Merge branch 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip

* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (162 commits)
  tracing/kprobes: unregister_trace_probe needs to be called under mutex
  perf: expose event__process function
  perf events: Fix mmap offset determination
  perf, powerpc: fsl_emb: Restore setting perf_sample_data.period
  perf, powerpc: Convert the FSL driver to use local64_t
  perf tools: Don't keep unreferenced maps when unmaps are detected
  perf session: Invalidate last_match when removing threads from rb_tree
  perf session: Free the ref_reloc_sym memory at the right place
  x86,mmiotrace: Add support for tracing STOS instruction
  perf, sched migration: Librarize task states and event headers helpers
  perf, sched migration: Librarize the GUI class
  perf, sched migration: Make the GUI class client agnostic
  perf, sched migration: Make it vertically scrollable
  perf, sched migration: Parameterize cpu height and spacing
  perf, sched migration: Fix key bindings
  perf, sched migration: Ignore unhandled task states
  perf, sched migration: Handle ignored migrate out events
  perf: New migration tool overview
  tracing: Drop cpparg() macro
  perf: Use tracepoint_synchronize_unregister() to flush any pending tracepoint call
  ...

Fix up trivial conflicts in Makefile and drivers/cpufreq/cpufreq.c

33 files changed, 1451 insertions, 2925 deletions

diff --git a/kernel/Makefile b/kernel/Makefile
index 057472fbc27..ce53fb2bd1d 100644
--- a/kernel/Makefile
+++ b/kernel/Makefile
@@ -76,8 +76,8 @@ obj-$(CONFIG_GCOV_KERNEL) += gcov/
 obj-$(CONFIG_AUDIT_TREE) += audit_tree.o
 obj-$(CONFIG_KPROBES) += kprobes.o
 obj-$(CONFIG_KGDB) += debug/
-obj-$(CONFIG_DETECT_SOFTLOCKUP) += softlockup.o
 obj-$(CONFIG_DETECT_HUNG_TASK) += hung_task.o
+obj-$(CONFIG_LOCKUP_DETECTOR) += watchdog.o
 obj-$(CONFIG_GENERIC_HARDIRQS) += irq/
 obj-$(CONFIG_SECCOMP) += seccomp.o
 obj-$(CONFIG_RCU_TORTURE_TEST) += rcutorture.o
diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c
index 71ed3ce29e1..d71a987fd2b 100644
--- a/kernel/hw_breakpoint.c
+++ b/kernel/hw_breakpoint.c
@@ -41,6 +41,7 @@
 #include <linux/sched.h>
 #include <linux/init.h>
 #include <linux/slab.h>
+#include <linux/list.h>
 #include <linux/cpu.h>
 #include <linux/smp.h>
 
@@ -62,6 +63,9 @@ static DEFINE_PER_CPU(unsigned int, nr_bp_flexible[TYPE_MAX]);
 
 static int nr_slots[TYPE_MAX];
 
+/* Keep track of the breakpoints attached to tasks */
+static LIST_HEAD(bp_task_head);
+
 static int constraints_initialized;
 
 /* Gather the number of total pinned and un-pinned bp in a cpuset */
@@ -103,33 +107,21 @@ static unsigned int max_task_bp_pinned(int cpu, enum bp_type_idx type)
 	return 0;
 }
 
-static int task_bp_pinned(struct task_struct *tsk, enum bp_type_idx type)
+/*
+ * Count the number of breakpoints of the same type and same task.
+ * The given event must be not on the list.
+ */
+static int task_bp_pinned(struct perf_event *bp, enum bp_type_idx type)
 {
-	struct perf_event_context *ctx = tsk->perf_event_ctxp;
-	struct list_head *list;
-	struct perf_event *bp;
-	unsigned long flags;
+	struct perf_event_context *ctx = bp->ctx;
+	struct perf_event *iter;
 	int count = 0;
 
-	if (WARN_ONCE(!ctx, "No perf context for this task"))
-		return 0;
-
-	list = &ctx->event_list;
-
-	raw_spin_lock_irqsave(&ctx->lock, flags);
-
-	/*
-	 * The current breakpoint counter is not included in the list
-	 * at the open() callback time
-	 */
-	list_for_each_entry(bp, list, event_entry) {
-		if (bp->attr.type == PERF_TYPE_BREAKPOINT)
-			if (find_slot_idx(bp) == type)
-				count += hw_breakpoint_weight(bp);
+	list_for_each_entry(iter, &bp_task_head, hw.bp_list) {
+		if (iter->ctx == ctx && find_slot_idx(iter) == type)
+			count += hw_breakpoint_weight(iter);
 	}
 
-	raw_spin_unlock_irqrestore(&ctx->lock, flags);
-
 	return count;
 }
 
@@ -149,7 +141,7 @@ fetch_bp_busy_slots(struct bp_busy_slots *slots, struct perf_event *bp,
 		if (!tsk)
 			slots->pinned += max_task_bp_pinned(cpu, type);
 		else
-			slots->pinned += task_bp_pinned(tsk, type);
+			slots->pinned += task_bp_pinned(bp, type);
 		slots->flexible = per_cpu(nr_bp_flexible[type], cpu);
 
 		return;
@@ -162,7 +154,7 @@ fetch_bp_busy_slots(struct bp_busy_slots *slots, struct perf_event *bp,
 		if (!tsk)
 			nr += max_task_bp_pinned(cpu, type);
 		else
-			nr += task_bp_pinned(tsk, type);
+			nr += task_bp_pinned(bp, type);
 
 		if (nr > slots->pinned)
 			slots->pinned = nr;
@@ -188,7 +180,7 @@ fetch_this_slot(struct bp_busy_slots *slots, int weight)
 /*
  * Add a pinned breakpoint for the given task in our constraint table
  */
-static void toggle_bp_task_slot(struct task_struct *tsk, int cpu, bool enable,
+static void toggle_bp_task_slot(struct perf_event *bp, int cpu, bool enable,
 				enum bp_type_idx type, int weight)
 {
 	unsigned int *tsk_pinned;
@@ -196,10 +188,11 @@ static void toggle_bp_task_slot(struct task_struct *tsk, int cpu, bool enable,
 	int old_idx = 0;
 	int idx = 0;
 
-	old_count = task_bp_pinned(tsk, type);
+	old_count = task_bp_pinned(bp, type);
 	old_idx = old_count - 1;
 	idx = old_idx + weight;
 
+	/* tsk_pinned[n] is the number of tasks having n breakpoints */
 	tsk_pinned = per_cpu(nr_task_bp_pinned[type], cpu);
 	if (enable) {
 		tsk_pinned[idx]++;
@@ -222,23 +215,30 @@ toggle_bp_slot(struct perf_event *bp, bool enable, enum bp_type_idx type,
 	int cpu = bp->cpu;
 	struct task_struct *tsk = bp->ctx->task;
 
+	/* Pinned counter cpu profiling */
+	if (!tsk) {
+
+		if (enable)
+			per_cpu(nr_cpu_bp_pinned[type], bp->cpu) += weight;
+		else
+			per_cpu(nr_cpu_bp_pinned[type], bp->cpu) -= weight;
+		return;
+	}
+
 	/* Pinned counter task profiling */
-	if (tsk) {
-		if (cpu >= 0) {
-			toggle_bp_task_slot(tsk, cpu, enable, type, weight);
-			return;
-		}
 
+	if (!enable)
+		list_del(&bp->hw.bp_list);
+
+	if (cpu >= 0) {
+		toggle_bp_task_slot(bp, cpu, enable, type, weight);
+	} else {
 		for_each_online_cpu(cpu)
-			toggle_bp_task_slot(tsk, cpu, enable, type, weight);
-		return;
+			toggle_bp_task_slot(bp, cpu, enable, type, weight);
 	}
 
-	/* Pinned counter cpu profiling */
 	if (enable)
-		per_cpu(nr_cpu_bp_pinned[type], bp->cpu) += weight;
-	else
-		per_cpu(nr_cpu_bp_pinned[type], bp->cpu) -= weight;
+		list_add_tail(&bp->hw.bp_list, &bp_task_head);
 }
 
 /*
@@ -312,6 +312,10 @@ static int __reserve_bp_slot(struct perf_event *bp)
 	weight = hw_breakpoint_weight(bp);
 
 	fetch_bp_busy_slots(&slots, bp, type);
+	/*
+	 * Simulate the addition of this breakpoint to the constraints
+	 * and see the result.
+	 */
 	fetch_this_slot(&slots, weight);
 
 	/* Flexible counters need to keep at least one slot */
diff --git a/kernel/perf_event.c b/kernel/perf_event.c
index ff86c558af4..c772a3d4000 100644
--- a/kernel/perf_event.c
+++ b/kernel/perf_event.c
@@ -675,7 +675,6 @@ group_sched_in(struct perf_event *group_event,
 	struct perf_event *event, *partial_group = NULL;
 	const struct pmu *pmu = group_event->pmu;
 	bool txn = false;
-	int ret;
 
 	if (group_event->state == PERF_EVENT_STATE_OFF)
 		return 0;
@@ -703,14 +702,8 @@ group_sched_in(struct perf_event *group_event,
 		}
 	}
 
-	if (!txn)
-		return 0;
-
-	ret = pmu->commit_txn(pmu);
-	if (!ret) {
-		pmu->cancel_txn(pmu);
+	if (!txn || !pmu->commit_txn(pmu))
 		return 0;
-	}
 
 group_error:
 	/*
@@ -1155,9 +1148,9 @@ static void __perf_event_sync_stat(struct perf_event *event,
 	 * In order to keep per-task stats reliable we need to flip the event
 	 * values when we flip the contexts.
 	 */
-	value = atomic64_read(&next_event->count);
-	value = atomic64_xchg(&event->count, value);
-	atomic64_set(&next_event->count, value);
+	value = local64_read(&next_event->count);
+	value = local64_xchg(&event->count, value);
+	local64_set(&next_event->count, value);
 
 	swap(event->total_time_enabled, next_event->total_time_enabled);
 	swap(event->total_time_running, next_event->total_time_running);
@@ -1547,10 +1540,10 @@ static void perf_adjust_period(struct perf_event *event, u64 nsec, u64 count)
 
 	hwc->sample_period = sample_period;
 
-	if (atomic64_read(&hwc->period_left) > 8*sample_period) {
+	if (local64_read(&hwc->period_left) > 8*sample_period) {
 		perf_disable();
 		perf_event_stop(event);
-		atomic64_set(&hwc->period_left, 0);
+		local64_set(&hwc->period_left, 0);
 		perf_event_start(event);
 		perf_enable();
 	}
@@ -1591,7 +1584,7 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx)
 
 		perf_disable();
 		event->pmu->read(event);
-		now = atomic64_read(&event->count);
+		now = local64_read(&event->count);
 		delta = now - hwc->freq_count_stamp;
 		hwc->freq_count_stamp = now;
 
@@ -1743,6 +1736,11 @@ static void __perf_event_read(void *info)
 	event->pmu->read(event);
 }
 
+static inline u64 perf_event_count(struct perf_event *event)
+{
+	return local64_read(&event->count) + atomic64_read(&event->child_count);
+}
+
 static u64 perf_event_read(struct perf_event *event)
 {
 	/*
@@ -1762,7 +1760,7 @@ static u64 perf_event_read(struct perf_event *event)
 		raw_spin_unlock_irqrestore(&ctx->lock, flags);
 	}
 
-	return atomic64_read(&event->count);
+	return perf_event_count(event);
 }
 
 /*
@@ -1883,7 +1881,7 @@ static void free_event_rcu(struct rcu_head *head)
 }
 
 static void perf_pending_sync(struct perf_event *event);
-static void perf_mmap_data_put(struct perf_mmap_data *data);
+static void perf_buffer_put(struct perf_buffer *buffer);
 
 static void free_event(struct perf_event *event)
 {
@@ -1891,7 +1889,7 @@ static void free_event(struct perf_event *event)
 
 	if (!event->parent) {
 		atomic_dec(&nr_events);
-		if (event->attr.mmap)
+		if (event->attr.mmap || event->attr.mmap_data)
 			atomic_dec(&nr_mmap_events);
 		if (event->attr.comm)
 			atomic_dec(&nr_comm_events);
@@ -1899,9 +1897,9 @@ static void free_event(struct perf_event *event)
 			atomic_dec(&nr_task_events);
 	}
 
-	if (event->data) {
-		perf_mmap_data_put(event->data);
-		event->data = NULL;
+	if (event->buffer) {
+		perf_buffer_put(event->buffer);
+		event->buffer = NULL;
 	}
 
 	if (event->destroy)
@@ -2126,13 +2124,13 @@ perf_read(struct file *file, char __user *buf, size_t count, loff_t *ppos)
 static unsigned int perf_poll(struct file *file, poll_table *wait)
 {
 	struct perf_event *event = file->private_data;
-	struct perf_mmap_data *data;
+	struct perf_buffer *buffer;
 	unsigned int events = POLL_HUP;
 
 	rcu_read_lock();
-	data = rcu_dereference(event->data);
-	if (data)
-		events = atomic_xchg(&data->poll, 0);
+	buffer = rcu_dereference(event->buffer);
+	if (buffer)
+		events = atomic_xchg(&buffer->poll, 0);
 	rcu_read_unlock();
 
 	poll_wait(file, &event->waitq, wait);
@@ -2143,7 +2141,7 @@ static unsigned int perf_poll(struct file *file, poll_table *wait)
 static void perf_event_reset(struct perf_event *event)
 {
 	(void)perf_event_read(event);
-	atomic64_set(&event->count, 0);
+	local64_set(&event->count, 0);
 	perf_event_update_userpage(event);
 }
 
@@ -2342,14 +2340,14 @@ static int perf_event_index(struct perf_event *event)
 void perf_event_update_userpage(struct perf_event *event)
 {
 	struct perf_event_mmap_page *userpg;
-	struct perf_mmap_data *data;
+	struct perf_buffer *buffer;
 
 	rcu_read_lock();
-	data = rcu_dereference(event->data);
-	if (!data)
+	buffer = rcu_dereference(event->buffer);
+	if (!buffer)
 		goto unlock;
 
-	userpg = data->user_page;
+	userpg = buffer->user_page;
 
 	/*
 	 * Disable preemption so as to not let the corresponding user-space
@@ -2359,9 +2357,9 @@ void perf_event_update_userpage(struct perf_event *event)
 	++userpg->lock;
 	barrier();
 	userpg->index = perf_event_index(event);
-	userpg->offset = atomic64_read(&event->count);
+	userpg->offset = perf_event_count(event);
 	if (event->state == PERF_EVENT_STATE_ACTIVE)
-		userpg->offset -= atomic64_read(&event->hw.prev_count);
+		userpg->offset -= local64_read(&event->hw.prev_count);
 
 	userpg->time_enabled = event->total_time_enabled +
 			atomic64_read(&event->child_total_time_enabled);
@@ -2376,6 +2374,25 @@ unlock:
 	rcu_read_unlock();
 }
 
+static unsigned long perf_data_size(struct perf_buffer *buffer);
+
+static void
+perf_buffer_init(struct perf_buffer *buffer, long watermark, int flags)
+{
+	long max_size = perf_data_size(buffer);
+
+	if (watermark)
+		buffer->watermark = min(max_size, watermark);
+
+	if (!buffer->watermark)
+		buffer->watermark = max_size / 2;
+
+	if (flags & PERF_BUFFER_WRITABLE)
+		buffer->writable = 1;
+
+	atomic_set(&buffer->refcount, 1);
+}
+
 #ifndef CONFIG_PERF_USE_VMALLOC
 
 /*
@@ -2383,15 +2400,15 @@ unlock:
  */
 
 static struct page *
-perf_mmap_to_page(struct perf_mmap_data *data, unsigned long pgoff)
+perf_mmap_to_page(struct perf_buffer *buffer, unsigned long pgoff)
 {
-	if (pgoff > data->nr_pages)
+	if (pgoff > buffer->nr_pages)
 		return NULL;
 
 	if (pgoff == 0)
-		return virt_to_page(data->user_page);
+		return virt_to_page(buffer->user_page);
 
-	return virt_to_page(data->data_pages[pgoff - 1]);
+	return virt_to_page(buffer->data_pages[pgoff - 1]);
 }
 
 static void *perf_mmap_alloc_page(int cpu)
@@ -2407,42 +2424,44 @@ static void *perf_mmap_alloc_page(int cpu)
 	return page_address(page);
 }
 
-static struct perf_mmap_data *
-perf_mmap_data_alloc(struct perf_event *event, int nr_pages)
+static struct perf_buffer *
+perf_buffer_alloc(int nr_pages, long watermark, int cpu, int flags)
 {
-	struct perf_mmap_data *data;
+	struct perf_buffer *buffer;
 	unsigned long size;
 	int i;
 
-	size = sizeof(struct perf_mmap_data);
+	size = sizeof(struct perf_buffer);
 	size += nr_pages * sizeof(void *);
 
-	data = kzalloc(size, GFP_KERNEL);
-	if (!data)
+	buffer = kzalloc(size, GFP_KERNEL);
+	if (!buffer)
 		goto fail;
 
-	data->user_page = perf_mmap_alloc_page(event->cpu);
-	if (!data->user_page)
+	buffer->user_page = perf_mmap_alloc_page(cpu);
+	if (!buffer->user_page)
 		goto fail_user_page;
 
 	for (i = 0; i < nr_pages; i++) {
-		data->data_pages[i] = perf_mmap_alloc_page(event->cpu);
-		if (!data->data_pages[i])
+		buffer->data_pages[i] = perf_mmap_alloc_page(cpu);
+		if (!buffer->data_pages[i])
 			goto fail_data_pages;
 	}
 
-	data->nr_pages = nr_pages;
+	buffer->nr_pages = nr_pages;
+
+	perf_buffer_init(buffer, watermark, flags);
 
-	return data;
+	return buffer;
 
 fail_data_pages:
 	for (i--; i >= 0; i--)
-		free_page((unsigned long)data->data_pages[i]);
+		free_page((unsigned long)buffer->data_pages[i]);
 
-	free_page((unsigned long)data->user_page);
+	free_page((unsigned long)buffer->user_page);
 
 fail_user_page:
-	kfree(data);
+	kfree(buffer);
 
 fail:
 	return NULL;
@@ -2456,17 +2475,17 @@ static void perf_mmap_free_page(unsigned long addr)
 	__free_page(page);
 }
 
-static void perf_mmap_data_free(struct perf_mmap_data *data)
+static void perf_buffer_free(struct perf_buffer *buffer)
 {
 	int i;
 
-	perf_mmap_free_page((unsigned long)data->user_page);
-	for (i = 0; i < data->nr_pages; i++)
-		perf_mmap_free_page((unsigned long)data->data_pages[i]);
-	kfree(data);
+	perf_mmap_free_page((unsigned long)buffer->user_page);
+	for (i = 0; i < buffer->nr_pages; i++)
+		perf_mmap_free_page((unsigned long)buffer->data_pages[i]);
+	kfree(buffer);
 }
 
-static inline int page_order(struct perf_mmap_data *data)
+static inline int page_order(struct perf_buffer *buffer)
 {
 	return 0;
 }
@@ -2479,18 +2498,18 @@ static inline int page_order(struct perf_mmap_data *data)
  * Required for architectures that have d-cache aliasing issues.
  */
 
-static inline int page_order(struct perf_mmap_data *data)
+static inline int page_order(struct perf_buffer *buffer)
 {
-	return data->page_order;
+	return buffer->page_order;
 }
 
 static struct page *
-perf_mmap_to_page(struct perf_mmap_data *data, unsigned long pgoff)
+perf_mmap_to_page(struct perf_buffer *buffer, unsigned long pgoff)
 {
-	if (pgoff > (1UL << page_order(data)))
+	if (pgoff > (1UL << page_order(buffer)))
 		return NULL;
 
-	return vmalloc_to_page((void *)data->user_page + pgoff * PAGE_SIZE);
+	return vmalloc_to_page((void *)buffer->user_page + pgoff * PAGE_SIZE);
 }
 
 static void perf_mmap_unmark_page(void *addr)
@@ -2500,57 +2519,59 @@ static void perf_mmap_unmark_page(void *addr)
 	page->mapping = NULL;
 }
 
-static void perf_mmap_data_free_work(struct work_struct *work)
+static void perf_buffer_free_work(struct work_struct *work)
 {
-	struct perf_mmap_data *data;
+	struct perf_buffer *buffer;
 	void *base;
 	int i, nr;
 
-	data = container_of(work, struct perf_mmap_data, work);
-	nr = 1 << page_order(data);
+	buffer = container_of(work, struct perf_buffer, work);
+	nr = 1 << page_order(buffer);
 
-	base = data->user_page;
+	base = buffer->user_page;
 	for (i = 0; i < nr + 1; i++)
 		perf_mmap_unmark_page(base + (i * PAGE_SIZE));
 
 	vfree(base);
-	kfree(data);
+	kfree(buffer);
 }
 
-static void perf_mmap_data_free(struct perf_mmap_data *data)
+static void perf_buffer_free(struct perf_buffer *buffer)
 {
-	schedule_work(&data->work);
+	schedule_work(&buffer->work);
 }
 
-static struct perf_mmap_data *
-perf_mmap_data_alloc(struct perf_event *event, int nr_pages)
+static struct perf_buffer *
+perf_buffer_alloc(int nr_pages, long watermark, int cpu, int flags)
 {
-	struct perf_mmap_data *data;
+	struct perf_buffer *buffer;
 	unsigned long size;
 	void *all_buf;
 
-	size = sizeof(struct perf_mmap_data);
+	size = sizeof(struct perf_buffer);
 	size += sizeof(void *);
 
-	data = kzalloc(size, GFP_KERNEL);
-	if (!data)
+	buffer = kzalloc(size, GFP_KERNEL);
+	if (!buffer)
 		goto fail;
 
-	INIT_WORK(&data->work, perf_mmap_data_free_work);
+	INIT_WORK(&buffer->work, perf_buffer_free_work);
 
 	all_buf = vmalloc_user((nr_pages + 1) * PAGE_SIZE);
 	if (!all_buf)
 		goto fail_all_buf;
 
-	data->user_page = all_buf;
-	data->data_pages[0] = all_buf + PAGE_SIZE;
-	data->page_order = ilog2(nr_pages);
-	data->nr_pages = 1;
+	buffer->user_page = all_buf;
+	buffer->data_pages[0] = all_buf + PAGE_SIZE;
+	buffer->page_order = ilog2(nr_pages);
+	buffer->nr_pages = 1;
+
+	perf_buffer_init(buffer, watermark, flags);
 
-	return data;
+	return buffer;
 
 fail_all_buf:
-	kfree(data);
+	kfree(buffer);
 
 fail:
 	return NULL;
@@ -2558,15 +2579,15 @@ fail:
 
 #endif
 
-static unsigned long perf_data_size(struct perf_mmap_data *data)
+static unsigned long perf_data_size(struct perf_buffer *buffer)
 {
-	return data->nr_pages << (PAGE_SHIFT + page_order(data));
+	return buffer->nr_pages << (PAGE_SHIFT + page_order(buffer));
 }
 
 static int perf_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 {
 	struct perf_event *event = vma->vm_file->private_data;
-	struct perf_mmap_data *data;
+	struct perf_buffer *buffer;
 	int ret = VM_FAULT_SIGBUS;
 
 	if (vmf->flags & FAULT_FLAG_MKWRITE) {
@@ -2576,14 +2597,14 @@ static int perf_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 	}
 
 	rcu_read_lock();
-	data = rcu_dereference(event->data);
-	if (!data)
+	buffer = rcu_dereference(event->buffer);
+	if (!buffer)
 		goto unlock;
 
 	if (vmf->pgoff && (vmf->flags & FAULT_FLAG_WRITE))
 		goto unlock;
 
-	vmf->page = perf_mmap_to_page(data, vmf->pgoff);
+	vmf->page = perf_mmap_to_page(buffer, vmf->pgoff);
 	if (!vmf->page)
 		goto unlock;
 
@@ -2598,52 +2619,35 @@ unlock:
 	return ret;
 }
 
-static void
-perf_mmap_data_init(struct perf_event *event, struct perf_mmap_data *data)
-{
-	long max_size = perf_data_size(data);
-
-	if (event->attr.watermark) {
-		data->watermark = min_t(long, max_size,
-					event->attr.wakeup_watermark);
-	}
-
-	if (!data->watermark)
-		data->watermark = max_size / 2;
-
-	atomic_set(&data->refcount, 1);
-	rcu_assign_pointer(event->data, data);
-}
-
-static void perf_mmap_data_free_rcu(struct rcu_head *rcu_head)
+static void perf_buffer_free_rcu(struct rcu_head *rcu_head)
 {
-	struct perf_mmap_data *data;
+	struct perf_buffer *buffer;
 
-	data = container_of(rcu_head, struct perf_mmap_data, rcu_head);
-	perf_mmap_data_free(data);
+	buffer = container_of(rcu_head, struct perf_buffer, rcu_head);
+	perf_buffer_free(buffer);
 }
 
-static struct perf_mmap_data *perf_mmap_data_get(struct perf_event *event)
+static struct perf_buffer *perf_buffer_get(struct perf_event *event)
 {
-	struct perf_mmap_data *data;
+	struct perf_buffer *buffer;
 
 	rcu_read_lock();
-	data = rcu_dereference(event->data);
-	if (data) {
-		if (!atomic_inc_not_zero(&data->refcount))
-			data = NULL;
+	buffer = rcu_dereference(event->buffer);
+	if (buffer) {
+		if (!atomic_inc_not_zero(&buffer->refcount))
+			buffer = NULL;
 	}
 	rcu_read_unlock();
 
-	return data;
+	return buffer;
 }
 
-static void perf_mmap_data_put(struct perf_mmap_data *data)
+static void perf_buffer_put(struct perf_buffer *buffer)
 {
-	if (!atomic_dec_and_test(&data->refcount))
+	if (!atomic_dec_and_test(&buffer->refcount))
 		return;
 
-	call_rcu(&data->rcu_head, perf_mmap_data_free_rcu);
+	call_rcu(&buffer->rcu_head, perf_buffer_free_rcu);
 }
 
 static void perf_mmap_open(struct vm_area_struct *vma)
@@ -2658,16 +2662,16 @@ static void perf_mmap_close(struct vm_area_struct *vma)
 	struct perf_event *event = vma->vm_file->private_data;
 
 	if (atomic_dec_and_mutex_lock(&event->mmap_count, &event->mmap_mutex)) {
-		unsigned long size = perf_data_size(event->data);
+		unsigned long size = perf_data_size(event->buffer);
 		struct user_struct *user = event->mmap_user;
-		struct perf_mmap_data *data = event->data;
+		struct perf_buffer *buffer = event->buffer;
 
 		atomic_long_sub((size >> PAGE_SHIFT) + 1, &user->locked_vm);
 		vma->vm_mm->locked_vm -= event->mmap_locked;
-		rcu_assign_pointer(event->data, NULL);
+		rcu_assign_pointer(event->buffer, NULL);
 		mutex_unlock(&event->mmap_mutex);
 
-		perf_mmap_data_put(data);
+		perf_buffer_put(buffer);
 		free_uid(user);
 	}
 }
@@ -2685,11 +2689,11 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
 	unsigned long user_locked, user_lock_limit;
 	struct user_struct *user = current_user();
 	unsigned long locked, lock_limit;
-	struct perf_mmap_data *data;
+	struct perf_buffer *buffer;
 	unsigned long vma_size;
 	unsigned long nr_pages;
 	long user_extra, extra;
-	int ret = 0;
+	int ret = 0, flags = 0;
 
 	/*
 	 * Don't allow mmap() of inherited per-task counters. This would
@@ -2706,7 +2710,7 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
 	nr_pages = (vma_size / PAGE_SIZE) - 1;
 
 	/*
-	 * If we have data pages ensure they're a power-of-two number, so we
+	 * If we have buffer pages ensure they're a power-of-two number, so we
 	 * can do bitmasks instead of modulo.
 	 */
 	if (nr_pages != 0 && !is_power_of_2(nr_pages))
@@ -2720,9 +2724,9 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
 
 	WARN_ON_ONCE(event->ctx->parent_ctx);
 	mutex_lock(&event->mmap_mutex);
-	if (event->data) {
-		if (event->data->nr_pages == nr_pages)
-			atomic_inc(&event->data->refcount);
+	if (event->buffer) {
+		if (event->buffer->nr_pages == nr_pages)
+			atomic_inc(&event->buffer->refcount);
 		else
 			ret = -EINVAL;
 		goto unlock;
@@ -2752,17 +2756,18 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
 		goto unlock;
 	}
 
-	WARN_ON(event->data);
+	WARN_ON(event->buffer);
+
+	if (vma->vm_flags & VM_WRITE)
+		flags |= PERF_BUFFER_WRITABLE;
 
-	data = perf_mmap_data_alloc(event, nr_pages);
-	if (!data) {
+	buffer = perf_buffer_alloc(nr_pages, event->attr.wakeup_watermark,
+				   event->cpu, flags);
+	if (!buffer) {
 		ret = -ENOMEM;
 		goto unlock;
 	}
-
-	perf_mmap_data_init(event, data);
-	if (vma->vm_flags & VM_WRITE)
-		event->data->writable = 1;
+	rcu_assign_pointer(event->buffer, buffer);
 
 	atomic_long_add(user_extra, &user->locked_vm);
 	event->mmap_locked = extra;
@@ -2941,11 +2946,6 @@ __weak struct perf_callchain_entry *perf_callchain(struct pt_regs *regs)
 	return NULL;
 }
 
-__weak
-void perf_arch_fetch_caller_regs(struct pt_regs *regs, unsigned long ip, int skip)
-{
-}
-
 
 /*
  * We assume there is only KVM supporting the callbacks.
@@ -2971,15 +2971,15 @@ EXPORT_SYMBOL_GPL(perf_unregister_guest_info_callbacks);
 /*
  * Output
  */
-static bool perf_output_space(struct perf_mmap_data *data, unsigned long tail,
+static bool perf_output_space(struct perf_buffer *buffer, unsigned long tail,
 			      unsigned long offset, unsigned long head)
 {
 	unsigned long mask;
 
-	if (!data->writable)
+	if (!buffer->writable)
 		return true;
 
-	mask = perf_data_size(data) - 1;
+	mask = perf_data_size(buffer) - 1;
 
 	offset = (offset - tail) & mask;
 	head   = (head   - tail) & mask;
@@ -2992,7 +2992,7 @@ static bool perf_output_space(struct perf_mmap_data *data, unsigned long tail,
 
 static void perf_output_wakeup(struct perf_output_handle *handle)
 {
-	atomic_set(&handle->data->poll, POLL_IN);
+	atomic_set(&handle->buffer->poll, POLL_IN);
 
 	if (handle->nmi) {
 		handle->event->pending_wakeup = 1;
@@ -3012,45 +3012,45 @@ static void perf_output_wakeup(struct perf_output_handle *handle)
  */
 static void perf_output_get_handle(struct perf_output_handle *handle)
 {
-	struct perf_mmap_data *data = handle->data;
+	struct perf_buffer *buffer = handle->buffer;
 
 	preempt_disable();
-	local_inc(&data->nest);
-	handle->wakeup = local_read(&data->wakeup);
+	local_inc(&buffer->nest);
+	handle->wakeup = local_read(&buffer->wakeup);
 }
 
 static void perf_output_put_handle(struct perf_output_handle *handle)
 {
-	struct perf_mmap_data *data = handle->data;
+	struct perf_buffer *buffer = handle->buffer;
 	unsigned long head;
 
 again:
-	head = local_read(&data->head);
+	head = local_read(&buffer->head);
 
 	/*
 	 * IRQ/NMI can happen here, which means we can miss a head upd