From 85edae14e4ee5e68cf037e9e4bca7498ea16874d Mon Sep 17 00:00:00 2001 From: Michal Janusz Miroslaw Date: Thu, 23 Feb 2006 09:49:35 +0000 Subject: [SERIAL] Trivial comment fix: include/linux/serial_reg.h Trivial comment fix for include/linux/serial_reg.h Signed-off-by: Russell King --- include/linux/serial_reg.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/serial_reg.h b/include/linux/serial_reg.h index 6a2bb955844b..3c8a6aa77415 100644 --- a/include/linux/serial_reg.h +++ b/include/linux/serial_reg.h @@ -247,10 +247,10 @@ #define UART_CTR 0xFF /* - * The 16C950 Additional Control Reigster + * The 16C950 Additional Control Register */ #define UART_ACR_RXDIS 0x01 /* Receiver disable */ -#define UART_ACR_TXDIS 0x02 /* Receiver disable */ +#define UART_ACR_TXDIS 0x02 /* Transmitter disable */ #define UART_ACR_DSRFC 0x04 /* DSR Flow Control */ #define UART_ACR_TLENB 0x20 /* 950 trigger levels enable */ #define UART_ACR_ICRRD 0x40 /* ICR Read enable */ -- cgit v1.2.3-58-ga151 From 21380b81ef8699179b535e197a95b891a7badac7 Mon Sep 17 00:00:00 2001 From: Herbert Xu Date: Wed, 22 Feb 2006 14:47:13 -0800 Subject: [XFRM]: Eliminate refcounting confusion by creating __xfrm_state_put(). We often just do an atomic_dec(&x->refcnt) on an xfrm_state object because we know there is more than 1 reference remaining and thus we can elide the heavier xfrm_state_put() call. Do this behind an inline function called __xfrm_state_put() so that is more obvious and also to allow us to more cleanly add refcount debugging later. Signed-off-by: Herbert Xu Signed-off-by: David S. Miller --- include/net/xfrm.h | 5 +++++ net/key/af_key.c | 2 +- net/xfrm/xfrm_state.c | 8 ++++---- net/xfrm/xfrm_user.c | 2 +- 4 files changed, 11 insertions(+), 6 deletions(-) (limited to 'include') diff --git a/include/net/xfrm.h b/include/net/xfrm.h index d6111a2f0a23..004e645f3e18 100644 --- a/include/net/xfrm.h +++ b/include/net/xfrm.h @@ -403,6 +403,11 @@ unsigned xfrm_spi_hash(xfrm_address_t *addr, u32 spi, u8 proto, unsigned short f extern void __xfrm_state_destroy(struct xfrm_state *); +static inline void __xfrm_state_put(struct xfrm_state *x) +{ + atomic_dec(&x->refcnt); +} + static inline void xfrm_state_put(struct xfrm_state *x) { if (atomic_dec_and_test(&x->refcnt)) diff --git a/net/key/af_key.c b/net/key/af_key.c index ae86d237a456..b2d4d1dd2116 100644 --- a/net/key/af_key.c +++ b/net/key/af_key.c @@ -1423,7 +1423,7 @@ static int pfkey_add(struct sock *sk, struct sk_buff *skb, struct sadb_msg *hdr, if (err < 0) { x->km.state = XFRM_STATE_DEAD; - xfrm_state_put(x); + __xfrm_state_put(x); goto out; } diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c index e12d0be5f976..c656cbaf35e8 100644 --- a/net/xfrm/xfrm_state.c +++ b/net/xfrm/xfrm_state.c @@ -220,14 +220,14 @@ static int __xfrm_state_delete(struct xfrm_state *x) x->km.state = XFRM_STATE_DEAD; spin_lock(&xfrm_state_lock); list_del(&x->bydst); - atomic_dec(&x->refcnt); + __xfrm_state_put(x); if (x->id.spi) { list_del(&x->byspi); - atomic_dec(&x->refcnt); + __xfrm_state_put(x); } spin_unlock(&xfrm_state_lock); if (del_timer(&x->timer)) - atomic_dec(&x->refcnt); + __xfrm_state_put(x); /* The number two in this test is the reference * mentioned in the comment below plus the reference @@ -243,7 +243,7 @@ static int __xfrm_state_delete(struct xfrm_state *x) * The xfrm_state_alloc call gives a reference, and that * is what we are dropping here. */ - atomic_dec(&x->refcnt); + __xfrm_state_put(x); err = 0; } diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c index ac87a09ba83e..7de17559249a 100644 --- a/net/xfrm/xfrm_user.c +++ b/net/xfrm/xfrm_user.c @@ -345,7 +345,7 @@ static int xfrm_add_sa(struct sk_buff *skb, struct nlmsghdr *nlh, void **xfrma) if (err < 0) { x->km.state = XFRM_STATE_DEAD; - xfrm_state_put(x); + __xfrm_state_put(x); goto out; } -- cgit v1.2.3-58-ga151 From 337a7128dbe68ebe7627b6f954cb32d30d7b11c6 Mon Sep 17 00:00:00 2001 From: Michael Ellerman Date: Tue, 21 Feb 2006 17:22:55 +1100 Subject: [PATCH] powerpc: Only calculate htab_size in one place for kexec For kexec we need to know the size of the MMU hash table. Currently we calculate the size once in the htab code, and then twice more in the kexec code, once using htab_hash_mask and once using ppc64_pft_size. On some machines the ppc64_pft_size calculation is broken because ppc64_pft_size is not set. So we need to fix the second calculation, but better still we should just calculate the size once and use it everywhere else. Tested on Power5 LPAR, Power4 non-LPAR and Power3. Signed-off-by: Michael Ellerman Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/machine_kexec_64.c | 10 +++------- arch/powerpc/mm/hash_utils_64.c | 3 ++- include/asm-powerpc/mmu.h | 1 + 3 files changed, 6 insertions(+), 8 deletions(-) (limited to 'include') diff --git a/arch/powerpc/kernel/machine_kexec_64.c b/arch/powerpc/kernel/machine_kexec_64.c index d6431440c54f..ee166c586642 100644 --- a/arch/powerpc/kernel/machine_kexec_64.c +++ b/arch/powerpc/kernel/machine_kexec_64.c @@ -26,8 +26,6 @@ #include #include -#define HASH_GROUP_SIZE 0x80 /* size of each hash group, asm/mmu.h */ - int default_machine_kexec_prepare(struct kimage *image) { int i; @@ -61,7 +59,7 @@ int default_machine_kexec_prepare(struct kimage *image) */ if (htab_address) { low = __pa(htab_address); - high = low + (htab_hash_mask + 1) * HASH_GROUP_SIZE; + high = low + htab_size_bytes; for (i = 0; i < image->nr_segments; i++) { begin = image->segment[i].mem; @@ -294,7 +292,7 @@ void default_machine_kexec(struct kimage *image) } /* Values we need to export to the second kernel via the device tree. */ -static unsigned long htab_base, htab_size, kernel_end; +static unsigned long htab_base, kernel_end; static struct property htab_base_prop = { .name = "linux,htab-base", @@ -305,7 +303,7 @@ static struct property htab_base_prop = { static struct property htab_size_prop = { .name = "linux,htab-size", .length = sizeof(unsigned long), - .value = (unsigned char *)&htab_size, + .value = (unsigned char *)&htab_size_bytes, }; static struct property kernel_end_prop = { @@ -331,8 +329,6 @@ static void __init export_htab_values(void) htab_base = __pa(htab_address); prom_add_property(node, &htab_base_prop); - - htab_size = 1UL << ppc64_pft_size; prom_add_property(node, &htab_size_prop); out: diff --git a/arch/powerpc/mm/hash_utils_64.c b/arch/powerpc/mm/hash_utils_64.c index 149351a84b94..b1f614c612dd 100644 --- a/arch/powerpc/mm/hash_utils_64.c +++ b/arch/powerpc/mm/hash_utils_64.c @@ -88,6 +88,7 @@ static unsigned long _SDR1; struct mmu_psize_def mmu_psize_defs[MMU_PAGE_COUNT]; hpte_t *htab_address; +unsigned long htab_size_bytes; unsigned long htab_hash_mask; int mmu_linear_psize = MMU_PAGE_4K; int mmu_virtual_psize = MMU_PAGE_4K; @@ -399,7 +400,7 @@ void create_section_mapping(unsigned long start, unsigned long end) void __init htab_initialize(void) { - unsigned long table, htab_size_bytes; + unsigned long table; unsigned long pteg_count; unsigned long mode_rw; unsigned long base = 0, size = 0; diff --git a/include/asm-powerpc/mmu.h b/include/asm-powerpc/mmu.h index d096d9e76ad7..b0b9a3f8cdc2 100644 --- a/include/asm-powerpc/mmu.h +++ b/include/asm-powerpc/mmu.h @@ -112,6 +112,7 @@ typedef struct { } hpte_t; extern hpte_t *htab_address; +extern unsigned long htab_size_bytes; extern unsigned long htab_hash_mask; /* -- cgit v1.2.3-58-ga151 From 1775dbbcd02cab0c41329dd2cec5b69c7fafd13f Mon Sep 17 00:00:00 2001 From: Kumar Gala Date: Wed, 22 Feb 2006 09:46:02 -0600 Subject: [PATCH] powerpc: Enable coherency for all pages on 83xx to fix PCI data corruption On the 83xx platform to ensure the PCI inbound memory is handled properly we have to turn on coherency for all pages in the MMU. Otherwise we see corruption if inbound "prefetching/streaming" is enabled on the PCI controller. Signed-off-by: Randy Vinson Signed-off-by: Kumar Gala Signed-off-by: Paul Mackerras --- include/asm-powerpc/cputable.h | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'include') diff --git a/include/asm-powerpc/cputable.h b/include/asm-powerpc/cputable.h index 64210549f56b..90d005bb4d1c 100644 --- a/include/asm-powerpc/cputable.h +++ b/include/asm-powerpc/cputable.h @@ -159,9 +159,11 @@ extern void do_cpu_ftr_fixups(unsigned long offset); #endif /* We need to mark all pages as being coherent if we're SMP or we - * have a 74[45]x and an MPC107 host bridge. + * have a 74[45]x and an MPC107 host bridge. Also 83xx requires + * it for PCI "streaming/prefetch" to work properly. */ -#if defined(CONFIG_SMP) || defined(CONFIG_MPC10X_BRIDGE) +#if defined(CONFIG_SMP) || defined(CONFIG_MPC10X_BRIDGE) \ + || defined(CONFIG_PPC_83xx) #define CPU_FTR_COMMON CPU_FTR_NEED_COHERENT #else #define CPU_FTR_COMMON 0 @@ -277,7 +279,8 @@ enum { CPU_FTRS_G2_LE = CPU_FTR_SPLIT_ID_CACHE | CPU_FTR_MAYBE_CAN_DOZE | CPU_FTR_USE_TB | CPU_FTR_MAYBE_CAN_NAP | CPU_FTR_HAS_HIGH_BATS, CPU_FTRS_E300 = CPU_FTR_SPLIT_ID_CACHE | CPU_FTR_MAYBE_CAN_DOZE | - CPU_FTR_USE_TB | CPU_FTR_MAYBE_CAN_NAP | CPU_FTR_HAS_HIGH_BATS, + CPU_FTR_USE_TB | CPU_FTR_MAYBE_CAN_NAP | CPU_FTR_HAS_HIGH_BATS | + CPU_FTR_COMMON, CPU_FTRS_CLASSIC32 = CPU_FTR_COMMON | CPU_FTR_SPLIT_ID_CACHE | CPU_FTR_USE_TB | CPU_FTR_HPTE_TABLE, CPU_FTRS_POWER3_32 = CPU_FTR_COMMON | CPU_FTR_SPLIT_ID_CACHE | -- cgit v1.2.3-58-ga151 From cb2c9b2741346eb23b177187a51ff5abf08295bd Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Mon, 13 Feb 2006 14:48:35 +1100 Subject: [PATCH] powerpc: Fix runlatch performance issues The runlatch SPR can take a lot of time to write. My original runlatch code would set it on every exception entry even though most of the time this was not required. It would also continually set it in the idle loop, which is an issue on an SMT capable processor. Now we cache the runlatch value in a threadinfo bit, and only check for it in decrementer and hardware interrupt exceptions as well as the idle loop. Boot on POWER3, POWER5 and iseries, and compile tested on pmac32. Signed-off-by: Anton Blanchard Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/head_64.S | 12 ++---------- arch/powerpc/kernel/process.c | 32 ++++++++++++++++++++++++++++++++ arch/powerpc/platforms/iseries/setup.c | 1 + include/asm-powerpc/reg.h | 33 +++------------------------------ include/asm-powerpc/thread_info.h | 4 ++-- 5 files changed, 40 insertions(+), 42 deletions(-) (limited to 'include') diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S index 2b03a09fe5e9..bb845eed0e98 100644 --- a/arch/powerpc/kernel/head_64.S +++ b/arch/powerpc/kernel/head_64.S @@ -319,7 +319,6 @@ exception_marker: label##_pSeries: \ HMT_MEDIUM; \ mtspr SPRN_SPRG1,r13; /* save r13 */ \ - RUNLATCH_ON(r13); \ EXCEPTION_PROLOG_PSERIES(PACA_EXGEN, label##_common) #define STD_EXCEPTION_ISERIES(n, label, area) \ @@ -327,7 +326,6 @@ label##_pSeries: \ label##_iSeries: \ HMT_MEDIUM; \ mtspr SPRN_SPRG1,r13; /* save r13 */ \ - RUNLATCH_ON(r13); \ EXCEPTION_PROLOG_ISERIES_1(area); \ EXCEPTION_PROLOG_ISERIES_2; \ b label##_common @@ -337,7 +335,6 @@ label##_iSeries: \ label##_iSeries: \ HMT_MEDIUM; \ mtspr SPRN_SPRG1,r13; /* save r13 */ \ - RUNLATCH_ON(r13); \ EXCEPTION_PROLOG_ISERIES_1(PACA_EXGEN); \ lbz r10,PACAPROCENABLED(r13); \ cmpwi 0,r10,0; \ @@ -390,6 +387,7 @@ label##_common: \ label##_common: \ EXCEPTION_PROLOG_COMMON(trap, PACA_EXGEN); \ DISABLE_INTS; \ + bl .ppc64_runlatch_on; \ addi r3,r1,STACK_FRAME_OVERHEAD; \ bl hdlr; \ b .ret_from_except_lite @@ -407,7 +405,6 @@ __start_interrupts: _machine_check_pSeries: HMT_MEDIUM mtspr SPRN_SPRG1,r13 /* save r13 */ - RUNLATCH_ON(r13) EXCEPTION_PROLOG_PSERIES(PACA_EXMC, machine_check_common) . = 0x300 @@ -434,7 +431,6 @@ END_FTR_SECTION_IFCLR(CPU_FTR_SLB) data_access_slb_pSeries: HMT_MEDIUM mtspr SPRN_SPRG1,r13 - RUNLATCH_ON(r13) mfspr r13,SPRN_SPRG3 /* get paca address into r13 */ std r3,PACA_EXSLB+EX_R3(r13) mfspr r3,SPRN_DAR @@ -460,7 +456,6 @@ data_access_slb_pSeries: instruction_access_slb_pSeries: HMT_MEDIUM mtspr SPRN_SPRG1,r13 - RUNLATCH_ON(r13) mfspr r13,SPRN_SPRG3 /* get paca address into r13 */ std r3,PACA_EXSLB+EX_R3(r13) mfspr r3,SPRN_SRR0 /* SRR0 is faulting address */ @@ -491,7 +486,6 @@ instruction_access_slb_pSeries: .globl system_call_pSeries system_call_pSeries: HMT_MEDIUM - RUNLATCH_ON(r9) mr r9,r13 mfmsr r10 mfspr r13,SPRN_SPRG3 @@ -575,7 +569,6 @@ slb_miss_user_pseries: system_reset_fwnmi: HMT_MEDIUM mtspr SPRN_SPRG1,r13 /* save r13 */ - RUNLATCH_ON(r13) EXCEPTION_PROLOG_PSERIES(PACA_EXGEN, system_reset_common) .globl machine_check_fwnmi @@ -583,7 +576,6 @@ system_reset_fwnmi: machine_check_fwnmi: HMT_MEDIUM mtspr SPRN_SPRG1,r13 /* save r13 */ - RUNLATCH_ON(r13) EXCEPTION_PROLOG_PSERIES(PACA_EXMC, machine_check_common) #ifdef CONFIG_PPC_ISERIES @@ -894,7 +886,6 @@ unrecov_fer: .align 7 .globl data_access_common data_access_common: - RUNLATCH_ON(r10) /* It wont fit in the 0x300 handler */ mfspr r10,SPRN_DAR std r10,PACA_EXGEN+EX_DAR(r13) mfspr r10,SPRN_DSISR @@ -1042,6 +1033,7 @@ hardware_interrupt_common: EXCEPTION_PROLOG_COMMON(0x500, PACA_EXGEN) hardware_interrupt_entry: DISABLE_INTS + bl .ppc64_runlatch_on addi r3,r1,STACK_FRAME_OVERHEAD bl .do_IRQ b .ret_from_except_lite diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c index 57703994a063..c225cf154bfe 100644 --- a/arch/powerpc/kernel/process.c +++ b/arch/powerpc/kernel/process.c @@ -888,3 +888,35 @@ void dump_stack(void) show_stack(current, NULL); } EXPORT_SYMBOL(dump_stack); + +#ifdef CONFIG_PPC64 +void ppc64_runlatch_on(void) +{ + unsigned long ctrl; + + if (cpu_has_feature(CPU_FTR_CTRL) && !test_thread_flag(TIF_RUNLATCH)) { + HMT_medium(); + + ctrl = mfspr(SPRN_CTRLF); + ctrl |= CTRL_RUNLATCH; + mtspr(SPRN_CTRLT, ctrl); + + set_thread_flag(TIF_RUNLATCH); + } +} + +void ppc64_runlatch_off(void) +{ + unsigned long ctrl; + + if (cpu_has_feature(CPU_FTR_CTRL) && test_thread_flag(TIF_RUNLATCH)) { + HMT_medium(); + + clear_thread_flag(TIF_RUNLATCH); + + ctrl = mfspr(SPRN_CTRLF); + ctrl &= ~CTRL_RUNLATCH; + mtspr(SPRN_CTRLT, ctrl); + } +} +#endif diff --git a/arch/powerpc/platforms/iseries/setup.c b/arch/powerpc/platforms/iseries/setup.c index 3f8790146b00..3ecc4a652d82 100644 --- a/arch/powerpc/platforms/iseries/setup.c +++ b/arch/powerpc/platforms/iseries/setup.c @@ -648,6 +648,7 @@ static void yield_shared_processor(void) * here and let the timer_interrupt code sort out the actual time. */ get_lppaca()->int_dword.fields.decr_int = 1; + ppc64_runlatch_on(); process_iSeries_events(); } diff --git a/include/asm-powerpc/reg.h b/include/asm-powerpc/reg.h index 12ecc9b9f285..72bfe3af0460 100644 --- a/include/asm-powerpc/reg.h +++ b/include/asm-powerpc/reg.h @@ -615,27 +615,9 @@ #define proc_trap() asm volatile("trap") #ifdef CONFIG_PPC64 -static inline void ppc64_runlatch_on(void) -{ - unsigned long ctrl; - - if (cpu_has_feature(CPU_FTR_CTRL)) { - ctrl = mfspr(SPRN_CTRLF); - ctrl |= CTRL_RUNLATCH; - mtspr(SPRN_CTRLT, ctrl); - } -} - -static inline void ppc64_runlatch_off(void) -{ - unsigned long ctrl; - - if (cpu_has_feature(CPU_FTR_CTRL)) { - ctrl = mfspr(SPRN_CTRLF); - ctrl &= ~CTRL_RUNLATCH; - mtspr(SPRN_CTRLT, ctrl); - } -} + +extern void ppc64_runlatch_on(void); +extern void ppc64_runlatch_off(void); extern unsigned long scom970_read(unsigned int address); extern void scom970_write(unsigned int address, unsigned long value); @@ -645,15 +627,6 @@ extern void scom970_write(unsigned int address, unsigned long value); #define __get_SP() ({unsigned long sp; \ asm volatile("mr %0,1": "=r" (sp)); sp;}) -#else /* __ASSEMBLY__ */ - -#define RUNLATCH_ON(REG) \ -BEGIN_FTR_SECTION \ - mfspr (REG),SPRN_CTRLF; \ - ori (REG),(REG),CTRL_RUNLATCH; \ - mtspr SPRN_CTRLT,(REG); \ -END_FTR_SECTION_IFSET(CPU_FTR_CTRL) - #endif /* __ASSEMBLY__ */ #endif /* __KERNEL__ */ #endif /* _ASM_POWERPC_REG_H */ diff --git a/include/asm-powerpc/thread_info.h b/include/asm-powerpc/thread_info.h index c044ec16a879..237fc2b72974 100644 --- a/include/asm-powerpc/thread_info.h +++ b/include/asm-powerpc/thread_info.h @@ -113,7 +113,7 @@ static inline struct thread_info *current_thread_info(void) #define TIF_POLLING_NRFLAG 4 /* true if poll_idle() is polling TIF_NEED_RESCHED */ #define TIF_32BIT 5 /* 32 bit binary */ -/* #define SPARE 6 */ +#define TIF_RUNLATCH 6 /* Is the runlatch enabled? */ #define TIF_ABI_PENDING 7 /* 32/64 bit switch needed */ #define TIF_SYSCALL_AUDIT 8 /* syscall auditing active */ #define TIF_SINGLESTEP 9 /* singlestepping active */ @@ -131,7 +131,7 @@ static inline struct thread_info *current_thread_info(void) #define _TIF_NEED_RESCHED (1< Date: Fri, 24 Feb 2006 13:03:51 -0800 Subject: [PATCH] m32r: fix and update for gcc-4.0 Fix and update for gcc-4.0. - arch/m32r/kernel/signal.c: Change type of the 8th parameter of sys_rt_sigsuspend() from 'struct pt_regs' to 'struct pt_regs *'. This functions make use of the 'regs' parameter to return status value, but gcc-4.0 optimizes and removes it as a dead code. Functions, sys_sigaltstack() and sys_rt_sigreturn(), have also modified. - arch/m32r/lib/usercopy.c, include/asm-m32r/uaccess.h: Add early-clobber constraints('&') to output values of asm statements; these constraints seems to be required for gcc-4.0 register assignment. Signed-off-by: Hayato Fujiwara Signed-off-by: Hirokazu Takata Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/m32r/kernel/signal.c | 24 ++++++++++-------------- arch/m32r/lib/usercopy.c | 4 ++-- include/asm-m32r/uaccess.h | 8 ++++---- 3 files changed, 16 insertions(+), 20 deletions(-) (limited to 'include') diff --git a/arch/m32r/kernel/signal.c b/arch/m32r/kernel/signal.c index 71763f7a1d19..cb33097fefc4 100644 --- a/arch/m32r/kernel/signal.c +++ b/arch/m32r/kernel/signal.c @@ -36,7 +36,7 @@ int do_signal(struct pt_regs *, sigset_t *); asmlinkage int sys_rt_sigsuspend(sigset_t *unewset, size_t sigsetsize, unsigned long r2, unsigned long r3, unsigned long r4, - unsigned long r5, unsigned long r6, struct pt_regs regs) + unsigned long r5, unsigned long r6, struct pt_regs *regs) { sigset_t saveset, newset; @@ -54,21 +54,21 @@ sys_rt_sigsuspend(sigset_t *unewset, size_t sigsetsize, recalc_sigpending(); spin_unlock_irq(¤t->sighand->siglock); - regs.r0 = -EINTR; + regs->r0 = -EINTR; while (1) { current->state = TASK_INTERRUPTIBLE; schedule(); - if (do_signal(®s, &saveset)) - return regs.r0; + if (do_signal(regs, &saveset)) + return regs->r0; } } asmlinkage int sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, unsigned long r2, unsigned long r3, unsigned long r4, - unsigned long r5, unsigned long r6, struct pt_regs regs) + unsigned long r5, unsigned long r6, struct pt_regs *regs) { - return do_sigaltstack(uss, uoss, regs.spu); + return do_sigaltstack(uss, uoss, regs->spu); } @@ -140,11 +140,10 @@ restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, asmlinkage int sys_rt_sigreturn(unsigned long r0, unsigned long r1, unsigned long r2, unsigned long r3, unsigned long r4, - unsigned long r5, unsigned long r6, struct pt_regs regs) + unsigned long r5, unsigned long r6, struct pt_regs *regs) { - struct rt_sigframe __user *frame = (struct rt_sigframe __user *)regs.spu; + struct rt_sigframe __user *frame = (struct rt_sigframe __user *)regs->spu; sigset_t set; - stack_t st; int result; if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) @@ -158,14 +157,11 @@ sys_rt_sigreturn(unsigned long r0, unsigned long r1, recalc_sigpending(); spin_unlock_irq(¤t->sighand->siglock); - if (restore_sigcontext(®s, &frame->uc.uc_mcontext, &result)) + if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &result)) goto badframe; - if (__copy_from_user(&st, &frame->uc.uc_stack, sizeof(st))) + if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->spu) == -EFAULT) goto badframe; - /* It is more difficult to avoid calling this function than to - call it and ignore errors. */ - do_sigaltstack(&st, NULL, regs.spu); return result; diff --git a/arch/m32r/lib/usercopy.c b/arch/m32r/lib/usercopy.c index ce16bbe26a52..2d1dd2106c4d 100644 --- a/arch/m32r/lib/usercopy.c +++ b/arch/m32r/lib/usercopy.c @@ -64,7 +64,7 @@ do { \ " .balign 4\n" \ " .long 0b,3b\n" \ ".previous" \ - : "=r"(res), "=r"(count), "=&r" (__d0), "=&r" (__d1), \ + : "=&r"(res), "=&r"(count), "=&r" (__d0), "=&r" (__d1), \ "=&r" (__d2) \ : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), \ "4"(dst) \ @@ -101,7 +101,7 @@ do { \ " .balign 4\n" \ " .long 0b,3b\n" \ ".previous" \ - : "=r"(res), "=r"(count), "=&r" (__d0), "=&r" (__d1), \ + : "=&r"(res), "=&r"(count), "=&r" (__d0), "=&r" (__d1), \ "=&r" (__d2) \ : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), \ "4"(dst) \ diff --git a/include/asm-m32r/uaccess.h b/include/asm-m32r/uaccess.h index 0da7c47d2f01..e8ae61956a51 100644 --- a/include/asm-m32r/uaccess.h +++ b/include/asm-m32r/uaccess.h @@ -328,7 +328,7 @@ extern void __put_user_bad(void); " .long 1b,4b\n" \ " .long 2b,4b\n" \ ".previous" \ - : "=r"(err) \ + : "=&r"(err) \ : "r"(x), "r"(addr), "i"(-EFAULT), "0"(err) \ : "r14", "memory") @@ -353,7 +353,7 @@ extern void __put_user_bad(void); " .long 1b,4b\n" \ " .long 2b,4b\n" \ ".previous" \ - : "=r"(err) \ + : "=&r"(err) \ : "r"(x), "r"(addr), "i"(-EFAULT), "0"(err) \ : "r14", "memory") #else @@ -398,7 +398,7 @@ struct __large_struct { unsigned long buf[100]; }; " .balign 4\n" \ " .long 1b,3b\n" \ ".previous" \ - : "=r"(err) \ + : "=&r"(err) \ : "r"(x), "r"(addr), "i"(-EFAULT), "0"(err) \ : "r14", "memory") @@ -442,7 +442,7 @@ do { \ " .balign 4\n" \ " .long 1b,3b\n" \ ".previous" \ - : "=r"(err), "=&r"(x) \ + : "=&r"(err), "=&r"(x) \ : "r"(addr), "i"(-EFAULT), "0"(err) \ : "r14", "memory") -- cgit v1.2.3-58-ga151 From 124d90be62343f71bbb7a6b4a907b5584181e6d5 Mon Sep 17 00:00:00 2001 From: Prasanna S Panchamukhi Date: Fri, 24 Feb 2006 13:04:08 -0800 Subject: [PATCH] Kprobes causes NX protection fault on i686 SMP Fix a problem seen on i686 machine with NX support where the instruction could not be single stepped because of NX bit set on the memory pages allocated by kprobes module. This patch provides allocation of instruction solt so that the processor can execute the instruction from that location similar to x86_64 architecture. Thanks to Bibo and Masami for testing this patch. Signed-off-by: Prasanna S Panchamukhi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/i386/kernel/kprobes.c | 16 ++++++++++++++-- include/asm-i386/kprobes.h | 7 +++++-- 2 files changed, 19 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/arch/i386/kernel/kprobes.c b/arch/i386/kernel/kprobes.c index 6483eeb1a4e8..694a13997637 100644 --- a/arch/i386/kernel/kprobes.c +++ b/arch/i386/kernel/kprobes.c @@ -58,6 +58,11 @@ static inline int is_IF_modifier(kprobe_opcode_t opcode) int __kprobes arch_prepare_kprobe(struct kprobe *p) { + /* insn: must be on special executable page on i386. */ + p->ainsn.insn = get_insn_slot(); + if (!p->ainsn.insn) + return -ENOMEM; + memcpy(p->ainsn.insn, p->addr, MAX_INSN_SIZE * sizeof(kprobe_opcode_t)); p->opcode = *p->addr; return 0; @@ -77,6 +82,13 @@ void __kprobes arch_disarm_kprobe(struct kprobe *p) (unsigned long) p->addr + sizeof(kprobe_opcode_t)); } +void __kprobes arch_remove_kprobe(struct kprobe *p) +{ + down(&kprobe_mutex); + free_insn_slot(p->ainsn.insn); + up(&kprobe_mutex); +} + static inline void save_previous_kprobe(struct kprobe_ctlblk *kcb) { kcb->prev_kprobe.kp = kprobe_running(); @@ -111,7 +123,7 @@ static inline void prepare_singlestep(struct kprobe *p, struct pt_regs *regs) if (p->opcode == BREAKPOINT_INSTRUCTION) regs->eip = (unsigned long)p->addr; else - regs->eip = (unsigned long)&p->ainsn.insn; + regs->eip = (unsigned long)p->ainsn.insn; } /* Called with kretprobe_lock held */ @@ -351,7 +363,7 @@ static void __kprobes resume_execution(struct kprobe *p, { unsigned long *tos = (unsigned long *)®s->esp; unsigned long next_eip = 0; - unsigned long copy_eip = (unsigned long)&p->ainsn.insn; + unsigned long copy_eip = (unsigned long)p->ainsn.insn; unsigned long orig_eip = (unsigned long)p->addr; switch (p->ainsn.insn[0]) { diff --git a/include/asm-i386/kprobes.h b/include/asm-i386/kprobes.h index 27cac050a60e..a0d2d74a7dda 100644 --- a/include/asm-i386/kprobes.h +++ b/include/asm-i386/kprobes.h @@ -27,6 +27,9 @@ #include #include +#define __ARCH_WANT_KPROBES_INSN_SLOT + +struct kprobe; struct pt_regs; typedef u8 kprobe_opcode_t; @@ -40,14 +43,14 @@ typedef u8 kprobe_opcode_t; #define JPROBE_ENTRY(pentry) (kprobe_opcode_t *)pentry #define ARCH_SUPPORTS_KRETPROBES -#define arch_remove_kprobe(p) do {} while (0) +void arch_remove_kprobe(struct kprobe *p); void kretprobe_trampoline(void); /* Architecture specific copy of original instruction*/ struct arch_specific_insn { /* copy of the original instruction */ - kprobe_opcode_t insn[MAX_INSN_SIZE]; + kprobe_opcode_t *insn; }; struct prev_kprobe { -- cgit v1.2.3-58-ga151 From 2b932f6cf052920fb3a6281499e08209b08f5086 Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Fri, 24 Feb 2006 13:04:14 -0800 Subject: [PATCH] x86: fix broken SMP boot sequence Recent GDT changes broke the SMP boot sequence if the booting CPU is numbered anything other than zero. There's also a subtle source of error in that the boot time CPU now uses cpu_gdt_table (which is actually the GDT for booting CPUs in head.S). This patch fixes both problems by making GDT descriptors themselves allocated from a per_cpu area and switching to them in cpu_init(), which now means that cpu_gdt_table is exclusively used for booting CPUs again. Signed-off-by: James Bottomley Cc: Zachary Amsden Cc: Matt Tolentino Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/i386/kernel/cpu/common.c | 32 ++++++++++++++++++++++++++++---- arch/i386/kernel/efi.c | 12 +++++++----- arch/i386/kernel/head.S | 2 -- arch/i386/kernel/i386_ksyms.c | 2 -- arch/i386/kernel/smpboot.c | 6 ------ include/asm-i386/desc.h | 6 ++++-- 6 files changed, 39 insertions(+), 21 deletions(-) (limited to 'include') diff --git a/arch/i386/kernel/cpu/common.c b/arch/i386/kernel/cpu/common.c index 7eb9213734a3..4ecd4b326ded 100644 --- a/arch/i386/kernel/cpu/common.c +++ b/arch/i386/kernel/cpu/common.c @@ -4,6 +4,7 @@ #include #include #include +#include #include #include #include @@ -18,6 +19,9 @@ #include "cpu.h" +DEFINE_PER_CPU(struct Xgt_desc_struct, cpu_gdt_descr); +EXPORT_PER_CPU_SYMBOL(cpu_gdt_descr); + DEFINE_PER_CPU(unsigned char, cpu_16bit_stack[CPU_16BIT_STACK_SIZE]); EXPORT_PER_CPU_SYMBOL(cpu_16bit_stack); @@ -571,8 +575,9 @@ void __devinit cpu_init(void) int cpu = smp_processor_id(); struct tss_struct * t = &per_cpu(init_tss, cpu); struct thread_struct *thread = ¤t->thread; - struct desc_struct *gdt = get_cpu_gdt_table(cpu); + struct desc_struct *gdt; __u32 stk16_off = (__u32)&per_cpu(cpu_16bit_stack, cpu); + struct Xgt_desc_struct *cpu_gdt_descr = &per_cpu(cpu_gdt_descr, cpu); if (cpu_test_and_set(cpu, cpu_initialized)) { printk(KERN_WARNING "CPU#%d already initialized!\n", cpu); @@ -589,6 +594,25 @@ void __devinit cpu_init(void) set_in_cr4(X86_CR4_TSD); } + /* + * This is a horrible hack to allocate the GDT. The problem + * is that cpu_init() is called really early for the boot CPU + * (and hence needs bootmem) but much later for the secondary + * CPUs, when bootmem will have gone away + */ + if (NODE_DATA(0)->bdata->node_bootmem_map) { + gdt = (struct desc_struct *)alloc_bootmem_pages(PAGE_SIZE); + /* alloc_bootmem_pages panics on failure, so no check */ + memset(gdt, 0, PAGE_SIZE); + } else { + gdt = (struct desc_struct *)get_zeroed_page(GFP_KERNEL); + if (unlikely(!gdt)) { + printk(KERN_CRIT "CPU%d failed to allocate GDT\n", cpu); + for (;;) + local_irq_enable(); + } + } + /* * Initialize the per-CPU GDT with the boot GDT, * and set up the GDT descriptor: @@ -601,10 +625,10 @@ void __devinit cpu_init(void) ((((__u64)stk16_off) << 32) & 0xff00000000000000ULL) | (CPU_16BIT_STACK_SIZE - 1); - cpu_gdt_descr[cpu].size = GDT_SIZE - 1; - cpu_gdt_descr[cpu].address = (unsigned long)gdt; + cpu_gdt_descr->size = GDT_SIZE - 1; + cpu_gdt_descr->address = (unsigned long)gdt; - load_gdt(&cpu_gdt_descr[cpu]); + load_gdt(cpu_gdt_descr); load_idt(&idt_descr); /* diff --git a/arch/i386/kernel/efi.c b/arch/i386/kernel/efi.c index ecad519fd395..e3e42fd62401 100644 --- a/arch/i386/kernel/efi.c +++ b/arch/i386/kernel/efi.c @@ -103,17 +103,19 @@ static void efi_call_phys_prelog(void) */ local_flush_tlb(); - cpu_gdt_descr[0].address = __pa(cpu_gdt_descr[0].address); - load_gdt((struct Xgt_desc_struct *) __pa(&cpu_gdt_descr[0])); + per_cpu(cpu_gdt_descr, 0).address = + __pa(per_cpu(cpu_gdt_descr, 0).address); + load_gdt((struct Xgt_desc_struct *)__pa(&per_cpu(cpu_gdt_descr, 0))); } static void efi_call_phys_epilog(void) { unsigned long cr4; - cpu_gdt_descr[0].address = - (unsigned long) __va(cpu_gdt_descr[0].address); - load_gdt(&cpu_gdt_descr[0]); + per_cpu(cpu_gdt_descr, 0).address = + (unsigned long)__va(per_cpu(cpu_gdt_descr, 0).address); + load_gdt((struct Xgt_desc_struct *)__va(&per_cpu(cpu_gdt_descr, 0))); + cr4 = read_cr4(); if (cr4 & X86_CR4_PSE) { diff --git a/arch/i386/kernel/head.S b/arch/i386/kernel/head.S index 2bee6499edd9..e0b7c632efbc 100644 --- a/arch/i386/kernel/head.S +++ b/arch/i386/kernel/head.S @@ -534,5 +534,3 @@ ENTRY(cpu_gdt_table) .quad 0x0000000000000000 /* 0xf0 - unused */ .quad 0x0000000000000000 /* 0xf8 - GDT entry 31: double-fault TSS */ - /* Be sure this is zeroed to avoid false validations in Xen */ - .fill PAGE_SIZE_asm / 8 - GDT_ENTRIES,8,0 diff --git a/arch/i386/kernel/i386_ksyms.c b/arch/i386/kernel/i386_ksyms.c index 3999bec50c33..055325056a74 100644 --- a/arch/i386/kernel/i386_ksyms.c +++ b/arch/i386/kernel/i386_ksyms.c @@ -3,8 +3,6 @@ #include #include -EXPORT_SYMBOL_GPL(cpu_gdt_descr); - EXPORT_SYMBOL(__down_failed); EXPORT_SYMBOL(__down_failed_interruptible); EXPORT_SYMBOL(__down_failed_trylock); diff --git a/arch/i386/kernel/smpboot.c b/arch/i386/kernel/smpboot.c index fb00ab7b7612..eba7f53f8b4a 100644 --- a/arch/i386/kernel/smpboot.c +++ b/arch/i386/kernel/smpboot.c @@ -898,12 +898,6 @@ static int __devinit do_boot_cpu(int apicid, int cpu) unsigned long start_eip; unsigned short nmi_high = 0, nmi_low = 0; - if (!cpu_gdt_descr[cpu].address && - !(cpu_gdt_descr[cpu].address = get_zeroed_page(GFP_KERNEL))) { - printk("Failed to allocate GDT for CPU %d\n", cpu); - return 1; - } - ++cpucount; /* diff --git a/include/asm-i386/desc.h b/include/asm-i386/desc.h index 494e73bca095..89b8b82c82b3 100644 --- a/include/asm-i386/desc.h +++ b/include/asm-i386/desc.h @@ -24,11 +24,13 @@ struct Xgt_desc_struct { unsigned short pad; } __attribute__ ((packed)); -extern struct Xgt_desc_struct idt_descr, cpu_gdt_descr[NR_CPUS]; +extern struct Xgt_desc_struct idt_descr; +DECLARE_PER_CPU(struct Xgt_desc_struct, cpu_gdt_descr); + static inline struct desc_struct *get_cpu_gdt_table(unsigned int cpu) { - return ((struct desc_struct *)cpu_gdt_descr[cpu].address); + return (struct desc_struct *)per_cpu(cpu_gdt_descr, cpu).address; } #define load_TR_desc() __asm__ __volatile__("ltr %w0"::"q" (GDT_ENTRY_TSS*8)) -- cgit v1.2.3-58-ga151 From c04030e16dbea2f7581f82cc6688695927f6ac5b Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Fri, 24 Feb 2006 13:04:21 -0800 Subject: [PATCH] flags parameter for linkat I'm currently at the POSIX meeting and one thing covered was the incompatibility of Linux's link() with the POSIX definition. The name. Linux does not follow symlinks, POSIX requires it does. Even if somebody thinks this is a good default behavior we cannot change this because it would break the ABI. But the fact remains that some application might want this behavior. We have one chance to help implementing this without breaking the behavior. For this we could use the new linkat interface which would need a new flags parameter. If the new parameter is AT_SYMLINK_FOLLOW the new behavior could be invoked. I do not want to introduce such a patch now. But we could add the parameter now, just don't use it. The patch below would do this. Can we get this late patch applied before the release more or less fixes the syscall API? Signed-off-by: Ulrich Drepper Signed-off-by: Ralf Baechle Cc: Heiko Carstens Cc: Martin Schwidefsky Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/mips/kernel/scall32-o32.S | 2 +- arch/s390/kernel/compat_wrapper.S | 1 + fs/namei.c | 8 ++++++-- include/linux/syscalls.h | 2 +- 4 files changed, 9 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/arch/mips/kernel/scall32-o32.S b/arch/mips/kernel/scall32-o32.S index d83e033dbc87..2f2dc54b2e26 100644 --- a/arch/mips/kernel/scall32-o32.S +++ b/arch/mips/kernel/scall32-o32.S @@ -626,7 +626,7 @@ einval: li v0, -EINVAL sys sys_fstatat64 4 sys sys_unlinkat 3 sys sys_renameat 4 /* 4295 */ - sys sys_linkat 4 + sys sys_linkat 5 sys sys_symlinkat 3 sys sys_readlinkat 4 sys sys_fchmodat 3 diff --git a/arch/s390/kernel/compat_wrapper.S b/arch/s390/kernel/compat_wrapper.S index 615964cca15f..50e80138e7ad 100644 --- a/arch/s390/kernel/compat_wrapper.S +++ b/arch/s390/kernel/compat_wrapper.S @@ -1552,6 +1552,7 @@ sys_linkat_wrapper: llgtr %r3,%r3 # const char * lgfr %r4,%r4 # int llgtr %r5,%r5 # const char * + lgfr %r6,%r6 # int jg sys_linkat .globl sys_symlinkat_wrapper diff --git a/fs/namei.c b/fs/namei.c index e28de846c591..557dcf395ca1 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -2224,13 +2224,17 @@ int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_de * and other special files. --ADM */ asmlinkage long sys_linkat(int olddfd, const char __user *oldname, - int newdfd, const char __user *newname) + int newdfd, const char __user *newname, + int flags) { struct dentry *new_dentry; struct nameidata nd, old_nd; int error; char * to; + if (flags != 0) + return -EINVAL; + to = getname(newname); if (IS_ERR(to)) return PTR_ERR(to); @@ -2263,7 +2267,7 @@ exit: asmlinkage long sys_link(const char __user *oldname, const char __user *newname) { - return sys_linkat(AT_FDCWD, oldname, AT_FDCWD, newname); + return sys_linkat(AT_FDCWD, oldname, AT_FDCWD, newname, 0); } /* diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index d73501ba7e44..b9ea44ac0ddb 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -543,7 +543,7 @@ asmlinkage long sys_unlinkat(int dfd, const char __user * pathname, int flag); asmlinkage long sys_symlinkat(const char __user * oldname, int newdfd, const char __user * newname); asmlinkage long sys_linkat(int olddfd, const char __user *oldname, - int newdfd, const char __user *newname); + int newdfd, const char __user *newname, int flags); asmlinkage long sys_renameat(int olddfd, const char __user * oldname, int newdfd, const char __user * newname); asmlinkage long sys_futimesat(int dfd, char __user *filename, -- cgit v1.2.3-58-ga151 From 60b08c67220cf6faef7410ac6adba23a8a743bf7 Mon Sep 17 00:00:00 2001 From: Jon Mason Date: Sun, 26 Feb 2006 04:18:22 +0100 Subject: [PATCH] x86_64: no_iommu removal in pci-gart.c In previous versions of pci-gart.c, no_iommu was used to determine if IOMMU was disabled in the GART DMA mapping functions. This changed in 2.6.16 and now gart_xxx() functions are only called if gart is enabled. Therefore, uses of no_iommu in the GART code are no longer necessary and can be removed. Also, it removes double deceleration of no_iommu and force_iommu in pci.h and proto.h, by removing the deceleration in pci.h. Lastly, end_pfn off by one error. Tested (along with patch 1/2) on dual opteron with gart enabled, iommu=soft, and iommu=off. Signed-off-by: Jon Mason Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/kernel/aperture.c | 2 +- arch/x86_64/kernel/pci-gart.c | 22 +++++----------------- include/asm-x86_64/pci.h | 2 -- 3 files changed, 6 insertions(+), 20 deletions(-) (limited to 'include') diff --git a/arch/x86_64/kernel/aperture.c b/arch/x86_64/kernel/aperture.c index e4e2b7d01f89..a0f955b9995f 100644 --- a/arch/x86_64/kernel/aperture.c +++ b/arch/x86_64/kernel/aperture.c @@ -248,7 +248,7 @@ void __init iommu_hole_init(void) /* Got the aperture from the AGP bridge */ } else if (swiotlb && !valid_agp) { /* Do nothing */ - } else if ((!no_iommu && end_pfn >= MAX_DMA32_PFN) || + } else if ((!no_iommu && end_pfn > MAX_DMA32_PFN) || force_iommu || valid_agp || fallback_aper_force) { diff --git a/arch/x86_64/kernel/pci-gart.c b/arch/x86_64/kernel/pci-gart.c index dd0718dc178b..0c3f052ba6ce 100644 --- a/arch/x86_64/kernel/pci-gart.c +++ b/arch/x86_64/kernel/pci-gart.c @@ -228,11 +228,6 @@ static inline int need_iommu(struct device *dev, unsigned long addr, size_t size int mmu = high; if (force_iommu) mmu = 1; - if (no_iommu) { - if (high) - panic("PCI-DMA: high address but no IOMMU.\n"); - mmu = 0; - } return mmu; } @@ -241,11 +236,6 @@ static inline int nonforced_iommu(struct device *dev, unsigned long addr, size_t u64 mask = *dev->dma_mask; int high = addr + size >= mask; int mmu = high; - if (no_iommu) { - if (high) - panic("PCI-DMA: high address but no IOMMU.\n"); - mmu = 0; - } return mmu; } @@ -310,7 +300,7 @@ void gart_unmap_sg(struct device *dev, struct scatterlist *sg, int nents, int di for (i = 0; i < nents; i++) { struct scatterlist *s = &sg[i]; - if (!s->dma_length) + if (!s->dma_length || !s->length) break; dma_unmap_single(dev, s->dma_address, s->dma_length, dir); } @@ -364,6 +354,7 @@ static int __dma_map_cont(struct scatterlist *sg, int start, int stopat, BUG_ON(i > start && s->offset); if (i == start) { + *sout = *s; sout->dma_address = iommu_bus_base; sout->dma_address += iommu_page*PAGE_SIZE + s->offset; sout->dma_length = s->length; @@ -390,6 +381,7 @@ static inline int dma_map_cont(struct scatterlist *sg, int start, int stopat, { if (!need) { BUG_ON(stopat - start != 1); + *sout = sg[start]; sout->dma_length = sg[start].length; return 0; } @@ -632,17 +624,13 @@ static int __init pci_iommu_init(void) (agp_copy_info(agp_bridge, &info) < 0); #endif - if (swiotlb) { - no_iommu = 1; + if (swiotlb) return -1; - } - + if (no_iommu || (!force_iommu && end_pfn <= MAX_DMA32_PFN) || !iommu_aperture || (no_agp && init_k8_gatt(&info) < 0)) { - no_iommu = 1; - no_iommu_init(); printk(KERN_INFO "PCI-DMA: Disabling IOMMU.\n"); if (end_pfn > MAX_DMA32_PFN) { printk(KERN_ERR "WARNING more than 4GB of memory " diff --git a/include/asm-x86_64/pci.h b/include/asm-x86_64/pci.h index fd03e15d7ea6..8a05af264d18 100644 --- a/include/asm-x86_64/pci.h +++ b/include/asm-x86_64/pci.h @@ -19,8 +19,6 @@ extern unsigned int pcibios_assign_all_busses(void); #endif #define pcibios_scan_all_fns(a, b) 0 -extern int no_iommu, force_iommu; - extern unsigned long pci_mem_start; #define PCIBIOS_MIN_IO 0x1000 #define PCIBIOS_MIN_MEM (pci_mem_start) -- cgit v1.2.3-58-ga151 From f83f2b5fbab4585f4de4523c7879d60e3f85a248 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Sun, 26 Feb 2006 04:18:25 +0100 Subject: [PATCH] x86_64: fix USER_PTRS_PER_PGD The value, while currently unused in the native kernel, was off by one. Signed-Off-By: Jan Beulich Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- include/asm-x86_64/pgtable.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include') diff --git a/include/asm-x86_64/pgtable.h b/include/asm-x86_64/pgtable.h index 8fbf4dd72115..715fd94cf577 100644 --- a/include/asm-x86_64/pgtable.h +++ b/include/asm-x86_64/pgtable.h @@ -131,7 +131,7 @@ static inline pte_t ptep_get_and_clear_full(struct mm_struct *mm, unsigned long #define PGDIR_SIZE (1UL << PGDIR_SHIFT) #define PGDIR_MASK (~(PGDIR_SIZE-1)) -#define USER_PTRS_PER_PGD (TASK_SIZE/PGDIR_SIZE) +#define USER_PTRS_PER_PGD ((TASK_SIZE-1)/PGDIR_SIZE+1) #define FIRST_USER_ADDRESS 0 #ifndef __ASSEMBLY__ -- cgit v1.2.3-58-ga151 From 13a229abc25640813f1480c0478dfc6bdbc1c19e Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sun, 26 Feb 2006 04:18:31 +0100 Subject: [PATCH] x86_64: Only do the clustered systems have unsynchronized TSC assumption on IBM systems Big Unisys systems have multiple clusters too, but they have an synchronized TSC. I'm using the SMBIOS to check for vendor == IBM. Cc: Chris McDermott Cc: "Protasevich, Natalie" Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/i386/kernel/acpi/boot.c | 3 --- arch/x86_64/kernel/apic.c | 9 ++++++++- include/asm-x86_64/acpi.h | 14 ++++++++++++++ 3 files changed, 22 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/arch/i386/kernel/acpi/boot.c b/arch/i386/kernel/acpi/boot.c index 79577f0ace98..8309a7b2cd63 100644 --- a/arch/i386/kernel/acpi/boot.c +++ b/arch/i386/kernel/acpi/boot.c @@ -44,9 +44,6 @@ extern void __init clustered_apic_check(void); extern int gsi_irq_sharing(int gsi); #include -static inline int acpi_madt_oem_check(char *oem_id, char *oem_table_id) { return 0; } - - #else /* X86 */ #ifdef CONFIG_X86_LOCAL_APIC diff --git a/arch/x86_64/kernel/apic.c b/arch/x86_64/kernel/apic.c index e5b14c57eaa0..d70605eda333 100644 --- a/arch/x86_64/kernel/apic.c +++ b/arch/x86_64/kernel/apic.c @@ -962,12 +962,14 @@ void smp_apic_timer_interrupt(struct pt_regs *regs) irq_exit(); } +int __initdata unsync_tsc_on_multicluster; + /* * oem_force_hpet_timer -- force HPET mode for some boxes. * * Thus far, the major user of this is IBM's Summit2 series: * - * Clustered boxes may have unsynced TSC problems if they are + * Some clustered boxes may have unsynced TSC problems if they are * multi-chassis. Use available data to take a good guess. * If in doubt, go HPET. */ @@ -977,6 +979,11 @@ __cpuinit int oem_force_hpet_timer(void) unsigned id; DECLARE_BITMAP(clustermap, NUM_APIC_CLUSTERS); + /* Only do this check on IBM machines - big Unisys systems + use multiple clusters too, but have synchronized TSC */ + if (!unsync_tsc_on_multicluster) + return 0; + bitmap_zero(clustermap, NUM_APIC_CLUSTERS); for (i = 0; i < NR_CPUS; i++) { diff --git a/include/asm-x86_64/acpi.h b/include/asm-x86_64/acpi.h index aa1c7b2e438c..e2b9923189a0 100644 --- a/include/asm-x86_64/acpi.h +++ b/include/asm-x86_64/acpi.h @@ -164,6 +164,20 @@ extern u8 x86_acpiid_to_apicid[]; extern int acpi_skip_timer_override; +extern int unsync_tsc_on_multicluster; + +static inline int acpi_madt_oem_check(char *oem, char *productid) +{ + /* Copied from i386. Probably has too many entries. */ + if (!strncmp(oem, "IBM ENSW", 8) && + (!strncmp(productid, "VIGIL SMP", 9) + || !strncmp(productid, "EXA", 3) + || !strncmp(productid, "RUTHLESS SMP", 12))) { + unsync_tsc_on_multicluster = 1; + } + return 0; +} + #endif /*__KERNEL__*/ #endif /*_ASM_ACPI_H*/ -- cgit v1.2.3-58-ga151 From e2c0388866dc12bef56b178b958f9b778fe6c687 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sun, 26 Feb 2006 04:18:46 +0100 Subject: [PATCH] x86_64: Fix the additional_cpus=.. option It didn't set up the CPU possible map early enough, so the option didn't actually work. Noticed by Heiko Carstens Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/kernel/setup.c | 6 ++++++ arch/x86_64/kernel/smpboot.c | 2 +- include/asm-x86_64/proto.h | 1 + 3 files changed, 8 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/arch/x86_64/kernel/setup.c b/arch/x86_64/kernel/setup.c index 9435ab7d6fb8..5de7eaf5d97c 100644 --- a/arch/x86_64/kernel/setup.c +++ b/arch/x86_64/kernel/setup.c @@ -423,6 +423,12 @@ static __init void parse_cmdline_early (char ** cmdline_p) else if(!memcmp(from, "elfcorehdr=", 11)) elfcorehdr_addr = memparse(from+11, &from); #endif + +#ifdef CONFIG_SMP + else if (!memcmp(from, "additional_cpus=", 16)) + setup_additional_cpus(from+16); +#endif + next_char: c = *(from++); if (!c) diff --git a/arch/x86_64/kernel/smpboot.c b/arch/x86_64/kernel/smpboot.c index 67e4e28f4df8..b82eb86e4f5d 100644 --- a/arch/x86_64/kernel/smpboot.c +++ b/arch/x86_64/kernel/smpboot.c @@ -1244,7 +1244,7 @@ void __cpu_die(unsigned int cpu) printk(KERN_ERR "CPU %u didn't die...\n", cpu); } -static __init int setup_additional_cpus(char *s) +__init int setup_additional_cpus(char *s) { return get_option(&s, &additional_cpus); } diff --git a/include/asm-x86_64/proto.h b/include/asm-x86_64/proto.h index eca3f2d633db..8bdcbd0aa03f 100644 --- a/include/asm-x86_64/proto.h +++ b/include/asm-x86_64/proto.h @@ -134,6 +134,7 @@ extern int force_iommu; extern int reboot_force; extern int notsc_setup(char *); +extern int setup_additional_cpus(char *); extern void smp_local_timer_interrupt(struct pt_regs * regs); -- cgit v1.2.3-58-ga151 From e8b917775b572bc27de105f1317c2de4335db5b3 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sun, 26 Feb 2006 04:18:49 +0100 Subject: [PATCH] x86_64: Move the SMP time selection earlier SMP time selection originally ran after all CPUs were brought up because it needed to know the number of CPUs to decide if it needs an MP safe timer or not. This is not needed anymore because we know present CPUs early. This fixes a couple of problems: - apicmaintimer didn't always work because it relied on state that was set up time_init_gtod too late. - The output for the used timer in early kernel log was misleading because time_init_gtod could actually change it later. Now always print the final timer choice Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/kernel/smpboot.c | 2 -- arch/x86_64/kernel/time.c | 22 +++++++++++----------- include/asm-x86_64/proto.h | 1 - 3 files changed, 11 insertions(+), 14 deletions(-) (limited to 'include') diff --git a/arch/x86_64/kernel/smpboot.c b/arch/x86_64/kernel/smpboot.c index b82eb86e4f5d..66e98659d077 100644 --- a/arch/x86_64/kernel/smpboot.c +++ b/arch/x86_64/kernel/smpboot.c @@ -1152,8 +1152,6 @@ void __init smp_cpus_done(unsigned int max_cpus) setup_ioapic_dest(); #endif - time_init_gtod(); - check_nmi_watchdog(); } diff --git a/arch/x86_64/kernel/time.c b/arch/x86_64/kernel/time.c index 67841d11ed1f..3080f84bf7b7 100644 --- a/arch/x86_64/kernel/time.c +++ b/arch/x86_64/kernel/time.c @@ -48,6 +48,8 @@ static void cpufreq_delayed_get(void); extern void i8254_timer_resume(void); extern int using_apic_timer; +static char *time_init_gtod(void); + DEFINE_SPINLOCK(rtc_lock); DEFINE_SPINLOCK(i8253_lock); @@ -901,6 +903,7 @@ static struct irqaction irq0 = { void __init time_init(void) { char *timename; + char *gtod; #ifdef HPET_HACK_ENABLE_DANGEROUS if (!vxtime.hpet_address) { @@ -945,21 +948,19 @@ void __init time_init(void) timename = "PIT"; } - printk(KERN_INFO "time.c: Using %ld.%06ld MHz %s timer.\n", - vxtime_hz / 1000000, vxtime_hz % 1000000, timename); + vxtime.mode = VXTIME_TSC; + gtod = time_init_gtod(); + + printk(KERN_INFO "time.c: Using %ld.%06ld MHz WALL %s GTOD %s timer.\n", + vxtime_hz / 1000000, vxtime_hz % 1000000, timename, gtod); printk(KERN_INFO "time.c: Detected %d.%03d MHz processor.\n", cpu_khz / 1000, cpu_khz % 1000); - vxtime.mode = VXTIME_TSC; vxtime.quot = (1000000L << 32) / vxtime_hz; vxtime.tsc_quot = (1000L << 32) / cpu_khz; vxtime.last_tsc = get_cycles_sync(); setup_irq(0, &irq0); set_cyc2ns_scale(cpu_khz); - -#ifndef CONFIG_SMP - time_init_gtod(); -#endif } /* @@ -981,9 +982,9 @@ __cpuinit int unsynchronized_tsc(void) } /* - * Decide after all CPUs are booted what mode gettimeofday should use. + * Decide what mode gettimeofday should use. */ -void __init time_init_gtod(void) +__init static char *time_init_gtod(void) { char *timetype; @@ -1011,8 +1012,7 @@ void __init time_init_gtod(void) timetype = hpet_use_timer ? "HPET/TSC" : "PIT/TSC"; vxtime.mode = VXTIME_TSC; } - - printk(KERN_INFO "time.c: Using %s based timekeeping.\n", timetype); + return timetype; } __setup("report_lost_ticks", time_setup); diff --git a/include/asm-x86_64/proto.h b/include/asm-x86_64/proto.h index 8bdcbd0aa03f..3ba8fd45fcb3 100644 --- a/include/asm-x86_64/proto.h +++ b/include/asm-x86_64/proto.h @@ -39,7 +39,6 @@ extern void config_acpi_tables(void); extern void ia32_syscall(void); extern void iommu_hole_init(void); -extern void time_init_gtod(void); extern int pmtimer_mark_offset(void); extern void pmtimer_resume(void); extern void pmtimer_wait(unsigned); -- cgit v1.2.3-58-ga151 From 043df59eb3798c094e6ba47136f3d3b34a6791a7 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Sat, 25 Feb 2006 12:15:31 -0800 Subject: [SPARC64]: Implement futex_atomic_op_inuser(). Signed-off-by: David S. Miller --- include/asm-sparc64/futex.h | 88 ++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 84 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/include/asm-sparc64/futex.h b/include/asm-sparc64/futex.h index 6a332a9f099c..0caf60147e97 100644 --- a/include/asm-sparc64/futex.h +++ b/include/asm-sparc64/futex.h @@ -1,6 +1,86 @@ -#ifndef _ASM_FUTEX_H -#define _ASM_FUTEX_H +#ifndef _SPARC64_FUTEX_H +#define _SPARC64_FUTEX_H -#include +#include +#include +#include +#include -#endif +#define __futex_cas_op(insn, ret, oldval, uaddr, oparg) \ + __asm__ __volatile__( \ + "\n1: lduwa [%3] %%asi, %2\n" \ + " " insn "\n" \ + "2: casa [%3] %%asi, %2, %1\n" \ + " cmp %2, %1\n" \ + " bne,pn %%icc, 1b\n" \ + " mov 0, %0\n" \ + "3:\n" \ + " .section .fixup,#alloc,#execinstr\n" \ + " .align 4\n" \ + "4: ba 3b\n" \ + " mov %5, %0\n" \ + " .previous\n" \ + " .section __ex_table,#alloc\n" \ + " .align 4\n" \ + " .word 1b, 4b\n" \ + " .word 2b, 4b\n" \ + " .previous\n" \ + : "=&r" (ret), "=&r" (oldval), "=&r" (tem) \ + : "r" (uaddr), "r" (oparg), "i" (-EFAULT) \ + : "memory") + +static inline int futex_atomic_op_inuser(int encoded_op, int __user *uaddr) +{ + int op = (encoded_op >> 28) & 7; + int cmp = (encoded_op >> 24) & 15; + int oparg = (encoded_op << 8) >> 20; + int cmparg = (encoded_op << 20) >> 20; + int oldval = 0, ret, tem; + + if (unlikely(!access_ok(VERIFY_WRITE, uaddr, sizeof(int)))) + return -EFAULT; + if (unlikely((((unsigned long) uaddr) & 0x3UL))) + return -EINVAL; + + if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28)) + oparg = 1 << oparg; + + inc_preempt_count(); + + switch (op) { + case FUTEX_OP_SET: + __futex_cas_op("mov\t%4, %1", ret, oldval, uaddr, oparg); + break; + case FUTEX_OP_ADD: + __futex_cas_op("add\t%2, %4, %1", ret, oldval, uaddr, oparg); + break; + case FUTEX_OP_OR: + __futex_cas_op("or\t%2, %4, %1", ret, oldval, uaddr, oparg); + break; + case FUTEX_OP_ANDN: + __futex_cas_op("and\t%2, %4, %1", ret, oldval, uaddr, oparg); + break; + case FUTEX_OP_XOR: + __futex_cas_op("xor\t%2, %4, %1", ret, oldval, uaddr, oparg); + break; + default: + ret = -ENOSYS; + } + + dec_preempt_count(); + + if (!ret) { + switch (cmp) { + case FUTEX_OP_CMP_EQ: ret = (oldval == cmparg); break; + case FUTEX_OP_CMP_NE: ret = (oldval != cmparg); break; + case FUTEX_OP_CMP_LT: ret = (oldval < cmparg); break; + case FUTEX_OP_CMP_GE: ret = (oldval >= cmparg); break; + case FUTEX_OP_CMP_LE: ret = (oldval <= cmparg); break; + case FUTEX_OP_CMP_GT: ret = (oldval > cmparg); break; + default: ret = -ENOSYS; + } + } + return ret; +} + +#endif /* !(_SPARC64_FUTEX_H) */ -- cgit v1.2.3-58-ga151 From 7abea9214585823f7f19d91872d7c6f8874bef9a Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Sat, 25 Feb 2006 13:39:56 -0800 Subject: [SPARC64]: Make cpu_present_map available earlier. The change to kernel/sched.c's init code to use for_each_cpu() requires that the cpu_possible_map be setup much earlier. Set it up via setup_arch(), constrained to NR_CPUS, and later constrain it to max_cpus in smp_prepare_cpus(). This fixes SMP booting on sparc64. Signed-off-by: David S. Miller --- arch/sparc64/kernel/setup.c | 2 ++ arch/sparc64/kernel/smp.c | 28 +++++++++++++++++++--------- include/asm-sparc64/smp.h | 6 ++++++ 3 files changed, 27 insertions(+), 9 deletions(-) (limited to 'include') diff --git a/arch/sparc64/kernel/setup.c b/arch/sparc64/kernel/setup.c index 054461e6946d..158bd31e15b7 100644 --- a/arch/sparc64/kernel/setup.c +++ b/arch/sparc64/kernel/setup.c @@ -542,6 +542,8 @@ void __init setup_arch(char **cmdline_p) } #endif + smp_setup_cpu_possible_map(); + paging_init(); } diff --git a/arch/sparc64/kernel/smp.c b/arch/sparc64/kernel/smp.c index 1fb6323e65a4..1f7ad8a69052 100644 --- a/arch/sparc64/kernel/smp.c +++ b/arch/sparc64/kernel/smp.c @@ -1079,18 +1079,12 @@ int setup_profiling_timer(unsigned int multiplier) return 0; } +/* Constrain the number of cpus to max_cpus. */ void __init smp_prepare_cpus(unsigned int max_cpus) { - int instance, mid; - - instance = 0; - while (!cpu_find_by_instance(instance, NULL, &mid)) { - if (mid < max_cpus) - cpu_set(mid, phys_cpu_present_map); - instance++; - } - if (num_possible_cpus() > max_cpus) { + int instance, mid; + instance = 0; while (!cpu_find_by_instance(instance, NULL, &mid)) { if (mid != boot_cpu_id) { @@ -1105,6 +1099,22 @@ void __init smp_prepare_cpus(unsigned int max_cpus) smp_store_cpu_info(boot_cpu_id); } +/* Set this up early so that things like the scheduler can init + * properly. We use the same cpu mask for both the present and + * possible cpu map. + */ +void __init smp_setup_cpu_possible_map(void) +{ + int instance, mid; + + instance = 0; + while (!cpu_find_by_instance(instance, NULL, &mid)) { + if (mid < NR_CPUS) + cpu_set(mid, phys_cpu_present_map); + instance++; + } +} + void __devinit smp_prepare_boot_cpu(void) { if (hard_smp_processor_id() >= NR_CPUS) { diff --git a/include/asm-sparc64/smp.h b/include/asm-sparc64/smp.h index 110a2de89123..473edb2603ec 100644 --- a/include/asm-sparc64/smp.h +++ b/include/asm-sparc64/smp.h @@ -66,8 +66,14 @@ static __inline__ int hard_smp_processor_id(void) #define raw_smp_processor_id() (current_thread_info()->cpu) +extern void smp_setup_cpu_possible_map(void); + #endif /* !(__ASSEMBLY__) */ +#else + +#define smp_setup_cpu_possible_map() do { } while (0) + #endif /* !(CONFIG_SMP) */ #define NO_PROC_ID 0xFF -- cgit v1.2.3-58-ga151