Merge branch 'perf/core' into perf/urgent

Merge reason: Switch from pre-merge topical split to the post-merge urgent track

Signed-off-by: Ingo Molnar <mingo@elte.hu>

18 files changed, 666 insertions, 517 deletions

diff --git a/kernel/futex.c b/kernel/futex.c
index d9b3a2228f9..e7a35f1039e 100644
--- a/kernel/futex.c
+++ b/kernel/futex.c
@@ -530,8 +530,25 @@ lookup_pi_state(u32 uval, struct futex_hash_bucket *hb,
 				return -EINVAL;
 
 			WARN_ON(!atomic_read(&pi_state->refcount));
-			WARN_ON(pid && pi_state->owner &&
-				pi_state->owner->pid != pid);
+
+			/*
+			 * When pi_state->owner is NULL then the owner died
+			 * and another waiter is on the fly. pi_state->owner
+			 * is fixed up by the task which acquires
+			 * pi_state->rt_mutex.
+			 *
+			 * We do not check for pid == 0 which can happen when
+			 * the owner died and robust_list_exit() cleared the
+			 * TID.
+			 */
+			if (pid && pi_state->owner) {
+				/*
+				 * Bail out if user space manipulated the
+				 * futex value.
+				 */
+				if (pid != task_pid_vnr(pi_state->owner))
+					return -EINVAL;
+			}
 
 			atomic_inc(&pi_state->refcount);
 			*ps = pi_state;
@@ -758,6 +775,13 @@ static int wake_futex_pi(u32 __user *uaddr, u32 uval, struct futex_q *this)
 	if (!pi_state)
 		return -EINVAL;
 
+	/*
+	 * If current does not own the pi_state then the futex is
+	 * inconsistent and user space fiddled with the futex value.
+	 */
+	if (pi_state->owner != current)
+		return -EINVAL;
+
 	raw_spin_lock(&pi_state->pi_mutex.wait_lock);
 	new_owner = rt_mutex_next_owner(&pi_state->pi_mutex);
 
@@ -1971,7 +1995,7 @@ retry_private:
 	/* Unqueue and drop the lock */
 	unqueue_me_pi(&q);
 
-	goto out;
+	goto out_put_key;
 
 out_unlock_put_key:
 	queue_unlock(&q, hb);
diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c
index 4d99512ee14..03808ed342a 100644
--- a/kernel/hw_breakpoint.c
+++ b/kernel/hw_breakpoint.c
@@ -413,17 +413,17 @@ EXPORT_SYMBOL_GPL(unregister_hw_breakpoint);
  *
  * @return a set of per_cpu pointers to perf events
  */
-struct perf_event **
+struct perf_event * __percpu *
 register_wide_hw_breakpoint(struct perf_event_attr *attr,
 			    perf_overflow_handler_t triggered)
 {
-	struct perf_event **cpu_events, **pevent, *bp;
+	struct perf_event * __percpu *cpu_events, **pevent, *bp;
 	long err;
 	int cpu;
 
 	cpu_events = alloc_percpu(typeof(*cpu_events));
 	if (!cpu_events)
-		return ERR_PTR(-ENOMEM);
+		return (void __percpu __force *)ERR_PTR(-ENOMEM);
 
 	get_online_cpus();
 	for_each_online_cpu(cpu) {
@@ -451,7 +451,7 @@ fail:
 	put_online_cpus();
 
 	free_percpu(cpu_events);
-	return ERR_PTR(err);
+	return (void __percpu __force *)ERR_PTR(err);
 }
 EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint);
 
@@ -459,7 +459,7 @@ EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint);
  * unregister_wide_hw_breakpoint - unregister a wide breakpoint in the kernel
  * @cpu_events: the per cpu set of events to unregister
  */
-void unregister_wide_hw_breakpoint(struct perf_event **cpu_events)
+void unregister_wide_hw_breakpoint(struct perf_event * __percpu *cpu_events)
 {
 	int cpu;
 	struct perf_event **pevent;
diff --git a/kernel/kfifo.c b/kernel/kfifo.c
index 498cabba225..35edbe22e9a 100644
--- a/kernel/kfifo.c
+++ b/kernel/kfifo.c
@@ -80,7 +80,7 @@ int kfifo_alloc(struct kfifo *fifo, unsigned int size, gfp_t gfp_mask)
 
 	buffer = kmalloc(size, gfp_mask);
 	if (!buffer) {
-		_kfifo_init(fifo, 0, 0);
+		_kfifo_init(fifo, NULL, 0);
 		return -ENOMEM;
 	}
 
@@ -97,6 +97,7 @@ EXPORT_SYMBOL(kfifo_alloc);
 void kfifo_free(struct kfifo *fifo)
 {
 	kfree(fifo->buffer);
+	_kfifo_init(fifo, NULL, 0);
 }
 EXPORT_SYMBOL(kfifo_free);
 
diff --git a/kernel/kgdb.c b/kernel/kgdb.c
index c7ade62e4ef..761fdd2b303 100644
--- a/kernel/kgdb.c
+++ b/kernel/kgdb.c
@@ -599,7 +599,7 @@ static void kgdb_wait(struct pt_regs *regs)
 
 	/* Signal the primary CPU that we are done: */
 	atomic_set(&cpu_in_kgdb[cpu], 0);
-	touch_softlockup_watchdog();
+	touch_softlockup_watchdog_sync();
 	clocksource_touch_watchdog();
 	local_irq_restore(flags);
 }
@@ -1453,7 +1453,7 @@ acquirelock:
 	    (kgdb_info[cpu].task &&
 	     kgdb_info[cpu].task->pid != kgdb_sstep_pid) && --sstep_tries) {
 		atomic_set(&kgdb_active, -1);
-		touch_softlockup_watchdog();
+		touch_softlockup_watchdog_sync();
 		clocksource_touch_watchdog();
 		local_irq_restore(flags);
 
@@ -1553,7 +1553,7 @@ kgdb_restore:
 	}
 	/* Free kgdb_active */
 	atomic_set(&kgdb_active, -1);
-	touch_softlockup_watchdog();
+	touch_softlockup_watchdog_sync();
 	clocksource_touch_watchdog();
 	local_irq_restore(flags);
 
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index b7df302a020..ccec774c716 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -44,6 +44,7 @@
 #include <linux/debugfs.h>
 #include <linux/kdebug.h>
 #include <linux/memory.h>
+#include <linux/ftrace.h>
 
 #include <asm-generic/sections.h>
 #include <asm/cacheflush.h>
@@ -93,6 +94,7 @@ static struct kprobe_blackpoint kprobe_blacklist[] = {
 	{"native_get_debugreg",},
 	{"irq_entries_start",},
 	{"common_interrupt",},
+	{"mcount",},	/* mcount can be called from everywhere */
 	{NULL}    /* Terminator */
 };
 
@@ -124,30 +126,6 @@ static LIST_HEAD(kprobe_insn_pages);
 static int kprobe_garbage_slots;
 static int collect_garbage_slots(void);
 
-static int __kprobes check_safety(void)
-{
-	int ret = 0;
-#if defined(CONFIG_PREEMPT) && defined(CONFIG_FREEZER)
-	ret = freeze_processes();
-	if (ret == 0) {
-		struct task_struct *p, *q;
-		do_each_thread(p, q) {
-			if (p != current && p->state == TASK_RUNNING &&
-			    p->pid != 0) {
-				printk("Check failed: %s is running\n",p->comm);
-				ret = -1;
-				goto loop_end;
-			}
-		} while_each_thread(p, q);
-	}
-loop_end:
-	thaw_processes();
-#else
-	synchronize_sched();
-#endif
-	return ret;
-}
-
 /**
  * __get_insn_slot() - Find a slot on an executable page for an instruction.
  * We allocate an executable page if there's no room on existing ones.
@@ -235,9 +213,8 @@ static int __kprobes collect_garbage_slots(void)
 {
 	struct kprobe_insn_page *kip, *next;
 
-	/* Ensure no-one is preepmted on the garbages */
-	if (check_safety())
-		return -EAGAIN;
+	/* Ensure no-one is interrupted on the garbages */
+	synchronize_sched();
 
 	list_for_each_entry_safe(kip, next, &kprobe_insn_pages, list) {
 		int i;
@@ -728,7 +705,8 @@ int __kprobes register_kprobe(struct kprobe *p)
 
 	preempt_disable();
 	if (!kernel_text_address((unsigned long) p->addr) ||
-	    in_kprobes_functions((unsigned long) p->addr)) {
+	    in_kprobes_functions((unsigned long) p->addr) ||
+	    ftrace_text_reserved(p->addr, p->addr)) {
 		preempt_enable();
 		return -EINVAL;
 	}
diff --git a/kernel/perf_event.c b/kernel/perf_event.c
index 2ae7409bf38..482d5e1d376 100644
--- a/kernel/perf_event.c
+++ b/kernel/perf_event.c
@@ -56,21 +56,6 @@ static atomic_t nr_task_events __read_mostly;
  */
 int sysctl_perf_event_paranoid __read_mostly = 1;
 
-static inline bool perf_paranoid_tracepoint_raw(void)
-{
-	return sysctl_perf_event_paranoid > -1;
-}
-
-static inline bool perf_paranoid_cpu(void)
-{
-	return sysctl_perf_event_paranoid > 0;
-}
-
-static inline bool perf_paranoid_kernel(void)
-{
-	return sysctl_perf_event_paranoid > 1;
-}
-
 int sysctl_perf_event_mlock __read_mostly = 512; /* 'free' kb per user */
 
 /*
@@ -98,11 +83,12 @@ void __weak hw_perf_enable(void)		{ barrier(); }
 
 void __weak hw_perf_event_setup(int cpu)	{ barrier(); }
 void __weak hw_perf_event_setup_online(int cpu)	{ barrier(); }
+void __weak hw_perf_event_setup_offline(int cpu)	{ barrier(); }
 
 int __weak
 hw_perf_group_sched_in(struct perf_event *group_leader,
 	       struct perf_cpu_context *cpuctx,
-	       struct perf_event_context *ctx, int cpu)
+	       struct perf_event_context *ctx)
 {
 	return 0;
 }
@@ -248,7 +234,7 @@ static void perf_unpin_context(struct perf_event_context *ctx)
 
 static inline u64 perf_clock(void)
 {
-	return cpu_clock(smp_processor_id());
+	return cpu_clock(raw_smp_processor_id());
 }
 
 /*
@@ -289,6 +275,15 @@ static void update_event_times(struct perf_event *event)
 	event->total_time_running = run_end - event->tstamp_running;
 }
 
+static struct list_head *
+ctx_group_list(struct perf_event *event, struct perf_event_context *ctx)
+{
+	if (event->attr.pinned)
+		return &ctx->pinned_groups;
+	else
+		return &ctx->flexible_groups;
+}
+
 /*
  * Add a event from the lists for its context.
  * Must be called with ctx->mutex and ctx->lock held.
@@ -303,9 +298,19 @@ list_add_event(struct perf_event *event, struct perf_event_context *ctx)
 	 * add it straight to the context's event list, or to the group
 	 * leader's sibling list:
 	 */
-	if (group_leader == event)
-		list_add_tail(&event->group_entry, &ctx->group_list);
-	else {
+	if (group_leader == event) {
+		struct list_head *list;
+
+		if (is_software_event(event))
+			event->group_flags |= PERF_GROUP_SOFTWARE;
+
+		list = ctx_group_list(event, ctx);
+		list_add_tail(&event->group_entry, list);
+	} else {
+		if (group_leader->group_flags & PERF_GROUP_SOFTWARE &&
+		    !is_software_event(event))
+			group_leader->group_flags &= ~PERF_GROUP_SOFTWARE;
+
 		list_add_tail(&event->group_entry, &group_leader->sibling_list);
 		group_leader->nr_siblings++;
 	}
@@ -355,9 +360,14 @@ list_del_event(struct perf_event *event, struct perf_event_context *ctx)
 	 * to the context list directly:
 	 */
 	list_for_each_entry_safe(sibling, tmp, &event->sibling_list, group_entry) {
+		struct list_head *list;
 
-		list_move_tail(&sibling->group_entry, &ctx->group_list);
+		list = ctx_group_list(event, ctx);
+		list_move_tail(&sibling->group_entry, list);
 		sibling->group_leader = sibling;
+
+		/* Inherit group flags from the previous leader */
+		sibling->group_flags = event->group_flags;
 	}
 }
 
@@ -608,14 +618,13 @@ void perf_event_disable(struct perf_event *event)
 static int
 event_sched_in(struct perf_event *event,
 		 struct perf_cpu_context *cpuctx,
-		 struct perf_event_context *ctx,
-		 int cpu)
+		 struct perf_event_context *ctx)
 {
 	if (event->state <= PERF_EVENT_STATE_OFF)
 		return 0;
 
 	event->state = PERF_EVENT_STATE_ACTIVE;
-	event->oncpu = cpu;	/* TODO: put 'cpu' into cpuctx->cpu */
+	event->oncpu = smp_processor_id();
 	/*
 	 * The new state must be visible before we turn it on in the hardware:
 	 */
@@ -642,8 +651,7 @@ event_sched_in(struct perf_event *event,
 static int
 group_sched_in(struct perf_event *group_event,
 	       struct perf_cpu_context *cpuctx,
-	       struct perf_event_context *ctx,
-	       int cpu)
+	       struct perf_event_context *ctx)
 {
 	struct perf_event *event, *partial_group;
 	int ret;
@@ -651,18 +659,18 @@ group_sched_in(struct perf_event *group_event,
 	if (group_event->state == PERF_EVENT_STATE_OFF)
 		return 0;
 
-	ret = hw_perf_group_sched_in(group_event, cpuctx, ctx, cpu);
+	ret = hw_perf_group_sched_in(group_event, cpuctx, ctx);
 	if (ret)
 		return ret < 0 ? ret : 0;
 
-	if (event_sched_in(group_event, cpuctx, ctx, cpu))
+	if (event_sched_in(group_event, cpuctx, ctx))
 		return -EAGAIN;
 
 	/*
 	 * Schedule in siblings as one group (if any):
 	 */
 	list_for_each_entry(event, &group_event->sibling_list, group_entry) {
-		if (event_sched_in(event, cpuctx, ctx, cpu)) {
+		if (event_sched_in(event, cpuctx, ctx)) {
 			partial_group = event;
 			goto group_error;
 		}
@@ -686,24 +694,6 @@ group_error:
 }
 
 /*
- * Return 1 for a group consisting entirely of software events,
- * 0 if the group contains any hardware events.
- */
-static int is_software_only_group(struct perf_event *leader)
-{
-	struct perf_event *event;
-
-	if (!is_software_event(leader))
-		return 0;
-
-	list_for_each_entry(event, &leader->sibling_list, group_entry)
-		if (!is_software_event(event))
-			return 0;
-
-	return 1;
-}
-
-/*
  * Work out whether we can put this event group on the CPU now.
  */
 static int group_can_go_on(struct perf_event *event,
@@ -713,7 +703,7 @@ static int group_can_go_on(struct perf_event *event,
 	/*
 	 * Groups consisting entirely of software events can always go on.
 	 */
-	if (is_software_only_group(event))
+	if (event->group_flags & PERF_GROUP_SOFTWARE)
 		return 1;
 	/*
 	 * If an exclusive group is already on, no other hardware
@@ -754,7 +744,6 @@ static void __perf_install_in_context(void *info)
 	struct perf_event *event = info;
 	struct perf_event_context *ctx = event->ctx;
 	struct perf_event *leader = event->group_leader;
-	int cpu = smp_processor_id();
 	int err;
 
 	/*
@@ -801,7 +790,7 @@ static void __perf_install_in_context(void *info)
 	if (!group_can_go_on(event, cpuctx, 1))
 		err = -EEXIST;
 	else
-		err = event_sched_in(event, cpuctx, ctx, cpu);
+		err = event_sched_in(event, cpuctx, ctx);
 
 	if (err) {
 		/*
@@ -943,11 +932,9 @@ static void __perf_event_enable(void *info)
 	} else {
 		perf_disable();
 		if (event == leader)
-			err = group_sched_in(event, cpuctx, ctx,
-					     smp_processor_id());
+			err = group_sched_in(event, cpuctx, ctx);
 		else
-			err = event_sched_in(event, cpuctx, ctx,
-					       smp_processor_id());
+			err = event_sched_in(event, cpuctx, ctx);
 		perf_enable();
 	}
 
@@ -1043,8 +1030,15 @@ static int perf_event_refresh(struct perf_event *event, int refresh)
 	return 0;
 }
 
-void __perf_event_sched_out(struct perf_event_context *ctx,
-			      struct perf_cpu_context *cpuctx)
+enum event_type_t {
+	EVENT_FLEXIBLE = 0x1,
+	EVENT_PINNED = 0x2,
+	EVENT_ALL = EVENT_FLEXIBLE | EVENT_PINNED,
+};
+
+static void ctx_sched_out(struct perf_event_context *ctx,
+			  struct perf_cpu_context *cpuctx,
+			  enum event_type_t event_type)
 {
 	struct perf_event *event;
 
@@ -1055,10 +1049,18 @@ void __perf_event_sched_out(struct perf_event_context *ctx,
 	update_context_time(ctx);
 
 	perf_disable();
-	if (ctx->nr_active) {
-		list_for_each_entry(event, &ctx->group_list, group_entry)
+	if (!ctx->nr_active)
+		goto out_enable;
+
+	if (event_type & EVENT_PINNED)
+		list_for_each_entry(event, &ctx->pinned_groups, group_entry)
 			group_sched_out(event, cpuctx, ctx);
-	}
+
+	if (event_type & EVENT_FLEXIBLE)
+		list_for_each_entry(event, &ctx->flexible_groups, group_entry)
+			group_sched_out(event, cpuctx, ctx);
+
+ out_enable:
 	perf_enable();
  out:
 	raw_spin_unlock(&ctx->lock);
@@ -1170,9 +1172,9 @@ static void perf_event_sync_stat(struct perf_event_context *ctx,
  * not restart the event.
  */
 void perf_event_task_sched_out(struct task_struct *task,
-				 struct task_struct *next, int cpu)
+				 struct task_struct *next)
 {
-	struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu);
+	struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context);
 	struct perf_event_context *ctx = task->perf_event_ctxp;
 	struct perf_event_context *next_ctx;
 	struct perf_event_context *parent;
@@ -1220,15 +1222,13 @@ void perf_event_task_sched_out(struct task_struct *task,
 	rcu_read_unlock();
 
 	if (do_switch) {
-		__perf_event_sched_out(ctx, cpuctx);
+		ctx_sched_out(ctx, cpuctx, EVENT_ALL);
 		cpuctx->task_ctx = NULL;
 	}
 }
 
-/*
- * Called with IRQs disabled
- */
-static void __perf_event_task_sched_out(struct perf_event_context *ctx)
+static void task_ctx_sched_out(struct perf_event_context *ctx,
+			       enum event_type_t event_type)
 {
 	struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context);
 
@@ -1238,47 +1238,41 @@ static void __perf_event_task_sched_out(struct perf_event_context *ctx)
 	if (WARN_ON_ONCE(ctx != cpuctx->task_ctx))
 		return;
 
-	__perf_event_sched_out(ctx, cpuctx);
+	ctx_sched_out(ctx, cpuctx, event_type);
 	cpuctx->task_ctx = NULL;
 }
 
 /*
  * Called with IRQs disabled
  */
-static void perf_event_cpu_sched_out(struct perf_cpu_context *cpuctx)
+static void __perf_event_task_sched_out(struct perf_event_context *ctx)
+{
+	task_ctx_sched_out(ctx, EVENT_ALL);
+}
+
+/*
+ * Called with IRQs disabled
+ */
+static void cpu_ctx_sched_out(struct perf_cpu_context *cpuctx,
+			      enum event_type_t event_type)
 {
-	__perf_event_sched_out(&cpuctx->ctx, cpuctx);
+	ctx_sched_out(&cpuctx->ctx, cpuctx, event_type);
 }
 
 static void
-__perf_event_sched_in(struct perf_event_context *ctx,
-			struct perf_cpu_context *cpuctx, int cpu)
+ctx_pinned_sched_in(struct perf_event_context *ctx,
+		    struct perf_cpu_context *cpuctx)
 {
 	struct perf_event *event;
-	int can_add_hw = 1;
-
-	raw_spin_lock(&ctx->lock);
-	ctx->is_active = 1;
-	if (likely(!ctx->nr_events))
-		goto out;
-
-	ctx->timestamp = perf_clock();
-
-	perf_disable();
 
-	/*
-	 * First go through the list and put on any pinned groups
-	 * in order to give them the best chance of going on.
-	 */
-	list_for_each_entry(event, &ctx->group_list, group_entry) {
-		if (event->state <= PERF_EVENT_STATE_OFF ||
-		    !event->attr.pinned)
+	list_for_each_entry(event, &ctx->pinned_groups, group_entry) {
+		if (event->state <= PERF_EVENT_STATE_OFF)
 			continue;
-		if (event->cpu != -1 && event->cpu != cpu)
+		if (event->cpu != -1 && event->cpu != smp_processor_id())
 			continue;
 
 		if (group_can_go_on(event, cpuctx, 1))
-			group_sched_in(event, cpuctx, ctx, cpu);
+			group_sched_in(event, cpuctx, ctx);
 
 		/*
 		 * If this pinned group hasn't been scheduled,
@@ -1289,32 +1283,83 @@ __perf_event_sched_in(struct perf_event_context *ctx,
 			event->state = PERF_EVENT_STATE_ERROR;
 		}
 	}
+}
 
-	list_for_each_entry(event, &ctx->group_list, group_entry) {
-		/*
-		 * Ignore events in OFF or ERROR state, and
-		 * ignore pinned events since we did them already.
-		 */
-		if (event->state <= PERF_EVENT_STATE_OFF ||
-		    event->attr.pinned)
-			continue;
+static void
+ctx_flexible_sched_in(struct perf_event_context *ctx,
+		      struct perf_cpu_context *cpuctx)
+{
+	struct perf_event *event;
+	int can_add_hw = 1;
 
+	list_for_each_entry(event, &ctx->flexible_groups, group_entry) {
+		/* Ignore events in OFF or ERROR state */
+		if (event->state <= PERF_EVENT_STATE_OFF)
+			continue;
 		/*
 		 * Listen to the 'cpu' scheduling filter constraint
 		 * of events:
 		 */
-		if (event->cpu != -1 && event->cpu != cpu)
+		if (event->cpu != -1 && event->cpu != smp_processor_id())
 			continue;
 
 		if (group_can_go_on(event, cpuctx, can_add_hw))
-			if (group_sched_in(event, cpuctx, ctx, cpu))
+			if (group_sched_in(event, cpuctx, ctx))
 				can_add_hw = 0;
 	}
+}
+
+static void
+ctx_sched_in(struct perf_event_context *ctx,
+	     struct perf_cpu_context *cpuctx,
+	     enum event_type_t event_type)
+{
+	raw_spin_lock(&ctx->lock);
+	ctx->is_active = 1;
+	if (likely(!ctx->nr_events))
+		goto out;
+
+	ctx->timestamp = perf_clock();
+
+	perf_disable();
+
+	/*
+	 * First go through the list and put on any pinned groups
+	 * in order to give them the best chance of going on.
+	 */
+	if (event_type & EVENT_PINNED)
+		ctx_pinned_sched_in(ctx, cpuctx);
+
+	/* Then walk through the lower prio flexible groups */
+	if (event_type & EVENT_FLEXIBLE)
+		ctx_flexible_sched_in(ctx, cpuctx);
+
 	perf_enable();
  out:
 	raw_spin_unlock(&ctx->lock);
 }
 
+static void cpu_ctx_sched_in(struct perf_cpu_context *cpuctx,
+			     enum event_type_t event_type)
+{
+	struct perf_event_context *ctx = &cpuctx->ctx;
+
+	ctx_sched_in(ctx, cpuctx, event_type);
+}
+
+static void task_ctx_sched_in(struct task_struct *task,
+			      enum event_type_t event_type)
+{
+	struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context);
+	struct perf_event_context *ctx = task->perf_event_ctxp;
+
+	if (likely(!ctx))
+		return;
+	if (cpuctx->task_ctx == ctx)
+		return;
+	ctx_sched_in(ctx, cpuctx, event_type);
+	cpuctx->task_ctx = ctx;
+}
 /*
  * Called from scheduler to add the events of the current task
  * with interrupts disabled.
@@ -1326,38 +1371,128 @@ __perf_event_sched_in(struct perf_event_context *ctx,
  * accessing the event control register. If a NMI hits, then it will
  * keep the event running.
  */
-void perf_event_task_sched_in(struct task_struct *task, int cpu)
+void perf_event_task_sched_in(struct task_struct *task)
 {
-	struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu);
+	struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context);
 	struct perf_event_context *ctx = task->perf_event_ctxp;
 
 	if (likely(!ctx))
 		return;
+
 	if (cpuctx->task_ctx == ctx)
 		return;
-	__perf_event_sched_in(ctx, cpuctx, cpu);
+
+	/*
+	 * We want to keep the following priority order:
+	 * cpu pinned (that don't need to move), task pinned,
+	 * cpu flexible, task flexible.
+	 */
+	cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE);
+
+	ctx_sched_in(ctx, cpuctx, EVENT_PINNED);
+	cpu_ctx_sched_in(cpuctx, EVENT_FLEXIBLE);
+	ctx_sched_in(ctx, cpuctx, EVENT_FLEXIBLE);
+
 	cpuctx->task_ctx = ctx;
 }
 
-static void perf_event_cpu_sched_in(struct perf_cpu_context *cpuctx, int cpu)
+#define MAX_INTERRUPTS (~0ULL)
+
+static void perf_log_throttle(struct perf_event *event, int enable);
+
+static u64 perf_calculate_period(struct perf_event *event, u64 nsec, u64 count)
 {
-	struct perf_event_context *ctx = &cpuctx->ctx;
+	u64 frequency = event->attr.sample_freq;
+	u64 sec = NSEC_PER_SEC;
+	u64 divisor, dividend;
+
+	int count_fls, nsec_fls, frequency_fls, sec_fls;
+
+	count_fls = fls64(count);
+	nsec_fls = fls64(nsec);
+	frequency_fls = fls64(frequency);
+	sec_fls = 30;
+
+	/*
+	 * We got @count in @nsec, with a target of sample_freq HZ
+	 * the target period becomes:
+	 *
+	 *             @count * 10^9
+	 * period = -------------------
+	 *          @nsec * sample_freq
+	 *
+	 */
+
+	/*
+	 * Reduce accuracy by one bit such that @a and @b converge
+	 * to a similar magnitude.
+	 */
+#define REDUCE_FLS(a, b) 		\
+do {					\
+	if (a##_fls > b##_fls) {	\
+		a >>= 1;		\
+		a##_fls--;		\
+	} else {			\
+		b >>= 1;		\
+		b##_fls--;		\
+	}				\
+} while (0)
+
+	/*
+	 * Reduce accuracy until either term fits in a u64, then proceed with
+	 * the other, so that finally we can do a u64/u64 division.
+	 */
+	while (count_fls + sec_fls > 64 && nsec_fls + frequency_fls > 64) {
+		REDUCE_FLS(nsec, frequency);
+		REDUCE_FLS(sec, count);
+	}
+
+	if (count_fls + sec_fls > 64) {
+		divisor = nsec * frequency;
+
+		while (count_fls + sec_fls > 64) {
+			REDUCE_FLS(count, sec);
+			divisor >>= 1;
+		}
 
-	__perf_event_sched_in(ctx, cpuctx, cpu);
+		dividend = count * sec;
+	} else {
+		dividend = count * sec;
+
+		while (nsec_fls + frequency_fls > 64) {
+			REDUCE_FLS(nsec, frequency);
+			dividend >>= 1;
+		}
+
+		divisor = nsec * frequency;
+	}
+
+	return div64_u64(dividend, divisor);
 }
 
-#define MAX_INTERRUPTS (~0ULL)
+static void perf_event_stop(struct perf_event *event)
+{
+	if (!event->pmu->stop)
+		return event->pmu->disable(event);
 
-static void perf_log_throttle(struct perf_event *event, int enable);
+	return event->pmu->stop(event);
+}
+
+static int perf_event_start(struct perf_event *event)
+{
+	if (!event->pmu->start)
+		return event->pmu->enable(event);
 
-static void perf_adjust_period(struct perf_event *event, u64 events)
+	return event->pmu->start(event);
+}
+
+static void perf_adjust_period(struct perf_event *event, u64 nsec, u64 count)
 {
 	struct hw_perf_event *hwc = &event->hw;
 	u64 period, sample_period;
 	s64 delta;
 
-	events *= hwc->sample_period;
-	period = div64_u64(events, event->attr.sample_freq);
+	period = perf_calculate_period(event, nsec, count);
 
 	delta = (s64)(period - hwc->sample_period);
 	delta = (delta + 7) / 8; /* low pass filter */
@@ -1368,13 +1503,22 @@ static void perf_adjust_period(struct perf_event *event, u64 events)
 		sample_period = 1;
 
 	hwc->sample_period = sample_period;
+
+	if (atomic64_read(&hwc->period_left) > 8*sample_period) {
+		perf_disable();
+		perf_event_stop(event);
+		atomic64_set(&hwc->period_left, 0);
+		perf_event_start(event);
+		perf_enable();
+	}
 }
 
 static void perf_ctx_adjust_freq(struct perf_event_context *ctx)
 {
 	struct perf_event *event;
 	struct hw_perf_event *hwc;
-	u64 interrupts, freq;
+	u64 interrupts, now;
+	s64 delta;
 
 	raw_spin_lock(&ctx->lock);
 	list_for_each_entry_rcu(event, &ctx->event_list, event_entry) {
@@ -1395,44 +1539,18 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx)
 		if (interrupts == MAX_INTERRUPTS) {
 			perf_log_throttle(event, 1);
 			event->pmu->unthrottle(event);
-			interrupts = 2*sysctl_perf_event_sample_rate/HZ;
 		}
 
 		if (!event->attr.freq || !event->attr.sample_freq)
 			continue;
 
-		/*
-		 * if the specified freq < HZ then we need to skip ticks
-		 */
-		if (event->attr.sample_freq < HZ) {
-			freq = event->attr.sample_freq;
-
-			hwc->freq_count += freq;
-			hwc->freq_interrupts += interrupts;
-
-			if (hwc->freq_count < HZ)
-				continue;
-
-			interrupts = hwc->freq_interrupts;
-			hwc->freq_interrupts = 0;
-			hwc->freq_count -= HZ;
-		} else
-			freq = HZ;
-
-		perf_adjust_period(event, freq * interrupts);
+		event->pmu->read(event);
+		now = atomic64_read(&event->count);
+		delta = now - hwc->freq_count_stamp;
+		hwc->freq_count_stamp = now;
 
-		/*
-		 * In order to avoid being stalled by an (accidental) huge
-		 * sample period, force reset the sample period if we didn't
-		 * get any events in this freq period.
-		 */
-		if (!interrupts) {
-			perf_disable();
-			event->pmu->disable(event);
-			atomic64_set(&hwc->period_left, 0);
-			event->pmu->enable(event);
-			perf_enable();
-		}
+		if (delta > 0)
+			perf_adjust_period(event, TICK_NSEC, delta);
 	}
 	raw_spin_unlock(&ctx->lock);
 }
@@ -1442,26 +1560,18 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx)
  */
 static void rotate_ctx(struct perf_event_context *ctx)
 {
-	struct perf_event *event;
-
 	if (!ctx->nr_events)
 		return;
 
 	raw_spin_lock(&ctx->lock);
-	/*
-	 * Rotate the first entry last (works just fine for group events too):
-	 */
-	perf_disable();
-	list_for_each_entry(event, &ctx->group_list, group_entry) {
-		list_move_tail(&event->group_entry, &ctx->group_list);
-		break;
-	}
-	perf_enable();
+
+	/* Rotate the first entry last of non-pinned groups */
+	list_rotate_left(&ctx->flexible_groups);
 
 	raw_spin_unlock(&ctx->lock);
 }
 
-void perf_event_task_tick(struct task_struct *curr, int cpu)
+void perf_event_task_tick(struct task_struct *curr)
 {
 	struct perf_cpu_context *cpuctx;
 	struct perf_event_context *ctx;
@@ -1469,24 +1579,43 @@ void perf_event_task_tick(struct task_struct *curr, int cpu)
 	if (!atomic_read(&nr_events))
 		return;
 
-	cpuctx = &per_cpu(perf_cpu_context, cpu);
+	cpuctx = &__get_cpu_var(perf_cpu_context);
 	ctx = curr->perf_event_ctxp;
 
+	perf_disable();
+
 	perf_ctx_adjust_freq(&cpuctx->ctx);
 	if (ctx)
 		perf_ctx_adjust_freq(ctx);
 
-	perf_event_cpu_sched_out(cpuctx);
+	cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE);
 	if (ctx)
-		__perf_event_task_sched_out(ctx);
+		task_ctx_sched_out(ctx, EVENT_FLEXIBLE);
 
 	rotate_ctx(&cpuctx->ctx);
 	if (ctx)
 		rotate_ctx(ctx);
 
-	perf_event_cpu_sched_in(cpuctx, cpu);
+	cpu_ctx_sched_in(cpuctx, EVENT_FLEXIBLE);
 	if (ctx)
-		perf_event_task_sched_in(curr, cpu);
+		task_ctx_sched_in(curr, EVENT_FLEXIBLE);
+
+	perf_enable();
+}
+
+static int event_enable_on_exec(struct perf_event *event,
+				struct perf_event_context *ctx)
+{
+	if (!event->attr.enable_on_exec)
+		return 0;
+
+	event->attr.enable_on_exec = 0;
+	if (event->state >= PERF_EVENT_STATE_INACTIVE)
+		return 0;
+
+	__perf_event_mark_enabled(event, ctx);
+
+	return 1;
 }
 
 /*
@@ -1499,6 +1628,7 @@ static void perf_event_enable_on_exec(struct task_struct *task)
 	struct perf_event *event;
 	unsigned long flags;
 	int enabled = 0;
+	int ret;
 
 	local_irq_save(flags);
 	ctx = task->perf_event_ctxp;
@@ -1509,14 +1639,16 @@ static void perf_event_enable_on_exec(struct task_struct *task)
 
 	raw_spin_lock(&ctx->lock);
 
-	list_for_each_entry(event, &ctx->group_list, group_entry) {
-		if (!event->attr.enable_on_exec)
-			continue;
-		event->attr.enable_on_exec = 0;
-		if (event->state >= PERF_EVENT_STATE_INACTIVE)
-			continue;
-		__perf_event_mark_enabled(event, ctx);
-		enabled = 1;
+	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) {
+		ret = event_enable_on_exec(event, ctx);
+