1 files changed, 334 insertions, 183 deletions

diff --git a/kernel/perf_event.c b/kernel/perf_event.c
index 3256e36ad251..accfd7bfe387 100644
--- a/kernel/perf_event.c
+++ b/kernel/perf_event.c
@@ -29,6 +29,7 @@
 #include <linux/kernel_stat.h>
 #include <linux/perf_event.h>
 #include <linux/ftrace_event.h>
+#include <linux/hw_breakpoint.h>
 
 #include <asm/irq_regs.h>
 
@@ -245,6 +246,49 @@ static void perf_unpin_context(struct perf_event_context *ctx)
 	put_ctx(ctx);
 }
 
+static inline u64 perf_clock(void)
+{
+	return cpu_clock(smp_processor_id());
+}
+
+/*
+ * Update the record of the current time in a context.
+ */
+static void update_context_time(struct perf_event_context *ctx)
+{
+	u64 now = perf_clock();
+
+	ctx->time += now - ctx->timestamp;
+	ctx->timestamp = now;
+}
+
+/*
+ * Update the total_time_enabled and total_time_running fields for a event.
+ */
+static void update_event_times(struct perf_event *event)
+{
+	struct perf_event_context *ctx = event->ctx;
+	u64 run_end;
+
+	if (event->state < PERF_EVENT_STATE_INACTIVE ||
+	    event->group_leader->state < PERF_EVENT_STATE_INACTIVE)
+		return;
+
+	if (ctx->is_active)
+		run_end = ctx->time;
+	else
+		run_end = event->tstamp_stopped;
+
+	event->total_time_enabled = run_end - event->tstamp_enabled;
+
+	if (event->state == PERF_EVENT_STATE_INACTIVE)
+		run_end = event->tstamp_stopped;
+	else
+		run_end = ctx->time;
+
+	event->total_time_running = run_end - event->tstamp_running;
+}
+
 /*
  * Add a event from the lists for its context.
  * Must be called with ctx->mutex and ctx->lock held.
@@ -293,6 +337,9 @@ list_del_event(struct perf_event *event, struct perf_event_context *ctx)
 	if (event->group_leader != event)
 		event->group_leader->nr_siblings--;
 
+	update_event_times(event);
+	event->state = PERF_EVENT_STATE_OFF;
+
 	/*
 	 * If this was a group event with sibling events then
 	 * upgrade the siblings to singleton events by adding them
@@ -446,50 +493,11 @@ retry:
 	 * can remove the event safely, if the call above did not
 	 * succeed.
 	 */
-	if (!list_empty(&event->group_entry)) {
+	if (!list_empty(&event->group_entry))
 		list_del_event(event, ctx);
-	}
 	spin_unlock_irq(&ctx->lock);
 }
 
-static inline u64 perf_clock(void)
-{
-	return cpu_clock(smp_processor_id());
-}
-
-/*
- * Update the record of the current time in a context.
- */
-static void update_context_time(struct perf_event_context *ctx)
-{
-	u64 now = perf_clock();
-
-	ctx->time += now - ctx->timestamp;
-	ctx->timestamp = now;
-}
-
-/*
- * Update the total_time_enabled and total_time_running fields for a event.
- */
-static void update_event_times(struct perf_event *event)
-{
-	struct perf_event_context *ctx = event->ctx;
-	u64 run_end;
-
-	if (event->state < PERF_EVENT_STATE_INACTIVE ||
-	    event->group_leader->state < PERF_EVENT_STATE_INACTIVE)
-		return;
-
-	event->total_time_enabled = ctx->time - event->tstamp_enabled;
-
-	if (event->state == PERF_EVENT_STATE_INACTIVE)
-		run_end = event->tstamp_stopped;
-	else
-		run_end = ctx->time;
-
-	event->total_time_running = run_end - event->tstamp_running;
-}
-
 /*
  * Update total_time_enabled and total_time_running for all events in a group.
  */
@@ -1032,10 +1040,10 @@ void __perf_event_sched_out(struct perf_event_context *ctx,
 	update_context_time(ctx);
 
 	perf_disable();
-	if (ctx->nr_active)
+	if (ctx->nr_active) {
 		list_for_each_entry(event, &ctx->group_list, group_entry)
 			group_sched_out(event, cpuctx, ctx);
-
+	}
 	perf_enable();
  out:
 	spin_unlock(&ctx->lock);
@@ -1060,8 +1068,6 @@ static int context_equiv(struct perf_event_context *ctx1,
 		&& !ctx1->pin_count && !ctx2->pin_count;
 }
 
-static void __perf_event_read(void *event);
-
 static void __perf_event_sync_stat(struct perf_event *event,
 				     struct perf_event *next_event)
 {
@@ -1079,8 +1085,8 @@ static void __perf_event_sync_stat(struct perf_event *event,
 	 */
 	switch (event->state) {
 	case PERF_EVENT_STATE_ACTIVE:
-		__perf_event_read(event);
-		break;
+		event->pmu->read(event);
+		/* fall-through */
 
 	case PERF_EVENT_STATE_INACTIVE:
 		update_event_times(event);
@@ -1119,6 +1125,8 @@ static void perf_event_sync_stat(struct perf_event_context *ctx,
 	if (!ctx->nr_stat)
 		return;
 
+	update_context_time(ctx);
+
 	event = list_first_entry(&ctx->event_list,
 				   struct perf_event, event_entry);
 
@@ -1162,8 +1170,6 @@ void perf_event_task_sched_out(struct task_struct *task,
 	if (likely(!ctx || !cpuctx->task_ctx))
 		return;
 
-	update_context_time(ctx);
-
 	rcu_read_lock();
 	parent = rcu_dereference(ctx->parent_ctx);
 	next_ctx = next->perf_event_ctxp;
@@ -1516,7 +1522,6 @@ static void __perf_event_read(void *info)
 	struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context);
 	struct perf_event *event = info;
 	struct perf_event_context *ctx = event->ctx;
-	unsigned long flags;
 
 	/*
 	 * If this is a task context, we need to check whether it is
@@ -1528,12 +1533,12 @@ static void __perf_event_read(void *info)
 	if (ctx->task && cpuctx->task_ctx != ctx)
 		return;
 
-	local_irq_save(flags);
-	if (ctx->is_active)
-		update_context_time(ctx);
-	event->pmu->read(event);
+	spin_lock(&ctx->lock);
+	update_context_time(ctx);
 	update_event_times(event);
-	local_irq_restore(flags);
+	spin_unlock(&ctx->lock);
+
+	event->pmu->read(event);
 }
 
 static u64 perf_event_read(struct perf_event *event)
@@ -1546,7 +1551,13 @@ static u64 perf_event_read(struct perf_event *event)
 		smp_call_function_single(event->oncpu,
 					 __perf_event_read, event, 1);
 	} else if (event->state == PERF_EVENT_STATE_INACTIVE) {
+		struct perf_event_context *ctx = event->ctx;
+		unsigned long flags;
+
+		spin_lock_irqsave(&ctx->lock, flags);
+		update_context_time(ctx);
 		update_event_times(event);
+		spin_unlock_irqrestore(&ctx->lock, flags);
 	}
 
 	return atomic64_read(&event->count);
@@ -1700,16 +1711,10 @@ static void free_event(struct perf_event *event)
 	call_rcu(&event->rcu_head, free_event_rcu);
 }
 
-/*
- * Called when the last reference to the file is gone.
- */
-static int perf_release(struct inode *inode, struct file *file)
+int perf_event_release_kernel(struct perf_event *event)
 {
-	struct perf_event *event = file->private_data;
 	struct perf_event_context *ctx = event->ctx;
 
-	file->private_data = NULL;
-
 	WARN_ON_ONCE(ctx->parent_ctx);
 	mutex_lock(&ctx->mutex);
 	perf_event_remove_from_context(event);
@@ -1724,6 +1729,19 @@ static int perf_release(struct inode *inode, struct file *file)
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(perf_event_release_kernel);
+
+/*
+ * Called when the last reference to the file is gone.
+ */
+static int perf_release(struct inode *inode, struct file *file)
+{
+	struct perf_event *event = file->private_data;
+
+	file->private_data = NULL;
+
+	return perf_event_release_kernel(event);
+}
 
 static int perf_event_read_size(struct perf_event *event)
 {
@@ -1750,91 +1768,94 @@ static int perf_event_read_size(struct perf_event *event)
 	return size;
 }
 
-static u64 perf_event_read_value(struct perf_event *event)
+u64 perf_event_read_value(struct perf_event *event, u64 *enabled, u64 *running)
 {
 	struct perf_event *child;
 	u64 total = 0;
 
+	*enabled = 0;
+	*running = 0;
+
+	mutex_lock(&event->child_mutex);
 	total += perf_event_read(event);
-	list_for_each_entry(child, &event->child_list, child_list)
+	*enabled += event->total_time_enabled +
+			atomic64_read(&event->child_total_time_enabled);
+	*running += event->total_time_running +
+			atomic64_read(&event->child_total_time_running);
+
+	list_for_each_entry(child, &event->child_list, child_list) {
 		total += perf_event_read(child);
+		*enabled += child->total_time_enabled;
+		*running += child->total_time_running;
+	}
+	mutex_unlock(&event->child_mutex);
 
 	return total;
 }
-
-static int perf_event_read_entry(struct perf_event *event,
-				   u64 read_format, char __user *buf)
-{
-	int n = 0, count = 0;
-	u64 values[2];
-
-	values[n++] = perf_event_read_value(event);
-	if (read_format & PERF_FORMAT_ID)
-		values[n++] = primary_event_id(event);
-
-	count = n * sizeof(u64);
-
-	if (copy_to_user(buf, values, count))
-		return -EFAULT;
-
-	return count;
-}
+EXPORT_SYMBOL_GPL(perf_event_read_value);
 
 static int perf_event_read_group(struct perf_event *event,
 				   u64 read_format, char __user *buf)
 {
 	struct perf_event *leader = event->group_leader, *sub;
-	int n = 0, size = 0, err = -EFAULT;
-	u64 values[3];
+	int n = 0, size = 0, ret = -EFAULT;
+	struct perf_event_context *ctx = leader->ctx;
+	u64 values[5];
+	u64 count, enabled, running;
+
+	mutex_lock(&ctx->mutex);
+	count = perf_event_read_value(leader, &enabled, &running);
 
 	values[n++] = 1 + leader->nr_siblings;
-	if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED) {
-		values[n++] = leader->total_time_enabled +
-			atomic64_read(&leader->child_total_time_enabled);
-	}
-	if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING) {
-		values[n++] = leader->total_time_running +
-			atomic64_read(&leader->child_total_time_running);
-	}
+	if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
+		values[n++] = enabled;
+	if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
+		values[n++] = running;
+	values[n++] = count;
+	if (read_format & PERF_FORMAT_ID)
+		values[n++] = primary_event_id(leader);
 
 	size = n * sizeof(u64);
 
 	if (copy_to_user(buf, values, size))
-		return -EFAULT;
-
-	err = perf_event_read_entry(leader, read_format, buf + size);
-	if (err < 0)
-		return err;
+		goto unlock;
 
-	size += err;
+	ret = size;
 
 	list_for_each_entry(sub, &leader->sibling_list, group_entry) {
-		err = perf_event_read_entry(sub, read_format,
-				buf + size);
-		if (err < 0)
-			return err;
+		n = 0;
 
-		size += err;
+		values[n++] = perf_event_read_value(sub, &enabled, &running);
+		if (read_format & PERF_FORMAT_ID)
+			values[n++] = primary_event_id(sub);
+
+		size = n * sizeof(u64);
+
+		if (copy_to_user(buf + ret, values, size)) {
+			ret = -EFAULT;
+			goto unlock;
+		}
+
+		ret += size;
 	}
+unlock:
+	mutex_unlock(&ctx->mutex);
 
-	return size;
+	return ret;
 }
 
 static int perf_event_read_one(struct perf_event *event,
 				 u64 read_format, char __user *buf)
 {
+	u64 enabled, running;
 	u64 values[4];
 	int n = 0;
 
-	values[n++] = perf_event_read_value(event);
-	if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED) {
-		values[n++] = event->total_time_enabled +
-			atomic64_read(&event->child_total_time_enabled);
-	}
-	if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING) {
-		values[n++] = event->total_time_running +
-			atomic64_read(&event->child_total_time_running);
-	}
+	values[n++] = perf_event_read_value(event, &enabled, &running);
+	if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
+		values[n++] = enabled;
+	if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
+		values[n++] = running;
 	if (read_format & PERF_FORMAT_ID)
 		values[n++] = primary_event_id(event);
 
@@ -1865,12 +1886,10 @@ perf_read_hw(struct perf_event *event, char __user *buf, size_t count)
 		return -ENOSPC;
 
 	WARN_ON_ONCE(event->ctx->parent_ctx);
-	mutex_lock(&event->child_mutex);
 	if (read_format & PERF_FORMAT_GROUP)
 		ret = perf_event_read_group(event, read_format, buf);
 	else
 		ret = perf_event_read_one(event, read_format, buf);
-	mutex_unlock(&event->child_mutex);
 
 	return ret;
 }
@@ -2315,7 +2334,7 @@ perf_mmap_data_init(struct perf_event *event, struct perf_mmap_data *data)
 	}
 
 	if (!data->watermark)
-		data->watermark = max_t(long, PAGE_SIZE, max_size / 2);
+		data->watermark = max_size / 2;
 
 
 	rcu_assign_pointer(event->data, data);
@@ -3245,15 +3264,10 @@ static void perf_event_task_ctx(struct perf_event_context *ctx,
 {
 	struct perf_event *event;
 
-	if (system_state != SYSTEM_RUNNING || list_empty(&ctx->event_list))
-		return;
-
-	rcu_read_lock();
 	list_for_each_entry_rcu(event, &ctx->event_list, event_entry) {
 		if (perf_event_task_match(event))
 			perf_event_task_output(event, task_event);
 	}
-	rcu_read_unlock();
 }
 
 static void perf_event_task_event(struct perf_task_event *task_event)
@@ -3261,11 +3275,11 @@ static void perf_event_task_event(struct perf_task_event *task_event)
 	struct perf_cpu_context *cpuctx;
 	struct perf_event_context *ctx = task_event->task_ctx;
 
+	rcu_read_lock();
 	cpuctx = &get_cpu_var(perf_cpu_context);
 	perf_event_task_ctx(&cpuctx->ctx, task_event);
 	put_cpu_var(perf_cpu_context);
 
-	rcu_read_lock();
 	if (!ctx)
 		ctx = rcu_dereference(task_event->task->perf_event_ctxp);
 	if (ctx)
@@ -3357,15 +3371,10 @@ static void perf_event_comm_ctx(struct perf_event_context *ctx,
 {
 	struct perf_event *event;
 
-	if (system_state != SYSTEM_RUNNING || list_empty(&ctx->event_list))
-		return;
-
-	rcu_read_lock();
 	list_for_each_entry_rcu(event, &ctx->event_list, event_entry) {
 		if (perf_event_comm_match(event))
 			perf_event_comm_output(event, comm_event);
 	}
-	rcu_read_unlock();
 }
 
 static void perf_event_comm_event(struct perf_comm_event *comm_event)
@@ -3376,7 +3385,7 @@ static void perf_event_comm_event(struct perf_comm_event *comm_event)
 	char comm[TASK_COMM_LEN];
 
 	memset(comm, 0, sizeof(comm));
-	strncpy(comm, comm_event->task->comm, sizeof(comm));
+	strlcpy(comm, comm_event->task->comm, sizeof(comm));
 	size = ALIGN(strlen(comm)+1, sizeof(u64));
 
 	comm_event->comm = comm;
@@ -3384,11 +3393,11 @@ static void perf_event_comm_event(struct perf_comm_event *comm_event)
 
 	comm_event->event_id.header.size = sizeof(comm_event->event_id) + size;
 
+	rcu_read_lock();
 	cpuctx = &get_cpu_var(perf_cpu_context);
 	perf_event_comm_ctx(&cpuctx->ctx, comm_event);
 	put_cpu_var(perf_cpu_context);
 
-	rcu_read_lock();
 	/*
 	 * doesn't really matter which of the child contexts the
 	 * events ends up in.
@@ -3481,15 +3490,10 @@ static void perf_event_mmap_ctx(struct perf_event_context *ctx,
 {
 	struct perf_event *event;
 
-	if (system_state != SYSTEM_RUNNING || list_empty(&ctx->event_list))
-		return;
-
-	rcu_read_lock();
 	list_for_each_entry_rcu(event, &ctx->event_list, event_entry) {
 		if (perf_event_mmap_match(event, mmap_event))
 			perf_event_mmap_output(event, mmap_event);
 	}
-	rcu_read_unlock();
 }
 
 static void perf_event_mmap_event(struct perf_mmap_event *mmap_event)
@@ -3545,11 +3549,11 @@ got_name:
 
 	mmap_event->event_id.header.size = sizeof(mmap_event->event_id) + size;
 
+	rcu_read_lock();
 	cpuctx = &get_cpu_var(perf_cpu_context);
 	perf_event_mmap_ctx(&cpuctx->ctx, mmap_event);
 	put_cpu_var(perf_cpu_context);
 
-	rcu_read_lock();
 	/*
 	 * doesn't really matter which of the child contexts the
 	 * events ends up in.
@@ -3688,7 +3692,11 @@ static int __perf_event_overflow(struct perf_event *event, int nmi,
 			perf_event_disable(event);
 	}
 
-	perf_event_output(event, nmi, data, regs);
+	if (event->overflow_handler)
+		event->overflow_handler(event, nmi, data, regs);
+	else
+		perf_event_output(event, nmi, data, regs);
+
 	return ret;
 }
 
@@ -3733,16 +3741,16 @@ again:
 	return nr;
 }
 
-static void perf_swevent_overflow(struct perf_event *event,
+static void perf_swevent_overflow(struct perf_event *event, u64 overflow,
 				    int nmi, struct perf_sample_data *data,
 				    struct pt_regs *regs)
 {
 	struct hw_perf_event *hwc = &event->hw;
 	int throttle = 0;
-	u64 overflow;
 
 	data->period = event->hw.last_period;
-	overflow = perf_swevent_set_period(event);
+	if (!overflow)
+		overflow = perf_swevent_set_period(event);
 
 	if (hwc->interrupts == MAX_INTERRUPTS)
 		return;
@@ -3775,14 +3783,19 @@ static void perf_swevent_add(struct perf_event *event, u64 nr,
 
 	atomic64_add(nr, &event->count);
 
+	if (!regs)
+		return;
+
 	if (!hwc->sample_period)
 		return;
 
-	if (!regs)
+	if (nr == 1 && hwc->sample_period == 1 && !event->attr.freq)
+		return perf_swevent_overflow(event, 1, nmi, data, regs);
+
+	if (atomic64_add_negative(nr, &hwc->period_left))
 		return;
 
-	if (!atomic64_add_negative(nr, &hwc->period_left))
-		perf_swevent_overflow(event, nmi, data, regs);
+	perf_swevent_overflow(event, 0, nmi, data, regs);
 }
 
 static int perf_swevent_is_counting(struct perf_event *event)
@@ -3818,6 +3831,20 @@ static int perf_swevent_is_counting(struct perf_event *event)
 static int perf_tp_event_match(struct perf_event *event,
 				struct perf_sample_data *data);
 
+static int perf_exclude_event(struct perf_event *event,
+			      struct pt_regs *regs)
+{
+	if (regs) {
+		if (event->attr.exclude_user && user_mode(regs))
+			return 1;
+
+		if (event->attr.exclude_kernel && !user_mode(regs))
+			return 1;
+	}
+
+	return 0;
+}
+
 static int perf_swevent_match(struct perf_event *event,
 				enum perf_type_id type,
 				u32 event_id,
@@ -3829,16 +3856,12 @@ static int perf_swevent_match(struct perf_event *event,
 
 	if (event->attr.type != type)
 		return 0;
+
 	if (event->attr.config != event_id)
 		return 0;
 
-	if (regs) {
-		if (event->attr.exclude_user && user_mode(regs))
-			return 0;
-
-		if (event->attr.exclude_kernel && !user_mode(regs))
-			return 0;
-	}
+	if (perf_exclude_event(event, regs))
+		return 0;
 
 	if (event->attr.type == PERF_TYPE_TRACEPOINT &&
 	    !perf_tp_event_match(event, data))
@@ -3855,49 +3878,59 @@ static void perf_swevent_ctx_event(struct perf_event_context *ctx,
 {
 	struct perf_event *event;
 
-	if (system_state != SYSTEM_RUNNING || list_empty(&ctx->event_list))
-		return;
-
-	rcu_read_lock();
 	list_for_each_entry_rcu(event, &ctx->event_list, event_entry) {
 		if (perf_swevent_match(event, type, event_id, data, regs))
 			perf_swevent_add(event, nr, nmi, data, regs);
 	}
-	rcu_read_unlock();
 }
 
-static int *perf_swevent_recursion_context(struct perf_cpu_context *cpuctx)
+int perf_swevent_get_recursion_context(void)
 {
+	struct perf_cpu_context *cpuctx = &get_cpu_var(perf_cpu_context);
+	int rctx;
+
 	if (in_nmi())
-		return &cpuctx->recursion[3];
+		rctx = 3;
+	else if (in_irq())
+		rctx = 2;
+	else if (in_softirq())
+		rctx = 1;
+	else
+		rctx = 0;
+
+	if (cpuctx->recursion[rctx]) {
+		put_cpu_var(perf_cpu_context);
+		return -1;
+	}
 
-	if (in_irq())
-		return &cpuctx->recursion[2];
+	cpuctx->recursion[rctx]++;
+	barrier();
 
-	if (in_softirq())
-		return &cpuctx->recursion[1];
+	return rctx;
+}
+EXPORT_SYMBOL_GPL(perf_swevent_get_recursion_context);
 
-	return &cpuctx->recursion[0];
+void perf_swevent_put_recursion_context(int rctx)
+{
+	struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context);
+	barrier();
+	cpuctx->recursion[rctx]++;
+	put_cpu_var(perf_cpu_context);
 }
+EXPORT_SYMBOL_GPL(perf_swevent_put_recursion_context);
 
 static void do_perf_sw_event(enum perf_type_id type, u32 event_id,
 				    u64 nr, int nmi,
 				    struct perf_sample_data *data,
 				    struct pt_regs *regs)
 {
-	struct perf_cpu_context *cpuctx = &get_cpu_var(perf_cpu_context);
-	int *recursion = perf_swevent_recursion_context(cpuctx);
+	struct perf_cpu_context *cpuctx;
 	struct perf_event_context *ctx;
 
-	if (*recursion)
-		goto out;
-
-	(*recursion)++;
-	barrier();
-
+	cpuctx = &__get_cpu_var(perf_cpu_context);
+	rcu_read_lock();
 	perf_swevent_ctx_event(&cpuctx->ctx, type, event_id,
 				 nr, nmi, data, regs);
-	rcu_read_lock();
 	/*
 	 * doesn't really matter which of the child contexts the
 	 * events ends up in.
@@ -3906,23 +3939,24 @@ static void do_perf_sw_event(enum perf_type_id type, u32 event_id,
 	if (ctx)
 		perf_swevent_ctx_event(ctx, type, event_id, nr, nmi, data, regs);
 	rcu_read_unlock();
-
-	barrier();
-	(*recursion)--;
-
-out:
-	put_cpu_var(perf_cpu_context);
 }
 
 void __perf_sw_event(u32 event_id, u64 nr, int nmi,
 			    struct pt_regs *regs, u64 addr)
 {
-	struct perf_sample_data data = {
-		.addr = addr,
-	};
+	struct perf_sample_data data;
+	int rctx;
 
-	do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, nmi,
-				&data, regs);
+	rctx = perf_swevent_get_recursion_context();
+	if (rctx < 0)
+		return;
+
+	data.addr = addr;
+	data.raw  = NULL;
+
+	do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, nmi, &data, regs);
+
+	perf_swevent_put_recursion_context(rctx);
 }
 
 static void perf_swevent_read(struct perf_event *event)
@@ -4145,6 +4179,7 @@ void perf_tp_event(int event_id, u64 addr, u64 count, void *record,
 	if (!regs)
 		regs = task_pt_regs(current);
 
+	/* Trace events already protected against recursion */
 	do_perf_sw_event(PERF_TYPE_TRACEPOINT, event_id, count, 1,
 				&data, regs);
 }
@@ -4231,6 +4266,57 @@ static void perf_event_free_filter(struct perf_event *event)
 
 #endif /* CONFIG_EVENT_PROFILE */
 
+#ifdef CONFIG_HAVE_HW_BREAKPOINT
+static void bp_perf_event_destroy(struct perf_event *event)
+{
+	release_bp_slot(event);
+}
+
+static const struct pmu *bp_perf_event_init(struct perf_event *bp)
+{
+	int err;
+	/*
+	 * The breakpoint is already filled if we haven't created the counter
+	 * through perf syscall
+	 * FIXME: manage to get trigerred to NULL if it comes from syscalls
+	 */
+	if (!bp->callback)
+		err = register_perf_hw_breakpoint(bp);
+	else
+		err = __register_perf_hw_breakpoint(bp);
+	if (err)
+		return ERR_PTR(err);
+
+	bp->destroy = bp_perf_event_destroy;
+
+	return &perf_ops_bp;
+}
+
+void perf_bp_event(struct perf_event *bp, void *data)
+{
+	struct perf_sample_data sample;
+	struct pt_regs *regs = data;
+
+	sample.addr = bp->attr.bp_addr;
+
+	if (!perf_exclude_event(bp, regs))
+		perf_swevent_add(bp, 1, 1, &sample, regs);
+}
+#else
+static void bp_perf_event_destroy(struct perf_event *event)
+{
+}
+
+static const struct pmu *bp_perf_event_init(struct perf_event *bp)
+{
+	return NULL;
+}
+
+void perf_bp_event(struct perf_event *bp, void *regs)
+{
+}
+#endif
+
 atomic_t perf_swevent_enabled[PERF_COUNT_SW_MAX];
 
 static void sw_perf_event_destroy(struct perf_event *event)
@@ -4297,6 +4383,7 @@ perf_event_alloc(struct perf_event_attr *attr,
 		   struct perf_event_context *ctx,
 		   struct perf_event *group_leader,
 		   struct perf_event *parent_event,
+		   perf_callback_t callback,
 		   gfp_t gfpflags)
 {
 	const struct pmu *pmu;
@@ -4339,6 +4426,11 @@ perf_event_alloc(struct perf_event_attr *attr,
 
 	event->state		= PERF_EVENT_STATE_INACTIVE;
 
+	if (!callback && parent_event)
+		callback = parent_event->callback;
+	
+	event->callback	= callback;
+
 	if (attr->disabled)
 		event->state = PERF_EVENT_STATE_OFF;
 
@@ -4373,6 +4465,11 @@ perf_event_alloc(struct perf_event_attr *attr,
 		pmu = tp_perf_event_init(event);
 		break;
 
+	case PERF_TYPE_BREAKPOINT:
+		pmu = bp_perf_event_init(event);
+		break;
+
+
 	default:
 		break;
 	}
@@ -4615,7 +4712,7 @@ SYSCALL_DEFINE5(perf_event_open,
 	}
 
 	event = perf_event_alloc(&attr, cpu, ctx, group_leader,
-				     NULL, GFP_KERNEL);
+				     NULL, NULL, GFP_KERNEL);
 	err = PTR_ERR(event);
 	if (IS_ERR(event))
 		goto err_put_context;
@@ -4663,6 +4760,58 @@ err_put_context:
 	return err;
 }
 
+/**
+ * perf_event_create_kernel_counter
+ *
+ * @attr: attributes of the counter to create
+ * @cpu: cpu in which the counter is bound
+ * @pid: task to profile
+ */
+struct perf_event *
+perf_event_create_kernel_counter(struct perf_event_attr *attr, int cpu,
+				 pid_t pid, perf_callback_t callback)
+{
+	struct perf_event *event;
+	struct perf_event_context *ctx;
+	int err;
+
+	/*
+	 * Get the target context (task or percpu):
+	 */
+
+	ctx = find_get_context(pid, cpu);
+	if (IS_ERR(ctx))
+		return NULL;
+
+	event = perf_event_alloc(attr, cpu, ctx, NULL,
+				     NULL, callback, GFP_KERNEL);
+	err = PTR_ERR(event);
+	if (IS_ERR(event))
+		goto err_put_context;
+
+	event->filp = NULL;
+	WARN_ON_ONCE(ctx->parent_ctx);
+	mutex_lock(&ctx->mutex);
+	perf_install_in_context(ctx, event, cpu);
+	++ctx->generation;
+	mutex_unlock(&ctx->mutex);
+
+	event->owner = current;
+	get_task_struct(current);
+	mutex_lock(&current->perf_event_mutex);
+	list_add_tail(&event->owner_entry, &current->perf_event_list);
+	mutex_unlock(&current->perf_event_mutex);
+
+	return event;
+
+err_put_context:
+	if (err < 0)
+		put_ctx(ctx);
+
+	return NULL;
+}
+EXPORT_SYMBOL_GPL(perf_event_create_kernel_counter);
+
 /*
  * inherit a event from parent task to child task:
  */
@@ -4688,7 +4837,7 @@ inherit_event(struct perf_event *parent_event,
 	child_event = perf_event_alloc(&parent_event->attr,
 					   parent_event->cpu, child_ctx,
 					   group_leader, parent_event,
-					   GFP_KERNEL);
+					   NULL, GFP_KERNEL);
 	if (IS_ERR(child_event))
 		return child_event;
 	get_ctx(child_ctx);
@@ -4706,6 +4855,8 @@ inherit_event(struct perf_event *parent_event,
 	if (parent_event->attr.freq)
 		child_event->hw.sample_period = parent_event->hw.sample_period;
 
+	child_event->overflow_handler = parent_event->overflow_handler;
+
 	/*
 	 * Link it up in the child's context:
 	 */
@@ -4795,7 +4946,6 @@ __perf_event_exit_task(struct perf_event *child_event,
 {
 	struct perf_event *parent_event;
 
-	update_event_times(child_event);
 	perf_event_remove_from_context(child_event);
 
 	parent_event = child_event->parent;
@@ -4847,6 +4997,7 @@ void perf_event_exit_task(struct task_struct *child)
 	 * the events from it.
 	 */
 	unclone_ctx(child_ctx);
+	update_context_time(child_ctx);
 	spin_unlock_irqrestore(&child_ctx->lock, flags);
 
 	/*