[PATCH v2] cgroup: Use data_race() for task->flags in task_css_set_check()
From: Guopeng Zhang <hidden>
Date: 2026-06-25 01:40:17
Also in:
lkml
Subsystem:
control group (cgroup), the rest · Maintainers:
Tejun Heo, Johannes Weiner, Michal Koutný, Linus Torvalds
From: Guopeng Zhang <redacted> task_css_set_check() uses rcu_dereference_check() to verify that task->cgroups can be dereferenced. One accepted condition is that the task is already exiting, tested by checking PF_EXITING in task->flags. This check is only part of the CONFIG_PROVE_RCU lockdep predicate. This was found by KCSAN during fuzz testing. KCSAN can report a data race when another task flag bit is updated concurrently. One report shows pids_release() reading task->flags through task_css_set_check() while do_task_dead() sets PF_NOFREEZE: KCSAN: data-race in task_css() [inline] KCSAN: data-race in pids_release() task_css() pids_release() cgroup_release() release_task() wait_task_zombie() value changed: 0x0040004c -> 0x0040804c The changed bit is PF_NOFREEZE, not PF_EXITING. PF_EXITING remains set before and after the update, so the task_css_set_check() condition does not change. This is not a race on task->cgroups and does not indicate incorrect pids charging or uncharging. tools/memory-model/Documentation/access-marking.txt recommends data_race() for data-racy loads used only for diagnostic purposes. Use data_race() here to mark the intended diagnostic-only access. No functional change intended. Suggested-by: Tejun Heo <tj@kernel.org> Signed-off-by: Guopeng Zhang <redacted> --- Changes in v2: - Use data_race() instead of READ_ONCE() for the diagnostic-only CONFIG_PROVE_RCU predicate, as suggested by Tejun. - Update the changelog to match access-marking.txt guidance. include/linux/cgroup.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
index f2aa46a4f871..b905208942bf 100644
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h@@ -480,7 +480,7 @@ static inline void cgroup_unlock(void) rcu_read_lock_sched_held() || \ lockdep_is_held(&cgroup_mutex) || \ lockdep_is_held(&css_set_lock) || \ - ((task)->flags & PF_EXITING) || (__c)) + (data_race((task)->flags) & PF_EXITING) || (__c)) #else #define task_css_set_check(task, __c) \ rcu_dereference((task)->cgroups)
--
2.25.1