diff options
| author | Peter Zijlstra <peterz@infradead.org> | 2025-12-20 14:14:41 +0100 |
|---|---|---|
| committer | Peter Zijlstra <peterz@infradead.org> | 2026-01-05 08:55:54 +0100 |
| commit | ff5860f5088e9076ebcccf05a6ca709d5935cfa9 (patch) | |
| tree | 9b97db528b00f27e705674b81abfd3e2b1a995ee /kernel | |
| parent | 9ace4753a5202b02191d54e9fdf7f9e3d02b85eb (diff) | |
perf: Ensure swevent hrtimer is properly destroyed
With the change to hrtimer_try_to_cancel() in
perf_swevent_cancel_hrtimer() it appears possible for the hrtimer to
still be active by the time the event gets freed.
Make sure the event does a full hrtimer_cancel() on the free path by
installing a perf_event::destroy handler.
Fixes: eb3182ef0405 ("perf/core: Fix system hang caused by cpu-clock usage")
Reported-by: CyberUnicorns <a101e_iotvul@163.com>
Tested-by: CyberUnicorns <a101e_iotvul@163.com>
Debugged-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Diffstat (limited to 'kernel')
| -rw-r--r-- | kernel/events/core.c | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/kernel/events/core.c b/kernel/events/core.c index dad0d3d2e85f..e3d8338fd51c 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -11906,6 +11906,11 @@ static void perf_swevent_cancel_hrtimer(struct perf_event *event) } } +static void perf_swevent_destroy_hrtimer(struct perf_event *event) +{ + hrtimer_cancel(&event->hw.hrtimer); +} + static void perf_swevent_init_hrtimer(struct perf_event *event) { struct hw_perf_event *hwc = &event->hw; @@ -11914,6 +11919,7 @@ static void perf_swevent_init_hrtimer(struct perf_event *event) return; hrtimer_setup(&hwc->hrtimer, perf_swevent_hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL_HARD); + event->destroy = perf_swevent_destroy_hrtimer; /* * Since hrtimers have a fixed rate, we can do a static freq->period |