[PATCH Fix perf cgroup problem 0/2]

[PATCH Fix perf cgroup problem 0/2]

[Luo Gengkun]

Fixed the issue that nr_cgroup and cpuctx->cgrp of the perf cgroup event
are not correctly updated. And Fixed the possible concurrency of
perf_cgroup_switch and perf_cgroup_event_disable.

Luo Gengkun (2):
  perf/core: Fix nr_cgroups/cpuctx->cgrp is not updated correctly
  perf/core: Fix WARN_ON_ONCE(cpuctx->ctx.nr_cgroups == 0) in
    perf_cgroup_switch

 kernel/events/core.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)

-- 
2.34.1

[PATCH Fix perf cgroup problem 1/2] perf/core: Fix nr_cgroups/cpuctx->cgrp is not updated correctly

[Luo Gengkun]

Since __perf_remove_from_context updates event->state before list_del_event,
this prevents list_del_event from calling perf_cgroup_event_disable,
resulting in will not update nr_cgroups and cpuctx->cgrp.

To resolve this issue:
 1. First update the event time to ensure perf_child_detach uses
    accurate event time.
 2. Then update event->state after list_del_event completes.

Fixes: a3c3c66670ce ("perf/core: Fix child_total_time_enabled accounting bug at task exit")
Signed-off-by: Luo Gengkun <luogengkun@huaweicloud.com>
---
 kernel/events/core.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/kernel/events/core.c b/kernel/events/core.c
index 95e703891b24..ecb4d852a006 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -2482,12 +2482,17 @@ __perf_remove_from_context(struct perf_event *event,
 		state = PERF_EVENT_STATE_DEAD;
 	}
 	event_sched_out(event, ctx);
-	perf_event_set_state(event, min(event->state, state));
+	perf_event_update_time(event);
 	if (flags & DETACH_GROUP)
 		perf_group_detach(event);
 	if (flags & DETACH_CHILD)
 		perf_child_detach(event);
 	list_del_event(event, ctx);
+	/*
+	 * Update the event->state after list_del_event.
+	 */
+	if (flags & (DETACH_DEAD | DETACH_EXIT))
+		event->state = state;
 
 	if (!pmu_ctx->nr_events) {
 		pmu_ctx->rotate_necessary = 0;
-- 
2.34.1

[PATCH Fix perf cgroup problem 2/2] perf/core: Fix WARN_ON_ONCE(cpuctx->ctx.nr_cgroups == 0) in perf_cgroup_switch

[Luo Gengkun]

There may be concurrency between perf_cgroup_switch and
perf_cgroup_event_disable. Consider the following scenario: after a new
perf cgroup event is created on CPU0, the new event might not trigger
reprogramm, leaving ctx->is_active as 0. In this case, when CPU1 disables
the perf event, it executes __perf_remove_from_context->list_del_event->
perf_cgroup_event_disable on CPU1, resulting in a race with
perf_cgroup_switch running on CPU0.

To fix this problem, expand the lock-holding critical section in
perf_cgroup_switch.

Fixes: db4a835601b7 ("perf/core: Set cgroup in CPU contexts for new cgroup events")
Signed-off-by: Luo Gengkun <luogengkun@huaweicloud.com>
---
 kernel/events/core.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/kernel/events/core.c b/kernel/events/core.c
index ecb4d852a006..fae1f68cbca5 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -931,20 +931,20 @@ static void perf_cgroup_switch(struct task_struct *task)
 	struct perf_cpu_context *cpuctx = this_cpu_ptr(&perf_cpu_context);
 	struct perf_cgroup *cgrp;
 
+	cgrp = perf_cgroup_from_task(task, NULL);
+	perf_ctx_lock(cpuctx, cpuctx->task_ctx);
 	/*
 	 * cpuctx->cgrp is set when the first cgroup event enabled,
 	 * and is cleared when the last cgroup event disabled.
 	 */
 	if (READ_ONCE(cpuctx->cgrp) == NULL)
-		return;
+		goto unlock;
 
 	WARN_ON_ONCE(cpuctx->ctx.nr_cgroups == 0);
 
-	cgrp = perf_cgroup_from_task(task, NULL);
 	if (READ_ONCE(cpuctx->cgrp) == cgrp)
-		return;
+		goto unlock;
 
-	perf_ctx_lock(cpuctx, cpuctx->task_ctx);
 	perf_ctx_disable(&cpuctx->ctx, true);
 
 	ctx_sched_out(&cpuctx->ctx, NULL, EVENT_ALL|EVENT_CGROUP);
@@ -962,6 +962,7 @@ static void perf_cgroup_switch(struct task_struct *task)
 	ctx_sched_in(&cpuctx->ctx, NULL, EVENT_ALL|EVENT_CGROUP);
 
 	perf_ctx_enable(&cpuctx->ctx, true);
+unlock:
 	perf_ctx_unlock(cpuctx, cpuctx->task_ctx);
 }
 
-- 
2.34.1