Merge branch 'perf/core' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip into trace/tip/tracing/core-7

Conflicts:
	include/linux/ftrace_event.h
	include/trace/ftrace.h
	kernel/trace/trace_event_perf.c
	kernel/trace/trace_kprobe.c
	kernel/trace/trace_syscalls.c

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>

25 files changed, 762 insertions, 489 deletions

diff --git a/kernel/acct.c b/kernel/acct.c
index 24f8c81fc48d..e4c0e1fee9b0 100644
--- a/kernel/acct.c
+++ b/kernel/acct.c
@@ -353,17 +353,18 @@ restart:
 
 void acct_exit_ns(struct pid_namespace *ns)
 {
-	struct bsd_acct_struct *acct;
+	struct bsd_acct_struct *acct = ns->bacct;
 
-	spin_lock(&acct_lock);
-	acct = ns->bacct;
-	if (acct != NULL) {
-		if (acct->file != NULL)
-			acct_file_reopen(acct, NULL, NULL);
+	if (acct == NULL)
+		return;
 
-		kfree(acct);
-	}
+	del_timer_sync(&acct->timer);
+	spin_lock(&acct_lock);
+	if (acct->file != NULL)
+		acct_file_reopen(acct, NULL, NULL);
 	spin_unlock(&acct_lock);
+
+	kfree(acct);
 }
 
 /*
diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index 4a07d057a265..e9ec642932ee 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -1646,7 +1646,9 @@ static inline struct cftype *__d_cft(struct dentry *dentry)
 int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen)
 {
 	char *start;
-	struct dentry *dentry = rcu_dereference(cgrp->dentry);
+	struct dentry *dentry = rcu_dereference_check(cgrp->dentry,
+						      rcu_read_lock_held() ||
+						      cgroup_lock_is_held());
 
 	if (!dentry || cgrp == dummytop) {
 		/*
@@ -1662,13 +1664,17 @@ int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen)
 	*--start = '\0';
 	for (;;) {
 		int len = dentry->d_name.len;
+
 		if ((start -= len) < buf)
 			return -ENAMETOOLONG;
-		memcpy(start, cgrp->dentry->d_name.name, len);
+		memcpy(start, dentry->d_name.name, len);
 		cgrp = cgrp->parent;
 		if (!cgrp)
 			break;
-		dentry = rcu_dereference(cgrp->dentry);
+
+		dentry = rcu_dereference_check(cgrp->dentry,
+					       rcu_read_lock_held() ||
+					       cgroup_lock_is_held());
 		if (!cgrp->parent)
 			continue;
 		if (--start < buf)
@@ -4429,7 +4435,15 @@ __setup("cgroup_disable=", cgroup_disable);
  */
 unsigned short css_id(struct cgroup_subsys_state *css)
 {
-	struct css_id *cssid = rcu_dereference(css->id);
+	struct css_id *cssid;
+
+	/*
+	 * This css_id() can return correct value when somone has refcnt
+	 * on this or this is under rcu_read_lock(). Once css->id is allocated,
+	 * it's unchanged until freed.
+	 */
+	cssid = rcu_dereference_check(css->id,
+			rcu_read_lock_held() || atomic_read(&css->refcnt));
 
 	if (cssid)
 		return cssid->id;
@@ -4439,7 +4453,10 @@ EXPORT_SYMBOL_GPL(css_id);
 
 unsigned short css_depth(struct cgroup_subsys_state *css)
 {
-	struct css_id *cssid = rcu_dereference(css->id);
+	struct css_id *cssid;
+
+	cssid = rcu_dereference_check(css->id,
+			rcu_read_lock_held() || atomic_read(&css->refcnt));
 
 	if (cssid)
 		return cssid->depth;
@@ -4447,15 +4464,36 @@ unsigned short css_depth(struct cgroup_subsys_state *css)
 }
 EXPORT_SYMBOL_GPL(css_depth);
 
+/**
+ *  css_is_ancestor - test "root" css is an ancestor of "child"
+ * @child: the css to be tested.
+ * @root: the css supporsed to be an ancestor of the child.
+ *
+ * Returns true if "root" is an ancestor of "child" in its hierarchy. Because
+ * this function reads css->id, this use rcu_dereference() and rcu_read_lock().
+ * But, considering usual usage, the csses should be valid objects after test.
+ * Assuming that the caller will do some action to the child if this returns
+ * returns true, the caller must take "child";s reference count.
+ * If "child" is valid object and this returns true, "root" is valid, too.
+ */
+
 bool css_is_ancestor(struct cgroup_subsys_state *child,
 		    const struct cgroup_subsys_state *root)
 {
-	struct css_id *child_id = rcu_dereference(child->id);
-	struct css_id *root_id = rcu_dereference(root->id);
+	struct css_id *child_id;
+	struct css_id *root_id;
+	bool ret = true;
 
-	if (!child_id || !root_id || (child_id->depth < root_id->depth))
-		return false;
-	return child_id->stack[root_id->depth] == root_id->id;
+	rcu_read_lock();
+	child_id  = rcu_dereference(child->id);
+	root_id = rcu_dereference(root->id);
+	if (!child_id
+	    || !root_id
+	    || (child_id->depth < root_id->depth)
+	    || (child_id->stack[root_id->depth] != root_id->id))
+		ret = false;
+	rcu_read_unlock();
+	return ret;
 }
 
 static void __free_css_id_cb(struct rcu_head *head)
@@ -4555,13 +4593,13 @@ static int alloc_css_id(struct cgroup_subsys *ss, struct cgroup *parent,
 {
 	int subsys_id, i, depth = 0;
 	struct cgroup_subsys_state *parent_css, *child_css;
-	struct css_id *child_id, *parent_id = NULL;
+	struct css_id *child_id, *parent_id;
 
 	subsys_id = ss->subsys_id;
 	parent_css = parent->subsys[subsys_id];
 	child_css = child->subsys[subsys_id];
-	depth = css_depth(parent_css) + 1;
 	parent_id = parent_css->id;
+	depth = parent_id->depth;
 
 	child_id = get_new_cssid(ss, depth);
 	if (IS_ERR(child_id))
diff --git a/kernel/fork.c b/kernel/fork.c
index 5d3592deaf71..4d57d9e3a6e9 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1111,7 +1111,6 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	p->memcg_batch.do_batch = 0;
 	p->memcg_batch.memcg = NULL;
 #endif
-	p->stack_start = stack_start;
 
 	/* Perform scheduler related setup. Assign this task to a CPU. */
 	sched_fork(p, clone_flags);
diff --git a/kernel/kexec.c b/kernel/kexec.c
index 87ebe8adc474..474a84715eac 100644
--- a/kernel/kexec.c
+++ b/kernel/kexec.c
@@ -1134,11 +1134,9 @@ int crash_shrink_memory(unsigned long new_size)
 
 	free_reserved_phys_range(end, crashk_res.end);
 
-	if (start == end) {
-		crashk_res.end = end;
+	if (start == end)
 		release_resource(&crashk_res);
-	} else
-		crashk_res.end = end - 1;
+	crashk_res.end = end - 1;
 
 unlock:
 	mutex_unlock(&kexec_mutex);
diff --git a/kernel/lockdep.c b/kernel/lockdep.c
index e9c759f06c1d..ec21304856d1 100644
--- a/kernel/lockdep.c
+++ b/kernel/lockdep.c
@@ -431,20 +431,7 @@ static struct stack_trace lockdep_init_trace = {
 /*
  * Various lockdep statistics:
  */
-atomic_t chain_lookup_hits;
-atomic_t chain_lookup_misses;
-atomic_t hardirqs_on_events;
-atomic_t hardirqs_off_events;
-atomic_t redundant_hardirqs_on;
-atomic_t redundant_hardirqs_off;
-atomic_t softirqs_on_events;
-atomic_t softirqs_off_events;
-atomic_t redundant_softirqs_on;
-atomic_t redundant_softirqs_off;
-atomic_t nr_unused_locks;
-atomic_t nr_cyclic_checks;
-atomic_t nr_find_usage_forwards_checks;
-atomic_t nr_find_usage_backwards_checks;
+DEFINE_PER_CPU(struct lockdep_stats, lockdep_stats);
 #endif
 
 /*
@@ -748,7 +735,7 @@ register_lock_class(struct lockdep_map *lock, unsigned int subclass, int force)
 		return NULL;
 	}
 	class = lock_classes + nr_lock_classes++;
-	debug_atomic_inc(&nr_unused_locks);
+	debug_atomic_inc(nr_unused_locks);
 	class->key = key;
 	class->name = lock->name;
 	class->subclass = subclass;
@@ -818,7 +805,8 @@ static struct lock_list *alloc_list_entry(void)
  * Add a new dependency to the head of the list:
  */
 static int add_lock_to_list(struct lock_class *class, struct lock_class *this,
-			    struct list_head *head, unsigned long ip, int distance)
+			    struct list_head *head, unsigned long ip,
+			    int distance, struct stack_trace *trace)
 {
 	struct lock_list *entry;
 	/*
@@ -829,11 +817,9 @@ static int add_lock_to_list(struct lock_class *class, struct lock_class *this,
 	if (!entry)
 		return 0;
 
-	if (!save_trace(&entry->trace))
-		return 0;
-
 	entry->class = this;
 	entry->distance = distance;
+	entry->trace = *trace;
 	/*
 	 * Since we never remove from the dependency list, the list can
 	 * be walked lockless by other CPUs, it's only allocation
@@ -1205,7 +1191,7 @@ check_noncircular(struct lock_list *root, struct lock_class *target,
 {
 	int result;
 
-	debug_atomic_inc(&nr_cyclic_checks);
+	debug_atomic_inc(nr_cyclic_checks);
 
 	result = __bfs_forwards(root, target, class_equal, target_entry);
 
@@ -1242,7 +1228,7 @@ find_usage_forwards(struct lock_list *root, enum lock_usage_bit bit,
 {
 	int result;
 
-	debug_atomic_inc(&nr_find_usage_forwards_checks);
+	debug_atomic_inc(nr_find_usage_forwards_checks);
 
 	result = __bfs_forwards(root, (void *)bit, usage_match, target_entry);
 
@@ -1265,7 +1251,7 @@ find_usage_backwards(struct lock_list *root, enum lock_usage_bit bit,
 {
 	int result;
 
-	debug_atomic_inc(&nr_find_usage_backwards_checks);
+	debug_atomic_inc(nr_find_usage_backwards_checks);
 
 	result = __bfs_backwards(root, (void *)bit, usage_match, target_entry);
 
@@ -1635,12 +1621,20 @@ check_deadlock(struct task_struct *curr, struct held_lock *next,
  */
 static int
 check_prev_add(struct task_struct *curr, struct held_lock *prev,
-	       struct held_lock *next, int distance)
+	       struct held_lock *next, int distance, int trylock_loop)
 {
 	struct lock_list *entry;
 	int ret;
 	struct lock_list this;
 	struct lock_list *uninitialized_var(target_entry);
+	/*
+	 * Static variable, serialized by the graph_lock().
+	 *
+	 * We use this static variable to save the stack trace in case
+	 * we call into this function multiple times due to encountering
+	 * trylocks in the held lock stack.
+	 */
+	static struct stack_trace trace;
 
 	/*
 	 * Prove that the new <prev> -> <next> dependency would not
@@ -1688,20 +1682,23 @@ check_prev_add(struct task_struct *curr, struct held_lock *prev,
 		}
 	}
 
+	if (!trylock_loop && !save_trace(&trace))
+		return 0;
+
 	/*
 	 * Ok, all validations passed, add the new lock
 	 * to the previous lock's dependency list:
 	 */
 	ret = add_lock_to_list(hlock_class(prev), hlock_class(next),
 			       &hlock_class(prev)->locks_after,
-			       next->acquire_ip, distance);
+			       next->acquire_ip, distance, &trace);
 
 	if (!ret)
 		return 0;
 
 	ret = add_lock_to_list(hlock_class(next), hlock_class(prev),
 			       &hlock_class(next)->locks_before,
-			       next->acquire_ip, distance);
+			       next->acquire_ip, distance, &trace);
 	if (!ret)
 		return 0;
 
@@ -1731,6 +1728,7 @@ static int
 check_prevs_add(struct task_struct *curr, struct held_lock *next)
 {
 	int depth = curr->lockdep_depth;
+	int trylock_loop = 0;
 	struct held_lock *hlock;
 
 	/*
@@ -1756,7 +1754,8 @@ check_prevs_add(struct task_struct *curr, struct held_lock *next)
 		 * added:
 		 */
 		if (hlock->read != 2) {
-			if (!check_prev_add(curr, hlock, next, distance))
+			if (!check_prev_add(curr, hlock, next,
+						distance, trylock_loop))
 				return 0;
 			/*
 			 * Stop after the first non-trylock entry,
@@ -1779,6 +1778,7 @@ check_prevs_add(struct task_struct *curr, struct held_lock *next)
 		if (curr->held_locks[depth].irq_context !=
 				curr->held_locks[depth-1].irq_context)
 			break;
+		trylock_loop = 1;
 	}
 	return 1;
 out_bug:
@@ -1825,7 +1825,7 @@ static inline int lookup_chain_cache(struct task_struct *curr,
 	list_for_each_entry(chain, hash_head, entry) {
 		if (chain->chain_key == chain_key) {
 cache_hit:
-			debug_atomic_inc(&chain_lookup_hits);
+			debug_atomic_inc(chain_lookup_hits);
 			if (very_verbose(class))
 				printk("\nhash chain already cached, key: "
 					"%016Lx tail class: [%p] %s\n",
@@ -1890,7 +1890,7 @@ cache_hit:
 		chain_hlocks[chain->base + j] = class - lock_classes;
 	}
 	list_add_tail_rcu(&chain->entry, hash_head);
-	debug_atomic_inc(&chain_lookup_misses);
+	debug_atomic_inc(chain_lookup_misses);
 	inc_chains();
 
 	return 1;
@@ -2311,7 +2311,12 @@ void trace_hardirqs_on_caller(unsigned long ip)
 		return;
 
 	if (unlikely(curr->hardirqs_enabled)) {
-		debug_atomic_inc(&redundant_hardirqs_on);
+		/*
+		 * Neither irq nor preemption are disabled here
+		 * so this is racy by nature but loosing one hit
+		 * in a stat is not a big deal.
+		 */
+		__debug_atomic_inc(redundant_hardirqs_on);
 		return;
 	}
 	/* we'll do an OFF -> ON transition: */
@@ -2338,7 +2343,7 @@ void trace_hardirqs_on_caller(unsigned long ip)
 
 	curr->hardirq_enable_ip = ip;
 	curr->hardirq_enable_event = ++curr->irq_events;
-	debug_atomic_inc(&hardirqs_on_events);
+	debug_atomic_inc(hardirqs_on_events);
 }
 EXPORT_SYMBOL(trace_hardirqs_on_caller);
 
@@ -2370,9 +2375,9 @@ void trace_hardirqs_off_caller(unsigned long ip)
 		curr->hardirqs_enabled = 0;
 		curr->hardirq_disable_ip = ip;
 		curr->hardirq_disable_event = ++curr->irq_events;
-		debug_atomic_inc(&hardirqs_off_events);
+		debug_atomic_inc(hardirqs_off_events);
 	} else
-		debug_atomic_inc(&redundant_hardirqs_off);
+		debug_atomic_inc(redundant_hardirqs_off);
 }
 EXPORT_SYMBOL(trace_hardirqs_off_caller);
 
@@ -2396,7 +2401,7 @@ void trace_softirqs_on(unsigned long ip)
 		return;
 
 	if (curr->softirqs_enabled) {
-		debug_atomic_inc(&redundant_softirqs_on);
+		debug_atomic_inc(redundant_softirqs_on);
 		return;
 	}
 
@@ -2406,7 +2411,7 @@ void trace_softirqs_on(unsigned long ip)
 	curr->softirqs_enabled = 1;
 	curr->softirq_enable_ip = ip;
 	curr->softirq_enable_event = ++curr->irq_events;
-	debug_atomic_inc(&softirqs_on_events);
+	debug_atomic_inc(softirqs_on_events);
 	/*
 	 * We are going to turn softirqs on, so set the
 	 * usage bit for all held locks, if hardirqs are
@@ -2436,10 +2441,10 @@ void trace_softirqs_off(unsigned long ip)
 		curr->softirqs_enabled = 0;
 		curr->softirq_disable_ip = ip;
 		curr->softirq_disable_event = ++curr->irq_events;
-		debug_atomic_inc(&softirqs_off_events);
+		debug_atomic_inc(softirqs_off_events);
 		DEBUG_LOCKS_WARN_ON(!softirq_count());
 	} else
-		debug_atomic_inc(&redundant_softirqs_off);
+		debug_atomic_inc(redundant_softirqs_off);
 }
 
 static void __lockdep_trace_alloc(gfp_t gfp_mask, unsigned long flags)
@@ -2644,7 +2649,7 @@ static int mark_lock(struct task_struct *curr, struct held_lock *this,
 			return 0;
 		break;
 	case LOCK_USED:
-		debug_atomic_dec(&nr_unused_locks);
+		debug_atomic_dec(nr_unused_locks);
 		break;
 	default:
 		if (!debug_locks_off_graph_unlock())
@@ -2750,7 +2755,7 @@ static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass,
 		if (!class)
 			return 0;
 	}
-	debug_atomic_inc((atomic_t *)&class->ops);
+	atomic_inc((atomic_t *)&class->ops);
 	if (very_verbose(class)) {
 		printk("\nacquire class [%p] %s", class->key, class->name);
 		if (class->name_version > 1)
@@ -3801,8 +3806,11 @@ void lockdep_rcu_dereference(const char *file, const int line)
 {
 	struct task_struct *curr = current;
 
+#ifndef CONFIG_PROVE_RCU_REPEATEDLY
 	if (!debug_locks_off())
 		return;
+#endif /* #ifdef CONFIG_PROVE_RCU_REPEATEDLY */
+	/* Note: the following can be executed concurrently, so be careful. */
 	printk("\n===================================================\n");
 	printk(  "[ INFO: suspicious rcu_dereference_check() usage. ]\n");
 	printk(  "---------------------------------------------------\n");
diff --git a/kernel/lockdep_internals.h b/kernel/lockdep_internals.h
index a2ee95ad1313..4f560cfedc8f 100644
--- a/kernel/lockdep_internals.h
+++ b/kernel/lockdep_internals.h
@@ -110,30 +110,60 @@ lockdep_count_backward_deps(struct lock_class *class)
 #endif
 
 #ifdef CONFIG_DEBUG_LOCKDEP
+
+#include <asm/local.h>
 /*
- * Various lockdep statistics:
+ * Various lockdep statistics.
+ * We want them per cpu as they are often accessed in fast path
+ * and we want to avoid too much cache bouncing.
  */
-extern atomic_t chain_lookup_hits;
-extern atomic_t chain_lookup_misses;
-extern atomic_t hardirqs_on_events;
-extern atomic_t hardirqs_off_events;
-extern atomic_t redundant_hardirqs_on;
-extern atomic_t redundant_hardirqs_off;
-extern atomic_t softirqs_on_events;
-extern atomic_t softirqs_off_events;
-extern atomic_t redundant_softirqs_on;
-extern atomic_t redundant_softirqs_off;
-extern atomic_t nr_unused_locks;
-extern atomic_t nr_cyclic_checks;
-extern atomic_t nr_cyclic_check_recursions;
-extern atomic_t nr_find_usage_forwards_checks;
-extern atomic_t nr_find_usage_forwards_recursions;
-extern atomic_t nr_find_usage_backwards_checks;
-extern atomic_t nr_find_usage_backwards_recursions;
-# define debug_atomic_inc(ptr)		atomic_inc(ptr)
-# define debug_atomic_dec(ptr)		atomic_dec(ptr)
-# define debug_atomic_read(ptr)		atomic_read(ptr)
+struct lockdep_stats {
+	int	chain_lookup_hits;
+	int	chain_lookup_misses;
+	int	hardirqs_on_events;
+	int	hardirqs_off_events;
+	int	redundant_hardirqs_on;
+	int	redundant_hardirqs_off;
+	int	softirqs_on_events;
+	int	softirqs_off_events;
+	int	redundant_softirqs_on;
+	int	redundant_softirqs_off;
+	int	nr_unused_locks;
+	int	nr_cyclic_checks;
+	int	nr_cyclic_check_recursions;
+	int	nr_find_usage_forwards_checks;
+	int	nr_find_usage_forwards_recursions;
+	int	nr_find_usage_backwards_checks;
+	int	nr_find_usage_backwards_recursions;
+};
+
+DECLARE_PER_CPU(struct lockdep_stats, lockdep_stats);
+
+#define __debug_atomic_inc(ptr)					\
+	this_cpu_inc(lockdep_stats.ptr);
+
+#define debug_atomic_inc(ptr)			{		\
+	WARN_ON_ONCE(!irqs_disabled());				\
+	__this_cpu_inc(lockdep_stats.ptr);			\
+}
+
+#define debug_atomic_dec(ptr)			{		\
+	WARN_ON_ONCE(!irqs_disabled());				\
+	__this_cpu_dec(lockdep_stats.ptr);			\
+}
+
+#define debug_atomic_read(ptr)		({				\
+	struct lockdep_stats *__cpu_lockdep_stats;			\
+	unsigned long long __total = 0;					\
+	int __cpu;							\
+	for_each_possible_cpu(__cpu) {					\
+		__cpu_lockdep_stats = &per_cpu(lockdep_stats, __cpu);	\
+		__total += __cpu_lockdep_stats->ptr;			\
+	}								\
+	__total;							\
+})
 #else
+# define __debug_atomic_inc(ptr)	do { } while (0)
 # define debug_atomic_inc(ptr)		do { } while (0)
 # define debug_atomic_dec(ptr)		do { } while (0)
 # define debug_atomic_read(ptr)		0
diff --git a/kernel/lockdep_proc.c b/kernel/lockdep_proc.c
index d4aba4f3584c..59b76c8ce9d7 100644
--- a/kernel/lockdep_proc.c
+++ b/kernel/lockdep_proc.c
@@ -184,34 +184,34 @@ static const struct file_operations proc_lockdep_chains_operations = {
 static void lockdep_stats_debug_show(struct seq_file *m)
 {
 #ifdef CONFIG_DEBUG_LOCKDEP
-	unsigned int hi1 = debug_atomic_read(&hardirqs_on_events),
-		     hi2 = debug_atomic_read(&hardirqs_off_events),
-		     hr1 = debug_atomic_read(&redundant_hardirqs_on),
-		     hr2 = debug_atomic_read(&redundant_hardirqs_off),
-		     si1 = debug_atomic_read(&softirqs_on_events),
-		     si2 = debug_atomic_read(&softirqs_off_events),
-		     sr1 = debug_atomic_read(&redundant_softirqs_on),
-		     sr2 = debug_atomic_read(&redundant_softirqs_off);
-
-	seq_printf(m, " chain lookup misses:           %11u\n",
-		debug_atomic_read(&chain_lookup_misses));
-	seq_printf(m, " chain lookup hits:             %11u\n",
-		debug_atomic_read(&chain_lookup_hits));
-	seq_printf(m, " cyclic checks:                 %11u\n",
-		debug_atomic_read(&nr_cyclic_checks));
-	seq_printf(m, " find-mask forwards checks:     %11u\n",
-		debug_atomic_read(&nr_find_usage_forwards_checks));
-	seq_printf(m, " find-mask backwards checks:    %11u\n",
-		debug_atomic_read(&nr_find_usage_backwards_checks));
-
-	seq_printf(m, " hardirq on events:             %11u\n", hi1);
-	seq_printf(m, " hardirq off events:            %11u\n", hi2);
-	seq_printf(m, " redundant hardirq ons:         %11u\n", hr1);
-	seq_printf(m, " redundant hardirq offs:        %11u\n", hr2);
-	seq_printf(m, " softirq on events:             %11u\n", si1);
-	seq_printf(m, " softirq off events:            %11u\n", si2);
-	seq_printf(m, " redundant softirq ons:         %11u\n", sr1);
-	seq_printf(m, " redundant softirq offs:        %11u\n", sr2);
+	unsigned long long hi1 = debug_atomic_read(hardirqs_on_events),
+			   hi2 = debug_atomic_read(hardirqs_off_events),
+			   hr1 = debug_atomic_read(redundant_hardirqs_on),
+			   hr2 = debug_atomic_read(redundant_hardirqs_off),
+			   si1 = debug_atomic_read(softirqs_on_events),
+			   si2 = debug_atomic_read(softirqs_off_events),
+			   sr1 = debug_atomic_read(redundant_softirqs_on),
+			   sr2 = debug_atomic_read(redundant_softirqs_off);
+
+	seq_printf(m, " chain lookup misses:           %11llu\n",
+		debug_atomic_read(chain_lookup_misses));
+	seq_printf(m, " chain lookup hits:             %11llu\n",
+		debug_atomic_read(chain_lookup_hits));
+	seq_printf(m, " cyclic checks:                 %11llu\n",
+		debug_atomic_read(nr_cyclic_checks));
+	seq_printf(m, " find-mask forwards checks:     %11llu\n",
+		debug_atomic_read(nr_find_usage_forwards_checks));
+	seq_printf(m, " find-mask backwards checks:    %11llu\n",
+		debug_atomic_read(nr_find_usage_backwards_checks));
+
+	seq_printf(m, " hardirq on events:             %11llu\n", hi1);
+	seq_printf(m, " hardirq off events:            %11llu\n", hi2);
+	seq_printf(m, " redundant hardirq ons:         %11llu\n", hr1);
+	seq_printf(m, " redundant hardirq offs:        %11llu\n", hr2);
+	seq_printf(m, " softirq on events:             %11llu\n", si1);
+	seq_printf(m, " softirq off events:            %11llu\n", si2);
+	seq_printf(m, " redundant softirq ons:         %11llu\n", sr1);
+	seq_printf(m, " redundant softirq offs:        %11llu\n", sr2);
 #endif
 }
 
@@ -263,7 +263,7 @@ static int lockdep_stats_show(struct seq_file *m, void *v)
 #endif
 	}
 #ifdef CONFIG_DEBUG_LOCKDEP
-	DEBUG_LOCKS_WARN_ON(debug_atomic_read(&nr_unused_locks) != nr_unused);
+	DEBUG_LOCKS_WARN_ON(debug_atomic_read(nr_unused_locks) != nr_unused);
 #endif
 	seq_printf(m, " lock-classes:                  %11lu [max: %lu]\n",
 			nr_lock_classes, MAX_LOCKDEP_KEYS);
diff --git a/kernel/perf_event.c b/kernel/perf_event.c
index a4fa381db3c2..e099650cd249 100644
--- a/kernel/perf_event.c
+++ b/kernel/perf_event.c
@@ -2297,11 +2297,6 @@ unlock:
 	rcu_read_unlock();
 }
 
-static unsigned long perf_data_size(struct perf_mmap_data *data)
-{
-	return data->nr_pages << (PAGE_SHIFT + data->data_order);
-}
-
 #ifndef CONFIG_PERF_USE_VMALLOC
 
 /*
@@ -2320,6 +2315,19 @@ perf_mmap_to_page(struct perf_mmap_data *data, unsigned long pgoff)
 	return virt_to_page(data->data_pages[pgoff - 1]);
 }
 
+static void *perf_mmap_alloc_page(int cpu)
+{
+	struct page *page;
+	int node;
+
+	node = (cpu == -1) ? cpu : cpu_to_node(cpu);
+	page = alloc_pages_node(node, GFP_KERNEL | __GFP_ZERO, 0);
+	if (!page)
+		return NULL;
+
+	return page_address(page);
+}
+
 static struct perf_mmap_data *
 perf_mmap_data_alloc(struct perf_event *event, int nr_pages)
 {
@@ -2336,17 +2344,16 @@ perf_mmap_data_alloc(struct perf_event *event, int nr_pages)
 	if (!data)
 		goto fail;
 
-	data->user_page = (void *)get_zeroed_page(GFP_KERNEL);
+	data->user_page = perf_mmap_alloc_page(event->cpu);
 	if (!data->user_page)
 		goto fail_user_page;
 
 	for (i = 0; i < nr_pages; i++) {
-		data->data_pages[i] = (void *)get_zeroed_page(GFP_KERNEL);
+		data->data_pages[i] = perf_mmap_alloc_page(event->cpu);
 		if (!data->data_pages[i])
 			goto fail_data_pages;
 	}
 
-	data->data_order = 0;
 	data->nr_pages = nr_pages;
 
 	return data;
@@ -2382,6 +2389,11 @@ static void perf_mmap_data_free(struct perf_mmap_data *data)
 	kfree(data);
 }
 
+static inline int page_order(struct perf_mmap_data *data)
+{
+	return 0;
+}
+
 #else
 
 /*
@@ -2390,10 +2402,15 @@ static void perf_mmap_data_free(struct perf_mmap_data *data)
  * Required for architectures that have d-cache aliasing issues.
  */
 
+static inline int page_order(struct perf_mmap_data *data)
+{
+	return data->page_order;
+}
+
 static struct page *
 perf_mmap_to_page(struct perf_mmap_data *data, unsigned long pgoff)
 {
-	if (pgoff > (1UL << data->data_order))
+	if (pgoff > (1UL << page_order(data)))
 		return NULL;
 
 	return vmalloc_to_page((void *)data->user_page + pgoff * PAGE_SIZE);
@@ -2413,7 +2430,7 @@ static void perf_mmap_data_free_work(struct work_struct *work)
 	int i, nr;
 
 	data = container_of(work, struct perf_mmap_data, work);
-	nr = 1 << data->data_order;
+	nr = 1 << page_order(data);
 
 	base = data->user_page;
 	for (i = 0; i < nr + 1; i++)
@@ -2452,7 +2469,7 @@ perf_mmap_data_alloc(struct perf_event *event, int nr_pages)
 
 	data->user_page = all_buf;
 	data->data_pages[0] = all_buf + PAGE_SIZE;
-	data->data_order = ilog2(nr_pages);
+	data->page_order = ilog2(nr_pages);
 	data->nr_pages = 1;
 
 	return data;
@@ -2466,6 +2483,11 @@ fail:
 
 #endif
 
+static unsigned long perf_data_size(struct perf_mmap_data *data)
+{
+	return data->nr_pages << (PAGE_SHIFT + page_order(data));
+}
+
 static int perf_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 {
 	struct perf_event *event = vma->vm_file->private_data;
@@ -2506,8 +2528,6 @@ perf_mmap_data_init(struct perf_event *event, struct perf_mmap_data *data)
 {
 	long max_size = perf_data_size(data);
 
-	atomic_set(&data->lock, -1);
-
 	if (event->attr.watermark) {
 		data->watermark = min_t(long, max_size,
 					event->attr.wakeup_watermark);
@@ -2580,6 +2600,14 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
 	long user_extra, extra;
 	int ret = 0;
 
+	/*
+	 * Don't allow mmap() of inherited per-task counters. This would
+	 * create a performance issue due to all children writing to the
+	 * same buffer.
+	 */
+	if (event->cpu == -1 && event->attr.inherit)
+		return -EINVAL;
+
 	if (!(vma->vm_flags & VM_SHARED))
 		return -EINVAL;
 
@@ -2885,120 +2913,80 @@ static void perf_output_wakeup(struct perf_output_handle *handle)
 }
 
 /*
- * Curious locking construct.
- *
  * We need to ensure a later event_id doesn't publish a head when a former
- * event_id isn't done writing. However since we need to deal with NMIs we
+ * event isn't done writing. However since we need to deal with NMIs we
  * cannot fully serialize things.
  *
- * What we do is serialize between CPUs so we only have to deal with NMI
- * nesting on a single CPU.
- *
  * We only publish the head (and generate a wakeup) when the outer-most
- * event_id completes.
+ * event completes.
  */
-static void perf_output_lock(struct perf_output_handle *handle)
+static void perf_output_get_handle(struct perf_output_handle *handle)
 {
 	struct perf_mmap_data *data = handle->data;
-	int cur, cpu = get_cpu();
-
-	handle->locked = 0;
-
-	for (;;) {
-		cur = atomic_cmpxchg(&data->lock, -1, cpu);
-		if (cur == -1) {
-			handle->locked = 1;
-			break;
-		}
-		if (cur == cpu)
-			break;
 
-		cpu_relax();
-	}
+	preempt_disable();
+	local_inc(&data->nest);
+	handle->wakeup = local_read(&data->wakeup);
 }
 
-static void perf_output_unlock(struct perf_output_handle *handle)
+static void perf_output_put_handle(struct perf_output_handle *handle)
 {
 	struct perf_mmap_data *data = handle->data;
 	unsigned long head;
-	int cpu;
-
-	data->done_head = data->head;
-
-	if (!handle->locked)
-		goto out;
 
 again:
-	/*
-	 * The xchg implies a full barrier that ensures all writes are done
-	 * before we publish the new head, matched by a rmb() in userspace when
-	 * reading this position.
-	 */
-	while ((head = atomic_long_xchg(&data->done_head, 0)))
-		data->user_page->data_head = head;
+	head = local_read(&data->head);
 
 	/*
-	 * NMI can happen here, which means we can miss a done_head update.
+	 * IRQ/NMI can happen here, which means we can miss a head update.
 	 */
 
-	cpu = atomic_xchg(&data->lock, -1);
-	WARN_ON_ONCE(cpu != smp_processor_id());
+	if (!local_dec_and_test(&data->nest))
+		goto out;
 
 	/*
-	 * Therefore we have to validate we did not indeed do so.
+	 * Publish the known good head. Rely on the full barrier implied
+	 * by atomic_dec_and_test() order the data->head read and this
+	 * write.
 	 */
-	if (unlikely(atomic_long_read(&data->done_head))) {
-		/*
-		 * Since we had it locked, we can lock it again.
-		 */
-		while (atomic_cmpxchg(&data->lock, -1, cpu) != -1)
-			cpu_relax();
+	data->user_page->data_head = head;
 
+	/*
+	 * Now check if we missed an update, rely on the (compiler)
+	 * barrier in atomic_dec_and_test() to re-read data->head.
+	 */
+	if (unlikely(head != local_read(&data->head))) {
+		local_inc(&data->nest);
 		goto again;
 	}
 
-	if (atomic_xchg(&data->wakeup, 0))
+	if (handle->wakeup != local_read(&data->wakeup))
 		perf_output_wakeup(handle);
-out:
-	put_cpu();
+
+ out:
+	preempt_enable();
 }
 
-void perf_output_copy(struct perf_output_handle *handle,
+__always_inline void perf_output_copy(struct perf_output_handle *handle,
 		      const void *buf, unsigned int len)
 {
-	unsigned int pages_mask;
-	unsigned long offset;
-	unsigned int size;
-	void **pages;
-
-	offset		= handle->offset;
-	pages_mask	= handle->data->nr_pages - 1;
-	pages		= handle->data->data_pages;
-
 	do {
-		unsigned long page_offset;
-		unsigned long page_size;
-		int nr;
+		unsigned long size = min_t(unsigned long, handle->size, len);
 
-		nr	    = (offset >> PAGE_SHIFT) & pages_mask;
-		page_size   = 1UL << (handle->data->data_order + PAGE_SHIFT);
-		page_offset = offset & (page_size - 1);
-		size	    = min_t(unsigned int, page_size - page_offset, len);
+		memcpy(handle->addr, buf, size);
 
-		memcpy(pages[nr] + page_offset, buf, size);
+		len -= size;
+		handle->addr += size;
+		handle->size -= size;
+		if (!handle->size) {
+			struct perf_mmap_data *data = handle->data;
 
-		len	    -= size;
-		buf	    += size;
-		offset	    += size;
+			handle->page++;
+			handle->page &= data->nr_pages - 1;
+			handle->addr = data->data_pages[handle->page];
+			handle->size = PAGE_SIZE << page_order(data);
+		}
 	} while (len);
-
-	handle->offset = offset;
-
-	/*
-	 * Check we didn't copy past our reservation window, taking the
-	 * possible unsigned int wrap into account.
-	 */
-	WARN_ON_ONCE(((long)(handle->head - handle->offset)) < 0);
 }
 
 int perf_output_begin(struct perf_output_handle *handle,
@@ -3036,13 +3024,13 @@ int perf_output_begin(struct perf_output_handle *handle,
 	handle->sample	= sample;
 
 	if (!data->nr_pages)
-		goto fail;
+		goto out;
 
-	have_lost = atomic_read(&data->lost);
+	have_lost = local_read(&data->lost);
 	if (have_lost)
 		size += sizeof(lost_event);
 
-	perf_output_lock(handle);
+	perf_output_get_handle(handle);
 
 	do {
 		/*
@@ -3052,24 +3040,28 @@ int perf_output_begin(struct perf_output_handle *handle,
 		 */
 		tail = ACCESS_ONCE(data->user_page->data_tail);
 		smp_rmb();
-		offset = head = atomic_long_read(&data->head);
+		offset = head = local_read(&data->head);
 		head += size;
 		if (unlikely(!perf_output_space(data, tail, offset, head)))
 			goto fail;
-	} while (atomic_long_cmpxchg(&data->head, offset, head) != offset);
+	} while (local_cmpxchg(&data->head, offset, head) != offset);
 
-	handle->offset	= offset;
-	handle->head	= head;
+	if (head - local_read(&data->wakeup) > data->watermark)
+		local_add(data->watermark, &data->wakeup);
 
-	if (head - tail > data->watermark)
-		atomic_set(&data->wakeup, 1);
+	handle->page = offset >> (PAGE_SHIFT + page_order(data));
+	handle->page &= data->nr_pages - 1;
+	handle->size = offset & ((PAGE_SIZE << page_order(data)) - 1);
+	handle->addr = data->data_pages[handle->page];
+	handle->addr += handle->size;
+	handle->size = (PAGE_SIZE << page_order(data)) - handle->size;
 
 	if (have_lost) {
 		lost_event.header.type = PERF_RECORD_LOST;
 		lost_event.header.misc = 0;
 		lost_event.header.size = sizeof(lost_event);
 		lost_event.id          = event->id;
-		lost_event.lost        = atomic_xchg(&data->lost, 0);
+		lost_event.lost        = local_xchg(&data->lost, 0);
 
 		perf_output_put(handle, lost_event);
 	}
@@ -3077,8 +3069,8 @@ int perf_output_begin(struct perf_output_handle *handle,
 	return 0;
 
 fail:
-	atomic_inc(&data->lost);
-	perf_output_unlock(handle);
+	local_inc(&data->lost);
+	perf_output_put_handle(handle);
 out:
 	rcu_read_unlock();
 
@@ -3093,14 +3085,14 @@ void perf_output_end(struct perf_output_handle *handle)
 	int wakeup_events = event->attr.wakeup_events;
 
 	if (handle->sample && wakeup_events) {
-		int events = atomic_inc_return(&data->events);
+		int events = local_inc_return(&data->events);
 		if (events >= wakeup_events) {
-			atomic_sub(wakeup_events, &data->events);
-			atomic_set(&data->wakeup, 1);
+			local_sub(wakeup_events, &data->events);
+			local_inc(&data->wakeup);
 		}
 	}
 
-	perf_output_unlock(handle);
+	perf_output_put_handle(handle);
 	rcu_read_unlock();
 }
 
@@ -3436,22 +3428,13 @@ static void perf_event_task_output(struct perf_event *event,
 {
 	struct perf_output_handle handle;
 	struct task_struct *task = task_event->task;
-	unsigned long flags;
 	int size, ret;
 
-	/*
-	 * If this CPU attempts to acquire an rq lock held by a CPU spinning
-	 * in perf_output_lock() from interrupt context, it's game over.
-	 */
-	local_irq_save(flags);
-
 	size  = task_event->event_id.header.size;
 	ret = perf_output_begin(&handle, event, size, 0, 0);
 
-	if (ret) {
-		local_irq_restore(flags);
+	if (ret)
 		return;
-	}
 
 	task_event->event_id.pid = perf_event_pid(event, task);
 	task_event->event_id.ppid = perf_event_pid(event, current);
@@ -3462,7 +3445,6 @@ static void perf_event_task_output(struct perf_event *event,
 	perf_output_put(&handle, task_event->event_id);
 
 	perf_output_end(&handle);
-	local_irq_restore(flags);
 }
 
 static int perf_event_task_match(struct perf_event *event)
@@ -4020,9 +4002,6 @@ static void perf_swevent_add(struct perf_event *event, u64 nr,
 	perf_swevent_overflow(event, 0, nmi, data, regs);
 }
 
-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)
 {
@@ -4052,10 +4031,6 @@ static int perf_swevent_match(struct perf_event *event,
 	if (perf_exclude_event(event, regs))
 		return 0;
 
-	if (event->attr.type == PERF_TYPE_TRACEPOINT &&
-	    !perf_tp_event_match(event, data))
-		return 0;
-
 	return 1;
 }
 
@@ -4066,19 +4041,46 @@ static inline u64 swevent_hash(u64 type, u32 event_id)
 	return hash_64(val, SWEVENT_HLIST_BITS);
 }
 
-static struct hlist_head *
-find_swevent_head(struct perf_cpu_context *ctx, u64 type, u32 event_id)
+static inline struct hlist_head *
+__find_swevent_head(struct swevent_hlist *hlist, u64 type, u32 event_id)
 {
-	u64 hash;
-	struct swevent_hlist *hlist;
+	u64 hash = swevent_hash(type, event_id);
+
+	return &hlist->heads[hash];
+}
 
-	hash = swevent_hash(type, event_id);
+/* For the read side: events when they trigger */
+static inline struct hlist_head *
+find_swevent_head_rcu(struct perf_cpu_context *ctx, u64 type, u32 event_id)
+{
+	struct swevent_hlist *hlist;
 
 	hlist = rcu_dereference(ctx->swevent_hlist);
 	if (!hlist)
 		return NULL;
 
-	return &hlist->heads[hash];
+	return __find_swevent_head(hlist, type, event_id);
+}
+
+/* For the event head insertion and removal in the hlist */
+static inline struct hlist_head *
+find_swevent_head(struct perf_cpu_context *ctx, struct perf_event *event)
+{
+	struct swevent_hlist *hlist;
+	u32 event_id = event->attr.config;
+	u64 type = event->attr.type;
+
+	/*
+	 * Event scheduling is always serialized against hlist allocation
+	 * and release. Which makes the protected version suitable here.
+	 * The context lock guarantees that.
+	 */
+	hlist = rcu_dereference_protected(ctx->swevent_hlist,
+					  lockdep_is_held(&event->ctx->lock));
+	if (!hlist)
+		return NULL;
+
+	return __find_swevent_head(hlist, type, event_id);
 }
 
 static void do_perf_sw_event(enum perf_type_id type, u32 event_id,
@@ -4095,7 +4097,7 @@ static void do_perf_sw_event(enum perf_type_id type, u32 event_id,
 
 	rcu_read_lock();
 
-	head = find_swevent_head(cpuctx, type, event_id);
+	head = find_swevent_head_rcu(cpuctx, type, event_id);
 
 	if (!head)
 		goto end;
@@ -4110,7 +4112,7 @@ end:
 
 int perf_swevent_get_recursion_context(void)
 {
-	struct perf_cpu_context *cpuctx = &get_cpu_var(perf_cpu_context);
+	struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context);
 	int rctx;
 
 	if (in_nmi())
@@ -4122,10 +4124,8 @@ int perf_swevent_get_recursion_context(void)
 	else
 		rctx = 0;
 
-	if (cpuctx->recursion[rctx]) {
-		put_cpu_var(perf_cpu_context);
+	if (cpuctx->recursion[rctx])
 		return -1;
-	}
 
 	cpuctx->recursion[rctx]++;
 	barrier();
@@ -4139,7 +4139,6 @@ 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);
 
@@ -4150,6 +4149,7 @@ void __perf_sw_event(u32 event_id, u64 nr, int nmi,
 	struct perf_sample_data data;
 	int rctx;
 
+	preempt_disable_notrace();
 	rctx = perf_swevent_get_recursion_context();
 	if (rctx < 0)
 		return;
@@ -4159,6 +4159,7 @@ void __perf_sw_event(u32 event_id, u64 nr, int nmi,
 	do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, nmi, &data, regs);
 
 	perf_swevent_put_recursion_context(rctx);
+	preempt_enable_notrace();
 }
 
 static void perf_swevent_read(struct perf_event *event)
@@ -4178,7 +4179,7 @@ static int perf_swevent_enable(struct perf_event *event)
 		perf_swevent_set_period(event);
 	}
 
-	head = find_swevent_head(cpuctx, event->attr.type, event->attr.config);
+	head = find_swevent_head(cpuctx, event);
 	if (WARN_ON_ONCE(!head))
 		return -EINVAL;
 
@@ -4366,6 +4367,14 @@ static const struct pmu perf_ops_task_clock = {
 	.read		= task_clock_perf_event_read,
 };
 
+/* Deref the hlist from the update side */
+static inline struct swevent_hlist *
+swevent_hlist_deref(struct perf_cpu_context *cpuctx)
+{
+	return rcu_dereference_protected(cpuctx->swevent_hlist,
+					 lockdep_is_held(&cpuctx->hlist_mutex));
+}
+
 static void swevent_hlist_release_rcu(struct rcu_head *rcu_head)
 {
 	struct swevent_hlist *hlist;
@@ -4376,12 +4385,11 @@ static void swevent_hlist_release_rcu(struct rcu_head *rcu_head)
 
 static void swevent_hlist_release(struct perf_cpu_context *cpuctx)
 {
-	struct swevent_hlist *hlist;
+	struct swevent_hlist *hlist = swevent_hlist_deref(cpuctx);
 
-	if (!cpuctx->swevent_hlist)
+	if (!hlist)
 		return;
 
-	hlist = cpuctx->swevent_hlist;
 	rcu_assign_pointer(cpuctx->swevent_hlist, NULL);
 	call_rcu(&hlist->rcu_head, swevent_hlist_release_rcu);
 }
@@ -4418,7 +4426,7 @@ static int swevent_hlist_get_cpu(struct perf_event *event, int cpu)
 
 	mutex_lock(&cpuctx->hlist_mutex);
 
-	if (!cpuctx->swevent_hlist && cpu_online(cpu)) {
+	if (!swevent_hlist_deref(cpuctx) && cpu_online(cpu)) {
 		struct swevent_hlist *hlist;
 
 		hlist = kzalloc(sizeof(*hlist), GFP_KERNEL);
@@ -4467,10 +4475,46 @@ static int swevent_hlist_get(struct perf_event *event)
 
 #ifdef CONFIG_EVENT_TRACING
 
-void perf_tp_event(int event_id, u64 addr, u64 count, void *record,
-		   int entry_size, struct pt_regs *regs)
+static const struct pmu perf_ops_tracepoint = {
+	.enable		= perf_trace_enable,
+	.disable	= perf_trace_disable,
+	.read		= perf_swevent_read,
+	.unthrottle	= perf_swevent_unthrottle,
+};
+
+static int perf_tp_filter_match(struct perf_event *event,
+				struct perf_sample_data *data)
+{
+	void *record = data->raw->data;
+
+	if (likely(!event->filter) || filter_match_preds(event->filter, record))
+		return 1;
+	return 0;
+}
+
+static int perf_tp_event_match(struct perf_event *event,
+				struct perf_sample_data *data,
+				struct pt_regs *regs)
+{
+	/*
+	 * All tracepoints are from kernel-space.
+	 */
+	if (event->attr.exclude_kernel)
+		return 0;
+
+	if (!perf_tp_filter_match(event, data))
+		return 0;
+
+	return 1;
+}
+
+void perf_tp_event(u64 addr, u64 count, void *record, int entry_size,
+		   struct pt_regs *regs, struct hlist_head *head)
 {
 	struct perf_sample_data data;
+	struct perf_event *event;
+	struct hlist_node *node;
+
 	struct perf_raw_record raw = {
 		.size = entry_size,
 		.data = record,
@@ -4479,26 +4523,18 @@ void perf_tp_event(int event_id, u64 addr, u64 count, void *record,
 	perf_sample_data_init(&data, addr);
 	data.raw = &raw;
 
-	/* Trace events already protected against recursion */
-	do_perf_sw_event(PERF_TYPE_TRACEPOINT, event_id, count, 1,
-			 &data, regs);
+	rcu_read_lock();
+	hlist_for_each_entry_rcu(event, node, head, hlist_entry) {
+		if (perf_tp_event_match(event, &data, regs))
+			perf_swevent_add(event, count, 1, &data, regs);
+	}
+	rcu_read_unlock();
 }
 EXPORT_SYMBOL_GPL(perf_tp_event);
 
-static int perf_tp_event_match(struct perf_event *event,
-				struct perf_sample_data *data)
-{
-	void *record = data->raw->data;
-
-	if (likely(!event->filter) || filter_match_preds(event->filter, record))
-		return 1;
-	return 0;
-}
-
 static void tp_perf_event_destroy(struct perf_event *event)
 {
-	perf_trace_disable(event->attr.config);
-	swevent_hlist_put(event);
+	perf_trace_destroy(event);
 }
 
 static const struct pmu *tp_perf_event_init(struct perf_event *event)
@@ -4514,17 +4550,13 @@ static const struct pmu *tp_perf_event_init(struct perf_event *event)
 			!capable(CAP_SYS_ADMIN))
 		return ERR_PTR(-EPERM);
 
-	if (perf_trace_enable(event->attr.config))
+	err = perf_trace_init(event);
+	if (err)
 		return NULL;
 
 	event->destroy = tp_perf_event_destroy;
-	err = swevent_hlist_get(event);
-	if (err) {
-		perf_trace_disable(event->attr.config);
-		return ERR_PTR(err);
-	}
 
-	return &perf_ops_generic;
+	return &perf_ops_tracepoint;
 }
 
 static int perf_event_set_filter(struct perf_event *event, void __user *arg)
@@ -4552,12 +4584,6 @@ static void perf_event_free_filter(struct perf_event *event)
 
 #else
 
-static int perf_tp_event_match(struct perf_event *event,
-				struct perf_sample_data *data)
-{
-	return 1;
-}
-
 static const struct pmu *tp_perf_event_init(struct perf_event *event)
 {
 	return NULL;
@@ -4894,6 +4920,13 @@ static int perf_event_set_output(struct perf_event *event, int output_fd)
 	int fput_needed = 0;
 	int ret = -EINVAL;
 
+	/*
+	 * Don't allow output of inherited per-task events. This would
+	 * create performance issues due to cross cpu access.
+	 */
+	if (event->cpu == -1 && event->attr.inherit)
+		return -EINVAL;
+
 	if (!output_fd)
 		goto set;
 
@@ -4914,6 +4947,18 @@ static int perf_event_set_output(struct perf_event *event, int output_fd)
 	if (event->data)
 		goto out;
 
+	/*
+	 * Don't allow cross-cpu buffers
+	 */
+	if (output_event->cpu != event->cpu)
+		goto out;
+
+	/*
+	 * If its not a per-cpu buffer, it must be the same task.
+	 */
+	if (output_event->cpu == -1 && output_event->ctx != event->ctx)
+		goto out;
+
 	atomic_long_inc(&output_file->f_count);
 
 set:
diff --git a/kernel/profile.c b/kernel/profile.c
index a55d3a367ae8..dfadc5b729f1 100644
--- a/kernel/profile.c
+++ b/kernel/profile.c
@@ -127,8 +127,10 @@ int __ref profile_init(void)
 		return 0;
 
 	prof_buffer = vmalloc(buffer_bytes);
-	if (prof_buffer)
+	if (prof_buffer) {
+		memset(prof_buffer, 0, buffer_bytes);
 		return 0;
+	}
 
 	free_cpumask_var(prof_cpu_mask);
 	return -ENOMEM;
diff --git a/kernel/ptrace.c b/kernel/ptrace.c
index 9fb51237b18c..6af9cdd558b7 100644
--- a/kernel/ptrace.c
+++ b/kernel/ptrace.c
@@ -14,7 +14,6 @@
 #include <linux/mm.h>
 #include <linux/highmem.h>
 #include <linux/pagemap.h>
-#include <linux/smp_lock.h>
 #include <linux/ptrace.h>
 #include <linux/security.h>
 #include <linux/signal.h>
@@ -665,10 +664,6 @@ SYSCALL_DEFINE4(ptrace, long, request, long, pid, long, addr, long, data)
 	struct task_struct *child;
 	long ret;
 
-	/*
-	 * This lock_kernel fixes a subtle race with suid exec
-	 */
-	lock_kernel();
 	if (request == PTRACE_TRACEME) {
 		ret = ptrace_traceme();
 		if (!ret)
@@ -702,7 +697,6 @@ SYSCALL_DEFINE4(ptrace, long, request, long, pid, long, addr, long, data)
  out_put_task_struct:
 	put_task_struct(child);
  out:
-	unlock_kernel();
 	return ret;
 }
 
@@ -812,10 +806,6 @@ asmlinkage long compat_sys_ptrace(compat_long_t request, compat_long_t pid,
 	struct task_struct *child;
 	long ret;
 
-	/*
-	 * This lock_kernel fixes a subtle race with suid exec
-	 */
-	lock_kernel();
 	if (request == PTRACE_TRACEME) {
 		ret = ptrace_traceme();
 		goto out;
@@ -845,7 +835,6 @@ asmlinkage long compat_sys_ptrace(compat_long_t request, compat_long_t pid,
  out_put_task_struct:
 	put_task_struct(child);
  out:
-	unlock_kernel();
 	return ret;
 }
 #endif	/* CONFIG_COMPAT */
diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c
index 03a7ea1579f6..72a8dc9567f5 100644
--- a/kernel/rcupdate.c
+++ b/kernel/rcupdate.c
@@ -44,7 +44,6 @@
 #include <linux/cpu.h>
 #include <linux/mutex.h>
 #include <linux/module.h>
-#include <linux/kernel_stat.h>
 #include <linux/hardirq.h>
 
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
@@ -64,9 +63,6 @@ struct lockdep_map rcu_sched_lock_map =
 EXPORT_SYMBOL_GPL(rcu_sched_lock_map);
 #endif
 
-int rcu_scheduler_active __read_mostly;
-EXPORT_SYMBOL_GPL(rcu_scheduler_active);
-
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
 
 int debug_lockdep_rcu_enabled(void)
@@ -97,21 +93,6 @@ EXPORT_SYMBOL_GPL(rcu_read_lock_bh_held);
 #endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */
 
 /*
- * This function is invoked towards the end of the scheduler's initialization
- * process.  Before this is called, the idle task might contain
- * RCU read-side critical sections (during which time, this idle
- * task is booting the system).  After this function is called, the
- * idle tasks are prohibited from containing RCU read-side critical
- * sections.
- */
-void rcu_scheduler_starting(void)
-{
-	WARN_ON(num_online_cpus() != 1);
-	WARN_ON(nr_context_switches() > 0);
-	rcu_scheduler_active = 1;
-}
-
-/*
  * Awaken the corresponding synchronize_rcu() instance now that a
  * grace period has elapsed.
  */
@@ -122,3 +103,14 @@ void wakeme_after_rcu(struct rcu_head  *head)
 	rcu = container_of(head, struct rcu_synchronize, head);
 	complete(&rcu->completion);
 }
+
+#ifdef CONFIG_PROVE_RCU
+/*
+ * wrapper function to avoid #include problems.
+ */
+int rcu_my_thread_group_empty(void)
+{
+	return thread_group_empty(current);
+}
+EXPORT_SYMBOL_GPL(rcu_my_thread_group_empty);
+#endif /* #ifdef CONFIG_PROVE_RCU */
diff --git a/kernel/rcutiny.c b/kernel/rcutiny.c
index 9f6d9ff2572c..38729d3cd236 100644
--- a/kernel/rcutiny.c
+++ b/kernel/rcutiny.c
@@ -44,9 +44,9 @@ struct rcu_ctrlblk {
 };
 
 /* Definition for rcupdate control block. */
-static struct rcu_ctrlblk rcu_ctrlblk = {
-	.donetail	= &rcu_ctrlblk.rcucblist,
-	.curtail	= &rcu_ctrlblk.rcucblist,
+static struct rcu_ctrlblk rcu_sched_ctrlblk = {
+	.donetail	= &rcu_sched_ctrlblk.rcucblist,
+	.curtail	= &rcu_sched_ctrlblk.rcucblist,
 };
 
 static struct rcu_ctrlblk rcu_bh_ctrlblk = {
@@ -54,6 +54,11 @@ static struct rcu_ctrlblk rcu_bh_ctrlblk = {
 	.curtail	= &rcu_bh_ctrlblk.rcucblist,
 };
 
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+int rcu_scheduler_active __read_mostly;
+EXPORT_SYMBOL_GPL(rcu_scheduler_active);
+#endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */
+
 #ifdef CONFIG_NO_HZ
 
 static long rcu_dynticks_nesting = 1;
@@ -108,7 +113,8 @@ static int rcu_qsctr_help(struct rcu_ctrlblk *rcp)
  */
 void rcu_sched_qs(int cpu)
 {
-	if (rcu_qsctr_help(&rcu_ctrlblk) + rcu_qsctr_help(&rcu_bh_ctrlblk))
+	if (rcu_qsctr_help(&rcu_sched_ctrlblk) +
+	    rcu_qsctr_help(&rcu_bh_ctrlblk))
 		raise_softirq(RCU_SOFTIRQ);
 }
 
@@ -173,7 +179,7 @@ static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp)
  */
 static void rcu_process_callbacks(struct softirq_action *unused)
 {
-	__rcu_process_callbacks(&rcu_ctrlblk);
+	__rcu_process_callbacks(&rcu_sched_ctrlblk);
 	__rcu_process_callbacks(&rcu_bh_ctrlblk);
 }
 
@@ -187,7 +193,8 @@ static void rcu_process_callbacks(struct softirq_action *unused)
  *
  * Cool, huh?  (Due to Josh Triplett.)
  *
- * But we want to make this a static inline later.
+ * But we want to make this a static inline later.  The cond_resched()
+ * currently makes this problematic.
  */
 void synchronize_sched(void)
 {
@@ -195,12 +202,6 @@ void synchronize_sched(void)
 }
 EXPORT_SYMBOL_GPL(synchronize_sched);
 
-void synchronize_rcu_bh(void)
-{
-	synchronize_sched();
-}
-EXPORT_SYMBOL_GPL(synchronize_rcu_bh);
-
 /*
  * Helper function for call_rcu() and call_rcu_bh().
  */
@@ -226,7 +227,7 @@ static void __call_rcu(struct rcu_head *head,
  */
 void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu))
 {
-	__call_rcu(head, func, &rcu_ctrlblk);
+	__call_rcu(head, func, &rcu_sched_ctrlblk);
 }
 EXPORT_SYMBOL_GPL(call_rcu);
 
@@ -244,11 +245,13 @@ void rcu_barrier(void)
 {
 	struct rcu_synchronize rcu;
 
+	init_rcu_head_on_stack(&rcu.head);
 	init_completion(&rcu.completion);
 	/* Will wake me after RCU finished. */
 	call_rcu(&rcu.head, wakeme_after_rcu);
 	/* Wait for it. */
 	wait_for_completion(&rcu.completion);
+	destroy_rcu_head_on_stack(&rcu.head);
 }
 EXPORT_SYMBOL_GPL(rcu_barrier);
 
@@ -256,11 +259,13 @@ void rcu_barrier_bh(void)
 {
 	struct rcu_synchronize rcu;
 
+	init_rcu_head_on_stack(&rcu.head);
 	init_completion(&rcu.completion);
 	/* Will wake me after RCU finished. */
 	call_rcu_bh(&rcu.head, wakeme_after_rcu);
 	/* Wait for it. */
 	wait_for_completion(&rcu.completion);
+	destroy_rcu_head_on_stack(&rcu.head);
 }
 EXPORT_SYMBOL_GPL(rcu_barrier_bh);
 
@@ -268,11 +273,13 @@ void rcu_barrier_sched(void)
 {
 	struct rcu_synchronize rcu;
 
+	init_rcu_head_on_stack(&rcu.head);
 	init_completion(&rcu.completion);
 	/* Will wake me after RCU finished. */
 	call_rcu_sched(&rcu.head, wakeme_after_rcu);
 	/* Wait for it. */
 	wait_for_completion(&rcu.completion);
+	destroy_rcu_head_on_stack(&rcu.head);
 }
 EXPORT_SYMBOL_GPL(rcu_barrier_sched);
 
@@ -280,3 +287,5 @@ void __init rcu_init(void)
 {
 	open_softirq(RCU_SOFTIRQ, rcu_process_callbacks);
 }
+
+#include "rcutiny_plugin.h"
diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h
new file mode 100644
index 000000000000..d223a92bc742
--- /dev/null
+++ b/kernel/rcutiny_plugin.h
@@ -0,0 +1,39 @@
+/*
+ * Read-Copy Update mechanism for mutual exclusion (tree-based version)
+ * Internal non-public definitions that provide either classic
+ * or preemptable semantics.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Copyright IBM Corporation, 2009
+ *
+ * Author: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+ */
+
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+
+#include <linux/kernel_stat.h>
+
+/*
+ * During boot, we forgive RCU lockdep issues.  After this function is
+ * invoked, we start taking RCU lockdep issues seriously.
+ */
+void rcu_scheduler_starting(void)
+{
+	WARN_ON(nr_context_switches() > 0);
+	rcu_scheduler_active = 1;
+}
+
+#endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */
diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c
index 2b676f3a0f26..6535ac8bc6a5 100644
--- a/kernel/rcutorture.c
+++ b/kernel/rcutorture.c
@@ -464,9 +464,11 @@ static void rcu_bh_torture_synchronize(void)
 {
 	struct rcu_bh_torture_synchronize rcu;
 
+	init_rcu_head_on_stack(&rcu.head);
 	init_completion(&rcu.completion);
 	call_rcu_bh(&rcu.head, rcu_bh_torture_wakeme_after_cb);
 	wait_for_completion(&rcu.completion);
+	destroy_rcu_head_on_stack(&rcu.head);
 }
 
 static struct rcu_torture_ops rcu_bh_ops = {
diff --git a/kernel/rcutree.c b/kernel/rcutree.c
index 3ec8160fc75f..d4437345706f 100644
--- a/kernel/rcutree.c
+++ b/kernel/rcutree.c
@@ -46,6 +46,7 @@
 #include <linux/cpu.h>
 #include <linux/mutex.h>
 #include <linux/time.h>
+#include <linux/kernel_stat.h>
 
 #include "rcutree.h"
 
@@ -53,8 +54,8 @@
 
 static struct lock_class_key rcu_node_class[NUM_RCU_LVLS];
 
-#define RCU_STATE_INITIALIZER(name) { \
-	.level = { &name.node[0] }, \
+#define RCU_STATE_INITIALIZER(structname) { \
+	.level = { &structname.node[0] }, \
 	.levelcnt = { \
 		NUM_RCU_LVL_0,  /* root of hierarchy. */ \
 		NUM_RCU_LVL_1, \
@@ -65,13 +66,14 @@ static struct lock_class_key rcu_node_class[NUM_RCU_LVLS];
 	.signaled = RCU_GP_IDLE, \
 	.gpnum = -300, \
 	.completed = -300, \
-	.onofflock = __RAW_SPIN_LOCK_UNLOCKED(&name.onofflock), \
+	.onofflock = __RAW_SPIN_LOCK_UNLOCKED(&structname.onofflock), \
 	.orphan_cbs_list = NULL, \
-	.orphan_cbs_tail = &name.orphan_cbs_list, \
+	.orphan_cbs_tail = &structname.orphan_cbs_list, \
 	.orphan_qlen = 0, \
-	.fqslock = __RAW_SPIN_LOCK_UNLOCKED(&name.fqslock), \
+	.fqslock = __RAW_SPIN_LOCK_UNLOCKED(&structname.fqslock), \
 	.n_force_qs = 0, \
 	.n_force_qs_ngp = 0, \
+	.name = #structname, \
 }
 
 struct rcu_state rcu_sched_state = RCU_STATE_INITIALIZER(rcu_sched_state);
@@ -80,6 +82,9 @@ DEFINE_PER_CPU(struct rcu_data, rcu_sched_data);
 struct rcu_state rcu_bh_state = RCU_STATE_INITIALIZER(rcu_bh_state);
 DEFINE_PER_CPU(struct rcu_data, rcu_bh_data);
 
+int rcu_scheduler_active __read_mostly;
+EXPORT_SYMBOL_GPL(rcu_scheduler_active);
+
 /*
  * Return true if an RCU grace period is in progress.  The ACCESS_ONCE()s
  * permit this function to be invoked without holding the root rcu_node
@@ -97,25 +102,32 @@ static int rcu_gp_in_progress(struct rcu_state *rsp)
  */
 void rcu_sched_qs(int cpu)
 {
-	struct rcu_data *rdp;
+	struct rcu_data *rdp = &per_cpu(rcu_sched_data, cpu);
 
-	rdp = &per_cpu(rcu_sched_data, cpu);
 	rdp->passed_quiesc_completed = rdp->gpnum - 1;
 	barrier();
 	rdp->passed_quiesc = 1;
-	rcu_preempt_note_context_switch(cpu);
 }
 
 void rcu_bh_qs(int cpu)
 {
-	struct rcu_data *rdp;
+	struct rcu_data *rdp = &per_cpu(rcu_bh_data, cpu);
 
-	rdp = &per_cpu(rcu_bh_data, cpu);
 	rdp->passed_quiesc_completed = rdp->gpnum - 1;
 	barrier();
 	rdp->passed_quiesc = 1;
 }
 
+/*
+ * Note a context switch.  This is a quiescent state for RCU-sched,
+ * and requires special handling for preemptible RCU.
+ */
+void rcu_note_context_switch(int cpu)
+{
+	rcu_sched_qs(cpu);
+	rcu_preempt_note_context_switch(cpu);
+}
+
 #ifdef CONFIG_NO_HZ
 DEFINE_PER_CPU(struct rcu_dynticks, rcu_dynticks) = {
 	.dynticks_nesting = 1,
@@ -438,6 +450,8 @@ static int rcu_implicit_dynticks_qs(struct rcu_data *rdp)
 
 #ifdef CONFIG_RCU_CPU_STALL_DETECTOR
 
+int rcu_cpu_stall_panicking __read_mostly;
+
 static void record_gp_stall_check_time(struct rcu_state *rsp)
 {
 	rsp->gp_start = jiffies;
@@ -470,7 +484,8 @@ static void print_other_cpu_stall(struct rcu_state *rsp)
 
 	/* OK, time to rat on our buddy... */
 
-	printk(KERN_ERR "INFO: RCU detected CPU stalls:");
+	printk(KERN_ERR "INFO: %s detected stalls on CPUs/tasks: {",
+	       rsp->name);
 	rcu_for_each_leaf_node(rsp, rnp) {
 		raw_spin_lock_irqsave(&rnp->lock, flags);
 		rcu_print_task_stall(rnp);
@@ -481,7 +496,7 @@ static void print_other_cpu_stall(struct rcu_state *rsp)
 			if (rnp->qsmask & (1UL << cpu))
 				printk(" %d", rnp->grplo + cpu);
 	}
-	printk(" (detected by %d, t=%ld jiffies)\n",
+	printk("} (detected by %d, t=%ld jiffies)\n",
 	       smp_processor_id(), (long)(jiffies - rsp->gp_start));
 	trigger_all_cpu_backtrace();
 
@@ -497,8 +512,8 @@ static void print_cpu_stall(struct rcu_state *rsp)
 	unsigned long flags;
 	struct rcu_node *rnp = rcu_get_root(rsp);
 
-	printk(KERN_ERR "INFO: RCU detected CPU %d stall (t=%lu jiffies)\n",
-			smp_processor_id(), jiffies - rsp->gp_start);
+	printk(KERN_ERR "INFO: %s detected stall on CPU %d (t=%lu jiffies)\n",
+	       rsp->name, smp_processor_id(), jiffies - rsp->gp_start);
 	trigger_all_cpu_backtrace();
 
 	raw_spin_lock_irqsave(&rnp->lock, flags);
@@ -515,6 +530,8 @@ static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp)
 	long delta;
 	struct rcu_node *rnp;
 
+	if (rcu_cpu_stall_panicking)
+		return;
 	delta = jiffies - rsp->jiffies_stall;
 	rnp = rdp->mynode;
 	if ((rnp->qsmask & rdp->grpmask) && delta >= 0) {
@@ -529,6 +546,21 @@ static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp)
 	}
 }
 
+static int rcu_panic(struct notifier_block *this, unsigned long ev, void *ptr)
+{
+	rcu_cpu_stall_panicking = 1;
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block rcu_panic_block = {
+	.notifier_call = rcu_panic,
+};
+
+static void __init check_cpu_stall_init(void)
+{
+	atomic_notifier_chain_register(&panic_notifier_list, &rcu_panic_block);
+}
+
 #else /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */
 
 static void record_gp_stall_check_time(struct rcu_state *rsp)
@@ -539,6 +571,10 @@ static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp)
 {
 }
 
+static void __init check_cpu_stall_init(void)
+{
+}
+
 #endif /* #else #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */
 
 /*
@@ -1125,8 +1161,6 @@ static void rcu_do_batch(struct rcu_state *rsp, struct rcu_data *rdp)
  */
 void rcu_check_callbacks(int cpu, int user)
 {
-	if (!rcu_pending(cpu))
-		return; /* if nothing for RCU to do. */
 	if (user ||
 	    (idle_cpu(cpu) && rcu_scheduler_active &&
 	     !in_softirq() && hardirq_count() <= (1 << HARDIRQ_SHIFT))) {
@@ -1158,7 +1192,8 @@ void rcu_check_callbacks(int cpu, int user)
 		rcu_bh_qs(cpu);
 	}
 	rcu_preempt_check_callbacks(cpu);
-	raise_softirq(RCU_SOFTIRQ);
+	if (rcu_pending(cpu))
+		raise_softirq(RCU_SOFTIRQ);
 }
 
 #ifdef CONFIG_SMP
@@ -1236,11 +1271,11 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed)
 		break; /* grace period idle or initializing, ignore. */
 
 	case RCU_SAVE_DYNTICK:
-
-		raw_spin_unlock(&rnp->lock);  /* irqs remain disabled */
 		if (RCU_SIGNAL_INIT != RCU_SAVE_DYNTICK)
 			break; /* So gcc recognizes the dead code. */
 
+		raw_spin_unlock(&rnp->lock);  /* irqs remain disabled */
+
 		/* Record dyntick-idle state. */
 		force_qs_rnp(rsp, dyntick_save_progress_counter);
 		raw_spin_lock(&rnp->lock);  /* irqs already disabled */
@@ -1449,11 +1484,13 @@ void synchronize_sched(void)
 	if (rcu_blocking_is_gp())
 		return;
 
+	init_rcu_head_on_stack(&rcu.head);
 	init_completion(&rcu.completion);
 	/* Will wake me after RCU finished. */
 	call_rcu_sched(&rcu.head, wakeme_after_rcu);
 	/* Wait for it. */
 	wait_for_completion(&rcu.completion);
+	destroy_rcu_head_on_stack(&rcu.head);
 }
 EXPORT_SYMBOL_GPL(synchronize_sched);
 
@@ -1473,11 +1510,13 @@ void synchronize_rcu_bh(void)
 	if (rcu_blocking_is_gp())
 		return;
 
+	init_rcu_head_on_stack(&rcu.head);
 	init_completion(&rcu.completion);
 	/* Will wake me after RCU finished. */
 	call_rcu_bh(&rcu.head, wakeme_after_rcu);
 	/* Wait for it. */
 	wait_for_completion(&rcu.completion);
+	destroy_rcu_head_on_stack(&rcu.head);
 }
 EXPORT_SYMBOL_GPL(synchronize_rcu_bh);
 
@@ -1498,8 +1537,20 @@ static int __rcu_pending(struct rcu_state *rsp, struct rcu_data *rdp)
 	check_cpu_stall(rsp, rdp);
 
 	/* Is the RCU core waiting for a quiescent state from this CPU? */
-	if (rdp->qs_pending) {
+	if (rdp->qs_pending && !rdp->passed_quiesc) {
+
+		/*
+		 * If force_quiescent_state() coming soon and this CPU
+		 * needs a quiescent state, and this is either RCU-sched
+		 * or RCU-bh, force a local reschedule.
+		 */
 		rdp->n_rp_qs_pending++;
+		if (!rdp->preemptable &&
+		    ULONG_CMP_LT(ACCESS_ONCE(rsp->jiffies_force_qs) - 1,
+				 jiffies))
+			set_need_resched();
+	} else if (rdp->qs_pending && rdp->passed_quiesc) {
+		rdp->n_rp_report_qs++;
 		return 1;
 	}
 
@@ -1767,6 +1818,21 @@ static int __cpuinit rcu_cpu_notify(struct notifier_block *self,
 }
 
 /*
+ * This function is invoked towards the end of the scheduler's initialization
+ * process.  Before this is called, the idle task might contain
+ * RCU read-side critical sections (during which time, this idle
+ * task is booting the system).  After this function is called, the
+ * idle tasks are prohibited from containing RCU read-side critical
+ * sections.  This function also enables RCU lockdep checking.
+ */
+void rcu_scheduler_starting(void)
+{
+	WARN_ON(num_online_cpus() != 1);
+	WARN_ON(nr_context_switches() > 0);
+	rcu_scheduler_active = 1;
+}
+
+/*
  * Compute the per-level fanout, either using the exact fanout specified
  * or balancing the tree, depending on CONFIG_RCU_FANOUT_EXACT.
  */
@@ -1849,6 +1915,14 @@ static void __init rcu_init_one(struct rcu_state *rsp)
 			INIT_LIST_HEAD(&rnp->blocked_tasks[3]);
 		}
 	}
+
+	rnp = rsp->level[NUM_RCU_LVLS - 1];
+	for_each_possible_cpu(i) {
+		while (i > rnp->grphi)
+			rnp++;
+		rsp->rda[i]->mynode = rnp;
+		rcu_boot_init_percpu_data(i, rsp);
+	}
 }
 
 /*
@@ -1859,19 +1933,11 @@ static void __init rcu_init_one(struct rcu_state *rsp)
 #define RCU_INIT_FLAVOR(rsp, rcu_data) \
 do { \
 	int i; \
-	int j; \
-	struct rcu_node *rnp; \
 	\
-	rcu_init_one(rsp); \
-	rnp = (rsp)->level[NUM_RCU_LVLS - 1]; \
-	j = 0; \
 	for_each_possible_cpu(i) { \
-		if (i > rnp[j].grphi) \
-			j++; \
-		per_cpu(rcu_data, i).mynode = &rnp[j]; \
 		(rsp)->rda[i] = &per_cpu(rcu_data, i); \
-		rcu_boot_init_percpu_data(i, rsp); \
 	} \
+	rcu_init_one(rsp); \
 } while (0)
 
 void __init rcu_init(void)
@@ -1879,12 +1945,6 @@ void __init rcu_init(void)
 	int cpu;
 
 	rcu_bootup_announce();
-#ifdef CONFIG_RCU_CPU_STALL_DETECTOR
-	printk(KERN_INFO "RCU-based detection of stalled CPUs is enabled.\n");
-#endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */
-#if NUM_RCU_LVL_4 != 0
-	printk(KERN_INFO "Experimental four-level hierarchy is enabled.\n");
-#endif /* #if NUM_RCU_LVL_4 != 0 */
 	RCU_INIT_FLAVOR(&rcu_sched_state, rcu_sched_data);
 	RCU_INIT_FLAVOR(&rcu_bh_state, rcu_bh_data);
 	__rcu_init_preempt();
@@ -1898,6 +1958,7 @@ void __init rcu_init(void)
 	cpu_notifier(rcu_cpu_notify, 0);
 	for_each_online_cpu(cpu)
 		rcu_cpu_notify(NULL, CPU_UP_PREPARE, (void *)(long)cpu);
+	check_cpu_stall_init();
 }
 
 #include "rcutree_plugin.h"
diff --git a/kernel/rcutree.h b/kernel/rcutree.h
index 4a525a30e08e..14c040b18ed0 100644
--- a/kernel/rcutree.h
+++ b/kernel/rcutree.h
@@ -223,6 +223,7 @@ struct rcu_data {
 	/* 5) __rcu_pending() statistics. */
 	unsigned long n_rcu_pending;	/* rcu_pending() calls since boot. */
 	unsigned long n_rp_qs_pending;
+	unsigned long n_rp_report_qs;
 	unsigned long n_rp_cb_ready;
 	unsigned long n_rp_cpu_needs_gp;
 	unsigned long n_rp_gp_completed;
@@ -326,6 +327,7 @@ struct rcu_state {
 	unsigned long jiffies_stall;		/* Time at which to check */
 						/*  for CPU stalls. */
 #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */
+	char *name;				/* Name of structure. */
 };
 
 /* Return values for rcu_preempt_offline_tasks(). */
diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h
index 79b53bda8943..0e4f420245d9 100644
--- a/kernel/rcutree_plugin.h
+++ b/kernel/rcutree_plugin.h
@@ -26,6 +26,45 @@
 
 #include <linux/delay.h>
 
+/*
+ * Check the RCU kernel configuration parameters and print informative
+ * messages about anything out of the ordinary.  If you like #ifdef, you
+ * will love this function.
+ */
+static void __init rcu_bootup_announce_oddness(void)
+{
+#ifdef CONFIG_RCU_TRACE
+	printk(KERN_INFO "\tRCU debugfs-based tracing is enabled.\n");
+#endif
+#if (defined(CONFIG_64BIT) && CONFIG_RCU_FANOUT != 64) || (!defined(CONFIG_64BIT) && CONFIG_RCU_FANOUT != 32)
+	printk(KERN_INFO "\tCONFIG_RCU_FANOUT set to non-default value of %d\n",
+	       CONFIG_RCU_FANOUT);
+#endif
+#ifdef CONFIG_RCU_FANOUT_EXACT
+	printk(KERN_INFO "\tHierarchical RCU autobalancing is disabled.\n");
+#endif
+#ifdef CONFIG_RCU_FAST_NO_HZ
+	printk(KERN_INFO
+	       "\tRCU dyntick-idle grace-period acceleration is enabled.\n");
+#endif
+#ifdef CONFIG_PROVE_RCU
+	printk(KERN_INFO "\tRCU lockdep checking is enabled.\n");
+#endif
+#ifdef CONFIG_RCU_TORTURE_TEST_RUNNABLE
+	printk(KERN_INFO "\tRCU torture testing starts during boot.\n");
+#endif
+#ifndef CONFIG_RCU_CPU_STALL_DETECTOR
+	printk(KERN_INFO
+	       "\tRCU-based detection of stalled CPUs is disabled.\n");
+#endif
+#ifndef CONFIG_RCU_CPU_STALL_VERBOSE
+	printk(KERN_INFO "\tVerbose stalled-CPUs detection is disabled.\n");
+#endif
+#if NUM_RCU_LVL_4 != 0
+	printk(KERN_INFO "\tExperimental four-level hierarchy is enabled.\n");
+#endif
+}
+
 #ifdef CONFIG_TREE_PREEMPT_RCU
 
 struct rcu_state rcu_preempt_state = RCU_STATE_INITIALIZER(rcu_preempt_state);
@@ -38,8 +77,8 @@ static int rcu_preempted_readers_exp(struct rcu_node *rnp);
  */
 static void __init rcu_bootup_announce(void)
 {
-	printk(KERN_INFO
-	       "Experimental preemptable hierarchical RCU implementation.\n");
+	printk(KERN_INFO "Preemptable hierarchical RCU implementation.\n");
+	rcu_bootup_announce_oddness();
 }
 
 /*
@@ -75,13 +114,19 @@ EXPORT_SYMBOL_GPL(rcu_force_quiescent_state);
  * that this just means that the task currently running on the CPU is
  * not in a quiescent state.  There might be any number of tasks blocked
  * while in an RCU read-side critical section.
+ *
+ * Unlike the other rcu_*_qs() functions, callers to this function
+ * must disable irqs in order to protect the assignment to
+ * ->rcu_read_unlock_special.
  */
 static void rcu_preempt_qs(int cpu)
 {
 	struct rcu_data *rdp = &per_cpu(rcu_preempt_data, cpu);
+
 	rdp->passed_quiesc_completed = rdp->gpnum - 1;
 	barrier();
 	rdp->passed_quiesc = 1;
+	current->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_NEED_QS;
 }
 
 /*
@@ -144,9 +189,8 @@ static void rcu_preempt_note_context_switch(int cpu)
 	 * grace period, then the fact that the task has been enqueued
 	 * means that we continue to block the current grace period.
 	 */
-	rcu_preempt_qs(cpu);
 	local_irq_save(flags);
-	t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_NEED_QS;
+	rcu_preempt_qs(cpu);
 	local_irq_restore(flags);
 }
 
@@ -236,7 +280,6 @@ static void rcu_read_unlock_special(struct task_struct *t)
 	 */
 	special = t->rcu_read_unlock_special;
 	if (special & RCU_READ_UNLOCK_NEED_QS) {
-		t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_NEED_QS;
 		rcu_preempt_qs(smp_processor_id());
 	}
 
@@ -473,7 +516,6 @@ static void rcu_preempt_check_callbacks(int cpu)
 	struct task_struct *t = current;
 
 	if (t->rcu_read_lock_nesting == 0) {
-		t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_NEED_QS;
 		rcu_preempt_qs(cpu);
 		return;
 	}
@@ -515,11 +557,13 @@ void synchronize_rcu(void)
 	if (!rcu_scheduler_active)
 		return;
 
+	init_rcu_head_on_stack(&rcu.head);
 	init_completion(&rcu.completion);
 	/* Will wake me after RCU finished. */
 	call_rcu(&rcu.head, wakeme_after_rcu);
 	/* Wait for it. */
 	wait_for_completion(&rcu.completion);
+	destroy_rcu_head_on_stack(&rcu.head);
 }
 EXPORT_SYMBOL_GPL(synchronize_rcu);
 
@@ -754,6 +798,7 @@ void exit_rcu(void)
 static void __init rcu_bootup_announce(void)
 {
 	printk(KERN_INFO "Hierarchical RCU implementation.\n");
+	rcu_bootup_announce_oddness();
 }
 
 /*
@@ -1008,6 +1053,8 @@ static DEFINE_PER_CPU(unsigned long, rcu_dyntick_holdoff);
 int rcu_needs_cpu(int cpu)
 {
 	int c = 0;
+	int snap;
+	int snap_nmi;
 	int thatcpu;
 
 	/* Check for being in the holdoff period. */
@@ -1015,12 +1062,18 @@ int rcu_needs_cpu(int cpu)
 		return rcu_needs_cpu_quick_check(cpu);
 
 	/* Don't bother unless we are the last non-dyntick-idle CPU. */
-	for_each_cpu_not(thatcpu, nohz_cpu_mask)
-		if (thatcpu != cpu) {
+	for_each_online_cpu(thatcpu) {
+		if (thatcpu == cpu)
+			continue;
+		snap = per_cpu(rcu_dynticks, thatcpu).dynticks;
+		snap_nmi = per_cpu(rcu_dynticks, thatcpu).dynticks_nmi;
+		smp_mb(); /* Order sampling of snap with end of grace period. */
+		if (((snap & 0x1) != 0) || ((snap_nmi & 0x1) != 0)) {
 			per_cpu(rcu_dyntick_drain, cpu) = 0;
 			per_cpu(rcu_dyntick_holdoff, cpu) = jiffies - 1;
 			return rcu_needs_cpu_quick_check(cpu);
 		}
+	}
 
 	/* Check and update the rcu_dyntick_drain sequencing. */
 	if (per_cpu(rcu_dyntick_drain, cpu) <= 0) {
diff --git a/kernel/rcutree_trace.c b/kernel/rcutree_trace.c
index d45db2e35d27..36c95b45738e 100644
--- a/kernel/rcutree_trace.c
+++ b/kernel/rcutree_trace.c
@@ -241,11 +241,13 @@ static const struct file_operations rcugp_fops = {
 static void print_one_rcu_pending(struct seq_file *m, struct rcu_data *rdp)
 {
 	seq_printf(m, "%3d%cnp=%ld "
-		   "qsp=%ld cbr=%ld cng=%ld gpc=%ld gps=%ld nf=%ld nn=%ld\n",
+		   "qsp=%ld rpq=%ld cbr=%ld cng=%ld "
+		   "gpc=%ld gps=%ld nf=%ld nn=%ld\n",
 		   rdp->cpu,
 		   cpu_is_offline(rdp->cpu) ? '!' : ' ',
 		   rdp->n_rcu_pending,
 		   rdp->n_rp_qs_pending,
+		   rdp->n_rp_report_qs,
 		   rdp->n_rp_cb_ready,
 		   rdp->n_rp_cpu_needs_gp,
 		   rdp->n_rp_gp_completed,
diff --git a/kernel/sched.c b/kernel/sched.c
index 78554dd0d1a4..1d93cd0ae4d3 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -3608,7 +3608,7 @@ need_resched:
 	preempt_disable();
 	cpu = smp_processor_id();
 	rq = cpu_rq(cpu);
-	rcu_sched_qs(cpu);
+	rcu_note_context_switch(cpu);
 	prev = rq->curr;
 	switch_count = &prev->nivcsw;
 
diff --git a/kernel/sched_debug.c b/kernel/sched_debug.c
index 9cf1baf6616a..87a330a7185f 100644
--- a/kernel/sched_debug.c
+++ b/kernel/sched_debug.c
@@ -114,7 +114,9 @@ print_task(struct seq_file *m, struct rq *rq, struct task_struct *p)
 	{
 		char path[64];
 
+		rcu_read_lock();
 		cgroup_path(task_group(p)->css.cgroup, path, sizeof(path));
+		rcu_read_unlock();
 		SEQ_printf(m, " %s", path);
 	}
 #endif
diff --git a/kernel/softirq.c b/kernel/softirq.c
index 7c1a67ef0274..0db913a5c60f 100644
--- a/kernel/softirq.c
+++ b/kernel/softirq.c
@@ -716,7 +716,7 @@ static int run_ksoftirqd(void * __bind_cpu)
 			preempt_enable_no_resched();
 			cond_resched();
 			preempt_disable();
-			rcu_sched_qs((long)__bind_cpu);
+			rcu_note_context_switch((long)__bind_cpu);
 		}
 		preempt_enable();
 		set_current_state(TASK_INTERRUPTIBLE);
diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c
index ef51d1fcf5e6..b4e7431e7c78 100644
--- a/kernel/stop_machine.c
+++ b/kernel/stop_machine.c
@@ -294,7 +294,6 @@ static int __cpuinit cpu_stop_cpu_callback(struct notifier_block *nfb,
 	struct sched_param param = { .sched_priority = MAX_RT_PRIO - 1 };
 	unsigned int cpu = (unsigned long)hcpu;
 	struct cpu_stopper *stopper = &per_cpu(cpu_stopper, cpu);
-	struct cpu_stop_work *work;
 	struct task_struct *p;
 
 	switch (action & ~CPU_TASKS_FROZEN) {
@@ -323,6 +322,9 @@ static int __cpuinit cpu_stop_cpu_callback(struct notifier_block *nfb,
 #ifdef CONFIG_HOTPLUG_CPU
 	case CPU_UP_CANCELED:
 	case CPU_DEAD:
+	{
+		struct cpu_stop_work *work;
+
 		/* kill the stopper */
 		kthread_stop(stopper->thread);
 		/* drain remaining works */
@@ -335,6 +337,7 @@ static int __cpuinit cpu_stop_cpu_callback(struct notifier_block *nfb,
 		put_task_struct(stopper->thread);
 		stopper->thread = NULL;
 		break;
+	}
 #endif
 	}
 
diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c
index 0a47e8d6b491..26b8607a0abc 100644
--- a/kernel/trace/trace_event_perf.c
+++ b/kernel/trace/trace_event_perf.c
@@ -9,13 +9,9 @@
 #include <linux/kprobes.h>
 #include "trace.h"
 
-DEFINE_PER_CPU(struct pt_regs, perf_trace_regs);
-EXPORT_PER_CPU_SYMBOL_GPL(perf_trace_regs);
-
 EXPORT_SYMBOL_GPL(perf_arch_fetch_caller_regs);
 
-static char *perf_trace_buf;
-static char *perf_trace_buf_nmi;
+static char *perf_trace_buf[4];
 
 /*
  * Force it to be aligned to unsigned long to avoid misaligned accesses
@@ -27,63 +23,82 @@ typedef typeof(unsigned long [PERF_MAX_TRACE_SIZE / sizeof(unsigned long)])
 /* Count the events in use (per event id, not per instance) */
 static int	total_ref_count;
 
-static int perf_trace_event_enable(struct ftrace_event_call *event)
+static int perf_trace_event_init(struct ftrace_event_call *tp_event,
+				 struct perf_event *p_event)
 {
-	char *buf;
+	struct hlist_head *list;
 	int ret = -ENOMEM;
+	int cpu;
 
-	if (event->perf_refcount++ > 0)
+	p_event->tp_event = tp_event;
+	if (tp_event->perf_refcount++ > 0)
 		return 0;
 
-	if (!total_ref_count) {
-		buf = (char *)alloc_percpu(perf_trace_t);
-		if (!buf)
-			goto fail_buf;
+	list = alloc_percpu(struct hlist_head);
+	if (!list)
+		goto fail;
+
+	for_each_possible_cpu(cpu)
+		INIT_HLIST_HEAD(per_cpu_ptr(list, cpu));
 
-		rcu_assign_pointer(perf_trace_buf, buf);
+	tp_event->perf_events = list;
+
+	if (!total_ref_count) {
+		char *buf;
+		int i;
 
-		buf = (char *)alloc_percpu(perf_trace_t);
-		if (!buf)
-			goto fail_buf_nmi;
+		for (i = 0; i < 4; i++) {
+			buf = (char *)alloc_percpu(perf_trace_t);
+			if (!buf)
+				goto fail;
 
-		rcu_assign_pointer(perf_trace_buf_nmi, buf);
+			perf_trace_buf[i] = buf;
+		}
 	}
 
-	if (event->class->reg)
-		ret = event->class->reg(event, TRACE_REG_PERF_REGISTER);
+	if (tp_event->class->reg)
+		ret = tp_event->class->reg(tp_event, TRACE_REG_PERF_REGISTER);
 	else
-		ret = tracepoint_probe_register(event->name,
-						event->class->perf_probe,
-						event);
-	if (!ret) {
-		total_ref_count++;
-		return 0;
-	}
+		ret = tracepoint_probe_register(tp_event->name,
+						tp_event->class->perf_probe,
+						tp_event);
+
+	if (ret)
+		goto fail;
 
-fail_buf_nmi:
+	total_ref_count++;
+	return 0;
+
+fail:
 	if (!total_ref_count) {
-		free_percpu(perf_trace_buf_nmi);
-		free_percpu(perf_trace_buf);
-		perf_trace_buf_nmi = NULL;
-		perf_trace_buf = NULL;
+		int i;
+
+		for (i = 0; i < 4; i++) {
+			free_percpu(perf_trace_buf[i]);
+			perf_trace_buf[i] = NULL;
+		}
+	}
+
+	if (!--tp_event->perf_refcount) {
+		free_percpu(tp_event->perf_events);
+		tp_event->perf_events = NULL;
 	}
-fail_buf:
-	event->perf_refcount--;
 
 	return ret;
 }
 
-int perf_trace_enable(int event_id)
+int perf_trace_init(struct perf_event *p_event)
 {
-	struct ftrace_event_call *event;
+	struct ftrace_event_call *tp_event;
+	int event_id = p_event->attr.config;
 	int ret = -EINVAL;
 
 	mutex_lock(&event_mutex);
-	list_for_each_entry(event, &ftrace_events, list) {
-		if (event->event.type == event_id &&
-		    event->class && event->class->perf_probe &&
-		    try_module_get(event->mod)) {
-			ret = perf_trace_event_enable(event);
+	list_for_each_entry(tp_event, &ftrace_events, list) {
+		if (tp_event->event.type == event_id &&
+		    tp_event->class && tp_event->class->perf_probe &&
+		    try_module_get(tp_event->mod)) {
+			ret = perf_trace_event_init(tp_event, p_event);
 			break;
 		}
 	}
@@ -92,93 +107,76 @@ int perf_trace_enable(int event_id)
 	return ret;
 }
 
-static void perf_trace_event_disable(struct ftrace_event_call *event)
+int perf_trace_enable(struct perf_event *p_event)
 {
-	char *buf, *nmi_buf;
-
-	if (--event->perf_refcount > 0)
-		return;
-
-	if (event->class->reg)
-		event->class->reg(event, TRACE_REG_PERF_UNREGISTER);
-	else
-		tracepoint_probe_unregister(event->name, event->class->perf_probe, event);
+	struct ftrace_event_call *tp_event = p_event->tp_event;
+	struct hlist_head *list;
 
-	if (!--total_ref_count) {
-		buf = perf_trace_buf;
-		rcu_assign_pointer(perf_trace_buf, NULL);
+	list = tp_event->perf_events;
+	if (WARN_ON_ONCE(!list))
+		return -EINVAL;
 
-		nmi_buf = perf_trace_buf_nmi;
-		rcu_assign_pointer(perf_trace_buf_nmi, NULL);
+	list = per_cpu_ptr(list, smp_processor_id());
+	hlist_add_head_rcu(&p_event->hlist_entry, list);
 
-		/*
-		 * Ensure every events in profiling have finished before
-		 * releasing the buffers
-		 */
-		synchronize_sched();
+	return 0;
+}
 
-		free_percpu(buf);
-		free_percpu(nmi_buf);
-	}
+void perf_trace_disable(struct perf_event *p_event)
+{
+	hlist_del_rcu(&p_event->hlist_entry);
 }
 
-void perf_trace_disable(int event_id)
+void perf_trace_destroy(struct perf_event *p_event)
 {
-	struct ftrace_event_call *event;
+	struct ftrace_event_call *tp_event = p_event->tp_event;
+	int i;
 
-	mutex_lock(&event_mutex);
-	list_for_each_entry(event, &ftrace_events, list) {
-		if (event->event.type == event_id) {
-			perf_trace_event_disable(event);
-			module_put(event->mod);
-			break;
+	if (--tp_event->perf_refcount > 0)
+		return;
+
+	if (tp_event->class->reg)
+		tp_event->class->reg(tp_event, TRACE_REG_PERF_UNREGISTER);
+	else
+		tracepoint_probe_unregister(tp_event->name,
+					    tp_event->class->perf_probe,
+					    tp_event);
+
+	free_percpu(tp_event->perf_events);
+	tp_event->perf_events = NULL;
+
+	if (!--total_ref_count) {
+		for (i = 0; i < 4; i++) {
+			free_percpu(perf_trace_buf[i]);
+			perf_trace_buf[i] = NULL;
 		}
 	}
-	mutex_unlock(&event_mutex);
 }
 
 __kprobes void *perf_trace_buf_prepare(int size, unsigned short type,
-				       int *rctxp, unsigned long *irq_flags)
+				       struct pt_regs *regs, int *rctxp)
 {
 	struct trace_entry *entry;
-	char *trace_buf, *raw_data;
-	int pc, cpu;
+	char *raw_data;
+	int pc;
 
 	BUILD_BUG_ON(PERF_MAX_TRACE_SIZE % sizeof(unsigned long));
 
 	pc = preempt_count();
 
-	/* Protect the per cpu buffer, begin the rcu read side */
-	local_irq_save(*irq_flags);
-
 	*rctxp = perf_swevent_get_recursion_context();
 	if (*rctxp < 0)
-		goto err_recursion;
-
-	cpu = smp_processor_id();
-
-	if (in_nmi())
-		trace_buf = rcu_dereference_sched(perf_trace_buf_nmi);
-	else
-		trace_buf = rcu_dereference_sched(perf_trace_buf);
-
-	if (!trace_buf)
-		goto err;
+		return NULL;
 
-	raw_data = per_cpu_ptr(trace_buf, cpu);
+	raw_data = per_cpu_ptr(perf_trace_buf[*rctxp], smp_processor_id());
 
 	/* zero the dead bytes from align to not leak stack to user */
 	memset(&raw_data[size - sizeof(u64)], 0, sizeof(u64));
 
 	entry = (struct trace_entry *)raw_data;
-	tracing_generic_entry_update(entry, *irq_flags, pc);
+	tracing_generic_entry_update(entry, regs->flags, pc);
 	entry->type = type;
 
 	return raw_data;
-err:
-	perf_swevent_put_recursion_context(*rctxp);
-err_recursion:
-	local_irq_restore(*irq_flags);
-	return NULL;
 }
 EXPORT_SYMBOL_GPL(perf_trace_buf_prepare);
diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c
index 9a082bba9537..faf7cefd15da 100644
--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -1338,9 +1338,9 @@ static __kprobes void kprobe_perf_func(struct kprobe *kp,
 	struct trace_probe *tp = container_of(kp, struct trace_probe, rp.kp);
 	struct ftrace_event_call *call = &tp->call;
 	struct kprobe_trace_entry_head *entry;
+	struct hlist_head *head;
 	u8 *data;
 	int size, __size, i;
-	unsigned long irq_flags;
 	int rctx;
 
 	__size = sizeof(*entry) + tp->size;
@@ -1350,8 +1350,7 @@ static __kprobes void kprobe_perf_func(struct kprobe *kp,
 		     "profile buffer not large enough"))
 		return;
 
-	entry = perf_trace_buf_prepare(size, call->event.type,
-				       &rctx, &irq_flags);
+	entry = perf_trace_buf_prepare(size, call->event.type, regs, &rctx);
 	if (!entry)
 		return;
 
@@ -1360,7 +1359,8 @@ static __kprobes void kprobe_perf_func(struct kprobe *kp,
 	for (i = 0; i < tp->nr_args; i++)
 		call_fetch(&tp->args[i].fetch, regs, data + tp->args[i].offset);
 
-	perf_trace_buf_submit(entry, size, rctx, entry->ip, 1, irq_flags, regs);
+	head = per_cpu_ptr(call->perf_events, smp_processor_id());
+	perf_trace_buf_submit(entry, size, rctx, entry->ip, 1, regs, head);
 }
 
 /* Kretprobe profile handler */
@@ -1370,9 +1370,9 @@ static __kprobes void kretprobe_perf_func(struct kretprobe_instance *ri,
 	struct trace_probe *tp = container_of(ri->rp, struct trace_probe, rp);
 	struct ftrace_event_call *call = &tp->call;
 	struct kretprobe_trace_entry_head *entry;
+	struct hlist_head *head;
 	u8 *data;
 	int size, __size, i;
-	unsigned long irq_flags;
 	int rctx;
 
 	__size = sizeof(*entry) + tp->size;
@@ -1382,8 +1382,7 @@ static __kprobes void kretprobe_perf_func(struct kretprobe_instance *ri,
 		     "profile buffer not large enough"))
 		return;
 
-	entry = perf_trace_buf_prepare(size, call->event.type,
-				       &rctx, &irq_flags);
+	entry = perf_trace_buf_prepare(size, call->event.type, regs, &rctx);
 	if (!entry)
 		return;
 
@@ -1393,8 +1392,8 @@ static __kprobes void kretprobe_perf_func(struct kretprobe_instance *ri,
 	for (i = 0; i < tp->nr_args; i++)
 		call_fetch(&tp->args[i].fetch, regs, data + tp->args[i].offset);
 
-	perf_trace_buf_submit(entry, size, rctx, entry->ret_ip, 1,
-			       irq_flags, regs);
+	head = per_cpu_ptr(call->perf_events, smp_processor_id());
+	perf_trace_buf_submit(entry, size, rctx, entry->ret_ip, 1, regs, head);
 }
 
 static int probe_perf_enable(struct ftrace_event_call *call)
diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index 9d358301ae3e..d2c859cec9ea 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -488,7 +488,7 @@ static void perf_syscall_enter(void *ignore, struct pt_regs *regs, long id)
 {
 	struct syscall_metadata *sys_data;
 	struct syscall_trace_enter *rec;
-	unsigned long flags;
+	struct hlist_head *head;
 	int syscall_nr;
 	int rctx;
 	int size;
@@ -511,15 +511,16 @@ static void perf_syscall_enter(void *ignore, struct pt_regs *regs, long id)
 		return;
 
 	rec = (struct syscall_trace_enter *)perf_trace_buf_prepare(size,
-				sys_data->enter_event->event.type,
-				&rctx, &flags);
+				sys_data->enter_event->event.type, regs, &rctx);
 	if (!rec)
 		return;
 
 	rec->nr = syscall_nr;
 	syscall_get_arguments(current, regs, 0, sys_data->nb_args,
 			       (unsigned long *)&rec->args);
-	perf_trace_buf_submit(rec, size, rctx, 0, 1, flags, regs);
+
+	head = per_cpu_ptr(sys_data->enter_event->perf_events, smp_processor_id());
+	perf_trace_buf_submit(rec, size, rctx, 0, 1, regs, head);
 }
 
 int perf_sysenter_enable(struct ftrace_event_call *call)
@@ -561,7 +562,7 @@ static void perf_syscall_exit(void *ignore, struct pt_regs *regs, long ret)
 {
 	struct syscall_metadata *sys_data;
 	struct syscall_trace_exit *rec;
-	unsigned long flags;
+	struct hlist_head *head;
 	int syscall_nr;
 	int rctx;
 	int size;
@@ -587,15 +588,15 @@ static void perf_syscall_exit(void *ignore, struct pt_regs *regs, long ret)
 		return;
 
 	rec = (struct syscall_trace_exit *)perf_trace_buf_prepare(size,
-				sys_data->exit_event->event.type,
-				&rctx, &flags);
+				sys_data->exit_event->event.type, regs, &rctx);
 	if (!rec)
 		return;
 
 	rec->nr = syscall_nr;
 	rec->ret = syscall_get_return_value(current, regs);
 
-	perf_trace_buf_submit(rec, size, rctx, 0, 1, flags, regs);
+	head = per_cpu_ptr(sys_data->exit_event->perf_events, smp_processor_id());
+	perf_trace_buf_submit(rec, size, rctx, 0, 1, regs, head);
 }
 
 int perf_sysexit_enable(struct ftrace_event_call *call)