Merge branch 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull perf updates from Ingo Molnar:
 "Kernel improvements:

   - watchdog driver improvements by Li Zefan
   - Power7 CPI stack events related improvements by Sukadev Bhattiprolu
   - event multiplexing via hrtimers and other improvements by Stephane
     Eranian
   - kernel stack use optimization by Andrew Hunter
   - AMD IOMMU uncore PMU support by Suravee Suthikulpanit
   - NMI handling rate-limits by Dave Hansen
   - various hw_breakpoint fixes by Oleg Nesterov
   - hw_breakpoint overflow period sampling and related signal handling
     fixes by Jiri Olsa
   - Intel Haswell PMU support by Andi Kleen

  Tooling improvements:

   - Reset SIGTERM handler in workload child process, fix from David
     Ahern.
   - Makefile reorganization, prep work for Kconfig patches, from Jiri
     Olsa.
   - Add automated make test suite, from Jiri Olsa.
   - Add --percent-limit option to 'top' and 'report', from Namhyung
     Kim.
   - Sorting improvements, from Namhyung Kim.
   - Expand definition of sysfs format attribute, from Michael Ellerman.

  Tooling fixes:

   - 'perf tests' fixes from Jiri Olsa.
   - Make Power7 CPI stack events available in sysfs, from Sukadev
     Bhattiprolu.
   - Handle death by SIGTERM in 'perf record', fix from David Ahern.
   - Fix printing of perf_event_paranoid message, from David Ahern.
   - Handle realloc failures in 'perf kvm', from David Ahern.
   - Fix divide by 0 in variance, from David Ahern.
   - Save parent pid in thread struct, from David Ahern.
   - Handle JITed code in shared memory, from Andi Kleen.
   - Fixes for 'perf diff', from Jiri Olsa.
   - Remove some unused struct members, from Jiri Olsa.
   - Add missing liblk.a dependency for python/perf.so, fix from Jiri
     Olsa.
   - Respect CROSS_COMPILE in liblk.a, from Rabin Vincent.
   - No need to do locking when adding hists in perf report, only 'top'
     needs that, from Namhyung Kim.
   - Fix alignment of symbol column in in the hists browser (top,
     report) when -v is given, from NAmhyung Kim.
   - Fix 'perf top' -E option behavior, from Namhyung Kim.
   - Fix bug in isupper() and islower(), from Sukadev Bhattiprolu.
   - Fix compile errors in bp_signal 'perf test', from Sukadev
     Bhattiprolu.

  ... and more things"

* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (102 commits)
  perf/x86: Disable PEBS-LL in intel_pmu_pebs_disable()
  perf/x86: Fix shared register mutual exclusion enforcement
  perf/x86/intel: Support full width counting
  x86: Add NMI duration tracepoints
  perf: Drop sample rate when sampling is too slow
  x86: Warn when NMI handlers take large amounts of time
  hw_breakpoint: Introduce "struct bp_cpuinfo"
  hw_breakpoint: Simplify *register_wide_hw_breakpoint()
  hw_breakpoint: Introduce cpumask_of_bp()
  hw_breakpoint: Simplify the "weight" usage in toggle_bp_slot() paths
  hw_breakpoint: Simplify list/idx mess in toggle_bp_slot() paths
  perf/x86/intel: Add mem-loads/stores support for Haswell
  perf/x86/intel: Support Haswell/v4 LBR format
  perf/x86/intel: Move NMI clearing to end of PMI handler
  perf/x86/intel: Add Haswell PEBS support
  perf/x86/intel: Add simple Haswell PMU support
  perf/x86/intel: Add Haswell PEBS record support
  perf/x86/intel: Fix sparse warning
  perf/x86/amd: AMD IOMMU Performance Counter PERF uncore PMU implementation
  perf/x86/amd: Add IOMMU Performance Counter resource management
  ...

kernel/events/core.c

@@ -165,10 +165,28 @@ int sysctl_perf_event_mlock __read_mostly = 512 + (PAGE_SIZE / 1024); /* 'free'
 /*
  * max perf event sample rate
  */
-#define DEFAULT_MAX_SAMPLE_RATE 100000
-int sysctl_perf_event_sample_rate __read_mostly = DEFAULT_MAX_SAMPLE_RATE;
-static int max_samples_per_tick __read_mostly =
-	DIV_ROUND_UP(DEFAULT_MAX_SAMPLE_RATE, HZ);
+#define DEFAULT_MAX_SAMPLE_RATE		100000
+#define DEFAULT_SAMPLE_PERIOD_NS	(NSEC_PER_SEC / DEFAULT_MAX_SAMPLE_RATE)
+#define DEFAULT_CPU_TIME_MAX_PERCENT	25
+
+int sysctl_perf_event_sample_rate __read_mostly	= DEFAULT_MAX_SAMPLE_RATE;
+
+static int max_samples_per_tick __read_mostly	= DIV_ROUND_UP(DEFAULT_MAX_SAMPLE_RATE, HZ);
+static int perf_sample_period_ns __read_mostly	= DEFAULT_SAMPLE_PERIOD_NS;
+
+static atomic_t perf_sample_allowed_ns __read_mostly =
+	ATOMIC_INIT( DEFAULT_SAMPLE_PERIOD_NS * DEFAULT_CPU_TIME_MAX_PERCENT / 100);
+
+void update_perf_cpu_limits(void)
+{
+	u64 tmp = perf_sample_period_ns;
+
+	tmp *= sysctl_perf_cpu_time_max_percent;
+	tmp = do_div(tmp, 100);
+	atomic_set(&perf_sample_allowed_ns, tmp);
+}
+
+static int perf_rotate_context(struct perf_cpu_context *cpuctx);
 
 int perf_proc_update_handler(struct ctl_table *table, int write,
 		void __user *buffer, size_t *lenp,
@@ -180,10 +198,78 @@ int perf_proc_update_handler(struct ctl_table *table, int write,
 		return ret;
 
 	max_samples_per_tick = DIV_ROUND_UP(sysctl_perf_event_sample_rate, HZ);
+	perf_sample_period_ns = NSEC_PER_SEC / sysctl_perf_event_sample_rate;
+	update_perf_cpu_limits();
 
 	return 0;
 }
 
+int sysctl_perf_cpu_time_max_percent __read_mostly = DEFAULT_CPU_TIME_MAX_PERCENT;
+
+int perf_cpu_time_max_percent_handler(struct ctl_table *table, int write,
+				void __user *buffer, size_t *lenp,
+				loff_t *ppos)
+{
+	int ret = proc_dointvec(table, write, buffer, lenp, ppos);
+
+	if (ret || !write)
+		return ret;
+
+	update_perf_cpu_limits();
+
+	return 0;
+}
+
+/*
+ * perf samples are done in some very critical code paths (NMIs).
+ * If they take too much CPU time, the system can lock up and not
+ * get any real work done.  This will drop the sample rate when
+ * we detect that events are taking too long.
+ */
+#define NR_ACCUMULATED_SAMPLES 128
+DEFINE_PER_CPU(u64, running_sample_length);
+
+void perf_sample_event_took(u64 sample_len_ns)
+{
+	u64 avg_local_sample_len;
+	u64 local_samples_len = __get_cpu_var(running_sample_length);
+
+	if (atomic_read(&perf_sample_allowed_ns) == 0)
+		return;
+
+	/* decay the counter by 1 average sample */
+	local_samples_len = __get_cpu_var(running_sample_length);
+	local_samples_len -= local_samples_len/NR_ACCUMULATED_SAMPLES;
+	local_samples_len += sample_len_ns;
+	__get_cpu_var(running_sample_length) = local_samples_len;
+
+	/*
+	 * note: this will be biased artifically low until we have
+	 * seen NR_ACCUMULATED_SAMPLES.  Doing it this way keeps us
+	 * from having to maintain a count.
+	 */
+	avg_local_sample_len = local_samples_len/NR_ACCUMULATED_SAMPLES;
+
+	if (avg_local_sample_len <= atomic_read(&perf_sample_allowed_ns))
+		return;
+
+	if (max_samples_per_tick <= 1)
+		return;
+
+	max_samples_per_tick = DIV_ROUND_UP(max_samples_per_tick, 2);
+	sysctl_perf_event_sample_rate = max_samples_per_tick * HZ;
+	perf_sample_period_ns = NSEC_PER_SEC / sysctl_perf_event_sample_rate;
+
+	printk_ratelimited(KERN_WARNING
+			"perf samples too long (%lld > %d), lowering "
+			"kernel.perf_event_max_sample_rate to %d\n",
+			avg_local_sample_len,
+			atomic_read(&perf_sample_allowed_ns),
+			sysctl_perf_event_sample_rate);
+
+	update_perf_cpu_limits();
+}
+
 static atomic64_t perf_event_id;
 
 static void cpu_ctx_sched_out(struct perf_cpu_context *cpuctx,
@@ -655,6 +741,106 @@ perf_cgroup_mark_enabled(struct perf_event *event,
 }
 #endif
 
+/*
+ * set default to be dependent on timer tick just
+ * like original code
+ */
+#define PERF_CPU_HRTIMER (1000 / HZ)
+/*
+ * function must be called with interrupts disbled
+ */
+static enum hrtimer_restart perf_cpu_hrtimer_handler(struct hrtimer *hr)
+{
+	struct perf_cpu_context *cpuctx;
+	enum hrtimer_restart ret = HRTIMER_NORESTART;
+	int rotations = 0;
+
+	WARN_ON(!irqs_disabled());
+
+	cpuctx = container_of(hr, struct perf_cpu_context, hrtimer);
+
+	rotations = perf_rotate_context(cpuctx);
+
+	/*
+	 * arm timer if needed
+	 */
+	if (rotations) {
+		hrtimer_forward_now(hr, cpuctx->hrtimer_interval);
+		ret = HRTIMER_RESTART;
+	}
+
+	return ret;
+}
+
+/* CPU is going down */
+void perf_cpu_hrtimer_cancel(int cpu)
+{
+	struct perf_cpu_context *cpuctx;
+	struct pmu *pmu;
+	unsigned long flags;
+
+	if (WARN_ON(cpu != smp_processor_id()))
+		return;
+
+	local_irq_save(flags);
+
+	rcu_read_lock();
+
+	list_for_each_entry_rcu(pmu, &pmus, entry) {
+		cpuctx = this_cpu_ptr(pmu->pmu_cpu_context);
+
+		if (pmu->task_ctx_nr == perf_sw_context)
+			continue;
+
+		hrtimer_cancel(&cpuctx->hrtimer);
+	}
+
+	rcu_read_unlock();
+
+	local_irq_restore(flags);
+}
+
+static void __perf_cpu_hrtimer_init(struct perf_cpu_context *cpuctx, int cpu)
+{
+	struct hrtimer *hr = &cpuctx->hrtimer;
+	struct pmu *pmu = cpuctx->ctx.pmu;
+	int timer;
+
+	/* no multiplexing needed for SW PMU */
+	if (pmu->task_ctx_nr == perf_sw_context)
+		return;
+
+	/*
+	 * check default is sane, if not set then force to
+	 * default interval (1/tick)
+	 */
+	timer = pmu->hrtimer_interval_ms;
+	if (timer < 1)
+		timer = pmu->hrtimer_interval_ms = PERF_CPU_HRTIMER;
+
+	cpuctx->hrtimer_interval = ns_to_ktime(NSEC_PER_MSEC * timer);
+
+	hrtimer_init(hr, CLOCK_MONOTONIC, HRTIMER_MODE_REL_PINNED);
+	hr->function = perf_cpu_hrtimer_handler;
+}
+
+static void perf_cpu_hrtimer_restart(struct perf_cpu_context *cpuctx)
+{
+	struct hrtimer *hr = &cpuctx->hrtimer;
+	struct pmu *pmu = cpuctx->ctx.pmu;
+
+	/* not for SW PMU */
+	if (pmu->task_ctx_nr == perf_sw_context)
+		return;
+
+	if (hrtimer_active(hr))
+		return;
+
+	if (!hrtimer_callback_running(hr))
+		__hrtimer_start_range_ns(hr, cpuctx->hrtimer_interval,
+					 0, HRTIMER_MODE_REL_PINNED, 0);
+}
+
 void perf_pmu_disable(struct pmu *pmu)
 {
 	int *count = this_cpu_ptr(pmu->pmu_disable_count);
@@ -1503,6 +1689,7 @@ group_sched_in(struct perf_event *group_event,
 
 	if (event_sched_in(group_event, cpuctx, ctx)) {
 		pmu->cancel_txn(pmu);
+		perf_cpu_hrtimer_restart(cpuctx);
 		return -EAGAIN;
 	}
 
@@ -1549,6 +1736,8 @@ group_error:
 
 	pmu->cancel_txn(pmu);
 
+	perf_cpu_hrtimer_restart(cpuctx);
+
 	return -EAGAIN;
 }
 
@@ -1804,8 +1993,10 @@ static int __perf_event_enable(void *info)
 		 * If this event can't go on and it's part of a
 		 * group, then the whole group has to come off.
 		 */
-		if (leader != event)
+		if (leader != event) {
 			group_sched_out(leader, cpuctx, ctx);
+			perf_cpu_hrtimer_restart(cpuctx);
+		}
 		if (leader->attr.pinned) {
 			update_group_times(leader);
 			leader->state = PERF_EVENT_STATE_ERROR;
@@ -2552,7 +2743,7 @@ static void rotate_ctx(struct perf_event_context *ctx)
  * because they're strictly cpu affine and rotate_start is called with IRQs
  * disabled, while rotate_context is called from IRQ context.
  */
-static void perf_rotate_context(struct perf_cpu_context *cpuctx)
+static int perf_rotate_context(struct perf_cpu_context *cpuctx)
 {
 	struct perf_event_context *ctx = NULL;
 	int rotate = 0, remove = 1;
@@ -2591,6 +2782,8 @@ static void perf_rotate_context(struct perf_cpu_context *cpuctx)
 done:
 	if (remove)
 		list_del_init(&cpuctx->rotation_list);
+
+	return rotate;
 }
 
 #ifdef CONFIG_NO_HZ_FULL
@@ -2622,10 +2815,6 @@ void perf_event_task_tick(void)
 		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);
 	}
 }
 
@@ -5036,7 +5225,7 @@ static DEFINE_PER_CPU(struct swevent_htable, swevent_htable);
  * sign as trigger.
  */
 
-static u64 perf_swevent_set_period(struct perf_event *event)
+u64 perf_swevent_set_period(struct perf_event *event)
 {
 	struct hw_perf_event *hwc = &event->hw;
 	u64 period = hwc->last_period;
@@ -5979,9 +6168,56 @@ type_show(struct device *dev, struct device_attribute *attr, char *page)
 	return snprintf(page, PAGE_SIZE-1, "%d\n", pmu->type);
 }
 
+static ssize_t
+perf_event_mux_interval_ms_show(struct device *dev,
+				struct device_attribute *attr,
+				char *page)
+{
+	struct pmu *pmu = dev_get_drvdata(dev);
+
+	return snprintf(page, PAGE_SIZE-1, "%d\n", pmu->hrtimer_interval_ms);
+}
+
+static ssize_t
+perf_event_mux_interval_ms_store(struct device *dev,
+				 struct device_attribute *attr,
+				 const char *buf, size_t count)
+{
+	struct pmu *pmu = dev_get_drvdata(dev);
+	int timer, cpu, ret;
+
+	ret = kstrtoint(buf, 0, &timer);
+	if (ret)
+		return ret;
+
+	if (timer < 1)
+		return -EINVAL;
+
+	/* same value, noting to do */
+	if (timer == pmu->hrtimer_interval_ms)
+		return count;
+
+	pmu->hrtimer_interval_ms = timer;
+
+	/* update all cpuctx for this PMU */
+	for_each_possible_cpu(cpu) {
+		struct perf_cpu_context *cpuctx;
+		cpuctx = per_cpu_ptr(pmu->pmu_cpu_context, cpu);
+		cpuctx->hrtimer_interval = ns_to_ktime(NSEC_PER_MSEC * timer);
+
+		if (hrtimer_active(&cpuctx->hrtimer))
+			hrtimer_forward_now(&cpuctx->hrtimer, cpuctx->hrtimer_interval);
+	}
+
+	return count;
+}
+
+#define __ATTR_RW(attr) __ATTR(attr, 0644, attr##_show, attr##_store)
+
 static struct device_attribute pmu_dev_attrs[] = {
-       __ATTR_RO(type),
-       __ATTR_NULL,
+	__ATTR_RO(type),
+	__ATTR_RW(perf_event_mux_interval_ms),
+	__ATTR_NULL,
 };
 
 static int pmu_bus_running;
@@ -6027,7 +6263,7 @@ free_dev:
 static struct lock_class_key cpuctx_mutex;
 static struct lock_class_key cpuctx_lock;
 
-int perf_pmu_register(struct pmu *pmu, char *name, int type)
+int perf_pmu_register(struct pmu *pmu, const char *name, int type)
 {
 	int cpu, ret;
 
@@ -6076,7 +6312,9 @@ skip_type:
 		lockdep_set_class(&cpuctx->ctx.lock, &cpuctx_lock);
 		cpuctx->ctx.type = cpu_context;
 		cpuctx->ctx.pmu = pmu;
-		cpuctx->jiffies_interval = 1;
+
+		__perf_cpu_hrtimer_init(cpuctx, cpu);
+
 		INIT_LIST_HEAD(&cpuctx->rotation_list);
 		cpuctx->unique_pmu = pmu;
 	}
@@ -6402,11 +6640,6 @@ static int perf_copy_attr(struct perf_event_attr __user *uattr,
 		if (!(mask & ~PERF_SAMPLE_BRANCH_PLM_ALL))
 			return -EINVAL;
 
-		/* kernel level capture: check permissions */
-		if ((mask & PERF_SAMPLE_BRANCH_PERM_PLM)
-		    && perf_paranoid_kernel() && !capable(CAP_SYS_ADMIN))
-			return -EACCES;
-
 		/* propagate priv level, when not set for branch */
 		if (!(mask & PERF_SAMPLE_BRANCH_PLM_ALL)) {
 
@@ -6424,6 +6657,10 @@ static int perf_copy_attr(struct perf_event_attr __user *uattr,
 			 */
 			attr->branch_sample_type = mask;
 		}
+		/* privileged levels capture (kernel, hv): check permissions */
+		if ((mask & PERF_SAMPLE_BRANCH_PERM_PLM)
+		    && perf_paranoid_kernel() && !capable(CAP_SYS_ADMIN))
+			return -EACCES;
 	}
 
 	if (attr->sample_type & PERF_SAMPLE_REGS_USER) {
@@ -7476,7 +7713,6 @@ perf_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu)
 	case CPU_DOWN_PREPARE:
 		perf_event_exit_cpu(cpu);
 		break;
-
 	default:
 		break;
 	}

kernel/events/hw_breakpoint.c

@@ -46,23 +46,26 @@
 #include <linux/smp.h>
 
 #include <linux/hw_breakpoint.h>
-
-
 /*
  * Constraints data
  */
+struct bp_cpuinfo {
+	/* Number of pinned cpu breakpoints in a cpu */
+	unsigned int	cpu_pinned;
+	/* tsk_pinned[n] is the number of tasks having n+1 breakpoints */
+	unsigned int	*tsk_pinned;
+	/* Number of non-pinned cpu/task breakpoints in a cpu */
+	unsigned int	flexible; /* XXX: placeholder, see fetch_this_slot() */
+};
 
-/* Number of pinned cpu breakpoints in a cpu */
-static DEFINE_PER_CPU(unsigned int, nr_cpu_bp_pinned[TYPE_MAX]);
-
-/* Number of pinned task breakpoints in a cpu */
-static DEFINE_PER_CPU(unsigned int *, nr_task_bp_pinned[TYPE_MAX]);
-
-/* Number of non-pinned cpu/task breakpoints in a cpu */
-static DEFINE_PER_CPU(unsigned int, nr_bp_flexible[TYPE_MAX]);
-
+static DEFINE_PER_CPU(struct bp_cpuinfo, bp_cpuinfo[TYPE_MAX]);
 static int nr_slots[TYPE_MAX];
 
+static struct bp_cpuinfo *get_bp_info(int cpu, enum bp_type_idx type)
+{
+	return per_cpu_ptr(bp_cpuinfo + type, cpu);
+}
+
 /* Keep track of the breakpoints attached to tasks */
 static LIST_HEAD(bp_task_head);
 
@@ -96,8 +99,8 @@ static inline enum bp_type_idx find_slot_idx(struct perf_event *bp)
  */
 static unsigned int max_task_bp_pinned(int cpu, enum bp_type_idx type)
 {
+	unsigned int *tsk_pinned = get_bp_info(cpu, type)->tsk_pinned;
 	int i;
-	unsigned int *tsk_pinned = per_cpu(nr_task_bp_pinned[type], cpu);
 
 	for (i = nr_slots[type] - 1; i >= 0; i--) {
 		if (tsk_pinned[i] > 0)
@@ -127,6 +130,13 @@ static int task_bp_pinned(int cpu, struct perf_event *bp, enum bp_type_idx type)
 	return count;
 }
 
+static const struct cpumask *cpumask_of_bp(struct perf_event *bp)
+{
+	if (bp->cpu >= 0)
+		return cpumask_of(bp->cpu);
+	return cpu_possible_mask;
+}
+
 /*
  * Report the number of pinned/un-pinned breakpoints we have in
  * a given cpu (cpu > -1) or in all of them (cpu = -1).
@@ -135,25 +145,15 @@ static void
 fetch_bp_busy_slots(struct bp_busy_slots *slots, struct perf_event *bp,
 		    enum bp_type_idx type)
 {
-	int cpu = bp->cpu;
-	struct task_struct *tsk = bp->hw.bp_target;
+	const struct cpumask *cpumask = cpumask_of_bp(bp);
+	int cpu;
 
-	if (cpu >= 0) {
-		slots->pinned = per_cpu(nr_cpu_bp_pinned[type], cpu);
-		if (!tsk)
-			slots->pinned += max_task_bp_pinned(cpu, type);
-		else
-			slots->pinned += task_bp_pinned(cpu, bp, type);
-		slots->flexible = per_cpu(nr_bp_flexible[type], cpu);
+	for_each_cpu(cpu, cpumask) {
+		struct bp_cpuinfo *info = get_bp_info(cpu, type);
+		int nr;
 
-		return;
-	}
-
-	for_each_possible_cpu(cpu) {
-		unsigned int nr;
-
-		nr = per_cpu(nr_cpu_bp_pinned[type], cpu);
-		if (!tsk)
+		nr = info->cpu_pinned;
+		if (!bp->hw.bp_target)
 			nr += max_task_bp_pinned(cpu, type);
 		else
 			nr += task_bp_pinned(cpu, bp, type);
@@ -161,8 +161,7 @@ fetch_bp_busy_slots(struct bp_busy_slots *slots, struct perf_event *bp,
 		if (nr > slots->pinned)
 			slots->pinned = nr;
 
-		nr = per_cpu(nr_bp_flexible[type], cpu);
-
+		nr = info->flexible;
 		if (nr > slots->flexible)
 			slots->flexible = nr;
 	}
@@ -182,29 +181,19 @@ fetch_this_slot(struct bp_busy_slots *slots, int weight)
 /*
  * Add a pinned breakpoint for the given task in our constraint table
  */
-static void toggle_bp_task_slot(struct perf_event *bp, int cpu, bool enable,
+static void toggle_bp_task_slot(struct perf_event *bp, int cpu,
 				enum bp_type_idx type, int weight)
 {
-	unsigned int *tsk_pinned;
-	int old_count = 0;
-	int old_idx = 0;
-	int idx = 0;
+	unsigned int *tsk_pinned = get_bp_info(cpu, type)->tsk_pinned;
+	int old_idx, new_idx;
 
-	old_count = task_bp_pinned(cpu, bp, type);
-	old_idx = old_count - 1;
-	idx = old_idx + weight;
+	old_idx = task_bp_pinned(cpu, bp, type) - 1;
+	new_idx = old_idx + weight;
 
-	/* tsk_pinned[n] is the number of tasks having n breakpoints */
-	tsk_pinned = per_cpu(nr_task_bp_pinned[type], cpu);
-	if (enable) {
-		tsk_pinned[idx]++;
-		if (old_count > 0)
-			tsk_pinned[old_idx]--;
-	} else {
-		tsk_pinned[idx]--;
-		if (old_count > 0)
-			tsk_pinned[old_idx]++;
-	}
+	if (old_idx >= 0)
+		tsk_pinned[old_idx]--;
+	if (new_idx >= 0)
+		tsk_pinned[new_idx]++;
 }
 
 /*
@@ -214,33 +203,26 @@ static void
 toggle_bp_slot(struct perf_event *bp, bool enable, enum bp_type_idx type,
 	       int weight)
 {
-	int cpu = bp->cpu;
-	struct task_struct *tsk = bp->hw.bp_target;
+	const struct cpumask *cpumask = cpumask_of_bp(bp);
+	int cpu;
+
+	if (!enable)
+		weight = -weight;
 
 	/* Pinned counter cpu profiling */
-	if (!tsk) {
-
-		if (enable)
-			per_cpu(nr_cpu_bp_pinned[type], bp->cpu) += weight;
-		else
-			per_cpu(nr_cpu_bp_pinned[type], bp->cpu) -= weight;
+	if (!bp->hw.bp_target) {
+		get_bp_info(bp->cpu, type)->cpu_pinned += weight;
 		return;
 	}
 
 	/* Pinned counter task profiling */
-
-	if (!enable)
-		list_del(&bp->hw.bp_list);
-
-	if (cpu >= 0) {
-		toggle_bp_task_slot(bp, cpu, enable, type, weight);
-	} else {
-		for_each_possible_cpu(cpu)
-			toggle_bp_task_slot(bp, cpu, enable, type, weight);
-	}
+	for_each_cpu(cpu, cpumask)
+		toggle_bp_task_slot(bp, cpu, type, weight);
 
 	if (enable)
 		list_add_tail(&bp->hw.bp_list, &bp_task_head);
+	else
+		list_del(&bp->hw.bp_list);
 }
 
 /*
@@ -261,8 +243,8 @@ __weak void arch_unregister_hw_breakpoint(struct perf_event *bp)
  *
  *   - If attached to a single cpu, check:
  *
- *       (per_cpu(nr_bp_flexible, cpu) || (per_cpu(nr_cpu_bp_pinned, cpu)
- *           + max(per_cpu(nr_task_bp_pinned, cpu)))) < HBP_NUM
+ *       (per_cpu(info->flexible, cpu) || (per_cpu(info->cpu_pinned, cpu)
+ *           + max(per_cpu(info->tsk_pinned, cpu)))) < HBP_NUM
  *
  *       -> If there are already non-pinned counters in this cpu, it means
  *          there is already a free slot for them.
@@ -272,8 +254,8 @@ __weak void arch_unregister_hw_breakpoint(struct perf_event *bp)
  *
  *   - If attached to every cpus, check:
  *
- *       (per_cpu(nr_bp_flexible, *) || (max(per_cpu(nr_cpu_bp_pinned, *))
- *           + max(per_cpu(nr_task_bp_pinned, *)))) < HBP_NUM
+ *       (per_cpu(info->flexible, *) || (max(per_cpu(info->cpu_pinned, *))
+ *           + max(per_cpu(info->tsk_pinned, *)))) < HBP_NUM
  *
  *       -> This is roughly the same, except we check the number of per cpu
  *          bp for every cpu and we keep the max one. Same for the per tasks
@@ -284,16 +266,16 @@ __weak void arch_unregister_hw_breakpoint(struct perf_event *bp)
  *
  *   - If attached to a single cpu, check:
  *
- *       ((per_cpu(nr_bp_flexible, cpu) > 1) + per_cpu(nr_cpu_bp_pinned, cpu)
- *            + max(per_cpu(nr_task_bp_pinned, cpu))) < HBP_NUM
+ *       ((per_cpu(info->flexible, cpu) > 1) + per_cpu(info->cpu_pinned, cpu)
+ *            + max(per_cpu(info->tsk_pinned, cpu))) < HBP_NUM
  *
- *       -> Same checks as before. But now the nr_bp_flexible, if any, must keep
+ *       -> Same checks as before. But now the info->flexible, if any, must keep
  *          one register at least (or they will never be fed).
  *
  *   - If attached to every cpus, check:
  *
- *       ((per_cpu(nr_bp_flexible, *) > 1) + max(per_cpu(nr_cpu_bp_pinned, *))
- *            + max(per_cpu(nr_task_bp_pinned, *))) < HBP_NUM
+ *       ((per_cpu(info->flexible, *) > 1) + max(per_cpu(info->cpu_pinned, *))
+ *            + max(per_cpu(info->tsk_pinned, *))) < HBP_NUM
  */
 static int __reserve_bp_slot(struct perf_event *bp)
 {
@@ -518,8 +500,8 @@ register_wide_hw_breakpoint(struct perf_event_attr *attr,
 			    perf_overflow_handler_t triggered,
 			    void *context)
 {
-	struct perf_event * __percpu *cpu_events, **pevent, *bp;
-	long err;
+	struct perf_event * __percpu *cpu_events, *bp;
+	long err = 0;
 	int cpu;
 
 	cpu_events = alloc_percpu(typeof(*cpu_events));
@@ -528,31 +510,21 @@ 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, context);
-
-		*pevent = bp;
-
 		if (IS_ERR(bp)) {
 			err = PTR_ERR(bp);
-			goto fail;
-		}
-	}
-	put_online_cpus();
-
-	return cpu_events;
-
-fail:
-	for_each_online_cpu(cpu) {
-		pevent = per_cpu_ptr(cpu_events, cpu);
-		if (IS_ERR(*pevent))
 			break;
-		unregister_hw_breakpoint(*pevent);
+		}
+
+		per_cpu(*cpu_events, cpu) = bp;
 	}
 	put_online_cpus();
 
-	free_percpu(cpu_events);
+	if (likely(!err))
+		return cpu_events;
+
+	unregister_wide_hw_breakpoint(cpu_events);
 	return (void __percpu __force *)ERR_PTR(err);
 }
 EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint);
@@ -564,12 +536,10 @@ EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint);
 void unregister_wide_hw_breakpoint(struct perf_event * __percpu *cpu_events)
 {
 	int cpu;
-	struct perf_event **pevent;
 
-	for_each_possible_cpu(cpu) {
-		pevent = per_cpu_ptr(cpu_events, cpu);
-		unregister_hw_breakpoint(*pevent);
-	}
+	for_each_possible_cpu(cpu)
+		unregister_hw_breakpoint(per_cpu(*cpu_events, cpu));
+
 	free_percpu(cpu_events);
 }
 EXPORT_SYMBOL_GPL(unregister_wide_hw_breakpoint);
@@ -612,6 +582,11 @@ static int hw_breakpoint_add(struct perf_event *bp, int flags)
 	if (!(flags & PERF_EF_START))
 		bp->hw.state = PERF_HES_STOPPED;
 
+	if (is_sampling_event(bp)) {
+		bp->hw.last_period = bp->hw.sample_period;
+		perf_swevent_set_period(bp);
+	}
+
 	return arch_install_hw_breakpoint(bp);
 }
 
@@ -650,7 +625,6 @@ static struct pmu perf_breakpoint = {
 
 int __init init_hw_breakpoint(void)
 {
-	unsigned int **task_bp_pinned;
 	int cpu, err_cpu;
 	int i;
 
@@ -659,10 +633,11 @@ int __init init_hw_breakpoint(void)
 
 	for_each_possible_cpu(cpu) {
 		for (i = 0; i < TYPE_MAX; i++) {
-			task_bp_pinned = &per_cpu(nr_task_bp_pinned[i], cpu);
-			*task_bp_pinned = kzalloc(sizeof(int) * nr_slots[i],
-						  GFP_KERNEL);
-			if (!*task_bp_pinned)
+			struct bp_cpuinfo *info = get_bp_info(cpu, i);
+
+			info->tsk_pinned = kcalloc(nr_slots[i], sizeof(int),
+							GFP_KERNEL);
+			if (!info->tsk_pinned)
 				goto err_alloc;
 		}
 	}
@@ -676,7 +651,7 @@ int __init init_hw_breakpoint(void)
  err_alloc:
 	for_each_possible_cpu(err_cpu) {
 		for (i = 0; i < TYPE_MAX; i++)
-			kfree(per_cpu(nr_task_bp_pinned[i], err_cpu));
+			kfree(get_bp_info(err_cpu, i)->tsk_pinned);
 		if (err_cpu == cpu)
 			break;
 	}

kernel/sysctl.c

@@ -120,7 +120,6 @@ extern int blk_iopoll_enabled;
 /* Constants used for minimum and  maximum */
 #ifdef CONFIG_LOCKUP_DETECTOR
 static int sixty = 60;
-static int neg_one = -1;
 #endif
 
 static int zero;
@@ -814,7 +813,7 @@ static struct ctl_table kern_table[] = {
 		.maxlen		= sizeof(int),
 		.mode		= 0644,
 		.proc_handler	= proc_dowatchdog,
-		.extra1		= &neg_one,
+		.extra1		= &zero,
 		.extra2		= &sixty,
 	},
 	{
@@ -1044,6 +1043,15 @@ static struct ctl_table kern_table[] = {
 		.mode		= 0644,
 		.proc_handler	= perf_proc_update_handler,
 	},
+	{
+		.procname	= "perf_cpu_time_max_percent",
+		.data		= &sysctl_perf_cpu_time_max_percent,
+		.maxlen		= sizeof(sysctl_perf_cpu_time_max_percent),
+		.mode		= 0644,
+		.proc_handler	= perf_cpu_time_max_percent_handler,
+		.extra1		= &zero,
+		.extra2		= &one_hundred,
+	},
 #endif
 #ifdef CONFIG_KMEMCHECK
 	{