1 files changed, 100 insertions, 259 deletions

diff --git a/kernel/events/core.c b/kernel/events/core.c
index 58690af323e..ba36013cfb2 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -4,7 +4,7 @@
  *  Copyright (C) 2008 Thomas Gleixner <tglx@linutronix.de>
  *  Copyright (C) 2008-2011 Red Hat, Inc., Ingo Molnar
  *  Copyright (C) 2008-2011 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
- *  Copyright  ©  2009 Paul Mackerras, IBM Corp. <paulus@au1.ibm.com>
+ *  Copyright  ©  2009 Paul Mackerras, IBM Corp. <paulus@au1.ibm.com>
  *
  * For licensing details see kernel-base/COPYING
  */
@@ -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;
@@ -815,7 +815,7 @@ static void update_event_times(struct perf_event *event)
 	 * here.
 	 */
 	if (is_cgroup_event(event))
-		run_end = perf_event_time(event);
+		run_end = perf_cgroup_event_time(event);
 	else if (ctx->is_active)
 		run_end = ctx->time;
 	else
@@ -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;
 }
@@ -1325,6 +1327,7 @@ 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,
@@ -1406,6 +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;
@@ -1662,8 +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);
@@ -1701,7 +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))
@@ -1782,7 +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;
 	}
 
@@ -1809,6 +1813,7 @@ retry:
 out:
 	raw_spin_unlock_irq(&ctx->lock);
 }
+EXPORT_SYMBOL_GPL(perf_event_enable);
 
 int perf_event_refresh(struct perf_event *event, int refresh)
 {
@@ -2295,6 +2300,9 @@ do {					\
 	return div64_u64(dividend, divisor);
 }
 
+static DEFINE_PER_CPU(int, perf_throttled_count);
+static DEFINE_PER_CPU(u64, perf_throttled_seq);
+
 static void perf_adjust_period(struct perf_event *event, u64 nsec, u64 count)
 {
 	struct hw_perf_event *hwc = &event->hw;
@@ -2320,13 +2328,29 @@ static void perf_adjust_period(struct perf_event *event, u64 nsec, u64 count)
 	}
 }
 
-static void perf_ctx_adjust_freq(struct perf_event_context *ctx, u64 period)
+/*
+ * combine freq adjustment with unthrottling to avoid two passes over the
+ * events. At the same time, make sure, having freq events does not change
+ * the rate of unthrottling as that would introduce bias.
+ */
+static void perf_adjust_freq_unthr_context(struct perf_event_context *ctx,
+					   int needs_unthr)
 {
 	struct perf_event *event;
 	struct hw_perf_event *hwc;
-	u64 interrupts, now;
+	u64 now, period = TICK_NSEC;
 	s64 delta;
 
+	/*
+	 * only need to iterate over all events iff:
+	 * - context have events in frequency mode (needs freq adjust)
+	 * - there are events to unthrottle on this cpu
+	 */
+	if (!(ctx->nr_freq || needs_unthr))
+		return;
+
+	raw_spin_lock(&ctx->lock);
+
 	list_for_each_entry_rcu(event, &ctx->event_list, event_entry) {
 		if (event->state != PERF_EVENT_STATE_ACTIVE)
 			continue;
@@ -2336,13 +2360,8 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx, u64 period)
 
 		hwc = &event->hw;
 
-		interrupts = hwc->interrupts;
-		hwc->interrupts = 0;
-
-		/*
-		 * unthrottle events on the tick
-		 */
-		if (interrupts == MAX_INTERRUPTS) {
+		if (needs_unthr && hwc->interrupts == MAX_INTERRUPTS) {
+			hwc->interrupts = 0;
 			perf_log_throttle(event, 1);
 			event->pmu->start(event, 0);
 		}
@@ -2350,14 +2369,26 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx, u64 period)
 		if (!event->attr.freq || !event->attr.sample_freq)
 			continue;
 
-		event->pmu->read(event);
+		/*
+		 * stop the event and update event->count
+		 */
+		event->pmu->stop(event, PERF_EF_UPDATE);
+
 		now = local64_read(&event->count);
 		delta = now - hwc->freq_count_stamp;
 		hwc->freq_count_stamp = now;
 
+		/*
+		 * restart the event
+		 * reload only if value has changed
+		 */
 		if (delta > 0)
 			perf_adjust_period(event, period, delta);
+
+		event->pmu->start(event, delta > 0 ? PERF_EF_RELOAD : 0);
 	}
+
+	raw_spin_unlock(&ctx->lock);
 }
 
 /*
@@ -2380,7 +2411,6 @@ static void rotate_ctx(struct perf_event_context *ctx)
  */
 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;
 
@@ -2397,15 +2427,12 @@ static void perf_rotate_context(struct perf_cpu_context *cpuctx)
 			rotate = 1;
 	}
 
-	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;
 
+	perf_ctx_lock(cpuctx, cpuctx->task_ctx);
+	perf_pmu_disable(cpuctx->ctx.pmu);
+
 	cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE);
 	if (ctx)
 		ctx_sched_out(ctx, cpuctx, EVENT_FLEXIBLE);
@@ -2416,22 +2443,33 @@ static void perf_rotate_context(struct perf_cpu_context *cpuctx)
 
 	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)
 {
 	struct list_head *head = &__get_cpu_var(rotation_list);
 	struct perf_cpu_context *cpuctx, *tmp;
+	struct perf_event_context *ctx;
+	int throttled;
 
 	WARN_ON(!irqs_disabled());
 
+	__this_cpu_inc(perf_throttled_seq);
+	throttled = __this_cpu_xchg(perf_throttled_count, 0);
+
 	list_for_each_entry_safe(cpuctx, tmp, head, rotation_list) {
+		ctx = &cpuctx->ctx;
+		perf_adjust_freq_unthr_context(ctx, throttled);
+
+		ctx = cpuctx->task_ctx;
+		if (ctx)
+			perf_adjust_freq_unthr_context(ctx, throttled);
+
 		if (cpuctx->jiffies_interval == 1 ||
 				!(jiffies % cpuctx->jiffies_interval))
 			perf_rotate_context(cpuctx);
@@ -2448,7 +2486,7 @@ 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;
 }
@@ -2480,13 +2518,7 @@ 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;
@@ -2574,215 +2606,6 @@ static u64 perf_event_read(struct perf_event *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:
  */
 static void __perf_event_init_context(struct perf_event_context *ctx)
@@ -2946,7 +2769,7 @@ 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)
@@ -2957,7 +2780,7 @@ 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);
 		}
 	}
 
@@ -4707,6 +4530,7 @@ static int __perf_event_overflow(struct perf_event *event,
 {
 	int events = atomic_read(&event->event_limit);
 	struct hw_perf_event *hwc = &event->hw;
+	u64 seq;
 	int ret = 0;
 
 	/*
@@ -4716,14 +4540,20 @@ static int __perf_event_overflow(struct perf_event *event,
 	if (unlikely(!is_sampling_event(event)))
 		return 0;
 
-	if (unlikely(hwc->interrupts >= max_samples_per_tick)) {
-		if (throttle) {
+	seq = __this_cpu_read(perf_throttled_seq);
+	if (seq != hwc->interrupts_seq) {
+		hwc->interrupts_seq = seq;
+		hwc->interrupts = 1;
+	} else {
+		hwc->interrupts++;
+		if (unlikely(throttle
+			     && hwc->interrupts >= max_samples_per_tick)) {
+			__this_cpu_inc(perf_throttled_count);
 			hwc->interrupts = MAX_INTERRUPTS;
 			perf_log_throttle(event, 0);
 			ret = 1;
 		}
-	} else
-		hwc->interrupts++;
+	}
 
 	if (event->attr.freq) {
 		u64 now = perf_clock();
@@ -4820,7 +4650,6 @@ 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);
 
@@ -4854,6 +4683,12 @@ 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);
 
@@ -5366,7 +5201,7 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer)
 	regs = get_irq_regs();
 
 	if (regs && !perf_exclude_event(event, regs)) {
-		if (!(event->attr.exclude_idle && current->pid == 0))
+		if (!(event->attr.exclude_idle && is_idle_task(current)))
 			if (perf_event_overflow(event, &data, regs))
 				ret = HRTIMER_NORESTART;
 	}
@@ -5981,7 +5816,7 @@ 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)
@@ -6219,7 +6054,7 @@ 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);
 	}
 
 	/*
@@ -7065,6 +6900,9 @@ 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)
@@ -7131,10 +6969,13 @@ static int __perf_cgroup_move(void *info)
 	return 0;
 }
 
-static void
-perf_cgroup_attach_task(struct cgroup *cgrp, struct task_struct *task)
+static void perf_cgroup_attach(struct cgroup_subsys *ss, struct cgroup *cgrp,
+			       struct cgroup_taskset *tset)
 {
-	task_function_call(task, __perf_cgroup_move, task);
+	struct task_struct *task;
+
+	cgroup_taskset_for_each(task, cgrp, tset)
+		task_function_call(task, __perf_cgroup_move, task);
 }
 
 static void perf_cgroup_exit(struct cgroup_subsys *ss, struct cgroup *cgrp,
@@ -7148,7 +6989,7 @@ static void perf_cgroup_exit(struct cgroup_subsys *ss, struct cgroup *cgrp,
 	if (!(task->flags & PF_EXITING))
 		return;
 
-	perf_cgroup_attach_task(cgrp, task);
+	task_function_call(task, __perf_cgroup_move, task);
 }
 
 struct cgroup_subsys perf_subsys = {
@@ -7157,6 +6998,6 @@ struct cgroup_subsys perf_subsys = {
 	.create		= perf_cgroup_create,
 	.destroy	= perf_cgroup_destroy,
 	.exit		= perf_cgroup_exit,
-	.attach_task	= perf_cgroup_attach_task,
+	.attach		= perf_cgroup_attach,
 };
 #endif /* CONFIG_CGROUP_PERF */