static int max_samples_per_tick __read_mostly =
DIV_ROUND_UP(DEFAULT_MAX_SAMPLE_RATE, HZ);
+static int perf_rotate_context(struct perf_cpu_context *cpuctx);
+
int perf_proc_update_handler(struct ctl_table *table, int write,
void __user *buffer, size_t *lenp,
loff_t *ppos)
static void update_context_time(struct perf_event_context *ctx);
static u64 perf_event_time(struct perf_event *event);
-static void ring_buffer_attach(struct perf_event *event,
- struct ring_buffer *rb);
-
void __weak perf_event_print_debug(void) { }
extern __weak const char *perf_pmu_name(void)
}
#endif
+/*
+ * set default to be dependent on timer tick just
+ * like original code
+ */
+#define PERF_CPU_HRTIMER (1000 / HZ)
+/*
+ * function must be called with interrupts disbled
+ */
+static enum hrtimer_restart perf_cpu_hrtimer_handler(struct hrtimer *hr)
+{
+ struct perf_cpu_context *cpuctx;
+ enum hrtimer_restart ret = HRTIMER_NORESTART;
+ int rotations = 0;
+
+ WARN_ON(!irqs_disabled());
+
+ cpuctx = container_of(hr, struct perf_cpu_context, hrtimer);
+
+ rotations = perf_rotate_context(cpuctx);
+
+ /*
+ * arm timer if needed
+ */
+ if (rotations) {
+ hrtimer_forward_now(hr, cpuctx->hrtimer_interval);
+ ret = HRTIMER_RESTART;
+ }
+
+ return ret;
+}
+
+/* CPU is going down */
+void perf_cpu_hrtimer_cancel(int cpu)
+{
+ struct perf_cpu_context *cpuctx;
+ struct pmu *pmu;
+ unsigned long flags;
+
+ if (WARN_ON(cpu != smp_processor_id()))
+ return;
+
+ local_irq_save(flags);
+
+ rcu_read_lock();
+
+ list_for_each_entry_rcu(pmu, &pmus, entry) {
+ cpuctx = this_cpu_ptr(pmu->pmu_cpu_context);
+
+ if (pmu->task_ctx_nr == perf_sw_context)
+ continue;
+
+ hrtimer_cancel(&cpuctx->hrtimer);
+ }
+
+ rcu_read_unlock();
+
+ local_irq_restore(flags);
+}
+
+static void __perf_cpu_hrtimer_init(struct perf_cpu_context *cpuctx, int cpu)
+{
+ struct hrtimer *hr = &cpuctx->hrtimer;
+ struct pmu *pmu = cpuctx->ctx.pmu;
+ int timer;
+
+ /* no multiplexing needed for SW PMU */
+ if (pmu->task_ctx_nr == perf_sw_context)
+ return;
+
+ /*
+ * check default is sane, if not set then force to
+ * default interval (1/tick)
+ */
+ timer = pmu->hrtimer_interval_ms;
+ if (timer < 1)
+ timer = pmu->hrtimer_interval_ms = PERF_CPU_HRTIMER;
+
+ cpuctx->hrtimer_interval = ns_to_ktime(NSEC_PER_MSEC * timer);
+
+ hrtimer_init(hr, CLOCK_MONOTONIC, HRTIMER_MODE_REL_PINNED);
+ hr->function = perf_cpu_hrtimer_handler;
+}
+
+static void perf_cpu_hrtimer_restart(struct perf_cpu_context *cpuctx)
+{
+ struct hrtimer *hr = &cpuctx->hrtimer;
+ struct pmu *pmu = cpuctx->ctx.pmu;
+
+ /* not for SW PMU */
+ if (pmu->task_ctx_nr == perf_sw_context)
+ return;
+
+ if (hrtimer_active(hr))
+ return;
+
+ if (!hrtimer_callback_running(hr))
+ __hrtimer_start_range_ns(hr, cpuctx->hrtimer_interval,
+ 0, HRTIMER_MODE_REL_PINNED, 0);
+}
+
void perf_pmu_disable(struct pmu *pmu)
{
int *count = this_cpu_ptr(pmu->pmu_disable_count);
if (event_sched_in(group_event, cpuctx, ctx)) {
pmu->cancel_txn(pmu);
+ perf_cpu_hrtimer_restart(cpuctx);
return -EAGAIN;
}
pmu->cancel_txn(pmu);
+ perf_cpu_hrtimer_restart(cpuctx);
+
return -EAGAIN;
}
* If this event can't go on and it's part of a
* group, then the whole group has to come off.
*/
- if (leader != event)
+ if (leader != event) {
group_sched_out(leader, cpuctx, ctx);
+ perf_cpu_hrtimer_restart(cpuctx);
+ }
if (leader->attr.pinned) {
update_group_times(leader);
leader->state = PERF_EVENT_STATE_ERROR;
* because they're strictly cpu affine and rotate_start is called with IRQs
* disabled, while rotate_context is called from IRQ context.
*/
-static void perf_rotate_context(struct perf_cpu_context *cpuctx)
+static int perf_rotate_context(struct perf_cpu_context *cpuctx)
{
struct perf_event_context *ctx = NULL;
int rotate = 0, remove = 1;
done:
if (remove)
list_del_init(&cpuctx->rotation_list);
+
+ return rotate;
}
#ifdef CONFIG_NO_HZ_FULL
ctx = cpuctx->task_ctx;
if (ctx)
perf_adjust_freq_unthr_context(ctx, throttled);
-
- if (cpuctx->jiffies_interval == 1 ||
- !(jiffies % cpuctx->jiffies_interval))
- perf_rotate_context(cpuctx);
}
}
}
static void ring_buffer_put(struct ring_buffer *rb);
+static void ring_buffer_detach(struct perf_event *event, struct ring_buffer *rb);
static void free_event(struct perf_event *event)
{
if (has_branch_stack(event)) {
static_key_slow_dec_deferred(&perf_sched_events);
/* is system-wide event */
- if (!(event->attach_state & PERF_ATTACH_TASK))
+ if (!(event->attach_state & PERF_ATTACH_TASK)) {
atomic_dec(&per_cpu(perf_branch_stack_events,
event->cpu));
+ }
}
}
if (event->rb) {
- ring_buffer_put(event->rb);
- event->rb = NULL;
+ struct ring_buffer *rb;
+
+ /*
+ * Can happen when we close an event with re-directed output.
+ *
+ * Since we have a 0 refcount, perf_mmap_close() will skip
+ * over us; possibly making our ring_buffer_put() the last.
+ */
+ mutex_lock(&event->mmap_mutex);
+ rb = event->rb;
+ if (rb) {
+ rcu_assign_pointer(event->rb, NULL);
+ ring_buffer_detach(event, rb);
+ ring_buffer_put(rb); /* could be last */
+ }
+ mutex_unlock(&event->mmap_mutex);
}
if (is_cgroup_event(event))
unsigned int events = POLL_HUP;
/*
- * Race between perf_event_set_output() and perf_poll(): perf_poll()
- * grabs the rb reference but perf_event_set_output() overrides it.
- * Here is the timeline for two threads T1, T2:
- * t0: T1, rb = rcu_dereference(event->rb)
- * t1: T2, old_rb = event->rb
- * t2: T2, event->rb = new rb
- * t3: T2, ring_buffer_detach(old_rb)
- * t4: T1, ring_buffer_attach(rb1)
- * t5: T1, poll_wait(event->waitq)
- *
- * To avoid this problem, we grab mmap_mutex in perf_poll()
- * thereby ensuring that the assignment of the new ring buffer
- * and the detachment of the old buffer appear atomic to perf_poll()
+ * Pin the event->rb by taking event->mmap_mutex; otherwise
+ * perf_event_set_output() can swizzle our rb and make us miss wakeups.
*/
mutex_lock(&event->mmap_mutex);
-
- rcu_read_lock();
- rb = rcu_dereference(event->rb);
- if (rb) {
- ring_buffer_attach(event, rb);
+ rb = event->rb;
+ if (rb)
events = atomic_xchg(&rb->poll, 0);
- }
- rcu_read_unlock();
-
mutex_unlock(&event->mmap_mutex);
poll_wait(file, &event->waitq, wait);
return;
spin_lock_irqsave(&rb->event_lock, flags);
- if (!list_empty(&event->rb_entry))
- goto unlock;
-
- list_add(&event->rb_entry, &rb->event_list);
-unlock:
+ if (list_empty(&event->rb_entry))
+ list_add(&event->rb_entry, &rb->event_list);
spin_unlock_irqrestore(&rb->event_lock, flags);
}
-static void ring_buffer_detach(struct perf_event *event,
- struct ring_buffer *rb)
+static void ring_buffer_detach(struct perf_event *event, struct ring_buffer *rb)
{
unsigned long flags;
rcu_read_lock();
rb = rcu_dereference(event->rb);
- if (!rb)
- goto unlock;
-
- list_for_each_entry_rcu(event, &rb->event_list, rb_entry)
- wake_up_all(&event->waitq);
-
-unlock:
+ if (rb) {
+ list_for_each_entry_rcu(event, &rb->event_list, rb_entry)
+ wake_up_all(&event->waitq);
+ }
rcu_read_unlock();
}
static void ring_buffer_put(struct ring_buffer *rb)
{
- struct perf_event *event, *n;
- unsigned long flags;
-
if (!atomic_dec_and_test(&rb->refcount))
return;
- spin_lock_irqsave(&rb->event_lock, flags);
- list_for_each_entry_safe(event, n, &rb->event_list, rb_entry) {
- list_del_init(&event->rb_entry);
- wake_up_all(&event->waitq);
- }
- spin_unlock_irqrestore(&rb->event_lock, flags);
+ WARN_ON_ONCE(!list_empty(&rb->event_list));
call_rcu(&rb->rcu_head, rb_free_rcu);
}
struct perf_event *event = vma->vm_file->private_data;
atomic_inc(&event->mmap_count);
+ atomic_inc(&event->rb->mmap_count);
}
+/*
+ * A buffer can be mmap()ed multiple times; either directly through the same
+ * event, or through other events by use of perf_event_set_output().
+ *
+ * In order to undo the VM accounting done by perf_mmap() we need to destroy
+ * the buffer here, where we still have a VM context. This means we need
+ * to detach all events redirecting to us.
+ */
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->rb);
- struct user_struct *user = event->mmap_user;
- struct ring_buffer *rb = event->rb;
+ struct ring_buffer *rb = event->rb;
+ struct user_struct *mmap_user = rb->mmap_user;
+ int mmap_locked = rb->mmap_locked;
+ unsigned long size = perf_data_size(rb);
+
+ atomic_dec(&rb->mmap_count);
+
+ if (!atomic_dec_and_mutex_lock(&event->mmap_count, &event->mmap_mutex))
+ return;
+
+ /* Detach current event from the buffer. */
+ rcu_assign_pointer(event->rb, NULL);
+ ring_buffer_detach(event, rb);
+ mutex_unlock(&event->mmap_mutex);
+
+ /* If there's still other mmap()s of this buffer, we're done. */
+ if (atomic_read(&rb->mmap_count)) {
+ ring_buffer_put(rb); /* can't be last */
+ return;
+ }
+
+ /*
+ * No other mmap()s, detach from all other events that might redirect
+ * into the now unreachable buffer. Somewhat complicated by the
+ * fact that rb::event_lock otherwise nests inside mmap_mutex.
+ */
+again:
+ rcu_read_lock();
+ list_for_each_entry_rcu(event, &rb->event_list, rb_entry) {
+ if (!atomic_long_inc_not_zero(&event->refcount)) {
+ /*
+ * This event is en-route to free_event() which will
+ * detach it and remove it from the list.
+ */
+ continue;
+ }
+ rcu_read_unlock();
- atomic_long_sub((size >> PAGE_SHIFT) + 1, &user->locked_vm);
- vma->vm_mm->pinned_vm -= event->mmap_locked;
- rcu_assign_pointer(event->rb, NULL);
- ring_buffer_detach(event, rb);
+ mutex_lock(&event->mmap_mutex);
+ /*
+ * Check we didn't race with perf_event_set_output() which can
+ * swizzle the rb from under us while we were waiting to
+ * acquire mmap_mutex.
+ *
+ * If we find a different rb; ignore this event, a next
+ * iteration will no longer find it on the list. We have to
+ * still restart the iteration to make sure we're not now
+ * iterating the wrong list.
+ */
+ if (event->rb == rb) {
+ rcu_assign_pointer(event->rb, NULL);
+ ring_buffer_detach(event, rb);
+ ring_buffer_put(rb); /* can't be last, we still have one */
+ }
mutex_unlock(&event->mmap_mutex);
+ put_event(event);
- ring_buffer_put(rb);
- free_uid(user);
+ /*
+ * Restart the iteration; either we're on the wrong list or
+ * destroyed its integrity by doing a deletion.
+ */
+ goto again;
}
+ rcu_read_unlock();
+
+ /*
+ * It could be there's still a few 0-ref events on the list; they'll
+ * get cleaned up by free_event() -- they'll also still have their
+ * ref on the rb and will free it whenever they are done with it.
+ *
+ * Aside from that, this buffer is 'fully' detached and unmapped,
+ * undo the VM accounting.
+ */
+
+ atomic_long_sub((size >> PAGE_SHIFT) + 1, &mmap_user->locked_vm);
+ vma->vm_mm->pinned_vm -= mmap_locked;
+ free_uid(mmap_user);
+
+ ring_buffer_put(rb); /* could be last */
}
static const struct vm_operations_struct perf_mmap_vmops = {
return -EINVAL;
WARN_ON_ONCE(event->ctx->parent_ctx);
+again:
mutex_lock(&event->mmap_mutex);
if (event->rb) {
- if (event->rb->nr_pages == nr_pages)
- atomic_inc(&event->rb->refcount);
- else
+ if (event->rb->nr_pages != nr_pages) {
ret = -EINVAL;
+ goto unlock;
+ }
+
+ if (!atomic_inc_not_zero(&event->rb->mmap_count)) {
+ /*
+ * Raced against perf_mmap_close() through
+ * perf_event_set_output(). Try again, hope for better
+ * luck.
+ */
+ mutex_unlock(&event->mmap_mutex);
+ goto again;
+ }
+
goto unlock;
}
ret = -ENOMEM;
goto unlock;
}
- rcu_assign_pointer(event->rb, rb);
+
+ atomic_set(&rb->mmap_count, 1);
+ rb->mmap_locked = extra;
+ rb->mmap_user = get_current_user();
atomic_long_add(user_extra, &user->locked_vm);
- event->mmap_locked = extra;
- event->mmap_user = get_current_user();
- vma->vm_mm->pinned_vm += event->mmap_locked;
+ vma->vm_mm->pinned_vm += extra;
+
+ ring_buffer_attach(event, rb);
+ rcu_assign_pointer(event->rb, rb);
perf_event_update_userpage(event);
atomic_inc(&event->mmap_count);
mutex_unlock(&event->mmap_mutex);
- vma->vm_flags |= VM_DONTEXPAND | VM_DONTDUMP;
+ /*
+ * Since pinned accounting is per vm we cannot allow fork() to copy our
+ * vma.
+ */
+ vma->vm_flags |= VM_DONTCOPY | VM_DONTEXPAND | VM_DONTDUMP;
vma->vm_ops = &perf_mmap_vmops;
return ret;
* sign as trigger.
*/
-static u64 perf_swevent_set_period(struct perf_event *event)
+u64 perf_swevent_set_period(struct perf_event *event)
{
struct hw_perf_event *hwc = &event->hw;
u64 period = hwc->last_period;
return snprintf(page, PAGE_SIZE-1, "%d\n", pmu->type);
}
+static ssize_t
+perf_event_mux_interval_ms_show(struct device *dev,
+ struct device_attribute *attr,
+ char *page)
+{
+ struct pmu *pmu = dev_get_drvdata(dev);
+
+ return snprintf(page, PAGE_SIZE-1, "%d\n", pmu->hrtimer_interval_ms);
+}
+
+static ssize_t
+perf_event_mux_interval_ms_store(struct device *dev,
+ struct device_attribute *attr,
+ const char *buf, size_t count)
+{
+ struct pmu *pmu = dev_get_drvdata(dev);
+ int timer, cpu, ret;
+
+ ret = kstrtoint(buf, 0, &timer);
+ if (ret)
+ return ret;
+
+ if (timer < 1)
+ return -EINVAL;
+
+ /* same value, noting to do */
+ if (timer == pmu->hrtimer_interval_ms)
+ return count;
+
+ pmu->hrtimer_interval_ms = timer;
+
+ /* update all cpuctx for this PMU */
+ for_each_possible_cpu(cpu) {
+ struct perf_cpu_context *cpuctx;
+ cpuctx = per_cpu_ptr(pmu->pmu_cpu_context, cpu);
+ cpuctx->hrtimer_interval = ns_to_ktime(NSEC_PER_MSEC * timer);
+
+ if (hrtimer_active(&cpuctx->hrtimer))
+ hrtimer_forward_now(&cpuctx->hrtimer, cpuctx->hrtimer_interval);
+ }
+
+ return count;
+}
+
+#define __ATTR_RW(attr) __ATTR(attr, 0644, attr##_show, attr##_store)
+
static struct device_attribute pmu_dev_attrs[] = {
- __ATTR_RO(type),
- __ATTR_NULL,
+ __ATTR_RO(type),
+ __ATTR_RW(perf_event_mux_interval_ms),
+ __ATTR_NULL,
};
static int pmu_bus_running;
lockdep_set_class(&cpuctx->ctx.lock, &cpuctx_lock);
cpuctx->ctx.type = cpu_context;
cpuctx->ctx.pmu = pmu;
- cpuctx->jiffies_interval = 1;
+
+ __perf_cpu_hrtimer_init(cpuctx, cpu);
+
INIT_LIST_HEAD(&cpuctx->rotation_list);
cpuctx->unique_pmu = pmu;
}
if (!(mask & ~PERF_SAMPLE_BRANCH_PLM_ALL))
return -EINVAL;
- /* kernel level capture: check permissions */
- if ((mask & PERF_SAMPLE_BRANCH_PERM_PLM)
- && perf_paranoid_kernel() && !capable(CAP_SYS_ADMIN))
- return -EACCES;
-
/* propagate priv level, when not set for branch */
if (!(mask & PERF_SAMPLE_BRANCH_PLM_ALL)) {
*/
attr->branch_sample_type = mask;
}
+ /* privileged levels capture (kernel, hv): check permissions */
+ if ((mask & PERF_SAMPLE_BRANCH_PERM_PLM)
+ && perf_paranoid_kernel() && !capable(CAP_SYS_ADMIN))
+ return -EACCES;
}
if (attr->sample_type & PERF_SAMPLE_REGS_USER) {
if (atomic_read(&event->mmap_count))
goto unlock;
+ old_rb = event->rb;
+
if (output_event) {
/* get the rb we want to redirect to */
rb = ring_buffer_get(output_event);
goto unlock;
}
- old_rb = event->rb;
- rcu_assign_pointer(event->rb, rb);
if (old_rb)
ring_buffer_detach(event, old_rb);
+
+ if (rb)
+ ring_buffer_attach(event, rb);
+
+ rcu_assign_pointer(event->rb, rb);
+
+ if (old_rb) {
+ ring_buffer_put(old_rb);
+ /*
+ * Since we detached before setting the new rb, so that we
+ * could attach the new rb, we could have missed a wakeup.
+ * Provide it now.
+ */
+ wake_up_all(&event->waitq);
+ }
+
ret = 0;
unlock:
mutex_unlock(&event->mmap_mutex);
- if (old_rb)
- ring_buffer_put(old_rb);
out:
return ret;
}
case CPU_DOWN_PREPARE:
perf_event_exit_cpu(cpu);
break;
-
default:
break;
}