From dcfce4a095932e6e95d83ad982be3609947963bc Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Tue, 11 Oct 2011 17:11:08 +0200 Subject: oprofile, x86: Reimplement nmi timer mode using perf event The legacy x86 nmi watchdog code was removed with the implementation of the perf based nmi watchdog. This broke Oprofile's nmi timer mode. To run nmi timer mode we relied on a continuous ticking nmi source which the nmi watchdog provided. The nmi tick was no longer available and current watchdog can not be used anymore since it runs with very long periods in the range of seconds. This patch reimplements the nmi timer mode using a perf counter nmi source. V2: * removing pr_info() * fix undefined reference to `__udivdi3' for 32 bit build * fix section mismatch of .cpuinit.data:nmi_timer_cpu_nb * removed nmi timer setup in arch/x86 * implemented function stubs for op_nmi_init/exit() * made code more readable in oprofile_init() V3: * fix architectural initialization in oprofile_init() * fix CONFIG_OPROFILE_NMI_TIMER dependencies Acked-by: Peter Zijlstra Signed-off-by: Robert Richter --- kernel/events/core.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel/events') diff --git a/kernel/events/core.c b/kernel/events/core.c index d1a1bee3522..d2e28bdd523 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -1322,6 +1322,7 @@ retry: } raw_spin_unlock_irq(&ctx->lock); } +EXPORT_SYMBOL_GPL(perf_event_disable); static void perf_set_shadow_time(struct perf_event *event, struct perf_event_context *ctx, @@ -1806,6 +1807,7 @@ retry: out: raw_spin_unlock_irq(&ctx->lock); } +EXPORT_SYMBOL_GPL(perf_event_enable); int perf_event_refresh(struct perf_event *event, int refresh) { -- cgit v1.2.3-70-g09d2 From 9251f904f95175b4a1d8cbc0449e748f9edd7629 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Sun, 16 Oct 2011 17:15:04 +0200 Subject: perf: Carve out callchain functionality Split the callchain code from the perf events core into a new kernel/events/callchain.c file. This simplifies a bit the big core.c Signed-off-by: Borislav Petkov Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian [keep ctx recursion handling inline and use internal headers] Signed-off-by: Frederic Weisbecker Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1318778104-17152-1-git-send-email-fweisbec@gmail.com Signed-off-by: Ingo Molnar --- kernel/events/Makefile | 2 +- kernel/events/callchain.c | 191 ++++++++++++++++++++++++++++++++++++++++++ kernel/events/core.c | 209 ---------------------------------------------- kernel/events/internal.h | 39 ++++++++- 4 files changed, 230 insertions(+), 211 deletions(-) create mode 100644 kernel/events/callchain.c (limited to 'kernel/events') diff --git a/kernel/events/Makefile b/kernel/events/Makefile index 89e5e8aa4c3..22d901f9caf 100644 --- a/kernel/events/Makefile +++ b/kernel/events/Makefile @@ -2,5 +2,5 @@ ifdef CONFIG_FUNCTION_TRACER CFLAGS_REMOVE_core.o = -pg endif -obj-y := core.o ring_buffer.o +obj-y := core.o ring_buffer.o callchain.o obj-$(CONFIG_HAVE_HW_BREAKPOINT) += hw_breakpoint.o diff --git a/kernel/events/callchain.c b/kernel/events/callchain.c new file mode 100644 index 00000000000..057e24b665c --- /dev/null +++ b/kernel/events/callchain.c @@ -0,0 +1,191 @@ +/* + * Performance events callchain code, extracted from core.c: + * + * Copyright (C) 2008 Thomas Gleixner + * Copyright (C) 2008-2011 Red Hat, Inc., Ingo Molnar + * Copyright (C) 2008-2011 Red Hat, Inc., Peter Zijlstra + * Copyright © 2009 Paul Mackerras, IBM Corp. + * + * For licensing details see kernel-base/COPYING + */ + +#include +#include +#include "internal.h" + +struct callchain_cpus_entries { + struct rcu_head rcu_head; + struct perf_callchain_entry *cpu_entries[0]; +}; + +static DEFINE_PER_CPU(int, callchain_recursion[PERF_NR_CONTEXTS]); +static atomic_t nr_callchain_events; +static DEFINE_MUTEX(callchain_mutex); +static struct callchain_cpus_entries *callchain_cpus_entries; + + +__weak void perf_callchain_kernel(struct perf_callchain_entry *entry, + struct pt_regs *regs) +{ +} + +__weak void perf_callchain_user(struct perf_callchain_entry *entry, + struct pt_regs *regs) +{ +} + +static void release_callchain_buffers_rcu(struct rcu_head *head) +{ + struct callchain_cpus_entries *entries; + int cpu; + + entries = container_of(head, struct callchain_cpus_entries, rcu_head); + + for_each_possible_cpu(cpu) + kfree(entries->cpu_entries[cpu]); + + kfree(entries); +} + +static void release_callchain_buffers(void) +{ + struct callchain_cpus_entries *entries; + + entries = callchain_cpus_entries; + rcu_assign_pointer(callchain_cpus_entries, NULL); + call_rcu(&entries->rcu_head, release_callchain_buffers_rcu); +} + +static int alloc_callchain_buffers(void) +{ + int cpu; + int size; + struct callchain_cpus_entries *entries; + + /* + * We can't use the percpu allocation API for data that can be + * accessed from NMI. Use a temporary manual per cpu allocation + * until that gets sorted out. + */ + size = offsetof(struct callchain_cpus_entries, cpu_entries[nr_cpu_ids]); + + entries = kzalloc(size, GFP_KERNEL); + if (!entries) + return -ENOMEM; + + size = sizeof(struct perf_callchain_entry) * PERF_NR_CONTEXTS; + + for_each_possible_cpu(cpu) { + entries->cpu_entries[cpu] = kmalloc_node(size, GFP_KERNEL, + cpu_to_node(cpu)); + if (!entries->cpu_entries[cpu]) + goto fail; + } + + rcu_assign_pointer(callchain_cpus_entries, entries); + + return 0; + +fail: + for_each_possible_cpu(cpu) + kfree(entries->cpu_entries[cpu]); + kfree(entries); + + return -ENOMEM; +} + +int get_callchain_buffers(void) +{ + int err = 0; + int count; + + mutex_lock(&callchain_mutex); + + count = atomic_inc_return(&nr_callchain_events); + if (WARN_ON_ONCE(count < 1)) { + err = -EINVAL; + goto exit; + } + + if (count > 1) { + /* If the allocation failed, give up */ + if (!callchain_cpus_entries) + err = -ENOMEM; + goto exit; + } + + err = alloc_callchain_buffers(); + if (err) + release_callchain_buffers(); +exit: + mutex_unlock(&callchain_mutex); + + return err; +} + +void put_callchain_buffers(void) +{ + if (atomic_dec_and_mutex_lock(&nr_callchain_events, &callchain_mutex)) { + release_callchain_buffers(); + mutex_unlock(&callchain_mutex); + } +} + +static struct perf_callchain_entry *get_callchain_entry(int *rctx) +{ + int cpu; + struct callchain_cpus_entries *entries; + + *rctx = get_recursion_context(__get_cpu_var(callchain_recursion)); + if (*rctx == -1) + return NULL; + + entries = rcu_dereference(callchain_cpus_entries); + if (!entries) + return NULL; + + cpu = smp_processor_id(); + + return &entries->cpu_entries[cpu][*rctx]; +} + +static void +put_callchain_entry(int rctx) +{ + put_recursion_context(__get_cpu_var(callchain_recursion), rctx); +} + +struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) +{ + int rctx; + struct perf_callchain_entry *entry; + + + entry = get_callchain_entry(&rctx); + if (rctx == -1) + return NULL; + + if (!entry) + goto exit_put; + + entry->nr = 0; + + if (!user_mode(regs)) { + perf_callchain_store(entry, PERF_CONTEXT_KERNEL); + perf_callchain_kernel(entry, regs); + if (current->mm) + regs = task_pt_regs(current); + else + regs = NULL; + } + + if (regs) { + perf_callchain_store(entry, PERF_CONTEXT_USER); + perf_callchain_user(entry, regs); + } + +exit_put: + put_callchain_entry(rctx); + + return entry; +} diff --git a/kernel/events/core.c b/kernel/events/core.c index 0e8457da6f9..eadac69265f 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -2569,215 +2569,6 @@ static u64 perf_event_read(struct perf_event *event) return perf_event_count(event); } -/* - * Callchain support - */ - -struct callchain_cpus_entries { - struct rcu_head rcu_head; - struct perf_callchain_entry *cpu_entries[0]; -}; - -static DEFINE_PER_CPU(int, callchain_recursion[PERF_NR_CONTEXTS]); -static atomic_t nr_callchain_events; -static DEFINE_MUTEX(callchain_mutex); -struct callchain_cpus_entries *callchain_cpus_entries; - - -__weak void perf_callchain_kernel(struct perf_callchain_entry *entry, - struct pt_regs *regs) -{ -} - -__weak void perf_callchain_user(struct perf_callchain_entry *entry, - struct pt_regs *regs) -{ -} - -static void release_callchain_buffers_rcu(struct rcu_head *head) -{ - struct callchain_cpus_entries *entries; - int cpu; - - entries = container_of(head, struct callchain_cpus_entries, rcu_head); - - for_each_possible_cpu(cpu) - kfree(entries->cpu_entries[cpu]); - - kfree(entries); -} - -static void release_callchain_buffers(void) -{ - struct callchain_cpus_entries *entries; - - entries = callchain_cpus_entries; - rcu_assign_pointer(callchain_cpus_entries, NULL); - call_rcu(&entries->rcu_head, release_callchain_buffers_rcu); -} - -static int alloc_callchain_buffers(void) -{ - int cpu; - int size; - struct callchain_cpus_entries *entries; - - /* - * We can't use the percpu allocation API for data that can be - * accessed from NMI. Use a temporary manual per cpu allocation - * until that gets sorted out. - */ - size = offsetof(struct callchain_cpus_entries, cpu_entries[nr_cpu_ids]); - - entries = kzalloc(size, GFP_KERNEL); - if (!entries) - return -ENOMEM; - - size = sizeof(struct perf_callchain_entry) * PERF_NR_CONTEXTS; - - for_each_possible_cpu(cpu) { - entries->cpu_entries[cpu] = kmalloc_node(size, GFP_KERNEL, - cpu_to_node(cpu)); - if (!entries->cpu_entries[cpu]) - goto fail; - } - - rcu_assign_pointer(callchain_cpus_entries, entries); - - return 0; - -fail: - for_each_possible_cpu(cpu) - kfree(entries->cpu_entries[cpu]); - kfree(entries); - - return -ENOMEM; -} - -static int get_callchain_buffers(void) -{ - int err = 0; - int count; - - mutex_lock(&callchain_mutex); - - count = atomic_inc_return(&nr_callchain_events); - if (WARN_ON_ONCE(count < 1)) { - err = -EINVAL; - goto exit; - } - - if (count > 1) { - /* If the allocation failed, give up */ - if (!callchain_cpus_entries) - err = -ENOMEM; - goto exit; - } - - err = alloc_callchain_buffers(); - if (err) - release_callchain_buffers(); -exit: - mutex_unlock(&callchain_mutex); - - return err; -} - -static void put_callchain_buffers(void) -{ - if (atomic_dec_and_mutex_lock(&nr_callchain_events, &callchain_mutex)) { - release_callchain_buffers(); - mutex_unlock(&callchain_mutex); - } -} - -static int get_recursion_context(int *recursion) -{ - int rctx; - - if (in_nmi()) - rctx = 3; - else if (in_irq()) - rctx = 2; - else if (in_softirq()) - rctx = 1; - else - rctx = 0; - - if (recursion[rctx]) - return -1; - - recursion[rctx]++; - barrier(); - - return rctx; -} - -static inline void put_recursion_context(int *recursion, int rctx) -{ - barrier(); - recursion[rctx]--; -} - -static struct perf_callchain_entry *get_callchain_entry(int *rctx) -{ - int cpu; - struct callchain_cpus_entries *entries; - - *rctx = get_recursion_context(__get_cpu_var(callchain_recursion)); - if (*rctx == -1) - return NULL; - - entries = rcu_dereference(callchain_cpus_entries); - if (!entries) - return NULL; - - cpu = smp_processor_id(); - - return &entries->cpu_entries[cpu][*rctx]; -} - -static void -put_callchain_entry(int rctx) -{ - put_recursion_context(__get_cpu_var(callchain_recursion), rctx); -} - -static struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) -{ - int rctx; - struct perf_callchain_entry *entry; - - - entry = get_callchain_entry(&rctx); - if (rctx == -1) - return NULL; - - if (!entry) - goto exit_put; - - entry->nr = 0; - - if (!user_mode(regs)) { - perf_callchain_store(entry, PERF_CONTEXT_KERNEL); - perf_callchain_kernel(entry, regs); - if (current->mm) - regs = task_pt_regs(current); - else - regs = NULL; - } - - if (regs) { - perf_callchain_store(entry, PERF_CONTEXT_USER); - perf_callchain_user(entry, regs); - } - -exit_put: - put_callchain_entry(rctx); - - return entry; -} - /* * Initialize the perf_event context in a task_struct: */ diff --git a/kernel/events/internal.h b/kernel/events/internal.h index 09097dd8116..be4a43f6de4 100644 --- a/kernel/events/internal.h +++ b/kernel/events/internal.h @@ -1,6 +1,10 @@ #ifndef _KERNEL_EVENTS_INTERNAL_H #define _KERNEL_EVENTS_INTERNAL_H +#include + +/* Buffer handling */ + #define RING_BUFFER_WRITABLE 0x01 struct ring_buffer { @@ -64,7 +68,7 @@ static inline int page_order(struct ring_buffer *rb) } #endif -static unsigned long perf_data_size(struct ring_buffer *rb) +static inline unsigned long perf_data_size(struct ring_buffer *rb) { return rb->nr_pages << (PAGE_SHIFT + page_order(rb)); } @@ -93,4 +97,37 @@ __output_copy(struct perf_output_handle *handle, } while (len); } +/* Callchain handling */ +extern struct perf_callchain_entry *perf_callchain(struct pt_regs *regs); +extern int get_callchain_buffers(void); +extern void put_callchain_buffers(void); + +static inline int get_recursion_context(int *recursion) +{ + int rctx; + + if (in_nmi()) + rctx = 3; + else if (in_irq()) + rctx = 2; + else if (in_softirq()) + rctx = 1; + else + rctx = 0; + + if (recursion[rctx]) + return -1; + + recursion[rctx]++; + barrier(); + + return rctx; +} + +static inline void put_recursion_context(int *recursion, int rctx) +{ + barrier(); + recursion[rctx]--; +} + #endif /* _KERNEL_EVENTS_INTERNAL_H */ -- cgit v1.2.3-70-g09d2 From 5d81e5cfb37a174e8ddc0413e2e70cdf05807ace Mon Sep 17 00:00:00 2001 From: Andrew Vagin Date: Mon, 7 Nov 2011 15:54:12 +0300 Subject: events: Don't divide events if it has field period This patch solves the following problem: Now some samples may be lost due to throttling. The number of samples is restricted by sysctl_perf_event_sample_rate/HZ. A trace event is divided on some samples according to event's period. I don't sure, that we should generate more than one sample on each trace event. I think the better way to use SAMPLE_PERIOD. E.g.: I want to trace when a process sleeps. I created a process, which sleeps for 1ms and for 4ms. perf got 100 events in both cases. swapper 0 [000] 1141.371830: sched_stat_sleep: comm=foo pid=1801 delay=1386750 [ns] swapper 0 [000] 1141.369444: sched_stat_sleep: comm=foo pid=1801 delay=4499585 [ns] In the first case a kernel want to send 4499585 events and in the second case it wants to send 1386750 events. perf-reports shows that process sleeps in both places equal time. It's bug. With this patch kernel generates one event on each "sleep" and the time slice is saved in the field "period". Perf knows how handle it. Signed-off-by: Andrew Vagin Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1320670457-2633428-3-git-send-email-avagin@openvz.org Signed-off-by: Ingo Molnar --- kernel/events/core.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'kernel/events') diff --git a/kernel/events/core.c b/kernel/events/core.c index eadac69265f..8d9dea56c26 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -4528,7 +4528,6 @@ static void perf_swevent_overflow(struct perf_event *event, u64 overflow, struct hw_perf_event *hwc = &event->hw; int throttle = 0; - data->period = event->hw.last_period; if (!overflow) overflow = perf_swevent_set_period(event); @@ -4562,6 +4561,12 @@ static void perf_swevent_event(struct perf_event *event, u64 nr, if (!is_sampling_event(event)) return; + if ((event->attr.sample_type & PERF_SAMPLE_PERIOD) && !event->attr.freq) { + data->period = nr; + return perf_swevent_overflow(event, 1, data, regs); + } else + data->period = event->hw.last_period; + if (nr == 1 && hwc->sample_period == 1 && !event->attr.freq) return perf_swevent_overflow(event, 1, data, regs); -- cgit v1.2.3-70-g09d2 From 0f5a2601284237e2ba089389fd75d67f77626cef Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 16 Nov 2011 14:38:16 +0100 Subject: perf: Avoid a useless pmu_disable() in the perf-tick Gleb writes: > Currently pmu is disabled and re-enabled on each timer interrupt even > when no rotation or frequency adjustment is needed. On Intel CPU this > results in two writes into PERF_GLOBAL_CTRL MSR per tick. On bare metal > it does not cause significant slowdown, but when running perf in a virtual > machine it leads to 20% slowdown on my machine. Cure this by keeping a perf_event_context::nr_freq counter that counts the number of active events that require frequency adjustments and use this in a similar fashion to the already existing nr_events != nr_active test in perf_rotate_context(). By being able to exclude both rotation and frequency adjustments a-priory for the common case we can avoid the otherwise superfluous PMU disable. Suggested-by: Gleb Natapov Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-515yhoatehd3gza7we9fapaa@git.kernel.org Signed-off-by: Ingo Molnar --- include/linux/perf_event.h | 1 + kernel/events/core.c | 48 ++++++++++++++++++++++++++++++---------------- 2 files changed, 33 insertions(+), 16 deletions(-) (limited to 'kernel/events') diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index b1f89122bf6..cb44c9e7566 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -890,6 +890,7 @@ struct perf_event_context { int nr_active; int is_active; int nr_stat; + int nr_freq; int rotate_disable; atomic_t refcount; struct task_struct *task; diff --git a/kernel/events/core.c b/kernel/events/core.c index a355ffb0b28..b3fed52aaf2 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -1130,6 +1130,8 @@ event_sched_out(struct perf_event *event, if (!is_software_event(event)) cpuctx->active_oncpu--; ctx->nr_active--; + if (event->attr.freq && event->attr.sample_freq) + ctx->nr_freq--; if (event->attr.exclusive || !cpuctx->active_oncpu) cpuctx->exclusive = 0; } @@ -1407,6 +1409,8 @@ event_sched_in(struct perf_event *event, if (!is_software_event(event)) cpuctx->active_oncpu++; ctx->nr_active++; + if (event->attr.freq && event->attr.sample_freq) + ctx->nr_freq++; if (event->attr.exclusive) cpuctx->exclusive = 1; @@ -2329,6 +2333,9 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx, u64 period) u64 interrupts, now; s64 delta; + if (!ctx->nr_freq) + return; + list_for_each_entry_rcu(event, &ctx->event_list, event_entry) { if (event->state != PERF_EVENT_STATE_ACTIVE) continue; @@ -2384,12 +2391,14 @@ static void perf_rotate_context(struct perf_cpu_context *cpuctx) { u64 interval = (u64)cpuctx->jiffies_interval * TICK_NSEC; struct perf_event_context *ctx = NULL; - int rotate = 0, remove = 1; + int rotate = 0, remove = 1, freq = 0; if (cpuctx->ctx.nr_events) { remove = 0; if (cpuctx->ctx.nr_events != cpuctx->ctx.nr_active) rotate = 1; + if (cpuctx->ctx.nr_freq) + freq = 1; } ctx = cpuctx->task_ctx; @@ -2397,33 +2406,40 @@ static void perf_rotate_context(struct perf_cpu_context *cpuctx) remove = 0; if (ctx->nr_events != ctx->nr_active) rotate = 1; + if (ctx->nr_freq) + freq = 1; } + if (!rotate && !freq) + goto done; + perf_ctx_lock(cpuctx, cpuctx->task_ctx); perf_pmu_disable(cpuctx->ctx.pmu); - perf_ctx_adjust_freq(&cpuctx->ctx, interval); - if (ctx) - perf_ctx_adjust_freq(ctx, interval); - if (!rotate) - goto done; + if (freq) { + perf_ctx_adjust_freq(&cpuctx->ctx, interval); + if (ctx) + perf_ctx_adjust_freq(ctx, interval); + } - cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE); - if (ctx) - ctx_sched_out(ctx, cpuctx, EVENT_FLEXIBLE); + if (rotate) { + cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE); + if (ctx) + ctx_sched_out(ctx, cpuctx, EVENT_FLEXIBLE); - rotate_ctx(&cpuctx->ctx); - if (ctx) - rotate_ctx(ctx); + rotate_ctx(&cpuctx->ctx); + if (ctx) + rotate_ctx(ctx); + + perf_event_sched_in(cpuctx, ctx, current); + } - perf_event_sched_in(cpuctx, ctx, current); + perf_pmu_enable(cpuctx->ctx.pmu); + perf_ctx_unlock(cpuctx, cpuctx->task_ctx); done: if (remove) list_del_init(&cpuctx->rotation_list); - - perf_pmu_enable(cpuctx->ctx.pmu); - perf_ctx_unlock(cpuctx, cpuctx->task_ctx); } void perf_event_task_tick(void) -- cgit v1.2.3-70-g09d2 From 1d9b482e78d3b16f97f85a82849f82db7eed3102 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 23 Nov 2011 12:34:20 +0100 Subject: perf: Remove superfluous arguments Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-yv4o74vh90suyghccgykbnry@git.kernel.org Signed-off-by: Ingo Molnar --- kernel/events/core.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) (limited to 'kernel/events') diff --git a/kernel/events/core.c b/kernel/events/core.c index b3fed52aaf2..eeda5403590 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -1667,8 +1667,7 @@ retry: * Note: this works for group members as well as group leaders * since the non-leader members' sibling_lists will be empty. */ -static void __perf_event_mark_enabled(struct perf_event *event, - struct perf_event_context *ctx) +static void __perf_event_mark_enabled(struct perf_event *event) { struct perf_event *sub; u64 tstamp = perf_event_time(event); @@ -1706,7 +1705,7 @@ static int __perf_event_enable(void *info) */ perf_cgroup_set_timestamp(current, ctx); - __perf_event_mark_enabled(event, ctx); + __perf_event_mark_enabled(event); if (!event_filter_match(event)) { if (is_cgroup_event(event)) @@ -1787,7 +1786,7 @@ void perf_event_enable(struct perf_event *event) retry: if (!ctx->is_active) { - __perf_event_mark_enabled(event, ctx); + __perf_event_mark_enabled(event); goto out; } @@ -2466,7 +2465,7 @@ static int event_enable_on_exec(struct perf_event *event, if (event->state >= PERF_EVENT_STATE_INACTIVE) return 0; - __perf_event_mark_enabled(event, ctx); + __perf_event_mark_enabled(event); return 1; } -- cgit v1.2.3-70-g09d2 From b79387ef185af2323594920923cecba5753c3817 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 22 Nov 2011 11:25:43 +0100 Subject: perf: Fix enable_on_exec for sibling events Deng-Cheng Zhu reported that sibling events that were created disabled with enable_on_exec would never get enabled. Iterate all events instead of the group lists. Reported-by: Deng-Cheng Zhu Tested-by: Deng-Cheng Zhu Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1322048382.14799.41.camel@twins Signed-off-by: Ingo Molnar --- kernel/events/core.c | 8 +------- 1 file changed, 1 insertion(+), 7 deletions(-) (limited to 'kernel/events') diff --git a/kernel/events/core.c b/kernel/events/core.c index eeda5403590..3c1541d7a53 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -2497,13 +2497,7 @@ static void perf_event_enable_on_exec(struct perf_event_context *ctx) raw_spin_lock(&ctx->lock); task_ctx_sched_out(ctx); - list_for_each_entry(event, &ctx->pinned_groups, group_entry) { - ret = event_enable_on_exec(event, ctx); - if (ret) - enabled = 1; - } - - list_for_each_entry(event, &ctx->flexible_groups, group_entry) { + list_for_each_entry(event, &ctx->event_list, event_entry) { ret = event_enable_on_exec(event, ctx); if (ret) enabled = 1; -- cgit v1.2.3-70-g09d2 From b202952075f62603bea9bfb6ebc6b0420db11949 Mon Sep 17 00:00:00 2001 From: Gleb Natapov Date: Sun, 27 Nov 2011 17:59:09 +0200 Subject: perf, core: Rate limit perf_sched_events jump_label patching jump_lable patching is very expensive operation that involves pausing all cpus. The patching of perf_sched_events jump_label is easily controllable from userspace by unprivileged user. When te user runs a loop like this: "while true; do perf stat -e cycles true; done" ... the performance of my test application that just increments a counter for one second drops by 4%. This is on a 16 cpu box with my test application using only one of them. An impact on a real server doing real work will be worse. Performance of KVM PMU drops nearly 50% due to jump_lable for "perf record" since KVM PMU implementation creates and destroys perf event frequently. This patch introduces a way to rate limit jump_label patching and uses it to fix the above problem. I believe that as jump_label use will spread the problem will become more common and thus solving it in a generic code is appropriate. Also fixing it in the perf code would result in moving jump_label accounting logic to perf code with all the ifdefs in case of JUMP_LABEL=n kernel. With this patch all details are nicely hidden inside jump_label code. Signed-off-by: Gleb Natapov Acked-by: Jason Baron Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20111127155909.GO2557@redhat.com Signed-off-by: Ingo Molnar --- include/linux/jump_label.h | 24 ++++++++++++++++++++++++ include/linux/perf_event.h | 6 +++--- kernel/events/core.c | 13 ++++++++----- kernel/jump_label.c | 35 +++++++++++++++++++++++++++++++++-- 4 files changed, 68 insertions(+), 10 deletions(-) (limited to 'kernel/events') diff --git a/include/linux/jump_label.h b/include/linux/jump_label.h index 388b0d425b5..a1e7f909c80 100644 --- a/include/linux/jump_label.h +++ b/include/linux/jump_label.h @@ -3,6 +3,7 @@ #include #include +#include #if defined(CC_HAVE_ASM_GOTO) && defined(CONFIG_JUMP_LABEL) @@ -14,6 +15,12 @@ struct jump_label_key { #endif }; +struct jump_label_key_deferred { + struct jump_label_key key; + unsigned long timeout; + struct delayed_work work; +}; + # include # define HAVE_JUMP_LABEL #endif /* CC_HAVE_ASM_GOTO && CONFIG_JUMP_LABEL */ @@ -51,8 +58,11 @@ extern void arch_jump_label_transform_static(struct jump_entry *entry, extern int jump_label_text_reserved(void *start, void *end); extern void jump_label_inc(struct jump_label_key *key); extern void jump_label_dec(struct jump_label_key *key); +extern void jump_label_dec_deferred(struct jump_label_key_deferred *key); extern bool jump_label_enabled(struct jump_label_key *key); extern void jump_label_apply_nops(struct module *mod); +extern void jump_label_rate_limit(struct jump_label_key_deferred *key, + unsigned long rl); #else /* !HAVE_JUMP_LABEL */ @@ -68,6 +78,10 @@ static __always_inline void jump_label_init(void) { } +struct jump_label_key_deferred { + struct jump_label_key key; +}; + static __always_inline bool static_branch(struct jump_label_key *key) { if (unlikely(atomic_read(&key->enabled))) @@ -85,6 +99,11 @@ static inline void jump_label_dec(struct jump_label_key *key) atomic_dec(&key->enabled); } +static inline void jump_label_dec_deferred(struct jump_label_key_deferred *key) +{ + jump_label_dec(&key->key); +} + static inline int jump_label_text_reserved(void *start, void *end) { return 0; @@ -102,6 +121,11 @@ static inline int jump_label_apply_nops(struct module *mod) { return 0; } + +static inline void jump_label_rate_limit(struct jump_label_key_deferred *key, + unsigned long rl) +{ +} #endif /* HAVE_JUMP_LABEL */ #endif /* _LINUX_JUMP_LABEL_H */ diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index cb44c9e7566..564769cdb47 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -1064,12 +1064,12 @@ perf_sw_event(u32 event_id, u64 nr, struct pt_regs *regs, u64 addr) } } -extern struct jump_label_key perf_sched_events; +extern struct jump_label_key_deferred perf_sched_events; static inline void perf_event_task_sched_in(struct task_struct *prev, struct task_struct *task) { - if (static_branch(&perf_sched_events)) + if (static_branch(&perf_sched_events.key)) __perf_event_task_sched_in(prev, task); } @@ -1078,7 +1078,7 @@ static inline void perf_event_task_sched_out(struct task_struct *prev, { perf_sw_event(PERF_COUNT_SW_CONTEXT_SWITCHES, 1, NULL, 0); - if (static_branch(&perf_sched_events)) + if (static_branch(&perf_sched_events.key)) __perf_event_task_sched_out(prev, next); } diff --git a/kernel/events/core.c b/kernel/events/core.c index 3c1541d7a53..3a3b1a18f49 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -128,7 +128,7 @@ enum event_type_t { * perf_sched_events : >0 events exist * perf_cgroup_events: >0 per-cpu cgroup events exist on this cpu */ -struct jump_label_key perf_sched_events __read_mostly; +struct jump_label_key_deferred perf_sched_events __read_mostly; static DEFINE_PER_CPU(atomic_t, perf_cgroup_events); static atomic_t nr_mmap_events __read_mostly; @@ -2748,7 +2748,7 @@ static void free_event(struct perf_event *event) if (!event->parent) { if (event->attach_state & PERF_ATTACH_TASK) - jump_label_dec(&perf_sched_events); + jump_label_dec_deferred(&perf_sched_events); if (event->attr.mmap || event->attr.mmap_data) atomic_dec(&nr_mmap_events); if (event->attr.comm) @@ -2759,7 +2759,7 @@ static void free_event(struct perf_event *event) put_callchain_buffers(); if (is_cgroup_event(event)) { atomic_dec(&per_cpu(perf_cgroup_events, event->cpu)); - jump_label_dec(&perf_sched_events); + jump_label_dec_deferred(&perf_sched_events); } } @@ -5784,7 +5784,7 @@ done: if (!event->parent) { if (event->attach_state & PERF_ATTACH_TASK) - jump_label_inc(&perf_sched_events); + jump_label_inc(&perf_sched_events.key); if (event->attr.mmap || event->attr.mmap_data) atomic_inc(&nr_mmap_events); if (event->attr.comm) @@ -6022,7 +6022,7 @@ SYSCALL_DEFINE5(perf_event_open, * - that may need work on context switch */ atomic_inc(&per_cpu(perf_cgroup_events, event->cpu)); - jump_label_inc(&perf_sched_events); + jump_label_inc(&perf_sched_events.key); } /* @@ -6868,6 +6868,9 @@ void __init perf_event_init(void) ret = init_hw_breakpoint(); WARN(ret, "hw_breakpoint initialization failed with: %d", ret); + + /* do not patch jump label more than once per second */ + jump_label_rate_limit(&perf_sched_events, HZ); } static int __init perf_event_sysfs_init(void) diff --git a/kernel/jump_label.c b/kernel/jump_label.c index 66ff7109f69..51a175ab0a0 100644 --- a/kernel/jump_label.c +++ b/kernel/jump_label.c @@ -72,15 +72,46 @@ void jump_label_inc(struct jump_label_key *key) jump_label_unlock(); } -void jump_label_dec(struct jump_label_key *key) +static void __jump_label_dec(struct jump_label_key *key, + unsigned long rate_limit, struct delayed_work *work) { if (!atomic_dec_and_mutex_lock(&key->enabled, &jump_label_mutex)) return; - jump_label_update(key, JUMP_LABEL_DISABLE); + if (rate_limit) { + atomic_inc(&key->enabled); + schedule_delayed_work(work, rate_limit); + } else + jump_label_update(key, JUMP_LABEL_DISABLE); + jump_label_unlock(); } +static void jump_label_update_timeout(struct work_struct *work) +{ + struct jump_label_key_deferred *key = + container_of(work, struct jump_label_key_deferred, work.work); + __jump_label_dec(&key->key, 0, NULL); +} + +void jump_label_dec(struct jump_label_key *key) +{ + __jump_label_dec(key, 0, NULL); +} + +void jump_label_dec_deferred(struct jump_label_key_deferred *key) +{ + __jump_label_dec(&key->key, key->timeout, &key->work); +} + + +void jump_label_rate_limit(struct jump_label_key_deferred *key, + unsigned long rl) +{ + key->timeout = rl; + INIT_DELAYED_WORK(&key->work, jump_label_update_timeout); +} + static int addr_conflict(struct jump_entry *entry, void *start, void *end) { if (entry->code <= (unsigned long)end && -- cgit v1.2.3-70-g09d2