[50/55] perf: Fix tear-down of inherited group events

From: Greg KH
Date: Fri Mar 25 2011 - 19:57:19 EST


2.6.37-stable review patch. If anyone has any objections, please let us know.

------------------

From: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>

commit 38b435b16c36b0d863efcf3f07b34a6fac9873fd upstream.

When destroying inherited events, we need to destroy groups too,
otherwise the event iteration in perf_event_exit_task_context() will
miss group siblings and we leak events with all the consequences.

Reported-and-tested-by: Vince Weaver <vweaver1@xxxxxxxxxxxx>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
LKML-Reference: <1300196470.2203.61.camel@twins>
Signed-off-by: Ingo Molnar <mingo@xxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxx>

---
kernel/perf_event.c | 13 ++++++++-----
1 file changed, 8 insertions(+), 5 deletions(-)

--- a/kernel/perf_event.c
+++ b/kernel/perf_event.c
@@ -5863,17 +5863,20 @@ __perf_event_exit_task(struct perf_event
struct perf_event_context *child_ctx,
struct task_struct *child)
{
- struct perf_event *parent_event;
+ if (child_event->parent) {
+ raw_spin_lock_irq(&child_ctx->lock);
+ perf_group_detach(child_event);
+ raw_spin_unlock_irq(&child_ctx->lock);
+ }

perf_event_remove_from_context(child_event);

- parent_event = child_event->parent;
/*
- * It can happen that parent exits first, and has events
+ * It can happen that the parent exits first, and has events
* that are still around due to the child reference. These
- * events need to be zapped - but otherwise linger.
+ * events need to be zapped.
*/
- if (parent_event) {
+ if (child_event->parent) {
sync_child_event(child_event, child);
free_event(child_event);
}


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/