workqueue: Put the pwq after detaching the rescuer from the pool

BugLink: https://bugs.launchpad.net/bugs/2104873

[ Upstream commit e769461 ]

The commit 68f8305("workqueue: Reap workers via kthread_stop() and
remove detach_completion") adds code to reap the normal workers but
mistakenly does not handle the rescuer and also removes the code waiting
for the rescuer in put_unbound_pool(), which caused a use-after-free bug
reported by Cheung Wall.

To avoid the use-after-free bug, the pool’s reference must be held until
the detachment is complete. Therefore, move the code that puts the pwq
after detaching the rescuer from the pool.

Reported-by: cheung wall <zzqq0103.hey@gmail.com>
Cc: cheung wall <zzqq0103.hey@gmail.com>
Link: https://lore.kernel.org/lkml/CAKHoSAvP3iQW+GwmKzWjEAOoPvzeWeoMO0Gz7Pp3_4kxt-RMoA@mail.gmail.com/
Fixes: 68f8305("workqueue: Reap workers via kthread_stop() and remove detach_completion")
Signed-off-by: Lai Jiangshan <jiangshan.ljs@antgroup.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Noah Wager <noah.wager@canonical.com>

Author: Lai Jiangshan

kernel/workqueue.c

@@ -3511,12 +3511,6 @@ static int rescuer_thread(void *__rescuer)
 			}
 		}
 
-		/*
-		 * Put the reference grabbed by send_mayday().  @pool won't
-		 * go away while we're still attached to it.
-		 */
-		put_pwq(pwq);
-
 		/*
 		 * Leave this pool. Notify regular workers; otherwise, we end up
 		 * with 0 concurrency and stalling the execution.
@@ -3527,6 +3521,12 @@ static int rescuer_thread(void *__rescuer)
 
 		worker_detach_from_pool(rescuer);
 
+		/*
+		 * Put the reference grabbed by send_mayday().  @pool might
+		 * go away any time after it.
+		 */
+		put_pwq_unlocked(pwq);
+
 		raw_spin_lock_irq(&wq_mayday_lock);
 	}