From 395810627b6a43c8d0ec948884043946fa162308 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Wed, 8 Jun 2011 16:09:21 +0900 Subject: x86: Swap save_stack_trace_regs parameters Swap the 1st and 2nd parameters of save_stack_trace_regs() as same as the parameters of save_stack_trace_tsk(). Signed-off-by: Masami Hiramatsu Cc: yrl.pp-manager.tt@hitachi.com Cc: Frederic Weisbecker Cc: Peter Zijlstra Cc: Namhyung Kim Link: http://lkml.kernel.org/r/20110608070921.17777.31103.stgit@fedora15 Signed-off-by: Steven Rostedt --- arch/x86/kernel/stacktrace.c | 2 +- arch/x86/mm/kmemcheck/error.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'arch') diff --git a/arch/x86/kernel/stacktrace.c b/arch/x86/kernel/stacktrace.c index 55d9bc03f696..fdd0c6430e5a 100644 --- a/arch/x86/kernel/stacktrace.c +++ b/arch/x86/kernel/stacktrace.c @@ -66,7 +66,7 @@ void save_stack_trace(struct stack_trace *trace) } EXPORT_SYMBOL_GPL(save_stack_trace); -void save_stack_trace_regs(struct stack_trace *trace, struct pt_regs *regs) +void save_stack_trace_regs(struct pt_regs *regs, struct stack_trace *trace) { dump_trace(current, regs, NULL, 0, &save_stack_ops, trace); if (trace->nr_entries < trace->max_entries) diff --git a/arch/x86/mm/kmemcheck/error.c b/arch/x86/mm/kmemcheck/error.c index 704a37cedddb..dab41876cdd5 100644 --- a/arch/x86/mm/kmemcheck/error.c +++ b/arch/x86/mm/kmemcheck/error.c @@ -185,7 +185,7 @@ void kmemcheck_error_save(enum kmemcheck_shadow state, e->trace.entries = e->trace_entries; e->trace.max_entries = ARRAY_SIZE(e->trace_entries); e->trace.skip = 0; - save_stack_trace_regs(&e->trace, regs); + save_stack_trace_regs(regs, &e->trace); /* Round address down to nearest 16 bytes */ shadow_copy = kmemcheck_shadow_lookup(address -- cgit v1.2.3 From a0e3e70243f5b270bc3eca718f0a9fa5e6b8262e Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Fri, 3 Jun 2011 16:37:47 +0200 Subject: oprofile, x86: Fix nmi-unsafe callgraph support Current oprofile's x86 callgraph support may trigger page faults throwing the BUG_ON(in_nmi()) message below. This patch fixes this by using the same nmi-safe copy-from-user code as in perf. ------------[ cut here ]------------ kernel BUG at .../arch/x86/kernel/traps.c:436! invalid opcode: 0000 [#1] SMP last sysfs file: /sys/devices/pci0000:00/0000:00:0a.0/0000:07:00.0/0000:08:04.0/net/eth0/broadcast CPU 5 Modules linked in: Pid: 8611, comm: opcontrol Not tainted 2.6.39-00007-gfe47ae7 #1 Advanced Micro Device Anaheim/Anaheim RIP: 0010:[] [] do_nmi+0x22/0x1ee RSP: 0000:ffff88042fd47f28 EFLAGS: 00010002 RAX: ffff88042c0a7fd8 RBX: 0000000000000001 RCX: 00000000c0000101 RDX: 00000000ffff8804 RSI: ffffffffffffffff RDI: ffff88042fd47f58 RBP: ffff88042fd47f48 R08: 0000000000000004 R09: 0000000000001484 R10: 0000000000000001 R11: 0000000000000000 R12: ffff88042fd47f58 R13: 0000000000000000 R14: ffff88042fd47d98 R15: 0000000000000020 FS: 00007fca25e56700(0000) GS:ffff88042fd40000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000074 CR3: 000000042d28b000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process opcontrol (pid: 8611, threadinfo ffff88042c0a6000, task ffff88042c532310) Stack: 0000000000000000 0000000000000001 ffff88042c0a7fd8 0000000000000000 ffff88042fd47de8 ffffffff813e897a 0000000000000020 ffff88042fd47d98 0000000000000000 ffff88042c0a7fd8 ffff88042fd47de8 0000000000000074 Call Trace: [] nmi+0x1a/0x20 [] ? bad_to_user+0x25/0x771 <> Code: ff 59 5b 41 5c 41 5d c9 c3 55 65 48 8b 04 25 88 b5 00 00 48 89 e5 41 55 41 54 49 89 fc 53 48 83 ec 08 f6 80 47 e0 ff ff 04 74 04 <0f> 0b eb fe 81 80 44 e0 ff ff 00 00 01 04 65 ff 04 25 c4 0f 01 RIP [] do_nmi+0x22/0x1ee RSP ---[ end trace ed6752185092104b ]--- Kernel panic - not syncing: Fatal exception in interrupt Pid: 8611, comm: opcontrol Tainted: G D 2.6.39-00007-gfe47ae7 #1 Call Trace: [] panic+0x8c/0x188 [] oops_end+0x81/0x8e [] die+0x55/0x5e [] do_trap+0x11c/0x12b [] do_invalid_op+0x91/0x9a [] ? do_nmi+0x22/0x1ee [] ? oprofile_add_sample+0x83/0x95 [] ? op_amd_check_ctrs+0x4f/0x2cf [] invalid_op+0x15/0x20 [] ? do_nmi+0x22/0x1ee [] ? do_nmi+0x67/0x1ee [] nmi+0x1a/0x20 [] ? bad_to_user+0x25/0x771 <> Cc: John Lumby Cc: Maynard Johnson Cc: # .37+ Signed-off-by: Robert Richter --- arch/x86/oprofile/backtrace.c | 56 ++++++++++++++++++++++++++++++++++++------- 1 file changed, 47 insertions(+), 9 deletions(-) (limited to 'arch') diff --git a/arch/x86/oprofile/backtrace.c b/arch/x86/oprofile/backtrace.c index a5b64ab4cd6e..32f78eb46744 100644 --- a/arch/x86/oprofile/backtrace.c +++ b/arch/x86/oprofile/backtrace.c @@ -11,10 +11,12 @@ #include #include #include +#include +#include + #include #include #include -#include static int backtrace_stack(void *data, char *name) { @@ -36,17 +38,53 @@ static struct stacktrace_ops backtrace_ops = { .walk_stack = print_context_stack, }; +/* from arch/x86/kernel/cpu/perf_event.c: */ + +/* + * best effort, GUP based copy_from_user() that assumes IRQ or NMI context + */ +static unsigned long +copy_from_user_nmi(void *to, const void __user *from, unsigned long n) +{ + unsigned long offset, addr = (unsigned long)from; + unsigned long size, len = 0; + struct page *page; + void *map; + int ret; + + do { + ret = __get_user_pages_fast(addr, 1, 0, &page); + if (!ret) + break; + + offset = addr & (PAGE_SIZE - 1); + size = min(PAGE_SIZE - offset, n - len); + + map = kmap_atomic(page); + memcpy(to, map+offset, size); + kunmap_atomic(map); + put_page(page); + + len += size; + to += size; + addr += size; + + } while (len < n); + + return len; +} + #ifdef CONFIG_COMPAT static struct stack_frame_ia32 * dump_user_backtrace_32(struct stack_frame_ia32 *head) { + /* Also check accessibility of one struct frame_head beyond: */ struct stack_frame_ia32 bufhead[2]; struct stack_frame_ia32 *fp; + unsigned long bytes; - /* Also check accessibility of one struct frame_head beyond */ - if (!access_ok(VERIFY_READ, head, sizeof(bufhead))) - return NULL; - if (__copy_from_user_inatomic(bufhead, head, sizeof(bufhead))) + bytes = copy_from_user_nmi(bufhead, head, sizeof(bufhead)); + if (bytes != sizeof(bufhead)) return NULL; fp = (struct stack_frame_ia32 *) compat_ptr(bufhead[0].next_frame); @@ -87,12 +125,12 @@ x86_backtrace_32(struct pt_regs * const regs, unsigned int depth) static struct stack_frame *dump_user_backtrace(struct stack_frame *head) { + /* Also check accessibility of one struct frame_head beyond: */ struct stack_frame bufhead[2]; + unsigned long bytes; - /* Also check accessibility of one struct stack_frame beyond */ - if (!access_ok(VERIFY_READ, head, sizeof(bufhead))) - return NULL; - if (__copy_from_user_inatomic(bufhead, head, sizeof(bufhead))) + bytes = copy_from_user_nmi(bufhead, head, sizeof(bufhead)); + if (bytes != sizeof(bufhead)) return NULL; oprofile_add_trace(bufhead[0].return_address); -- cgit v1.2.3 From 94ea09c6a8e6c5ffb59bb1d5ca10008d37544e1a Mon Sep 17 00:00:00 2001 From: Daniel Kalmar Date: Fri, 13 May 2011 08:38:04 -0400 Subject: MIPS: Add new unwind_stack variant The unwind_stack_by_address variant supports unwinding based on any kernel code address. This symbol is also exported so it can be called from modules. Signed-off-by: Daniel Kalmar Signed-off-by: Gergely Kis Signed-off-by: Robert Richter --- arch/mips/include/asm/stacktrace.h | 4 ++++ arch/mips/kernel/process.c | 19 ++++++++++++++----- 2 files changed, 18 insertions(+), 5 deletions(-) (limited to 'arch') diff --git a/arch/mips/include/asm/stacktrace.h b/arch/mips/include/asm/stacktrace.h index 0bf82818aa53..780ee2c2a2ac 100644 --- a/arch/mips/include/asm/stacktrace.h +++ b/arch/mips/include/asm/stacktrace.h @@ -7,6 +7,10 @@ extern int raw_show_trace; extern unsigned long unwind_stack(struct task_struct *task, unsigned long *sp, unsigned long pc, unsigned long *ra); +extern unsigned long unwind_stack_by_address(unsigned long stack_page, + unsigned long *sp, + unsigned long pc, + unsigned long *ra); #else #define raw_show_trace 1 static inline unsigned long unwind_stack(struct task_struct *task, diff --git a/arch/mips/kernel/process.c b/arch/mips/kernel/process.c index d2112d3cf115..c28fbe6107bc 100644 --- a/arch/mips/kernel/process.c +++ b/arch/mips/kernel/process.c @@ -373,18 +373,18 @@ unsigned long thread_saved_pc(struct task_struct *tsk) #ifdef CONFIG_KALLSYMS -/* used by show_backtrace() */ -unsigned long unwind_stack(struct task_struct *task, unsigned long *sp, - unsigned long pc, unsigned long *ra) +/* generic stack unwinding function */ +unsigned long notrace unwind_stack_by_address(unsigned long stack_page, + unsigned long *sp, + unsigned long pc, + unsigned long *ra) { - unsigned long stack_page; struct mips_frame_info info; unsigned long size, ofs; int leaf; extern void ret_from_irq(void); extern void ret_from_exception(void); - stack_page = (unsigned long)task_stack_page(task); if (!stack_page) return 0; @@ -443,6 +443,15 @@ unsigned long unwind_stack(struct task_struct *task, unsigned long *sp, *ra = 0; return __kernel_text_address(pc) ? pc : 0; } +EXPORT_SYMBOL(unwind_stack_by_address); + +/* used by show_backtrace() */ +unsigned long unwind_stack(struct task_struct *task, unsigned long *sp, + unsigned long pc, unsigned long *ra) +{ + unsigned long stack_page = (unsigned long)task_stack_page(task); + return unwind_stack_by_address(stack_page, sp, pc, ra); +} #endif /* -- cgit v1.2.3 From 5f307491f3a0c8551cedf5d90d660d656e0d56ec Mon Sep 17 00:00:00 2001 From: Daniel Kalmar Date: Fri, 13 May 2011 08:38:05 -0400 Subject: MIPS: oprofile: Add callgraph support Stack unwinding is done by code examination. For kernelspace, the already existing unwind function is utilized that uses kallsyms to quickly find the beginning of functions. For userspace a new function was added that examines code at and before the pc. Signed-off-by: Daniel Kalmar Signed-off-by: Gergely Kis Signed-off-by: Robert Richter --- arch/mips/oprofile/Makefile | 2 +- arch/mips/oprofile/backtrace.c | 175 +++++++++++++++++++++++++++++++++++++++++ arch/mips/oprofile/common.c | 1 + arch/mips/oprofile/op_impl.h | 2 + 4 files changed, 179 insertions(+), 1 deletion(-) create mode 100644 arch/mips/oprofile/backtrace.c (limited to 'arch') diff --git a/arch/mips/oprofile/Makefile b/arch/mips/oprofile/Makefile index 4b9d7044e26c..29f2f13eb31c 100644 --- a/arch/mips/oprofile/Makefile +++ b/arch/mips/oprofile/Makefile @@ -8,7 +8,7 @@ DRIVER_OBJS = $(addprefix ../../../drivers/oprofile/, \ oprofilefs.o oprofile_stats.o \ timer_int.o ) -oprofile-y := $(DRIVER_OBJS) common.o +oprofile-y := $(DRIVER_OBJS) common.o backtrace.o oprofile-$(CONFIG_CPU_MIPS32) += op_model_mipsxx.o oprofile-$(CONFIG_CPU_MIPS64) += op_model_mipsxx.o diff --git a/arch/mips/oprofile/backtrace.c b/arch/mips/oprofile/backtrace.c new file mode 100644 index 000000000000..6854ed5097d2 --- /dev/null +++ b/arch/mips/oprofile/backtrace.c @@ -0,0 +1,175 @@ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +struct stackframe { + unsigned long sp; + unsigned long pc; + unsigned long ra; +}; + +static inline int get_mem(unsigned long addr, unsigned long *result) +{ + unsigned long *address = (unsigned long *) addr; + if (!access_ok(VERIFY_READ, addr, sizeof(unsigned long))) + return -1; + if (__copy_from_user_inatomic(result, address, sizeof(unsigned long))) + return -3; + return 0; +} + +/* + * These two instruction helpers were taken from process.c + */ +static inline int is_ra_save_ins(union mips_instruction *ip) +{ + /* sw / sd $ra, offset($sp) */ + return (ip->i_format.opcode == sw_op || ip->i_format.opcode == sd_op) + && ip->i_format.rs == 29 && ip->i_format.rt == 31; +} + +static inline int is_sp_move_ins(union mips_instruction *ip) +{ + /* addiu/daddiu sp,sp,-imm */ + if (ip->i_format.rs != 29 || ip->i_format.rt != 29) + return 0; + if (ip->i_format.opcode == addiu_op || ip->i_format.opcode == daddiu_op) + return 1; + return 0; +} + +/* + * Looks for specific instructions that mark the end of a function. + * This usually means we ran into the code area of the previous function. + */ +static inline int is_end_of_function_marker(union mips_instruction *ip) +{ + /* jr ra */ + if (ip->r_format.func == jr_op && ip->r_format.rs == 31) + return 1; + /* lui gp */ + if (ip->i_format.opcode == lui_op && ip->i_format.rt == 28) + return 1; + return 0; +} + +/* + * TODO for userspace stack unwinding: + * - handle cases where the stack is adjusted inside a function + * (generally doesn't happen) + * - find optimal value for max_instr_check + * - try to find a way to handle leaf functions + */ + +static inline int unwind_user_frame(struct stackframe *old_frame, + const unsigned int max_instr_check) +{ + struct stackframe new_frame = *old_frame; + off_t ra_offset = 0; + size_t stack_size = 0; + unsigned long addr; + + if (old_frame->pc == 0 || old_frame->sp == 0 || old_frame->ra == 0) + return -9; + + for (addr = new_frame.pc; (addr + max_instr_check > new_frame.pc) + && (!ra_offset || !stack_size); --addr) { + union mips_instruction ip; + + if (get_mem(addr, (unsigned long *) &ip)) + return -11; + + if (is_sp_move_ins(&ip)) { + int stack_adjustment = ip.i_format.simmediate; + if (stack_adjustment > 0) + /* This marks the end of the previous function, + which means we overran. */ + break; + stack_size = (unsigned) stack_adjustment; + } else if (is_ra_save_ins(&ip)) { + int ra_slot = ip.i_format.simmediate; + if (ra_slot < 0) + /* This shouldn't happen. */ + break; + ra_offset = ra_slot; + } else if (is_end_of_function_marker(&ip)) + break; + } + + if (!ra_offset || !stack_size) + return -1; + + if (ra_offset) { + new_frame.ra = old_frame->sp + ra_offset; + if (get_mem(new_frame.ra, &(new_frame.ra))) + return -13; + } + + if (stack_size) { + new_frame.sp = old_frame->sp + stack_size; + if (get_mem(new_frame.sp, &(new_frame.sp))) + return -14; + } + + if (new_frame.sp > old_frame->sp) + return -2; + + new_frame.pc = old_frame->ra; + *old_frame = new_frame; + + return 0; +} + +static inline void do_user_backtrace(unsigned long low_addr, + struct stackframe *frame, + unsigned int depth) +{ + const unsigned int max_instr_check = 512; + const unsigned long high_addr = low_addr + THREAD_SIZE; + + while (depth-- && !unwind_user_frame(frame, max_instr_check)) { + oprofile_add_trace(frame->ra); + if (frame->sp < low_addr || frame->sp > high_addr) + break; + } +} + +#ifndef CONFIG_KALLSYMS +static inline void do_kernel_backtrace(unsigned long low_addr, + struct stackframe *frame, + unsigned int depth) { } +#else +static inline void do_kernel_backtrace(unsigned long low_addr, + struct stackframe *frame, + unsigned int depth) +{ + while (depth-- && frame->pc) { + frame->pc = unwind_stack_by_address(low_addr, + &(frame->sp), + frame->pc, + &(frame->ra)); + oprofile_add_trace(frame->ra); + } +} +#endif + +void notrace op_mips_backtrace(struct pt_regs *const regs, unsigned int depth) +{ + struct stackframe frame = { .sp = regs->regs[29], + .pc = regs->cp0_epc, + .ra = regs->regs[31] }; + const int userspace = user_mode(regs); + const unsigned long low_addr = ALIGN(frame.sp, THREAD_SIZE); + + if (userspace) + do_user_backtrace(low_addr, &frame, depth); + else + do_kernel_backtrace(low_addr, &frame, depth); +} diff --git a/arch/mips/oprofile/common.c b/arch/mips/oprofile/common.c index f9eb1aba6345..d1f2d4c52d42 100644 --- a/arch/mips/oprofile/common.c +++ b/arch/mips/oprofile/common.c @@ -115,6 +115,7 @@ int __init oprofile_arch_init(struct oprofile_operations *ops) ops->start = op_mips_start; ops->stop = op_mips_stop; ops->cpu_type = lmodel->cpu_type; + ops->backtrace = op_mips_backtrace; printk(KERN_INFO "oprofile: using %s performance monitoring.\n", lmodel->cpu_type); diff --git a/arch/mips/oprofile/op_impl.h b/arch/mips/oprofile/op_impl.h index f04b54fb37d1..7c2da27ece04 100644 --- a/arch/mips/oprofile/op_impl.h +++ b/arch/mips/oprofile/op_impl.h @@ -36,4 +36,6 @@ struct op_mips_model { unsigned char num_counters; }; +void op_mips_backtrace(struct pt_regs * const regs, unsigned int depth); + #endif -- cgit v1.2.3 From 4f8b50bbbe63ae4ec6bea28a90a9a603c745ea71 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 27 Jun 2011 17:22:43 +0200 Subject: irq_work, ppc: Fix up arch hooks Commit e360adbe29 ("irq_work: Add generic hardirq context callbacks") fouled up the ppc bit, not properly naming the arch specific function that raises the 'self-IPI'. Cc: Huang Ying Cc: Benjamin Herrenschmidt Cc: Anton Blanchard Cc: Eric B Munson Cc: stable@kernel.org # 37+ Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-eg0aqien8p1aqvzu9dft6dtv@git.kernel.org Signed-off-by: Ingo Molnar --- arch/powerpc/kernel/time.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index f33acfd872ad..03b29a6759ab 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -544,7 +544,7 @@ DEFINE_PER_CPU(u8, irq_work_pending); #endif /* 32 vs 64 bit */ -void set_irq_work_pending(void) +void arch_irq_work_raise(void) { preempt_disable(); set_irq_work_pending_flag(); -- cgit v1.2.3 From 0f933625e7b6c3d91878ae95e341bf1984db7eaf Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 28 Jun 2011 12:15:51 +0200 Subject: irq_work, alpha: Fix up arch hooks Commit e360adbe29 ("irq_work: Add generic hardirq context callbacks") fouled up the Alpha bit, not properly naming the arch specific function that raises the 'self-IPI'. Signed-off-by: Peter Zijlstra Cc: Michael Cree Cc: stable@kernel.org # 37+ Link: http://lkml.kernel.org/n/tip-gukh0txmql2l4thgrekzzbfy@git.kernel.org Signed-off-by: Ingo Molnar --- arch/alpha/kernel/time.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/alpha/kernel/time.c b/arch/alpha/kernel/time.c index 818e74ed45dc..f20d1b5396b8 100644 --- a/arch/alpha/kernel/time.c +++ b/arch/alpha/kernel/time.c @@ -91,7 +91,7 @@ DEFINE_PER_CPU(u8, irq_work_pending); #define test_irq_work_pending() __get_cpu_var(irq_work_pending) #define clear_irq_work_pending() __get_cpu_var(irq_work_pending) = 0 -void set_irq_work_pending(void) +void arch_irq_work_raise(void) { set_irq_work_pending_flag(); } -- cgit v1.2.3 From 1880c4ae182afb5650c5678949ecfe7ff66a724e Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Thu, 23 Jun 2011 16:49:18 +0400 Subject: perf, x86: Add hw_watchdog_set_attr() in a sake of nmi-watchdog on P4 Due to restriction and specifics of Netburst PMU we need a separated event for NMI watchdog. In particular every Netburst event consumes not just a counter and a config register, but also an additional ESCR register. Since ESCR registers are grouped upon counters (i.e. if ESCR is occupied for some event there is no room for another event to enter until its released) we need to pick up the "least" used ESCR (or the most available one) for nmi-watchdog purposes -- so MSR_P4_CRU_ESCR2/3 was chosen. With this patch nmi-watchdog and perf top should be able to run simultaneously. Signed-off-by: Cyrill Gorcunov CC: Lin Ming CC: Arnaldo Carvalho de Melo CC: Frederic Weisbecker Tested-and-reviewed-by: Don Zickus Tested-and-reviewed-by: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20110623124918.GC13050@sun Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event.c | 7 +++++++ arch/x86/kernel/cpu/perf_event_p4.c | 26 ++++++++++++++++++++++++++ kernel/watchdog.c | 6 +++++- 3 files changed, 38 insertions(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 3a0338b4b179..8a57f9aa8e36 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -233,6 +233,7 @@ struct x86_pmu { void (*enable_all)(int added); void (*enable)(struct perf_event *); void (*disable)(struct perf_event *); + void (*hw_watchdog_set_attr)(struct perf_event_attr *attr); int (*hw_config)(struct perf_event *event); int (*schedule_events)(struct cpu_hw_events *cpuc, int n, int *assign); unsigned eventsel; @@ -315,6 +316,12 @@ static u64 __read_mostly hw_cache_extra_regs [PERF_COUNT_HW_CACHE_OP_MAX] [PERF_COUNT_HW_CACHE_RESULT_MAX]; +void hw_nmi_watchdog_set_attr(struct perf_event_attr *wd_attr) +{ + if (x86_pmu.hw_watchdog_set_attr) + x86_pmu.hw_watchdog_set_attr(wd_attr); +} + /* * Propagate event elapsed time into the generic event. * Can only be executed on the CPU where the event is active. diff --git a/arch/x86/kernel/cpu/perf_event_p4.c b/arch/x86/kernel/cpu/perf_event_p4.c index ead584fb6a7d..f76fddf63381 100644 --- a/arch/x86/kernel/cpu/perf_event_p4.c +++ b/arch/x86/kernel/cpu/perf_event_p4.c @@ -705,6 +705,31 @@ static int p4_validate_raw_event(struct perf_event *event) return 0; } +static void p4_hw_watchdog_set_attr(struct perf_event_attr *wd_attr) +{ + /* + * Watchdog ticks are special on Netburst, we use + * that named "non-sleeping" ticks as recommended + * by Intel SDM Vol3b. + */ + WARN_ON_ONCE(wd_attr->type != PERF_TYPE_HARDWARE || + wd_attr->config != PERF_COUNT_HW_CPU_CYCLES); + + wd_attr->type = PERF_TYPE_RAW; + wd_attr->config = + p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_EXECUTION_EVENT) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS0) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS1) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS2) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS3) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS0) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS1) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS2) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS3)) | + p4_config_pack_cccr(P4_CCCR_THRESHOLD(15) | P4_CCCR_COMPLEMENT | + P4_CCCR_COMPARE); +} + static int p4_hw_config(struct perf_event *event) { int cpu = get_cpu(); @@ -1179,6 +1204,7 @@ static __initconst const struct x86_pmu p4_pmu = { .cntval_bits = ARCH_P4_CNTRVAL_BITS, .cntval_mask = ARCH_P4_CNTRVAL_MASK, .max_period = (1ULL << (ARCH_P4_CNTRVAL_BITS - 1)) - 1, + .hw_watchdog_set_attr = p4_hw_watchdog_set_attr, .hw_config = p4_hw_config, .schedule_events = p4_pmu_schedule_events, /* diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 3d0c56ad4792..752b75ba662b 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -200,6 +200,8 @@ static int is_softlockup(unsigned long touch_ts) } #ifdef CONFIG_HARDLOCKUP_DETECTOR +void __weak hw_nmi_watchdog_set_attr(struct perf_event_attr *wd_attr) { } + static struct perf_event_attr wd_hw_attr = { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_CPU_CYCLES, @@ -368,9 +370,11 @@ static int watchdog_nmi_enable(int cpu) if (event != NULL) goto out_enable; - /* Try to register using hardware perf events */ wd_attr = &wd_hw_attr; wd_attr->sample_period = hw_nmi_get_sample_period(watchdog_thresh); + hw_nmi_watchdog_set_attr(wd_attr); + + /* Try to register using hardware perf events */ event = perf_event_create_kernel_counter(wd_attr, cpu, NULL, watchdog_overflow_callback); if (!IS_ERR(event)) { printk(KERN_INFO "NMI watchdog enabled, takes one hw-pmu counter.\n"); -- cgit v1.2.3 From a8b0ca17b80e92faab46ee7179ba9e99ccb61233 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 27 Jun 2011 14:41:57 +0200 Subject: perf: Remove the nmi parameter from the swevent and overflow interface The nmi parameter indicated if we could do wakeups from the current context, if not, we would set some state and self-IPI and let the resulting interrupt do the wakeup. For the various event classes: - hardware: nmi=0; PMI is in fact an NMI or we run irq_work_run from the PMI-tail (ARM etc.) - tracepoint: nmi=0; since tracepoint could be from NMI context. - software: nmi=[0,1]; some, like the schedule thing cannot perform wakeups, and hence need 0. As one can see, there is very little nmi=1 usage, and the down-side of not using it is that on some platforms some software events can have a jiffy delay in wakeup (when arch_irq_work_raise isn't implemented). The up-side however is that we can remove the nmi parameter and save a bunch of conditionals in fast paths. Signed-off-by: Peter Zijlstra Cc: Michael Cree Cc: Will Deacon Cc: Deng-Cheng Zhu Cc: Anton Blanchard Cc: Eric B Munson Cc: Heiko Carstens Cc: Paul Mundt Cc: David S. Miller Cc: Frederic Weisbecker Cc: Jason Wessel Cc: Don Zickus Link: http://lkml.kernel.org/n/tip-agjev8eu666tvknpb3iaj0fg@git.kernel.org Signed-off-by: Ingo Molnar --- arch/alpha/kernel/perf_event.c | 2 +- arch/arm/kernel/perf_event_v6.c | 2 +- arch/arm/kernel/perf_event_v7.c | 2 +- arch/arm/kernel/perf_event_xscale.c | 4 +- arch/arm/kernel/ptrace.c | 2 +- arch/arm/kernel/swp_emulate.c | 2 +- arch/arm/mm/fault.c | 6 +-- arch/mips/kernel/perf_event.c | 2 +- arch/mips/kernel/traps.c | 8 ++-- arch/mips/kernel/unaligned.c | 5 +-- arch/mips/math-emu/cp1emu.c | 3 +- arch/mips/mm/fault.c | 8 ++-- arch/powerpc/include/asm/emulated_ops.h | 4 +- arch/powerpc/kernel/perf_event.c | 6 +-- arch/powerpc/kernel/perf_event_fsl_emb.c | 6 +-- arch/powerpc/kernel/ptrace.c | 2 +- arch/powerpc/mm/fault.c | 6 +-- arch/s390/mm/fault.c | 6 +-- arch/sh/kernel/ptrace_32.c | 2 +- arch/sh/kernel/traps_32.c | 2 +- arch/sh/kernel/traps_64.c | 8 ++-- arch/sh/math-emu/math.c | 2 +- arch/sh/mm/fault_32.c | 6 +-- arch/sh/mm/tlbflush_64.c | 6 +-- arch/sparc/kernel/perf_event.c | 2 +- arch/sparc/kernel/unaligned_32.c | 4 +- arch/sparc/kernel/unaligned_64.c | 12 +++--- arch/sparc/kernel/visemul.c | 2 +- arch/sparc/math-emu/math_32.c | 2 +- arch/sparc/math-emu/math_64.c | 2 +- arch/sparc/mm/fault_32.c | 8 ++-- arch/sparc/mm/fault_64.c | 8 ++-- arch/x86/kernel/cpu/perf_event.c | 2 +- arch/x86/kernel/cpu/perf_event_intel.c | 2 +- arch/x86/kernel/cpu/perf_event_intel_ds.c | 4 +- arch/x86/kernel/cpu/perf_event_p4.c | 2 +- arch/x86/kernel/kgdb.c | 2 +- arch/x86/kernel/ptrace.c | 2 +- arch/x86/mm/fault.c | 6 +-- include/linux/perf_event.h | 18 ++++----- kernel/events/core.c | 63 ++++++++++++++----------------- kernel/events/internal.h | 1 - kernel/events/ring_buffer.c | 10 ++--- kernel/sched.c | 2 +- kernel/watchdog.c | 2 +- samples/hw_breakpoint/data_breakpoint.c | 2 +- 46 files changed, 119 insertions(+), 141 deletions(-) (limited to 'arch') diff --git a/arch/alpha/kernel/perf_event.c b/arch/alpha/kernel/perf_event.c index 90561c45e7d8..8e47709160f8 100644 --- a/arch/alpha/kernel/perf_event.c +++ b/arch/alpha/kernel/perf_event.c @@ -847,7 +847,7 @@ static void alpha_perf_event_irq_handler(unsigned long la_ptr, data.period = event->hw.last_period; if (alpha_perf_event_set_period(event, hwc, idx)) { - if (perf_event_overflow(event, 1, &data, regs)) { + if (perf_event_overflow(event, &data, regs)) { /* Interrupts coming too quickly; "throttle" the * counter, i.e., disable it for a little while. */ diff --git a/arch/arm/kernel/perf_event_v6.c b/arch/arm/kernel/perf_event_v6.c index f1e8dd94afe8..38dc4da1d612 100644 --- a/arch/arm/kernel/perf_event_v6.c +++ b/arch/arm/kernel/perf_event_v6.c @@ -479,7 +479,7 @@ armv6pmu_handle_irq(int irq_num, if (!armpmu_event_set_period(event, hwc, idx)) continue; - if (perf_event_overflow(event, 0, &data, regs)) + if (perf_event_overflow(event, &data, regs)) armpmu->disable(hwc, idx); } diff --git a/arch/arm/kernel/perf_event_v7.c b/arch/arm/kernel/perf_event_v7.c index 4960686afb58..6e5f8752303b 100644 --- a/arch/arm/kernel/perf_event_v7.c +++ b/arch/arm/kernel/perf_event_v7.c @@ -787,7 +787,7 @@ static irqreturn_t armv7pmu_handle_irq(int irq_num, void *dev) if (!armpmu_event_set_period(event, hwc, idx)) continue; - if (perf_event_overflow(event, 0, &data, regs)) + if (perf_event_overflow(event, &data, regs)) armpmu->disable(hwc, idx); } diff --git a/arch/arm/kernel/perf_event_xscale.c b/arch/arm/kernel/perf_event_xscale.c index 39affbe4fdb2..99b6b85c7e49 100644 --- a/arch/arm/kernel/perf_event_xscale.c +++ b/arch/arm/kernel/perf_event_xscale.c @@ -251,7 +251,7 @@ xscale1pmu_handle_irq(int irq_num, void *dev) if (!armpmu_event_set_period(event, hwc, idx)) continue; - if (perf_event_overflow(event, 0, &data, regs)) + if (perf_event_overflow(event, &data, regs)) armpmu->disable(hwc, idx); } @@ -583,7 +583,7 @@ xscale2pmu_handle_irq(int irq_num, void *dev) if (!armpmu_event_set_period(event, hwc, idx)) continue; - if (perf_event_overflow(event, 0, &data, regs)) + if (perf_event_overflow(event, &data, regs)) armpmu->disable(hwc, idx); } diff --git a/arch/arm/kernel/ptrace.c b/arch/arm/kernel/ptrace.c index 97260060bf26..0c9b1054f790 100644 --- a/arch/arm/kernel/ptrace.c +++ b/arch/arm/kernel/ptrace.c @@ -396,7 +396,7 @@ static long ptrace_hbp_idx_to_num(int idx) /* * Handle hitting a HW-breakpoint. */ -static void ptrace_hbptriggered(struct perf_event *bp, int unused, +static void ptrace_hbptriggered(struct perf_event *bp, struct perf_sample_data *data, struct pt_regs *regs) { diff --git a/arch/arm/kernel/swp_emulate.c b/arch/arm/kernel/swp_emulate.c index 40ee7e5045e4..5f452f8fde05 100644 --- a/arch/arm/kernel/swp_emulate.c +++ b/arch/arm/kernel/swp_emulate.c @@ -183,7 +183,7 @@ static int swp_handler(struct pt_regs *regs, unsigned int instr) unsigned int address, destreg, data, type; unsigned int res = 0; - perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, regs->ARM_pc); + perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, regs->ARM_pc); if (current->pid != previous_pid) { pr_debug("\"%s\" (%ld) uses deprecated SWP{B} instruction\n", diff --git a/arch/arm/mm/fault.c b/arch/arm/mm/fault.c index bc0e1d88fd3b..9ea4f7ddd665 100644 --- a/arch/arm/mm/fault.c +++ b/arch/arm/mm/fault.c @@ -318,11 +318,11 @@ do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs) fault = __do_page_fault(mm, addr, fsr, tsk); up_read(&mm->mmap_sem); - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, addr); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, addr); if (fault & VM_FAULT_MAJOR) - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0, regs, addr); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, addr); else if (fault & VM_FAULT_MINOR) - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0, regs, addr); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, addr); /* * Handle the "normal" case first - VM_FAULT_MAJOR / VM_FAULT_MINOR diff --git a/arch/mips/kernel/perf_event.c b/arch/mips/kernel/perf_event.c index a8244854d3dc..d0deaab9ace2 100644 --- a/arch/mips/kernel/perf_event.c +++ b/arch/mips/kernel/perf_event.c @@ -527,7 +527,7 @@ handle_associated_event(struct cpu_hw_events *cpuc, if (!mipspmu_event_set_period(event, hwc, idx)) return; - if (perf_event_overflow(event, 0, data, regs)) + if (perf_event_overflow(event, data, regs)) mipspmu->disable_event(idx); } diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c index e9b3af27d844..b7517e3abc85 100644 --- a/arch/mips/kernel/traps.c +++ b/arch/mips/kernel/traps.c @@ -578,12 +578,12 @@ static int simulate_llsc(struct pt_regs *regs, unsigned int opcode) { if ((opcode & OPCODE) == LL) { perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, - 1, 0, regs, 0); + 1, regs, 0); return simulate_ll(regs, opcode); } if ((opcode & OPCODE) == SC) { perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, - 1, 0, regs, 0); + 1, regs, 0); return simulate_sc(regs, opcode); } @@ -602,7 +602,7 @@ static int simulate_rdhwr(struct pt_regs *regs, unsigned int opcode) int rd = (opcode & RD) >> 11; int rt = (opcode & RT) >> 16; perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, - 1, 0, regs, 0); + 1, regs, 0); switch (rd) { case 0: /* CPU number */ regs->regs[rt] = smp_processor_id(); @@ -640,7 +640,7 @@ static int simulate_sync(struct pt_regs *regs, unsigned int opcode) { if ((opcode & OPCODE) == SPEC0 && (opcode & FUNC) == SYNC) { perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, - 1, 0, regs, 0); + 1, regs, 0); return 0; } diff --git a/arch/mips/kernel/unaligned.c b/arch/mips/kernel/unaligned.c index cfea1adfa153..eb319b580353 100644 --- a/arch/mips/kernel/unaligned.c +++ b/arch/mips/kernel/unaligned.c @@ -111,8 +111,7 @@ static void emulate_load_store_insn(struct pt_regs *regs, unsigned long value; unsigned int res; - perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, - 1, 0, regs, 0); + perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0); /* * This load never faults. @@ -517,7 +516,7 @@ asmlinkage void do_ade(struct pt_regs *regs) mm_segment_t seg; perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, - 1, 0, regs, regs->cp0_badvaddr); + 1, regs, regs->cp0_badvaddr); /* * Did we catch a fault trying to load an instruction? * Or are we running in MIPS16 mode? diff --git a/arch/mips/math-emu/cp1emu.c b/arch/mips/math-emu/cp1emu.c index d32cb0503110..dbf2f93a5091 100644 --- a/arch/mips/math-emu/cp1emu.c +++ b/arch/mips/math-emu/cp1emu.c @@ -272,8 +272,7 @@ static int cop1Emulate(struct pt_regs *xcp, struct mips_fpu_struct *ctx, } emul: - perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, - 1, 0, xcp, 0); + perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, xcp, 0); MIPS_FPU_EMU_INC_STATS(emulated); switch (MIPSInst_OPCODE(ir)) { case ldc1_op:{ diff --git a/arch/mips/mm/fault.c b/arch/mips/mm/fault.c index 137ee76a0045..937cf3368164 100644 --- a/arch/mips/mm/fault.c +++ b/arch/mips/mm/fault.c @@ -145,7 +145,7 @@ good_area: * the fault. */ fault = handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0); - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address); if (unlikely(fault & VM_FAULT_ERROR)) { if (fault & VM_FAULT_OOM) goto out_of_memory; @@ -154,12 +154,10 @@ good_area: BUG(); } if (fault & VM_FAULT_MAJOR) { - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, - 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, address); tsk->maj_flt++; } else { - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, - 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, address); tsk->min_flt++; } diff --git a/arch/powerpc/include/asm/emulated_ops.h b/arch/powerpc/include/asm/emulated_ops.h index 45921672b97a..2cc41c715d2b 100644 --- a/arch/powerpc/include/asm/emulated_ops.h +++ b/arch/powerpc/include/asm/emulated_ops.h @@ -78,14 +78,14 @@ extern void ppc_warn_emulated_print(const char *type); #define PPC_WARN_EMULATED(type, regs) \ do { \ perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, \ - 1, 0, regs, 0); \ + 1, regs, 0); \ __PPC_WARN_EMULATED(type); \ } while (0) #define PPC_WARN_ALIGNMENT(type, regs) \ do { \ perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, \ - 1, 0, regs, regs->dar); \ + 1, regs, regs->dar); \ __PPC_WARN_EMULATED(type); \ } while (0) diff --git a/arch/powerpc/kernel/perf_event.c b/arch/powerpc/kernel/perf_event.c index 822f63008ae1..14967de98876 100644 --- a/arch/powerpc/kernel/perf_event.c +++ b/arch/powerpc/kernel/perf_event.c @@ -1207,7 +1207,7 @@ struct pmu power_pmu = { * here so there is no possibility of being interrupted. */ static void record_and_restart(struct perf_event *event, unsigned long val, - struct pt_regs *regs, int nmi) + struct pt_regs *regs) { u64 period = event->hw.sample_period; s64 prev, delta, left; @@ -1258,7 +1258,7 @@ static void record_and_restart(struct perf_event *event, unsigned long val, if (event->attr.sample_type & PERF_SAMPLE_ADDR) perf_get_data_addr(regs, &data.addr); - if (perf_event_overflow(event, nmi, &data, regs)) + if (perf_event_overflow(event, &data, regs)) power_pmu_stop(event, 0); } } @@ -1346,7 +1346,7 @@ static void perf_event_interrupt(struct pt_regs *regs) if ((int)val < 0) { /* event has overflowed */ found = 1; - record_and_restart(event, val, regs, nmi); + record_and_restart(event, val, regs); } } diff --git a/arch/powerpc/kernel/perf_event_fsl_emb.c b/arch/powerpc/kernel/perf_event_fsl_emb.c index b0dc8f7069cd..0a6d2a9d569c 100644 --- a/arch/powerpc/kernel/perf_event_fsl_emb.c +++ b/arch/powerpc/kernel/perf_event_fsl_emb.c @@ -568,7 +568,7 @@ static struct pmu fsl_emb_pmu = { * here so there is no possibility of being interrupted. */ static void record_and_restart(struct perf_event *event, unsigned long val, - struct pt_regs *regs, int nmi) + struct pt_regs *regs) { u64 period = event->hw.sample_period; s64 prev, delta, left; @@ -616,7 +616,7 @@ static void record_and_restart(struct perf_event *event, unsigned long val, perf_sample_data_init(&data, 0); data.period = event->hw.last_period; - if (perf_event_overflow(event, nmi, &data, regs)) + if (perf_event_overflow(event, &data, regs)) fsl_emb_pmu_stop(event, 0); } } @@ -644,7 +644,7 @@ static void perf_event_interrupt(struct pt_regs *regs) if (event) { /* event has overflowed */ found = 1; - record_and_restart(event, val, regs, nmi); + record_and_restart(event, val, regs); } else { /* * Disabled counter is negative, diff --git a/arch/powerpc/kernel/ptrace.c b/arch/powerpc/kernel/ptrace.c index cb22024f2b42..3177617af2ef 100644 --- a/arch/powerpc/kernel/ptrace.c +++ b/arch/powerpc/kernel/ptrace.c @@ -882,7 +882,7 @@ void user_disable_single_step(struct task_struct *task) } #ifdef CONFIG_HAVE_HW_BREAKPOINT -void ptrace_triggered(struct perf_event *bp, int nmi, +void ptrace_triggered(struct perf_event *bp, struct perf_sample_data *data, struct pt_regs *regs) { struct perf_event_attr attr; diff --git a/arch/powerpc/mm/fault.c b/arch/powerpc/mm/fault.c index 54f4fb994e99..dbc48254c6cc 100644 --- a/arch/powerpc/mm/fault.c +++ b/arch/powerpc/mm/fault.c @@ -173,7 +173,7 @@ int __kprobes do_page_fault(struct pt_regs *regs, unsigned long address, die("Weird page fault", regs, SIGSEGV); } - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address); /* When running in the kernel we expect faults to occur only to * addresses in user space. All other faults represent errors in the @@ -319,7 +319,7 @@ good_area: } if (ret & VM_FAULT_MAJOR) { current->maj_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0, + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, address); #ifdef CONFIG_PPC_SMLPAR if (firmware_has_feature(FW_FEATURE_CMO)) { @@ -330,7 +330,7 @@ good_area: #endif } else { current->min_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0, + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, address); } up_read(&mm->mmap_sem); diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c index fe103e891e7a..095f782a5512 100644 --- a/arch/s390/mm/fault.c +++ b/arch/s390/mm/fault.c @@ -299,7 +299,7 @@ static inline int do_exception(struct pt_regs *regs, int access, goto out; address = trans_exc_code & __FAIL_ADDR_MASK; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address); flags = FAULT_FLAG_ALLOW_RETRY; if (access == VM_WRITE || (trans_exc_code & store_indication) == 0x400) flags |= FAULT_FLAG_WRITE; @@ -345,11 +345,11 @@ retry: if (flags & FAULT_FLAG_ALLOW_RETRY) { if (fault & VM_FAULT_MAJOR) { tsk->maj_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0, + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, address); } else { tsk->min_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0, + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, address); } if (fault & VM_FAULT_RETRY) { diff --git a/arch/sh/kernel/ptrace_32.c b/arch/sh/kernel/ptrace_32.c index 3d7b209b2178..8051976100a6 100644 --- a/arch/sh/kernel/ptrace_32.c +++ b/arch/sh/kernel/ptrace_32.c @@ -63,7 +63,7 @@ static inline int put_stack_long(struct task_struct *task, int offset, return 0; } -void ptrace_triggered(struct perf_event *bp, int nmi, +void ptrace_triggered(struct perf_event *bp, struct perf_sample_data *data, struct pt_regs *regs) { struct perf_event_attr attr; diff --git a/arch/sh/kernel/traps_32.c b/arch/sh/kernel/traps_32.c index b51a17104b5f..d9006f8ffc14 100644 --- a/arch/sh/kernel/traps_32.c +++ b/arch/sh/kernel/traps_32.c @@ -393,7 +393,7 @@ int handle_unaligned_access(insn_size_t instruction, struct pt_regs *regs, */ if (!expected) { unaligned_fixups_notify(current, instruction, regs); - perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, + perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, address); } diff --git a/arch/sh/kernel/traps_64.c b/arch/sh/kernel/traps_64.c index 6713ca97e553..67110be83fd7 100644 --- a/arch/sh/kernel/traps_64.c +++ b/arch/sh/kernel/traps_64.c @@ -434,7 +434,7 @@ static int misaligned_load(struct pt_regs *regs, return error; } - perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, address); destreg = (opcode >> 4) & 0x3f; if (user_mode(regs)) { @@ -512,7 +512,7 @@ static int misaligned_store(struct pt_regs *regs, return error; } - perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, address); srcreg = (opcode >> 4) & 0x3f; if (user_mode(regs)) { @@ -588,7 +588,7 @@ static int misaligned_fpu_load(struct pt_regs *regs, return error; } - perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, address); destreg = (opcode >> 4) & 0x3f; if (user_mode(regs)) { @@ -665,7 +665,7 @@ static int misaligned_fpu_store(struct pt_regs *regs, return error; } - perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, address); srcreg = (opcode >> 4) & 0x3f; if (user_mode(regs)) { diff --git a/arch/sh/math-emu/math.c b/arch/sh/math-emu/math.c index f76a5090d5d1..977195210653 100644 --- a/arch/sh/math-emu/math.c +++ b/arch/sh/math-emu/math.c @@ -620,7 +620,7 @@ int do_fpu_inst(unsigned short inst, struct pt_regs *regs) struct task_struct *tsk = current; struct sh_fpu_soft_struct *fpu = &(tsk->thread.xstate->softfpu); - perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0); + perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0); if (!(task_thread_info(tsk)->status & TS_USEDFPU)) { /* initialize once. */ diff --git a/arch/sh/mm/fault_32.c b/arch/sh/mm/fault_32.c index d4c34d757f0d..7bebd044f2a1 100644 --- a/arch/sh/mm/fault_32.c +++ b/arch/sh/mm/fault_32.c @@ -160,7 +160,7 @@ asmlinkage void __kprobes do_page_fault(struct pt_regs *regs, if ((regs->sr & SR_IMASK) != SR_IMASK) local_irq_enable(); - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address); /* * If we're in an interrupt, have no user context or are running @@ -210,11 +210,11 @@ good_area: } if (fault & VM_FAULT_MAJOR) { tsk->maj_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0, + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, address); } else { tsk->min_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0, + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, address); } diff --git a/arch/sh/mm/tlbflush_64.c b/arch/sh/mm/tlbflush_64.c index 7f5810f5dfdc..e3430e093d43 100644 --- a/arch/sh/mm/tlbflush_64.c +++ b/arch/sh/mm/tlbflush_64.c @@ -116,7 +116,7 @@ asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long writeaccess, /* Not an IO address, so reenable interrupts */ local_irq_enable(); - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address); /* * If we're in an interrupt or have no user @@ -200,11 +200,11 @@ good_area: if (fault & VM_FAULT_MAJOR) { tsk->maj_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0, + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, address); } else { tsk->min_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0, + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, address); } diff --git a/arch/sparc/kernel/perf_event.c b/arch/sparc/kernel/perf_event.c index 2cb0e1c001e2..0b32f2e9e08d 100644 --- a/arch/sparc/kernel/perf_event.c +++ b/arch/sparc/kernel/perf_event.c @@ -1277,7 +1277,7 @@ static int __kprobes perf_event_nmi_handler(struct notifier_block *self, if (!sparc_perf_event_set_period(event, hwc, idx)) continue; - if (perf_event_overflow(event, 1, &data, regs)) + if (perf_event_overflow(event, &data, regs)) sparc_pmu_stop(event, 0); } diff --git a/arch/sparc/kernel/unaligned_32.c b/arch/sparc/kernel/unaligned_32.c index 4491f4cb2695..7efbb2f9e77f 100644 --- a/arch/sparc/kernel/unaligned_32.c +++ b/arch/sparc/kernel/unaligned_32.c @@ -247,7 +247,7 @@ asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn) unsigned long addr = compute_effective_address(regs, insn); int err; - perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, addr); + perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, addr); switch (dir) { case load: err = do_int_load(fetch_reg_addr(((insn>>25)&0x1f), @@ -338,7 +338,7 @@ asmlinkage void user_unaligned_trap(struct pt_regs *regs, unsigned int insn) } addr = compute_effective_address(regs, insn); - perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, addr); + perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, addr); switch(dir) { case load: err = do_int_load(fetch_reg_addr(((insn>>25)&0x1f), diff --git a/arch/sparc/kernel/unaligned_64.c b/arch/sparc/kernel/unaligned_64.c index b2b019ea8caa..35cff1673aa4 100644 --- a/arch/sparc/kernel/unaligned_64.c +++ b/arch/sparc/kernel/unaligned_64.c @@ -317,7 +317,7 @@ asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn) addr = compute_effective_address(regs, insn, ((insn >> 25) & 0x1f)); - perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, addr); + perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, addr); switch (asi) { case ASI_NL: case ASI_AIUPL: @@ -384,7 +384,7 @@ int handle_popc(u32 insn, struct pt_regs *regs) int ret, i, rd = ((insn >> 25) & 0x1f); int from_kernel = (regs->tstate & TSTATE_PRIV) != 0; - perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0); + perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0); if (insn & 0x2000) { maybe_flush_windows(0, 0, rd, from_kernel); value = sign_extend_imm13(insn); @@ -431,7 +431,7 @@ int handle_ldf_stq(u32 insn, struct pt_regs *regs) int asi = decode_asi(insn, regs); int flag = (freg < 32) ? FPRS_DL : FPRS_DU; - perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0); + perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0); save_and_clear_fpu(); current_thread_info()->xfsr[0] &= ~0x1c000; @@ -554,7 +554,7 @@ void handle_ld_nf(u32 insn, struct pt_regs *regs) int from_kernel = (regs->tstate & TSTATE_PRIV) != 0; unsigned long *reg; - perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0); + perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0); maybe_flush_windows(0, 0, rd, from_kernel); reg = fetch_reg_addr(rd, regs); @@ -586,7 +586,7 @@ void handle_lddfmna(struct pt_regs *regs, unsigned long sfar, unsigned long sfsr if (tstate & TSTATE_PRIV) die_if_kernel("lddfmna from kernel", regs); - perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, sfar); + perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, sfar); if (test_thread_flag(TIF_32BIT)) pc = (u32)pc; if (get_user(insn, (u32 __user *) pc) != -EFAULT) { @@ -647,7 +647,7 @@ void handle_stdfmna(struct pt_regs *regs, unsigned long sfar, unsigned long sfsr if (tstate & TSTATE_PRIV) die_if_kernel("stdfmna from kernel", regs); - perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, sfar); + perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, sfar); if (test_thread_flag(TIF_32BIT)) pc = (u32)pc; if (get_user(insn, (u32 __user *) pc) != -EFAULT) { diff --git a/arch/sparc/kernel/visemul.c b/arch/sparc/kernel/visemul.c index 36357717d691..32b626c9d815 100644 --- a/arch/sparc/kernel/visemul.c +++ b/arch/sparc/kernel/visemul.c @@ -802,7 +802,7 @@ int vis_emul(struct pt_regs *regs, unsigned int insn) BUG_ON(regs->tstate & TSTATE_PRIV); - perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0); + perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0); if (test_thread_flag(TIF_32BIT)) pc = (u32)pc; diff --git a/arch/sparc/math-emu/math_32.c b/arch/sparc/math-emu/math_32.c index a3fccde894ec..aa4d55b0bdf0 100644 --- a/arch/sparc/math-emu/math_32.c +++ b/arch/sparc/math-emu/math_32.c @@ -164,7 +164,7 @@ int do_mathemu(struct pt_regs *regs, struct task_struct *fpt) int retcode = 0; /* assume all succeed */ unsigned long insn; - perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0); + perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0); #ifdef DEBUG_MATHEMU printk("In do_mathemu()... pc is %08lx\n", regs->pc); diff --git a/arch/sparc/math-emu/math_64.c b/arch/sparc/math-emu/math_64.c index 56d2c44747b8..e575bd2fe381 100644 --- a/arch/sparc/math-emu/math_64.c +++ b/arch/sparc/math-emu/math_64.c @@ -184,7 +184,7 @@ int do_mathemu(struct pt_regs *regs, struct fpustate *f) if (tstate & TSTATE_PRIV) die_if_kernel("unfinished/unimplemented FPop from kernel", regs); - perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0); + perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0); if (test_thread_flag(TIF_32BIT)) pc = (u32)pc; if (get_user(insn, (u32 __user *) pc) != -EFAULT) { diff --git a/arch/sparc/mm/fault_32.c b/arch/sparc/mm/fault_32.c index 7543ddbdadb2..aa1c1b1ce5cc 100644 --- a/arch/sparc/mm/fault_32.c +++ b/arch/sparc/mm/fault_32.c @@ -251,7 +251,7 @@ asmlinkage void do_sparc_fault(struct pt_regs *regs, int text_fault, int write, if (in_atomic() || !mm) goto no_context; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address); down_read(&mm->mmap_sem); @@ -301,12 +301,10 @@ good_area: } if (fault & VM_FAULT_MAJOR) { current->maj_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0, - regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, address); } else { current->min_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0, - regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, address); } up_read(&mm->mmap_sem); return; diff --git a/arch/sparc/mm/fault_64.c b/arch/sparc/mm/fault_64.c index f92ce56a8b22..504c0622f729 100644 --- a/arch/sparc/mm/fault_64.c +++ b/arch/sparc/mm/fault_64.c @@ -325,7 +325,7 @@ asmlinkage void __kprobes do_sparc64_fault(struct pt_regs *regs) if (in_atomic() || !mm) goto intr_or_no_mm; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address); if (!down_read_trylock(&mm->mmap_sem)) { if ((regs->tstate & TSTATE_PRIV) && @@ -433,12 +433,10 @@ good_area: } if (fault & VM_FAULT_MAJOR) { current->maj_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0, - regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, address); } else { current->min_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0, - regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, address); } up_read(&mm->mmap_sem); diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 8a57f9aa8e36..5b86ec51534c 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1339,7 +1339,7 @@ static int x86_pmu_handle_irq(struct pt_regs *regs) if (!x86_perf_event_set_period(event)) continue; - if (perf_event_overflow(event, 1, &data, regs)) + if (perf_event_overflow(event, &data, regs)) x86_pmu_stop(event, 0); } diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index 41178c826c48..d38b0020f775 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -1003,7 +1003,7 @@ again: data.period = event->hw.last_period; - if (perf_event_overflow(event, 1, &data, regs)) + if (perf_event_overflow(event, &data, regs)) x86_pmu_stop(event, 0); } diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c index bab491b8ee25..0941f93f2940 100644 --- a/arch/x86/kernel/cpu/perf_event_intel_ds.c +++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c @@ -340,7 +340,7 @@ static int intel_pmu_drain_bts_buffer(void) */ perf_prepare_sample(&header, &data, event, ®s); - if (perf_output_begin(&handle, event, header.size * (top - at), 1, 1)) + if (perf_output_begin(&handle, event, header.size * (top - at), 1)) return 1; for (; at < top; at++) { @@ -616,7 +616,7 @@ static void __intel_pmu_pebs_event(struct perf_event *event, else regs.flags &= ~PERF_EFLAGS_EXACT; - if (perf_event_overflow(event, 1, &data, ®s)) + if (perf_event_overflow(event, &data, ®s)) x86_pmu_stop(event, 0); } diff --git a/arch/x86/kernel/cpu/perf_event_p4.c b/arch/x86/kernel/cpu/perf_event_p4.c index f76fddf63381..d6e6a67b9608 100644 --- a/arch/x86/kernel/cpu/perf_event_p4.c +++ b/arch/x86/kernel/cpu/perf_event_p4.c @@ -970,7 +970,7 @@ static int p4_pmu_handle_irq(struct pt_regs *regs) if (!x86_perf_event_set_period(event)) continue; - if (perf_event_overflow(event, 1, &data, regs)) + if (perf_event_overflow(event, &data, regs)) x86_pmu_stop(event, 0); } diff --git a/arch/x86/kernel/kgdb.c b/arch/x86/kernel/kgdb.c index 5f9ecff328b5..98da6a7b5e82 100644 --- a/arch/x86/kernel/kgdb.c +++ b/arch/x86/kernel/kgdb.c @@ -608,7 +608,7 @@ int kgdb_arch_init(void) return register_die_notifier(&kgdb_notifier); } -static void kgdb_hw_overflow_handler(struct perf_event *event, int nmi, +static void kgdb_hw_overflow_handler(struct perf_event *event, struct perf_sample_data *data, struct pt_regs *regs) { struct task_struct *tsk = current; diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index 807c2a2b80f1..11db2e9b860a 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -528,7 +528,7 @@ static int genregs_set(struct task_struct *target, return ret; } -static void ptrace_triggered(struct perf_event *bp, int nmi, +static void ptrace_triggered(struct perf_event *bp, struct perf_sample_data *data, struct pt_regs *regs) { diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index 2dbf6bf4c7e5..4d09df054e39 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -1059,7 +1059,7 @@ do_page_fault(struct pt_regs *regs, unsigned long error_code) if (unlikely(error_code & PF_RSVD)) pgtable_bad(regs, error_code, address); - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address); + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address); /* * If we're in an interrupt, have no user context or are running @@ -1161,11 +1161,11 @@ good_area: if (flags & FAULT_FLAG_ALLOW_RETRY) { if (fault & VM_FAULT_MAJOR) { tsk->maj_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0, + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, address); } else { tsk->min_flt++; - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0, + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, address); } if (fault & VM_FAULT_RETRY) { diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 2f7b5d42ab41..0946a8bc098d 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -682,7 +682,7 @@ enum perf_event_active_state { struct file; struct perf_sample_data; -typedef void (*perf_overflow_handler_t)(struct perf_event *, int, +typedef void (*perf_overflow_handler_t)(struct perf_event *, struct perf_sample_data *, struct pt_regs *regs); @@ -925,7 +925,6 @@ struct perf_output_handle { unsigned long size; void *addr; int page; - int nmi; int sample; }; @@ -993,7 +992,7 @@ extern void perf_prepare_sample(struct perf_event_header *header, struct perf_event *event, struct pt_regs *regs); -extern int perf_event_overflow(struct perf_event *event, int nmi, +extern int perf_event_overflow(struct perf_event *event, struct perf_sample_data *data, struct pt_regs *regs); @@ -1012,7 +1011,7 @@ static inline int is_software_event(struct perf_event *event) extern struct jump_label_key perf_swevent_enabled[PERF_COUNT_SW_MAX]; -extern void __perf_sw_event(u32, u64, int, struct pt_regs *, u64); +extern void __perf_sw_event(u32, u64, struct pt_regs *, u64); #ifndef perf_arch_fetch_caller_regs static inline void perf_arch_fetch_caller_regs(struct pt_regs *regs, unsigned long ip) { } @@ -1034,7 +1033,7 @@ static inline void perf_fetch_caller_regs(struct pt_regs *regs) } static __always_inline void -perf_sw_event(u32 event_id, u64 nr, int nmi, struct pt_regs *regs, u64 addr) +perf_sw_event(u32 event_id, u64 nr, struct pt_regs *regs, u64 addr) { struct pt_regs hot_regs; @@ -1043,7 +1042,7 @@ perf_sw_event(u32 event_id, u64 nr, int nmi, struct pt_regs *regs, u64 addr) perf_fetch_caller_regs(&hot_regs); regs = &hot_regs; } - __perf_sw_event(event_id, nr, nmi, regs, addr); + __perf_sw_event(event_id, nr, regs, addr); } } @@ -1057,7 +1056,7 @@ static inline void perf_event_task_sched_in(struct task_struct *task) static inline void perf_event_task_sched_out(struct task_struct *task, struct task_struct *next) { - perf_sw_event(PERF_COUNT_SW_CONTEXT_SWITCHES, 1, 1, NULL, 0); + perf_sw_event(PERF_COUNT_SW_CONTEXT_SWITCHES, 1, NULL, 0); __perf_event_task_sched_out(task, next); } @@ -1119,7 +1118,7 @@ extern void perf_bp_event(struct perf_event *event, void *data); extern int perf_output_begin(struct perf_output_handle *handle, struct perf_event *event, unsigned int size, - int nmi, int sample); + int sample); extern void perf_output_end(struct perf_output_handle *handle); extern void perf_output_copy(struct perf_output_handle *handle, const void *buf, unsigned int len); @@ -1143,8 +1142,7 @@ static inline int perf_event_task_disable(void) { return -EINVAL; } static inline int perf_event_task_enable(void) { return -EINVAL; } static inline void -perf_sw_event(u32 event_id, u64 nr, int nmi, - struct pt_regs *regs, u64 addr) { } +perf_sw_event(u32 event_id, u64 nr, struct pt_regs *regs, u64 addr) { } static inline void perf_bp_event(struct perf_event *event, void *data) { } diff --git a/kernel/events/core.c b/kernel/events/core.c index 270e32f9fc06..dbd1ca75bd3c 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -3972,7 +3972,7 @@ void perf_prepare_sample(struct perf_event_header *header, } } -static void perf_event_output(struct perf_event *event, int nmi, +static void perf_event_output(struct perf_event *event, struct perf_sample_data *data, struct pt_regs *regs) { @@ -3984,7 +3984,7 @@ static void perf_event_output(struct perf_event *event, int nmi, perf_prepare_sample(&header, data, event, regs); - if (perf_output_begin(&handle, event, header.size, nmi, 1)) + if (perf_output_begin(&handle, event, header.size, 1)) goto exit; perf_output_sample(&handle, &header, data, event); @@ -4024,7 +4024,7 @@ perf_event_read_event(struct perf_event *event, int ret; perf_event_header__init_id(&read_event.header, &sample, event); - ret = perf_output_begin(&handle, event, read_event.header.size, 0, 0); + ret = perf_output_begin(&handle, event, read_event.header.size, 0); if (ret) return; @@ -4067,7 +4067,7 @@ static void perf_event_task_output(struct perf_event *event, perf_event_header__init_id(&task_event->event_id.header, &sample, event); ret = perf_output_begin(&handle, event, - task_event->event_id.header.size, 0, 0); + task_event->event_id.header.size, 0); if (ret) goto out; @@ -4204,7 +4204,7 @@ static void perf_event_comm_output(struct perf_event *event, perf_event_header__init_id(&comm_event->event_id.header, &sample, event); ret = perf_output_begin(&handle, event, - comm_event->event_id.header.size, 0, 0); + comm_event->event_id.header.size, 0); if (ret) goto out; @@ -4351,7 +4351,7 @@ static void perf_event_mmap_output(struct perf_event *event, perf_event_header__init_id(&mmap_event->event_id.header, &sample, event); ret = perf_output_begin(&handle, event, - mmap_event->event_id.header.size, 0, 0); + mmap_event->event_id.header.size, 0); if (ret) goto out; @@ -4546,7 +4546,7 @@ static void perf_log_throttle(struct perf_event *event, int enable) perf_event_header__init_id(&throttle_event.header, &sample, event); ret = perf_output_begin(&handle, event, - throttle_event.header.size, 1, 0); + throttle_event.header.size, 0); if (ret) return; @@ -4559,7 +4559,7 @@ static void perf_log_throttle(struct perf_event *event, int enable) * Generic event overflow handling, sampling. */ -static int __perf_event_overflow(struct perf_event *event, int nmi, +static int __perf_event_overflow(struct perf_event *event, int throttle, struct perf_sample_data *data, struct pt_regs *regs) { @@ -4602,34 +4602,28 @@ static int __perf_event_overflow(struct perf_event *event, int nmi, if (events && atomic_dec_and_test(&event->event_limit)) { ret = 1; event->pending_kill = POLL_HUP; - if (nmi) { - event->pending_disable = 1; - irq_work_queue(&event->pending); - } else - perf_event_disable(event); + event->pending_disable = 1; + irq_work_queue(&event->pending); } if (event->overflow_handler) - event->overflow_handler(event, nmi, data, regs); + event->overflow_handler(event, data, regs); else - perf_event_output(event, nmi, data, regs); + perf_event_output(event, data, regs); if (event->fasync && event->pending_kill) { - if (nmi) { - event->pending_wakeup = 1; - irq_work_queue(&event->pending); - } else - perf_event_wakeup(event); + event->pending_wakeup = 1; + irq_work_queue(&event->pending); } return ret; } -int perf_event_overflow(struct perf_event *event, int nmi, +int perf_event_overflow(struct perf_event *event, struct perf_sample_data *data, struct pt_regs *regs) { - return __perf_event_overflow(event, nmi, 1, data, regs); + return __perf_event_overflow(event, 1, data, regs); } /* @@ -4678,7 +4672,7 @@ again: } static void perf_swevent_overflow(struct perf_event *event, u64 overflow, - int nmi, struct perf_sample_data *data, + struct perf_sample_data *data, struct pt_regs *regs) { struct hw_perf_event *hwc = &event->hw; @@ -4692,7 +4686,7 @@ static void perf_swevent_overflow(struct perf_event *event, u64 overflow, return; for (; overflow; overflow--) { - if (__perf_event_overflow(event, nmi, throttle, + if (__perf_event_overflow(event, throttle, data, regs)) { /* * We inhibit the overflow from happening when @@ -4705,7 +4699,7 @@ static void perf_swevent_overflow(struct perf_event *event, u64 overflow, } static void perf_swevent_event(struct perf_event *event, u64 nr, - int nmi, struct perf_sample_data *data, + struct perf_sample_data *data, struct pt_regs *regs) { struct hw_perf_event *hwc = &event->hw; @@ -4719,12 +4713,12 @@ static void perf_swevent_event(struct perf_event *event, u64 nr, return; if (nr == 1 && hwc->sample_period == 1 && !event->attr.freq) - return perf_swevent_overflow(event, 1, nmi, data, regs); + return perf_swevent_overflow(event, 1, data, regs); if (local64_add_negative(nr, &hwc->period_left)) return; - perf_swevent_overflow(event, 0, nmi, data, regs); + perf_swevent_overflow(event, 0, data, regs); } static int perf_exclude_event(struct perf_event *event, @@ -4812,7 +4806,7 @@ find_swevent_head(struct swevent_htable *swhash, struct perf_event *event) } static void do_perf_sw_event(enum perf_type_id type, u32 event_id, - u64 nr, int nmi, + u64 nr, struct perf_sample_data *data, struct pt_regs *regs) { @@ -4828,7 +4822,7 @@ static void do_perf_sw_event(enum perf_type_id type, u32 event_id, hlist_for_each_entry_rcu(event, node, head, hlist_entry) { if (perf_swevent_match(event, type, event_id, data, regs)) - perf_swevent_event(event, nr, nmi, data, regs); + perf_swevent_event(event, nr, data, regs); } end: rcu_read_unlock(); @@ -4849,8 +4843,7 @@ inline void perf_swevent_put_recursion_context(int rctx) put_recursion_context(swhash->recursion, rctx); } -void __perf_sw_event(u32 event_id, u64 nr, int nmi, - struct pt_regs *regs, u64 addr) +void __perf_sw_event(u32 event_id, u64 nr, struct pt_regs *regs, u64 addr) { struct perf_sample_data data; int rctx; @@ -4862,7 +4855,7 @@ void __perf_sw_event(u32 event_id, u64 nr, int nmi, perf_sample_data_init(&data, addr); - do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, nmi, &data, regs); + do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, &data, regs); perf_swevent_put_recursion_context(rctx); preempt_enable_notrace(); @@ -5110,7 +5103,7 @@ void perf_tp_event(u64 addr, u64 count, void *record, int entry_size, hlist_for_each_entry_rcu(event, node, head, hlist_entry) { if (perf_tp_event_match(event, &data, regs)) - perf_swevent_event(event, count, 1, &data, regs); + perf_swevent_event(event, count, &data, regs); } perf_swevent_put_recursion_context(rctx); @@ -5203,7 +5196,7 @@ void perf_bp_event(struct perf_event *bp, void *data) perf_sample_data_init(&sample, bp->attr.bp_addr); if (!bp->hw.state && !perf_exclude_event(bp, regs)) - perf_swevent_event(bp, 1, 1, &sample, regs); + perf_swevent_event(bp, 1, &sample, regs); } #endif @@ -5232,7 +5225,7 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer) if (regs && !perf_exclude_event(event, regs)) { if (!(event->attr.exclude_idle && current->pid == 0)) - if (perf_event_overflow(event, 0, &data, regs)) + if (perf_event_overflow(event, &data, regs)) ret = HRTIMER_NORESTART; } diff --git a/kernel/events/internal.h b/kernel/events/internal.h index 114f27f3a624..09097dd8116c 100644 --- a/kernel/events/internal.h +++ b/kernel/events/internal.h @@ -27,7 +27,6 @@ struct ring_buffer { void *data_pages[0]; }; - extern void rb_free(struct ring_buffer *rb); extern struct ring_buffer * rb_alloc(int nr_pages, long watermark, int cpu, int flags); diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c index fc2701c99207..8b3b73630fa4 100644 --- a/kernel/events/ring_buffer.c +++ b/kernel/events/ring_buffer.c @@ -38,11 +38,8 @@ static void perf_output_wakeup(struct perf_output_handle *handle) { atomic_set(&handle->rb->poll, POLL_IN); - if (handle->nmi) { - handle->event->pending_wakeup = 1; - irq_work_queue(&handle->event->pending); - } else - perf_event_wakeup(handle->event); + handle->event->pending_wakeup = 1; + irq_work_queue(&handle->event->pending); } /* @@ -102,7 +99,7 @@ out: int perf_output_begin(struct perf_output_handle *handle, struct perf_event *event, unsigned int size, - int nmi, int sample) + int sample) { struct ring_buffer *rb; unsigned long tail, offset, head; @@ -127,7 +124,6 @@ int perf_output_begin(struct perf_output_handle *handle, handle->rb = rb; handle->event = event; - handle->nmi = nmi; handle->sample = sample; if (!rb->nr_pages) diff --git a/kernel/sched.c b/kernel/sched.c index 3f2e502d609b..d08d110b8976 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2220,7 +2220,7 @@ void set_task_cpu(struct task_struct *p, unsigned int new_cpu) if (task_cpu(p) != new_cpu) { p->se.nr_migrations++; - perf_sw_event(PERF_COUNT_SW_CPU_MIGRATIONS, 1, 1, NULL, 0); + perf_sw_event(PERF_COUNT_SW_CPU_MIGRATIONS, 1, NULL, 0); } __set_task_cpu(p, new_cpu); diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 752b75ba662b..a6708e677a0a 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -211,7 +211,7 @@ static struct perf_event_attr wd_hw_attr = { }; /* Callback function for perf event subsystem */ -static void watchdog_overflow_callback(struct perf_event *event, int nmi, +static void watchdog_overflow_callback(struct perf_event *event, struct perf_sample_data *data, struct pt_regs *regs) { diff --git a/samples/hw_breakpoint/data_breakpoint.c b/samples/hw_breakpoint/data_breakpoint.c index 063653955f9f..7b164d3200ff 100644 --- a/samples/hw_breakpoint/data_breakpoint.c +++ b/samples/hw_breakpoint/data_breakpoint.c @@ -41,7 +41,7 @@ module_param_string(ksym, ksym_name, KSYM_NAME_LEN, S_IRUGO); MODULE_PARM_DESC(ksym, "Kernel symbol to monitor; this module will report any" " write operations on the kernel symbol"); -static void sample_hbp_handler(struct perf_event *bp, int nmi, +static void sample_hbp_handler(struct perf_event *bp, struct perf_sample_data *data, struct pt_regs *regs) { -- cgit v1.2.3 From a7ac67ea021b4603095d2aa458bc41641238f22c Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 27 Jun 2011 16:47:16 +0200 Subject: perf: Remove the perf_output_begin(.sample) argument Since only samples call perf_output_sample() its much saner (and more correct) to put the sample logic in there than in the perf_output_begin()/perf_output_end() pair. Saves a useless argument, reduces conditionals and shrinks struct perf_output_handle, win! Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-2crpvsx3cqu67q3zqjbnlpsc@git.kernel.org Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event_intel_ds.c | 2 +- include/linux/perf_event.h | 4 +--- kernel/events/core.c | 26 ++++++++++++++++++++------ kernel/events/ring_buffer.c | 19 +------------------ 4 files changed, 23 insertions(+), 28 deletions(-) (limited to 'arch') diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c index 0941f93f2940..1b1ef3addcfd 100644 --- a/arch/x86/kernel/cpu/perf_event_intel_ds.c +++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c @@ -340,7 +340,7 @@ static int intel_pmu_drain_bts_buffer(void) */ perf_prepare_sample(&header, &data, event, ®s); - if (perf_output_begin(&handle, event, header.size * (top - at), 1)) + if (perf_output_begin(&handle, event, header.size * (top - at))) return 1; for (; at < top; at++) { diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 0946a8bc098d..771b0b2845e4 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -925,7 +925,6 @@ struct perf_output_handle { unsigned long size; void *addr; int page; - int sample; }; #ifdef CONFIG_PERF_EVENTS @@ -1117,8 +1116,7 @@ extern void perf_bp_event(struct perf_event *event, void *data); #endif extern int perf_output_begin(struct perf_output_handle *handle, - struct perf_event *event, unsigned int size, - int sample); + struct perf_event *event, unsigned int size); extern void perf_output_end(struct perf_output_handle *handle); extern void perf_output_copy(struct perf_output_handle *handle, const void *buf, unsigned int len); diff --git a/kernel/events/core.c b/kernel/events/core.c index dbd1ca75bd3c..81de28dcca8c 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -3928,6 +3928,20 @@ void perf_output_sample(struct perf_output_handle *handle, perf_output_put(handle, raw); } } + + if (!event->attr.watermark) { + int wakeup_events = event->attr.wakeup_events; + + if (wakeup_events) { + struct ring_buffer *rb = handle->rb; + int events = local_inc_return(&rb->events); + + if (events >= wakeup_events) { + local_sub(wakeup_events, &rb->events); + local_inc(&rb->wakeup); + } + } + } } void perf_prepare_sample(struct perf_event_header *header, @@ -3984,7 +3998,7 @@ static void perf_event_output(struct perf_event *event, perf_prepare_sample(&header, data, event, regs); - if (perf_output_begin(&handle, event, header.size, 1)) + if (perf_output_begin(&handle, event, header.size)) goto exit; perf_output_sample(&handle, &header, data, event); @@ -4024,7 +4038,7 @@ perf_event_read_event(struct perf_event *event, int ret; perf_event_header__init_id(&read_event.header, &sample, event); - ret = perf_output_begin(&handle, event, read_event.header.size, 0); + ret = perf_output_begin(&handle, event, read_event.header.size); if (ret) return; @@ -4067,7 +4081,7 @@ static void perf_event_task_output(struct perf_event *event, perf_event_header__init_id(&task_event->event_id.header, &sample, event); ret = perf_output_begin(&handle, event, - task_event->event_id.header.size, 0); + task_event->event_id.header.size); if (ret) goto out; @@ -4204,7 +4218,7 @@ static void perf_event_comm_output(struct perf_event *event, perf_event_header__init_id(&comm_event->event_id.header, &sample, event); ret = perf_output_begin(&handle, event, - comm_event->event_id.header.size, 0); + comm_event->event_id.header.size); if (ret) goto out; @@ -4351,7 +4365,7 @@ static void perf_event_mmap_output(struct perf_event *event, perf_event_header__init_id(&mmap_event->event_id.header, &sample, event); ret = perf_output_begin(&handle, event, - mmap_event->event_id.header.size, 0); + mmap_event->event_id.header.size); if (ret) goto out; @@ -4546,7 +4560,7 @@ static void perf_log_throttle(struct perf_event *event, int enable) perf_event_header__init_id(&throttle_event.header, &sample, event); ret = perf_output_begin(&handle, event, - throttle_event.header.size, 0); + throttle_event.header.size); if (ret) return; diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c index 8b3b73630fa4..a2a29205cc0f 100644 --- a/kernel/events/ring_buffer.c +++ b/kernel/events/ring_buffer.c @@ -98,8 +98,7 @@ out: } int perf_output_begin(struct perf_output_handle *handle, - struct perf_event *event, unsigned int size, - int sample) + struct perf_event *event, unsigned int size) { struct ring_buffer *rb; unsigned long tail, offset, head; @@ -124,7 +123,6 @@ int perf_output_begin(struct perf_output_handle *handle, handle->rb = rb; handle->event = event; - handle->sample = sample; if (!rb->nr_pages) goto out; @@ -192,21 +190,6 @@ void perf_output_copy(struct perf_output_handle *handle, void perf_output_end(struct perf_output_handle *handle) { - struct perf_event *event = handle->event; - struct ring_buffer *rb = handle->rb; - - if (handle->sample && !event->attr.watermark) { - int wakeup_events = event->attr.wakeup_events; - - if (wakeup_events) { - int events = local_inc_return(&rb->events); - if (events >= wakeup_events) { - local_sub(wakeup_events, &rb->events); - local_inc(&rb->wakeup); - } - } - } - perf_output_put_handle(handle); rcu_read_unlock(); } -- cgit v1.2.3 From efc9f05df2dd171280dcb736a4d973ffefd5508e Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Mon, 6 Jun 2011 16:57:03 +0200 Subject: perf_events: Update Intel extra regs shared constraints management This patch improves the code managing the extra shared registers used for offcore_response events on Intel Nehalem/Westmere. The idea is to use static allocation instead of dynamic allocation. This simplifies greatly the get and put constraint routines for those events. The patch also renames per_core to shared_regs because the same data structure gets used whether or not HT is on. When HT is off, those events still need to coordination because they use a extra MSR that has to be shared within an event group. Signed-off-by: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20110606145703.GA7258@quad Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event.c | 78 +++++++--- arch/x86/kernel/cpu/perf_event_intel.c | 260 ++++++++++++++++----------------- include/linux/perf_event.h | 14 +- 3 files changed, 200 insertions(+), 152 deletions(-) (limited to 'arch') diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 5b86ec51534c..019fda7489e7 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -44,6 +44,29 @@ do { \ } while (0) #endif +/* + * | NHM/WSM | SNB | + * register ------------------------------- + * | HT | no HT | HT | no HT | + *----------------------------------------- + * offcore | core | core | cpu | core | + * lbr_sel | core | core | cpu | core | + * ld_lat | cpu | core | cpu | core | + *----------------------------------------- + * + * Given that there is a small number of shared regs, + * we can pre-allocate their slot in the per-cpu + * per-core reg tables. + */ +enum extra_reg_type { + EXTRA_REG_NONE = -1, /* not used */ + + EXTRA_REG_RSP_0 = 0, /* offcore_response_0 */ + EXTRA_REG_RSP_1 = 1, /* offcore_response_1 */ + + EXTRA_REG_MAX /* number of entries needed */ +}; + /* * best effort, GUP based copy_from_user() that assumes IRQ or NMI context */ @@ -132,11 +155,10 @@ struct cpu_hw_events { struct perf_branch_entry lbr_entries[MAX_LBR_ENTRIES]; /* - * Intel percore register state. - * Coordinate shared resources between HT threads. + * manage shared (per-core, per-cpu) registers + * used on Intel NHM/WSM/SNB */ - int percore_used; /* Used by this CPU? */ - struct intel_percore *per_core; + struct intel_shared_regs *shared_regs; /* * AMD specific bits @@ -186,27 +208,46 @@ struct cpu_hw_events { #define for_each_event_constraint(e, c) \ for ((e) = (c); (e)->weight; (e)++) +/* + * Per register state. + */ +struct er_account { + raw_spinlock_t lock; /* per-core: protect structure */ + u64 config; /* extra MSR config */ + u64 reg; /* extra MSR number */ + atomic_t ref; /* reference count */ +}; + /* * Extra registers for specific events. + * * Some events need large masks and require external MSRs. - * Define a mapping to these extra registers. + * Those extra MSRs end up being shared for all events on + * a PMU and sometimes between PMU of sibling HT threads. + * In either case, the kernel needs to handle conflicting + * accesses to those extra, shared, regs. The data structure + * to manage those registers is stored in cpu_hw_event. */ struct extra_reg { unsigned int event; unsigned int msr; u64 config_mask; u64 valid_mask; + int idx; /* per_xxx->regs[] reg index */ }; -#define EVENT_EXTRA_REG(e, ms, m, vm) { \ +#define EVENT_EXTRA_REG(e, ms, m, vm, i) { \ .event = (e), \ .msr = (ms), \ .config_mask = (m), \ .valid_mask = (vm), \ + .idx = EXTRA_REG_##i \ } -#define INTEL_EVENT_EXTRA_REG(event, msr, vm) \ - EVENT_EXTRA_REG(event, msr, ARCH_PERFMON_EVENTSEL_EVENT, vm) -#define EVENT_EXTRA_END EVENT_EXTRA_REG(0, 0, 0, 0) + +#define INTEL_EVENT_EXTRA_REG(event, msr, vm, idx) \ + EVENT_EXTRA_REG(event, msr, ARCH_PERFMON_EVENTSEL_EVENT, vm, idx) + +#define EVENT_EXTRA_END EVENT_EXTRA_REG(0, 0, 0, 0, RSP_0) union perf_capabilities { struct { @@ -253,7 +294,6 @@ struct x86_pmu { void (*put_event_constraints)(struct cpu_hw_events *cpuc, struct perf_event *event); struct event_constraint *event_constraints; - struct event_constraint *percore_constraints; void (*quirks)(void); int perfctr_second_write; @@ -400,10 +440,10 @@ static inline unsigned int x86_pmu_event_addr(int index) */ static int x86_pmu_extra_regs(u64 config, struct perf_event *event) { + struct hw_perf_event_extra *reg; struct extra_reg *er; - event->hw.extra_reg = 0; - event->hw.extra_config = 0; + reg = &event->hw.extra_reg; if (!x86_pmu.extra_regs) return 0; @@ -413,8 +453,10 @@ static int x86_pmu_extra_regs(u64 config, struct perf_event *event) continue; if (event->attr.config1 & ~er->valid_mask) return -EINVAL; - event->hw.extra_reg = er->msr; - event->hw.extra_config = event->attr.config1; + + reg->idx = er->idx; + reg->config = event->attr.config1; + reg->reg = er->msr; break; } return 0; @@ -713,6 +755,9 @@ static int __x86_pmu_event_init(struct perf_event *event) event->hw.last_cpu = -1; event->hw.last_tag = ~0ULL; + /* mark unused */ + event->hw.extra_reg.idx = EXTRA_REG_NONE; + return x86_pmu.hw_config(event); } @@ -754,8 +799,8 @@ static void x86_pmu_disable(struct pmu *pmu) static inline void __x86_pmu_enable_event(struct hw_perf_event *hwc, u64 enable_mask) { - if (hwc->extra_reg) - wrmsrl(hwc->extra_reg, hwc->extra_config); + if (hwc->extra_reg.reg) + wrmsrl(hwc->extra_reg.reg, hwc->extra_reg.config); wrmsrl(hwc->config_base, hwc->config | enable_mask); } @@ -1692,7 +1737,6 @@ static int validate_group(struct perf_event *event) fake_cpuc = kmalloc(sizeof(*fake_cpuc), GFP_KERNEL | __GFP_ZERO); if (!fake_cpuc) goto out; - /* * the event is not yet connected with its * siblings therefore we must first collect diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index d38b0020f775..6ad95baff856 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -1,25 +1,15 @@ #ifdef CONFIG_CPU_SUP_INTEL -#define MAX_EXTRA_REGS 2 - -/* - * Per register state. - */ -struct er_account { - int ref; /* reference count */ - unsigned int extra_reg; /* extra MSR number */ - u64 extra_config; /* extra MSR config */ -}; - /* - * Per core state - * This used to coordinate shared registers for HT threads. + * Per core/cpu state + * + * Used to coordinate shared registers between HT threads or + * among events on a single PMU. */ -struct intel_percore { - raw_spinlock_t lock; /* protect structure */ - struct er_account regs[MAX_EXTRA_REGS]; - int refcnt; /* number of threads */ - unsigned core_id; +struct intel_shared_regs { + struct er_account regs[EXTRA_REG_MAX]; + int refcnt; /* per-core: #HT threads */ + unsigned core_id; /* per-core: core id */ }; /* @@ -88,16 +78,10 @@ static struct event_constraint intel_nehalem_event_constraints[] __read_mostly = static struct extra_reg intel_nehalem_extra_regs[] __read_mostly = { - INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff), + INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff, RSP_0), EVENT_EXTRA_END }; -static struct event_constraint intel_nehalem_percore_constraints[] __read_mostly = -{ - INTEL_EVENT_CONSTRAINT(0xb7, 0), - EVENT_CONSTRAINT_END -}; - static struct event_constraint intel_westmere_event_constraints[] __read_mostly = { FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */ @@ -125,18 +109,11 @@ static struct event_constraint intel_snb_event_constraints[] __read_mostly = static struct extra_reg intel_westmere_extra_regs[] __read_mostly = { - INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff), - INTEL_EVENT_EXTRA_REG(0xbb, MSR_OFFCORE_RSP_1, 0xffff), + INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff, RSP_0), + INTEL_EVENT_EXTRA_REG(0xbb, MSR_OFFCORE_RSP_1, 0xffff, RSP_1), EVENT_EXTRA_END }; -static struct event_constraint intel_westmere_percore_constraints[] __read_mostly = -{ - INTEL_EVENT_CONSTRAINT(0xb7, 0), - INTEL_EVENT_CONSTRAINT(0xbb, 0), - EVENT_CONSTRAINT_END -}; - static struct event_constraint intel_gen_event_constraints[] __read_mostly = { FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */ @@ -1037,65 +1014,89 @@ intel_bts_constraints(struct perf_event *event) return NULL; } +/* + * manage allocation of shared extra msr for certain events + * + * sharing can be: + * per-cpu: to be shared between the various events on a single PMU + * per-core: per-cpu + shared by HT threads + */ static struct event_constraint * -intel_percore_constraints(struct cpu_hw_events *cpuc, struct perf_event *event) +__intel_shared_reg_get_constraints(struct cpu_hw_events *cpuc, + struct hw_perf_event_extra *reg) { - struct hw_perf_event *hwc = &event->hw; - unsigned int e = hwc->config & ARCH_PERFMON_EVENTSEL_EVENT; - struct event_constraint *c; - struct intel_percore *pc; + struct event_constraint *c = &emptyconstraint; struct er_account *era; - int i; - int free_slot; - int found; - if (!x86_pmu.percore_constraints || hwc->extra_alloc) - return NULL; + /* already allocated shared msr */ + if (reg->alloc || !cpuc->shared_regs) + return &unconstrained; - for (c = x86_pmu.percore_constraints; c->cmask; c++) { - if (e != c->code) - continue; + era = &cpuc->shared_regs->regs[reg->idx]; + + raw_spin_lock(&era->lock); + + if (!atomic_read(&era->ref) || era->config == reg->config) { + + /* lock in msr value */ + era->config = reg->config; + era->reg = reg->reg; + + /* one more user */ + atomic_inc(&era->ref); + + /* no need to reallocate during incremental event scheduling */ + reg->alloc = 1; /* - * Allocate resource per core. + * All events using extra_reg are unconstrained. + * Avoids calling x86_get_event_constraints() + * + * Must revisit if extra_reg controlling events + * ever have constraints. Worst case we go through + * the regular event constraint table. */ - pc = cpuc->per_core; - if (!pc) - break; - c = &emptyconstraint; - raw_spin_lock(&pc->lock); - free_slot = -1; - found = 0; - for (i = 0; i < MAX_EXTRA_REGS; i++) { - era = &pc->regs[i]; - if (era->ref > 0 && hwc->extra_reg == era->extra_reg) { - /* Allow sharing same config */ - if (hwc->extra_config == era->extra_config) { - era->ref++; - cpuc->percore_used = 1; - hwc->extra_alloc = 1; - c = NULL; - } - /* else conflict */ - found = 1; - break; - } else if (era->ref == 0 && free_slot == -1) - free_slot = i; - } - if (!found && free_slot != -1) { - era = &pc->regs[free_slot]; - era->ref = 1; - era->extra_reg = hwc->extra_reg; - era->extra_config = hwc->extra_config; - cpuc->percore_used = 1; - hwc->extra_alloc = 1; - c = NULL; - } - raw_spin_unlock(&pc->lock); - return c; + c = &unconstrained; } + raw_spin_unlock(&era->lock); - return NULL; + return c; +} + +static void +__intel_shared_reg_put_constraints(struct cpu_hw_events *cpuc, + struct hw_perf_event_extra *reg) +{ + struct er_account *era; + + /* + * only put constraint if extra reg was actually + * allocated. Also takes care of event which do + * not use an extra shared reg + */ + if (!reg->alloc) + return; + + era = &cpuc->shared_regs->regs[reg->idx]; + + /* one fewer user */ + atomic_dec(&era->ref); + + /* allocate again next time */ + reg->alloc = 0; +} + +static struct event_constraint * +intel_shared_regs_constraints(struct cpu_hw_events *cpuc, + struct perf_event *event) +{ + struct event_constraint *c = NULL; + struct hw_perf_event_extra *xreg; + + xreg = &event->hw.extra_reg; + if (xreg->idx != EXTRA_REG_NONE) + c = __intel_shared_reg_get_constraints(cpuc, xreg); + return c; } static struct event_constraint * @@ -1111,49 +1112,28 @@ intel_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event if (c) return c; - c = intel_percore_constraints(cpuc, event); + c = intel_shared_regs_constraints(cpuc, event); if (c) return c; return x86_get_event_constraints(cpuc, event); } -static void intel_put_event_constraints(struct cpu_hw_events *cpuc, +static void +intel_put_shared_regs_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event) { - struct extra_reg *er; - struct intel_percore *pc; - struct er_account *era; - struct hw_perf_event *hwc = &event->hw; - int i, allref; - - if (!cpuc->percore_used) - return; + struct hw_perf_event_extra *reg; - for (er = x86_pmu.extra_regs; er->msr; er++) { - if (er->event != (hwc->config & er->config_mask)) - continue; + reg = &event->hw.extra_reg; + if (reg->idx != EXTRA_REG_NONE) + __intel_shared_reg_put_constraints(cpuc, reg); +} - pc = cpuc->per_core; - raw_spin_lock(&pc->lock); - for (i = 0; i < MAX_EXTRA_REGS; i++) { - era = &pc->regs[i]; - if (era->ref > 0 && - era->extra_config == hwc->extra_config && - era->extra_reg == er->msr) { - era->ref--; - hwc->extra_alloc = 0; - break; - } - } - allref = 0; - for (i = 0; i < MAX_EXTRA_REGS; i++) - allref += pc->regs[i].ref; - if (allref == 0) - cpuc->percore_used = 0; - raw_spin_unlock(&pc->lock); - break; - } +static void intel_put_event_constraints(struct cpu_hw_events *cpuc, + struct perf_event *event) +{ + intel_put_shared_regs_event_constraints(cpuc, event); } static int intel_pmu_hw_config(struct perf_event *event) @@ -1231,20 +1211,36 @@ static __initconst const struct x86_pmu core_pmu = { .event_constraints = intel_core_event_constraints, }; +static struct intel_shared_regs *allocate_shared_regs(int cpu) +{ + struct intel_shared_regs *regs; + int i; + + regs = kzalloc_node(sizeof(struct intel_shared_regs), + GFP_KERNEL, cpu_to_node(cpu)); + if (regs) { + /* + * initialize the locks to keep lockdep happy + */ + for (i = 0; i < EXTRA_REG_MAX; i++) + raw_spin_lock_init(®s->regs[i].lock); + + regs->core_id = -1; + } + return regs; +} + static int intel_pmu_cpu_prepare(int cpu) { struct cpu_hw_events *cpuc = &per_cpu(cpu_hw_events, cpu); - if (!cpu_has_ht_siblings()) + if (!x86_pmu.extra_regs) return NOTIFY_OK; - cpuc->per_core = kzalloc_node(sizeof(struct intel_percore), - GFP_KERNEL, cpu_to_node(cpu)); - if (!cpuc->per_core) + cpuc->shared_regs = allocate_shared_regs(cpu); + if (!cpuc->shared_regs) return NOTIFY_BAD; - raw_spin_lock_init(&cpuc->per_core->lock); - cpuc->per_core->core_id = -1; return NOTIFY_OK; } @@ -1260,32 +1256,34 @@ static void intel_pmu_cpu_starting(int cpu) */ intel_pmu_lbr_reset(); - if (!cpu_has_ht_siblings()) + if (!cpuc->shared_regs) return; for_each_cpu(i, topology_thread_cpumask(cpu)) { - struct intel_percore *pc = per_cpu(cpu_hw_events, i).per_core; + struct intel_shared_regs *pc; + pc = per_cpu(cpu_hw_events, i).shared_regs; if (pc && pc->core_id == core_id) { - kfree(cpuc->per_core); - cpuc->per_core = pc; + kfree(cpuc->shared_regs); + cpuc->shared_regs = pc; break; } } - cpuc->per_core->core_id = core_id; - cpuc->per_core->refcnt++; + cpuc->shared_regs->core_id = core_id; + cpuc->shared_regs->refcnt++; } static void intel_pmu_cpu_dying(int cpu) { struct cpu_hw_events *cpuc = &per_cpu(cpu_hw_events, cpu); - struct intel_percore *pc = cpuc->per_core; + struct intel_shared_regs *pc; + pc = cpuc->shared_regs; if (pc) { if (pc->core_id == -1 || --pc->refcnt == 0) kfree(pc); - cpuc->per_core = NULL; + cpuc->shared_regs = NULL; } fini_debug_store_on_cpu(cpu); @@ -1436,7 +1434,6 @@ static __init int intel_pmu_init(void) x86_pmu.event_constraints = intel_nehalem_event_constraints; x86_pmu.pebs_constraints = intel_nehalem_pebs_event_constraints; - x86_pmu.percore_constraints = intel_nehalem_percore_constraints; x86_pmu.enable_all = intel_pmu_nhm_enable_all; x86_pmu.extra_regs = intel_nehalem_extra_regs; @@ -1481,7 +1478,6 @@ static __init int intel_pmu_init(void) intel_pmu_lbr_init_nhm(); x86_pmu.event_constraints = intel_westmere_event_constraints; - x86_pmu.percore_constraints = intel_westmere_percore_constraints; x86_pmu.enable_all = intel_pmu_nhm_enable_all; x86_pmu.pebs_constraints = intel_westmere_pebs_event_constraints; x86_pmu.extra_regs = intel_westmere_extra_regs; diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 771b0b2845e4..069315eefb22 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -536,6 +536,16 @@ struct perf_branch_stack { struct task_struct; +/* + * extra PMU register associated with an event + */ +struct hw_perf_event_extra { + u64 config; /* register value */ + unsigned int reg; /* register address or index */ + int alloc; /* extra register already allocated */ + int idx; /* index in shared_regs->regs[] */ +}; + /** * struct hw_perf_event - performance event hardware details: */ @@ -549,9 +559,7 @@ struct hw_perf_event { unsigned long event_base; int idx; int last_cpu; - unsigned int extra_reg; - u64 extra_config; - int extra_alloc; + struct hw_perf_event_extra extra_reg; }; struct { /* software */ struct hrtimer hrtimer; -- cgit v1.2.3 From cd8a38d33e2528998998bae70a45ad27e442f114 Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Mon, 6 Jun 2011 16:57:08 +0200 Subject: perf_events: Fix validation of events using an extra reg The validate_group() function needs to validate events with extra shared regs. Within an event group, only events with the same value for the extra reg can co-exist. This was not checked by validate_group() because it was missing the shared_regs logic. This patch changes the allocation of the fake cpuc used for validation to also point to a fake shared_regs structure such that group events be properly testing. It modifies __intel_shared_reg_get_constraints() to use spin_lock_irqsave() to avoid lockdep issues. Signed-off-by: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20110606145708.GA7279@quad Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event.c | 59 ++++++++++++++++++++++++++-------- arch/x86/kernel/cpu/perf_event_intel.c | 16 ++++++--- 2 files changed, 57 insertions(+), 18 deletions(-) (limited to 'arch') diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 019fda7489e7..9a0f55c99b6e 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1689,6 +1689,40 @@ static int x86_pmu_commit_txn(struct pmu *pmu) perf_pmu_enable(pmu); return 0; } +/* + * a fake_cpuc is used to validate event groups. Due to + * the extra reg logic, we need to also allocate a fake + * per_core and per_cpu structure. Otherwise, group events + * using extra reg may conflict without the kernel being + * able to catch this when the last event gets added to + * the group. + */ +static void free_fake_cpuc(struct cpu_hw_events *cpuc) +{ + kfree(cpuc->shared_regs); + kfree(cpuc); +} + +static struct cpu_hw_events *allocate_fake_cpuc(void) +{ + struct cpu_hw_events *cpuc; + int cpu = raw_smp_processor_id(); + + cpuc = kzalloc(sizeof(*cpuc), GFP_KERNEL); + if (!cpuc) + return ERR_PTR(-ENOMEM); + + /* only needed, if we have extra_regs */ + if (x86_pmu.extra_regs) { + cpuc->shared_regs = allocate_shared_regs(cpu); + if (!cpuc->shared_regs) + goto error; + } + return cpuc; +error: + free_fake_cpuc(cpuc); + return ERR_PTR(-ENOMEM); +} /* * validate that we can schedule this event @@ -1699,9 +1733,9 @@ static int validate_event(struct perf_event *event) struct event_constraint *c; int ret = 0; - fake_cpuc = kmalloc(sizeof(*fake_cpuc), GFP_KERNEL | __GFP_ZERO); - if (!fake_cpuc) - return -ENOMEM; + fake_cpuc = allocate_fake_cpuc(); + if (IS_ERR(fake_cpuc)) + return PTR_ERR(fake_cpuc); c = x86_pmu.get_event_constraints(fake_cpuc, event); @@ -1711,7 +1745,7 @@ static int validate_event(struct perf_event *event) if (x86_pmu.put_event_constraints) x86_pmu.put_event_constraints(fake_cpuc, event); - kfree(fake_cpuc); + free_fake_cpuc(fake_cpuc); return ret; } @@ -1731,35 +1765,32 @@ static int validate_group(struct perf_event *event) { struct perf_event *leader = event->group_leader; struct cpu_hw_events *fake_cpuc; - int ret, n; + int ret = -ENOSPC, n; - ret = -ENOMEM; - fake_cpuc = kmalloc(sizeof(*fake_cpuc), GFP_KERNEL | __GFP_ZERO); - if (!fake_cpuc) - goto out; + fake_cpuc = allocate_fake_cpuc(); + if (IS_ERR(fake_cpuc)) + return PTR_ERR(fake_cpuc); /* * the event is not yet connected with its * siblings therefore we must first collect * existing siblings, then add the new event * before we can simulate the scheduling */ - ret = -ENOSPC; n = collect_events(fake_cpuc, leader, true); if (n < 0) - goto out_free; + goto out; fake_cpuc->n_events = n; n = collect_events(fake_cpuc, event, false); if (n < 0) - goto out_free; + goto out; fake_cpuc->n_events = n; ret = x86_pmu.schedule_events(fake_cpuc, n, NULL); -out_free: - kfree(fake_cpuc); out: + free_fake_cpuc(fake_cpuc); return ret; } diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index 6ad95baff856..ac02b83e8614 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -1027,14 +1027,18 @@ __intel_shared_reg_get_constraints(struct cpu_hw_events *cpuc, { struct event_constraint *c = &emptyconstraint; struct er_account *era; + unsigned long flags; /* already allocated shared msr */ - if (reg->alloc || !cpuc->shared_regs) + if (reg->alloc) return &unconstrained; era = &cpuc->shared_regs->regs[reg->idx]; - - raw_spin_lock(&era->lock); + /* + * we use spin_lock_irqsave() to avoid lockdep issues when + * passing a fake cpuc + */ + raw_spin_lock_irqsave(&era->lock, flags); if (!atomic_read(&era->ref) || era->config == reg->config) { @@ -1058,7 +1062,7 @@ __intel_shared_reg_get_constraints(struct cpu_hw_events *cpuc, */ c = &unconstrained; } - raw_spin_unlock(&era->lock); + raw_spin_unlock_irqrestore(&era->lock, flags); return c; } @@ -1524,4 +1528,8 @@ static int intel_pmu_init(void) return 0; } +static struct intel_shared_regs *allocate_shared_regs(int cpu) +{ + return NULL; +} #endif /* CONFIG_CPU_SUP_INTEL */ -- cgit v1.2.3 From ee89cbc2d48150c7c0e9f2aaac00afde99af098c Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Mon, 6 Jun 2011 16:57:12 +0200 Subject: perf_events: Add Intel Sandy Bridge offcore_response low-level support This patch adds Intel Sandy Bridge offcore_response support by providing the low-level constraint table for those events. On Sandy Bridge, there are two offcore_response events. Each uses its own dedictated extra register. But those registers are NOT shared between sibling CPUs when HT is on unlike Nehalem/Westmere. They are always private to each CPU. But they still need to be controlled within an event group. All events within an event group must use the same value for the extra MSR. That's not controlled by the second patch in this series. Furthermore on Sandy Bridge, the offcore_response events have NO counter constraints contrary to what the official documentation indicates, so drop the events from the contraint table. Signed-off-by: Stephane Eranian Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20110606145712.GA7304@quad Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event.c | 1 + arch/x86/kernel/cpu/perf_event_intel.c | 13 ++++++++++--- 2 files changed, 11 insertions(+), 3 deletions(-) (limited to 'arch') diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 9a0f55c99b6e..583f3113436d 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -327,6 +327,7 @@ struct x86_pmu { * Extra registers for events */ struct extra_reg *extra_regs; + bool regs_no_ht_sharing; }; static struct x86_pmu x86_pmu __read_mostly; diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index ac02b83e8614..a674ae45a472 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -100,8 +100,6 @@ static struct event_constraint intel_snb_event_constraints[] __read_mostly = FIXED_EVENT_CONSTRAINT(0x003c, 1), /* CPU_CLK_UNHALTED.CORE */ /* FIXED_EVENT_CONSTRAINT(0x013c, 2), CPU_CLK_UNHALTED.REF */ INTEL_EVENT_CONSTRAINT(0x48, 0x4), /* L1D_PEND_MISS.PENDING */ - INTEL_EVENT_CONSTRAINT(0xb7, 0x1), /* OFF_CORE_RESPONSE_0 */ - INTEL_EVENT_CONSTRAINT(0xbb, 0x8), /* OFF_CORE_RESPONSE_1 */ INTEL_UEVENT_CONSTRAINT(0x01c0, 0x2), /* INST_RETIRED.PREC_DIST */ INTEL_EVENT_CONSTRAINT(0xcd, 0x8), /* MEM_TRANS_RETIRED.LOAD_LATENCY */ EVENT_CONSTRAINT_END @@ -122,6 +120,12 @@ static struct event_constraint intel_gen_event_constraints[] __read_mostly = EVENT_CONSTRAINT_END }; +static struct extra_reg intel_snb_extra_regs[] __read_mostly = { + INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0x3fffffffffull, RSP_0), + INTEL_EVENT_EXTRA_REG(0xbb, MSR_OFFCORE_RSP_1, 0x3fffffffffull, RSP_1), + EVENT_EXTRA_END +}; + static u64 intel_pmu_event_map(int hw_event) { return intel_perfmon_event_map[hw_event]; @@ -1260,7 +1264,7 @@ static void intel_pmu_cpu_starting(int cpu) */ intel_pmu_lbr_reset(); - if (!cpuc->shared_regs) + if (!cpuc->shared_regs || x86_pmu.regs_no_ht_sharing) return; for_each_cpu(i, topology_thread_cpumask(cpu)) { @@ -1502,6 +1506,9 @@ static __init int intel_pmu_init(void) x86_pmu.event_constraints = intel_snb_event_constraints; x86_pmu.pebs_constraints = intel_snb_pebs_events; + x86_pmu.extra_regs = intel_snb_extra_regs; + /* all extra regs are per-cpu when HT is on */ + x86_pmu.regs_no_ht_sharing = true; /* UOPS_ISSUED.ANY,c=1,i=1 to count stall cycles */ intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] = 0x180010e; -- cgit v1.2.3 From b79e8941fb9af07d810da91b4e29da2bba331b6e Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 23 May 2011 11:08:15 +0200 Subject: perf, intel: Try alternative OFFCORE encodings Since the OFFCORE registers are fully symmetric, try the other one when the specified one is already in use. Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1306141897.18455.8.camel@twins Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event.c | 5 +++- arch/x86/kernel/cpu/perf_event_intel.c | 44 ++++++++++++++++++++++++++++------ 2 files changed, 41 insertions(+), 8 deletions(-) (limited to 'arch') diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 583f3113436d..c53d433c3dde 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -327,9 +327,12 @@ struct x86_pmu { * Extra registers for events */ struct extra_reg *extra_regs; - bool regs_no_ht_sharing; + unsigned int er_flags; }; +#define ERF_NO_HT_SHARING 1 +#define ERF_HAS_RSP_1 2 + static struct x86_pmu x86_pmu __read_mostly; static DEFINE_PER_CPU(struct cpu_hw_events, cpu_hw_events) = { diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index a674ae45a472..5c448622468c 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -1018,6 +1018,29 @@ intel_bts_constraints(struct perf_event *event) return NULL; } +static bool intel_try_alt_er(struct perf_event *event, int orig_idx) +{ + if (!(x86_pmu.er_flags & ERF_HAS_RSP_1)) + return false; + + if (event->hw.extra_reg.idx == EXTRA_REG_RSP_0) { + event->hw.config &= ~INTEL_ARCH_EVENT_MASK; + event->hw.config |= 0x01bb; + event->hw.extra_reg.idx = EXTRA_REG_RSP_1; + event->hw.extra_reg.reg = MSR_OFFCORE_RSP_1; + } else if (event->hw.extra_reg.idx == EXTRA_REG_RSP_1) { + event->hw.config &= ~INTEL_ARCH_EVENT_MASK; + event->hw.config |= 0x01b7; + event->hw.extra_reg.idx = EXTRA_REG_RSP_0; + event->hw.extra_reg.reg = MSR_OFFCORE_RSP_0; + } + + if (event->hw.extra_reg.idx == orig_idx) + return false; + + return true; +} + /* * manage allocation of shared extra msr for certain events * @@ -1027,16 +1050,19 @@ intel_bts_constraints(struct perf_event *event) */ static struct event_constraint * __intel_shared_reg_get_constraints(struct cpu_hw_events *cpuc, - struct hw_perf_event_extra *reg) + struct perf_event *event) { struct event_constraint *c = &emptyconstraint; + struct hw_perf_event_extra *reg = &event->hw.extra_reg; struct er_account *era; unsigned long flags; + int orig_idx = reg->idx; /* already allocated shared msr */ if (reg->alloc) return &unconstrained; +again: era = &cpuc->shared_regs->regs[reg->idx]; /* * we use spin_lock_irqsave() to avoid lockdep issues when @@ -1065,6 +1091,9 @@ __intel_shared_reg_get_constraints(struct cpu_hw_events *cpuc, * the regular event constraint table. */ c = &unconstrained; + } else if (intel_try_alt_er(event, orig_idx)) { + raw_spin_unlock(&era->lock); + goto again; } raw_spin_unlock_irqrestore(&era->lock, flags); @@ -1099,11 +1128,10 @@ intel_shared_regs_constraints(struct cpu_hw_events *cpuc, struct perf_event *event) { struct event_constraint *c = NULL; - struct hw_perf_event_extra *xreg; - xreg = &event->hw.extra_reg; - if (xreg->idx != EXTRA_REG_NONE) - c = __intel_shared_reg_get_constraints(cpuc, xreg); + if (event->hw.extra_reg.idx != EXTRA_REG_NONE) + c = __intel_shared_reg_get_constraints(cpuc, event); + return c; } @@ -1264,7 +1292,7 @@ static void intel_pmu_cpu_starting(int cpu) */ intel_pmu_lbr_reset(); - if (!cpuc->shared_regs || x86_pmu.regs_no_ht_sharing) + if (!cpuc->shared_regs || (x86_pmu.er_flags & ERF_NO_HT_SHARING)) return; for_each_cpu(i, topology_thread_cpumask(cpu)) { @@ -1489,6 +1517,7 @@ static __init int intel_pmu_init(void) x86_pmu.enable_all = intel_pmu_nhm_enable_all; x86_pmu.pebs_constraints = intel_westmere_pebs_event_constraints; x86_pmu.extra_regs = intel_westmere_extra_regs; + x86_pmu.er_flags |= ERF_HAS_RSP_1; /* UOPS_ISSUED.STALLED_CYCLES */ intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] = 0x180010e; @@ -1508,7 +1537,8 @@ static __init int intel_pmu_init(void) x86_pmu.pebs_constraints = intel_snb_pebs_events; x86_pmu.extra_regs = intel_snb_extra_regs; /* all extra regs are per-cpu when HT is on */ - x86_pmu.regs_no_ht_sharing = true; + x86_pmu.er_flags |= ERF_HAS_RSP_1; + x86_pmu.er_flags |= ERF_NO_HT_SHARING; /* UOPS_ISSUED.ANY,c=1,i=1 to count stall cycles */ intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] = 0x180010e; -- cgit v1.2.3 From 89d6c0b5bdbb1927775584dcf532d98b3efe1477 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 22 Apr 2011 23:37:06 +0200 Subject: perf, arch: Add generic NODE cache events Add a NODE level to the generic cache events which is used to measure local vs remote memory accesses. Like all other cache events, an ACCESS is HIT+MISS, if there is no way to distinguish between reads and writes do reads only etc.. The below needs filling out for !x86 (which I filled out with unsupported events). I'm fairly sure ARM can leave it like that since it doesn't strike me as an architecture that even has NUMA support. SH might have something since it does appear to have some NUMA bits. Sparc64, PowerPC and MIPS certainly want a good look there since they clearly are NUMA capable. Signed-off-by: Peter Zijlstra Cc: David Miller Cc: Anton Blanchard Cc: David Daney Cc: Deng-Cheng Zhu Cc: Paul Mundt Cc: Will Deacon Cc: Robert Richter Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1303508226.4865.8.camel@laptop Signed-off-by: Ingo Molnar --- arch/arm/kernel/perf_event_v6.c | 28 ++++++++++++++++ arch/arm/kernel/perf_event_v7.c | 28 ++++++++++++++++ arch/arm/kernel/perf_event_xscale.c | 14 ++++++++ arch/mips/kernel/perf_event_mipsxx.c | 28 ++++++++++++++++ arch/powerpc/kernel/e500-pmu.c | 5 +++ arch/powerpc/kernel/mpc7450-pmu.c | 5 +++ arch/powerpc/kernel/power4-pmu.c | 5 +++ arch/powerpc/kernel/power5+-pmu.c | 5 +++ arch/powerpc/kernel/power5-pmu.c | 5 +++ arch/powerpc/kernel/power6-pmu.c | 5 +++ arch/powerpc/kernel/power7-pmu.c | 5 +++ arch/powerpc/kernel/ppc970-pmu.c | 5 +++ arch/sh/kernel/cpu/sh4/perf_event.c | 15 +++++++++ arch/sh/kernel/cpu/sh4a/perf_event.c | 15 +++++++++ arch/sparc/kernel/perf_event.c | 42 ++++++++++++++++++++++++ arch/x86/kernel/cpu/perf_event_amd.c | 14 ++++++++ arch/x86/kernel/cpu/perf_event_intel.c | 59 +++++++++++++++++++++++++++++++++- arch/x86/kernel/cpu/perf_event_p4.c | 14 ++++++++ include/linux/perf_event.h | 3 +- 19 files changed, 298 insertions(+), 2 deletions(-) (limited to 'arch') diff --git a/arch/arm/kernel/perf_event_v6.c b/arch/arm/kernel/perf_event_v6.c index 38dc4da1d612..dd7f3b9f4cb3 100644 --- a/arch/arm/kernel/perf_event_v6.c +++ b/arch/arm/kernel/perf_event_v6.c @@ -173,6 +173,20 @@ static const unsigned armv6_perf_cache_map[PERF_COUNT_HW_CACHE_MAX] [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, }, }, + [C(NODE)] = { + [C(OP_READ)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + [C(OP_WRITE)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + [C(OP_PREFETCH)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + }, }; enum armv6mpcore_perf_types { @@ -310,6 +324,20 @@ static const unsigned armv6mpcore_perf_cache_map[PERF_COUNT_HW_CACHE_MAX] [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, }, }, + [C(NODE)] = { + [C(OP_READ)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + [C(OP_WRITE)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + [C(OP_PREFETCH)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + }, }; static inline unsigned long diff --git a/arch/arm/kernel/perf_event_v7.c b/arch/arm/kernel/perf_event_v7.c index 6e5f8752303b..e20ca9cafef5 100644 --- a/arch/arm/kernel/perf_event_v7.c +++ b/arch/arm/kernel/perf_event_v7.c @@ -255,6 +255,20 @@ static const unsigned armv7_a8_perf_cache_map[PERF_COUNT_HW_CACHE_MAX] [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, }, }, + [C(NODE)] = { + [C(OP_READ)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + [C(OP_WRITE)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + [C(OP_PREFETCH)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + }, }; /* @@ -371,6 +385,20 @@ static const unsigned armv7_a9_perf_cache_map[PERF_COUNT_HW_CACHE_MAX] [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, }, }, + [C(NODE)] = { + [C(OP_READ)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + [C(OP_WRITE)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + [C(OP_PREFETCH)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + }, }; /* diff --git a/arch/arm/kernel/perf_event_xscale.c b/arch/arm/kernel/perf_event_xscale.c index 99b6b85c7e49..3c4397491d08 100644 --- a/arch/arm/kernel/perf_event_xscale.c +++ b/arch/arm/kernel/perf_event_xscale.c @@ -144,6 +144,20 @@ static const unsigned xscale_perf_cache_map[PERF_COUNT_HW_CACHE_MAX] [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, }, }, + [C(NODE)] = { + [C(OP_READ)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + [C(OP_WRITE)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + [C(OP_PREFETCH)] = { + [C(RESULT_ACCESS)] = CACHE_OP_UNSUPPORTED, + [C(RESULT_MISS)] = CACHE_OP_UNSUPPORTED, + }, + }, }; #define XSCALE_PMU_ENABLE 0x001 diff --git a/arch/mips/kernel/perf_event_mipsxx.c b/arch/mips/kernel/perf_event_mipsxx.c index 75266ff4cc33..e5ad09a9baf7 100644 --- a/arch/mips/kernel/perf_event_mipsxx.c +++ b/arch/mips/kernel/perf_event_mipsxx.c @@ -377,6 +377,20 @@ static const struct mips_perf_event mipsxxcore_cache_map [C(RESULT_MISS)] = { UNSUPPORTED_PERF_EVENT_ID }, }, }, +[C(NODE)] = { + [C(OP_READ)] = { + [C(RESULT_ACCESS)] = { UNSUPPORTED_PERF_EVENT_ID }, + [C(RESULT_MISS)] = { UNSUPPORTED_PERF_EVENT_ID }, + }, + [C(OP_WRITE)] = { + [C(RESULT_ACCESS)] = { UNSUPPORTED_PERF_EVENT_ID }, + [C(RESULT_MISS)] = { UNSUPPORTED_PERF_EVENT_ID }, + }, + [C(OP_PREFETCH)] = { + [C(RESULT_ACCESS)] = { UNSUPPORTED_PERF_EVENT_ID }, + [C(RESULT_MISS)] = { UNSUPPORTED_PERF_EVENT_ID }, + }, +}, }; /* 74K core has completely different cache event map. */ @@ -480,6 +494,20 @@ static const struct mips_perf_event mipsxx74Kcore_cache_map [C(RESULT_MISS)] = { UNSUPPORTED_PERF_EVENT_ID }, }, }, +[C(NODE)] = { + [C(OP_READ)] = { + [C(RESULT_ACCESS)] = { UNSUPPORTED_PERF_EVENT_ID }, + [C(RESULT_MISS)] = { UNSUPPORTED_PERF_EVENT_ID }, + }, + [C(OP_WRITE)] = { + [C(RESULT_ACCESS)] = { UNSUPPORTED_PERF_EVENT_ID }, + [C(RESULT_MISS)] = { UNSUPPORTED_PERF_EVENT_ID }, + }, + [C(OP_PREFETCH)] = { + [C(RESULT_ACCESS)] = { UNSUPPORTED_PERF_EVENT_ID }, + [C(RESULT_MISS)] = { UNSUPPORTED_PERF_EVENT_ID }, + }, +}, }; #ifdef CONFIG_MIPS_MT_SMP diff --git a/arch/powerpc/kernel/e500-pmu.c b/arch/powerpc/kernel/e500-pmu.c index b150b510510f..cb2e2949c8d1 100644 --- a/arch/powerpc/kernel/e500-pmu.c +++ b/arch/powerpc/kernel/e500-pmu.c @@ -75,6 +75,11 @@ static int e500_cache_events[C(MAX)][C(OP_MAX)][C(RESULT_MAX)] = { [C(OP_WRITE)] = { -1, -1 }, [C(OP_PREFETCH)] = { -1, -1 }, }, + [C(NODE)] = { /* RESULT_ACCESS RESULT_MISS */ + [C(OP_READ)] = { -1, -1 }, + [C(OP_WRITE)] = { -1, -1 }, + [C(OP_PREFETCH)] = { -1, -1 }, + }, }; static int num_events = 128; diff --git a/arch/powerpc/kernel/mpc7450-pmu.c b/arch/powerpc/kernel/mpc7450-pmu.c index 2cc5e0301d0b..845a58478890 100644 --- a/arch/powerpc/kernel/mpc7450-pmu.c +++ b/arch/powerpc/kernel/mpc7450-pmu.c @@ -388,6 +388,11 @@ static int mpc7450_cache_events[C(MAX)][C(OP_MAX)][C(RESULT_MAX)] = { [C(OP_WRITE)] = { -1, -1 }, [C(OP_PREFETCH)] = { -1, -1 }, }, + [C(NODE)] = { /* RESULT_ACCESS RESULT_MISS */ + [C(OP_READ)] = { -1, -1 }, + [C(OP_WRITE)] = { -1, -1 }, + [C(OP_PREFETCH)] = { -1, -1 }, + }, }; struct power_pmu mpc7450_pmu = { diff --git a/arch/powerpc/kernel/power4-pmu.c b/arch/powerpc/kernel/power4-pmu.c index ead8b3c2649e..e9dbc2d35c9c 100644 --- a/arch/powerpc/kernel/power4-pmu.c +++ b/arch/powerpc/kernel/power4-pmu.c @@ -587,6 +587,11 @@ static int power4_cache_events[C(MAX)][C(OP_MAX)][C(RESULT_MAX)] = { [C(OP_WRITE)] = { -1, -1 }, [C(OP_PREFETCH)] = { -1, -1 }, }, + [C(NODE)] = { /* RESULT_ACCESS RESULT_MISS */ + [C(OP_READ)] = { -1, -1 }, + [C(OP_WRITE)] = { -1, -1 }, + [C(OP_PREFETCH)] = { -1, -1 }, + }, }; static struct power_pmu power4_pmu = { diff --git a/arch/powerpc/kernel/power5+-pmu.c b/arch/powerpc/kernel/power5+-pmu.c index eca0ac595cb6..f58a2bd41b59 100644 --- a/arch/powerpc/kernel/power5+-pmu.c +++ b/arch/powerpc/kernel/power5+-pmu.c @@ -653,6 +653,11 @@ static int power5p_cache_events[C(MAX)][C(OP_MAX)][C(RESULT_MAX)] = { [C(OP_WRITE)] = { -1, -1 }, [C(OP_PREFETCH)] = { -1, -1 }, }, + [C(NODE)] = { /* RESULT_ACCESS RESULT_MISS */ + [C(OP_READ)] = { -1, -1 }, + [C(OP_WRITE)] = { -1, -1 }, + [C(OP_PREFETCH)] = { -1, -1 }, + }, }; static struct power_pmu power5p_pmu = { diff --git a/arch/powerpc/kernel/power5-pmu.c b/arch/powerpc/kernel/power5-pmu.c index d5ff0f64a5e6..b1acab684142 100644 --- a/arch/powerpc/kernel/power5-pmu.c +++ b/arch/powerpc/kernel/power5-pmu.c @@ -595,6 +595,11 @@ static int power5_cache_events[C(MAX)][C(OP_MAX)][C(RESULT_MAX)] = { [C(OP_WRITE)] = { -1, -1 }, [C(OP_PREFETCH)] = { -1, -1 }, }, + [C(NODE)] = { /* RESULT_ACCESS RESULT_MISS */ + [C(OP_READ)] = { -1, -1 }, + [C(OP_WRITE)] = { -1, -1 }, + [C(OP_PREFETCH)] = { -1, -1 }, + }, }; static struct power_pmu power5_pmu = { diff --git a/arch/powerpc/kernel/power6-pmu.c b/arch/powerpc/kernel/power6-pmu.c index 31603927e376..b24a3a23d073 100644 --- a/arch/powerpc/kernel/power6-pmu.c +++ b/arch/powerpc/kernel/power6-pmu.c @@ -516,6 +516,11 @@ static int power6_cache_events[C(MAX)][C(OP_MAX)][C(RESULT_MAX)] = { [C(OP_WRITE)] = { -1, -1 }, [C(OP_PREFETCH)] = { -1, -1 }, }, + [C(NODE)] = { /* RESULT_ACCESS RESULT_MISS */ + [C(OP_READ)] = { -1, -1 }, + [C(OP_WRITE)] = { -1, -1 }, + [C(OP_PREFETCH)] = { -1, -1 }, + }, }; static struct power_pmu power6_pmu = { diff --git a/arch/powerpc/kernel/power7-pmu.c b/arch/powerpc/kernel/power7-pmu.c index 593740fcb799..6d9dccb2ea59 100644 --- a/arch/powerpc/kernel/power7-pmu.c +++ b/arch/powerpc/kernel/power7-pmu.c @@ -342,6 +342,11 @@ static int power7_cache_events[C(MAX)][C(OP_MAX)][C(RESULT_MAX)] = { [C(OP_WRITE)] = { -1, -1 }, [C(OP_PREFETCH)] = { -1, -1 }, }, + [C(NODE)] = { /* RESULT_ACCESS RESULT_MISS */ + [C(OP_READ)] = { -1, -1 }, + [C(OP_WRITE)] = { -1, -1 }, + [C(OP_PREFETCH)] = { -1, -1 }, + }, }; static struct power_pmu power7_pmu = { diff --git a/arch/powerpc/kernel/ppc970-pmu.c b/arch/powerpc/kernel/ppc970-pmu.c index 9a6e093858fe..b121de9658eb 100644 --- a/arch/powerpc/kernel/ppc970-pmu.c +++ b/arch/powerpc/kernel/ppc970-pmu.c @@ -467,6 +467,11 @@ static int ppc970_cache_events[C(MAX)][C(OP_MAX)][C(RESULT_MAX)] = { [C(OP_WRITE)] = { -1, -1 }, [C(OP_PREFETCH)] = { -1, -1 }, }, + [C(NODE)] = { /* RESULT_ACCESS RESULT_MISS */ + [C(OP_READ)] = { -1, -1 }, + [C(OP_WRITE)] = { -1, -1 }, + [C(OP_PREFETCH)] = { -1, -1 }, + }, }; static struct power_pmu ppc970_pmu = { diff --git a/arch/sh/kernel/cpu/sh4/perf_event.c b/arch/sh/kernel/cpu/sh4/perf_event.c index 748955df018d..fa4f724b295a 100644 --- a/arch/sh/kernel/cpu/sh4/perf_event.c +++ b/arch/sh/kernel/cpu/sh4/perf_event.c @@ -180,6 +180,21 @@ static const int sh7750_cache_events [ C(RESULT_MISS) ] = -1, }, }, + + [ C(NODE) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, }; static int sh7750_event_map(int event) diff --git a/arch/sh/kernel/cpu/sh4a/perf_event.c b/arch/sh/kernel/cpu/sh4a/perf_event.c index 17e6bebfede0..84a2c396ceee 100644 --- a/arch/sh/kernel/cpu/sh4a/perf_event.c +++ b/arch/sh/kernel/cpu/sh4a/perf_event.c @@ -205,6 +205,21 @@ static const int sh4a_cache_events [ C(RESULT_MISS) ] = -1, }, }, + + [ C(NODE) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, }; static int sh4a_event_map(int event) diff --git a/arch/sparc/kernel/perf_event.c b/arch/sparc/kernel/perf_event.c index 0b32f2e9e08d..62a034318b18 100644 --- a/arch/sparc/kernel/perf_event.c +++ b/arch/sparc/kernel/perf_event.c @@ -246,6 +246,20 @@ static const cache_map_t ultra3_cache_map = { [ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED }, }, }, +[C(NODE)] = { + [C(OP_READ)] = { + [C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED }, + [C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED }, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED }, + [ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED }, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED }, + [ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED }, + }, +}, }; static const struct sparc_pmu ultra3_pmu = { @@ -361,6 +375,20 @@ static const cache_map_t niagara1_cache_map = { [ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED }, }, }, +[C(NODE)] = { + [C(OP_READ)] = { + [C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED }, + [C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED }, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED }, + [ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED }, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED }, + [ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED }, + }, +}, }; static const struct sparc_pmu niagara1_pmu = { @@ -473,6 +501,20 @@ static const cache_map_t niagara2_cache_map = { [ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED }, }, }, +[C(NODE)] = { + [C(OP_READ)] = { + [C(RESULT_ACCESS)] = { CACHE_OP_UNSUPPORTED }, + [C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED }, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED }, + [ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED }, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = { CACHE_OP_UNSUPPORTED }, + [ C(RESULT_MISS) ] = { CACHE_OP_UNSUPPORTED }, + }, +}, }; static const struct sparc_pmu niagara2_pmu = { diff --git a/arch/x86/kernel/cpu/perf_event_amd.c b/arch/x86/kernel/cpu/perf_event_amd.c index fe29c1d2219e..941caa2e449b 100644 --- a/arch/x86/kernel/cpu/perf_event_amd.c +++ b/arch/x86/kernel/cpu/perf_event_amd.c @@ -89,6 +89,20 @@ static __initconst const u64 amd_hw_cache_event_ids [ C(RESULT_MISS) ] = -1, }, }, + [ C(NODE) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0xb8e9, /* CPU Request to Memory, l+r */ + [ C(RESULT_MISS) ] = 0x98e9, /* CPU Request to Memory, r */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, }; /* diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index 5c448622468c..bf6f92f7c19d 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -226,6 +226,21 @@ static __initconst const u64 snb_hw_cache_event_ids [ C(RESULT_MISS) ] = -1, }, }, + [ C(NODE) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, + }; static __initconst const u64 westmere_hw_cache_event_ids @@ -327,6 +342,20 @@ static __initconst const u64 westmere_hw_cache_event_ids [ C(RESULT_MISS) ] = -1, }, }, + [ C(NODE) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x01b7, + [ C(RESULT_MISS) ] = 0x01b7, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x01b7, + [ C(RESULT_MISS) ] = 0x01b7, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x01b7, + [ C(RESULT_MISS) ] = 0x01b7, + }, + }, }; /* @@ -379,7 +408,21 @@ static __initconst const u64 nehalem_hw_cache_extra_regs [ C(RESULT_ACCESS) ] = NHM_DMND_PREFETCH|NHM_L3_ACCESS, [ C(RESULT_MISS) ] = NHM_DMND_PREFETCH|NHM_L3_MISS, }, - } + }, + [ C(NODE) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = NHM_DMND_READ|NHM_ALL_DRAM, + [ C(RESULT_MISS) ] = NHM_DMND_READ|NHM_REMOTE_DRAM, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = NHM_DMND_WRITE|NHM_ALL_DRAM, + [ C(RESULT_MISS) ] = NHM_DMND_WRITE|NHM_REMOTE_DRAM, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = NHM_DMND_PREFETCH|NHM_ALL_DRAM, + [ C(RESULT_MISS) ] = NHM_DMND_PREFETCH|NHM_REMOTE_DRAM, + }, + }, }; static __initconst const u64 nehalem_hw_cache_event_ids @@ -481,6 +524,20 @@ static __initconst const u64 nehalem_hw_cache_event_ids [ C(RESULT_MISS) ] = -1, }, }, + [ C(NODE) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x01b7, + [ C(RESULT_MISS) ] = 0x01b7, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x01b7, + [ C(RESULT_MISS) ] = 0x01b7, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x01b7, + [ C(RESULT_MISS) ] = 0x01b7, + }, + }, }; static __initconst const u64 core2_hw_cache_event_ids diff --git a/arch/x86/kernel/cpu/perf_event_p4.c b/arch/x86/kernel/cpu/perf_event_p4.c index d6e6a67b9608..fb901c5080f7 100644 --- a/arch/x86/kernel/cpu/perf_event_p4.c +++ b/arch/x86/kernel/cpu/perf_event_p4.c @@ -554,6 +554,20 @@ static __initconst const u64 p4_hw_cache_event_ids [ C(RESULT_MISS) ] = -1, }, }, + [ C(NODE) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, }; static u64 p4_general_events[PERF_COUNT_HW_MAX] = { diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 069315eefb22..a5f54b973bdb 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -61,7 +61,7 @@ enum perf_hw_id { /* * Generalized hardware cache events: * - * { L1-D, L1-I, LLC, ITLB, DTLB, BPU } x + * { L1-D, L1-I, LLC, ITLB, DTLB, BPU, NODE } x * { read, write, prefetch } x * { accesses, misses } */ @@ -72,6 +72,7 @@ enum perf_hw_cache_id { PERF_COUNT_HW_CACHE_DTLB = 3, PERF_COUNT_HW_CACHE_ITLB = 4, PERF_COUNT_HW_CACHE_BPU = 5, + PERF_COUNT_HW_CACHE_NODE = 6, PERF_COUNT_HW_CACHE_MAX, /* non-ABI */ }; -- cgit v1.2.3 From 4dc0da86967d5463708631d02a70cfed5b104884 Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Wed, 29 Jun 2011 18:42:35 +0300 Subject: perf: Add context field to perf_event The perf_event overflow handler does not receive any caller-derived argument, so many callers need to resort to looking up the perf_event in their local data structure. This is ugly and doesn't scale if a single callback services many perf_events. Fix by adding a context parameter to perf_event_create_kernel_counter() (and derived hardware breakpoints APIs) and storing it in the perf_event. The field can be accessed from the callback as event->overflow_handler_context. All callers are updated. Signed-off-by: Avi Kivity Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1309362157-6596-2-git-send-email-avi@redhat.com Signed-off-by: Ingo Molnar --- arch/arm/kernel/ptrace.c | 3 ++- arch/powerpc/kernel/ptrace.c | 2 +- arch/sh/kernel/ptrace_32.c | 3 ++- arch/x86/kernel/kgdb.c | 2 +- arch/x86/kernel/ptrace.c | 3 ++- drivers/oprofile/oprofile_perf.c | 2 +- include/linux/hw_breakpoint.h | 10 ++++++++-- include/linux/perf_event.h | 4 +++- kernel/events/core.c | 21 +++++++++++++++------ kernel/events/hw_breakpoint.c | 10 +++++++--- kernel/watchdog.c | 2 +- samples/hw_breakpoint/data_breakpoint.c | 2 +- 12 files changed, 44 insertions(+), 20 deletions(-) (limited to 'arch') diff --git a/arch/arm/kernel/ptrace.c b/arch/arm/kernel/ptrace.c index 0c9b1054f790..5c199610719f 100644 --- a/arch/arm/kernel/ptrace.c +++ b/arch/arm/kernel/ptrace.c @@ -479,7 +479,8 @@ static struct perf_event *ptrace_hbp_create(struct task_struct *tsk, int type) attr.bp_type = type; attr.disabled = 1; - return register_user_hw_breakpoint(&attr, ptrace_hbptriggered, tsk); + return register_user_hw_breakpoint(&attr, ptrace_hbptriggered, NULL, + tsk); } static int ptrace_gethbpregs(struct task_struct *tsk, long num, diff --git a/arch/powerpc/kernel/ptrace.c b/arch/powerpc/kernel/ptrace.c index 3177617af2ef..05b7dd217f60 100644 --- a/arch/powerpc/kernel/ptrace.c +++ b/arch/powerpc/kernel/ptrace.c @@ -973,7 +973,7 @@ int ptrace_set_debugreg(struct task_struct *task, unsigned long addr, &attr.bp_type); thread->ptrace_bps[0] = bp = register_user_hw_breakpoint(&attr, - ptrace_triggered, task); + ptrace_triggered, NULL, task); if (IS_ERR(bp)) { thread->ptrace_bps[0] = NULL; ptrace_put_breakpoints(task); diff --git a/arch/sh/kernel/ptrace_32.c b/arch/sh/kernel/ptrace_32.c index 8051976100a6..92b3c276339a 100644 --- a/arch/sh/kernel/ptrace_32.c +++ b/arch/sh/kernel/ptrace_32.c @@ -91,7 +91,8 @@ static int set_single_step(struct task_struct *tsk, unsigned long addr) attr.bp_len = HW_BREAKPOINT_LEN_2; attr.bp_type = HW_BREAKPOINT_R; - bp = register_user_hw_breakpoint(&attr, ptrace_triggered, tsk); + bp = register_user_hw_breakpoint(&attr, ptrace_triggered, + NULL, tsk); if (IS_ERR(bp)) return PTR_ERR(bp); diff --git a/arch/x86/kernel/kgdb.c b/arch/x86/kernel/kgdb.c index 98da6a7b5e82..00354d4919a9 100644 --- a/arch/x86/kernel/kgdb.c +++ b/arch/x86/kernel/kgdb.c @@ -638,7 +638,7 @@ void kgdb_arch_late(void) for (i = 0; i < HBP_NUM; i++) { if (breakinfo[i].pev) continue; - breakinfo[i].pev = register_wide_hw_breakpoint(&attr, NULL); + breakinfo[i].pev = register_wide_hw_breakpoint(&attr, NULL, NULL); if (IS_ERR((void * __force)breakinfo[i].pev)) { printk(KERN_ERR "kgdb: Could not allocate hw" "breakpoints\nDisabling the kernel debugger\n"); diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index 11db2e9b860a..82528799c5de 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -715,7 +715,8 @@ static int ptrace_set_breakpoint_addr(struct task_struct *tsk, int nr, attr.bp_type = HW_BREAKPOINT_W; attr.disabled = 1; - bp = register_user_hw_breakpoint(&attr, ptrace_triggered, tsk); + bp = register_user_hw_breakpoint(&attr, ptrace_triggered, + NULL, tsk); /* * CHECKME: the previous code returned -EIO if the addr wasn't diff --git a/drivers/oprofile/oprofile_perf.c b/drivers/oprofile/oprofile_perf.c index 9046f7b2ed79..59acf9ef78a4 100644 --- a/drivers/oprofile/oprofile_perf.c +++ b/drivers/oprofile/oprofile_perf.c @@ -79,7 +79,7 @@ static int op_create_counter(int cpu, int event) pevent = perf_event_create_kernel_counter(&counter_config[event].attr, cpu, NULL, - op_overflow_handler); + op_overflow_handler, NULL); if (IS_ERR(pevent)) return PTR_ERR(pevent); diff --git a/include/linux/hw_breakpoint.h b/include/linux/hw_breakpoint.h index d1e55fed2c7d..6ae9c631a1be 100644 --- a/include/linux/hw_breakpoint.h +++ b/include/linux/hw_breakpoint.h @@ -73,6 +73,7 @@ static inline unsigned long hw_breakpoint_len(struct perf_event *bp) extern struct perf_event * register_user_hw_breakpoint(struct perf_event_attr *attr, perf_overflow_handler_t triggered, + void *context, struct task_struct *tsk); /* FIXME: only change from the attr, and don't unregister */ @@ -85,11 +86,13 @@ modify_user_hw_breakpoint(struct perf_event *bp, struct perf_event_attr *attr); extern struct perf_event * register_wide_hw_breakpoint_cpu(struct perf_event_attr *attr, perf_overflow_handler_t triggered, + void *context, int cpu); extern struct perf_event * __percpu * register_wide_hw_breakpoint(struct perf_event_attr *attr, - perf_overflow_handler_t triggered); + perf_overflow_handler_t triggered, + void *context); extern int register_perf_hw_breakpoint(struct perf_event *bp); extern int __register_perf_hw_breakpoint(struct perf_event *bp); @@ -115,6 +118,7 @@ static inline int __init init_hw_breakpoint(void) { return 0; } static inline struct perf_event * register_user_hw_breakpoint(struct perf_event_attr *attr, perf_overflow_handler_t triggered, + void *context, struct task_struct *tsk) { return NULL; } static inline int modify_user_hw_breakpoint(struct perf_event *bp, @@ -122,10 +126,12 @@ modify_user_hw_breakpoint(struct perf_event *bp, static inline struct perf_event * register_wide_hw_breakpoint_cpu(struct perf_event_attr *attr, perf_overflow_handler_t triggered, + void *context, int cpu) { return NULL; } static inline struct perf_event * __percpu * register_wide_hw_breakpoint(struct perf_event_attr *attr, - perf_overflow_handler_t triggered) { return NULL; } + perf_overflow_handler_t triggered, + void *context) { return NULL; } static inline int register_perf_hw_breakpoint(struct perf_event *bp) { return -ENOSYS; } static inline int diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index a5f54b973bdb..2a08cacb1628 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -839,6 +839,7 @@ struct perf_event { u64 id; perf_overflow_handler_t overflow_handler; + void *overflow_handler_context; #ifdef CONFIG_EVENT_TRACING struct ftrace_event_call *tp_event; @@ -960,7 +961,8 @@ extern struct perf_event * perf_event_create_kernel_counter(struct perf_event_attr *attr, int cpu, struct task_struct *task, - perf_overflow_handler_t callback); + perf_overflow_handler_t callback, + void *context); extern u64 perf_event_read_value(struct perf_event *event, u64 *enabled, u64 *running); diff --git a/kernel/events/core.c b/kernel/events/core.c index 81de28dcca8c..ba8e0f4a20e6 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -5745,7 +5745,8 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu, struct task_struct *task, struct perf_event *group_leader, struct perf_event *parent_event, - perf_overflow_handler_t overflow_handler) + perf_overflow_handler_t overflow_handler, + void *context) { struct pmu *pmu; struct perf_event *event; @@ -5803,10 +5804,13 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu, #endif } - if (!overflow_handler && parent_event) + if (!overflow_handler && parent_event) { overflow_handler = parent_event->overflow_handler; + context = parent_event->overflow_handler_context; + } event->overflow_handler = overflow_handler; + event->overflow_handler_context = context; if (attr->disabled) event->state = PERF_EVENT_STATE_OFF; @@ -6073,7 +6077,8 @@ SYSCALL_DEFINE5(perf_event_open, } } - event = perf_event_alloc(&attr, cpu, task, group_leader, NULL, NULL); + event = perf_event_alloc(&attr, cpu, task, group_leader, NULL, + NULL, NULL); if (IS_ERR(event)) { err = PTR_ERR(event); goto err_task; @@ -6258,7 +6263,8 @@ err_fd: struct perf_event * perf_event_create_kernel_counter(struct perf_event_attr *attr, int cpu, struct task_struct *task, - perf_overflow_handler_t overflow_handler) + perf_overflow_handler_t overflow_handler, + void *context) { struct perf_event_context *ctx; struct perf_event *event; @@ -6268,7 +6274,8 @@ perf_event_create_kernel_counter(struct perf_event_attr *attr, int cpu, * Get the target context (task or percpu): */ - event = perf_event_alloc(attr, cpu, task, NULL, NULL, overflow_handler); + event = perf_event_alloc(attr, cpu, task, NULL, NULL, + overflow_handler, context); if (IS_ERR(event)) { err = PTR_ERR(event); goto err; @@ -6552,7 +6559,7 @@ inherit_event(struct perf_event *parent_event, parent_event->cpu, child, group_leader, parent_event, - NULL); + NULL, NULL); if (IS_ERR(child_event)) return child_event; get_ctx(child_ctx); @@ -6579,6 +6586,8 @@ inherit_event(struct perf_event *parent_event, child_event->ctx = child_ctx; child_event->overflow_handler = parent_event->overflow_handler; + child_event->overflow_handler_context + = parent_event->overflow_handler_context; /* * Precalculate sample_data sizes diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c index 086adf25a55e..b7971d6f38bf 100644 --- a/kernel/events/hw_breakpoint.c +++ b/kernel/events/hw_breakpoint.c @@ -431,9 +431,11 @@ int register_perf_hw_breakpoint(struct perf_event *bp) struct perf_event * register_user_hw_breakpoint(struct perf_event_attr *attr, perf_overflow_handler_t triggered, + void *context, struct task_struct *tsk) { - return perf_event_create_kernel_counter(attr, -1, tsk, triggered); + return perf_event_create_kernel_counter(attr, -1, tsk, triggered, + context); } EXPORT_SYMBOL_GPL(register_user_hw_breakpoint); @@ -502,7 +504,8 @@ EXPORT_SYMBOL_GPL(unregister_hw_breakpoint); */ struct perf_event * __percpu * register_wide_hw_breakpoint(struct perf_event_attr *attr, - perf_overflow_handler_t triggered) + perf_overflow_handler_t triggered, + void *context) { struct perf_event * __percpu *cpu_events, **pevent, *bp; long err; @@ -515,7 +518,8 @@ register_wide_hw_breakpoint(struct perf_event_attr *attr, get_online_cpus(); for_each_online_cpu(cpu) { pevent = per_cpu_ptr(cpu_events, cpu); - bp = perf_event_create_kernel_counter(attr, cpu, NULL, triggered); + bp = perf_event_create_kernel_counter(attr, cpu, NULL, + triggered, context); *pevent = bp; diff --git a/kernel/watchdog.c b/kernel/watchdog.c index a6708e677a0a..a933e3a0398b 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -375,7 +375,7 @@ static int watchdog_nmi_enable(int cpu) hw_nmi_watchdog_set_attr(wd_attr); /* Try to register using hardware perf events */ - event = perf_event_create_kernel_counter(wd_attr, cpu, NULL, watchdog_overflow_callback); + event = perf_event_create_kernel_counter(wd_attr, cpu, NULL, watchdog_overflow_callback, NULL); if (!IS_ERR(event)) { printk(KERN_INFO "NMI watchdog enabled, takes one hw-pmu counter.\n"); goto out_save; diff --git a/samples/hw_breakpoint/data_breakpoint.c b/samples/hw_breakpoint/data_breakpoint.c index 7b164d3200ff..ef7f32291852 100644 --- a/samples/hw_breakpoint/data_breakpoint.c +++ b/samples/hw_breakpoint/data_breakpoint.c @@ -60,7 +60,7 @@ static int __init hw_break_module_init(void) attr.bp_len = HW_BREAKPOINT_LEN_4; attr.bp_type = HW_BREAKPOINT_W | HW_BREAKPOINT_R; - sample_hbp = register_wide_hw_breakpoint(&attr, sample_hbp_handler); + sample_hbp = register_wide_hw_breakpoint(&attr, sample_hbp_handler, NULL); if (IS_ERR((void __force *)sample_hbp)) { ret = PTR_ERR((void __force *)sample_hbp); goto fail; -- cgit v1.2.3 From 0af3ac1fdb9d5c297b4b07c9e0172531d42b6716 Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Wed, 29 Jun 2011 18:42:36 +0300 Subject: x86, perf: Add constraints for architectural PMU The v1 PMU does not have any fixed counters. Using the v2 constraints, which do have fixed counters, causes an additional choice to be present in the weight calculation, but not when actually scheduling the event, leading to an event being not scheduled at all. Signed-off-by: Avi Kivity Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1309362157-6596-3-git-send-email-avi@redhat.com Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event_intel.c | 23 ++++++++++++++++++----- 1 file changed, 18 insertions(+), 5 deletions(-) (limited to 'arch') diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index bf6f92f7c19d..45fbb8f7f549 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -112,6 +112,11 @@ static struct extra_reg intel_westmere_extra_regs[] __read_mostly = EVENT_EXTRA_END }; +static struct event_constraint intel_v1_event_constraints[] __read_mostly = +{ + EVENT_CONSTRAINT_END +}; + static struct event_constraint intel_gen_event_constraints[] __read_mostly = { FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */ @@ -1606,11 +1611,19 @@ static __init int intel_pmu_init(void) break; default: - /* - * default constraints for v2 and up - */ - x86_pmu.event_constraints = intel_gen_event_constraints; - pr_cont("generic architected perfmon, "); + switch (x86_pmu.version) { + case 1: + x86_pmu.event_constraints = intel_v1_event_constraints; + pr_cont("generic architected perfmon v1, "); + break; + default: + /* + * default constraints for v2 and up + */ + x86_pmu.event_constraints = intel_gen_event_constraints; + pr_cont("generic architected perfmon, "); + break; + } } return 0; } -- cgit v1.2.3 From 29dfc4fd7e792854b30d771d4e299fbdd2adb763 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 1 Jul 2011 15:24:53 +0200 Subject: perf, powerpc: Fix build borkage The patch a8b0ca17b80e ("perf: Remove the nmi parameter from the swevent and overflow interface") missed a spot in the ppc hw_breakpoint code, fix this up so things compile again. Reported-by: Ingo Molnar Cc: Anton Blanchard Cc: Eric B Munson Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-09pfip95g88s70iwkxu6nnbt@git.kernel.org Signed-off-by: Ingo Molnar --- arch/powerpc/include/asm/hw_breakpoint.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/powerpc/include/asm/hw_breakpoint.h b/arch/powerpc/include/asm/hw_breakpoint.h index 1c33ec17ca36..80fd4d2b4a62 100644 --- a/arch/powerpc/include/asm/hw_breakpoint.h +++ b/arch/powerpc/include/asm/hw_breakpoint.h @@ -57,7 +57,7 @@ void hw_breakpoint_pmu_read(struct perf_event *bp); extern void flush_ptrace_hw_breakpoint(struct task_struct *tsk); extern struct pmu perf_ops_bp; -extern void ptrace_triggered(struct perf_event *bp, int nmi, +extern void ptrace_triggered(struct perf_event *bp, struct perf_sample_data *data, struct pt_regs *regs); static inline void hw_breakpoint_disable(void) { -- cgit v1.2.3 From 9e46294dadedc0c04adcb8ce760bd2cd74f7332d Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 2 Jul 2011 15:00:52 +0200 Subject: x86: Save stack pointer in perf live regs savings In order to prepare for fetching the stack pointer from the regs when possible in dump_trace() instead of taking the local one, save the current stack pointer in perf live regs saving. Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Thomas Gleixner Cc: H. Peter Anvin Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo --- arch/x86/include/asm/perf_event.h | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'arch') diff --git a/arch/x86/include/asm/perf_event.h b/arch/x86/include/asm/perf_event.h index d9d4dae305f6..094fb30817ab 100644 --- a/arch/x86/include/asm/perf_event.h +++ b/arch/x86/include/asm/perf_event.h @@ -152,6 +152,11 @@ extern unsigned long perf_misc_flags(struct pt_regs *regs); (regs)->bp = caller_frame_pointer(); \ (regs)->cs = __KERNEL_CS; \ regs->flags = 0; \ + asm volatile( \ + _ASM_MOV "%%"_ASM_SP ", %0\n" \ + : "=m" ((regs)->sp) \ + :: "memory" \ + ); \ } #else -- cgit v1.2.3 From 47ce11a2b6519f9c7843223ea8e561eb71ea5896 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 30 Jun 2011 19:04:56 +0200 Subject: x86: Fetch stack from regs when possible in dump_trace() When regs are passed to dump_stack(), we fetch the frame pointer from the regs but the stack pointer is taken from the current frame. Thus the frame and stack pointers may not come from the same context. For example this can result in the unwinder to think the context is in irq, due to the current value of the stack, but the frame pointer coming from the regs points to a frame from another place. It then tries to fix up the irq link but ends up dereferencing a random frame pointer that doesn't belong to the irq stack: [ 9131.706906] ------------[ cut here ]------------ [ 9131.707003] WARNING: at arch/x86/kernel/dumpstack_64.c:129 dump_trace+0x2aa/0x330() [ 9131.707003] Hardware name: AMD690VM-FMH [ 9131.707003] Perf: bad frame pointer = 0000000000000005 in callchain [ 9131.707003] Modules linked in: [ 9131.707003] Pid: 1050, comm: perf Not tainted 3.0.0-rc3+ #181 [ 9131.707003] Call Trace: [ 9131.707003] [] warn_slowpath_common+0x7a/0xb0 [ 9131.707003] [] warn_slowpath_fmt+0x41/0x50 [ 9131.707003] [] ? bad_to_user+0x6d/0x10be [ 9131.707003] [] dump_trace+0x2aa/0x330 [ 9131.707003] [] ? native_sched_clock+0x13/0x50 [ 9131.707003] [] perf_callchain_kernel+0x54/0x70 [ 9131.707003] [] perf_prepare_sample+0x19f/0x2a0 [ 9131.707003] [] __perf_event_overflow+0x16c/0x290 [ 9131.707003] [] ? __perf_event_overflow+0x130/0x290 [ 9131.707003] [] ? native_sched_clock+0x13/0x50 [ 9131.707003] [] ? sched_clock+0x9/0x10 [ 9131.707003] [] ? T.375+0x15/0x90 [ 9131.707003] [] ? trace_hardirqs_on_caller+0x64/0x180 [ 9131.707003] [] ? trace_hardirqs_off+0xd/0x10 [ 9131.707003] [] perf_event_overflow+0x14/0x20 [ 9131.707003] [] perf_swevent_hrtimer+0x11c/0x130 [ 9131.707003] [] ? error_exit+0x51/0xb0 [ 9131.707003] [] __run_hrtimer+0x83/0x1e0 [ 9131.707003] [] ? perf_event_overflow+0x20/0x20 [ 9131.707003] [] hrtimer_interrupt+0x106/0x250 [ 9131.707003] [] ? trace_hardirqs_off_thunk+0x3a/0x3c [ 9131.707003] [] smp_apic_timer_interrupt+0x53/0x90 [ 9131.707003] [] apic_timer_interrupt+0x13/0x20 [ 9131.707003] [] ? error_exit+0x51/0xb0 [ 9131.707003] [] ? error_exit+0x4c/0xb0 [ 9131.707003] ---[ end trace b2560d4876709347 ]--- Fix this by simply taking the stack pointer from regs->sp when regs are provided. Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Thomas Gleixner Cc: H. Peter Anvin Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo --- arch/x86/kernel/dumpstack_64.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'arch') diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index e71c98d3c0d2..788295cbe4a7 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -155,9 +155,12 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, task = current; if (!stack) { - stack = &dummy; - if (task && task != current) + if (regs) + stack = (unsigned long *)regs->sp; + else if (task && task != current) stack = (unsigned long *)task->thread.sp; + else + stack = &dummy; } if (!bp) -- cgit v1.2.3 From 1871853f7abc3c727c4346539c5062cbeaf016a4 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 1 Jul 2011 01:51:22 +0200 Subject: x86,64: Simplify save_regs() The save_regs function that saves the regs on low level irq entry is complicated because of the fact it changes its stack in the middle and also because it manipulates data allocated in the caller frame and accesses there are directly calculated from callee rsp value with the return address in the middle of the way. This complicates the static stack offsets calculation and require more dynamic ones. It also needs a save/restore of the function's return address. To simplify and optimize this, turn save_regs() into a macro. Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Thomas Gleixner Cc: H. Peter Anvin Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Jan Beulich --- arch/x86/kernel/entry_64.S | 44 +++++++++++++++++--------------------------- 1 file changed, 17 insertions(+), 27 deletions(-) (limited to 'arch') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 8a445a0c989e..b6b2e85454cf 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -297,27 +297,22 @@ ENDPROC(native_usergs_sysret64) .endm /* save partial stack frame */ - .pushsection .kprobes.text, "ax" -ENTRY(save_args) - XCPT_FRAME + .macro SAVE_ARGS_IRQ cld - /* - * start from rbp in pt_regs and jump over - * return address. - */ - movq_cfi rdi, RDI+8-RBP - movq_cfi rsi, RSI+8-RBP - movq_cfi rdx, RDX+8-RBP - movq_cfi rcx, RCX+8-RBP - movq_cfi rax, RAX+8-RBP - movq_cfi r8, R8+8-RBP - movq_cfi r9, R9+8-RBP - movq_cfi r10, R10+8-RBP - movq_cfi r11, R11+8-RBP - - leaq -RBP+8(%rsp),%rdi /* arg1 for handler */ - movq_cfi rbp, 8 /* push %rbp */ - leaq 8(%rsp), %rbp /* mov %rsp, %ebp */ + /* start from rbp in pt_regs and jump over */ + movq_cfi rdi, RDI-RBP + movq_cfi rsi, RSI-RBP + movq_cfi rdx, RDX-RBP + movq_cfi rcx, RCX-RBP + movq_cfi rax, RAX-RBP + movq_cfi r8, R8-RBP + movq_cfi r9, R9-RBP + movq_cfi r10, R10-RBP + movq_cfi r11, R11-RBP + + leaq -RBP(%rsp),%rdi /* arg1 for handler */ + movq_cfi rbp, 0 /* push %rbp */ + movq %rsp, %rbp testl $3, CS(%rdi) je 1f SWAPGS @@ -329,19 +324,14 @@ ENTRY(save_args) */ 1: incl PER_CPU_VAR(irq_count) jne 2f - popq_cfi %rax /* move return address... */ mov PER_CPU_VAR(irq_stack_ptr),%rsp EMPTY_FRAME 0 pushq_cfi %rbp /* backlink for unwinder */ - pushq_cfi %rax /* ... to the new stack */ /* * We entered an interrupt context - irqs are off: */ 2: TRACE_IRQS_OFF - ret - CFI_ENDPROC -END(save_args) - .popsection + .endm ENTRY(save_rest) PARTIAL_FRAME 1 REST_SKIP+8 @@ -791,7 +781,7 @@ END(interrupt) /* reserve pt_regs for scratch regs and rbp */ subq $ORIG_RAX-RBP, %rsp CFI_ADJUST_CFA_OFFSET ORIG_RAX-RBP - call save_args + SAVE_ARGS_IRQ PARTIAL_FRAME 0 call \func .endm -- cgit v1.2.3 From 3b99a3ef55b292180473a221f3d6bc24455f0632 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 1 Jul 2011 02:25:17 +0200 Subject: x86,64: Separate arg1 from rbp handling in SAVE_REGS_IRQ Just for clarity in the code. Have a first block that handles the frame pointer and a separate one that handles pt_regs pointer and its use. Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Thomas Gleixner Cc: H. Peter Anvin Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Jan Beulich --- arch/x86/kernel/entry_64.S | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index b6b2e85454cf..20dc8e6b6d80 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -310,9 +310,10 @@ ENDPROC(native_usergs_sysret64) movq_cfi r10, R10-RBP movq_cfi r11, R11-RBP - leaq -RBP(%rsp),%rdi /* arg1 for handler */ movq_cfi rbp, 0 /* push %rbp */ movq %rsp, %rbp + + leaq -RBP(%rsp),%rdi /* arg1 for handler */ testl $3, CS(%rdi) je 1f SWAPGS -- cgit v1.2.3 From 48ffee7d9e6df51b4957bed64115b7beed671374 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 2 Jul 2011 15:03:44 +0200 Subject: x86: Remove useless unwinder backlink from irq regs saving The unwinder backlink in interrupt entry is very useless. It's actually not part of the stack frame chain and thus is never used. Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Thomas Gleixner Cc: H. Peter Anvin Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Jan Beulich --- arch/x86/kernel/entry_64.S | 1 - 1 file changed, 1 deletion(-) (limited to 'arch') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 20dc8e6b6d80..6131432c5afa 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -327,7 +327,6 @@ ENDPROC(native_usergs_sysret64) jne 2f mov PER_CPU_VAR(irq_stack_ptr),%rsp EMPTY_FRAME 0 - pushq_cfi %rbp /* backlink for unwinder */ /* * We entered an interrupt context - irqs are off: */ -- cgit v1.2.3 From a2bbe75089d5eb9a3a46d50dd5c215e213790288 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 2 Jul 2011 16:52:45 +0200 Subject: x86: Don't use frame pointer to save old stack on irq entry rbp is used in SAVE_ARGS_IRQ to save the old stack pointer in order to restore it later in ret_from_intr. It is convenient because we save its value in the irq regs and it's easily restored using the leave instruction. However this is a kind of abuse of the frame pointer which role is to help unwinding the kernel by chaining frames together, each node following the return address to the previous frame. But although we are breaking the frame by changing the stack pointer, there is no preceding return address before the new frame. Hence using the frame pointer to link the two stacks breaks the stack unwinders that find a random value instead of a return address here. There is no workaround that can work in every case. We are using the fixup_bp_irq_link() function to dereference that abused frame pointer in the case of non nesting interrupt (which means stack changed). But that doesn't fix the case of interrupts that don't change the stack (but we still have the unconditional frame link), which is the case of hardirq interrupting softirq. We have no way to detect this transition so the frame irq link is considered as a real frame pointer and the return address is dereferenced but it is still a spurious one. There are two possible results of this: either the spurious return address, a random stack value, luckily belongs to the kernel text and then the unwinding can continue and we just have a weird entry in the stack trace. Or it doesn't belong to the kernel text and unwinding stops there. This is the reason why stacktraces (including perf callchains) on irqs that interrupted softirqs don't work very well. To solve this, we don't save the old stack pointer on rbp anymore but we save it to a scratch register that we push on the new stack and that we pop back later on irq return. This preserves the whole frame chain without spurious return addresses in the middle and drops the need for the horrid fixup_bp_irq_link() workaround. And finally irqs that interrupt softirq are sanely unwinded. Before: 99.81% perf [kernel.kallsyms] [k] perf_pending_event | --- perf_pending_event irq_work_run smp_irq_work_interrupt irq_work_interrupt | |--41.60%-- __read | | | |--99.90%-- create_worker | | bench_sched_messaging | | cmd_bench | | run_builtin | | main | | __libc_start_main | --0.10%-- [...] After: 1.64% swapper [kernel.kallsyms] [k] perf_pending_event | --- perf_pending_event irq_work_run smp_irq_work_interrupt irq_work_interrupt | |--95.00%-- arch_irq_work_raise | irq_work_queue | __perf_event_overflow | perf_swevent_overflow | perf_swevent_event | perf_tp_event | perf_trace_softirq | __do_softirq | call_softirq | do_softirq | irq_exit | | | |--73.68%-- smp_apic_timer_interrupt | | apic_timer_interrupt | | | | | |--96.43%-- amd_e400_idle | | | cpu_idle | | | start_secondary Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Thomas Gleixner Cc: H. Peter Anvin Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Jan Beulich --- arch/x86/kernel/dumpstack_64.c | 30 ------------------------------ arch/x86/kernel/entry_64.S | 27 +++++++++++++++------------ 2 files changed, 15 insertions(+), 42 deletions(-) (limited to 'arch') diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index 788295cbe4a7..19853ad8afc5 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -104,34 +104,6 @@ in_irq_stack(unsigned long *stack, unsigned long *irq_stack, return (stack >= irq_stack && stack < irq_stack_end); } -/* - * We are returning from the irq stack and go to the previous one. - * If the previous stack is also in the irq stack, then bp in the first - * frame of the irq stack points to the previous, interrupted one. - * Otherwise we have another level of indirection: We first save - * the bp of the previous stack, then we switch the stack to the irq one - * and save a new bp that links to the previous one. - * (See save_args()) - */ -static inline unsigned long -fixup_bp_irq_link(unsigned long bp, unsigned long *stack, - unsigned long *irq_stack, unsigned long *irq_stack_end) -{ -#ifdef CONFIG_FRAME_POINTER - struct stack_frame *frame = (struct stack_frame *)bp; - unsigned long next; - - if (!in_irq_stack(stack, irq_stack, irq_stack_end)) { - if (!probe_kernel_address(&frame->next_frame, next)) - return next; - else - WARN_ONCE(1, "Perf: bad frame pointer = %p in " - "callchain\n", &frame->next_frame); - } -#endif - return bp; -} - /* * x86-64 can have up to three kernel stacks: * process stack @@ -208,8 +180,6 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, * pointer (index -1 to end) in the IRQ stack: */ stack = (unsigned long *) (irq_stack_end[-1]); - bp = fixup_bp_irq_link(bp, stack, irq_stack, - irq_stack_end); irq_stack_end = NULL; ops->stack(data, "EOI"); continue; diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 6131432c5afa..d656f68371a4 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -310,8 +310,11 @@ ENDPROC(native_usergs_sysret64) movq_cfi r10, R10-RBP movq_cfi r11, R11-RBP - movq_cfi rbp, 0 /* push %rbp */ - movq %rsp, %rbp + /* Save rbp so that we can unwind from get_irq_regs() */ + movq_cfi rbp, 0 + + /* Save previous stack value */ + movq %rsp, %rsi leaq -RBP(%rsp),%rdi /* arg1 for handler */ testl $3, CS(%rdi) @@ -327,10 +330,11 @@ ENDPROC(native_usergs_sysret64) jne 2f mov PER_CPU_VAR(irq_stack_ptr),%rsp EMPTY_FRAME 0 - /* - * We entered an interrupt context - irqs are off: - */ -2: TRACE_IRQS_OFF + +2: /* Store previous stack value */ + pushq %rsi + /* We entered an interrupt context - irqs are off: */ + TRACE_IRQS_OFF .endm ENTRY(save_rest) @@ -804,15 +808,14 @@ ret_from_intr: DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF decl PER_CPU_VAR(irq_count) - leaveq - CFI_RESTORE rbp + /* Restore saved previous stack */ + popq %rsi + leaq 16(%rsi), %rsp + CFI_DEF_CFA_REGISTER rsp - CFI_ADJUST_CFA_OFFSET -8 + CFI_ADJUST_CFA_OFFSET -16 - /* we did not save rbx, restore only from ARGOFFSET */ - addq $8, %rsp - CFI_ADJUST_CFA_OFFSET -8 exit_intr: GET_THREAD_INFO(%rcx) testl $3,CS-ARGOFFSET(%rsp) -- cgit v1.2.3 From e08fbb78f03fe2c4f88824faf6f51ce6af185e11 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 1 Jul 2011 23:04:36 -0400 Subject: tracing, x86/irq: Do not trace arch_local_{*,irq_*}() functions I triggered a triple fault with gcc 4.5.1 because it did not honor the inline annotation to arch_local_save_flags() function and that function was added to the pool of functions traced by the function tracer. When preempt_schedule() called arch_local_save_flags() (called by irqs_disabled()), it was traced, but the first thing the function tracer does is disable preemption. When it enables preemption, the NEED_RESCHED flag will not have been cleared and the preemption check will trigger the call to preempt_schedule() again. Although the dynamic function tracer crashed immediately, the static version of the function tracer (CONFIG_DYNAMIC_FTRACE is not set) actually was able to show where the problem was. swapper-1 3.N.. 103885us : arch_local_save_flags <-preempt_schedule swapper-1 3.N.. 103886us : arch_local_save_flags <-preempt_schedule swapper-1 3.N.. 103886us : arch_local_save_flags <-preempt_schedule swapper-1 3.N.. 103887us : arch_local_save_flags <-preempt_schedule swapper-1 3.N.. 103887us : arch_local_save_flags <-preempt_schedule swapper-1 3.N.. 103888us : arch_local_save_flags <-preempt_schedule swapper-1 3.N.. 103888us : arch_local_save_flags <-preempt_schedule It went on for a while before it triple faulted with a corrupted stack. The arch_local_save_flags and arch_local_irq_* functions should not be traced. Even though they are marked as inline, gcc may still make them a function and enable tracing of them. The simple solution is to just mark them as notrace. I had to add the for this file to include the notrace tag. Signed-off-by: Steven Rostedt Link: http://lkml.kernel.org/r/20110702033852.733414762@goodmis.org Signed-off-by: Ingo Molnar --- arch/x86/include/asm/irqflags.h | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) (limited to 'arch') diff --git a/arch/x86/include/asm/irqflags.h b/arch/x86/include/asm/irqflags.h index 5745ce8bf108..bba3cf88e624 100644 --- a/arch/x86/include/asm/irqflags.h +++ b/arch/x86/include/asm/irqflags.h @@ -60,23 +60,24 @@ static inline void native_halt(void) #include #else #ifndef __ASSEMBLY__ +#include -static inline unsigned long arch_local_save_flags(void) +static inline notrace unsigned long arch_local_save_flags(void) { return native_save_fl(); } -static inline void arch_local_irq_restore(unsigned long flags) +static inline notrace void arch_local_irq_restore(unsigned long flags) { native_restore_fl(flags); } -static inline void arch_local_irq_disable(void) +static inline notrace void arch_local_irq_disable(void) { native_irq_disable(); } -static inline void arch_local_irq_enable(void) +static inline notrace void arch_local_irq_enable(void) { native_irq_enable(); } @@ -102,7 +103,7 @@ static inline void halt(void) /* * For spinlocks, etc: */ -static inline unsigned long arch_local_irq_save(void) +static inline notrace unsigned long arch_local_irq_save(void) { unsigned long flags = arch_local_save_flags(); arch_local_irq_disable(); -- cgit v1.2.3 From f91298709790b9a483752ca3c967845537df2af3 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Sat, 9 Jul 2011 00:17:12 +0400 Subject: perf, x86: P4 PMU - Introduce event alias feature Instead of hw_nmi_watchdog_set_attr() weak function and appropriate x86_pmu::hw_watchdog_set_attr() call we introduce even alias mechanism which allow us to drop this routines completely and isolate quirks of Netburst architecture inside P4 PMU code only. The main idea remains the same though -- to allow nmi-watchdog and perf top run simultaneously. Note the aliasing mechanism applies to generic PERF_COUNT_HW_CPU_CYCLES event only because arbitrary event (say passed as RAW initially) might have some additional bits set inside ESCR register changing the behaviour of event and we can't guarantee anymore that alias event will give the same result. P.S. Thanks a huge to Don and Steven for for testing and early review. Acked-by: Don Zickus Tested-by: Steven Rostedt Signed-off-by: Cyrill Gorcunov CC: Ingo Molnar CC: Peter Zijlstra CC: Stephane Eranian CC: Lin Ming CC: Arnaldo Carvalho de Melo CC: Frederic Weisbecker Link: http://lkml.kernel.org/r/20110708201712.GS23657@sun Signed-off-by: Steven Rostedt --- arch/x86/include/asm/perf_event_p4.h | 33 +++++++++ arch/x86/kernel/cpu/perf_event.c | 7 -- arch/x86/kernel/cpu/perf_event_p4.c | 135 +++++++++++++++++++++++++++-------- kernel/watchdog.c | 2 - 4 files changed, 139 insertions(+), 38 deletions(-) (limited to 'arch') diff --git a/arch/x86/include/asm/perf_event_p4.h b/arch/x86/include/asm/perf_event_p4.h index 56fd9e3abbda..4d86c86178f2 100644 --- a/arch/x86/include/asm/perf_event_p4.h +++ b/arch/x86/include/asm/perf_event_p4.h @@ -101,6 +101,14 @@ #define P4_CONFIG_HT_SHIFT 63 #define P4_CONFIG_HT (1ULL << P4_CONFIG_HT_SHIFT) +/* + * If an event has alias it should be marked + * with a special bit. (Don't forget to check + * P4_PEBS_CONFIG_MASK and related bits on + * modification.) + */ +#define P4_CONFIG_ALIASABLE (1 << 9) + /* * The bits we allow to pass for RAW events */ @@ -123,6 +131,31 @@ (p4_config_pack_escr(P4_CONFIG_MASK_ESCR)) | \ (p4_config_pack_cccr(P4_CONFIG_MASK_CCCR)) +/* + * In case of event aliasing we need to preserve some + * caller bits otherwise the mapping won't be complete. + */ +#define P4_CONFIG_EVENT_ALIAS_MASK \ + (p4_config_pack_escr(P4_CONFIG_MASK_ESCR) | \ + p4_config_pack_cccr(P4_CCCR_EDGE | \ + P4_CCCR_THRESHOLD_MASK | \ + P4_CCCR_COMPLEMENT | \ + P4_CCCR_COMPARE)) + +#define P4_CONFIG_EVENT_ALIAS_IMMUTABLE_BITS \ + ((P4_CONFIG_HT) | \ + p4_config_pack_escr(P4_ESCR_T0_OS | \ + P4_ESCR_T0_USR | \ + P4_ESCR_T1_OS | \ + P4_ESCR_T1_USR) | \ + p4_config_pack_cccr(P4_CCCR_OVF | \ + P4_CCCR_CASCADE | \ + P4_CCCR_FORCE_OVF | \ + P4_CCCR_THREAD_ANY | \ + P4_CCCR_OVF_PMI_T0 | \ + P4_CCCR_OVF_PMI_T1 | \ + P4_CONFIG_ALIASABLE)) + static inline bool p4_is_event_cascaded(u64 config) { u32 cccr = p4_config_unpack_cccr(config); diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index c53d433c3dde..b7a010fce8c3 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -274,7 +274,6 @@ struct x86_pmu { void (*enable_all)(int added); void (*enable)(struct perf_event *); void (*disable)(struct perf_event *); - void (*hw_watchdog_set_attr)(struct perf_event_attr *attr); int (*hw_config)(struct perf_event *event); int (*schedule_events)(struct cpu_hw_events *cpuc, int n, int *assign); unsigned eventsel; @@ -360,12 +359,6 @@ static u64 __read_mostly hw_cache_extra_regs [PERF_COUNT_HW_CACHE_OP_MAX] [PERF_COUNT_HW_CACHE_RESULT_MAX]; -void hw_nmi_watchdog_set_attr(struct perf_event_attr *wd_attr) -{ - if (x86_pmu.hw_watchdog_set_attr) - x86_pmu.hw_watchdog_set_attr(wd_attr); -} - /* * Propagate event elapsed time into the generic event. * Can only be executed on the CPU where the event is active. diff --git a/arch/x86/kernel/cpu/perf_event_p4.c b/arch/x86/kernel/cpu/perf_event_p4.c index fb901c5080f7..8b7a0c306784 100644 --- a/arch/x86/kernel/cpu/perf_event_p4.c +++ b/arch/x86/kernel/cpu/perf_event_p4.c @@ -570,11 +570,92 @@ static __initconst const u64 p4_hw_cache_event_ids }, }; +/* + * Because of Netburst being quite restricted in now + * many same events can run simultaneously, we use + * event aliases, ie different events which have the + * same functionallity but use non-intersected resources + * (ESCR/CCCR/couter registers). This allow us to run + * two or more semi-same events together. It is done + * transparently to a user space. + * + * Never set any cusom internal bits such as P4_CONFIG_HT, + * P4_CONFIG_ALIASABLE or bits for P4_PEBS_METRIC, they are + * either up-to-dated automatically either not appliable + * at all. + * + * And be really carefull choosing aliases! + */ +struct p4_event_alias { + u64 orig; + u64 alter; +} p4_event_aliases[] = { + { + /* + * Non-halted cycles can be substituted with + * non-sleeping cycles (see Intel SDM Vol3b for + * details). + */ + .orig = + p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_GLOBAL_POWER_EVENTS) | + P4_ESCR_EMASK_BIT(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING)), + .alter = + p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_EXECUTION_EVENT) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS0)| + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS1)| + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS2)| + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS3)| + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS0) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS1) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS2) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS3))| + p4_config_pack_cccr(P4_CCCR_THRESHOLD(15) | P4_CCCR_COMPLEMENT | + P4_CCCR_COMPARE), + }, +}; + +static u64 p4_get_alias_event(u64 config) +{ + u64 config_match; + int i; + + /* + * Probably we're lucky and don't have to do + * matching over all config bits. + */ + if (!(config & P4_CONFIG_ALIASABLE)) + return 0; + + config_match = config & P4_CONFIG_EVENT_ALIAS_MASK; + + /* + * If an event was previously swapped to the alter config + * we should swap it back otherwise contnention on registers + * will return back. + */ + for (i = 0; i < ARRAY_SIZE(p4_event_aliases); i++) { + if (config_match == p4_event_aliases[i].orig) { + config_match = p4_event_aliases[i].alter; + break; + } else if (config_match == p4_event_aliases[i].alter) { + config_match = p4_event_aliases[i].orig; + break; + } + } + + if (i >= ARRAY_SIZE(p4_event_aliases)) + return 0; + + return config_match | + (config & P4_CONFIG_EVENT_ALIAS_IMMUTABLE_BITS); +} + static u64 p4_general_events[PERF_COUNT_HW_MAX] = { /* non-halted CPU clocks */ [PERF_COUNT_HW_CPU_CYCLES] = p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_GLOBAL_POWER_EVENTS) | - P4_ESCR_EMASK_BIT(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING)), + P4_ESCR_EMASK_BIT(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING)) | + P4_CONFIG_ALIASABLE, /* * retired instructions @@ -719,31 +800,6 @@ static int p4_validate_raw_event(struct perf_event *event) return 0; } -static void p4_hw_watchdog_set_attr(struct perf_event_attr *wd_attr) -{ - /* - * Watchdog ticks are special on Netburst, we use - * that named "non-sleeping" ticks as recommended - * by Intel SDM Vol3b. - */ - WARN_ON_ONCE(wd_attr->type != PERF_TYPE_HARDWARE || - wd_attr->config != PERF_COUNT_HW_CPU_CYCLES); - - wd_attr->type = PERF_TYPE_RAW; - wd_attr->config = - p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_EXECUTION_EVENT) | - P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS0) | - P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS1) | - P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS2) | - P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS3) | - P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS0) | - P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS1) | - P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS2) | - P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS3)) | - p4_config_pack_cccr(P4_CCCR_THRESHOLD(15) | P4_CCCR_COMPLEMENT | - P4_CCCR_COMPARE); -} - static int p4_hw_config(struct perf_event *event) { int cpu = get_cpu(); @@ -1159,6 +1215,8 @@ static int p4_pmu_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign struct p4_event_bind *bind; unsigned int i, thread, num; int cntr_idx, escr_idx; + u64 config_alias; + int pass; bitmap_zero(used_mask, X86_PMC_IDX_MAX); bitmap_zero(escr_mask, P4_ESCR_MSR_TABLE_SIZE); @@ -1167,6 +1225,17 @@ static int p4_pmu_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign hwc = &cpuc->event_list[i]->hw; thread = p4_ht_thread(cpu); + pass = 0; + +again: + /* + * Aliases are swappable so we may hit circular + * lock if both original config and alias need + * resources (MSR registers) which already busy. + */ + if (pass > 2) + goto done; + bind = p4_config_get_bind(hwc->config); escr_idx = p4_get_escr_idx(bind->escr_msr[thread]); if (unlikely(escr_idx == -1)) @@ -1180,8 +1249,17 @@ static int p4_pmu_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign } cntr_idx = p4_next_cntr(thread, used_mask, bind); - if (cntr_idx == -1 || test_bit(escr_idx, escr_mask)) - goto done; + if (cntr_idx == -1 || test_bit(escr_idx, escr_mask)) { + /* + * Probably an event alias is still available. + */ + config_alias = p4_get_alias_event(hwc->config); + if (!config_alias) + goto done; + hwc->config = config_alias; + pass++; + goto again; + } p4_pmu_swap_config_ts(hwc, cpu); if (assign) @@ -1218,7 +1296,6 @@ static __initconst const struct x86_pmu p4_pmu = { .cntval_bits = ARCH_P4_CNTRVAL_BITS, .cntval_mask = ARCH_P4_CNTRVAL_MASK, .max_period = (1ULL << (ARCH_P4_CNTRVAL_BITS - 1)) - 1, - .hw_watchdog_set_attr = p4_hw_watchdog_set_attr, .hw_config = p4_hw_config, .schedule_events = p4_pmu_schedule_events, /* diff --git a/kernel/watchdog.c b/kernel/watchdog.c index a933e3a0398b..36491cd5b7d4 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -200,7 +200,6 @@ static int is_softlockup(unsigned long touch_ts) } #ifdef CONFIG_HARDLOCKUP_DETECTOR -void __weak hw_nmi_watchdog_set_attr(struct perf_event_attr *wd_attr) { } static struct perf_event_attr wd_hw_attr = { .type = PERF_TYPE_HARDWARE, @@ -372,7 +371,6 @@ static int watchdog_nmi_enable(int cpu) wd_attr = &wd_hw_attr; wd_attr->sample_period = hw_nmi_get_sample_period(watchdog_thresh); - hw_nmi_watchdog_set_attr(wd_attr); /* Try to register using hardware perf events */ event = perf_event_create_kernel_counter(wd_attr, cpu, NULL, watchdog_overflow_callback, NULL); -- cgit v1.2.3 From f53173e47dee5f7514d264796bec58d43ed0f67f Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Thu, 21 Jul 2011 20:06:25 +0400 Subject: x86, perf: P4 PMU - Fix typos in comments and style cleanup This patch: - fixes typos in comments and clarifies the text - renames obscure p4_event_alias::original and ::alter members to ::original and ::alternative as appropriate - drops parenthesis from the return of p4_get_alias_event() No functional changes. Reported-by: Ingo Molnar Signed-off-by: Cyrill Gorcunov Link: http://lkml.kernel.org/r/20110721160625.GX7492@sun Signed-off-by: Ingo Molnar --- arch/x86/include/asm/perf_event_p4.h | 2 +- arch/x86/kernel/cpu/perf_event_p4.c | 66 ++++++++++++++++-------------------- 2 files changed, 31 insertions(+), 37 deletions(-) (limited to 'arch') diff --git a/arch/x86/include/asm/perf_event_p4.h b/arch/x86/include/asm/perf_event_p4.h index 4d86c86178f2..4f7e67e2345e 100644 --- a/arch/x86/include/asm/perf_event_p4.h +++ b/arch/x86/include/asm/perf_event_p4.h @@ -133,7 +133,7 @@ /* * In case of event aliasing we need to preserve some - * caller bits otherwise the mapping won't be complete. + * caller bits, otherwise the mapping won't be complete. */ #define P4_CONFIG_EVENT_ALIAS_MASK \ (p4_config_pack_escr(P4_CONFIG_MASK_ESCR) | \ diff --git a/arch/x86/kernel/cpu/perf_event_p4.c b/arch/x86/kernel/cpu/perf_event_p4.c index 8b7a0c306784..7809d2bcb209 100644 --- a/arch/x86/kernel/cpu/perf_event_p4.c +++ b/arch/x86/kernel/cpu/perf_event_p4.c @@ -571,35 +571,34 @@ static __initconst const u64 p4_hw_cache_event_ids }; /* - * Because of Netburst being quite restricted in now - * many same events can run simultaneously, we use - * event aliases, ie different events which have the - * same functionallity but use non-intersected resources - * (ESCR/CCCR/couter registers). This allow us to run - * two or more semi-same events together. It is done - * transparently to a user space. + * Because of Netburst being quite restricted in how many + * identical events may run simultaneously, we introduce event aliases, + * ie the different events which have the same functionality but + * utilize non-intersected resources (ESCR/CCCR/counter registers). * - * Never set any cusom internal bits such as P4_CONFIG_HT, - * P4_CONFIG_ALIASABLE or bits for P4_PEBS_METRIC, they are - * either up-to-dated automatically either not appliable - * at all. + * This allow us to relax restrictions a bit and run two or more + * identical events together. * - * And be really carefull choosing aliases! + * Never set any custom internal bits such as P4_CONFIG_HT, + * P4_CONFIG_ALIASABLE or bits for P4_PEBS_METRIC, they are + * either up to date automatically or not applicable at all. */ struct p4_event_alias { - u64 orig; - u64 alter; + u64 original; + u64 alternative; } p4_event_aliases[] = { { /* - * Non-halted cycles can be substituted with - * non-sleeping cycles (see Intel SDM Vol3b for - * details). + * Non-halted cycles can be substituted with non-sleeping cycles (see + * Intel SDM Vol3b for details). We need this alias to be able + * to run nmi-watchdog and 'perf top' (or any other user space tool + * which is interested in running PERF_COUNT_HW_CPU_CYCLES) + * simultaneously. */ - .orig = + .original = p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_GLOBAL_POWER_EVENTS) | P4_ESCR_EMASK_BIT(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING)), - .alter = + .alternative = p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_EXECUTION_EVENT) | P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS0)| P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS1)| @@ -620,25 +619,21 @@ static u64 p4_get_alias_event(u64 config) int i; /* - * Probably we're lucky and don't have to do - * matching over all config bits. + * Only event with special mark is allowed, + * we're to be sure it didn't come as malformed + * RAW event. */ if (!(config & P4_CONFIG_ALIASABLE)) return 0; config_match = config & P4_CONFIG_EVENT_ALIAS_MASK; - /* - * If an event was previously swapped to the alter config - * we should swap it back otherwise contnention on registers - * will return back. - */ for (i = 0; i < ARRAY_SIZE(p4_event_aliases); i++) { - if (config_match == p4_event_aliases[i].orig) { - config_match = p4_event_aliases[i].alter; + if (config_match == p4_event_aliases[i].original) { + config_match = p4_event_aliases[i].alternative; break; - } else if (config_match == p4_event_aliases[i].alter) { - config_match = p4_event_aliases[i].orig; + } else if (config_match == p4_event_aliases[i].alternative) { + config_match = p4_event_aliases[i].original; break; } } @@ -646,8 +641,7 @@ static u64 p4_get_alias_event(u64 config) if (i >= ARRAY_SIZE(p4_event_aliases)) return 0; - return config_match | - (config & P4_CONFIG_EVENT_ALIAS_IMMUTABLE_BITS); + return config_match | (config & P4_CONFIG_EVENT_ALIAS_IMMUTABLE_BITS); } static u64 p4_general_events[PERF_COUNT_HW_MAX] = { @@ -1229,9 +1223,9 @@ static int p4_pmu_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign again: /* - * Aliases are swappable so we may hit circular - * lock if both original config and alias need - * resources (MSR registers) which already busy. + * It's possible to hit a circular lock + * between original and alternative events + * if both are scheduled already. */ if (pass > 2) goto done; @@ -1251,7 +1245,7 @@ again: cntr_idx = p4_next_cntr(thread, used_mask, bind); if (cntr_idx == -1 || test_bit(escr_idx, escr_mask)) { /* - * Probably an event alias is still available. + * Check whether an event alias is still available. */ config_alias = p4_get_alias_event(hwc->config); if (!config_alias) -- cgit v1.2.3 From 1ac2e6ca44e13a087eb7438d284f887097ee7e84 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Tue, 7 Jun 2011 11:49:55 +0200 Subject: x86, perf: Make copy_from_user_nmi() a library function copy_from_user_nmi() is used in oprofile and perf. Moving it to other library functions like copy_from_user(). As this is x86 code for 32 and 64 bits, create a new file usercopy.c for unified code. Signed-off-by: Robert Richter Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20110607172413.GJ20052@erda.amd.com Signed-off-by: Ingo Molnar --- arch/x86/include/asm/uaccess.h | 3 +++ arch/x86/kernel/cpu/perf_event.c | 35 -------------------------------- arch/x86/lib/Makefile | 2 +- arch/x86/lib/usercopy.c | 43 ++++++++++++++++++++++++++++++++++++++++ arch/x86/oprofile/backtrace.c | 39 +----------------------------------- 5 files changed, 48 insertions(+), 74 deletions(-) create mode 100644 arch/x86/lib/usercopy.c (limited to 'arch') diff --git a/arch/x86/include/asm/uaccess.h b/arch/x86/include/asm/uaccess.h index 99ddd148a760..36361bf6fdd1 100644 --- a/arch/x86/include/asm/uaccess.h +++ b/arch/x86/include/asm/uaccess.h @@ -555,6 +555,9 @@ struct __large_struct { unsigned long buf[100]; }; #endif /* CONFIG_X86_WP_WORKS_OK */ +extern unsigned long +copy_from_user_nmi(void *to, const void __user *from, unsigned long n); + /* * movsl can be slow when source and dest are not both 8-byte aligned */ diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index b7a010fce8c3..4ee3abf20ed6 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -22,7 +22,6 @@ #include #include #include -#include #include #include @@ -67,40 +66,6 @@ enum extra_reg_type { EXTRA_REG_MAX /* number of entries needed */ }; -/* - * best effort, GUP based copy_from_user() that assumes IRQ or NMI context - */ -static unsigned long -copy_from_user_nmi(void *to, const void __user *from, unsigned long n) -{ - unsigned long offset, addr = (unsigned long)from; - unsigned long size, len = 0; - struct page *page; - void *map; - int ret; - - do { - ret = __get_user_pages_fast(addr, 1, 0, &page); - if (!ret) - break; - - offset = addr & (PAGE_SIZE - 1); - size = min(PAGE_SIZE - offset, n - len); - - map = kmap_atomic(page); - memcpy(to, map+offset, size); - kunmap_atomic(map); - put_page(page); - - len += size; - to += size; - addr += size; - - } while (len < n); - - return len; -} - struct event_constraint { union { unsigned long idxmsk[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; diff --git a/arch/x86/lib/Makefile b/arch/x86/lib/Makefile index f2479f19ddde..6ba477342b8e 100644 --- a/arch/x86/lib/Makefile +++ b/arch/x86/lib/Makefile @@ -18,7 +18,7 @@ obj-$(CONFIG_SMP) += msr-smp.o cache-smp.o lib-y := delay.o lib-y += thunk_$(BITS).o -lib-y += usercopy_$(BITS).o getuser.o putuser.o +lib-y += usercopy_$(BITS).o usercopy.o getuser.o putuser.o lib-y += memcpy_$(BITS).o lib-$(CONFIG_INSTRUCTION_DECODER) += insn.o inat.o diff --git a/arch/x86/lib/usercopy.c b/arch/x86/lib/usercopy.c new file mode 100644 index 000000000000..97be9cb54483 --- /dev/null +++ b/arch/x86/lib/usercopy.c @@ -0,0 +1,43 @@ +/* + * User address space access functions. + * + * For licencing details see kernel-base/COPYING + */ + +#include +#include + +/* + * best effort, GUP based copy_from_user() that is NMI-safe + */ +unsigned long +copy_from_user_nmi(void *to, const void __user *from, unsigned long n) +{ + unsigned long offset, addr = (unsigned long)from; + unsigned long size, len = 0; + struct page *page; + void *map; + int ret; + + do { + ret = __get_user_pages_fast(addr, 1, 0, &page); + if (!ret) + break; + + offset = addr & (PAGE_SIZE - 1); + size = min(PAGE_SIZE - offset, n - len); + + map = kmap_atomic(page); + memcpy(to, map+offset, size); + kunmap_atomic(map); + put_page(page); + + len += size; + to += size; + addr += size; + + } while (len < n); + + return len; +} +EXPORT_SYMBOL_GPL(copy_from_user_nmi); diff --git a/arch/x86/oprofile/backtrace.c b/arch/x86/oprofile/backtrace.c index 32f78eb46744..bff89dfe3619 100644 --- a/arch/x86/oprofile/backtrace.c +++ b/arch/x86/oprofile/backtrace.c @@ -12,10 +12,9 @@ #include #include #include -#include +#include #include -#include #include static int backtrace_stack(void *data, char *name) @@ -38,42 +37,6 @@ static struct stacktrace_ops backtrace_ops = { .walk_stack = print_context_stack, }; -/* from arch/x86/kernel/cpu/perf_event.c: */ - -/* - * best effort, GUP based copy_from_user() that assumes IRQ or NMI context - */ -static unsigned long -copy_from_user_nmi(void *to, const void __user *from, unsigned long n) -{ - unsigned long offset, addr = (unsigned long)from; - unsigned long size, len = 0; - struct page *page; - void *map; - int ret; - - do { - ret = __get_user_pages_fast(addr, 1, 0, &page); - if (!ret) - break; - - offset = addr & (PAGE_SIZE - 1); - size = min(PAGE_SIZE - offset, n - len); - - map = kmap_atomic(page); - memcpy(to, map+offset, size); - kunmap_atomic(map); - put_page(page); - - len += size; - to += size; - addr += size; - - } while (len < n); - - return len; -} - #ifdef CONFIG_COMPAT static struct stack_frame_ia32 * dump_user_backtrace_32(struct stack_frame_ia32 *head) -- cgit v1.2.3