[PATCH v2] cgroup: Don't expose dead tasks in cgroup

[Sebastian Andrzej Siewior <bigeasy@linutronix.de>]

Once a task exits it has its state set to TASK_DEAD and then it is
removed the cgroup it belonged to. The last step happens on the task
gets out of its last schedule() invocation and is delayed on PREEMPT_RT
due to locking constrains.

As a result it is possible to receive a pid via waitpid() of a task
which is still listed in cgroup.procs for the cgroup it belonged
to. This is something that systemd does not expect and as a result it
waits for its exit until a time out occurs.
This can also be reproduced on !PREEMPT_RT kernel with a significant
delay in do_exit() after exit_notify().

Hide the task from the output which have PF_EXITING set which is done
before the parent is notified. Keeping zombies with live threads
shouldn't break anything (suggested by Tejun).

Reported-by: Bert Karwatzki <spasswolf@web.de>
Closes: https://lore.kernel.org/all/20260219164648.3014-1-spasswolf@web.de/
Tested-by: Bert Karwatzki <spasswolf@web.de>
Fixes: 9311e6c29b348 ("cgroup: Fix sleeping from invalid context warning on PREEMPT_RT")
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 v1…v2: https://lore.kernel.org/all/20260302120738.6KkDipsR@linutronix.de/
   - Close the race window filtering out PF_EXITING tasks instead.
   - Document the possible race window after it has been verified.

 kernel/cgroup/cgroup.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
index c22cda7766d84..eef01b80ec933 100644
--- a/kernel/cgroup/cgroup.c
+++ b/kernel/cgroup/cgroup.c
@@ -5108,6 +5108,12 @@ static void css_task_iter_advance(struct css_task_iter *it)
 		return;
 
 	task = list_entry(it->task_pos, struct task_struct, cg_list);
+	/*
+	 * Hide task which are exitting but not yet removed. Keep zombie
+	 * leaders with live threads visible.
+	 */
+	if ((task->flags & PF_EXITING) && !atomic_read(&task->signal->live))
+		goto repeat;
 
 	if (it->flags & CSS_TASK_ITER_PROCS) {
 		/* if PROCS, skip over tasks which aren't group leaders */
-- 
2.53.0