perf/arm-dmc620: Fix lockdep assert in ->event_init()

for_each_sibling_event() checks leader's ctx but it doesn't have the ctx yet if it's the leader. Like in perf_event_validate_size(), we should skip checking siblings in that case. Acked-by: Mark Rutland <[email protected]> Fixes: f3c0eba ("perf: Add a few assertions") Reported-by: Greg Thelen <[email protected]> Cc: Robin Murphy <[email protected]> Cc: Tuan Phan <[email protected]> Signed-off-by: Namhyung Kim <[email protected]> Reviewed-by: Robin Murphy <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Will Deacon <[email protected]>
hardkernel · May 17, 2024 · a4c5a45 · a4c5a45
1 parent b8995a1
commit a4c5a45
Showing 1 changed file with 6 additions and 3 deletions.
diff --git a/drivers/perf/arm_dmc620_pmu.c b/drivers/perf/arm_dmc620_pmu.c
@@ -542,12 +542,16 @@ static int dmc620_pmu_event_init(struct perf_event *event)
 	if (event->cpu < 0)
 		return -EINVAL;
 
+	hwc->idx = -1;
+
+	if (event->group_leader == event)
+		return 0;
+
 	/*
 	 * We can't atomically disable all HW counters so only one event allowed,
 	 * although software events are acceptable.
 	 */
-	if (event->group_leader != event &&
-			!is_software_event(event->group_leader))
+	if (!is_software_event(event->group_leader))
 		return -EINVAL;
 
 	for_each_sibling_event(sibling, event->group_leader) {
@@ -556,7 +560,6 @@ static int dmc620_pmu_event_init(struct perf_event *event)
 			return -EINVAL;
 	}
 
-	hwc->idx = -1;
 	return 0;
 }