1 files changed, 116 insertions, 84 deletions

diff --git a/kernel/perf_event.c b/kernel/perf_event.c
index eac7e3364335..aede71245e9f 100644
--- a/kernel/perf_event.c
+++ b/kernel/perf_event.c
@@ -312,9 +312,75 @@ list_add_event(struct perf_event *event, struct perf_event_context *ctx)
                 ctx->nr_stat++;
 }
 
+/*
+ * Called at perf_event creation and when events are attached/detached from a
+ * group.
+ */
+static void perf_event__read_size(struct perf_event *event)
+{
+        int entry = sizeof(u64); /* value */
+        int size = 0;
+        int nr = 1;
+
+        if (event->attr.read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
+                size += sizeof(u64);
+
+        if (event->attr.read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
+                size += sizeof(u64);
+
+        if (event->attr.read_format & PERF_FORMAT_ID)
+                entry += sizeof(u64);
+
+        if (event->attr.read_format & PERF_FORMAT_GROUP) {
+                nr += event->group_leader->nr_siblings;
+                size += sizeof(u64);
+        }
+
+        size += entry * nr;
+        event->read_size = size;
+}
+
+static void perf_event__header_size(struct perf_event *event)
+{
+        struct perf_sample_data *data;
+        u64 sample_type = event->attr.sample_type;
+        u16 size = 0;
+
+        perf_event__read_size(event);
+
+        if (sample_type & PERF_SAMPLE_IP)
+                size += sizeof(data->ip);
+
+        if (sample_type & PERF_SAMPLE_TID)
+                size += sizeof(data->tid_entry);
+
+        if (sample_type & PERF_SAMPLE_TIME)
+                size += sizeof(data->time);
+
+        if (sample_type & PERF_SAMPLE_ADDR)
+                size += sizeof(data->addr);
+
+        if (sample_type & PERF_SAMPLE_ID)
+                size += sizeof(data->id);
+
+        if (sample_type & PERF_SAMPLE_STREAM_ID)
+                size += sizeof(data->stream_id);
+
+        if (sample_type & PERF_SAMPLE_CPU)
+                size += sizeof(data->cpu_entry);
+
+        if (sample_type & PERF_SAMPLE_PERIOD)
+                size += sizeof(data->period);
+
+        if (sample_type & PERF_SAMPLE_READ)
+                size += event->read_size;
+
+        event->header_size = size;
+}
+
 static void perf_group_attach(struct perf_event *event)
 {
-        struct perf_event *group_leader = event->group_leader;
+        struct perf_event *group_leader = event->group_leader, *pos;
 
         /*
          * We can have double attach due to group movement in perf_event_open.
@@ -333,6 +399,11 @@ static void perf_group_attach(struct perf_event *event)
 
         list_add_tail(&event->group_entry, &group_leader->sibling_list);
         group_leader->nr_siblings++;
+
+        perf_event__header_size(group_leader);
+
+        list_for_each_entry(pos, &group_leader->sibling_list, group_entry)
+                perf_event__header_size(pos);
 }
 
 /*
@@ -391,7 +462,7 @@ static void perf_group_detach(struct perf_event *event)
         if (event->group_leader != event) {
                 list_del_init(&event->group_entry);
                 event->group_leader->nr_siblings--;
-                return;
+                goto out;
         }
 
         if (!list_empty(&event->group_entry))
@@ -410,6 +481,12 @@ static void perf_group_detach(struct perf_event *event)
                 /* Inherit group flags from the previous leader */
                 sibling->group_flags = event->group_flags;
         }
+
+out:
+        perf_event__header_size(event->group_leader);
+
+        list_for_each_entry(tmp, &event->group_leader->sibling_list, group_entry)
+                perf_event__header_size(tmp);
 }
 
 static inline int
@@ -1073,7 +1150,7 @@ static int perf_event_refresh(struct perf_event *event, int refresh)
         /*
          * not supported on inherited events
          */
-        if (event->attr.inherit)
+        if (event->attr.inherit || !is_sampling_event(event))
                 return -EINVAL;
 
         atomic_add(refresh, &event->event_limit);
@@ -2289,31 +2366,6 @@ static int perf_release(struct inode *inode, struct file *file)
         return perf_event_release_kernel(event);
 }
 
-static int perf_event_read_size(struct perf_event *event)
-{
-        int entry = sizeof(u64); /* value */
-        int size = 0;
-        int nr = 1;
-
-        if (event->attr.read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
-                size += sizeof(u64);
-
-        if (event->attr.read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
-                size += sizeof(u64);
-
-        if (event->attr.read_format & PERF_FORMAT_ID)
-                entry += sizeof(u64);
-
-        if (event->attr.read_format & PERF_FORMAT_GROUP) {
-                nr += event->group_leader->nr_siblings;
-                size += sizeof(u64);
-        }
-
-        size += entry * nr;
-
-        return size;
-}
-
 u64 perf_event_read_value(struct perf_event *event, u64 *enabled, u64 *running)
 {
         struct perf_event *child;
@@ -2428,7 +2480,7 @@ perf_read_hw(struct perf_event *event, char __user *buf, size_t count)
         if (event->state == PERF_EVENT_STATE_ERROR)
                 return 0;
 
-        if (count < perf_event_read_size(event))
+        if (count < event->read_size)
                 return -ENOSPC;
 
         WARN_ON_ONCE(event->ctx->parent_ctx);
@@ -2514,7 +2566,7 @@ static int perf_event_period(struct perf_event *event, u64 __user *arg)
         int ret = 0;
         u64 value;
 
-        if (!event->attr.sample_period)
+        if (!is_sampling_event(event))
                 return -EINVAL;
 
         if (copy_from_user(&value, arg, sizeof(value)))
@@ -3606,59 +3658,34 @@ void perf_prepare_sample(struct perf_event_header *header,
         data->type = sample_type;
 
         header->type = PERF_RECORD_SAMPLE;
-        header->size = sizeof(*header);
+        header->size = sizeof(*header) + event->header_size;
 
         header->misc = 0;
         header->misc |= perf_misc_flags(regs);
 
-        if (sample_type & PERF_SAMPLE_IP) {
+        if (sample_type & PERF_SAMPLE_IP)
                 data->ip = perf_instruction_pointer(regs);
 
-                header->size += sizeof(data->ip);
-        }
-
         if (sample_type & PERF_SAMPLE_TID) {
                 /* namespace issues */
                 data->tid_entry.pid = perf_event_pid(event, current);
                 data->tid_entry.tid = perf_event_tid(event, current);
-
-                header->size += sizeof(data->tid_entry);
         }
 
-        if (sample_type & PERF_SAMPLE_TIME) {
+        if (sample_type & PERF_SAMPLE_TIME)
                 data->time = perf_clock();
 
-                header->size += sizeof(data->time);
-        }
-
-        if (sample_type & PERF_SAMPLE_ADDR)
-                header->size += sizeof(data->addr);
-
-        if (sample_type & PERF_SAMPLE_ID) {
+        if (sample_type & PERF_SAMPLE_ID)
                 data->id = primary_event_id(event);
 
-                header->size += sizeof(data->id);
-        }
-
-        if (sample_type & PERF_SAMPLE_STREAM_ID) {
+        if (sample_type & PERF_SAMPLE_STREAM_ID)
                 data->stream_id = event->id;
 
-                header->size += sizeof(data->stream_id);
-        }
-
         if (sample_type & PERF_SAMPLE_CPU) {
                 data->cpu_entry.cpu             = raw_smp_processor_id();
                 data->cpu_entry.reserved        = 0;
-
-                header->size += sizeof(data->cpu_entry);
         }
 
-        if (sample_type & PERF_SAMPLE_PERIOD)
-                header->size += sizeof(data->period);
-
-        if (sample_type & PERF_SAMPLE_READ)
-                header->size += perf_event_read_size(event);
-
         if (sample_type & PERF_SAMPLE_CALLCHAIN) {
                 int size = 1;
 
@@ -3726,7 +3753,7 @@ perf_event_read_event(struct perf_event *event,
                 .header = {
                         .type = PERF_RECORD_READ,
                         .misc = 0,
-                        .size = sizeof(read_event) + perf_event_read_size(event),
+                        .size = sizeof(read_event) + event->read_size,
                 },
                 .pid = perf_event_pid(event, task),
                 .tid = perf_event_tid(event, task),
@@ -4240,6 +4267,13 @@ static int __perf_event_overflow(struct perf_event *event, int nmi,
         struct hw_perf_event *hwc = &event->hw;
         int ret = 0;
 
+        /*
+         * Non-sampling counters might still use the PMI to fold short
+         * hardware counters, ignore those.
+         */
+        if (unlikely(!is_sampling_event(event)))
+                return 0;
+
         if (!throttle) {
                 hwc->interrupts++;
         } else {
@@ -4385,7 +4419,7 @@ static void perf_swevent_event(struct perf_event *event, u64 nr,
         if (!regs)
                 return;
 
-        if (!hwc->sample_period)
+        if (!is_sampling_event(event))
                 return;
 
         if (nr == 1 && hwc->sample_period == 1 && !event->attr.freq)
@@ -4548,7 +4582,7 @@ static int perf_swevent_add(struct perf_event *event, int flags)
         struct hw_perf_event *hwc = &event->hw;
         struct hlist_head *head;
 
-        if (hwc->sample_period) {
+        if (is_sampling_event(event)) {
                 hwc->last_period = hwc->sample_period;
                 perf_swevent_set_period(event);
         }
@@ -4805,15 +4839,6 @@ static int perf_tp_event_init(struct perf_event *event)
         if (event->attr.type != PERF_TYPE_TRACEPOINT)
                 return -ENOENT;
 
-        /*
-         * Raw tracepoint data is a severe data leak, only allow root to
-         * have these.
-         */
-        if ((event->attr.sample_type & PERF_SAMPLE_RAW) &&
-                        perf_paranoid_tracepoint_raw() &&
-                        !capable(CAP_SYS_ADMIN))
-                return -EPERM;
-
         err = perf_trace_init(event);
         if (err)
                 return err;
@@ -4926,31 +4951,33 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer)
 static void perf_swevent_start_hrtimer(struct perf_event *event)
 {
         struct hw_perf_event *hwc = &event->hw;
+        s64 period;
+
+        if (!is_sampling_event(event))
+                return;
 
         hrtimer_init(&hwc->hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
         hwc->hrtimer.function = perf_swevent_hrtimer;
-        if (hwc->sample_period) {
-                s64 period = local64_read(&hwc->period_left);
 
-                if (period) {
-                        if (period < 0)
-                                period = 10000;
+        period = local64_read(&hwc->period_left);
+        if (period) {
+                if (period < 0)
+                        period = 10000;
 
-                        local64_set(&hwc->period_left, 0);
-                } else {
-                        period = max_t(u64, 10000, hwc->sample_period);
-                }
-                __hrtimer_start_range_ns(&hwc->hrtimer,
+                local64_set(&hwc->period_left, 0);
+        } else {
+                period = max_t(u64, 10000, hwc->sample_period);
+        }
+        __hrtimer_start_range_ns(&hwc->hrtimer,
                                 ns_to_ktime(period), 0,
                                 HRTIMER_MODE_REL_PINNED, 0);
-        }
 }
 
 static void perf_swevent_cancel_hrtimer(struct perf_event *event)
 {
         struct hw_perf_event *hwc = &event->hw;
 
-        if (hwc->sample_period) {
+        if (is_sampling_event(event)) {
                 ktime_t remaining = hrtimer_get_remaining(&hwc->hrtimer);
                 local64_set(&hwc->period_left, ktime_to_ns(remaining));
 
@@ -5715,6 +5742,11 @@ SYSCALL_DEFINE5(perf_event_open,
         mutex_unlock(&current->perf_event_mutex);
 
         /*
+         * Precalculate sample_data sizes
+         */
+        perf_event__header_size(event);
+
+        /*
          * Drop the reference on the group_event after placing the
          * new event on the sibling_list. This ensures destruction
          * of the group leader will find the pointer to itself in