -947,7 +947,7 @@ dss_debugfs_create_file(struct dss_device *dss, const char *name, &dss_debug_fops); if (IS_ERR(d)) { kfree(entry); - return ERR_PTR(PTR_ERR(d)); + return ERR_CAST(d); } entry->dentry = d; diff --git a/drivers/gpu/drm/omapdrm/omap_connector.c b/drivers/gpu/drm/omapdrm/omap_connector.c index 98f5ca29444a..b81302c4bf9e 100644 --- a/drivers/gpu/drm/omapdrm/omap_connector.c +++ b/drivers/gpu/drm/omapdrm/omap_connector.c @@ -164,10 +164,11 @@ static void omap_connector_destroy(struct drm_connector *connector) drm_connector_unregister(connector); drm_connector_cleanup(connector); - kfree(omap_connector); omapdss_device_put(omap_connector->output); omapdss_device_put(omap_connector->display); + + kfree(omap_connector); } #define MAX_EDID 512 diff --git a/drivers/gpu/drm/omapdrm/omap_dmm_priv.h b/drivers/gpu/drm/omapdrm/omap_dmm_priv.h index c2785cc98dc9..60bb3f9297bc 100644 --- a/drivers/gpu/drm/omapdrm/omap_dmm_priv.h +++ b/drivers/gpu/drm/omapdrm/omap_dmm_priv.h @@ -159,6 +159,7 @@ struct dmm_platform_data { struct dmm { struct device *dev; + dma_addr_t phys_base; void __iomem *base; int irq; @@ -189,6 +190,12 @@ struct dmm { struct list_head alloc_head; const struct dmm_platform_data *plat_data; + + bool dmm_workaround; + spinlock_t wa_lock; + u32 *wa_dma_data; + dma_addr_t wa_dma_handle; + struct dma_chan *wa_dma_chan; }; #endif diff --git a/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c b/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c index f92fe205550b..252f5ebb1acc 100644 --- a/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c +++ b/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c @@ -18,6 +18,7 @@ #include <linux/completion.h> #include <linux/delay.h> #include <linux/dma-mapping.h> +#include <linux/dmaengine.h> #include <linux/errno.h> #include <linux/init.h> #include <linux/interrupt.h> @@ -79,14 +80,138 @@ static const u32 reg[][4] = { DMM_PAT_DESCR__2, DMM_PAT_DESCR__3}, }; +static int dmm_dma_copy(struct dmm *dmm, dma_addr_t src, dma_addr_t dst) +{ + struct dma_device *dma_dev = dmm->wa_dma_chan->device; + struct dma_async_tx_descriptor *tx; + enum dma_status status; + dma_cookie_t cookie; + + tx = dma_dev->device_prep_dma_memcpy(dmm->wa_dma_chan, dst, src, 4, 0); + if (!tx) { + dev_err(dmm->dev, "Failed to prepare DMA memcpy\n"); + return -EIO; + } + + cookie = tx->tx_submit(tx); + if (dma_submit_error(cookie)) { + dev_err(dmm->dev, "Failed to do DMA tx_submit\n"); + return -EIO; + } + + dma_async_issue_pending(dmm->wa_dma_chan); + status = dma_sync_wait(dmm->wa_dma_chan, cookie); + if (status != DMA_COMPLETE) + dev_err(dmm->dev, "i878 wa DMA copy failure\n"); + + dmaengine_terminate_all(dmm->wa_dma_chan); + return 0; +} + +static u32 dmm_read_wa(struct dmm *dmm, u32 reg) +{ + dma_addr_t src, dst; + int r; + + src = dmm->phys_base + reg; + dst = dmm->wa_dma_handle; + + r = dmm_dma_copy(dmm, src, dst); + if (r) { + dev_err(dmm->dev, "sDMA read transfer timeout\n"); + return readl(dmm->base + reg); + } + + /* + * As per i878 workaround, the DMA is used to access the DMM registers. + * Make sure that the readl is not moved by the compiler or the CPU + * earlier than the DMA finished writing the value to memory. + */ + rmb(); + return readl(dmm->wa_dma_data); +} + +static void dmm_write_wa(struct dmm *dmm, u32 val, u32 reg) +{ + dma_addr_t src, dst; + int r; + + writel(val, dmm->wa_dma_data); + /* + * As per i878 workaround, the DMA is used to access the DMM registers. + * Make sure that the writel is not moved by the compiler or the CPU, so + * the data will be in place before we start the DMA to do the actual + * register write. + */ + wmb(); + + src = dmm->wa_dma_handle; + dst = dmm->phys_base + reg; + + r = dmm_dma_copy(dmm, src, dst); + if (r) { + dev_err(dmm->dev, "sDMA write transfer timeout\n"); + writel(val, dmm->base + reg); + } +} + static u32 dmm_read(struct dmm *dmm, u32 reg) { - return readl(dmm->base + reg); + if (dmm->dmm_workaround) { + u32 v; + unsigned long flags; + + spin_lock_irqsave(&dmm->wa_lock, flags); + v = dmm_read_wa(dmm, reg); + spin_unlock_irqrestore(&dmm->wa_lock, flags); + + return v; + } else { + return readl(dmm->base + reg); + } } static void dmm_write(struct dmm *dmm, u32 val, u32 reg) { - writel(val, dmm->base + reg); + if (dmm->dmm_workaround) { + unsigned long flags; + + spin_lock_irqsave(&dmm->wa_lock, flags); + dmm_write_wa(dmm, val, reg); + spin_unlock_irqrestore(&dmm->wa_lock, flags); + } else { + writel(val, dmm->base + reg); + } +} + +static int dmm_workaround_init(struct dmm *dmm) +{ + dma_cap_mask_t mask; + + spin_lock_init(&dmm->wa_lock); + + dmm->wa_dma_data = dma_alloc_coherent(dmm->dev, sizeof(u32), + &dmm->wa_dma_handle, GFP_KERNEL); + if (!dmm->wa_dma_data) + return -ENOMEM; + + dma_cap_zero(mask); + dma_cap_set(DMA_MEMCPY, mask); + + dmm->wa_dma_chan = dma_request_channel(mask, NULL, NULL); + if (!dmm->wa_dma_chan) { + dma_free_coherent(dmm->dev, 4, dmm->wa_dma_data, dmm->wa_dma_handle); + return -ENODEV; + } + + return 0; +} + +static void dmm_workaround_uninit(struct dmm *dmm) +{ + dma_release_channel(dmm->wa_dma_chan); + + dma_free_coherent(dmm->dev, 4, dmm->wa_dma_data, dmm->wa_dma_handle); } /* simple allocator to grab next 16 byte aligned memory from txn */ @@ -285,6 +410,17 @@ static int dmm_txn_commit(struct dmm_txn *txn, bool wait) } txn->last_pat->next_pa = 0; + /* ensure that the written descriptors are visible to DMM */ + wmb(); + + /* + * NOTE: the wmb() above should be enough, but there seems to be a bug + * in OMAP's memory barrier implementation, which in some rare cases may + * cause the writes not to be observable after wmb(). + */ + + /* read back to ensure the data is in RAM */ + readl(&txn->last_pat->next_pa); /* write to PAT_DESCR to clear out any pending transaction */ dmm_write(dmm, 0x0, reg[PAT_DESCR][engine->id]); @@ -603,6 +739,10 @@ static int omap_dmm_remove(struct platform_device *dev) unsigned long flags; if (omap_dmm) { + /* Disable all enabled interrupts */ + dmm_write(omap_dmm, 0x7e7e7e7e, DMM_PAT_IRQENABLE_CLR); + free_irq(omap_dmm->irq, omap_dmm); + /* free all area regions */ spin_lock_irqsave(&list_lock, flags); list_for_each_entry_safe(block, _block, &omap_dmm->alloc_head, @@ -625,8 +765,8 @@ static int omap_dmm_remove(struct platform_device *dev) if (omap_dmm->dummy_page) __free_page(omap_dmm->dummy_page); - if (omap_dmm->irq > 0) - free_irq(omap_dmm->irq, omap_dmm); + if (omap_dmm->dmm_workaround) + dmm_workaround_uninit(omap_dmm); iounmap(omap_dmm->base); kfree(omap_dmm); @@ -673,6 +813,7 @@ static int omap_dmm_probe(struct platform_device *dev) goto fail; } + omap_dmm->phys_base = mem->start; omap_dmm->base = ioremap(mem->start, SZ_2K); if (!omap_dmm->base) { @@ -688,6 +829,22 @@ static int omap_dmm_probe(struct platform_device *dev) omap_dmm->dev = &dev->dev; + if (of_machine_is_compatible("ti,dra7")) { + /* + * DRA7 Errata i878 says that MPU should not be used to access + * RAM and DMM at the same time. As it's not possible to prevent + * MPU accessing RAM, we need to access DMM via a proxy. + */ + if (!dmm_workaround_init(omap_dmm)) { + omap_dmm->dmm_workaround = true; + dev_info(&dev->dev, + "workaround for errata i878 in use\n"); + } else { + dev_warn(&dev->dev, + "failed to initialize work-around for i878\n"); + } + } + hwinfo = dmm_read(omap_dmm, DMM_PAT_HWINFO); omap_dmm->num_engines = (hwinfo >> 24) & 0x1F; omap_dmm->num_lut = (hwinfo >> 16) & 0x1F; @@ -714,24 +871,6 @@ static int omap_dmm_probe(struct platform_device *dev) dmm_write(omap_dmm, 0x88888888, DMM_TILER_OR__0); dmm_write(omap_dmm, 0x88888888, DMM_TILER_OR__1); - ret = request_irq(omap_dmm->irq, omap_dmm_irq_handler, IRQF_SHARED, - "omap_dmm_irq_handler", omap_dmm); - - if (ret) { - dev_err(&dev->dev, "couldn't register IRQ %d, error %d\n", - omap_dmm->irq, ret); - omap_dmm->irq = -1; - goto fail; - } - - /* Enable all interrupts for each refill engine except - * ERR_LUT_MISS<n> (which is just advisory, and we don't care - * about because we want to be able to refill live scanout - * buffers for accelerated pan/scroll) and FILL_DSC<n> which - * we just generally don't care about. - */ - dmm_write(omap_dmm, 0x7e7e7e7e, DMM_PAT_IRQENABLE_SET); - omap_dmm->dummy_page = alloc_page(GFP_KERNEL | __GFP_DMA32); if (!omap_dmm->dummy_page) { dev_err(&dev->dev, "could not allocate dummy page\n"); @@ -823,6 +962,24 @@ static int omap_dmm_probe(struct platform_device *dev) .p1.y = omap_dmm->container_height - 1, }; + ret = request_irq(omap_dmm->irq, omap_dmm_irq_handler, IRQF_SHARED, + "omap_dmm_irq_handler", omap_dmm); + + if (ret) { + dev_err(&dev->dev, "couldn't register IRQ %d, error %d\n", + omap_dmm->irq, ret); + omap_dmm->irq = -1; + goto fail; + } + + /* Enable all interrupts for each refill engine except + * ERR_LUT_MISS<n> (which is just advisory, and we don't care + * about because we want to be able to refill live scanout + * buffers for accelerated pan/scroll) and FILL_DSC<n> which + * we just generally don't care about. + */ + dmm_write(omap_dmm, 0x7e7e7e7e, DMM_PAT_IRQENABLE_SET); + /* initialize all LUTs to dummy page entries */ for (i = 0; i < omap_dmm->num_lut; i++) { area.tcm = omap_dmm->tcm[i]; diff --git a/drivers/gpu/drm/omapdrm/omap_drv.c b/drivers/gpu/drm/omapdrm/omap_drv.c index 5f98506ac2c5..5e67d58cbc28 100644 --- a/drivers/gpu/drm/omapdrm/omap_drv.c +++ b/drivers/gpu/drm/omapdrm/omap_drv.c @@ -439,7 +439,7 @@ static int ioctl_gem_info(struct drm_device *dev, void *data, args->size = omap_gem_mmap_size(obj); args->offset = omap_gem_mmap_offset(obj); - drm_gem_object_unreference_unlocked(obj); + drm_gem_object_put_unlocked(obj); return ret; } @@ -614,7 +614,7 @@ err_gem_deinit: omap_disconnect_pipelines(ddev); err_crtc_uninit: omap_crtc_pre_uninit(priv); - drm_dev_unref(ddev); + drm_dev_put(ddev); return ret; } @@ -643,7 +643,7 @@ static void omapdrm_cleanup(struct omap_drm_private *priv) omap_disconnect_pipelines(ddev); omap_crtc_pre_uninit(priv); - drm_dev_unref(ddev); + drm_dev_put(ddev); } static int pdev_probe(struct platform_device *pdev) diff --git a/drivers/gpu/drm/omapdrm/omap_fb.c b/drivers/gpu/drm/omapdrm/omap_fb.c index 9f1e3d8f8488..4d264fd554d8 100644 --- a/drivers/gpu/drm/omapdrm/omap_fb.c +++ b/drivers/gpu/drm/omapdrm/omap_fb.c @@ -319,7 +319,7 @@ struct drm_framebuffer *omap_framebuffer_create(struct drm_device *dev, error: while (--i >= 0) - drm_gem_object_unreference_unlocked(bos[i]); + drm_gem_object_put_unlocked(bos[i]); return fb; } diff --git a/drivers/gpu/drm/omapdrm/omap_fbdev.c b/drivers/gpu/drm/omapdrm/omap_fbdev.c index b445309b0143..aee99194499f 100644 --- a/drivers/gpu/drm/omapdrm/omap_fbdev.c +++ b/drivers/gpu/drm/omapdrm/omap_fbdev.c @@ -150,7 +150,7 @@ static int omap_fbdev_create(struct drm_fb_helper *helper, /* note: if fb creation failed, we can't rely on fb destroy * to unref the bo: */ - drm_gem_object_unreference_unlocked(fbdev->bo); + drm_gem_object_put_unlocked(fbdev->bo); ret = PTR_ERR(fb); goto fail; } diff --git a/drivers/gpu/drm/omapdrm/omap_gem.c b/drivers/gpu/drm/omapdrm/omap_gem.c index 4ba5d035c590..8dcaf9f4aa75 100644 --- a/drivers/gpu/drm/omapdrm/omap_gem.c +++ b/drivers/gpu/drm/omapdrm/omap_gem.c @@ -638,7 +638,7 @@ int omap_gem_dumb_map_offset(struct drm_file *file, struct drm_device *dev, *offset = omap_gem_mmap_offset(obj); - drm_gem_object_unreference_unlocked(obj); + drm_gem_object_put_unlocked(obj); fail: return ret; @@ -1312,7 +1312,7 @@ int omap_gem_new_handle(struct drm_device *dev, struct drm_file *file, } /* drop reference from allocate - handle holds it now */ - drm_gem_object_unreference_unlocked(obj); + drm_gem_object_put_unlocked(obj); return 0; } diff --git a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c index ec04a69ade46..0f8b597ccd10 100644 --- a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c +++ b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c @@ -168,7 +168,7 @@ struct drm_gem_object *omap_gem_prime_import(struct drm_device *dev, * Importing dmabuf exported from out own gem increases * refcount on gem itself instead of f_count of dmabuf. */ - drm_gem_object_reference(obj); + drm_gem_object_get(obj); return obj; } } diff --git a/drivers/gpu/drm/omapdrm/tcm-sita.h b/drivers/gpu/drm/omapdrm/tcm-sita.h deleted file mode 100644 index 460e63dbf825..000000000000 --- a/drivers/gpu/drm/omapdrm/tcm-sita.h +++ /dev/null @@ -1,93 +0,0 @@ -/* - * SImple Tiler Allocator (SiTA) private structures. - * - * Copyright (C) 2009-2011 Texas Instruments Incorporated - http://www.ti.com/ - * Author: Ravi Ramachandra <r.ramachandra@ti.com> - * - * All rights reserved. - * - * Redistribution and use in source and binary forms, with or without - * modification, are permitted provided that the following conditions - * are met: - * - * * Redistributions of source code must retain the above copyright - * notice, this list of conditions and the following disclaimer. - * - * * Redistributions in binary form must reproduce the above copyright - * notice, this list of conditions and the following disclaimer in the - * documentation and/or other materials provided with the distribution. - * - * * Neither the name of Texas Instruments Incorporated nor the names of - * its contributors may be used to endorse or promote products derived - * from this software without specific prior written permission. - * - * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" - * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, - * THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR - * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR - * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, - * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, - * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; - * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, - * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR - * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, - * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. - */ - -#ifndef _TCM_SITA_H -#define _TCM_SITA_H - -#include "tcm.h" - -/* length between two coordinates */ -#define LEN(a, b) ((a) > (b) ? (a) - (b) + 1 : (b) - (a) + 1) - -enum criteria { - CR_MAX_NEIGHS = 0x01, - CR_FIRST_FOUND = 0x10, - CR_BIAS_HORIZONTAL = 0x20, - CR_BIAS_VERTICAL = 0x40, - CR_DIAGONAL_BALANCE = 0x80 -}; - -/* nearness to the beginning of the search field from 0 to 1000 */ -struct nearness_factor { - s32 x; - s32 y; -}; - -/* - * Statistics on immediately neighboring slots. Edge is the number of - * border segments that are also border segments of the scan field. Busy - * refers to the number of neighbors that are occupied. - */ -struct neighbor_stats { - u16 edge; - u16 busy; -}; - -/* structure to keep the score of a potential allocation */ -struct score { - struct nearness_factor f; - struct neighbor_stats n; - struct tcm_area a; - u16 neighs; /* number of busy neighbors */ -}; - -struct sita_pvt { - spinlock_t lock; /* spinlock to protect access */ - struct tcm_pt div_pt; /* divider point splitting container */ - struct tcm_area ***map; /* pointers to the parent area for each slot */ -}; - -/* assign coordinates to area */ -static inline -void assign(struct tcm_area *a, u16 x0, u16 y0, u16 x1, u16 y1) -{ - a->p0.x = x0; - a->p0.y = y0; - a->p1.x = x1; - a->p1.y = y1; -} - -#endif diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c index 54324330b91f..f471537c852f 100644 --- a/drivers/gpu/drm/radeon/evergreen_cs.c +++ b/drivers/gpu/drm/radeon/evergreen_cs.c @@ -2416,7 +2416,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, size = radeon_get_ib_value(p, idx+1+(i*8)+1); if (p->rdev && (size + offset) > radeon_bo_size(reloc->robj)) { /* force size to size of the buffer */ - dev_warn(p->dev, "vbo resource seems too big for the bo\n"); + dev_warn_ratelimited(p->dev, "vbo resource seems too big for the bo\n"); ib[idx+1+(i*8)+1] = radeon_bo_size(reloc->robj) - offset; } diff --git a/drivers/gpu/drm/radeon/mkregtable.c b/drivers/gpu/drm/radeon/mkregtable.c index ba704633b072..52a7246fed9e 100644 --- a/drivers/gpu/drm/radeon/mkregtable.c +++ b/drivers/gpu/drm/radeon/mkregtable.c @@ -1,4 +1,4 @@ -// SPDX-License-Identifier: GPL-2.0 +// SPDX-License-Identifier: MIT /* utility to create the register check tables * this includes inlined list.h safe for userspace. * diff --git a/drivers/gpu/drm/radeon/r100_track.h b/drivers/gpu/drm/radeon/r100_track.h index ad16a925f8d5..57e2b09784be 100644 --- a/drivers/gpu/drm/radeon/r100_track.h +++ b/drivers/gpu/drm/radeon/r100_track.h @@ -1,4 +1,4 @@ -/* SPDX-License-Identifier: GPL-2.0 */ +/* SPDX-License-Identifier: MIT */ #define R100_TRACK_MAX_TEXTURE 3 #define R200_TRACK_MAX_TEXTURE 6 diff --git a/drivers/gpu/drm/radeon/radeon_dp_mst.c b/drivers/gpu/drm/radeon/radeon_dp_mst.c index f920be236cc9..84b3ad2172a3 100644 --- a/drivers/gpu/drm/radeon/radeon_dp_mst.c +++ b/drivers/gpu/drm/radeon/radeon_dp_mst.c @@ -1,4 +1,4 @@ -// SPDX-License-Identifier: GPL-2.0 +// SPDX-License-Identifier: MIT #include <drm/drmP.h> #include <drm/drm_dp_mst_helper.h> diff --git a/drivers/gpu/drm/radeon/radeon_legacy_tv.c b/drivers/gpu/drm/radeon/radeon_legacy_tv.c index 611cf934b211..4278272e3191 100644 --- a/drivers/gpu/drm/radeon/radeon_legacy_tv.c +++ b/drivers/gpu/drm/radeon/radeon_legacy_tv.c @@ -1,4 +1,4 @@ -// SPDX-License-Identifier: GPL-2.0 +// SPDX-License-Identifier: MIT #include <drm/drmP.h> #include <drm/drm_crtc_helper.h> #include "radeon.h" diff --git a/drivers/gpu/drm/radeon/radeon_trace.h b/drivers/gpu/drm/radeon/radeon_trace.h index bc26efd1793e..0d84b8aafab3 100644 --- a/drivers/gpu/drm/radeon/radeon_trace.h +++ b/drivers/gpu/drm/radeon/radeon_trace.h @@ -1,4 +1,4 @@ -/* SPDX-License-Identifier: GPL-2.0 */ +/* SPDX-License-Identifier: MIT */ #if !defined(_RADEON_TRACE_H) || defined(TRACE_HEADER_MULTI_READ) #define _RADEON_TRACE_H_ diff --git a/drivers/gpu/drm/radeon/radeon_trace_points.c b/drivers/gpu/drm/radeon/radeon_trace_points.c index 66b3d5084662..65e92302f974 100644 --- a/drivers/gpu/drm/radeon/radeon_trace_points.c +++ b/drivers/gpu/drm/radeon/radeon_trace_points.c @@ -1,4 +1,4 @@ -// SPDX-License-Identifier: GPL-2.0 +// SPDX-License-Identifier: MIT /* Copyright Red Hat Inc 2010. * Author : Dave Airlie <airlied@redhat.com> */ diff --git a/drivers/gpu/drm/scheduler/sched_entity.c b/drivers/gpu/drm/scheduler/sched_entity.c index 4e5e95c0cab5..3e22a54a99c2 100644 --- a/drivers/gpu/drm/scheduler/sched_entity.c +++ b/drivers/gpu/drm/scheduler/sched_entity.c @@ -467,8 +467,7 @@ void drm_sched_entity_select_rq(struct drm_sched_entity *entity) struct dma_fence *fence; struct drm_sched_rq *rq; - if (!spsc_queue_count(&entity->job_queue) == 0 || - entity->num_rq_list <= 1) + if (spsc_queue_count(&entity->job_queue) || entity->num_rq_list <= 1) return; fence = READ_ONCE(entity->last_scheduled); diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c index 9ca741f3a0bc..44fe587aaef9 100644 --- a/drivers/gpu/drm/scheduler/sched_main.c +++ b/drivers/gpu/drm/scheduler/sched_main.c @@ -182,6 +182,20 @@ bool drm_sched_dependency_optimized(struct dma_fence* fence, } EXPORT_SYMBOL(drm_sched_dependency_optimized); +/** + * drm_sched_start_timeout - start timeout for reset worker + * + * @sched: scheduler instance to start the worker for + * + * Start the timeout for the given scheduler. + */ +static void drm_sched_start_timeout(struct drm_gpu_scheduler *sched) +{ + if (sched->timeout != MAX_SCHEDULE_TIMEOUT && + !list_empty(&sched->ring_mirror_list)) + schedule_delayed_work(&sched->work_tdr, sched->timeout); +} + /* job_finish is called after hw fence signaled */ static void drm_sched_job_finish(struct work_struct *work) @@ -197,19 +211,13 @@ static void drm_sched_job_finish(struct work_struct *work) * manages to find this job as the next job in the list, the fence * signaled check below will prevent the timeout to be restarted. */ - cancel_delayed_work_sync(&s_job->work_tdr); + cancel_delayed_work_sync(&sched->work_tdr); spin_lock(&sched->job_list_lock); - /* queue TDR for next job */ - if (sched->timeout != MAX_SCHEDULE_TIMEOUT && - !list_is_last(&s_job->node, &sched->ring_mirror_list)) { - struct drm_sched_job *next = list_next_entry(s_job, node); - - if (!dma_fence_is_signaled(&next->s_fence->finished)) - schedule_delayed_work(&next->work_tdr, sched->timeout); - } /* remove job from ring_mirror_list */ list_del(&s_job->node); + /* queue TDR for next job */ + drm_sched_start_timeout(sched); spin_unlock(&sched->job_list_lock); dma_fence_put(&s_job->s_fence->finished); @@ -233,19 +241,49 @@ static void drm_sched_job_begin(struct drm_sched_job *s_job) spin_lock(&sched->job_list_lock); list_add_tail(&s_job->node, &sched->ring_mirror_list); - if (sched->timeout != MAX_SCHEDULE_TIMEOUT && - list_first_entry_or_null(&sched->ring_mirror_list, - struct drm_sched_job, node) == s_job) - schedule_delayed_work(&s_job->work_tdr, sched->timeout); + drm_sched_start_timeout(sched); spin_unlock(&sched->job_list_lock); } static void drm_sched_job_timedout(struct work_struct *work) { - struct drm_sched_job *job = container_of(work, struct drm_sched_job, - work_tdr.work); + struct drm_gpu_scheduler *sched; + struct drm_sched_job *job; + int r; - job->sched->ops->timedout_job(job); + sched = container_of(work, struct drm_gpu_scheduler, work_tdr.work); + + spin_lock(&sched->job_list_lock); + list_for_each_entry_reverse(job, &sched->ring_mirror_list, node) { + struct drm_sched_fence *fence = job->s_fence; + + if (!dma_fence_remove_callback(fence->parent, &fence->cb)) + goto already_signaled; + } + + job = list_first_entry_or_null(&sched->ring_mirror_list, + struct drm_sched_job, node); + spin_unlock(&sched->job_list_lock); + + if (job) + sched->ops->timedout_job(job); + + spin_lock(&sched->job_list_lock); + list_for_each_entry(job, &sched->ring_mirror_list, node) { + struct drm_sched_fence *fence = job->s_fence; + + if (!fence->parent || !list_empty(&fence->cb.node)) + continue; + + r = dma_fence_add_callback(fence->parent, &fence->cb, + drm_sched_process_job); + if (r) + drm_sched_process_job(fence->parent, &fence->cb); + +already_signaled: + ; + } + spin_unlock(&sched->job_list_lock); } /** @@ -312,11 +350,6 @@ void drm_sched_job_recovery(struct drm_gpu_scheduler *sched) int r; spin_lock(&sched->job_list_lock); - s_job = list_first_entry_or_null(&sched->ring_mirror_list, - struct drm_sched_job, node); - if (s_job && sched->timeout != MAX_SCHEDULE_TIMEOUT) - schedule_delayed_work(&s_job->work_tdr, sched->timeout); - list_for_each_entry_safe(s_job, tmp, &sched->ring_mirror_list, node) { struct drm_sched_fence *s_fence = s_job->s_fence; struct dma_fence *fence; @@ -349,6 +382,7 @@ void drm_sched_job_recovery(struct drm_gpu_scheduler *sched) } spin_lock(&sched->job_list_lock); } + drm_sched_start_timeout(sched); spin_unlock(&sched->job_list_lock); } EXPORT_SYMBOL(drm_sched_job_recovery); @@ -384,7 +418,6 @@ int drm_sched_job_init(struct drm_sched_job *job, INIT_WORK(&job->finish_work, drm_sched_job_finish); INIT_LIST_HEAD(&job->node); - INIT_DELAYED_WORK(&job->work_tdr, drm_sched_job_timedout); return 0; } @@ -575,6 +608,7 @@ int drm_sched_init(struct drm_gpu_scheduler *sched, INIT_LIST_HEAD(&sched->ring_mirror_list); spin_lock_init(&sched->job_list_lock); atomic_set(&sched->hw_rq_count, 0); + INIT_DELAYED_WORK(&sched->work_tdr, drm_sched_job_timedout); atomic_set(&sched->num_jobs, 0); atomic64_set(&sched->job_id_count, 0); diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c index 965088afcfad..f80e82e16475 100644 --- a/drivers/gpu/drm/tegra/dc.c +++ b/drivers/gpu/drm/tegra/dc.c @@ -1988,6 +1988,28 @@ static int tegra_dc_init(struct host1x_client *client) struct drm_plane *cursor = NULL; int err; + /* + * XXX do not register DCs with no window groups because we cannot + * assign a primary plane to them, which in turn will cause KMS to + * crash. + */ + if (dc->soc->wgrps) { + bool has_wgrps = false; + unsigned int i; + + for (i = 0; i < dc->soc->num_wgrps; i++) { + const struct tegra_windowgroup_soc *wgrp = &dc->soc->wgrps[i]; + + if (wgrp->dc == dc->pipe && wgrp->num_windows > 0) { + has_wgrps = true; + break; + } + } + + if (!has_wgrps) + return 0; + } + dc->syncpt = host1x_syncpt_request(client, flags); if (!dc->syncpt) dev_warn(dc->dev, "failed to allocate syncpoint\n"); @@ -2234,8 +2256,59 @@ static const struct tegra_dc_soc_info tegra186_dc_soc_info = { .num_wgrps = ARRAY_SIZE(tegra186_dc_wgrps), }; +static const struct tegra_windowgroup_soc tegra194_dc_wgrps[] = { + { + .index = 0, + .dc = 0, + .windows = (const unsigned int[]) { 0 }, + .num_windows = 1, + }, { + .index = 1, + .dc = 1, + .windows = (const unsigned int[]) { 1 }, + .num_windows = 1, + }, { + .index = 2, + .dc = 1, + .windows = (const unsigned int[]) { 2 }, + .num_windows = 1, + }, { + .index = 3, + .dc = 2, + .windows = (const unsigned int[]) { 3 }, + .num_windows = 1, + }, { + .index = 4, + .dc = 2, + .windows = (const unsigned int[]) { 4 }, + .num_windows = 1, + }, { + .index = 5, + .dc = 2, + .windows = (const unsigned int[]) { 5 }, + .num_windows = 1, + }, +}; + +static const struct tegra_dc_soc_info tegra194_dc_soc_info = { + .supports_background_color = true, + .supports_interlacing = true, + .supports_cursor = true, + .supports_block_linear = true, + .has_legacy_blending = false, + .pitch_align = 64, + .has_powergate = false, + .coupled_pm = false, + .has_nvdisplay = true, + .wgrps = tegra194_dc_wgrps, + .num_wgrps = ARRAY_SIZE(tegra194_dc_wgrps), +}; + static const struct of_device_id tegra_dc_of_match[] = { { + .compatible = "nvidia,tegra194-dc", + .data = &tegra194_dc_soc_info, + }, { .compatible = "nvidia,tegra186-dc", .data = &tegra186_dc_soc_info, }, { diff --git a/drivers/gpu/drm/tegra/dc.h b/drivers/gpu/drm/tegra/dc.h index e96f582ca692..1256dfb6b2f5 100644 --- a/drivers/gpu/drm/tegra/dc.h +++ b/drivers/gpu/drm/tegra/dc.h @@ -300,7 +300,7 @@ int tegra_dc_rgb_exit(struct tegra_dc *dc); #define SOR1_TIMING_CYA (1 << 27) #define CURSOR_ENABLE (1 << 16) -#define SOR_ENABLE(x) (1 << (25 + (x))) +#define SOR_ENABLE(x) (1 << (25 + (((x) > 1) ? ((x) + 1) : (x)))) #define DC_DISP_DISP_MEM_HIGH_PRIORITY 0x403 #define CURSOR_THRESHOLD(x) (((x) & 0x03) << 24) diff --git a/drivers/gpu/drm/tegra/dpaux.c b/drivers/gpu/drm/tegra/dpaux.c index d84e81ff36ad..ee4180d8db14 100644 --- a/drivers/gpu/drm/tegra/dpaux.c +++ b/drivers/gpu/drm/tegra/dpaux.c @@ -521,7 +521,7 @@ static int tegra_dpaux_probe(struct platform_device *pdev) * is no possibility to perform the I2C mode configuration in the * HDMI path. */ - err = tegra_dpaux_pad_config(dpaux, DPAUX_HYBRID_PADCTL_MODE_I2C); + err = tegra_dpaux_pad_config(dpaux, DPAUX_PADCTL_FUNC_I2C); if (err < 0) return err; @@ -639,6 +639,7 @@ static const struct dev_pm_ops tegra_dpaux_pm_ops = { }; static const struct of_device_id tegra_dpaux_of_match[] = { + { .compatible = "nvidia,tegra194-dpaux", }, { .compatible = "nvidia,tegra186-dpaux", }, { .compatible = "nvidia,tegra210-dpaux", }, { .compatible = "nvidia,tegra124-dpaux", }, diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 8cdb610561ba..65ea4988b332 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -15,6 +15,10 @@ #include <drm/drm_atomic.h> #include <drm/drm_atomic_helper.h> +#if IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU) +#include <asm/dma-iommu.h> +#endif + #include "drm.h" #include "gem.h" @@ -1068,6 +1072,14 @@ struct iommu_group *host1x_client_iommu_attach(struct host1x_client *client, } if (!shared || (shared && (group != tegra->group))) { +#if IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU) + if (client->dev->archdata.mapping) { + struct dma_iommu_mapping *mapping = + to_dma_iommu_mapping(client->dev); + arm_iommu_detach_device(client->dev); + arm_iommu_release_mapping(mapping); + } +#endif err = iommu_attach_group(tegra->domain, group); if (err < 0) { iommu_group_put(group); @@ -1216,31 +1228,15 @@ static int host1x_drm_remove(struct host1x_device *dev) static int host1x_drm_suspend(struct device *dev) { struct drm_device *drm = dev_get_drvdata(dev); - struct tegra_drm *tegra = drm->dev_private; - - drm_kms_helper_poll_disable(drm); - tegra_drm_fb_suspend(drm); - tegra->state = drm_atomic_helper_suspend(drm); - if (IS_ERR(tegra->state)) { - tegra_drm_fb_resume(drm); - drm_kms_helper_poll_enable(drm); - return PTR_ERR(tegra->state); - } - - return 0; + return drm_mode_config_helper_suspend(drm); } static int host1x_drm_resume(struct device *dev) { struct drm_device *drm = dev_get_drvdata(dev); - struct tegra_drm *tegra = drm->dev_private; - - drm_atomic_helper_resume(drm, tegra->state); - tegra_drm_fb_resume(drm); - drm_kms_helper_poll_enable(drm); - return 0; + return drm_mode_config_helper_resume(drm); } #endif @@ -1275,6 +1271,9 @@ static const struct of_device_id host1x_drm_subdevs[] = { { .compatible = "nvidia,tegra186-sor", }, { .compatible = "nvidia,tegra186-sor1", }, { .compatible = "nvidia,tegra186-vic", }, + { .compatible = "nvidia,tegra194-display", }, + { .compatible = "nvidia,tegra194-dc", }, + { .compatible = "nvidia,tegra194-sor", }, { /* sentinel */ } }; diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h index 92d248784396..1012335bb489 100644 --- a/drivers/gpu/drm/tegra/drm.h +++ b/drivers/gpu/drm/tegra/drm.h @@ -60,8 +60,6 @@ struct tegra_drm { unsigned int pitch_align; struct tegra_display_hub *hub; - - struct drm_atomic_state *state; }; struct tegra_drm_client; @@ -186,8 +184,6 @@ int tegra_drm_fb_prepare(struct drm_device *drm); void tegra_drm_fb_free(struct drm_device *drm); int tegra_drm_fb_init(struct drm_device *drm); void tegra_drm_fb_exit(struct drm_device *drm); -void tegra_drm_fb_suspend(struct drm_device *drm); -void tegra_drm_fb_resume(struct drm_device *drm); extern struct platform_driver tegra_display_hub_driver; extern struct platform_driver tegra_dc_driver; diff --git a/drivers/gpu/drm/tegra/fb.c b/drivers/gpu/drm/tegra/fb.c index 4c22cdded3c2..b947e82bbeb1 100644 --- a/drivers/gpu/drm/tegra/fb.c +++ b/drivers/gpu/drm/tegra/fb.c @@ -356,7 +356,7 @@ static void tegra_fbdev_exit(struct tegra_fbdev *fbdev) /* Undo the special mapping we made in fbdev probe. */ if (bo && bo->pages) { vunmap(bo->vaddr); - bo->vaddr = 0; + bo->vaddr = NULL; } drm_framebuffer_remove(fbdev->fb); @@ -412,25 +412,3 @@ void tegra_drm_fb_exit(struct drm_device *drm) tegra_fbdev_exit(tegra->fbdev); #endif } - -void tegra_drm_fb_suspend(struct drm_device *drm) -{ -#ifdef CONFIG_DRM_FBDEV_EMULATION - struct tegra_drm *tegra = drm->dev_private; - - console_lock(); - drm_fb_helper_set_suspend(&tegra->fbdev->base, 1); - console_unlock(); -#endif -} - -void tegra_drm_fb_resume(struct drm_device *drm) -{ -#ifdef CONFIG_DRM_FBDEV_EMULATION - struct tegra_drm *tegra = drm->dev_private; - - console_lock(); - drm_fb_helper_set_suspend(&tegra->fbdev->base, 0); - console_unlock(); -#endif -} diff --git a/drivers/gpu/drm/tegra/hub.c b/drivers/gpu/drm/tegra/hub.c index 8f4fcbb515fb..6112d9042979 100644 --- a/drivers/gpu/drm/tegra/hub.c +++ b/drivers/gpu/drm/tegra/hub.c @@ -758,10 +758,12 @@ static int tegra_display_hub_probe(struct platform_device *pdev) return err; } - hub->clk_dsc = devm_clk_get(&pdev->dev, "dsc"); - if (IS_ERR(hub->clk_dsc)) { - err = PTR_ERR(hub->clk_dsc); - return err; + if (hub->soc->supports_dsc) { + hub->clk_dsc = devm_clk_get(&pdev->dev, "dsc"); + if (IS_ERR(hub->clk_dsc)) { + err = PTR_ERR(hub->clk_dsc); + return err; + } } hub->clk_hub = devm_clk_get(&pdev->dev, "hub"); @@ -890,10 +892,19 @@ static const struct dev_pm_ops tegra_display_hub_pm_ops = { static const struct tegra_display_hub_soc tegra186_display_hub = { .num_wgrps = 6, + .supports_dsc = true, +}; + +static const struct tegra_display_hub_soc tegra194_display_hub = { + .num_wgrps = 6, + .supports_dsc = false, }; static const struct of_device_id tegra_display_hub_of_match[] = { { + .compatible = "nvidia,tegra194-display", + .data = &tegra194_display_hub + }, { .compatible = "nvidia,tegra186-display", .data = &tegra186_display_hub }, { diff --git a/drivers/gpu/drm/tegra/hub.h b/drivers/gpu/drm/tegra/hub.h index 85b8bf41a395..6696a85fc1f2 100644 --- a/drivers/gpu/drm/tegra/hub.h +++ b/drivers/gpu/drm/tegra/hub.h @@ -38,6 +38,7 @@ to_tegra_shared_plane(struct drm_plane *plane) struct tegra_display_hub_soc { unsigned int num_wgrps; + bool supports_dsc; }; struct tegra_display_hub { diff --git a/drivers/gpu/drm/tegra/sor.c b/drivers/gpu/drm/tegra/sor.c index d7fe9f15def1..b129da2e5afd 100644 --- a/drivers/gpu/drm/tegra/sor.c +++ b/drivers/gpu/drm/tegra/sor.c @@ -282,6 +282,85 @@ static const struct tegra_sor_hdmi_settings tegra186_sor_hdmi_defaults[] = { } }; +static const struct tegra_sor_hdmi_settings tegra194_sor_hdmi_defaults[] = { + { + .frequency = 54000000, + .vcocap = 0, + .filter = 5, + .ichpmp = 5, + .loadadj = 3, + .tmds_termadj = 0xf, + .tx_pu_value = 0, + .bg_temp_coef = 3, + .bg_vref_level = 8, + .avdd10_level = 4, + .avdd14_level = 4, + .sparepll = 0x54, + .drive_current = { 0x3a, 0x3a, 0x3a, 0x33 }, + .preemphasis = { 0x00, 0x00, 0x00, 0x00 }, + }, { + .frequency = 75000000, + .vcocap = 1, + .filter = 5, + .ichpmp = 5, + .loadadj = 3, + .tmds_termadj = 0xf, + .tx_pu_value = 0, + .bg_temp_coef = 3, + .bg_vref_level = 8, + .avdd10_level = 4, + .avdd14_level = 4, + .sparepll = 0x44, + .drive_current = { 0x3a, 0x3a, 0x3a, 0x33 }, + .preemphasis = { 0x00, 0x00, 0x00, 0x00 }, + }, { + .frequency = 150000000, + .vcocap = 3, + .filter = 5, + .ichpmp = 5, + .loadadj = 3, + .tmds_termadj = 15, + .tx_pu_value = 0x66 /* 0 */, + .bg_temp_coef = 3, + .bg_vref_level = 8, + .avdd10_level = 4, + .avdd14_level = 4, + .sparepll = 0x00, /* 0x34 */ + .drive_current = { 0x3a, 0x3a, 0x3a, 0x37 }, + .preemphasis = { 0x00, 0x00, 0x00, 0x00 }, + }, { + .frequency = 300000000, + .vcocap = 3, + .filter = 5, + .ichpmp = 5, + .loadadj = 3, + .tmds_termadj = 15, + .tx_pu_value = 64, + .bg_temp_coef = 3, + .bg_vref_level = 8, + .avdd10_level = 4, + .avdd14_level = 4, + .sparepll = 0x34, + .drive_current = { 0x3d, 0x3d, 0x3d, 0x33 }, + .preemphasis = { 0x00, 0x00, 0x00, 0x00 }, + }, { + .frequency = 600000000, + .vcocap = 3, + .filter = 5, + .ichpmp = 5, + .loadadj = 3, + .tmds_termadj = 12, + .tx_pu_value = 96, + .bg_temp_coef = 3, + .bg_vref_level = 8, + .avdd10_level = 4, + .avdd14_level = 4, + .sparepll = 0x34, + .drive_current = { 0x3d, 0x3d, 0x3d, 0x33 }, + .preemphasis = { 0x00, 0x00, 0x00, 0x00 }, + } +}; + struct tegra_sor_regs { unsigned int head_state0; unsigned int head_state1; @@ -2894,7 +2973,38 @@ static const struct tegra_sor_soc tegra186_sor1 = { .xbar_cfg = tegra124_sor_xbar_cfg, }; +static const struct tegra_sor_regs tegra194_sor_regs = { + .head_state0 = 0x151, + .head_state1 = 0x155, + .head_state2 = 0x159, + .head_state3 = 0x15d, + .head_state4 = 0x161, + .head_state5 = 0x165, + .pll0 = 0x169, + .pll1 = 0x16a, + .pll2 = 0x16b, + .pll3 = 0x16c, + .dp_padctl0 = 0x16e, + .dp_padctl2 = 0x16f, +}; + +static const struct tegra_sor_soc tegra194_sor = { + .supports_edp = true, + .supports_lvds = false, + .supports_hdmi = true, + .supports_dp = true, + + .regs = &tegra194_sor_regs, + .has_nvdisplay = true, + + .num_settings = ARRAY_SIZE(tegra194_sor_hdmi_defaults), + .settings = tegra194_sor_hdmi_defaults, + + .xbar_cfg = tegra210_sor_xbar_cfg, +}; + static const struct of_device_id tegra_sor_of_match[] = { + { .compatible = "nvidia,tegra194-sor", .data = &tegra194_sor }, { .compatible = "nvidia,tegra186-sor1", .data = &tegra186_sor1 }, { .compatible = "nvidia,tegra186-sor", .data = &tegra186_sor }, { .compatible = "nvidia,tegra210-sor1", .data = &tegra210_sor1 }, diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.c b/drivers/gpu/drm/tilcdc/tilcdc_drv.c index 0fb300d41a09..33e533268488 100644 --- a/drivers/gpu/drm/tilcdc/tilcdc_drv.c +++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.c @@ -554,29 +554,23 @@ static struct drm_driver tilcdc_driver = { static int tilcdc_pm_suspend(struct device *dev) { struct drm_device *ddev = dev_get_drvdata(dev); - struct tilcdc_drm_private *priv = ddev->dev_private; + int ret = 0; - priv->saved_state = drm_atomic_helper_suspend(ddev); + ret = drm_mode_config_helper_suspend(ddev); /* Select sleep pin state */ pinctrl_pm_select_sleep_state(dev); - return 0; + return ret; } static int tilcdc_pm_resume(struct device *dev) { struct drm_device *ddev = dev_get_drvdata(dev); - struct tilcdc_drm_private *priv = ddev->dev_private; - int ret = 0; /* Select default pin state */ pinctrl_pm_select_default_state(dev); - - if (priv->saved_state) - ret = drm_atomic_helper_resume(ddev, priv->saved_state); - - return ret; + return drm_mode_config_helper_resume(ddev); } #endif diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.h b/drivers/gpu/drm/tilcdc/tilcdc_drv.h index ead512216669..62cea5ff5558 100644 --- a/drivers/gpu/drm/tilcdc/tilcdc_drv.h +++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.h @@ -70,9 +70,6 @@ struct tilcdc_drm_private { const uint32_t *pixelformats; uint32_t num_pixelformats; - /* The context for pm susped/resume cycle is stored here */ - struct drm_atomic_state *saved_state; - #ifdef CONFIG_CPU_FREQ struct notifier_block freq_transition; #endif diff --git a/drivers/gpu/drm/ttm/Makefile b/drivers/gpu/drm/ttm/Makefile index a60e560804e0..01fc670ce7a2 100644 --- a/drivers/gpu/drm/ttm/Makefile +++ b/drivers/gpu/drm/ttm/Makefile @@ -4,8 +4,8 @@ ttm-y := ttm_memory.o ttm_tt.o ttm_bo.o \ ttm_bo_util.o ttm_bo_vm.o ttm_module.o \ - ttm_object.o ttm_lock.o ttm_execbuf_util.o ttm_page_alloc.o \ - ttm_bo_manager.o ttm_page_alloc_dma.o + ttm_execbuf_util.o ttm_page_alloc.o ttm_bo_manager.o \ + ttm_page_alloc_dma.o ttm-$(CONFIG_AGP) += ttm_agp_backend.o obj-$(CONFIG_DRM_TTM) += ttm.o diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index b2a33bf1ef10..26b889f86670 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -247,20 +247,6 @@ void ttm_bo_move_to_lru_tail(struct ttm_buffer_object *bo, } EXPORT_SYMBOL(ttm_bo_move_to_lru_tail); -static void ttm_list_move_bulk_tail(struct list_head *list, - struct list_head *first, - struct list_head *last) -{ - first->prev->next = last->next; - last->next->prev = first->prev; - - list->prev->next = first; - first->prev = list->prev; - - last->next = list; - list->prev = last; -} - void ttm_bo_bulk_move_lru_tail(struct ttm_lru_bulk_move *bulk) { unsigned i; @@ -276,8 +262,8 @@ void ttm_bo_bulk_move_lru_tail(struct ttm_lru_bulk_move *bulk) reservation_object_assert_held(pos->last->resv); man = &pos->first->bdev->man[TTM_PL_TT]; - ttm_list_move_bulk_tail(&man->lru[i], &pos->first->lru, - &pos->last->lru); + list_bulk_move_tail(&man->lru[i], &pos->first->lru, + &pos->last->lru); } for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) { @@ -291,8 +277,8 @@ void ttm_bo_bulk_move_lru_tail(struct ttm_lru_bulk_move *bulk) reservation_object_assert_held(pos->last->resv); man = &pos->first->bdev->man[TTM_PL_VRAM]; - ttm_list_move_bulk_tail(&man->lru[i], &pos->first->lru, - &pos->last->lru); + list_bulk_move_tail(&man->lru[i], &pos->first->lru, + &pos->last->lru); } for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) { @@ -306,8 +292,7 @@ void ttm_bo_bulk_move_lru_tail(struct ttm_lru_bulk_move *bulk) reservation_object_assert_held(pos->last->resv); lru = &pos->first->bdev->glob->swap_lru[i]; - ttm_list_move_bulk_tail(lru, &pos->first->swap, - &pos->last->swap); + list_bulk_move_tail(lru, &pos->first->swap, &pos->last->swap); } } EXPORT_SYMBOL(ttm_bo_bulk_move_lru_tail); diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c index 6fe91c1b692d..a1d977fbade5 100644 --- a/drivers/gpu/drm/ttm/ttm_bo_vm.c +++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c @@ -409,8 +409,7 @@ static struct ttm_buffer_object *ttm_bo_vm_lookup(struct ttm_bo_device *bdev, node = drm_vma_offset_lookup_locked(&bdev->vma_manager, offset, pages); if (likely(node)) { bo = container_of(node, struct ttm_buffer_object, vma_node); - if (!kref_get_unless_zero(&bo->kref)) - bo = NULL; + bo = ttm_bo_get_unless_zero(bo); } drm_vma_offset_unlock_lookup(&bdev->vma_manager); diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index a5501581d96b..9243dea6e6ad 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -168,7 +168,7 @@ v3d_job_timedout(struct drm_sched_job *sched_job) job->timedout_ctca = ctca; job->timedout_ctra = ctra; - schedule_delayed_work(&job->base.work_tdr, + schedule_delayed_work(&job->base.sched->work_tdr, job->base.sched->timeout); return; } diff --git a/drivers/gpu/drm/vmwgfx/Makefile b/drivers/gpu/drm/vmwgfx/Makefile index 09b2aa08363e..8841bd30e1e5 100644 --- a/drivers/gpu/drm/vmwgfx/Makefile +++ b/drivers/gpu/drm/vmwgfx/Makefile @@ -7,6 +7,8 @@ vmwgfx-y := vmwgfx_execbuf.o vmwgfx_gmr.o vmwgfx_kms.o vmwgfx_drv.o \ vmwgfx_surface.o vmwgfx_prime.o vmwgfx_mob.o vmwgfx_shader.o \ vmwgfx_cmdbuf_res.o vmwgfx_cmdbuf.o vmwgfx_stdu.o \ vmwgfx_cotable.o vmwgfx_so.o vmwgfx_binding.o vmwgfx_msg.o \ - vmwgfx_simple_resource.o vmwgfx_va.o vmwgfx_blit.o + vmwgfx_simple_resource.o vmwgfx_va.o vmwgfx_blit.o \ + vmwgfx_validation.o \ + ttm_object.o ttm_lock.o obj-$(CONFIG_DRM_VMWGFX) := vmwgfx.o diff --git a/drivers/gpu/drm/ttm/ttm_lock.c b/drivers/gpu/drm/vmwgfx/ttm_lock.c index 20694b8a01ca..16b2083cb9d4 100644 --- a/drivers/gpu/drm/ttm/ttm_lock.c +++ b/drivers/gpu/drm/vmwgfx/ttm_lock.c @@ -29,13 +29,13 @@ * Authors: Thomas Hellstrom <thellstrom-at-vmware-dot-com> */ -#include <drm/ttm/ttm_lock.h> #include <drm/ttm/ttm_module.h> #include <linux/atomic.h> #include <linux/errno.h> #include <linux/wait.h> #include <linux/sched/signal.h> -#include <linux/module.h> +#include "ttm_lock.h" +#include "ttm_object.h" #define TTM_WRITE_LOCK_PENDING (1 << 0) #define TTM_VT_LOCK_PENDING (1 << 1) @@ -52,7 +52,6 @@ void ttm_lock_init(struct ttm_lock *lock) lock->kill_takers = false; lock->signal = SIGKILL; } -EXPORT_SYMBOL(ttm_lock_init); void ttm_read_unlock(struct ttm_lock *lock) { @@ -61,7 +60,6 @@ void ttm_read_unlock(struct ttm_lock *lock) wake_up_all(&lock->queue); spin_unlock(&lock->lock); } -EXPORT_SYMBOL(ttm_read_unlock); static bool __ttm_read_lock(struct ttm_lock *lock) { @@ -92,7 +90,6 @@ int ttm_read_lock(struct ttm_lock *lock, bool interruptible) wait_event(lock->queue, __ttm_read_lock(lock)); return ret; } -EXPORT_SYMBOL(ttm_read_lock); static bool __ttm_read_trylock(struct ttm_lock *lock, bool *locked) { @@ -144,7 +141,6 @@ void ttm_write_unlock(struct ttm_lock *lock) wake_up_all(&lock->queue); spin_unlock(&lock->lock); } -EXPORT_SYMBOL(ttm_write_unlock); static bool __ttm_write_lock(struct ttm_lock *lock) { @@ -185,7 +181,6 @@ int ttm_write_lock(struct ttm_lock *lock, bool interruptible) return ret; } -EXPORT_SYMBOL(ttm_write_lock); static int __ttm_vt_unlock(struct ttm_lock *lock) { @@ -262,14 +257,12 @@ int ttm_vt_lock(struct ttm_lock *lock, return ret; } -EXPORT_SYMBOL(ttm_vt_lock); int ttm_vt_unlock(struct ttm_lock *lock) { return ttm_ref_object_base_unref(lock->vt_holder, - lock->base.hash.key, TTM_REF_USAGE); + lock->base.handle, TTM_REF_USAGE); } -EXPORT_SYMBOL(ttm_vt_unlock); void ttm_suspend_unlock(struct ttm_lock *lock) { @@ -278,7 +271,6 @@ void ttm_suspend_unlock(struct ttm_lock *lock) wake_up_all(&lock->queue); spin_unlock(&lock->lock); } -EXPORT_SYMBOL(ttm_suspend_unlock); static bool __ttm_suspend_lock(struct ttm_lock *lock) { @@ -300,4 +292,3 @@ void ttm_suspend_lock(struct ttm_lock *lock) { wait_event(lock->queue, __ttm_suspend_lock(lock)); } -EXPORT_SYMBOL(ttm_suspend_lock); diff --git a/drivers/gpu/drm/vmwgfx/ttm_lock.h b/drivers/gpu/drm/vmwgfx/ttm_lock.h new file mode 100644 index 000000000000..0c3af9836863 --- /dev/null +++ b/drivers/gpu/drm/vmwgfx/ttm_lock.h @@ -0,0 +1,248 @@ +/************************************************************************** + * + * Copyright (c) 2007-2009 VMware, Inc., Palo Alto, CA., USA + * All Rights Reserved. + * + * 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, sub license, 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 (including the + * next paragraph) 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 NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS 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: Thomas Hellstrom <thellstrom-at-vmware-dot-com> + */ + +/** @file ttm_lock.h + * This file implements a simple replacement for the buffer manager use + * of the DRM heavyweight hardware lock. + * The lock is a read-write lock. Taking it in read mode and write mode + * is relatively fast, and intended for in-kernel use only. + * + * The vt mode is used only when there is a need to block all + * user-space processes from validating buffers. + * It's allowed to leave kernel space with the vt lock held. + * If a user-space process dies while having the vt-lock, + * it will be released during the file descriptor release. The vt lock + * excludes write lock and read lock. + * + * The suspend mode is used to lock out all TTM users when preparing for + * and executing suspend operations. + * + */ + +#ifndef _TTM_LOCK_H_ +#define _TTM_LOCK_H_ + +#include <linux/wait.h> +#include <linux/atomic.h> + +#include "ttm_object.h" + +/** + * struct ttm_lock + * + * @base: ttm base object used solely to release the lock if the client + * holding the lock dies. + * @queue: Queue for processes waiting for lock change-of-status. + * @lock: Spinlock protecting some lock members. + * @rw: Read-write lock counter. Protected by @lock. + * @flags: Lock state. Protected by @lock. + * @kill_takers: Boolean whether to kill takers of the lock. + * @signal: Signal to send when kill_takers is true. + */ + +struct ttm_lock { + struct ttm_base_object base; + wait_queue_head_t queue; + spinlock_t lock; + int32_t rw; + uint32_t flags; + bool kill_takers; + int signal; + struct ttm_object_file *vt_holder; +}; + + +/** + * ttm_lock_init + * + * @lock: Pointer to a struct ttm_lock + * Initializes the lock. + */ +extern void ttm_lock_init(struct ttm_lock *lock); + +/** + * ttm_read_unlock + * + * @lock: Pointer to a struct ttm_lock + * + * Releases a read lock. + */ +extern void ttm_read_unlock(struct ttm_lock *lock); + +/** + * ttm_read_lock + * + * @lock: Pointer to a struct ttm_lock + * @interruptible: Interruptible sleeping while waiting for a lock. + * + * Takes the lock in read mode. + * Returns: + * -ERESTARTSYS If interrupted by a signal and interruptible is true. + */ +extern int ttm_read_lock(struct ttm_lock *lock, bool interruptible); + +/** + * ttm_read_trylock + * + * @lock: Pointer to a struct ttm_lock + * @interruptible: Interruptible sleeping while waiting for a lock. + * + * Tries to take the lock in read mode. If the lock is already held + * in write mode, the function will return -EBUSY. If the lock is held + * in vt or suspend mode, the function will sleep until these modes + * are unlocked. + * + * Returns: + * -EBUSY The lock was already held in write mode. + * -ERESTARTSYS If interrupted by a signal and interruptible is true. + */ +extern int ttm_read_trylock(struct ttm_lock *lock, bool interruptible); + +/** + * ttm_write_unlock + * + * @lock: Pointer to a struct ttm_lock + * + * Releases a write lock. + */ +extern void ttm_write_unlock(struct ttm_lock *lock); + +/** + * ttm_write_lock + * + * @lock: Pointer to a struct ttm_lock + * @interruptible: Interruptible sleeping while waiting for a lock. + * + * Takes the lock in write mode. + * Returns: + * -ERESTARTSYS If interrupted by a signal and interruptible is true. + */ +extern int ttm_write_lock(struct ttm_lock *lock, bool interruptible); + +/** + * ttm_lock_downgrade + * + * @lock: Pointer to a struct ttm_lock + * + * Downgrades a write lock to a read lock. + */ +extern void ttm_lock_downgrade(struct ttm_lock *lock); + +/** + * ttm_suspend_lock + * + * @lock: Pointer to a struct ttm_lock + * + * Takes the lock in suspend mode. Excludes read and write mode. + */ +extern void ttm_suspend_lock(struct ttm_lock *lock); + +/** + * ttm_suspend_unlock + * + * @lock: Pointer to a struct ttm_lock + * + * Releases a suspend lock + */ +extern void ttm_suspend_unlock(struct ttm_lock *lock); + +/** + * ttm_vt_lock + * + * @lock: Pointer to a struct ttm_lock + * @interruptible: Interruptible sleeping while waiting for a lock. + * @tfile: Pointer to a struct ttm_object_file to register the lock with. + * + * Takes the lock in vt mode. + * Returns: + * -ERESTARTSYS If interrupted by a signal and interruptible is true. + * -ENOMEM: Out of memory when locking. + */ +extern int ttm_vt_lock(struct ttm_lock *lock, bool interruptible, + struct ttm_object_file *tfile); + +/** + * ttm_vt_unlock + * + * @lock: Pointer to a struct ttm_lock + * + * Releases a vt lock. + * Returns: + * -EINVAL If the lock was not held. + */ +extern int ttm_vt_unlock(struct ttm_lock *lock); + +/** + * ttm_write_unlock + * + * @lock: Pointer to a struct ttm_lock + * + * Releases a write lock. + */ +extern void ttm_write_unlock(struct ttm_lock *lock); + +/** + * ttm_write_lock + * + * @lock: Pointer to a struct ttm_lock + * @interruptible: Interruptible sleeping while waiting for a lock. + * + * Takes the lock in write mode. + * Returns: + * -ERESTARTSYS If interrupted by a signal and interruptible is true. + */ +extern int ttm_write_lock(struct ttm_lock *lock, bool interruptible); + +/** + * ttm_lock_set_kill + * + * @lock: Pointer to a struct ttm_lock + * @val: Boolean whether to kill processes taking the lock. + * @signal: Signal to send to the process taking the lock. + * + * The kill-when-taking-lock functionality is used to kill processes that keep + * on using the TTM functionality when its resources has been taken down, for + * example when the X server exits. A typical sequence would look like this: + * - X server takes lock in write mode. + * - ttm_lock_set_kill() is called with @val set to true. + * - As part of X server exit, TTM resources are taken down. + * - X server releases the lock on file release. + * - Another dri client wants to render, takes the lock and is killed. + * + */ +static inline void ttm_lock_set_kill(struct ttm_lock *lock, bool val, + int signal) +{ + lock->kill_takers = val; + if (val) + lock->signal = signal; +} + +#endif diff --git a/drivers/gpu/drm/ttm/ttm_object.c b/drivers/gpu/drm/vmwgfx/ttm_object.c index 74f1b1eb1f8e..36990b80e790 100644 --- a/drivers/gpu/drm/ttm/ttm_object.c +++ b/drivers/gpu/drm/vmwgfx/ttm_object.c @@ -59,13 +59,12 @@ #define pr_fmt(fmt) "[TTM] " fmt -#include <drm/ttm/ttm_object.h> #include <drm/ttm/ttm_module.h> #include <linux/list.h> #include <linux/spinlock.h> #include <linux/slab.h> -#include <linux/module.h> #include <linux/atomic.h> +#include "ttm_object.h" struct ttm_object_file { struct ttm_object_device *tdev; @@ -95,6 +94,7 @@ struct ttm_object_device { struct dma_buf_ops ops; void (*dmabuf_release)(struct dma_buf *dma_buf); size_t dma_buf_size; + struct idr idr; }; /** @@ -172,14 +172,15 @@ int ttm_base_object_init(struct ttm_object_file *tfile, base->ref_obj_release = ref_obj_release; base->object_type = object_type; kref_init(&base->refcount); + idr_preload(GFP_KERNEL); spin_lock(&tdev->object_lock); - ret = drm_ht_just_insert_please_rcu(&tdev->object_hash, - &base->hash, - (unsigned long)base, 31, 0, 0); + ret = idr_alloc(&tdev->idr, base, 0, 0, GFP_NOWAIT); spin_unlock(&tdev->object_lock); - if (unlikely(ret != 0)) - goto out_err0; + idr_preload_end(); + if (ret < 0) + return ret; + base->handle = ret; ret = ttm_ref_object_add(tfile, base, TTM_REF_USAGE, NULL, false); if (unlikely(ret != 0)) goto out_err1; @@ -189,12 +190,10 @@ int ttm_base_object_init(struct ttm_object_file *tfile, return 0; out_err1: spin_lock(&tdev->object_lock); - (void)drm_ht_remove_item_rcu(&tdev->object_hash, &base->hash); + idr_remove(&tdev->idr, base->handle); spin_unlock(&tdev->object_lock); -out_err0: return ret; } -EXPORT_SYMBOL(ttm_base_object_init); static void ttm_release_base(struct kref *kref) { @@ -203,7 +202,7 @@ static void ttm_release_base(struct kref *kref) struct ttm_object_device *tdev = base->tfile->tdev; spin_lock(&tdev->object_lock); - (void)drm_ht_remove_item_rcu(&tdev->object_hash, &base->hash); + idr_remove(&tdev->idr, base->handle); spin_unlock(&tdev->object_lock); /* @@ -225,7 +224,41 @@ void ttm_base_object_unref(struct ttm_base_object **p_base) kref_put(&base->refcount, ttm_release_base); } -EXPORT_SYMBOL(ttm_base_object_unref); + +/** + * ttm_base_object_noref_lookup - look up a base object without reference + * @tfile: The struct ttm_object_file the object is registered with. + * @key: The object handle. + * + * This function looks up a ttm base object and returns a pointer to it + * without refcounting the pointer. The returned pointer is only valid + * until ttm_base_object_noref_release() is called, and the object + * pointed to by the returned pointer may be doomed. Any persistent usage + * of the object requires a refcount to be taken using kref_get_unless_zero(). + * Iff this function returns successfully it needs to be paired with + * ttm_base_object_noref_release() and no sleeping- or scheduling functions + * may be called inbetween these function callse. + * + * Return: A pointer to the object if successful or NULL otherwise. + */ +struct ttm_base_object * +ttm_base_object_noref_lookup(struct ttm_object_file *tfile, uint32_t key) +{ + struct drm_hash_item *hash; + struct drm_open_hash *ht = &tfile->ref_hash[TTM_REF_USAGE]; + int ret; + + rcu_read_lock(); + ret = drm_ht_find_item_rcu(ht, key, &hash); + if (ret) { + rcu_read_unlock(); + return NULL; + } + + __release(RCU); + return drm_hash_entry(hash, struct ttm_ref_object, hash)->obj; +} +EXPORT_SYMBOL(ttm_base_object_noref_lookup); struct ttm_base_object *ttm_base_object_lookup(struct ttm_object_file *tfile, uint32_t key) @@ -247,29 +280,21 @@ struct ttm_base_object *ttm_base_object_lookup(struct ttm_object_file *tfile, return base; } -EXPORT_SYMBOL(ttm_base_object_lookup); struct ttm_base_object * ttm_base_object_lookup_for_ref(struct ttm_object_device *tdev, uint32_t key) { - struct ttm_base_object *base = NULL; - struct drm_hash_item *hash; - struct drm_open_hash *ht = &tdev->object_hash; - int ret; + struct ttm_base_object *base; rcu_read_lock(); - ret = drm_ht_find_item_rcu(ht, key, &hash); + base = idr_find(&tdev->idr, key); - if (likely(ret == 0)) { - base = drm_hash_entry(hash, struct ttm_base_object, hash); - if (!kref_get_unless_zero(&base->refcount)) - base = NULL; - } + if (base && !kref_get_unless_zero(&base->refcount)) + base = NULL; rcu_read_unlock(); return base; } -EXPORT_SYMBOL(ttm_base_object_lookup_for_ref); /** * ttm_ref_object_exists - Check whether a caller has a valid ref object @@ -289,7 +314,7 @@ bool ttm_ref_object_exists(struct ttm_object_file *tfile, struct ttm_ref_object *ref; rcu_read_lock(); - if (unlikely(drm_ht_find_item_rcu(ht, base->hash.key, &hash) != 0)) + if (unlikely(drm_ht_find_item_rcu(ht, base->handle, &hash) != 0)) goto out_false; /* @@ -315,7 +340,6 @@ bool ttm_ref_object_exists(struct ttm_object_file *tfile, rcu_read_unlock(); return false; } -EXPORT_SYMBOL(ttm_ref_object_exists); int ttm_ref_object_add(struct ttm_object_file *tfile, struct ttm_base_object *base, @@ -340,7 +364,7 @@ int ttm_ref_object_add(struct ttm_object_file *tfile, while (ret == -EINVAL) { rcu_read_lock(); - ret = drm_ht_find_item_rcu(ht, base->hash.key, &hash); + ret = drm_ht_find_item_rcu(ht, base->handle, &hash); if (ret == 0) { ref = drm_hash_entry(hash, struct ttm_ref_object, hash); @@ -364,7 +388,7 @@ int ttm_ref_object_add(struct ttm_object_file *tfile, return -ENOMEM; } - ref->hash.key = base->hash.key; + ref->hash.key = base->handle; ref->obj = base; ref->tfile = tfile; ref->ref_type = ref_type; @@ -391,9 +415,9 @@ int ttm_ref_object_add(struct ttm_object_file *tfile, return ret; } -EXPORT_SYMBOL(ttm_ref_object_add); -static void ttm_ref_object_release(struct kref *kref) +static void __releases(tfile->lock) __acquires(tfile->lock) +ttm_ref_object_release(struct kref *kref) { struct ttm_ref_object *ref = container_of(kref, struct ttm_ref_object, kref); @@ -435,7 +459,6 @@ int ttm_ref_object_base_unref(struct ttm_object_file *tfile, spin_unlock(&tfile->lock); return 0; } -EXPORT_SYMBOL(ttm_ref_object_base_unref); void ttm_object_file_release(struct ttm_object_file **p_tfile) { @@ -464,7 +487,6 @@ void ttm_object_file_release(struct ttm_object_file **p_tfile) ttm_object_file_unref(&tfile); } -EXPORT_SYMBOL(ttm_object_file_release); struct ttm_object_file *ttm_object_file_init(struct ttm_object_device *tdev, unsigned int hash_order) @@ -499,7 +521,6 @@ out_err: return NULL; } -EXPORT_SYMBOL(ttm_object_file_init); struct ttm_object_device * ttm_object_device_init(struct ttm_mem_global *mem_glob, @@ -519,6 +540,7 @@ ttm_object_device_init(struct ttm_mem_global *mem_glob, if (ret != 0) goto out_no_object_hash; + idr_init(&tdev->idr); tdev->ops = *ops; tdev->dmabuf_release = tdev->ops.release; tdev->ops.release = ttm_prime_dmabuf_release; @@ -530,7 +552,6 @@ out_no_object_hash: kfree(tdev); return NULL; } -EXPORT_SYMBOL(ttm_object_device_init); void ttm_object_device_release(struct ttm_object_device **p_tdev) { @@ -538,11 +559,12 @@ void ttm_object_device_release(struct ttm_object_device **p_tdev) *p_tdev = NULL; + WARN_ON_ONCE(!idr_is_empty(&tdev->idr)); + idr_destroy(&tdev->idr); drm_ht_remove(&tdev->object_hash); kfree(tdev); } -EXPORT_SYMBOL(ttm_object_device_release); /** * get_dma_buf_unless_doomed - get a dma_buf reference if possible. @@ -641,14 +663,13 @@ int ttm_prime_fd_to_handle(struct ttm_object_file *tfile, prime = (struct ttm_prime_object *) dma_buf->priv; base = &prime->base; - *handle = base->hash.key; + *handle = base->handle; ret = ttm_ref_object_add(tfile, base, TTM_REF_USAGE, NULL, false); dma_buf_put(dma_buf); return ret; } -EXPORT_SYMBOL_GPL(ttm_prime_fd_to_handle); /** * ttm_prime_handle_to_fd - Return a dma_buf fd from a ttm prime object @@ -739,7 +760,6 @@ out_unref: ttm_base_object_unref(&base); return ret; } -EXPORT_SYMBOL_GPL(ttm_prime_handle_to_fd); /** * ttm_prime_object_init - Initialize a ttm_prime_object @@ -772,4 +792,3 @@ int ttm_prime_object_init(struct ttm_object_file *tfile, size_t size, ttm_prime_refcount_release, ref_obj_release); } -EXPORT_SYMBOL(ttm_prime_object_init); diff --git a/drivers/gpu/drm/vmwgfx/ttm_object.h b/drivers/gpu/drm/vmwgfx/ttm_object.h new file mode 100644 index 000000000000..50d26c7ff42d --- /dev/null +++ b/drivers/gpu/drm/vmwgfx/ttm_object.h @@ -0,0 +1,375 @@ +/************************************************************************** + * + * Copyright (c) 2006-2009 VMware, Inc., Palo Alto, CA., USA + * All Rights Reserved. + * + * 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, sub license, 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 (including the + * next paragraph) 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 NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS 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: Thomas Hellstrom <thellstrom-at-vmware-dot-com> + */ +/** @file ttm_object.h + * + * Base- and reference object implementation for the various + * ttm objects. Implements reference counting, minimal security checks + * and release on file close. + */ + +#ifndef _TTM_OBJECT_H_ +#define _TTM_OBJECT_H_ + +#include <linux/list.h> +#include <drm/drm_hashtab.h> +#include <linux/kref.h> +#include <linux/rcupdate.h> +#include <linux/dma-buf.h> +#include <drm/ttm/ttm_memory.h> + +/** + * enum ttm_ref_type + * + * Describes what type of reference a ref object holds. + * + * TTM_REF_USAGE is a simple refcount on a base object. + * + * TTM_REF_SYNCCPU_READ is a SYNCCPU_READ reference on a + * buffer object. + * + * TTM_REF_SYNCCPU_WRITE is a SYNCCPU_WRITE reference on a + * buffer object. + * + */ + +enum ttm_ref_type { + TTM_REF_USAGE, + TTM_REF_SYNCCPU_READ, + TTM_REF_SYNCCPU_WRITE, + TTM_REF_NUM +}; + +/** + * enum ttm_object_type + * + * One entry per ttm object type. + * Device-specific types should use the + * ttm_driver_typex types. + */ + +enum ttm_object_type { + ttm_fence_type, + ttm_buffer_type, + ttm_lock_type, + ttm_prime_type, + ttm_driver_type0 = 256, + ttm_driver_type1, + ttm_driver_type2, + ttm_driver_type3, + ttm_driver_type4, + ttm_driver_type5 +}; + +struct ttm_object_file; +struct ttm_object_device; + +/** + * struct ttm_base_object + * + * @hash: hash entry for the per-device object hash. + * @type: derived type this object is base class for. + * @shareable: Other ttm_object_files can access this object. + * + * @tfile: Pointer to ttm_object_file of the creator. + * NULL if the object was not created by a user request. + * (kernel object). + * + * @refcount: Number of references to this object, not + * including the hash entry. A reference to a base object can + * only be held by a ref object. + * + * @refcount_release: A function to be called when there are + * no more references to this object. This function should + * destroy the object (or make sure destruction eventually happens), + * and when it is called, the object has + * already been taken out of the per-device hash. The parameter + * "base" should be set to NULL by the function. + * + * @ref_obj_release: A function to be called when a reference object + * with another ttm_ref_type than TTM_REF_USAGE is deleted. + * This function may, for example, release a lock held by a user-space + * process. + * + * This struct is intended to be used as a base struct for objects that + * are visible to user-space. It provides a global name, race-safe + * access and refcounting, minimal access contol and hooks for unref actions. + */ + +struct ttm_base_object { + struct rcu_head rhead; + struct ttm_object_file *tfile; + struct kref refcount; + void (*refcount_release) (struct ttm_base_object **base); + void (*ref_obj_release) (struct ttm_base_object *base, + enum ttm_ref_type ref_type); + u32 handle; + enum ttm_object_type object_type; + u32 shareable; +}; + + +/** + * struct ttm_prime_object - Modified base object that is prime-aware + * + * @base: struct ttm_base_object that we derive from + * @mutex: Mutex protecting the @dma_buf member. + * @size: Size of the dma_buf associated with this object + * @real_type: Type of the underlying object. Needed since we're setting + * the value of @base::object_type to ttm_prime_type + * @dma_buf: Non ref-coutned pointer to a struct dma_buf created from this + * object. + * @refcount_release: The underlying object's release method. Needed since + * we set @base::refcount_release to our own release method. + */ + +struct ttm_prime_object { + struct ttm_base_object base; + struct mutex mutex; + size_t size; + enum ttm_object_type real_type; + struct dma_buf *dma_buf; + void (*refcount_release) (struct ttm_base_object **); +}; + +/** + * ttm_base_object_init + * + * @tfile: Pointer to a struct ttm_object_file. + * @base: The struct ttm_base_object to initialize. + * @shareable: This object is shareable with other applcations. + * (different @tfile pointers.) + * @type: The object type. + * @refcount_release: See the struct ttm_base_object description. + * @ref_obj_release: See the struct ttm_base_object description. + * + * Initializes a struct ttm_base_object. + */ + +extern int ttm_base_object_init(struct ttm_object_file *tfile, + struct ttm_base_object *base, + bool shareable, + enum ttm_object_type type, + void (*refcount_release) (struct ttm_base_object + **), + void (*ref_obj_release) (struct ttm_base_object + *, + enum ttm_ref_type + ref_type)); + +/** + * ttm_base_object_lookup + * + * @tfile: Pointer to a struct ttm_object_file. + * @key: Hash key + * + * Looks up a struct ttm_base_object with the key @key. + */ + +extern struct ttm_base_object *ttm_base_object_lookup(struct ttm_object_file + *tfile, uint32_t key); + +/** + * ttm_base_object_lookup_for_ref + * + * @tdev: Pointer to a struct ttm_object_device. + * @key: Hash key + * + * Looks up a struct ttm_base_object with the key @key. + * This function should only be used when the struct tfile associated with the + * caller doesn't yet have a reference to the base object. + */ + +extern struct ttm_base_object * +ttm_base_object_lookup_for_ref(struct ttm_object_device *tdev, uint32_t key); + +/** + * ttm_base_object_unref + * + * @p_base: Pointer to a pointer referencing a struct ttm_base_object. + * + * Decrements the base object refcount and clears the pointer pointed to by + * p_base. + */ + +extern void ttm_base_object_unref(struct ttm_base_object **p_base); + +/** + * ttm_ref_object_add. + * + * @tfile: A struct ttm_object_file representing the application owning the + * ref_object. + * @base: The base object to reference. + * @ref_type: The type of reference. + * @existed: Upon completion, indicates that an identical reference object + * already existed, and the refcount was upped on that object instead. + * @require_existed: Fail with -EPERM if an identical ref object didn't + * already exist. + * + * Checks that the base object is shareable and adds a ref object to it. + * + * Adding a ref object to a base object is basically like referencing the + * base object, but a user-space application holds the reference. When the + * file corresponding to @tfile is closed, all its reference objects are + * deleted. A reference object can have different types depending on what + * it's intended for. It can be refcounting to prevent object destruction, + * When user-space takes a lock, it can add a ref object to that lock to + * make sure the lock is released if the application dies. A ref object + * will hold a single reference on a base object. + */ +extern int ttm_ref_object_add(struct ttm_object_file *tfile, + struct ttm_base_object *base, + enum ttm_ref_type ref_type, bool *existed, + bool require_existed); + +extern bool ttm_ref_object_exists(struct ttm_object_file *tfile, + struct ttm_base_object *base); + +/** + * ttm_ref_object_base_unref + * + * @key: Key representing the base object. + * @ref_type: Ref type of the ref object to be dereferenced. + * + * Unreference a ref object with type @ref_type + * on the base object identified by @key. If there are no duplicate + * references, the ref object will be destroyed and the base object + * will be unreferenced. + */ +extern int ttm_ref_object_base_unref(struct ttm_object_file *tfile, + unsigned long key, + enum ttm_ref_type ref_type); + +/** + * ttm_object_file_init - initialize a struct ttm_object file + * + * @tdev: A struct ttm_object device this file is initialized on. + * @hash_order: Order of the hash table used to hold the reference objects. + * + * This is typically called by the file_ops::open function. + */ + +extern struct ttm_object_file *ttm_object_file_init(struct ttm_object_device + *tdev, + unsigned int hash_order); + +/** + * ttm_object_file_release - release data held by a ttm_object_file + * + * @p_tfile: Pointer to pointer to the ttm_object_file object to release. + * *p_tfile will be set to NULL by this function. + * + * Releases all data associated by a ttm_object_file. + * Typically called from file_ops::release. The caller must + * ensure that there are no concurrent users of tfile. + */ + +extern void ttm_object_file_release(struct ttm_object_file **p_tfile); + +/** + * ttm_object device init - initialize a struct ttm_object_device + * + * @mem_glob: struct ttm_mem_global for memory accounting. + * @hash_order: Order of hash table used to hash the base objects. + * @ops: DMA buf ops for prime objects of this device. + * + * This function is typically called on device initialization to prepare + * data structures needed for ttm base and ref objects. + */ + +extern struct ttm_object_device * +ttm_object_device_init(struct ttm_mem_global *mem_glob, + unsigned int hash_order, + const struct dma_buf_ops *ops); + +/** + * ttm_object_device_release - release data held by a ttm_object_device + * + * @p_tdev: Pointer to pointer to the ttm_object_device object to release. + * *p_tdev will be set to NULL by this function. + * + * Releases all data associated by a ttm_object_device. + * Typically called from driver::unload before the destruction of the + * device private data structure. + */ + +extern void ttm_object_device_release(struct ttm_object_device **p_tdev); + +#define ttm_base_object_kfree(__object, __base)\ + kfree_rcu(__object, __base.rhead) + +extern int ttm_prime_object_init(struct ttm_object_file *tfile, + size_t size, + struct ttm_prime_object *prime, + bool shareable, + enum ttm_object_type type, + void (*refcount_release) + (struct ttm_base_object **), + void (*ref_obj_release) + (struct ttm_base_object *, + enum ttm_ref_type ref_type)); + +static inline enum ttm_object_type +ttm_base_object_type(struct ttm_base_object *base) +{ + return (base->object_type == ttm_prime_type) ? + container_of(base, struct ttm_prime_object, base)->real_type : + base->object_type; +} +extern int ttm_prime_fd_to_handle(struct ttm_object_file *tfile, + int fd, u32 *handle); +extern int ttm_prime_handle_to_fd(struct ttm_object_file *tfile, + uint32_t handle, uint32_t flags, + int *prime_fd); + +#define ttm_prime_object_kfree(__obj, __prime) \ + kfree_rcu(__obj, __prime.base.rhead) + +/* + * Extra memory required by the base object's idr storage, which is allocated + * separately from the base object itself. We estimate an on-average 128 bytes + * per idr. + */ +#define TTM_OBJ_EXTRA_SIZE 128 + +struct ttm_base_object * +ttm_base_object_noref_lookup(struct ttm_object_file *tfile, uint32_t key); + +/** + * ttm_base_object_noref_release - release a base object pointer looked up + * without reference + * + * Releases a base object pointer looked up with ttm_base_object_noref_lookup(). + */ +static inline void ttm_base_object_noref_release(void) +{ + __acquire(RCU); + rcu_read_unlock(); +} +#endif diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c index 2dda03345761..7ce1c2f87d9a 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c @@ -30,7 +30,7 @@ #include <drm/drmP.h> #include "vmwgfx_drv.h" -#include "drm/ttm/ttm_object.h" +#include "ttm_object.h" /** @@ -441,7 +441,8 @@ static size_t vmw_bo_acc_size(struct vmw_private *dev_priv, size_t size, struct_size = backend_size + ttm_round_pot(sizeof(struct vmw_buffer_object)); user_struct_size = backend_size + - ttm_round_pot(sizeof(struct vmw_user_buffer_object)); + ttm_round_pot(sizeof(struct vmw_user_buffer_object)) + + TTM_OBJ_EXTRA_SIZE; } if (dev_priv->map_mode == vmw_dma_alloc_coherent) @@ -631,7 +632,7 @@ int vmw_user_bo_alloc(struct vmw_private *dev_priv, *p_base = &user_bo->prime.base; kref_get(&(*p_base)->refcount); } - *handle = user_bo->prime.base.hash.key; + *handle = user_bo->prime.base.handle; out_no_base_object: return ret; @@ -920,6 +921,47 @@ int vmw_user_bo_lookup(struct ttm_object_file *tfile, return 0; } +/** + * vmw_user_bo_noref_lookup - Look up a vmw user buffer object without reference + * @tfile: The TTM object file the handle is registered with. + * @handle: The user buffer object handle. + * + * This function looks up a struct vmw_user_bo and returns a pointer to the + * struct vmw_buffer_object it derives from without refcounting the pointer. + * The returned pointer is only valid until vmw_user_bo_noref_release() is + * called, and the object pointed to by the returned pointer may be doomed. + * Any persistent usage of the object requires a refcount to be taken using + * ttm_bo_reference_unless_doomed(). Iff this function returns successfully it + * needs to be paired with vmw_user_bo_noref_release() and no sleeping- + * or scheduling functions may be called inbetween these function calls. + * + * Return: A struct vmw_buffer_object pointer if successful or negative + * error pointer on failure. + */ +struct vmw_buffer_object * +vmw_user_bo_noref_lookup(struct ttm_object_file *tfile, u32 handle) +{ + struct vmw_user_buffer_object *vmw_user_bo; + struct ttm_base_object *base; + + base = ttm_base_object_noref_lookup(tfile, handle); + if (!base) { + DRM_ERROR("Invalid buffer object handle 0x%08lx.\n", + (unsigned long)handle); + return ERR_PTR(-ESRCH); + } + + if (unlikely(ttm_base_object_type(base) != ttm_buffer_type)) { + ttm_base_object_noref_release(); + DRM_ERROR("Invalid buffer object handle 0x%08lx.\n", + (unsigned long)handle); + return ERR_PTR(-EINVAL); + } + + vmw_user_bo = container_of(base, struct vmw_user_buffer_object, + prime.base); + return &vmw_user_bo->vbo; +} /** * vmw_user_bo_reference - Open a handle to a vmw user buffer object. @@ -940,7 +982,7 @@ int vmw_user_bo_reference(struct ttm_object_file *tfile, user_bo = container_of(vbo, struct vmw_user_buffer_object, vbo); - *handle = user_bo->prime.base.hash.key; + *handle = user_bo->prime.base.handle; return ttm_ref_object_add(tfile, &user_bo->prime.base, TTM_REF_USAGE, NULL, false); } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf.c index e7e4655d3f36..48d1380a952e 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf.c @@ -660,7 +660,7 @@ static void __vmw_cmdbuf_cur_flush(struct vmw_cmdbuf_man *man) { struct vmw_cmdbuf_header *cur = man->cur; - WARN_ON(!mutex_is_locked(&man->cur_mutex)); + lockdep_assert_held_once(&man->cur_mutex); if (!cur) return; @@ -1045,7 +1045,7 @@ static void vmw_cmdbuf_commit_cur(struct vmw_cmdbuf_man *man, { struct vmw_cmdbuf_header *cur = man->cur; - WARN_ON(!mutex_is_locked(&man->cur_mutex)); + lockdep_assert_held_once(&man->cur_mutex); WARN_ON(size > cur->reserved); man->cur_pos += size; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf_res.c b/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf_res.c index 3b75af9bf85f..4ac55fc2bf97 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf_res.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf_res.c @@ -89,8 +89,7 @@ vmw_cmdbuf_res_lookup(struct vmw_cmdbuf_res_manager *man, if (unlikely(ret != 0)) return ERR_PTR(ret); - return vmw_resource_reference - (drm_hash_entry(hash, struct vmw_cmdbuf_res, hash)->res); + return drm_hash_entry(hash, struct vmw_cmdbuf_res, hash)->res; } /** diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_context.c b/drivers/gpu/drm/vmwgfx/vmwgfx_context.c index 7c3cb8efd11a..14bd760a62fd 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_context.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_context.c @@ -217,9 +217,7 @@ static int vmw_gb_context_init(struct vmw_private *dev_priv, } } - - - vmw_resource_activate(res, vmw_hw_context_destroy); + res->hw_destroy = vmw_hw_context_destroy; return 0; out_cotables: @@ -274,7 +272,7 @@ static int vmw_context_init(struct vmw_private *dev_priv, vmw_fifo_commit(dev_priv, sizeof(*cmd)); vmw_fifo_resource_inc(dev_priv); - vmw_resource_activate(res, vmw_hw_context_destroy); + res->hw_destroy = vmw_hw_context_destroy; return 0; out_early: @@ -757,14 +755,10 @@ static int vmw_context_define(struct drm_device *dev, void *data, return -EINVAL; } - /* - * Approximate idr memory usage with 128 bytes. It will be limited - * by maximum number_of contexts anyway. - */ - if (unlikely(vmw_user_context_size == 0)) - vmw_user_context_size = ttm_round_pot(sizeof(*ctx)) + 128 + - ((dev_priv->has_mob) ? vmw_cmdbuf_res_man_size() : 0); + vmw_user_context_size = ttm_round_pot(sizeof(*ctx)) + + ((dev_priv->has_mob) ? vmw_cmdbuf_res_man_size() : 0) + + + VMW_IDA_ACC_SIZE + TTM_OBJ_EXTRA_SIZE; ret = ttm_read_lock(&dev_priv->reservation_sem, true); if (unlikely(ret != 0)) @@ -809,7 +803,7 @@ static int vmw_context_define(struct drm_device *dev, void *data, goto out_err; } - arg->cid = ctx->base.hash.key; + arg->cid = ctx->base.handle; out_err: vmw_resource_unreference(&res); out_unlock: @@ -867,9 +861,8 @@ struct vmw_resource *vmw_context_cotable(struct vmw_resource *ctx, if (cotable_type >= SVGA_COTABLE_DX10_MAX) return ERR_PTR(-EINVAL); - return vmw_resource_reference - (container_of(ctx, struct vmw_user_context, res)-> - cotables[cotable_type]); + return container_of(ctx, struct vmw_user_context, res)-> + cotables[cotable_type]; } /** diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_cotable.c b/drivers/gpu/drm/vmwgfx/vmwgfx_cotable.c index 1d45714e1d5a..44f3f6f107d3 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_cotable.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_cotable.c @@ -615,7 +615,7 @@ struct vmw_resource *vmw_cotable_alloc(struct vmw_private *dev_priv, vcotbl->type = type; vcotbl->ctx = ctx; - vmw_resource_activate(&vcotbl->res, vmw_hw_cotable_destroy); + vcotbl->res.hw_destroy = vmw_hw_cotable_destroy; return &vcotbl->res; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c index bb6dbbe18835..61a84b958d67 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c @@ -30,9 +30,9 @@ #include <drm/drmP.h> #include "vmwgfx_drv.h" #include "vmwgfx_binding.h" +#include "ttm_object.h" #include <drm/ttm/ttm_placement.h> #include <drm/ttm/ttm_bo_driver.h> -#include <drm/ttm/ttm_object.h> #include <drm/ttm/ttm_module.h> #include <linux/dma_remapping.h> @@ -667,8 +667,8 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) mutex_init(&dev_priv->binding_mutex); mutex_init(&dev_priv->requested_layout_mutex); mutex_init(&dev_priv->global_kms_state_mutex); - rwlock_init(&dev_priv->resource_lock); ttm_lock_init(&dev_priv->reservation_sem); + spin_lock_init(&dev_priv->resource_lock); spin_lock_init(&dev_priv->hw_lock); spin_lock_init(&dev_priv->waiter_lock); spin_lock_init(&dev_priv->cap_lock); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h index 1abe21758b0d..59f614225bcd 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -28,6 +28,7 @@ #ifndef _VMWGFX_DRV_H_ #define _VMWGFX_DRV_H_ +#include "vmwgfx_validation.h" #include "vmwgfx_reg.h" #include <drm/drmP.h> #include <drm/vmwgfx_drm.h> @@ -35,11 +36,11 @@ #include <drm/drm_auth.h> #include <linux/suspend.h> #include <drm/ttm/ttm_bo_driver.h> -#include <drm/ttm/ttm_object.h> -#include <drm/ttm/ttm_lock.h> #include <drm/ttm/ttm_execbuf_util.h> #include <drm/ttm/ttm_module.h> #include "vmwgfx_fence.h" +#include "ttm_object.h" +#include "ttm_lock.h" #include <linux/sync_file.h> #define VMWGFX_DRIVER_NAME "vmwgfx" @@ -112,21 +113,49 @@ struct vmw_validate_buffer { }; struct vmw_res_func; + + +/** + * struct vmw-resource - base class for hardware resources + * + * @kref: For refcounting. + * @dev_priv: Pointer to the device private for this resource. Immutable. + * @id: Device id. Protected by @dev_priv::resource_lock. + * @backup_size: Backup buffer size. Immutable. + * @res_dirty: Resource contains data not yet in the backup buffer. Protected + * by resource reserved. + * @backup_dirty: Backup buffer contains data not yet in the HW resource. + * Protecte by resource reserved. + * @backup: The backup buffer if any. Protected by resource reserved. + * @backup_offset: Offset into the backup buffer if any. Protected by resource + * reserved. Note that only a few resource types can have a @backup_offset + * different from zero. + * @pin_count: The pin count for this resource. A pinned resource has a + * pin-count greater than zero. It is not on the resource LRU lists and its + * backup buffer is pinned. Hence it can't be evicted. + * @func: Method vtable for this resource. Immutable. + * @lru_head: List head for the LRU list. Protected by @dev_priv::resource_lock. + * @mob_head: List head for the MOB backup list. Protected by @backup reserved. + * @binding_head: List head for the context binding list. Protected by + * the @dev_priv::binding_mutex + * @res_free: The resource destructor. + * @hw_destroy: Callback to destroy the resource on the device, as part of + * resource destruction. + */ struct vmw_resource { struct kref kref; struct vmw_private *dev_priv; int id; - bool avail; unsigned long backup_size; - bool res_dirty; /* Protected by backup buffer reserved */ - bool backup_dirty; /* Protected by backup buffer reserved */ + bool res_dirty; + bool backup_dirty; struct vmw_buffer_object *backup; unsigned long backup_offset; - unsigned long pin_count; /* Protected by resource reserved */ + unsigned long pin_count; const struct vmw_res_func *func; - struct list_head lru_head; /* Protected by the resource lock */ - struct list_head mob_head; /* Protected by @backup reserved */ - struct list_head binding_head; /* Protected by binding_mutex */ + struct list_head lru_head; + struct list_head mob_head; + struct list_head binding_head; void (*res_free) (struct vmw_resource *res); void (*hw_destroy) (struct vmw_resource *res); }; @@ -204,29 +233,24 @@ struct vmw_fifo_state { bool dx; }; -struct vmw_relocation { - SVGAMobId *mob_loc; - SVGAGuestPtr *location; - uint32_t index; -}; - /** * struct vmw_res_cache_entry - resource information cache entry - * + * @handle: User-space handle of a resource. + * @res: Non-ref-counted pointer to the resource. + * @valid_handle: Whether the @handle member is valid. * @valid: Whether the entry is valid, which also implies that the execbuf * code holds a reference to the resource, and it's placed on the * validation list. - * @handle: User-space handle of a resource. - * @res: Non-ref-counted pointer to the resource. * * Used to avoid frequent repeated user-space handle lookups of the * same resource. */ struct vmw_res_cache_entry { - bool valid; uint32_t handle; struct vmw_resource *res; - struct vmw_resource_val_node *node; + void *private; + unsigned short valid_handle; + unsigned short valid; }; /** @@ -291,35 +315,63 @@ enum vmw_display_unit_type { vmw_du_screen_target }; +struct vmw_validation_context; +struct vmw_ctx_validation_info; +/** + * struct vmw_sw_context - Command submission context + * @res_ht: Pointer hash table used to find validation duplicates + * @kernel: Whether the command buffer originates from kernel code rather + * than from user-space + * @fp: If @kernel is false, points to the file of the client. Otherwise + * NULL + * @cmd_bounce: Command bounce buffer used for command validation before + * copying to fifo space + * @cmd_bounce_size: Current command bounce buffer size + * @cur_query_bo: Current buffer object used as query result buffer + * @bo_relocations: List of buffer object relocations + * @res_relocations: List of resource relocations + * @buf_start: Pointer to start of memory where command validation takes + * place + * @res_cache: Cache of recently looked up resources + * @last_query_ctx: Last context that submitted a query + * @needs_post_query_barrier: Whether a query barrier is needed after + * command submission + * @staged_bindings: Cached per-context binding tracker + * @staged_bindings_inuse: Whether the cached per-context binding tracker + * is in use + * @staged_cmd_res: List of staged command buffer managed resources in this + * command buffer + * @ctx_list: List of context resources referenced in this command buffer + * @dx_ctx_node: Validation metadata of the current DX context + * @dx_query_mob: The MOB used for DX queries + * @dx_query_ctx: The DX context used for the last DX query + * @man: Pointer to the command buffer managed resource manager + * @ctx: The validation context + */ struct vmw_sw_context{ struct drm_open_hash res_ht; bool res_ht_initialized; - bool kernel; /**< is the called made from the kernel */ + bool kernel; struct vmw_fpriv *fp; - struct list_head validate_nodes; - struct vmw_relocation relocs[VMWGFX_MAX_RELOCATIONS]; - uint32_t cur_reloc; - struct vmw_validate_buffer val_bufs[VMWGFX_MAX_VALIDATIONS]; - uint32_t cur_val_buf; uint32_t *cmd_bounce; uint32_t cmd_bounce_size; - struct list_head resource_list; - struct list_head ctx_resource_list; /* For contexts and cotables */ struct vmw_buffer_object *cur_query_bo; + struct list_head bo_relocations; struct list_head res_relocations; uint32_t *buf_start; struct vmw_res_cache_entry res_cache[vmw_res_max]; struct vmw_resource *last_query_ctx; bool needs_post_query_barrier; - struct vmw_resource *error_resource; struct vmw_ctx_binding_state *staged_bindings; bool staged_bindings_inuse; struct list_head staged_cmd_res; - struct vmw_resource_val_node *dx_ctx_node; + struct list_head ctx_list; + struct vmw_ctx_validation_info *dx_ctx_node; struct vmw_buffer_object *dx_query_mob; struct vmw_resource *dx_query_ctx; struct vmw_cmdbuf_res_manager *man; + struct vmw_validation_context *ctx; }; struct vmw_legacy_display; @@ -444,7 +496,7 @@ struct vmw_private { * Context and surface management. */ - rwlock_t resource_lock; + spinlock_t resource_lock; struct idr res_idr[vmw_res_max]; /* * Block lastclose from racing with firstopen. @@ -628,7 +680,7 @@ extern void vmw_resource_unreference(struct vmw_resource **p_res); extern struct vmw_resource *vmw_resource_reference(struct vmw_resource *res); extern struct vmw_resource * vmw_resource_reference_unless_doomed(struct vmw_resource *res); -extern int vmw_resource_validate(struct vmw_resource *res); +extern int vmw_resource_validate(struct vmw_resource *res, bool intr); extern int vmw_resource_reserve(struct vmw_resource *res, bool interruptible, bool no_backup); extern bool vmw_resource_needs_backup(const struct vmw_resource *res); @@ -643,6 +695,12 @@ extern int vmw_user_resource_lookup_handle( uint32_t handle, const struct vmw_user_resource_conv *converter, struct vmw_resource **p_res); +extern struct vmw_resource * +vmw_user_resource_noref_lookup_handle(struct vmw_private *dev_priv, + struct ttm_object_file *tfile, + uint32_t handle, + const struct vmw_user_resource_conv * + converter); extern int vmw_stream_claim_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int vmw_stream_unref_ioctl(struct drm_device *dev, void *data, @@ -662,6 +720,15 @@ extern void vmw_resource_evict_all(struct vmw_private *dev_priv); extern void vmw_resource_unbind_list(struct vmw_buffer_object *vbo); /** + * vmw_user_resource_noref_release - release a user resource pointer looked up + * without reference + */ +static inline void vmw_user_resource_noref_release(void) +{ + ttm_base_object_noref_release(); +} + +/** * Buffer object helper functions - vmwgfx_bo.c */ extern int vmw_bo_pin_in_placement(struct vmw_private *vmw_priv, @@ -717,6 +784,18 @@ extern void vmw_bo_unmap(struct vmw_buffer_object *vbo); extern void vmw_bo_move_notify(struct ttm_buffer_object *bo, struct ttm_mem_reg *mem); extern void vmw_bo_swap_notify(struct ttm_buffer_object *bo); +extern struct vmw_buffer_object * +vmw_user_bo_noref_lookup(struct ttm_object_file *tfile, u32 handle); + +/** + * vmw_user_bo_noref_release - release a buffer object pointer looked up + * without reference + */ +static inline void vmw_user_bo_noref_release(void) +{ + ttm_base_object_noref_release(); +} + /** * Misc Ioctl functionality - vmwgfx_ioctl.c @@ -864,10 +943,6 @@ extern void vmw_execbuf_copy_fence_user(struct vmw_private *dev_priv, uint32_t fence_handle, int32_t out_fence_fd, struct sync_file *sync_file); -extern int vmw_validate_single_buffer(struct vmw_private *dev_priv, - struct ttm_buffer_object *bo, - bool interruptible, - bool validate_as_mob); bool vmw_cmd_describe(const void *buf, u32 *size, char const **cmd); /** diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c index f0ab6b2313bb..5a6b70ba137a 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c @@ -35,6 +35,23 @@ #define VMW_RES_HT_ORDER 12 +/* + * struct vmw_relocation - Buffer object relocation + * + * @head: List head for the command submission context's relocation list + * @vbo: Non ref-counted pointer to buffer object + * @mob_loc: Pointer to location for mob id to be modified + * @location: Pointer to location for guest pointer to be modified + */ +struct vmw_relocation { + struct list_head head; + struct vmw_buffer_object *vbo; + union { + SVGAMobId *mob_loc; + SVGAGuestPtr *location; + }; +}; + /** * enum vmw_resource_relocation_type - Relocation type for resources * @@ -69,35 +86,18 @@ struct vmw_resource_relocation { enum vmw_resource_relocation_type rel_type:3; }; -/** - * struct vmw_resource_val_node - Validation info for resources - * - * @head: List head for the software context's resource list. - * @hash: Hash entry for quick resouce to val_node lookup. - * @res: Ref-counted pointer to the resource. - * @switch_backup: Boolean whether to switch backup buffer on unreserve. - * @new_backup: Refcounted pointer to the new backup buffer. - * @staged_bindings: If @res is a context, tracks bindings set up during - * the command batch. Otherwise NULL. - * @new_backup_offset: New backup buffer offset if @new_backup is non-NUll. - * @first_usage: Set to true the first time the resource is referenced in - * the command stream. - * @switching_backup: The command stream provides a new backup buffer for a - * resource. - * @no_buffer_needed: This means @switching_backup is true on first buffer - * reference. So resource reservation does not need to allocate a backup - * buffer for the resource. +/* + * struct vmw_ctx_validation_info - Extra validation metadata for contexts + * @head: List head of context list + * @ctx: The context resource + * @cur: The context's persistent binding state + * @staged: The binding state changes of this command buffer */ -struct vmw_resource_val_node { +struct vmw_ctx_validation_info { struct list_head head; - struct drm_hash_item hash; - struct vmw_resource *res; - struct vmw_buffer_object *new_backup; - struct vmw_ctx_binding_state *staged_bindings; - unsigned long new_backup_offset; - u32 first_usage : 1; - u32 switching_backup : 1; - u32 no_buffer_needed : 1; + struct vmw_resource *ctx; + struct vmw_ctx_binding_state *cur; + struct vmw_ctx_binding_state *staged; }; /** @@ -127,10 +127,6 @@ static int vmw_translate_mob_ptr(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, SVGAMobId *id, struct vmw_buffer_object **vmw_bo_p); -static int vmw_bo_to_validate_list(struct vmw_sw_context *sw_context, - struct vmw_buffer_object *vbo, - bool validate_as_mob, - uint32_t *p_val_node); /** * vmw_ptr_diff - Compute the offset from a to b in bytes * @@ -145,48 +141,38 @@ static size_t vmw_ptr_diff(void *a, void *b) } /** - * vmw_resources_unreserve - unreserve resources previously reserved for - * command submission. - * - * @sw_context: pointer to the software context - * @backoff: Whether command submission failed. + * vmw_execbuf_bindings_commit - Commit modified binding state + * @sw_context: The command submission context + * @backoff: Whether this is part of the error path and binding state + * changes should be ignored */ -static void vmw_resources_unreserve(struct vmw_sw_context *sw_context, - bool backoff) +static void vmw_execbuf_bindings_commit(struct vmw_sw_context *sw_context, + bool backoff) { - struct vmw_resource_val_node *val; - struct list_head *list = &sw_context->resource_list; + struct vmw_ctx_validation_info *entry; - if (sw_context->dx_query_mob && !backoff) - vmw_context_bind_dx_query(sw_context->dx_query_ctx, - sw_context->dx_query_mob); + list_for_each_entry(entry, &sw_context->ctx_list, head) { + if (!backoff) + vmw_binding_state_commit(entry->cur, entry->staged); + if (entry->staged != sw_context->staged_bindings) + vmw_binding_state_free(entry->staged); + else + sw_context->staged_bindings_inuse = false; + } - list_for_each_entry(val, list, head) { - struct vmw_resource *res = val->res; - bool switch_backup = - (backoff) ? false : val->switching_backup; - - /* - * Transfer staged context bindings to the - * persistent context binding tracker. - */ - if (unlikely(val->staged_bindings)) { - if (!backoff) { - vmw_binding_state_commit - (vmw_context_binding_state(val->res), - val->staged_bindings); - } + /* List entries are freed with the validation context */ + INIT_LIST_HEAD(&sw_context->ctx_list); +} - if (val->staged_bindings != sw_context->staged_bindings) - vmw_binding_state_free(val->staged_bindings); - else - sw_context->staged_bindings_inuse = false; - val->staged_bindings = NULL; - } - vmw_resource_unreserve(res, switch_backup, val->new_backup, - val->new_backup_offset); - vmw_bo_unreference(&val->new_backup); - } +/** + * vmw_bind_dx_query_mob - Bind the DX query MOB if referenced + * @sw_context: The command submission context + */ +static void vmw_bind_dx_query_mob(struct vmw_sw_context *sw_context) +{ + if (sw_context->dx_query_mob) + vmw_context_bind_dx_query(sw_context->dx_query_ctx, + sw_context->dx_query_mob); } /** @@ -194,16 +180,17 @@ static void vmw_resources_unreserve(struct vmw_sw_context *sw_context, * added to the validate list. * * @dev_priv: Pointer to the device private: - * @sw_context: The validation context: - * @node: The validation node holding this context. + * @sw_context: The command submission context + * @node: The validation node holding the context resource metadata */ static int vmw_cmd_ctx_first_setup(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, - struct vmw_resource_val_node *node) + struct vmw_resource *res, + struct vmw_ctx_validation_info *node) { int ret; - ret = vmw_resource_context_res_add(dev_priv, sw_context, node->res); + ret = vmw_resource_context_res_add(dev_priv, sw_context, res); if (unlikely(ret != 0)) goto out_err; @@ -220,91 +207,138 @@ static int vmw_cmd_ctx_first_setup(struct vmw_private *dev_priv, } if (sw_context->staged_bindings_inuse) { - node->staged_bindings = vmw_binding_state_alloc(dev_priv); - if (IS_ERR(node->staged_bindings)) { + node->staged = vmw_binding_state_alloc(dev_priv); + if (IS_ERR(node->staged)) { DRM_ERROR("Failed to allocate context binding " "information.\n"); - ret = PTR_ERR(node->staged_bindings); - node->staged_bindings = NULL; + ret = PTR_ERR(node->staged); + node->staged = NULL; goto out_err; } } else { - node->staged_bindings = sw_context->staged_bindings; + node->staged = sw_context->staged_bindings; sw_context->staged_bindings_inuse = true; } + node->ctx = res; + node->cur = vmw_context_binding_state(res); + list_add_tail(&node->head, &sw_context->ctx_list); + return 0; out_err: return ret; } /** - * vmw_resource_val_add - Add a resource to the software context's - * resource list if it's not already on it. + * vmw_execbuf_res_size - calculate extra size fore the resource validation + * node + * @dev_priv: Pointer to the device private struct. + * @res_type: The resource type. * - * @sw_context: Pointer to the software context. + * Guest-backed contexts and DX contexts require extra size to store + * execbuf private information in the validation node. Typically the + * binding manager associated data structures. + * + * Returns: The extra size requirement based on resource type. + */ +static unsigned int vmw_execbuf_res_size(struct vmw_private *dev_priv, + enum vmw_res_type res_type) +{ + return (res_type == vmw_res_dx_context || + (res_type == vmw_res_context && dev_priv->has_mob)) ? + sizeof(struct vmw_ctx_validation_info) : 0; +} + +/** + * vmw_execbuf_rcache_update - Update a resource-node cache entry + * + * @rcache: Pointer to the entry to update. * @res: Pointer to the resource. - * @p_node On successful return points to a valid pointer to a - * struct vmw_resource_val_node, if non-NULL on entry. + * @private: Pointer to the execbuf-private space in the resource + * validation node. + */ +static void vmw_execbuf_rcache_update(struct vmw_res_cache_entry *rcache, + struct vmw_resource *res, + void *private) +{ + rcache->res = res; + rcache->private = private; + rcache->valid = 1; + rcache->valid_handle = 0; +} + +/** + * vmw_execbuf_res_noref_val_add - Add a resource described by an + * unreferenced rcu-protected pointer to the validation list. + * @sw_context: Pointer to the software context. + * @res: Unreferenced rcu-protected pointer to the resource. + * + * Returns: 0 on success. Negative error code on failure. Typical error + * codes are %-EINVAL on inconsistency and %-ESRCH if the resource was + * doomed. */ -static int vmw_resource_val_add(struct vmw_sw_context *sw_context, - struct vmw_resource *res, - struct vmw_resource_val_node **p_node) +static int vmw_execbuf_res_noref_val_add(struct vmw_sw_context *sw_context, + struct vmw_resource *res) { struct vmw_private *dev_priv = res->dev_priv; - struct vmw_resource_val_node *node; - struct drm_hash_item *hash; int ret; - - if (likely(drm_ht_find_item(&sw_context->res_ht, (unsigned long) res, - &hash) == 0)) { - node = container_of(hash, struct vmw_resource_val_node, hash); - node->first_usage = false; - if (unlikely(p_node != NULL)) - *p_node = node; + enum vmw_res_type res_type = vmw_res_type(res); + struct vmw_res_cache_entry *rcache; + struct vmw_ctx_validation_info *ctx_info; + bool first_usage; + unsigned int priv_size; + + rcache = &sw_context->res_cache[res_type]; + if (likely(rcache->valid && rcache->res == res)) { + vmw_user_resource_noref_release(); return 0; } - node = kzalloc(sizeof(*node), GFP_KERNEL); - if (unlikely(!node)) { - DRM_ERROR("Failed to allocate a resource validation " - "entry.\n"); - return -ENOMEM; - } - - node->hash.key = (unsigned long) res; - ret = drm_ht_insert_item(&sw_context->res_ht, &node->hash); - if (unlikely(ret != 0)) { - DRM_ERROR("Failed to initialize a resource validation " - "entry.\n"); - kfree(node); + priv_size = vmw_execbuf_res_size(dev_priv, res_type); + ret = vmw_validation_add_resource(sw_context->ctx, res, priv_size, + (void **)&ctx_info, &first_usage); + vmw_user_resource_noref_release(); + if (ret) return ret; + + if (priv_size && first_usage) { + ret = vmw_cmd_ctx_first_setup(dev_priv, sw_context, res, + ctx_info); + if (ret) + return ret; } - node->res = vmw_resource_reference(res); - node->first_usage = true; - if (unlikely(p_node != NULL)) - *p_node = node; - if (!dev_priv->has_mob) { - list_add_tail(&node->head, &sw_context->resource_list); + vmw_execbuf_rcache_update(rcache, res, ctx_info); + return 0; +} + +/** + * vmw_execbuf_res_noctx_val_add - Add a non-context resource to the resource + * validation list if it's not already on it + * @sw_context: Pointer to the software context. + * @res: Pointer to the resource. + * + * Returns: Zero on success. Negative error code on failure. + */ +static int vmw_execbuf_res_noctx_val_add(struct vmw_sw_context *sw_context, + struct vmw_resource *res) +{ + struct vmw_res_cache_entry *rcache; + enum vmw_res_type res_type = vmw_res_type(res); + void *ptr; + int ret; + + rcache = &sw_context->res_cache[res_type]; + if (likely(rcache->valid && rcache->res == res)) return 0; - } - switch (vmw_res_type(res)) { - case vmw_res_context: - case vmw_res_dx_context: - list_add(&node->head, &sw_context->ctx_resource_list); - ret = vmw_cmd_ctx_first_setup(dev_priv, sw_context, node); - break; - case vmw_res_cotable: - list_add_tail(&node->head, &sw_context->ctx_resource_list); - break; - default: - list_add_tail(&node->head, &sw_context->resource_list); - break; - } + ret = vmw_validation_add_resource(sw_context->ctx, res, 0, &ptr, NULL); + if (ret) + return ret; - return ret; + vmw_execbuf_rcache_update(rcache, res, ptr); + + return 0; } /** @@ -325,11 +359,11 @@ static int vmw_view_res_val_add(struct vmw_sw_context *sw_context, * First add the resource the view is pointing to, otherwise * it may be swapped out when the view is validated. */ - ret = vmw_resource_val_add(sw_context, vmw_view_srf(view), NULL); + ret = vmw_execbuf_res_noctx_val_add(sw_context, vmw_view_srf(view)); if (ret) return ret; - return vmw_resource_val_add(sw_context, view, NULL); + return vmw_execbuf_res_noctx_val_add(sw_context, view); } /** @@ -342,28 +376,33 @@ static int vmw_view_res_val_add(struct vmw_sw_context *sw_context, * * The view is represented by a view id and the DX context it's created on, * or scheduled for creation on. If there is no DX context set, the function - * will return -EINVAL. Otherwise returns 0 on success and -EINVAL on failure. + * will return an -EINVAL error pointer. + * + * Returns: Unreferenced pointer to the resource on success, negative error + * pointer on failure. */ -static int vmw_view_id_val_add(struct vmw_sw_context *sw_context, - enum vmw_view_type view_type, u32 id) +static struct vmw_resource * +vmw_view_id_val_add(struct vmw_sw_context *sw_context, + enum vmw_view_type view_type, u32 id) { - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; struct vmw_resource *view; int ret; if (!ctx_node) { DRM_ERROR("DX Context not set.\n"); - return -EINVAL; + return ERR_PTR(-EINVAL); } view = vmw_view_lookup(sw_context->man, view_type, id); if (IS_ERR(view)) - return PTR_ERR(view); + return view; ret = vmw_view_res_val_add(sw_context, view); - vmw_resource_unreference(&view); + if (ret) + return ERR_PTR(ret); - return ret; + return view; } /** @@ -394,8 +433,7 @@ static int vmw_resource_context_res_add(struct vmw_private *dev_priv, if (IS_ERR(res)) continue; - ret = vmw_resource_val_add(sw_context, res, NULL); - vmw_resource_unreference(&res); + ret = vmw_execbuf_res_noctx_val_add(sw_context, res); if (unlikely(ret != 0)) return ret; } @@ -407,17 +445,11 @@ static int vmw_resource_context_res_add(struct vmw_private *dev_priv, binding_list = vmw_context_binding_list(ctx); list_for_each_entry(entry, binding_list, ctx_list) { - /* entry->res is not refcounted */ - res = vmw_resource_reference_unless_doomed(entry->res); - if (unlikely(res == NULL)) - continue; - if (vmw_res_type(entry->res) == vmw_res_view) ret = vmw_view_res_val_add(sw_context, entry->res); else - ret = vmw_resource_val_add(sw_context, entry->res, - NULL); - vmw_resource_unreference(&res); + ret = vmw_execbuf_res_noctx_val_add(sw_context, + entry->res); if (unlikely(ret != 0)) break; } @@ -427,9 +459,8 @@ static int vmw_resource_context_res_add(struct vmw_private *dev_priv, dx_query_mob = vmw_context_get_dx_query_mob(ctx); if (dx_query_mob) - ret = vmw_bo_to_validate_list(sw_context, - dx_query_mob, - true, NULL); + ret = vmw_validation_add_bo(sw_context->ctx, + dx_query_mob, true, false); } mutex_unlock(&dev_priv->binding_mutex); @@ -445,7 +476,7 @@ static int vmw_resource_context_res_add(struct vmw_private *dev_priv, * id that needs fixup is located. Granularity is one byte. * @rel_type: Relocation type. */ -static int vmw_resource_relocation_add(struct list_head *list, +static int vmw_resource_relocation_add(struct vmw_sw_context *sw_context, const struct vmw_resource *res, unsigned long offset, enum vmw_resource_relocation_type @@ -453,7 +484,7 @@ static int vmw_resource_relocation_add(struct list_head *list, { struct vmw_resource_relocation *rel; - rel = kmalloc(sizeof(*rel), GFP_KERNEL); + rel = vmw_validation_mem_alloc(sw_context->ctx, sizeof(*rel)); if (unlikely(!rel)) { DRM_ERROR("Failed to allocate a resource relocation.\n"); return -ENOMEM; @@ -462,7 +493,7 @@ static int vmw_resource_relocation_add(struct list_head *list, rel->res = res; rel->offset = offset; rel->rel_type = rel_type; - list_add_tail(&rel->head, list); + list_add_tail(&rel->head, &sw_context->res_relocations); return 0; } @@ -470,16 +501,13 @@ static int vmw_resource_relocation_add(struct list_head *list, /** * vmw_resource_relocations_free - Free all relocations on a list * - * @list: Pointer to the head of the relocation list. + * @list: Pointer to the head of the relocation list */ static void vmw_resource_relocations_free(struct list_head *list) { - struct vmw_resource_relocation *rel, *n; + /* Memory is validation context memory, so no need to free it */ - list_for_each_entry_safe(rel, n, list, head) { - list_del(&rel->head); - kfree(rel); - } + INIT_LIST_HEAD(list); } /** @@ -532,68 +560,6 @@ static int vmw_cmd_ok(struct vmw_private *dev_priv, } /** - * vmw_bo_to_validate_list - add a bo to a validate list - * - * @sw_context: The software context used for this command submission batch. - * @bo: The buffer object to add. - * @validate_as_mob: Validate this buffer as a MOB. - * @p_val_node: If non-NULL Will be updated with the validate node number - * on return. - * - * Returns -EINVAL if the limit of number of buffer objects per command - * submission is reached. - */ -static int vmw_bo_to_validate_list(struct vmw_sw_context *sw_context, - struct vmw_buffer_object *vbo, - bool validate_as_mob, - uint32_t *p_val_node) -{ - uint32_t val_node; - struct vmw_validate_buffer *vval_buf; - struct ttm_validate_buffer *val_buf; - struct drm_hash_item *hash; - int ret; - - if (likely(drm_ht_find_item(&sw_context->res_ht, (unsigned long) vbo, - &hash) == 0)) { - vval_buf = container_of(hash, struct vmw_validate_buffer, - hash); - if (unlikely(vval_buf->validate_as_mob != validate_as_mob)) { - DRM_ERROR("Inconsistent buffer usage.\n"); - return -EINVAL; - } - val_buf = &vval_buf->base; - val_node = vval_buf - sw_context->val_bufs; - } else { - val_node = sw_context->cur_val_buf; - if (unlikely(val_node >= VMWGFX_MAX_VALIDATIONS)) { - DRM_ERROR("Max number of DMA buffers per submission " - "exceeded.\n"); - return -EINVAL; - } - vval_buf = &sw_context->val_bufs[val_node]; - vval_buf->hash.key = (unsigned long) vbo; - ret = drm_ht_insert_item(&sw_context->res_ht, &vval_buf->hash); - if (unlikely(ret != 0)) { - DRM_ERROR("Failed to initialize a buffer validation " - "entry.\n"); - return ret; - } - ++sw_context->cur_val_buf; - val_buf = &vval_buf->base; - val_buf->bo = ttm_bo_reference(&vbo->base); - val_buf->shared = false; - list_add_tail(&val_buf->head, &sw_context->validate_nodes); - vval_buf->validate_as_mob = validate_as_mob; - } - - if (p_val_node) - *p_val_node = val_node; - - return 0; -} - -/** * vmw_resources_reserve - Reserve all resources on the sw_context's * resource list. * @@ -605,27 +571,11 @@ static int vmw_bo_to_validate_list(struct vmw_sw_context *sw_context, */ static int vmw_resources_reserve(struct vmw_sw_context *sw_context) { - struct vmw_resource_val_node *val; - int ret = 0; - - list_for_each_entry(val, &sw_context->resource_list, head) { - struct vmw_resource *res = val->res; - - ret = vmw_resource_reserve(res, true, val->no_buffer_needed); - if (unlikely(ret != 0)) - return ret; - - if (res->backup) { - struct vmw_buffer_object *vbo = res->backup; - - ret = vmw_bo_to_validate_list - (sw_context, vbo, - vmw_resource_needs_backup(res), NULL); + int ret; - if (unlikely(ret != 0)) - return ret; - } - } + ret = vmw_validation_res_reserve(sw_context->ctx, true); + if (ret) + return ret; if (sw_context->dx_query_mob) { struct vmw_buffer_object *expected_dx_query_mob; @@ -642,87 +592,6 @@ static int vmw_resources_reserve(struct vmw_sw_context *sw_context) } /** - * vmw_resources_validate - Validate all resources on the sw_context's - * resource list. - * - * @sw_context: Pointer to the software context. - * - * Before this function is called, all resource backup buffers must have - * been validated. - */ -static int vmw_resources_validate(struct vmw_sw_context *sw_context) -{ - struct vmw_resource_val_node *val; - int ret; - - list_for_each_entry(val, &sw_context->resource_list, head) { - struct vmw_resource *res = val->res; - struct vmw_buffer_object *backup = res->backup; - - ret = vmw_resource_validate(res); - if (unlikely(ret != 0)) { - if (ret != -ERESTARTSYS) - DRM_ERROR("Failed to validate resource.\n"); - return ret; - } - - /* Check if the resource switched backup buffer */ - if (backup && res->backup && (backup != res->backup)) { - struct vmw_buffer_object *vbo = res->backup; - - ret = vmw_bo_to_validate_list - (sw_context, vbo, - vmw_resource_needs_backup(res), NULL); - if (ret) { - ttm_bo_unreserve(&vbo->base); - return ret; - } - } - } - return 0; -} - -/** - * vmw_cmd_res_reloc_add - Add a resource to a software context's - * relocation- and validation lists. - * - * @dev_priv: Pointer to a struct vmw_private identifying the device. - * @sw_context: Pointer to the software context. - * @id_loc: Pointer to where the id that needs translation is located. - * @res: Valid pointer to a struct vmw_resource. - * @p_val: If non null, a pointer to the struct vmw_resource_validate_node - * used for this resource is returned here. - */ -static int vmw_cmd_res_reloc_add(struct vmw_private *dev_priv, - struct vmw_sw_context *sw_context, - uint32_t *id_loc, - struct vmw_resource *res, - struct vmw_resource_val_node **p_val) -{ - int ret; - struct vmw_resource_val_node *node; - - *p_val = NULL; - ret = vmw_resource_relocation_add(&sw_context->res_relocations, - res, - vmw_ptr_diff(sw_context->buf_start, - id_loc), - vmw_res_rel_normal); - if (unlikely(ret != 0)) - return ret; - - ret = vmw_resource_val_add(sw_context, res, &node); - if (unlikely(ret != 0)) - return ret; - - if (p_val) - *p_val = node; - - return 0; -} - - -/** * vmw_cmd_res_check - Check that a resource is present and if so, put it * on the resource validate list unless it's already there. * @@ -741,17 +610,16 @@ vmw_cmd_res_check(struct vmw_private *dev_priv, enum vmw_res_type res_type, const struct vmw_user_resource_conv *converter, uint32_t *id_loc, - struct vmw_resource_val_node **p_val) + struct vmw_resource **p_res) { - struct vmw_res_cache_entry *rcache = - &sw_context->res_cache[res_type]; + struct vmw_res_cache_entry *rcache = &sw_context->res_cache[res_type]; struct vmw_resource *res; - struct vmw_resource_val_node *node; int ret; + if (p_res) + *p_res = NULL; + if (*id_loc == SVGA3D_INVALID_ID) { - if (p_val) - *p_val = NULL; if (res_type == vmw_res_context) { DRM_ERROR("Illegal context invalid id.\n"); return -EINVAL; @@ -759,56 +627,41 @@ vmw_cmd_res_check(struct vmw_private *dev_priv, return 0; } - /* - * Fastpath in case of repeated commands referencing the same - * resource - */ + if (likely(rcache->valid_handle && *id_loc == rcache->handle)) { + res = rcache->res; + } else { + unsigned int size = vmw_execbuf_res_size(dev_priv, res_type); - if (likely(rcache->valid && *id_loc == rcache->handle)) { - const struct vmw_resource *res = rcache->res; + ret = vmw_validation_preload_res(sw_context->ctx, size); + if (ret) + return ret; - rcache->node->first_usage = false; - if (p_val) - *p_val = rcache->node; + res = vmw_user_resource_noref_lookup_handle + (dev_priv, sw_context->fp->tfile, *id_loc, converter); + if (unlikely(IS_ERR(res))) { + DRM_ERROR("Could not find or use resource 0x%08x.\n", + (unsigned int) *id_loc); + return PTR_ERR(res); + } - return vmw_resource_relocation_add - (&sw_context->res_relocations, res, - vmw_ptr_diff(sw_context->buf_start, id_loc), - vmw_res_rel_normal); - } + ret = vmw_execbuf_res_noref_val_add(sw_context, res); + if (unlikely(ret != 0)) + return ret; - ret = vmw_user_resource_lookup_handle(dev_priv, - sw_context->fp->tfile, - *id_loc, - converter, - &res); - if (unlikely(ret != 0)) { - DRM_ERROR("Could not find or use resource 0x%08x.\n", - (unsigned) *id_loc); - dump_stack(); - return ret; + if (rcache->valid && rcache->res == res) { + rcache->valid_handle = true; + rcache->handle = *id_loc; + } } - rcache->valid = true; - rcache->res = res; - rcache->handle = *id_loc; - - ret = vmw_cmd_res_reloc_add(dev_priv, sw_context, id_loc, - res, &node); - if (unlikely(ret != 0)) - goto out_no_reloc; + ret = vmw_resource_relocation_add(sw_context, res, + vmw_ptr_diff(sw_context->buf_start, + id_loc), + vmw_res_rel_normal); + if (p_res) + *p_res = res; - rcache->node = node; - if (p_val) - *p_val = node; - vmw_resource_unreference(&res); return 0; - -out_no_reloc: - BUG_ON(sw_context->error_resource != NULL); - sw_context->error_resource = res; - - return ret; } /** @@ -861,22 +714,18 @@ static int vmw_rebind_all_dx_query(struct vmw_resource *ctx_res) */ static int vmw_rebind_contexts(struct vmw_sw_context *sw_context) { - struct vmw_resource_val_node *val; + struct vmw_ctx_validation_info *val; int ret; - list_for_each_entry(val, &sw_context->resource_list, head) { - if (unlikely(!val->staged_bindings)) - break; - - ret = vmw_binding_rebind_all - (vmw_context_binding_state(val->res)); + list_for_each_entry(val, &sw_context->ctx_list, head) { + ret = vmw_binding_rebind_all(val->cur); if (unlikely(ret != 0)) { if (ret != -ERESTARTSYS) DRM_ERROR("Failed to rebind context.\n"); return ret; } - ret = vmw_rebind_all_dx_query(val->res); + ret = vmw_rebind_all_dx_query(val->ctx); if (ret != 0) return ret; } @@ -903,45 +752,33 @@ static int vmw_view_bindings_add(struct vmw_sw_context *sw_context, uint32 view_ids[], u32 num_views, u32 first_slot) { - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; - struct vmw_cmdbuf_res_manager *man; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; u32 i; - int ret; if (!ctx_node) { DRM_ERROR("DX Context not set.\n"); return -EINVAL; } - man = sw_context->man; for (i = 0; i < num_views; ++i) { struct vmw_ctx_bindinfo_view binding; struct vmw_resource *view = NULL; if (view_ids[i] != SVGA3D_INVALID_ID) { - view = vmw_view_lookup(man, view_type, view_ids[i]); + view = vmw_view_id_val_add(sw_context, view_type, + view_ids[i]); if (IS_ERR(view)) { DRM_ERROR("View not found.\n"); return PTR_ERR(view); } - - ret = vmw_view_res_val_add(sw_context, view); - if (ret) { - DRM_ERROR("Could not add view to " - "validation list.\n"); - vmw_resource_unreference(&view); - return ret; - } } - binding.bi.ctx = ctx_node->res; + binding.bi.ctx = ctx_node->ctx; binding.bi.res = view; binding.bi.bt = binding_type; binding.shader_slot = shader_slot; binding.slot = first_slot + i; - vmw_binding_add(ctx_node->staged_bindings, &binding.bi, + vmw_binding_add(ctx_node->staged, &binding.bi, shader_slot, binding.slot); - if (view) - vmw_resource_unreference(&view); } return 0; @@ -971,6 +808,34 @@ static int vmw_cmd_cid_check(struct vmw_private *dev_priv, user_context_converter, &cmd->cid, NULL); } +/** + * vmw_execbuf_info_from_res - Get the private validation metadata for a + * recently validated resource + * @sw_context: Pointer to the command submission context + * @res: The resource + * + * The resource pointed to by @res needs to be present in the command submission + * context's resource cache and hence the last resource of that type to be + * processed by the validation code. + * + * Return: a pointer to the private metadata of the resource, or NULL + * if it wasn't found + */ +static struct vmw_ctx_validation_info * +vmw_execbuf_info_from_res(struct vmw_sw_context *sw_context, + struct vmw_resource *res) +{ + struct vmw_res_cache_entry *rcache = + &sw_context->res_cache[vmw_res_type(res)]; + + if (rcache->valid && rcache->res == res) + return rcache->private; + + WARN_ON_ONCE(true); + return NULL; +} + + static int vmw_cmd_set_render_target_check(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, SVGA3dCmdHeader *header) @@ -979,8 +844,8 @@ static int vmw_cmd_set_render_target_check(struct vmw_private *dev_priv, SVGA3dCmdHeader header; SVGA3dCmdSetRenderTarget body; } *cmd; - struct vmw_resource_val_node *ctx_node; - struct vmw_resource_val_node *res_node; + struct vmw_resource *ctx; + struct vmw_resource *res; int ret; cmd = container_of(header, struct vmw_sid_cmd, header); @@ -993,25 +858,29 @@ static int vmw_cmd_set_render_target_check(struct vmw_private *dev_priv, ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_context, user_context_converter, &cmd->body.cid, - &ctx_node); + &ctx); if (unlikely(ret != 0)) return ret; ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface, - user_surface_converter, - &cmd->body.target.sid, &res_node); - if (unlikely(ret != 0)) + user_surface_converter, &cmd->body.target.sid, + &res); + if (unlikely(ret)) return ret; if (dev_priv->has_mob) { struct vmw_ctx_bindinfo_view binding; + struct vmw_ctx_validation_info *node; - binding.bi.ctx = ctx_node->res; - binding.bi.res = res_node ? res_node->res : NULL; + node = vmw_execbuf_info_from_res(sw_context, ctx); + if (!node) + return -EINVAL; + + binding.bi.ctx = ctx; + binding.bi.res = res; binding.bi.bt = vmw_ctx_binding_rt; binding.slot = cmd->body.type; - vmw_binding_add(ctx_node->staged_bindings, - &binding.bi, 0, binding.slot); + vmw_binding_add(node->staged, &binding.bi, 0, binding.slot); } return 0; @@ -1030,8 +899,8 @@ static int vmw_cmd_surface_copy_check(struct vmw_private *dev_priv, cmd = container_of(header, struct vmw_sid_cmd, header); ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface, - user_surface_converter, - &cmd->body.src.sid, NULL); + user_surface_converter, + &cmd->body.src.sid, NULL); if (ret) return ret; @@ -1171,17 +1040,17 @@ static int vmw_query_bo_switch_prepare(struct vmw_private *dev_priv, if (unlikely(sw_context->cur_query_bo != NULL)) { sw_context->needs_post_query_barrier = true; - ret = vmw_bo_to_validate_list(sw_context, - sw_context->cur_query_bo, - dev_priv->has_mob, NULL); + ret = vmw_validation_add_bo(sw_context->ctx, + sw_context->cur_query_bo, + dev_priv->has_mob, false); if (unlikely(ret != 0)) return ret; } sw_context->cur_query_bo = new_query_bo; - ret = vmw_bo_to_validate_list(sw_context, - dev_priv->dummy_query_bo, - dev_priv->has_mob, NULL); + ret = vmw_validation_add_bo(sw_context->ctx, + dev_priv->dummy_query_bo, + dev_priv->has_mob, false); if (unlikely(ret != 0)) return ret; @@ -1269,7 +1138,7 @@ static void vmw_query_bo_switch_commit(struct vmw_private *dev_priv, * @sw_context: The software context used for this command batch validation. * @id: Pointer to the user-space handle to be translated. * @vmw_bo_p: Points to a location that, on successful return will carry - * a reference-counted pointer to the DMA buffer identified by the + * a non-reference-counted pointer to the buffer object identified by the * user-space handle in @id. * * This function saves information needed to translate a user-space buffer @@ -1284,40 +1153,34 @@ static int vmw_translate_mob_ptr(struct vmw_private *dev_priv, SVGAMobId *id, struct vmw_buffer_object **vmw_bo_p) { - struct vmw_buffer_object *vmw_bo = NULL; + struct vmw_buffer_object *vmw_bo; uint32_t handle = *id; struct vmw_relocation *reloc; int ret; - ret = vmw_user_bo_lookup(sw_context->fp->tfile, handle, &vmw_bo, NULL); - if (unlikely(ret != 0)) { + vmw_validation_preload_bo(sw_context->ctx); + vmw_bo = vmw_user_bo_noref_lookup(sw_context->fp->tfile, handle); + if (IS_ERR(vmw_bo)) { DRM_ERROR("Could not find or use MOB buffer.\n"); - ret = -EINVAL; - goto out_no_reloc; + return PTR_ERR(vmw_bo); } - if (unlikely(sw_context->cur_reloc >= VMWGFX_MAX_RELOCATIONS)) { - DRM_ERROR("Max number relocations per submission" - " exceeded\n"); - ret = -EINVAL; - goto out_no_reloc; - } + ret = vmw_validation_add_bo(sw_context->ctx, vmw_bo, true, false); + vmw_user_bo_noref_release(); + if (unlikely(ret != 0)) + return ret; - reloc = &sw_context->relocs[sw_context->cur_reloc++]; - reloc->mob_loc = id; - reloc->location = NULL; + reloc = vmw_validation_mem_alloc(sw_context->ctx, sizeof(*reloc)); + if (!reloc) + return -ENOMEM; - ret = vmw_bo_to_validate_list(sw_context, vmw_bo, true, &reloc->index); - if (unlikely(ret != 0)) - goto out_no_reloc; + reloc->mob_loc = id; + reloc->vbo = vmw_bo; *vmw_bo_p = vmw_bo; - return 0; + list_add_tail(&reloc->head, &sw_context->bo_relocations); -out_no_reloc: - vmw_bo_unreference(&vmw_bo); - *vmw_bo_p = NULL; - return ret; + return 0; } /** @@ -1328,7 +1191,7 @@ out_no_reloc: * @sw_context: The software context used for this command batch validation. * @ptr: Pointer to the user-space handle to be translated. * @vmw_bo_p: Points to a location that, on successful return will carry - * a reference-counted pointer to the DMA buffer identified by the + * a non-reference-counted pointer to the DMA buffer identified by the * user-space handle in @id. * * This function saves information needed to translate a user-space buffer @@ -1344,39 +1207,33 @@ static int vmw_translate_guest_ptr(struct vmw_private *dev_priv, SVGAGuestPtr *ptr, struct vmw_buffer_object **vmw_bo_p) { - struct vmw_buffer_object *vmw_bo = NULL; + struct vmw_buffer_object *vmw_bo; uint32_t handle = ptr->gmrId; struct vmw_relocation *reloc; int ret; - ret = vmw_user_bo_lookup(sw_context->fp->tfile, handle, &vmw_bo, NULL); - if (unlikely(ret != 0)) { + vmw_validation_preload_bo(sw_context->ctx); + vmw_bo = vmw_user_bo_noref_lookup(sw_context->fp->tfile, handle); + if (IS_ERR(vmw_bo)) { DRM_ERROR("Could not find or use GMR region.\n"); - ret = -EINVAL; - goto out_no_reloc; - } - - if (unlikely(sw_context->cur_reloc >= VMWGFX_MAX_RELOCATIONS)) { - DRM_ERROR("Max number relocations per submission" - " exceeded\n"); - ret = -EINVAL; - goto out_no_reloc; + return PTR_ERR(vmw_bo); } - reloc = &sw_context->relocs[sw_context->cur_reloc++]; - reloc->location = ptr; - - ret = vmw_bo_to_validate_list(sw_context, vmw_bo, false, &reloc->index); + ret = vmw_validation_add_bo(sw_context->ctx, vmw_bo, false, false); + vmw_user_bo_noref_release(); if (unlikely(ret != 0)) - goto out_no_reloc; + return ret; + + reloc = vmw_validation_mem_alloc(sw_context->ctx, sizeof(*reloc)); + if (!reloc) + return -ENOMEM; + reloc->location = ptr; + reloc->vbo = vmw_bo; *vmw_bo_p = vmw_bo; - return 0; + list_add_tail(&reloc->head, &sw_context->bo_relocations); -out_no_reloc: - vmw_bo_unreference(&vmw_bo); - *vmw_bo_p = NULL; - return ret; + return 0; } @@ -1400,7 +1257,7 @@ static int vmw_cmd_dx_define_query(struct vmw_private *dev_priv, } *cmd; int ret; - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; struct vmw_resource *cotable_res; @@ -1415,9 +1272,8 @@ static int vmw_cmd_dx_define_query(struct vmw_private *dev_priv, cmd->q.type >= SVGA3D_QUERYTYPE_MAX) return -EINVAL; - cotable_res = vmw_context_cotable(ctx_node->res, SVGA_COTABLE_DXQUERY); + cotable_res = vmw_context_cotable(ctx_node->ctx, SVGA_COTABLE_DXQUERY); ret = vmw_cotable_notify(cotable_res, cmd->q.queryId); - vmw_resource_unreference(&cotable_res); return ret; } @@ -1462,11 +1318,8 @@ static int vmw_cmd_dx_bind_query(struct vmw_private *dev_priv, return ret; sw_context->dx_query_mob = vmw_bo; - sw_context->dx_query_ctx = sw_context->dx_ctx_node->res; - - vmw_bo_unreference(&vmw_bo); - - return ret; + sw_context->dx_query_ctx = sw_context->dx_ctx_node->ctx; + return 0; } @@ -1567,7 +1420,6 @@ static int vmw_cmd_end_gb_query(struct vmw_private *dev_priv, ret = vmw_query_bo_switch_prepare(dev_priv, vmw_bo, sw_context); - vmw_bo_unreference(&vmw_bo); return ret; } @@ -1621,7 +1473,6 @@ static int vmw_cmd_end_query(struct vmw_private *dev_priv, ret = vmw_query_bo_switch_prepare(dev_priv, vmw_bo, sw_context); - vmw_bo_unreference(&vmw_bo); return ret; } @@ -1654,7 +1505,6 @@ static int vmw_cmd_wait_gb_query(struct vmw_private *dev_priv, if (unlikely(ret != 0)) return ret; - vmw_bo_unreference(&vmw_bo); return 0; } @@ -1706,7 +1556,6 @@ static int vmw_cmd_wait_query(struct vmw_private *dev_priv, if (unlikely(ret != 0)) return ret; - vmw_bo_unreference(&vmw_bo); return 0; } @@ -1757,7 +1606,7 @@ static int vmw_cmd_dma(struct vmw_private *dev_priv, if (unlikely(ret != 0)) { if (unlikely(ret != -ERESTARTSYS)) DRM_ERROR("could not find surface for DMA.\n"); - goto out_no_surface; + return ret; } srf = vmw_res_to_srf(sw_context->res_cache[vmw_res_surface].res); @@ -1765,9 +1614,7 @@ static int vmw_cmd_dma(struct vmw_private *dev_priv, vmw_kms_cursor_snoop(srf, sw_context->fp->tfile, &vmw_bo->base, header); -out_no_surface: - vmw_bo_unreference(&vmw_bo); - return ret; + return 0; } static int vmw_cmd_draw(struct vmw_private *dev_priv, @@ -1837,8 +1684,8 @@ static int vmw_cmd_tex_state(struct vmw_private *dev_priv, ((unsigned long) header + header->size + sizeof(header)); SVGA3dTextureState *cur_state = (SVGA3dTextureState *) ((unsigned long) header + sizeof(struct vmw_tex_state_cmd)); - struct vmw_resource_val_node *ctx_node; - struct vmw_resource_val_node *res_node; + struct vmw_resource *ctx; + struct vmw_resource *res; int ret; cmd = container_of(header, struct vmw_tex_state_cmd, @@ -1846,7 +1693,7 @@ static int vmw_cmd_tex_state(struct vmw_private *dev_priv, ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_context, user_context_converter, &cmd->state.cid, - &ctx_node); + &ctx); if (unlikely(ret != 0)) return ret; @@ -1862,19 +1709,24 @@ static int vmw_cmd_tex_state(struct vmw_private *dev_priv, ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface, user_surface_converter, - &cur_state->value, &res_node); + &cur_state->value, &res); if (unlikely(ret != 0)) return ret; if (dev_priv->has_mob) { struct vmw_ctx_bindinfo_tex binding; + struct vmw_ctx_validation_info *node; + + node = vmw_execbuf_info_from_res(sw_context, ctx); + if (!node) + return -EINVAL; - binding.bi.ctx = ctx_node->res; - binding.bi.res = res_node ? res_node->res : NULL; + binding.bi.ctx = ctx; + binding.bi.res = res; binding.bi.bt = vmw_ctx_binding_tex; binding.texture_stage = cur_state->stage; - vmw_binding_add(ctx_node->staged_bindings, &binding.bi, - 0, binding.texture_stage); + vmw_binding_add(node->staged, &binding.bi, 0, + binding.texture_stage); } } @@ -1893,14 +1745,9 @@ static int vmw_cmd_check_define_gmrfb(struct vmw_private *dev_priv, SVGAFifoCmdDefineGMRFB body; } *cmd = buf; - ret = vmw_translate_guest_ptr(dev_priv, sw_context, - &cmd->body.ptr, - &vmw_bo); - if (unlikely(ret != 0)) - return ret; - - vmw_bo_unreference(&vmw_bo); - + return vmw_translate_guest_ptr(dev_priv, sw_context, + &cmd->body.ptr, + &vmw_bo); return ret; } @@ -1922,25 +1769,24 @@ static int vmw_cmd_check_define_gmrfb(struct vmw_private *dev_priv, */ static int vmw_cmd_res_switch_backup(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, - struct vmw_resource_val_node *val_node, + struct vmw_resource *res, uint32_t *buf_id, unsigned long backup_offset) { - struct vmw_buffer_object *dma_buf; + struct vmw_buffer_object *vbo; + void *info; int ret; - ret = vmw_translate_mob_ptr(dev_priv, sw_context, buf_id, &dma_buf); + info = vmw_execbuf_info_from_res(sw_context, res); + if (!info) + return -EINVAL; + + ret = vmw_translate_mob_ptr(dev_priv, sw_context, buf_id, &vbo); if (ret) return ret; - val_node->switching_backup = true; - if (val_node->first_usage) - val_node->no_buffer_needed = true; - - vmw_bo_unreference(&val_node->new_backup); - val_node->new_backup = dma_buf; - val_node->new_backup_offset = backup_offset; - + vmw_validation_res_switch_backup(sw_context->ctx, info, vbo, + backup_offset); return 0; } @@ -1970,15 +1816,15 @@ static int vmw_cmd_switch_backup(struct vmw_private *dev_priv, uint32_t *buf_id, unsigned long backup_offset) { - struct vmw_resource_val_node *val_node; + struct vmw_resource *res; int ret; ret = vmw_cmd_res_check(dev_priv, sw_context, res_type, - converter, res_id, &val_node); + converter, res_id, &res); if (ret) return ret; - return vmw_cmd_res_switch_backup(dev_priv, sw_context, val_node, + return vmw_cmd_res_switch_backup(dev_priv, sw_context, res, buf_id, backup_offset); } @@ -2170,14 +2016,14 @@ static int vmw_cmd_shader_define(struct vmw_private *dev_priv, } *cmd; int ret; size_t size; - struct vmw_resource_val_node *val; + struct vmw_resource *ctx; cmd = container_of(header, struct vmw_shader_define_cmd, header); ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_context, user_context_converter, &cmd->body.cid, - &val); + &ctx); if (unlikely(ret != 0)) return ret; @@ -2186,14 +2032,14 @@ static int vmw_cmd_shader_define(struct vmw_private *dev_priv, size = cmd->header.size - sizeof(cmd->body); ret = vmw_compat_shader_add(dev_priv, - vmw_context_res_man(val->res), + vmw_context_res_man(ctx), cmd->body.shid, cmd + 1, cmd->body.type, size, &sw_context->staged_cmd_res); if (unlikely(ret != 0)) return ret; - return vmw_resource_relocation_add(&sw_context->res_relocations, + return vmw_resource_relocation_add(sw_context, NULL, vmw_ptr_diff(sw_context->buf_start, &cmd->header.id), @@ -2217,28 +2063,28 @@ static int vmw_cmd_shader_destroy(struct vmw_private *dev_priv, SVGA3dCmdDestroyShader body; } *cmd; int ret; - struct vmw_resource_val_node *val; + struct vmw_resource *ctx; cmd = container_of(header, struct vmw_shader_destroy_cmd, header); ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_context, user_context_converter, &cmd->body.cid, - &val); + &ctx); if (unlikely(ret != 0)) return ret; if (unlikely(!dev_priv->has_mob)) return 0; - ret = vmw_shader_remove(vmw_context_res_man(val->res), + ret = vmw_shader_remove(vmw_context_res_man(ctx), cmd->body.shid, cmd->body.type, &sw_context->staged_cmd_res); if (unlikely(ret != 0)) return ret; - return vmw_resource_relocation_add(&sw_context->res_relocations, + return vmw_resource_relocation_add(sw_context, NULL, vmw_ptr_diff(sw_context->buf_start, &cmd->header.id), @@ -2261,9 +2107,9 @@ static int vmw_cmd_set_shader(struct vmw_private *dev_priv, SVGA3dCmdHeader header; SVGA3dCmdSetShader body; } *cmd; - struct vmw_resource_val_node *ctx_node, *res_node = NULL; struct vmw_ctx_bindinfo_shader binding; - struct vmw_resource *res = NULL; + struct vmw_resource *ctx, *res = NULL; + struct vmw_ctx_validation_info *ctx_info; int ret; cmd = container_of(header, struct vmw_set_shader_cmd, @@ -2277,7 +2123,7 @@ static int vmw_cmd_set_shader(struct vmw_private *dev_priv, ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_context, user_context_converter, &cmd->body.cid, - &ctx_node); + &ctx); if (unlikely(ret != 0)) return ret; @@ -2285,34 +2131,35 @@ static int vmw_cmd_set_shader(struct vmw_private *dev_priv, return 0; if (cmd->body.shid != SVGA3D_INVALID_ID) { - res = vmw_shader_lookup(vmw_context_res_man(ctx_node->res), + res = vmw_shader_lookup(vmw_context_res_man(ctx), cmd->body.shid, cmd->body.type); if (!IS_ERR(res)) { - ret = vmw_cmd_res_reloc_add(dev_priv, sw_context, - &cmd->body.shid, res, - &res_node); - vmw_resource_unreference(&res); + ret = vmw_execbuf_res_noctx_val_add(sw_context, res); if (unlikely(ret != 0)) return ret; } } - if (!res_node) { + if (IS_ERR_OR_NULL(res)) { ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_shader, user_shader_converter, - &cmd->body.shid, &res_node); + &cmd->body.shid, &res); if (unlikely(ret != 0)) return ret; } - binding.bi.ctx = ctx_node->res; - binding.bi.res = res_node ? res_node->res : NULL; + ctx_info = vmw_execbuf_info_from_res(sw_context, ctx); + if (!ctx_info) + return -EINVAL; + + binding.bi.ctx = ctx; + binding.bi.res = res; binding.bi.bt = vmw_ctx_binding_shader; binding.shader_slot = cmd->body.type - SVGA3D_SHADERTYPE_MIN; - vmw_binding_add(ctx_node->staged_bindings, &binding.bi, + vmw_binding_add(ctx_info->staged, &binding.bi, binding.shader_slot, 0); return 0; } @@ -2393,8 +2240,8 @@ vmw_cmd_dx_set_single_constant_buffer(struct vmw_private *dev_priv, SVGA3dCmdHeader header; SVGA3dCmdDXSetSingleConstantBuffer body; } *cmd; - struct vmw_resource_val_node *res_node = NULL; - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; + struct vmw_resource *res = NULL; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; struct vmw_ctx_bindinfo_cb binding; int ret; @@ -2406,12 +2253,12 @@ vmw_cmd_dx_set_single_constant_buffer(struct vmw_private *dev_priv, cmd = container_of(header, typeof(*cmd), header); ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface, user_surface_converter, - &cmd->body.sid, &res_node); + &cmd->body.sid, &res); if (unlikely(ret != 0)) return ret; - binding.bi.ctx = ctx_node->res; - binding.bi.res = res_node ? res_node->res : NULL; + binding.bi.ctx = ctx_node->ctx; + binding.bi.res = res; binding.bi.bt = vmw_ctx_binding_cb; binding.shader_slot = cmd->body.type - SVGA3D_SHADERTYPE_MIN; binding.offset = cmd->body.offsetInBytes; @@ -2426,7 +2273,7 @@ vmw_cmd_dx_set_single_constant_buffer(struct vmw_private *dev_priv, return -EINVAL; } - vmw_binding_add(ctx_node->staged_bindings, &binding.bi, + vmw_binding_add(ctx_node->staged, &binding.bi, binding.shader_slot, binding.slot); return 0; @@ -2482,7 +2329,7 @@ static int vmw_cmd_dx_set_shader(struct vmw_private *dev_priv, SVGA3dCmdDXSetShader body; } *cmd; struct vmw_resource *res = NULL; - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; struct vmw_ctx_bindinfo_shader binding; int ret = 0; @@ -2506,23 +2353,20 @@ static int vmw_cmd_dx_set_shader(struct vmw_private *dev_priv, return PTR_ERR(res); } - ret = vmw_resource_val_add(sw_context, res, NULL); + ret = vmw_execbuf_res_noctx_val_add(sw_context, res); if (ret) - goto out_unref; + return ret; } - binding.bi.ctx = ctx_node->res; + binding.bi.ctx = ctx_node->ctx; binding.bi.res = res; binding.bi.bt = vmw_ctx_binding_dx_shader; binding.shader_slot = cmd->body.type - SVGA3D_SHADERTYPE_MIN; - vmw_binding_add(ctx_node->staged_bindings, &binding.bi, + vmw_binding_add(ctx_node->staged, &binding.bi, binding.shader_slot, 0); -out_unref: - if (res) - vmw_resource_unreference(&res); - return ret; + return 0; } /** @@ -2537,9 +2381,9 @@ static int vmw_cmd_dx_set_vertex_buffers(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, SVGA3dCmdHeader *header) { - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; struct vmw_ctx_bindinfo_vb binding; - struct vmw_resource_val_node *res_node; + struct vmw_resource *res; struct { SVGA3dCmdHeader header; SVGA3dCmdDXSetVertexBuffers body; @@ -2564,18 +2408,18 @@ static int vmw_cmd_dx_set_vertex_buffers(struct vmw_private *dev_priv, for (i = 0; i < num; i++) { ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface, user_surface_converter, - &cmd->buf[i].sid, &res_node); + &cmd->buf[i].sid, &res); if (unlikely(ret != 0)) return ret; - binding.bi.ctx = ctx_node->res; + binding.bi.ctx = ctx_node->ctx; binding.bi.bt = vmw_ctx_binding_vb; - binding.bi.res = ((res_node) ? res_node->res : NULL); + binding.bi.res = res; binding.offset = cmd->buf[i].offset; binding.stride = cmd->buf[i].stride; binding.slot = i + cmd->body.startBuffer; - vmw_binding_add(ctx_node->staged_bindings, &binding.bi, + vmw_binding_add(ctx_node->staged, &binding.bi, 0, binding.slot); } @@ -2594,9 +2438,9 @@ static int vmw_cmd_dx_set_index_buffer(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, SVGA3dCmdHeader *header) { - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; struct vmw_ctx_bindinfo_ib binding; - struct vmw_resource_val_node *res_node; + struct vmw_resource *res; struct { SVGA3dCmdHeader header; SVGA3dCmdDXSetIndexBuffer body; @@ -2611,17 +2455,17 @@ static int vmw_cmd_dx_set_index_buffer(struct vmw_private *dev_priv, cmd = container_of(header, typeof(*cmd), header); ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface, user_surface_converter, - &cmd->body.sid, &res_node); + &cmd->body.sid, &res); if (unlikely(ret != 0)) return ret; - binding.bi.ctx = ctx_node->res; - binding.bi.res = ((res_node) ? res_node->res : NULL); + binding.bi.ctx = ctx_node->ctx; + binding.bi.res = res; binding.bi.bt = vmw_ctx_binding_ib; binding.offset = cmd->body.offset; binding.format = cmd->body.format; - vmw_binding_add(ctx_node->staged_bindings, &binding.bi, 0, 0); + vmw_binding_add(ctx_node->staged, &binding.bi, 0, 0); return 0; } @@ -2679,8 +2523,8 @@ static int vmw_cmd_dx_clear_rendertarget_view(struct vmw_private *dev_priv, SVGA3dCmdDXClearRenderTargetView body; } *cmd = container_of(header, typeof(*cmd), header); - return vmw_view_id_val_add(sw_context, vmw_view_rt, - cmd->body.renderTargetViewId); + return PTR_RET(vmw_view_id_val_add(sw_context, vmw_view_rt, + cmd->body.renderTargetViewId)); } /** @@ -2700,16 +2544,16 @@ static int vmw_cmd_dx_clear_depthstencil_view(struct vmw_private *dev_priv, SVGA3dCmdDXClearDepthStencilView body; } *cmd = container_of(header, typeof(*cmd), header); - return vmw_view_id_val_add(sw_context, vmw_view_ds, - cmd->body.depthStencilViewId); + return PTR_RET(vmw_view_id_val_add(sw_context, vmw_view_ds, + cmd->body.depthStencilViewId)); } static int vmw_cmd_dx_view_define(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, SVGA3dCmdHeader *header) { - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; - struct vmw_resource_val_node *srf_node; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; + struct vmw_resource *srf; struct vmw_resource *res; enum vmw_view_type view_type; int ret; @@ -2734,19 +2578,18 @@ static int vmw_cmd_dx_view_define(struct vmw_private *dev_priv, cmd = container_of(header, typeof(*cmd), header); ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface, user_surface_converter, - &cmd->sid, &srf_node); + &cmd->sid, &srf); if (unlikely(ret != 0)) return ret; - res = vmw_context_cotable(ctx_node->res, vmw_view_cotables[view_type]); + res = vmw_context_cotable(ctx_node->ctx, vmw_view_cotables[view_type]); ret = vmw_cotable_notify(res, cmd->defined_id); - vmw_resource_unreference(&res); if (unlikely(ret != 0)) return ret; return vmw_view_add(sw_context->man, - ctx_node->res, - srf_node->res, + ctx_node->ctx, + srf, view_type, cmd->defined_id, header, @@ -2766,9 +2609,9 @@ static int vmw_cmd_dx_set_so_targets(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, SVGA3dCmdHeader *header) { - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; struct vmw_ctx_bindinfo_so binding; - struct vmw_resource_val_node *res_node; + struct vmw_resource *res; struct { SVGA3dCmdHeader header; SVGA3dCmdDXSetSOTargets body; @@ -2793,18 +2636,18 @@ static int vmw_cmd_dx_set_so_targets(struct vmw_private *dev_priv, for (i = 0; i < num; i++) { ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface, user_surface_converter, - &cmd->targets[i].sid, &res_node); + &cmd->targets[i].sid, &res); if (unlikely(ret != 0)) return ret; - binding.bi.ctx = ctx_node->res; - binding.bi.res = ((res_node) ? res_node->res : NULL); + binding.bi.ctx = ctx_node->ctx; + binding.bi.res = res; binding.bi.bt = vmw_ctx_binding_so, binding.offset = cmd->targets[i].offset; binding.size = cmd->targets[i].sizeInBytes; binding.slot = i; - vmw_binding_add(ctx_node->staged_bindings, &binding.bi, + vmw_binding_add(ctx_node->staged, &binding.bi, 0, binding.slot); } @@ -2815,7 +2658,7 @@ static int vmw_cmd_dx_so_define(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, SVGA3dCmdHeader *header) { - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; struct vmw_resource *res; /* * This is based on the fact that all affected define commands have @@ -2834,10 +2677,9 @@ static int vmw_cmd_dx_so_define(struct vmw_private *dev_priv, } so_type = vmw_so_cmd_to_type(header->id); - res = vmw_context_cotable(ctx_node->res, vmw_so_cotables[so_type]); + res = vmw_context_cotable(ctx_node->ctx, vmw_so_cotables[so_type]); cmd = container_of(header, typeof(*cmd), header); ret = vmw_cotable_notify(res, cmd->defined_id); - vmw_resource_unreference(&res); return ret; } @@ -2882,7 +2724,7 @@ static int vmw_cmd_dx_cid_check(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, SVGA3dCmdHeader *header) { - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; if (unlikely(ctx_node == NULL)) { DRM_ERROR("DX Context not set.\n"); @@ -2907,7 +2749,7 @@ static int vmw_cmd_dx_view_remove(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, SVGA3dCmdHeader *header) { - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; struct { SVGA3dCmdHeader header; union vmw_view_destroy body; @@ -2934,7 +2776,7 @@ static int vmw_cmd_dx_view_remove(struct vmw_private *dev_priv, * relocation to conditionally make this command a NOP to avoid * device errors. */ - return vmw_resource_relocation_add(&sw_context->res_relocations, + return vmw_resource_relocation_add(sw_context, view, vmw_ptr_diff(sw_context->buf_start, &cmd->header.id), @@ -2953,7 +2795,7 @@ static int vmw_cmd_dx_define_shader(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, SVGA3dCmdHeader *header) { - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; struct vmw_resource *res; struct { SVGA3dCmdHeader header; @@ -2966,13 +2808,12 @@ static int vmw_cmd_dx_define_shader(struct vmw_private *dev_priv, return -EINVAL; } - res = vmw_context_cotable(ctx_node->res, SVGA_COTABLE_DXSHADER); + res = vmw_context_cotable(ctx_node->ctx, SVGA_COTABLE_DXSHADER); ret = vmw_cotable_notify(res, cmd->body.shaderId); - vmw_resource_unreference(&res); if (ret) return ret; - return vmw_dx_shader_add(sw_context->man, ctx_node->res, + return vmw_dx_shader_add(sw_context->man, ctx_node->ctx, cmd->body.shaderId, cmd->body.type, &sw_context->staged_cmd_res); } @@ -2989,7 +2830,7 @@ static int vmw_cmd_dx_destroy_shader(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, SVGA3dCmdHeader *header) { - struct vmw_resource_val_node *ctx_node = sw_context->dx_ctx_node; + struct vmw_ctx_validation_info *ctx_node = sw_context->dx_ctx_node; struct { SVGA3dCmdHeader header; SVGA3dCmdDXDestroyShader body; @@ -3021,8 +2862,7 @@ static int vmw_cmd_dx_bind_shader(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, SVGA3dCmdHeader *header) { - struct vmw_resource_val_node *ctx_node; - struct vmw_resource_val_node *res_node; + struct vmw_resource *ctx; struct vmw_resource *res; struct { SVGA3dCmdHeader header; @@ -3033,38 +2873,33 @@ static int vmw_cmd_dx_bind_shader(struct vmw_private *dev_priv, if (cmd->body.cid != SVGA3D_INVALID_ID) { ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_context, user_context_converter, - &cmd->body.cid, &ctx_node); + &cmd->body.cid, &ctx); if (ret) return ret; } else { - ctx_node = sw_context->dx_ctx_node; - if (!ctx_node) { + if (!sw_context->dx_ctx_node) { DRM_ERROR("DX Context not set.\n"); return -EINVAL; } + ctx = sw_context->dx_ctx_node->ctx; } - res = vmw_shader_lookup(vmw_context_res_man(ctx_node->res), + res = vmw_shader_lookup(vmw_context_res_man(ctx), cmd->body.shid, 0); if (IS_ERR(res)) { DRM_ERROR("Could not find shader to bind.\n"); return PTR_ERR(res); } - ret = vmw_resource_val_add(sw_context, res, &res_node); + ret = vmw_execbuf_res_noctx_val_add(sw_context, res); if (ret) { DRM_ERROR("Error creating resource validation node.\n"); - goto out_unref; + return ret; } - - ret = vmw_cmd_res_switch_backup(dev_priv, sw_context, res_node, - &cmd->body.mobid, - cmd->body.offsetInBytes); -out_unref: - vmw_resource_unreference(&res); - - return ret; + return vmw_cmd_res_switch_backup(dev_priv, sw_context, res, + &cmd->body.mobid, + cmd->body.offsetInBytes); } /** @@ -3083,8 +2918,8 @@ static int vmw_cmd_dx_genmips(struct vmw_private *dev_priv, SVGA3dCmdDXGenMips body; } *cmd = container_of(header, typeof(*cmd), header); - return vmw_view_id_val_add(sw_context, vmw_view_sr, - cmd->body.shaderResourceViewId); + return PTR_RET(vmw_view_id_val_add(sw_context, vmw_view_sr, + cmd->body.shaderResourceViewId)); } /** @@ -3638,20 +3473,18 @@ static int vmw_cmd_check_all(struct vmw_private *dev_priv, static void vmw_free_relocations(struct vmw_sw_context *sw_context) { - sw_context->cur_reloc = 0; + /* Memory is validation context memory, so no need to free it */ + + INIT_LIST_HEAD(&sw_context->bo_relocations); } static void vmw_apply_relocations(struct vmw_sw_context *sw_context) { - uint32_t i; struct vmw_relocation *reloc; - struct ttm_validate_buffer *validate; struct ttm_buffer_object *bo; - for (i = 0; i < sw_context->cur_reloc; ++i) { - reloc = &sw_context->relocs[i]; - validate = &sw_context->val_bufs[reloc->index].base; - bo = validate->bo; + list_for_each_entry(reloc, &sw_context->bo_relocations, head) { + bo = &reloc->vbo->base; switch (bo->mem.mem_type) { case TTM_PL_VRAM: reloc->location->offset += bo->offset; @@ -3670,110 +3503,6 @@ static void vmw_apply_relocations(struct vmw_sw_context *sw_context) vmw_free_relocations(sw_context); } -/** - * vmw_resource_list_unrefererence - Free up a resource list and unreference - * all resources referenced by it. - * - * @list: The resource list. - */ -static void vmw_resource_list_unreference(struct vmw_sw_context *sw_context, - struct list_head *list) -{ - struct vmw_resource_val_node *val, *val_next; - - /* - * Drop references to resources held during command submission. - */ - - list_for_each_entry_safe(val, val_next, list, head) { - list_del_init(&val->head); - vmw_resource_unreference(&val->res); - - if (val->staged_bindings) { - if (val->staged_bindings != sw_context->staged_bindings) - vmw_binding_state_free(val->staged_bindings); - else - sw_context->staged_bindings_inuse = false; - val->staged_bindings = NULL; - } - - kfree(val); - } -} - -static void vmw_clear_validations(struct vmw_sw_context *sw_context) -{ - struct vmw_validate_buffer *entry, *next; - struct vmw_resource_val_node *val; - - /* - * Drop references to DMA buffers held during command submission. - */ - list_for_each_entry_safe(entry, next, &sw_context->validate_nodes, - base.head) { - list_del(&entry->base.head); - ttm_bo_unref(&entry->base.bo); - (void) drm_ht_remove_item(&sw_context->res_ht, &entry->hash); - sw_context->cur_val_buf--; - } - BUG_ON(sw_context->cur_val_buf != 0); - - list_for_each_entry(val, &sw_context->resource_list, head) - (void) drm_ht_remove_item(&sw_context->res_ht, &val->hash); -} - -int vmw_validate_single_buffer(struct vmw_private *dev_priv, - struct ttm_buffer_object *bo, - bool interruptible, - bool validate_as_mob) -{ - struct vmw_buffer_object *vbo = - container_of(bo, struct vmw_buffer_object, base); - struct ttm_operation_ctx ctx = { interruptible, false }; - int ret; - - if (vbo->pin_count > 0) - return 0; - - if (validate_as_mob) - return ttm_bo_validate(bo, &vmw_mob_placement, &ctx); - - /** - * Put BO in VRAM if there is space, otherwise as a GMR. - * If there is no space in VRAM and GMR ids are all used up, - * start evicting GMRs to make room. If the DMA buffer can't be - * used as a GMR, this will return -ENOMEM. - */ - - ret = ttm_bo_validate(bo, &vmw_vram_gmr_placement, &ctx); - if (likely(ret == 0 || ret == -ERESTARTSYS)) - return ret; - - /** - * If that failed, try VRAM again, this time evicting - * previous contents. - */ - - ret = ttm_bo_validate(bo, &vmw_vram_placement, &ctx); - return ret; -} - -static int vmw_validate_buffers(struct vmw_private *dev_priv, - struct vmw_sw_context *sw_context) -{ - struct vmw_validate_buffer *entry; - int ret; - - list_for_each_entry(entry, &sw_context->validate_nodes, base.head) { - ret = vmw_validate_single_buffer(dev_priv, entry->base.bo, - true, - entry->validate_as_mob); - if (unlikely(ret != 0)) - return ret; - } - return 0; -} - static int vmw_resize_cmd_bounce(struct vmw_sw_context *sw_context, uint32_t size) { @@ -3946,7 +3675,7 @@ static int vmw_execbuf_submit_fifo(struct vmw_private *dev_priv, if (sw_context->dx_ctx_node) cmd = vmw_fifo_reserve_dx(dev_priv, command_size, - sw_context->dx_ctx_node->res->id); + sw_context->dx_ctx_node->ctx->id); else cmd = vmw_fifo_reserve(dev_priv, command_size); if (!cmd) { @@ -3980,7 +3709,7 @@ static int vmw_execbuf_submit_cmdbuf(struct vmw_private *dev_priv, u32 command_size, struct vmw_sw_context *sw_context) { - u32 id = ((sw_context->dx_ctx_node) ? sw_context->dx_ctx_node->res->id : + u32 id = ((sw_context->dx_ctx_node) ? sw_context->dx_ctx_node->ctx->id : SVGA3D_INVALID_ID); void *cmd = vmw_cmdbuf_reserve(dev_priv->cman, command_size, id, false, header); @@ -4057,31 +3786,35 @@ static int vmw_execbuf_tie_context(struct vmw_private *dev_priv, struct vmw_sw_context *sw_context, uint32_t handle) { - struct vmw_resource_val_node *ctx_node; struct vmw_resource *res; int ret; + unsigned int size; if (handle == SVGA3D_INVALID_ID) return 0; - ret = vmw_user_resource_lookup_handle(dev_priv, sw_context->fp->tfile, - handle, user_context_converter, - &res); - if (unlikely(ret != 0)) { + size = vmw_execbuf_res_size(dev_priv, vmw_res_dx_context); + ret = vmw_validation_preload_res(sw_context->ctx, size); + if (ret) + return ret; + + res = vmw_user_resource_noref_lookup_handle + (dev_priv, sw_context->fp->tfile, handle, + user_context_converter); + if (unlikely(IS_ERR(res))) { DRM_ERROR("Could not find or user DX context 0x%08x.\n", (unsigned) handle); - return ret; + return PTR_ERR(res); } - ret = vmw_resource_val_add(sw_context, res, &ctx_node); + ret = vmw_execbuf_res_noref_val_add(sw_context, res); if (unlikely(ret != 0)) - goto out_err; + return ret; - sw_context->dx_ctx_node = ctx_node; + sw_context->dx_ctx_node = vmw_execbuf_info_from_res(sw_context, res); sw_context->man = vmw_context_res_man(res); -out_err: - vmw_resource_unreference(&res); - return ret; + + return 0; } int vmw_execbuf_process(struct drm_file *file_priv, @@ -4097,15 +3830,12 @@ int vmw_execbuf_process(struct drm_file *file_priv, { struct vmw_sw_context *sw_context = &dev_priv->ctx; struct vmw_fence_obj *fence = NULL; - struct vmw_resource *error_resource; - struct list_head resource_list; struct vmw_cmdbuf_header *header; - struct ww_acquire_ctx ticket; uint32_t handle; int ret; int32_t out_fence_fd = -1; struct sync_file *sync_file = NULL; - + DECLARE_VAL_CONTEXT(val_ctx, &sw_context->res_ht, 1); if (flags & DRM_VMW_EXECBUF_FLAG_EXPORT_FENCE_FD) { out_fence_fd = get_unused_fd_flags(O_CLOEXEC); @@ -4157,10 +3887,7 @@ int vmw_execbuf_process(struct drm_file *file_priv, sw_context->kernel = true; sw_context->fp = vmw_fpriv(file_priv); - sw_context->cur_reloc = 0; - sw_context->cur_val_buf = 0; - INIT_LIST_HEAD(&sw_context->resource_list); - INIT_LIST_HEAD(&sw_context->ctx_resource_list); + INIT_LIST_HEAD(&sw_context->ctx_list); sw_context->cur_query_bo = dev_priv->pinned_bo; sw_context->last_query_ctx = NULL; sw_context->needs_post_query_barrier = false; @@ -4168,8 +3895,8 @@ int vmw_execbuf_process(struct drm_file *file_priv, sw_context->dx_query_mob = NULL; sw_context->dx_query_ctx = NULL; memset(sw_context->res_cache, 0, sizeof(sw_context->res_cache)); - INIT_LIST_HEAD(&sw_context->validate_nodes); INIT_LIST_HEAD(&sw_context->res_relocations); + INIT_LIST_HEAD(&sw_context->bo_relocations); if (sw_context->staged_bindings) vmw_binding_state_reset(sw_context->staged_bindings); @@ -4180,24 +3907,13 @@ int vmw_execbuf_process(struct drm_file *file_priv, sw_context->res_ht_initialized = true; } INIT_LIST_HEAD(&sw_context->staged_cmd_res); - INIT_LIST_HEAD(&resource_list); + sw_context->ctx = &val_ctx; ret = vmw_execbuf_tie_context(dev_priv, sw_context, dx_context_handle); - if (unlikely(ret != 0)) { - list_splice_init(&sw_context->ctx_resource_list, - &sw_context->resource_list); + if (unlikely(ret != 0)) goto out_err_nores; - } ret = vmw_cmd_check_all(dev_priv, sw_context, kernel_commands, command_size); - /* - * Merge the resource lists before checking the return status - * from vmd_cmd_check_all so that all the open hashtabs will - * be handled properly even if vmw_cmd_check_all fails. - */ - list_splice_init(&sw_context->ctx_resource_list, - &sw_context->resource_list); - if (unlikely(ret != 0)) goto out_err_nores; @@ -4205,18 +3921,18 @@ int vmw_execbuf_process(struct drm_file *file_priv, if (unlikely(ret != 0)) goto out_err_nores; - ret = ttm_eu_reserve_buffers(&ticket, &sw_context->validate_nodes, - true, NULL); + ret = vmw_validation_bo_reserve(&val_ctx, true); if (unlikely(ret != 0)) goto out_err_nores; - ret = vmw_validate_buffers(dev_priv, sw_context); + ret = vmw_validation_bo_validate(&val_ctx, true); if (unlikely(ret != 0)) goto out_err; - ret = vmw_resources_validate(sw_context); + ret = vmw_validation_res_validate(&val_ctx, true); if (unlikely(ret != 0)) goto out_err; + vmw_validation_drop_ht(&val_ctx); ret = mutex_lock_interruptible(&dev_priv->binding_mutex); if (unlikely(ret != 0)) { @@ -4255,17 +3971,16 @@ int vmw_execbuf_process(struct drm_file *file_priv, if (ret != 0) DRM_ERROR("Fence submission error. Syncing.\n"); - vmw_resources_unreserve(sw_context, false); + vmw_execbuf_bindings_commit(sw_context, false); + vmw_bind_dx_query_mob(sw_context); + vmw_validation_res_unreserve(&val_ctx, false); - ttm_eu_fence_buffer_objects(&ticket, &sw_context->validate_nodes, - (void *) fence); + vmw_validation_bo_fence(sw_context->ctx, fence); if (unlikely(dev_priv->pinned_bo != NULL && !dev_priv->query_cid_valid)) __vmw_execbuf_release_pinned_bo(dev_priv, fence); - vmw_clear_validations(sw_context); - /* * If anything fails here, give up trying to export the fence * and do a sync since the user mode will not be able to sync @@ -4300,7 +4015,6 @@ int vmw_execbuf_process(struct drm_file *file_priv, vmw_fence_obj_unreference(&fence); } - list_splice_init(&sw_context->resource_list, &resource_list); vmw_cmdbuf_res_commit(&sw_context->staged_cmd_res); mutex_unlock(&dev_priv->cmdbuf_mutex); @@ -4308,36 +4022,33 @@ int vmw_execbuf_process(struct drm_file *file_priv, * Unreference resources outside of the cmdbuf_mutex to * avoid deadlocks in resource destruction paths. */ - vmw_resource_list_unreference(sw_context, &resource_list); + vmw_validation_unref_lists(&val_ctx); return 0; out_unlock_binding: mutex_unlock(&dev_priv->binding_mutex); out_err: - ttm_eu_backoff_reservation(&ticket, &sw_context->validate_nodes); + vmw_validation_bo_backoff(&val_ctx); out_err_nores: - vmw_resources_unreserve(sw_context, true); + vmw_execbuf_bindings_commit(sw_context, true); + vmw_validation_res_unreserve(&val_ctx, true); vmw_resource_relocations_free(&sw_context->res_relocations); vmw_free_relocations(sw_context); - vmw_clear_validations(sw_context); if (unlikely(dev_priv->pinned_bo != NULL && !dev_priv->query_cid_valid)) __vmw_execbuf_release_pinned_bo(dev_priv, NULL); out_unlock: - list_splice_init(&sw_context->resource_list, &resource_list); - error_resource = sw_context->error_resource; - sw_context->error_resource = NULL; vmw_cmdbuf_res_revert(&sw_context->staged_cmd_res); + vmw_validation_drop_ht(&val_ctx); + WARN_ON(!list_empty(&sw_context->ctx_list)); mutex_unlock(&dev_priv->cmdbuf_mutex); /* * Unreference resources outside of the cmdbuf_mutex to * avoid deadlocks in resource destruction paths. */ - vmw_resource_list_unreference(sw_context, &resource_list); - if (unlikely(error_resource != NULL)) - vmw_resource_unreference(&error_resource); + vmw_validation_unref_lists(&val_ctx); out_free_header: if (header) vmw_cmdbuf_header_free(header); @@ -4398,38 +4109,31 @@ void __vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv, struct vmw_fence_obj *fence) { int ret = 0; - struct list_head validate_list; - struct ttm_validate_buffer pinned_val, query_val; struct vmw_fence_obj *lfence = NULL; - struct ww_acquire_ctx ticket; + DECLARE_VAL_CONTEXT(val_ctx, NULL, 0); if (dev_priv->pinned_bo == NULL) goto out_unlock; - INIT_LIST_HEAD(&validate_list); - - pinned_val.bo = ttm_bo_reference(&dev_priv->pinned_bo->base); - pinned_val.shared = false; - list_add_tail(&pinned_val.head, &validate_list); + ret = vmw_validation_add_bo(&val_ctx, dev_priv->pinned_bo, false, + false); + if (ret) + goto out_no_reserve; - query_val.bo = ttm_bo_reference(&dev_priv->dummy_query_bo->base); - query_val.shared = false; - list_add_tail(&query_val.head, &validate_list); + ret = vmw_validation_add_bo(&val_ctx, dev_priv->dummy_query_bo, false, + false); + if (ret) + goto out_no_reserve; - ret = ttm_eu_reserve_buffers(&ticket, &validate_list, - false, NULL); - if (unlikely(ret != 0)) { - vmw_execbuf_unpin_panic(dev_priv); + ret = vmw_validation_bo_reserve(&val_ctx, false); + if (ret) goto out_no_reserve; - } if (dev_priv->query_cid_valid) { BUG_ON(fence != NULL); ret = vmw_fifo_emit_dummy_query(dev_priv, dev_priv->query_cid); - if (unlikely(ret != 0)) { - vmw_execbuf_unpin_panic(dev_priv); + if (ret) goto out_no_emit; - } dev_priv->query_cid_valid = false; } @@ -4443,22 +4147,22 @@ void __vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv, NULL); fence = lfence; } - ttm_eu_fence_buffer_objects(&ticket, &validate_list, (void *) fence); + vmw_validation_bo_fence(&val_ctx, fence); if (lfence != NULL) vmw_fence_obj_unreference(&lfence); - ttm_bo_unref(&query_val.bo); - ttm_bo_unref(&pinned_val.bo); + vmw_validation_unref_lists(&val_ctx); vmw_bo_unreference(&dev_priv->pinned_bo); out_unlock: return; out_no_emit: - ttm_eu_backoff_reservation(&ticket, &validate_list); + vmw_validation_bo_backoff(&val_ctx); out_no_reserve: - ttm_bo_unref(&query_val.bo); - ttm_bo_unref(&pinned_val.bo); + vmw_validation_unref_lists(&val_ctx); + vmw_execbuf_unpin_panic(dev_priv); vmw_bo_unreference(&dev_priv->pinned_bo); + } /** diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c index 3d546d409334..f87261545f2c 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c @@ -306,7 +306,8 @@ struct vmw_fence_manager *vmw_fence_manager_init(struct vmw_private *dev_priv) INIT_LIST_HEAD(&fman->cleanup_list); INIT_WORK(&fman->work, &vmw_fence_work_func); fman->fifo_down = true; - fman->user_fence_size = ttm_round_pot(sizeof(struct vmw_user_fence)); + fman->user_fence_size = ttm_round_pot(sizeof(struct vmw_user_fence)) + + TTM_OBJ_EXTRA_SIZE; fman->fence_size = ttm_round_pot(sizeof(struct vmw_fence_obj)); fman->event_fence_action_size = ttm_round_pot(sizeof(struct vmw_event_fence_action)); @@ -650,7 +651,7 @@ int vmw_user_fence_create(struct drm_file *file_priv, } *p_fence = &ufence->fence; - *p_handle = ufence->base.hash.key; + *p_handle = ufence->base.handle; return 0; out_err: @@ -1137,7 +1138,7 @@ int vmw_fence_event_ioctl(struct drm_device *dev, void *data, "object.\n"); goto out_no_ref_obj; } - handle = base->hash.key; + handle = base->handle; } ttm_base_object_unref(&base); } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c index c769df9e257b..e6b11f6ae2e4 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c @@ -2575,88 +2575,31 @@ int vmw_kms_helper_dirty(struct vmw_private *dev_priv, } /** - * vmw_kms_helper_buffer_prepare - Reserve and validate a buffer object before - * command submission. - * - * @dev_priv. Pointer to a device private structure. - * @buf: The buffer object - * @interruptible: Whether to perform waits as interruptible. - * @validate_as_mob: Whether the buffer should be validated as a MOB. If false, - * The buffer will be validated as a GMR. Already pinned buffers will not be - * validated. - * - * Returns 0 on success, negative error code on failure, -ERESTARTSYS if - * interrupted by a signal. + * vmw_kms_helper_validation_finish - Helper for post KMS command submission + * cleanup and fencing + * @dev_priv: Pointer to the device-private struct + * @file_priv: Pointer identifying the client when user-space fencing is used + * @ctx: Pointer to the validation context + * @out_fence: If non-NULL, returned refcounted fence-pointer + * @user_fence_rep: If non-NULL, pointer to user-space address area + * in which to copy user-space fence info */ -int vmw_kms_helper_buffer_prepare(struct vmw_private *dev_priv, - struct vmw_buffer_object *buf, - bool interruptible, - bool validate_as_mob, - bool for_cpu_blit) -{ - struct ttm_operation_ctx ctx = { - .interruptible = interruptible, - .no_wait_gpu = false}; - struct ttm_buffer_object *bo = &buf->base; - int ret; - - ttm_bo_reserve(bo, false, false, NULL); - if (for_cpu_blit) - ret = ttm_bo_validate(bo, &vmw_nonfixed_placement, &ctx); - else - ret = vmw_validate_single_buffer(dev_priv, bo, interruptible, - validate_as_mob); - if (ret) - ttm_bo_unreserve(bo); - - return ret; -} - -/** - * vmw_kms_helper_buffer_revert - Undo the actions of - * vmw_kms_helper_buffer_prepare. - * - * @res: Pointer to the buffer object. - * - * Helper to be used if an error forces the caller to undo the actions of - * vmw_kms_helper_buffer_prepare. - */ -void vmw_kms_helper_buffer_revert(struct vmw_buffer_object *buf) -{ - if (buf) - ttm_bo_unreserve(&buf->base); -} - -/** - * vmw_kms_helper_buffer_finish - Unreserve and fence a buffer object after - * kms command submission. - * - * @dev_priv: Pointer to a device private structure. - * @file_priv: Pointer to a struct drm_file representing the caller's - * connection. Must be set to NULL if @user_fence_rep is NULL, and conversely - * if non-NULL, @user_fence_rep must be non-NULL. - * @buf: The buffer object. - * @out_fence: Optional pointer to a fence pointer. If non-NULL, a - * ref-counted fence pointer is returned here. - * @user_fence_rep: Optional pointer to a user-space provided struct - * drm_vmw_fence_rep. If provided, @file_priv must also be provided and the - * function copies fence data to user-space in a fail-safe manner. - */ -void vmw_kms_helper_buffer_finish(struct vmw_private *dev_priv, - struct drm_file *file_priv, - struct vmw_buffer_object *buf, - struct vmw_fence_obj **out_fence, - struct drm_vmw_fence_rep __user * - user_fence_rep) -{ - struct vmw_fence_obj *fence; +void vmw_kms_helper_validation_finish(struct vmw_private *dev_priv, + struct drm_file *file_priv, + struct vmw_validation_context *ctx, + struct vmw_fence_obj **out_fence, + struct drm_vmw_fence_rep __user * + user_fence_rep) +{ + struct vmw_fence_obj *fence = NULL; uint32_t handle; int ret; - ret = vmw_execbuf_fence_commands(file_priv, dev_priv, &fence, - file_priv ? &handle : NULL); - if (buf) - vmw_bo_fence_single(&buf->base, fence); + if (file_priv || user_fence_rep || vmw_validation_has_bos(ctx) || + out_fence) + ret = vmw_execbuf_fence_commands(file_priv, dev_priv, &fence, + file_priv ? &handle : NULL); + vmw_validation_done(ctx, fence); if (file_priv) vmw_execbuf_copy_fence_user(dev_priv, vmw_fpriv(file_priv), ret, user_fence_rep, fence, @@ -2665,106 +2608,6 @@ void vmw_kms_helper_buffer_finish(struct vmw_private *dev_priv, *out_fence = fence; else vmw_fence_obj_unreference(&fence); - - vmw_kms_helper_buffer_revert(buf); -} - - -/** - * vmw_kms_helper_resource_revert - Undo the actions of - * vmw_kms_helper_resource_prepare. - * - * @res: Pointer to the resource. Typically a surface. - * - * Helper to be used if an error forces the caller to undo the actions of - * vmw_kms_helper_resource_prepare. - */ -void vmw_kms_helper_resource_revert(struct vmw_validation_ctx *ctx) -{ - struct vmw_resource *res = ctx->res; - - vmw_kms_helper_buffer_revert(ctx->buf); - vmw_bo_unreference(&ctx->buf); - vmw_resource_unreserve(res, false, NULL, 0); - mutex_unlock(&res->dev_priv->cmdbuf_mutex); -} - -/** - * vmw_kms_helper_resource_prepare - Reserve and validate a resource before - * command submission. - * - * @res: Pointer to the resource. Typically a surface. - * @interruptible: Whether to perform waits as interruptible. - * - * Reserves and validates also the backup buffer if a guest-backed resource. - * Returns 0 on success, negative error code on failure. -ERESTARTSYS if - * interrupted by a signal. - */ -int vmw_kms_helper_resource_prepare(struct vmw_resource *res, - bool interruptible, - struct vmw_validation_ctx *ctx) -{ - int ret = 0; - - ctx->buf = NULL; - ctx->res = res; - - if (interruptible) - ret = mutex_lock_interruptible(&res->dev_priv->cmdbuf_mutex); - else - mutex_lock(&res->dev_priv->cmdbuf_mutex); - - if (unlikely(ret != 0)) - return -ERESTARTSYS; - - ret = vmw_resource_reserve(res, interruptible, false); - if (ret) - goto out_unlock; - - if (res->backup) { - ret = vmw_kms_helper_buffer_prepare(res->dev_priv, res->backup, - interruptible, - res->dev_priv->has_mob, - false); - if (ret) - goto out_unreserve; - - ctx->buf = vmw_bo_reference(res->backup); - } - ret = vmw_resource_validate(res); - if (ret) - goto out_revert; - return 0; - -out_revert: - vmw_kms_helper_buffer_revert(ctx->buf); -out_unreserve: - vmw_resource_unreserve(res, false, NULL, 0); -out_unlock: - mutex_unlock(&res->dev_priv->cmdbuf_mutex); - return ret; -} - -/** - * vmw_kms_helper_resource_finish - Unreserve and fence a resource after - * kms command submission. - * - * @res: Pointer to the resource. Typically a surface. - * @out_fence: Optional pointer to a fence pointer. If non-NULL, a - * ref-counted fence pointer is returned here. - */ -void vmw_kms_helper_resource_finish(struct vmw_validation_ctx *ctx, - struct vmw_fence_obj **out_fence) -{ - struct vmw_resource *res = ctx->res; - - if (ctx->buf || out_fence) - vmw_kms_helper_buffer_finish(res->dev_priv, NULL, ctx->buf, - out_fence, NULL); - - vmw_bo_unreference(&ctx->buf); - vmw_resource_unreserve(res, false, NULL, 0); - mutex_unlock(&res->dev_priv->cmdbuf_mutex); } /** diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h index 31311298ec0b..76ec570c0684 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h @@ -308,24 +308,12 @@ int vmw_kms_helper_dirty(struct vmw_private *dev_priv, int increment, struct vmw_kms_dirty *dirty); -int vmw_kms_helper_buffer_prepare(struct vmw_private *dev_priv, - struct vmw_buffer_object *buf, - bool interruptible, - bool validate_as_mob, - bool for_cpu_blit); -void vmw_kms_helper_buffer_revert(struct vmw_buffer_object *buf); -void vmw_kms_helper_buffer_finish(struct vmw_private *dev_priv, - struct drm_file *file_priv, - struct vmw_buffer_object *buf, - struct vmw_fence_obj **out_fence, - struct drm_vmw_fence_rep __user * - user_fence_rep); -int vmw_kms_helper_resource_prepare(struct vmw_resource *res, - bool interruptible, - struct vmw_validation_ctx *ctx); -void vmw_kms_helper_resource_revert(struct vmw_validation_ctx *ctx); -void vmw_kms_helper_resource_finish(struct vmw_validation_ctx *ctx, - struct vmw_fence_obj **out_fence); +void vmw_kms_helper_validation_finish(struct vmw_private *dev_priv, + struct drm_file *file_priv, + struct vmw_validation_context *ctx, + struct vmw_fence_obj **out_fence, + struct drm_vmw_fence_rep __user * + user_fence_rep); int vmw_kms_readback(struct vmw_private *dev_priv, struct drm_file *file_priv, struct vmw_framebuffer *vfb, diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_prime.c b/drivers/gpu/drm/vmwgfx/vmwgfx_prime.c index 0861c821a7fe..e420675e8db3 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_prime.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_prime.c @@ -31,8 +31,8 @@ */ #include "vmwgfx_drv.h" +#include "ttm_object.h" #include <linux/dma-buf.h> -#include <drm/ttm/ttm_object.h> /* * DMA-BUF attach- and mapping methods. No need to implement diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c index 92003ea5a219..8a029bade32a 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c @@ -58,11 +58,11 @@ void vmw_resource_release_id(struct vmw_resource *res) struct vmw_private *dev_priv = res->dev_priv; struct idr *idr = &dev_priv->res_idr[res->func->res_type]; - write_lock(&dev_priv->resource_lock); + spin_lock(&dev_priv->resource_lock); if (res->id != -1) idr_remove(idr, res->id); res->id = -1; - write_unlock(&dev_priv->resource_lock); + spin_unlock(&dev_priv->resource_lock); } static void vmw_resource_release(struct kref *kref) @@ -73,10 +73,9 @@ static void vmw_resource_release(struct kref *kref) int id; struct idr *idr = &dev_priv->res_idr[res->func->res_type]; - write_lock(&dev_priv->resource_lock); - res->avail = false; + spin_lock(&dev_priv->resource_lock); list_del_init(&res->lru_head); - write_unlock(&dev_priv->resource_lock); + spin_unlock(&dev_priv->resource_lock); if (res->backup) { struct ttm_buffer_object *bo = &res->backup->base; @@ -108,10 +107,10 @@ static void vmw_resource_release(struct kref *kref) else kfree(res); - write_lock(&dev_priv->resource_lock); + spin_lock(&dev_priv->resource_lock); if (id != -1) idr_remove(idr, id); - write_unlock(&dev_priv->resource_lock); + spin_unlock(&dev_priv->resource_lock); } void vmw_resource_unreference(struct vmw_resource **p_res) @@ -140,13 +139,13 @@ int vmw_resource_alloc_id(struct vmw_resource *res) BUG_ON(res->id != -1); idr_preload(GFP_KERNEL); - write_lock(&dev_priv->resource_lock); + spin_lock(&dev_priv->resource_lock); ret = idr_alloc(idr, res, 1, 0, GFP_NOWAIT); if (ret >= 0) res->id = ret; - write_unlock(&dev_priv->resource_lock); + spin_unlock(&dev_priv->resource_lock); idr_preload_end(); return ret < 0 ? ret : 0; } @@ -170,7 +169,6 @@ int vmw_resource_init(struct vmw_private *dev_priv, struct vmw_resource *res, kref_init(&res->kref); res->hw_destroy = NULL; res->res_free = res_free; - res->avail = false; res->dev_priv = dev_priv; res->func = func; INIT_LIST_HEAD(&res->lru_head); @@ -187,28 +185,6 @@ int vmw_resource_init(struct vmw_private *dev_priv, struct vmw_resource *res, return vmw_resource_alloc_id(res); } -/** - * vmw_resource_activate - * - * @res: Pointer to the newly created resource - * @hw_destroy: Destroy function. NULL if none. - * - * Activate a resource after the hardware has been made aware of it. - * Set tye destroy function to @destroy. Typically this frees the - * resource and destroys the hardware resources associated with it. - * Activate basically means that the function vmw_resource_lookup will - * find it. - */ -void vmw_resource_activate(struct vmw_resource *res, - void (*hw_destroy) (struct vmw_resource *)) -{ - struct vmw_private *dev_priv = res->dev_priv; - - write_lock(&dev_priv->resource_lock); - res->avail = true; - res->hw_destroy = hw_destroy; - write_unlock(&dev_priv->resource_lock); -} /** * vmw_user_resource_lookup_handle - lookup a struct resource from a @@ -243,15 +219,7 @@ int vmw_user_resource_lookup_handle(struct vmw_private *dev_priv, goto out_bad_resource; res = converter->base_obj_to_res(base); - - read_lock(&dev_priv->resource_lock); - if (!res->avail || res->res_free != converter->res_free) { - read_unlock(&dev_priv->resource_lock); - goto out_bad_resource; - } - kref_get(&res->kref); - read_unlock(&dev_priv->resource_lock); *p_res = res; ret = 0; @@ -263,6 +231,41 @@ out_bad_resource: } /** + * vmw_user_resource_lookup_handle - lookup a struct resource from a + * TTM user-space handle and perform basic type checks + * + * @dev_priv: Pointer to a device private struct + * @tfile: Pointer to a struct ttm_object_file identifying the caller + * @handle: The TTM user-space handle + * @converter: Pointer to an object describing the resource type + * @p_res: On successful return the location pointed to will contain + * a pointer to a refcounted struct vmw_resource. + * + * If the handle can't be found or is associated with an incorrect resource + * type, -EINVAL will be returned. + */ +struct vmw_resource * +vmw_user_resource_noref_lookup_handle(struct vmw_private *dev_priv, + struct ttm_object_file *tfile, + uint32_t handle, + const struct vmw_user_resource_conv + *converter) +{ + struct ttm_base_object *base; + + base = ttm_base_object_noref_lookup(tfile, handle); + if (!base) + return ERR_PTR(-ESRCH); + + if (unlikely(ttm_base_object_type(base) != converter->object_type)) { + ttm_base_object_noref_release(); + return ERR_PTR(-EINVAL); + } + + return converter->base_obj_to_res(base); +} + +/** * Helper function that looks either a surface or bo. * * The pointer this pointed at by out_surf and out_buf needs to be null. @@ -422,10 +425,10 @@ void vmw_resource_unreserve(struct vmw_resource *res, if (!res->func->may_evict || res->id == -1 || res->pin_count) return; - write_lock(&dev_priv->resource_lock); + spin_lock(&dev_priv->resource_lock); list_add_tail(&res->lru_head, &res->dev_priv->res_lru[res->func->res_type]); - write_unlock(&dev_priv->resource_lock); + spin_unlock(&dev_priv->resource_lock); } /** @@ -504,9 +507,9 @@ int vmw_resource_reserve(struct vmw_resource *res, bool interruptible, struct vmw_private *dev_priv = res->dev_priv; int ret; - write_lock(&dev_priv->resource_lock); + spin_lock(&dev_priv->resource_lock); list_del_init(&res->lru_head); - write_unlock(&dev_priv->resource_lock); + spin_unlock(&dev_priv->resource_lock); if (res->func->needs_backup && res->backup == NULL && !no_backup) { @@ -587,15 +590,18 @@ out_no_unbind: /** * vmw_resource_validate - Make a resource up-to-date and visible * to the device. - * - * @res: The resource to make visible to the device. + * @res: The resource to make visible to the device. + * @intr: Perform waits interruptible if possible. * * On succesful return, any backup DMA buffer pointed to by @res->backup will * be reserved and validated. * On hardware resource shortage, this function will repeatedly evict * resources of the same type until the validation succeeds. + * + * Return: Zero on success, -ERESTARTSYS if interrupted, negative error code + * on failure. */ -int vmw_resource_validate(struct vmw_resource *res) +int vmw_resource_validate(struct vmw_resource *res, bool intr) { int ret; struct vmw_resource *evict_res; @@ -616,12 +622,12 @@ int vmw_resource_validate(struct vmw_resource *res) if (likely(ret != -EBUSY)) break; - write_lock(&dev_priv->resource_lock); + spin_lock(&dev_priv->resource_lock); if (list_empty(lru_list) || !res->func->may_evict) { DRM_ERROR("Out of device device resources " "for %s.\n", res->func->type_name); ret = -EBUSY; - write_unlock(&dev_priv->resource_lock); + spin_unlock(&dev_priv->resource_lock); break; } @@ -630,14 +636,14 @@ int vmw_resource_validate(struct vmw_resource *res) lru_head)); list_del_init(&evict_res->lru_head); - write_unlock(&dev_priv->resource_lock); + spin_unlock(&dev_priv->resource_lock); /* Trylock backup buffers with a NULL ticket. */ - ret = vmw_resource_do_evict(NULL, evict_res, true); + ret = vmw_resource_do_evict(NULL, evict_res, intr); if (unlikely(ret != 0)) { - write_lock(&dev_priv->resource_lock); + spin_lock(&dev_priv->resource_lock); list_add_tail(&evict_res->lru_head, lru_list); - write_unlock(&dev_priv->resource_lock); + spin_unlock(&dev_priv->resource_lock); if (ret == -ERESTARTSYS || ++err_count > VMW_RES_EVICT_ERR_COUNT) { vmw_resource_unreference(&evict_res); @@ -819,7 +825,7 @@ static void vmw_resource_evict_type(struct vmw_private *dev_priv, struct ww_acquire_ctx ticket; do { - write_lock(&dev_priv->resource_lock); + spin_lock(&dev_priv->resource_lock); if (list_empty(lru_list)) goto out_unlock; @@ -828,14 +834,14 @@ static void vmw_resource_evict_type(struct vmw_private *dev_priv, list_first_entry(lru_list, struct vmw_resource, lru_head)); list_del_init(&evict_res->lru_head); - write_unlock(&dev_priv->resource_lock); + spin_unlock(&dev_priv->resource_lock); /* Wait lock backup buffers with a ticket. */ ret = vmw_resource_do_evict(&ticket, evict_res, false); if (unlikely(ret != 0)) { - write_lock(&dev_priv->resource_lock); + spin_lock(&dev_priv->resource_lock); list_add_tail(&evict_res->lru_head, lru_list); - write_unlock(&dev_priv->resource_lock); + spin_unlock(&dev_priv->resource_lock); if (++err_count > VMW_RES_EVICT_ERR_COUNT) { vmw_resource_unreference(&evict_res); return; @@ -846,7 +852,7 @@ static void vmw_resource_evict_type(struct vmw_private *dev_priv, } while (1); out_unlock: - write_unlock(&dev_priv->resource_lock); + spin_unlock(&dev_priv->resource_lock); } /** @@ -914,7 +920,7 @@ int vmw_resource_pin(struct vmw_resource *res, bool interruptible) /* Do we really need to pin the MOB as well? */ vmw_bo_pin_reserved(vbo, true); } - ret = vmw_resource_validate(res); + ret = vmw_resource_validate(res, interruptible); if (vbo) ttm_bo_unreserve(&vbo->base); if (ret) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource_priv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_resource_priv.h index a8c1c5ebd71d..7e19eba0b0b8 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource_priv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource_priv.h @@ -30,6 +30,11 @@ #include "vmwgfx_drv.h" +/* + * Extra memory required by the resource id's ida storage, which is allocated + * separately from the base object itself. We estimate an on-average 128 bytes + * per ida. + */ #define VMW_IDA_ACC_SIZE 128 enum vmw_cmdbuf_res_state { @@ -120,8 +125,6 @@ int vmw_resource_init(struct vmw_private *dev_priv, struct vmw_resource *res, bool delay_id, void (*res_free) (struct vmw_resource *res), const struct vmw_res_func *func); -void vmw_resource_activate(struct vmw_resource *res, - void (*hw_destroy) (struct vmw_resource *)); int vmw_simple_resource_create_ioctl(struct drm_device *dev, void *data, diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c b/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c index 4c68ad6f3605..333418dc259f 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c @@ -945,16 +945,20 @@ int vmw_kms_sou_do_surface_dirty(struct vmw_private *dev_priv, struct vmw_framebuffer_surface *vfbs = container_of(framebuffer, typeof(*vfbs), base); struct vmw_kms_sou_surface_dirty sdirty; - struct vmw_validation_ctx ctx; + DECLARE_VAL_CONTEXT(val_ctx, NULL, 0); int ret; if (!srf) srf = &vfbs->surface->res; - ret = vmw_kms_helper_resource_prepare(srf, true, &ctx); + ret = vmw_validation_add_resource(&val_ctx, srf, 0, NULL, NULL); if (ret) return ret; + ret = vmw_validation_prepare(&val_ctx, &dev_priv->cmdbuf_mutex, true); + if (ret) + goto out_unref; + sdirty.base.fifo_commit = vmw_sou_surface_fifo_commit; sdirty.base.clip = vmw_sou_surface_clip; sdirty.base.dev_priv = dev_priv; @@ -971,9 +975,14 @@ int vmw_kms_sou_do_surface_dirty(struct vmw_private *dev_priv, ret = vmw_kms_helper_dirty(dev_priv, framebuffer, clips, vclips, dest_x, dest_y, num_clips, inc, &sdirty.base); - vmw_kms_helper_resource_finish(&ctx, out_fence); + vmw_kms_helper_validation_finish(dev_priv, NULL, &val_ctx, out_fence, + NULL); return ret; + +out_unref: + vmw_validation_unref_lists(&val_ctx); + return ret; } /** @@ -1050,13 +1059,17 @@ int vmw_kms_sou_do_bo_dirty(struct vmw_private *dev_priv, container_of(framebuffer, struct vmw_framebuffer_bo, base)->buffer; struct vmw_kms_dirty dirty; + DECLARE_VAL_CONTEXT(val_ctx, NULL, 0); int ret; - ret = vmw_kms_helper_buffer_prepare(dev_priv, buf, interruptible, - false, false); + ret = vmw_validation_add_bo(&val_ctx, buf, false, false); if (ret) return ret; + ret = vmw_validation_prepare(&val_ctx, NULL, interruptible); + if (ret) + goto out_unref; + ret = do_bo_define_gmrfb(dev_priv, framebuffer); if (unlikely(ret != 0)) goto out_revert; @@ -1068,12 +1081,15 @@ int vmw_kms_sou_do_bo_dirty(struct vmw_private *dev_priv, num_clips; ret = vmw_kms_helper_dirty(dev_priv, framebuffer, clips, vclips, 0, 0, num_clips, increment, &dirty); - vmw_kms_helper_buffer_finish(dev_priv, NULL, buf, out_fence, NULL); + vmw_kms_helper_validation_finish(dev_priv, NULL, &val_ctx, out_fence, + NULL); return ret; out_revert: - vmw_kms_helper_buffer_revert(buf); + vmw_validation_revert(&val_ctx); +out_unref: + vmw_validation_unref_lists(&val_ctx); return ret; } @@ -1149,13 +1165,17 @@ int vmw_kms_sou_readback(struct vmw_private *dev_priv, struct vmw_buffer_object *buf = container_of(vfb, struct vmw_framebuffer_bo, base)->buffer; struct vmw_kms_dirty dirty; + DECLARE_VAL_CONTEXT(val_ctx, NULL, 0); int ret; - ret = vmw_kms_helper_buffer_prepare(dev_priv, buf, true, false, - false); + ret = vmw_validation_add_bo(&val_ctx, buf, false, false); if (ret) return ret; + ret = vmw_validation_prepare(&val_ctx, NULL, true); + if (ret) + goto out_unref; + ret = do_bo_define_gmrfb(dev_priv, vfb); if (unlikely(ret != 0)) goto out_revert; @@ -1167,13 +1187,15 @@ int vmw_kms_sou_readback(struct vmw_private *dev_priv, num_clips; ret = vmw_kms_helper_dirty(dev_priv, vfb, NULL, vclips, 0, 0, num_clips, 1, &dirty); - vmw_kms_helper_buffer_finish(dev_priv, file_priv, buf, NULL, - user_fence_rep); + vmw_kms_helper_validation_finish(dev_priv, file_priv, &val_ctx, NULL, + user_fence_rep); return ret; out_revert: - vmw_kms_helper_buffer_revert(buf); - + vmw_validation_revert(&val_ctx); +out_unref: + vmw_validation_unref_lists(&val_ctx); + return ret; } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c b/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c index fe4842ca3b6e..bf32fe446219 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c @@ -186,7 +186,7 @@ static int vmw_gb_shader_init(struct vmw_private *dev_priv, shader->num_input_sig = num_input_sig; shader->num_output_sig = num_output_sig; - vmw_resource_activate(res, vmw_hw_shader_destroy); + res->hw_destroy = vmw_hw_shader_destroy; return 0; } @@ -562,7 +562,7 @@ void vmw_dx_shader_cotable_list_scrub(struct vmw_private *dev_priv, { struct vmw_dx_shader *entry, *next; - WARN_ON_ONCE(!mutex_is_locked(&dev_priv->binding_mutex)); + lockdep_assert_held_once(&dev_priv->binding_mutex); list_for_each_entry_safe(entry, next, list, cotable_head) { WARN_ON(vmw_dx_shader_scrub(&entry->res)); @@ -636,7 +636,8 @@ int vmw_dx_shader_add(struct vmw_cmdbuf_res_manager *man, res = &shader->res; shader->ctx = ctx; - shader->cotable = vmw_context_cotable(ctx, SVGA_COTABLE_DXSHADER); + shader->cotable = vmw_resource_reference + (vmw_context_cotable(ctx, SVGA_COTABLE_DXSHADER)); shader->id = user_key; shader->committed = false; INIT_LIST_HEAD(&shader->cotable_head); @@ -656,7 +657,7 @@ int vmw_dx_shader_add(struct vmw_cmdbuf_res_manager *man, goto out_resource_init; res->id = shader->id; - vmw_resource_activate(res, vmw_hw_shader_destroy); + res->hw_destroy = vmw_hw_shader_destroy; out_resource_init: vmw_resource_unreference(&res); @@ -740,13 +741,10 @@ static int vmw_user_shader_alloc(struct vmw_private *dev_priv, }; int ret; - /* - * Approximate idr memory usage with 128 bytes. It will be limited - * by maximum number_of shaders anyway. - */ if (unlikely(vmw_user_shader_size == 0)) vmw_user_shader_size = - ttm_round_pot(sizeof(struct vmw_user_shader)) + 128; + ttm_round_pot(sizeof(struct vmw_user_shader)) + + VMW_IDA_ACC_SIZE + TTM_OBJ_EXTRA_SIZE; ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv), vmw_user_shader_size, @@ -792,7 +790,7 @@ static int vmw_user_shader_alloc(struct vmw_private *dev_priv, } if (handle) - *handle = ushader->base.hash.key; + *handle = ushader->base.handle; out_err: vmw_resource_unreference(&res); out: @@ -814,13 +812,10 @@ static struct vmw_resource *vmw_shader_alloc(struct vmw_private *dev_priv, }; int ret; - /* - * Approximate idr memory usage with 128 bytes. It will be limited - * by maximum number_of shaders anyway. - */ if (unlikely(vmw_shader_size == 0)) vmw_shader_size = - ttm_round_pot(sizeof(struct vmw_shader)) + 128; + ttm_round_pot(sizeof(struct vmw_shader)) + + VMW_IDA_ACC_SIZE; ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv), vmw_shader_size, diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_simple_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_simple_resource.c index 6ebc5affde14..6a6865384e91 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_simple_resource.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_simple_resource.c @@ -81,7 +81,7 @@ static int vmw_simple_resource_init(struct vmw_private *dev_priv, return ret; } - vmw_resource_activate(&simple->res, simple->func->hw_destroy); + simple->res.hw_destroy = simple->func->hw_destroy; return 0; } @@ -159,7 +159,8 @@ vmw_simple_resource_create_ioctl(struct drm_device *dev, void *data, alloc_size = offsetof(struct vmw_user_simple_resource, simple) + func->size; - account_size = ttm_round_pot(alloc_size) + VMW_IDA_ACC_SIZE; + account_size = ttm_round_pot(alloc_size) + VMW_IDA_ACC_SIZE + + TTM_OBJ_EXTRA_SIZE; ret = ttm_read_lock(&dev_priv->reservation_sem, true); if (ret) @@ -208,7 +209,7 @@ vmw_simple_resource_create_ioctl(struct drm_device *dev, void *data, goto out_err; } - func->set_arg_handle(data, usimple->base.hash.key); + func->set_arg_handle(data, usimple->base.handle); out_err: vmw_resource_unreference(&res); out_ret: diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_so.c b/drivers/gpu/drm/vmwgfx/vmwgfx_so.c index e9b6b7baa009..bc8bb690f1ea 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_so.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_so.c @@ -208,7 +208,7 @@ static int vmw_view_destroy(struct vmw_resource *res) union vmw_view_destroy body; } *cmd; - WARN_ON_ONCE(!mutex_is_locked(&dev_priv->binding_mutex)); + lockdep_assert_held_once(&dev_priv->binding_mutex); vmw_binding_res_list_scrub(&res->binding_head); if (!view->committed || res->id == -1) @@ -366,7 +366,8 @@ int vmw_view_add(struct vmw_cmdbuf_res_manager *man, res = &view->res; view->ctx = ctx; view->srf = vmw_resource_reference(srf); - view->cotable = vmw_context_cotable(ctx, vmw_view_cotables[view_type]); + view->cotable = vmw_resource_reference + (vmw_context_cotable(ctx, vmw_view_cotables[view_type])); view->view_type = view_type; view->view_id = user_key; view->cmd_size = cmd_size; @@ -386,7 +387,7 @@ int vmw_view_add(struct vmw_cmdbuf_res_manager *man, goto out_resource_init; res->id = view->view_id; - vmw_resource_activate(res, vmw_hw_view_destroy); + res->hw_destroy = vmw_hw_view_destroy; out_resource_init: vmw_resource_unreference(&res); @@ -439,7 +440,7 @@ void vmw_view_cotable_list_destroy(struct vmw_private *dev_priv, { struct vmw_view *entry, *next; - WARN_ON_ONCE(!mutex_is_locked(&dev_priv->binding_mutex)); + lockdep_assert_held_once(&dev_priv->binding_mutex); list_for_each_entry_safe(entry, next, list, cotable_head) WARN_ON(vmw_view_destroy(&entry->res)); @@ -459,7 +460,7 @@ void vmw_view_surface_list_destroy(struct vmw_private *dev_priv, { struct vmw_view *entry, *next; - WARN_ON_ONCE(!mutex_is_locked(&dev_priv->binding_mutex)); + lockdep_assert_held_once(&dev_priv->binding_mutex); list_for_each_entry_safe(entry, next, list, srf_head) WARN_ON(vmw_view_destroy(&entry->res)); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c index e28bb08114a5..c3e435f444c1 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c @@ -759,17 +759,21 @@ int vmw_kms_stdu_dma(struct vmw_private *dev_priv, struct vmw_stdu_dirty ddirty; int ret; bool cpu_blit = !(dev_priv->capabilities & SVGA_CAP_3D); + DECLARE_VAL_CONTEXT(val_ctx, NULL, 0); /* * VMs without 3D support don't have the surface DMA command and * we'll be using a CPU blit, and the framebuffer should be moved out * of VRAM. */ - ret = vmw_kms_helper_buffer_prepare(dev_priv, buf, interruptible, - false, cpu_blit); + ret = vmw_validation_add_bo(&val_ctx, buf, false, cpu_blit); if (ret) return ret; + ret = vmw_validation_prepare(&val_ctx, NULL, interruptible); + if (ret) + goto out_unref; + ddirty.transfer = (to_surface) ? SVGA3D_WRITE_HOST_VRAM : SVGA3D_READ_HOST_VRAM; ddirty.left = ddirty.top = S32_MAX; @@ -796,9 +800,13 @@ int vmw_kms_stdu_dma(struct vmw_private *dev_priv, ret = vmw_kms_helper_dirty(dev_priv, vfb, clips, vclips, 0, 0, num_clips, increment, &ddirty.base); - vmw_kms_helper_buffer_finish(dev_priv, file_priv, buf, NULL, - user_fence_rep); + vmw_kms_helper_validation_finish(dev_priv, file_priv, &val_ctx, NULL, + user_fence_rep); + return ret; + +out_unref: + vmw_validation_unref_lists(&val_ctx); return ret; } @@ -924,16 +932,20 @@ int vmw_kms_stdu_surface_dirty(struct vmw_private *dev_priv, struct vmw_framebuffer_surface *vfbs = container_of(framebuffer, typeof(*vfbs), base); struct vmw_stdu_dirty sdirty; - struct vmw_validation_ctx ctx; + DECLARE_VAL_CONTEXT(val_ctx, NULL, 0); int ret; if (!srf) srf = &vfbs->surface->res; - ret = vmw_kms_helper_resource_prepare(srf, true, &ctx); + ret = vmw_validation_add_resource(&val_ctx, srf, 0, NULL, NULL); if (ret) return ret; + ret = vmw_validation_prepare(&val_ctx, &dev_priv->cmdbuf_mutex, true); + if (ret) + goto out_unref; + if (vfbs->is_bo_proxy) { ret = vmw_kms_update_proxy(srf, clips, num_clips, inc); if (ret) @@ -954,8 +966,13 @@ int vmw_kms_stdu_surface_dirty(struct vmw_private *dev_priv, dest_x, dest_y, num_clips, inc, &sdirty.base); out_finish: - vmw_kms_helper_resource_finish(&ctx, out_fence); + vmw_kms_helper_validation_finish(dev_priv, NULL, &val_ctx, out_fence, + NULL); + + return ret; +out_unref: + vmw_validation_unref_lists(&val_ctx); return ret; } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c b/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c index 80a01cd4c051..ef09f7edf931 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c @@ -614,7 +614,7 @@ static int vmw_surface_init(struct vmw_private *dev_priv, */ INIT_LIST_HEAD(&srf->view_list); - vmw_resource_activate(res, vmw_hw_surface_destroy); + res->hw_destroy = vmw_hw_surface_destroy; return ret; } @@ -731,7 +731,7 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data, if (unlikely(vmw_user_surface_size == 0)) vmw_user_surface_size = ttm_round_pot(sizeof(*user_srf)) + - 128; + VMW_IDA_ACC_SIZE + TTM_OBJ_EXTRA_SIZE; num_sizes = 0; for (i = 0; i < DRM_VMW_MAX_SURFACE_FACES; ++i) { @@ -744,7 +744,7 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data, num_sizes == 0) return -EINVAL; - size = vmw_user_surface_size + 128 + + size = vmw_user_surface_size + ttm_round_pot(num_sizes * sizeof(struct drm_vmw_size)) + ttm_round_pot(num_sizes * sizeof(struct vmw_surface_offset)); @@ -886,7 +886,7 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data, goto out_unlock; } - rep->sid = user_srf->prime.base.hash.key; + rep->sid = user_srf->prime.base.handle; vmw_resource_unreference(&res); ttm_read_unlock(&dev_priv->reservation_sem); @@ -1024,7 +1024,7 @@ int vmw_surface_reference_ioctl(struct drm_device *dev, void *data, if (unlikely(ret != 0)) { DRM_ERROR("copy_to_user failed %p %u\n", user_sizes, srf->num_sizes); - ttm_ref_object_base_unref(tfile, base->hash.key, TTM_REF_USAGE); + ttm_ref_object_base_unref(tfile, base->handle, TTM_REF_USAGE); ret = -EFAULT; } @@ -1613,9 +1613,9 @@ vmw_gb_surface_define_internal(struct drm_device *dev, if (unlikely(vmw_user_surface_size == 0)) vmw_user_surface_size = ttm_round_pot(sizeof(*user_srf)) + - 128; + VMW_IDA_ACC_SIZE + TTM_OBJ_EXTRA_SIZE; - size = vmw_user_surface_size + 128; + size = vmw_user_surface_size; /* Define a surface based on the parameters. */ ret = vmw_surface_gb_priv_define(dev, @@ -1687,7 +1687,7 @@ vmw_gb_surface_define_internal(struct drm_device *dev, goto out_unlock; } - rep->handle = user_srf->prime.base.hash.key; + rep->handle = user_srf->prime.base.handle; rep->backup_size = res->backup_size; if (res->backup) { rep->buffer_map_handle = @@ -1749,7 +1749,7 @@ vmw_gb_surface_reference_internal(struct drm_device *dev, if (unlikely(ret != 0)) { DRM_ERROR("Could not add a reference to a GB surface " "backup buffer.\n"); - (void) ttm_ref_object_base_unref(tfile, base->hash.key, + (void) ttm_ref_object_base_unref(tfile, base->handle, TTM_REF_USAGE); goto out_bad_resource; } @@ -1763,7 +1763,7 @@ vmw_gb_surface_reference_internal(struct drm_device *dev, rep->creq.base.array_size = srf->array_size; rep->creq.base.buffer_handle = backup_handle; rep->creq.base.base_size = srf->base_size; - rep->crep.handle = user_srf->prime.base.hash.key; + rep->crep.handle = user_srf->prime.base.handle; rep->crep.backup_size = srf->res.backup_size; rep->crep.buffer_handle = backup_handle; rep->crep.buffer_map_handle = diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c new file mode 100644 index 000000000000..184025fa938e --- /dev/null +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c @@ -0,0 +1,770 @@ +// SPDX-License-Identifier: GPL-2.0 OR MIT +/************************************************************************** + * + * Copyright © 2018 VMware, Inc., Palo Alto, CA., USA + * All Rights Reserved. + * + * 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, sub license, 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 (including the + * next paragraph) 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 NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS 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/slab.h> +#include "vmwgfx_validation.h" +#include "vmwgfx_drv.h" + +/** + * struct vmw_validation_bo_node - Buffer object validation metadata. + * @base: Metadata used for TTM reservation- and validation. + * @hash: A hash entry used for the duplicate detection hash table. + * @as_mob: Validate as mob. + * @cpu_blit: Validate for cpu blit access. + * + * Bit fields are used since these structures are allocated and freed in + * large numbers and space conservation is desired. + */ +struct vmw_validation_bo_node { + struct ttm_validate_buffer base; + struct drm_hash_item hash; + u32 as_mob : 1; + u32 cpu_blit : 1; +}; + +/** + * struct vmw_validation_res_node - Resource validation metadata. + * @head: List head for the resource validation list. + * @hash: A hash entry used for the duplicate detection hash table. + * @res: Reference counted resource pointer. + * @new_backup: Non ref-counted pointer to new backup buffer to be assigned + * to a resource. + * @new_backup_offset: Offset into the new backup mob for resources that can + * share MOBs. + * @no_buffer_needed: Kernel does not need to allocate a MOB during validation, + * the command stream provides a mob bind operation. + * @switching_backup: The validation process is switching backup MOB. + * @first_usage: True iff the resource has been seen only once in the current + * validation batch. + * @reserved: Whether the resource is currently reserved by this process. + * @private: Optionally additional memory for caller-private data. + * + * Bit fields are used since these structures are allocated and freed in + * large numbers and space conservation is desired. + */ +struct vmw_validation_res_node { + struct list_head head; + struct drm_hash_item hash; + struct vmw_resource *res; + struct vmw_buffer_object *new_backup; + unsigned long new_backup_offset; + u32 no_buffer_needed : 1; + u32 switching_backup : 1; + u32 first_usage : 1; + u32 reserved : 1; + unsigned long private[0]; +}; + +/** + * vmw_validation_mem_alloc - Allocate kernel memory from the validation + * context based allocator + * @ctx: The validation context + * @size: The number of bytes to allocated. + * + * The memory allocated may not exceed PAGE_SIZE, and the returned + * address is aligned to sizeof(long). All memory allocated this way is + * reclaimed after validation when calling any of the exported functions: + * vmw_validation_unref_lists() + * vmw_validation_revert() + * vmw_validation_done() + * + * Return: Pointer to the allocated memory on success. NULL on failure. + */ +void *vmw_validation_mem_alloc(struct vmw_validation_context *ctx, + unsigned int size) +{ + void *addr; + + size = vmw_validation_align(size); + if (size > PAGE_SIZE) + return NULL; + + if (ctx->mem_size_left < size) { + struct page *page = alloc_page(GFP_KERNEL | __GFP_ZERO); + + if (!page) + return NULL; + + list_add_tail(&page->lru, &ctx->page_list); + ctx->page_address = page_address(page); + ctx->mem_size_left = PAGE_SIZE; + } + + addr = (void *) (ctx->page_address + (PAGE_SIZE - ctx->mem_size_left)); + ctx->mem_size_left -= size; + + return addr; +} + +/** + * vmw_validation_mem_free - Free all memory allocated using + * vmw_validation_mem_alloc() + * @ctx: The validation context + * + * All memory previously allocated for this context using + * vmw_validation_mem_alloc() is freed. + */ +static void vmw_validation_mem_free(struct vmw_validation_context *ctx) +{ + struct page *entry, *next; + + list_for_each_entry_safe(entry, next, &ctx->page_list, lru) { + list_del_init(&entry->lru); + __free_page(entry); + } + + ctx->mem_size_left = 0; +} + +/** + * vmw_validation_find_bo_dup - Find a duplicate buffer object entry in the + * validation context's lists. + * @ctx: The validation context to search. + * @vbo: The buffer object to search for. + * + * Return: Pointer to the struct vmw_validation_bo_node referencing the + * duplicate, or NULL if none found. + */ +static struct vmw_validation_bo_node * +vmw_validation_find_bo_dup(struct vmw_validation_context *ctx, + struct vmw_buffer_object *vbo) +{ + struct vmw_validation_bo_node *bo_node = NULL; + + if (!ctx->merge_dups) + return NULL; + + if (ctx->ht) { + struct drm_hash_item *hash; + + if (!drm_ht_find_item(ctx->ht, (unsigned long) vbo, &hash)) + bo_node = container_of(hash, typeof(*bo_node), hash); + } else { + struct vmw_validation_bo_node *entry; + + list_for_each_entry(entry, &ctx->bo_list, base.head) { + if (entry->base.bo == &vbo->base) { + bo_node = entry; + break; + } + } + } + + return bo_node; +} + +/** + * vmw_validation_find_res_dup - Find a duplicate resource entry in the + * validation context's lists. + * @ctx: The validation context to search. + * @vbo: The buffer object to search for. + * + * Return: Pointer to the struct vmw_validation_bo_node referencing the + * duplicate, or NULL if none found. + */ +static struct vmw_validation_res_node * +vmw_validation_find_res_dup(struct vmw_validation_context *ctx, + struct vmw_resource *res) +{ + struct vmw_validation_res_node *res_node = NULL; + + if (!ctx->merge_dups) + return NULL; + + if (ctx->ht) { + struct drm_hash_item *hash; + + if (!drm_ht_find_item(ctx->ht, (unsigned long) res, &hash)) + res_node = container_of(hash, typeof(*res_node), hash); + } else { + struct vmw_validation_res_node *entry; + + list_for_each_entry(entry, &ctx->resource_ctx_list, head) { + if (entry->res == res) { + res_node = entry; + goto out; + } + } + + list_for_each_entry(entry, &ctx->resource_list, head) { + if (entry->res == res) { + res_node = entry; + break; + } + } + + } +out: + return res_node; +} + +/** + * vmw_validation_add_bo - Add a buffer object to the validation context. + * @ctx: The validation context. + * @vbo: The buffer object. + * @as_mob: Validate as mob, otherwise suitable for GMR operations. + * @cpu_blit: Validate in a page-mappable location. + * + * Return: Zero on success, negative error code otherwise. + */ +int vmw_validation_add_bo(struct vmw_validation_context *ctx, + struct vmw_buffer_object *vbo, + bool as_mob, + bool cpu_blit) +{ + struct vmw_validation_bo_node *bo_node; + + bo_node = vmw_validation_find_bo_dup(ctx, vbo); + if (bo_node) { + if (bo_node->as_mob != as_mob || + bo_node->cpu_blit != cpu_blit) { + DRM_ERROR("Inconsistent buffer usage.\n"); + return -EINVAL; + } + } else { + struct ttm_validate_buffer *val_buf; + int ret; + + bo_node = vmw_validation_mem_alloc(ctx, sizeof(*bo_node)); + if (!bo_node) + return -ENOMEM; + + if (ctx->ht) { + bo_node->hash.key = (unsigned long) vbo; + ret = drm_ht_insert_item(ctx->ht, &bo_node->hash); + if (ret) { + DRM_ERROR("Failed to initialize a buffer " + "validation entry.\n"); + return ret; + } + } + val_buf = &bo_node->base; + val_buf->bo = ttm_bo_get_unless_zero(&vbo->base); + if (!val_buf->bo) + return -ESRCH; + val_buf->shared = false; + list_add_tail(&val_buf->head, &ctx->bo_list); + bo_node->as_mob = as_mob; + bo_node->cpu_blit = cpu_blit; + } + + return 0; +} + +/** + * vmw_validation_add_resource - Add a resource to the validation context. + * @ctx: The validation context. + * @res: The resource. + * @priv_size: Size of private, additional metadata. + * @p_node: Output pointer of additional metadata address. + * @first_usage: Whether this was the first time this resource was seen. + * + * Return: Zero on success, negative error code otherwise. + */ +int vmw_validation_add_resource(struct vmw_validation_context *ctx, + struct vmw_resource *res, + size_t priv_size, + void **p_node, + bool *first_usage) +{ + struct vmw_validation_res_node *node; + int ret; + + node = vmw_validation_find_res_dup(ctx, res); + if (node) { + node->first_usage = 0; + goto out_fill; + } + + node = vmw_validation_mem_alloc(ctx, sizeof(*node) + priv_size); + if (!node) { + DRM_ERROR("Failed to allocate a resource validation " + "entry.\n"); + return -ENOMEM; + } + + if (ctx->ht) { + node->hash.key = (unsigned long) res; + ret = drm_ht_insert_item(ctx->ht, &node->hash); + if (ret) { + DRM_ERROR("Failed to initialize a resource validation " + "entry.\n"); + return ret; + } + } + node->res = vmw_resource_reference_unless_doomed(res); + if (!node->res) + return -ESRCH; + + node->first_usage = 1; + if (!res->dev_priv->has_mob) { + list_add_tail(&node->head, &ctx->resource_list); + } else { + switch (vmw_res_type(res)) { + case vmw_res_context: + case vmw_res_dx_context: + list_add(&node->head, &ctx->resource_ctx_list); + break; + case vmw_res_cotable: + list_add_tail(&node->head, &ctx->resource_ctx_list); + break; + default: + list_add_tail(&node->head, &ctx->resource_list); + break; + } + } + +out_fill: + if (first_usage) + *first_usage = node->first_usage; + if (p_node) + *p_node = &node->private; + + return 0; +} + +/** + * vmw_validation_res_switch_backup - Register a backup MOB switch during + * validation. + * @ctx: The validation context. + * @val_private: The additional meta-data pointer returned when the + * resource was registered with the validation context. Used to identify + * the resource. + * @vbo: The new backup buffer object MOB. This buffer object needs to have + * already been registered with the validation context. + * @backup_offset: Offset into the new backup MOB. + */ +void vmw_validation_res_switch_backup(struct vmw_validation_context *ctx, + void *val_private, + struct vmw_buffer_object *vbo, + unsigned long backup_offset) +{ + struct vmw_validation_res_node *val; + + val = container_of(val_private, typeof(*val), private); + + val->switching_backup = 1; + if (val->first_usage) + val->no_buffer_needed = 1; + + val->new_backup = vbo; + val->new_backup_offset = backup_offset; +} + +/** + * vmw_validation_res_reserve - Reserve all resources registered with this + * validation context. + * @ctx: The validation context. + * @intr: Use interruptible waits when possible. + * + * Return: Zero on success, -ERESTARTSYS if interrupted. Negative error + * code on failure. + */ +int vmw_validation_res_reserve(struct vmw_validation_context *ctx, + bool intr) +{ + struct vmw_validation_res_node *val; + int ret = 0; + + list_splice_init(&ctx->resource_ctx_list, &ctx->resource_list); + + list_for_each_entry(val, &ctx->resource_list, head) { + struct vmw_resource *res = val->res; + + ret = vmw_resource_reserve(res, intr, val->no_buffer_needed); + if (ret) + goto out_unreserve; + + val->reserved = 1; + if (res->backup) { + struct vmw_buffer_object *vbo = res->backup; + + ret = vmw_validation_add_bo + (ctx, vbo, vmw_resource_needs_backup(res), + false); + if (ret) + goto out_unreserve; + } + } + + return 0; + +out_unreserve: + vmw_validation_res_unreserve(ctx, true); + return ret; +} + +/** + * vmw_validation_res_unreserve - Unreserve all reserved resources + * registered with this validation context. + * @ctx: The validation context. + * @backoff: Whether this is a backoff- of a commit-type operation. This + * is used to determine whether to switch backup MOBs or not. + */ +void vmw_validation_res_unreserve(struct vmw_validation_context *ctx, + bool backoff) +{ + struct vmw_validation_res_node *val; + + list_splice_init(&ctx->resource_ctx_list, &ctx->resource_list); + + list_for_each_entry(val, &ctx->resource_list, head) { + if (val->reserved) + vmw_resource_unreserve(val->res, + !backoff && + val->switching_backup, + val->new_backup, + val->new_backup_offset); + } +} + +/** + * vmw_validation_bo_validate_single - Validate a single buffer object. + * @bo: The TTM buffer object base. + * @interruptible: Whether to perform waits interruptible if possible. + * @validate_as_mob: Whether to validate in MOB memory. + * + * Return: Zero on success, -ERESTARTSYS if interrupted. Negative error + * code on failure. + */ +int vmw_validation_bo_validate_single(struct ttm_buffer_object *bo, + bool interruptible, + bool validate_as_mob) +{ + struct vmw_buffer_object *vbo = + container_of(bo, struct vmw_buffer_object, base); + struct ttm_operation_ctx ctx = { + .interruptible = interruptible, + .no_wait_gpu = false + }; + int ret; + + if (vbo->pin_count > 0) + return 0; + + if (validate_as_mob) + return ttm_bo_validate(bo, &vmw_mob_placement, &ctx); + + /** + * Put BO in VRAM if there is space, otherwise as a GMR. + * If there is no space in VRAM and GMR ids are all used up, + * start evicting GMRs to make room. If the DMA buffer can't be + * used as a GMR, this will return -ENOMEM. + */ + + ret = ttm_bo_validate(bo, &vmw_vram_gmr_placement, &ctx); + if (ret == 0 || ret == -ERESTARTSYS) + return ret; + + /** + * If that failed, try VRAM again, this time evicting + * previous contents. + */ + + ret = ttm_bo_validate(bo, &vmw_vram_placement, &ctx); + return ret; +} + +/** + * vmw_validation_bo_validate - Validate all buffer objects registered with + * the validation context. + * @ctx: The validation context. + * @intr: Whether to perform waits interruptible if possible. + * + * Return: Zero on success, -ERESTARTSYS if interrupted, + * negative error code on failure. + */ +int vmw_validation_bo_validate(struct vmw_validation_context *ctx, bool intr) +{ + struct vmw_validation_bo_node *entry; + int ret; + + list_for_each_entry(entry, &ctx->bo_list, base.head) { + if (entry->cpu_blit) { + struct ttm_operation_ctx ctx = { + .interruptible = intr, + .no_wait_gpu = false + }; + + ret = ttm_bo_validate(entry->base.bo, + &vmw_nonfixed_placement, &ctx); + } else { + ret = vmw_validation_bo_validate_single + (entry->base.bo, intr, entry->as_mob); + } + if (ret) + return ret; + } + return 0; +} + +/** + * vmw_validation_res_validate - Validate all resources registered with the + * validation context. + * @ctx: The validation context. + * @intr: Whether to perform waits interruptible if possible. + * + * Before this function is called, all resource backup buffers must have + * been validated. + * + * Return: Zero on success, -ERESTARTSYS if interrupted, + * negative error code on failure. + */ +int vmw_validation_res_validate(struct vmw_validation_context *ctx, bool intr) +{ + struct vmw_validation_res_node *val; + int ret; + + list_for_each_entry(val, &ctx->resource_list, head) { + struct vmw_resource *res = val->res; + struct vmw_buffer_object *backup = res->backup; + + ret = vmw_resource_validate(res, intr); + if (ret) { + if (ret != -ERESTARTSYS) + DRM_ERROR("Failed to validate resource.\n"); + return ret; + } + + /* Check if the resource switched backup buffer */ + if (backup && res->backup && (backup != res->backup)) { + struct vmw_buffer_object *vbo = res->backup; + + ret = vmw_validation_add_bo + (ctx, vbo, vmw_resource_needs_backup(res), + false); + if (ret) + return ret; + } + } + return 0; +} + +/** + * vmw_validation_drop_ht - Reset the hash table used for duplicate finding + * and unregister it from this validation context. + * @ctx: The validation context. + * + * The hash table used for duplicate finding is an expensive resource and + * may be protected by mutexes that may cause deadlocks during resource + * unreferencing if held. After resource- and buffer object registering, + * there is no longer any use for this hash table, so allow freeing it + * either to shorten any mutex locking time, or before resources- and + * buffer objects are freed during validation context cleanup. + */ +void vmw_validation_drop_ht(struct vmw_validation_context *ctx) +{ + struct vmw_validation_bo_node *entry; + struct vmw_validation_res_node *val; + + if (!ctx->ht) + return; + + list_for_each_entry(entry, &ctx->bo_list, base.head) + (void) drm_ht_remove_item(ctx->ht, &entry->hash); + + list_for_each_entry(val, &ctx->resource_list, head) + (void) drm_ht_remove_item(ctx->ht, &val->hash); + + list_for_each_entry(val, &ctx->resource_ctx_list, head) + (void) drm_ht_remove_item(ctx->ht, &val->hash); + + ctx->ht = NULL; +} + +/** + * vmw_validation_unref_lists - Unregister previously registered buffer + * object and resources. + * @ctx: The validation context. + * + * Note that this function may cause buffer object- and resource destructors + * to be invoked. + */ +void vmw_validation_unref_lists(struct vmw_validation_context *ctx) +{ + struct vmw_validation_bo_node *entry; + struct vmw_validation_res_node *val; + + list_for_each_entry(entry, &ctx->bo_list, base.head) + ttm_bo_unref(&entry->base.bo); + + list_splice_init(&ctx->resource_ctx_list, &ctx->resource_list); + list_for_each_entry(val, &ctx->resource_list, head) + vmw_resource_unreference(&val->res); + + /* + * No need to detach each list entry since they are all freed with + * vmw_validation_free_mem. Just make the inaccessible. + */ + INIT_LIST_HEAD(&ctx->bo_list); + INIT_LIST_HEAD(&ctx->resource_list); + + vmw_validation_mem_free(ctx); +} + +/** + * vmw_validation_prepare - Prepare a validation context for command + * submission. + * @ctx: The validation context. + * @mutex: The mutex used to protect resource reservation. + * @intr: Whether to perform waits interruptible if possible. + * + * Note that the single reservation mutex @mutex is an unfortunate + * construct. Ideally resource reservation should be moved to per-resource + * ww_mutexes. + * If this functions doesn't return Zero to indicate success, all resources + * are left unreserved but still referenced. + * Return: Zero on success, -ERESTARTSYS if interrupted, negative error code + * on error. + */ +int vmw_validation_prepare(struct vmw_validation_context *ctx, + struct mutex *mutex, + bool intr) +{ + int ret = 0; + + if (mutex) { + if (intr) + ret = mutex_lock_interruptible(mutex); + else + mutex_lock(mutex); + if (ret) + return -ERESTARTSYS; + } + + ctx->res_mutex = mutex; + ret = vmw_validation_res_reserve(ctx, intr); + if (ret) + goto out_no_res_reserve; + + ret = vmw_validation_bo_reserve(ctx, intr); + if (ret) + goto out_no_bo_reserve; + + ret = vmw_validation_bo_validate(ctx, intr); + if (ret) + goto out_no_validate; + + ret = vmw_validation_res_validate(ctx, intr); + if (ret) + goto out_no_validate; + + return 0; + +out_no_validate: + vmw_validation_bo_backoff(ctx); +out_no_bo_reserve: + vmw_validation_res_unreserve(ctx, true); +out_no_res_reserve: + if (mutex) + mutex_unlock(mutex); + + return ret; +} + +/** + * vmw_validation_revert - Revert validation actions if command submission + * failed. + * + * @ctx: The validation context. + * + * The caller still needs to unref resources after a call to this function. + */ +void vmw_validation_revert(struct vmw_validation_context *ctx) +{ + vmw_validation_bo_backoff(ctx); + vmw_validation_res_unreserve(ctx, true); + if (ctx->res_mutex) + mutex_unlock(ctx->res_mutex); + vmw_validation_unref_lists(ctx); +} + +/** + * vmw_validation_cone - Commit validation actions after command submission + * success. + * @ctx: The validation context. + * @fence: Fence with which to fence all buffer objects taking part in the + * command submission. + * + * The caller does NOT need to unref resources after a call to this function. + */ +void vmw_validation_done(struct vmw_validation_context *ctx, + struct vmw_fence_obj *fence) +{ + vmw_validation_bo_fence(ctx, fence); + vmw_validation_res_unreserve(ctx, false); + if (ctx->res_mutex) + mutex_unlock(ctx->res_mutex); + vmw_validation_unref_lists(ctx); +} + +/** + * vmw_validation_preload_bo - Preload the validation memory allocator for a + * call to vmw_validation_add_bo(). + * @ctx: Pointer to the validation context. + * + * Iff this function returns successfully, the next call to + * vmw_validation_add_bo() is guaranteed not to sleep. An error is not fatal + * but voids the guarantee. + * + * Returns: Zero if successful, %-EINVAL otherwise. + */ +int vmw_validation_preload_bo(struct vmw_validation_context *ctx) +{ + unsigned int size = sizeof(struct vmw_validation_bo_node); + + if (!vmw_validation_mem_alloc(ctx, size)) + return -ENOMEM; + + ctx->mem_size_left += size; + return 0; +} + +/** + * vmw_validation_preload_res - Preload the validation memory allocator for a + * call to vmw_validation_add_res(). + * @ctx: Pointer to the validation context. + * @size: Size of the validation node extra data. See below. + * + * Iff this function returns successfully, the next call to + * vmw_validation_add_res() with the same or smaller @size is guaranteed not to + * sleep. An error is not fatal but voids the guarantee. + * + * Returns: Zero if successful, %-EINVAL otherwise. + */ +int vmw_validation_preload_res(struct vmw_validation_context *ctx, + unsigned int size) +{ + size = vmw_validation_align(sizeof(struct vmw_validation_res_node) + + size) + + vmw_validation_align(sizeof(struct vmw_validation_bo_node)); + if (!vmw_validation_mem_alloc(ctx, size)) + return -ENOMEM; + + ctx->mem_size_left += size; + return 0; +} diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.h b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.h new file mode 100644 index 000000000000..b57e3292c386 --- /dev/null +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.h @@ -0,0 +1,227 @@ +/* SPDX-License-Identifier: GPL-2.0 OR MIT */ +/************************************************************************** + * + * Copyright © 2018 VMware, Inc., Palo Alto, CA., USA + * All Rights Reserved. + * + * 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, sub license, 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 (including the + * next paragraph) 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 NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS 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 _VMWGFX_VALIDATION_H_ +#define _VMWGFX_VALIDATION_H_ + +#include <drm/drm_hashtab.h> +#include <linux/list.h> +#include <linux/ww_mutex.h> +#include <drm/ttm/ttm_execbuf_util.h> + +/** + * struct vmw_validation_context - Per command submission validation context + * @ht: Hash table used to find resource- or buffer object duplicates + * @resource_list: List head for resource validation metadata + * @resource_ctx_list: List head for resource validation metadata for + * resources that need to be validated before those in @resource_list + * @bo_list: List head for buffer objects + * @page_list: List of pages used by the memory allocator + * @ticket: Ticked used for ww mutex locking + * @res_mutex: Pointer to mutex used for resource reserving + * @merge_dups: Whether to merge metadata for duplicate resources or + * buffer objects + * @mem_size_left: Free memory left in the last page in @page_list + * @page_address: Kernel virtual address of the last page in @page_list + */ +struct vmw_validation_context { + struct drm_open_hash *ht; + struct list_head resource_list; + struct list_head resource_ctx_list; + struct list_head bo_list; + struct list_head page_list; + struct ww_acquire_ctx ticket; + struct mutex *res_mutex; + unsigned int merge_dups; + unsigned int mem_size_left; + u8 *page_address; +}; + +struct vmw_buffer_object; +struct vmw_resource; +struct vmw_fence_obj; + +#if 0 +/** + * DECLARE_VAL_CONTEXT - Declare a validation context with initialization + * @_name: The name of the variable + * @_ht: The hash table used to find dups or NULL if none + * @_merge_dups: Whether to merge duplicate buffer object- or resource + * entries. If set to true, ideally a hash table pointer should be supplied + * as well unless the number of resources and buffer objects per validation + * is known to be very small + */ +#endif +#define DECLARE_VAL_CONTEXT(_name, _ht, _merge_dups) \ + struct vmw_validation_context _name = \ + { .ht = _ht, \ + .resource_list = LIST_HEAD_INIT((_name).resource_list), \ + .resource_ctx_list = LIST_HEAD_INIT((_name).resource_ctx_list), \ + .bo_list = LIST_HEAD_INIT((_name).bo_list), \ + .page_list = LIST_HEAD_INIT((_name).page_list), \ + .res_mutex = NULL, \ + .merge_dups = _merge_dups, \ + .mem_size_left = 0, \ + } + +/** + * vmw_validation_has_bos - return whether the validation context has + * any buffer objects registered. + * + * @ctx: The validation context + * Returns: Whether any buffer objects are registered + */ +static inline bool +vmw_validation_has_bos(struct vmw_validation_context *ctx) +{ + return !list_empty(&ctx->bo_list); +} + +/** + * vmw_validation_set_ht - Register a hash table for duplicate finding + * @ctx: The validation context + * @ht: Pointer to a hash table to use for duplicate finding + * This function is intended to be used if the hash table wasn't + * available at validation context declaration time + */ +static inline void vmw_validation_set_ht(struct vmw_validation_context *ctx, + struct drm_open_hash *ht) +{ + ctx->ht = ht; +} + +/** + * vmw_validation_bo_reserve - Reserve buffer objects registered with a + * validation context + * @ctx: The validation context + * @intr: Perform waits interruptible + * + * Return: Zero on success, -ERESTARTSYS when interrupted, negative error + * code on failure + */ +static inline int +vmw_validation_bo_reserve(struct vmw_validation_context *ctx, + bool intr) +{ + return ttm_eu_reserve_buffers(&ctx->ticket, &ctx->bo_list, intr, + NULL); +} + +/** + * vmw_validation_bo_backoff - Unreserve buffer objects registered with a + * validation context + * @ctx: The validation context + * + * This function unreserves the buffer objects previously reserved using + * vmw_validation_bo_reserve. It's typically used as part of an error path + */ +static inline void +vmw_validation_bo_backoff(struct vmw_validation_context *ctx) +{ + ttm_eu_backoff_reservation(&ctx->ticket, &ctx->bo_list); +} + +/** + * vmw_validation_bo_fence - Unreserve and fence buffer objects registered + * with a validation context + * @ctx: The validation context + * + * This function unreserves the buffer objects previously reserved using + * vmw_validation_bo_reserve, and fences them with a fence object. + */ +static inline void +vmw_validation_bo_fence(struct vmw_validation_context *ctx, + struct vmw_fence_obj *fence) +{ + ttm_eu_fence_buffer_objects(&ctx->ticket, &ctx->bo_list, + (void *) fence); +} + +/** + * vmw_validation_context_init - Initialize a validation context + * @ctx: Pointer to the validation context to initialize + * + * This function initializes a validation context with @merge_dups set + * to false + */ +static inline void +vmw_validation_context_init(struct vmw_validation_context *ctx) +{ + memset(ctx, 0, sizeof(*ctx)); + INIT_LIST_HEAD(&ctx->resource_list); + INIT_LIST_HEAD(&ctx->resource_ctx_list); + INIT_LIST_HEAD(&ctx->bo_list); +} + +/** + * vmw_validation_align - Align a validation memory allocation + * @val: The size to be aligned + * + * Returns: @val aligned to the granularity used by the validation memory + * allocator. + */ +static inline unsigned int vmw_validation_align(unsigned int val) +{ + return ALIGN(val, sizeof(long)); +} + +int vmw_validation_add_bo(struct vmw_validation_context *ctx, + struct vmw_buffer_object *vbo, + bool as_mob, bool cpu_blit); +int vmw_validation_bo_validate_single(struct ttm_buffer_object *bo, + bool interruptible, + bool validate_as_mob); +int vmw_validation_bo_validate(struct vmw_validation_context *ctx, bool intr); +void vmw_validation_unref_lists(struct vmw_validation_context *ctx); +int vmw_validation_add_resource(struct vmw_validation_context *ctx, + struct vmw_resource *res, + size_t priv_size, + void **p_node, + bool *first_usage); +void vmw_validation_drop_ht(struct vmw_validation_context *ctx); +int vmw_validation_res_reserve(struct vmw_validation_context *ctx, + bool intr); +void vmw_validation_res_unreserve(struct vmw_validation_context *ctx, + bool backoff); +void vmw_validation_res_switch_backup(struct vmw_validation_context *ctx, + void *val_private, + struct vmw_buffer_object *vbo, + unsigned long backup_offset); +int vmw_validation_res_validate(struct vmw_validation_context *ctx, bool intr); + +int vmw_validation_prepare(struct vmw_validation_context *ctx, + struct mutex *mutex, bool intr); +void vmw_validation_revert(struct vmw_validation_context *ctx); +void vmw_validation_done(struct vmw_validation_context *ctx, + struct vmw_fence_obj *fence); + +void *vmw_validation_mem_alloc(struct vmw_validation_context *ctx, + unsigned int size); +int vmw_validation_preload_bo(struct vmw_validation_context *ctx); +int vmw_validation_preload_res(struct vmw_validation_context *ctx, + unsigned int size); +#endif |