Skip to contentSkip to conversation

signals: fix sigqueue_free() vs __exit_signal() race

PATCH In review Oleg Nesterov 2 participants 3 messages May 17, 2008 – May 20, 2008

lore
Share & tools
lore
#0 lore raw

PATCH 1/3 signals: fix sigqueue_free() vs __exit_signal() race

__exit_signal() does flush_sigqueue(tsk->pending) outside of ->siglock. This can race with another thread doing sigqueue_free(), we can free the same SIGQUEUE_PREALLOC sigqueue twice or corrupt the pending->list.

Note that even sys_exit_group() can trigger this race, not only sys_timer_delete().

Move the callsite of flush_sigqueue(tsk->pending) under ->siglock.

This patch doesn't touch flush_sigqueue(->shared_pending) below, it is called when there are no other threads which can play with signals, and sigqueue_free() can't be used outside of our thread group.

kernel/exit.c renamed from kernel/exit.c~1_SQFREE_EXIT_RACE +6 -1
@@ -125,6 +125,12 @@ static void __exit_signal(struct task_st
 
 	__unhash_process(tsk);
 
+	/*
+	 * Do this under ->siglock, we can race with another thread
+	 * doing sigqueue_free() if we have SIGQUEUE_PREALLOC signals.
+	 */
+	flush_sigqueue(&tsk->pending);
+
 	tsk->signal = NULL;
 	tsk->sighand = NULL;
 	spin_unlock(&sighand->siglock);
@@ -132,7 +138,6 @@ static void __exit_signal(struct task_st
 
 	__cleanup_sighand(sighand);
 	clear_tsk_thread_flag(tsk,TIF_SIGPENDING);
-	flush_sigqueue(&tsk->pending);
 	if (sig) {
 		flush_sigqueue(&sig->shared_pending);
 		taskstats_tgid_free(sig);
kernel/signal.c renamed from kernel/signal.c~1_SQFREE_EXIT_RACE +2 -1
@@ -1242,7 +1242,8 @@ void sigqueue_free(struct sigqueue *q)
 	/*
 	 * If the signal is still pending remove it from the
 	 * pending queue. We must hold ->siglock while testing
-	 * q->list to serialize with collect_signal().
+	 * q->list to serialize with collect_signal() or with
+	 * __exit_signal()->flush_sigqueue().
 	 */
 	spin_lock_irqsave(lock, flags);
 	if (!list_empty(&q->list))
#1 lore raw
Quoted 4 lines · On Sat, 17 May 2008 19:14:18 +0400 Oleg Nesterov <[email protected]> wrote:

__exit_signal() does flush_sigqueue(tsk->pending) outside of ->siglock. This can race with another thread doing sigqueue_free(), we can free the same SIGQUEUE_PREALLOC sigqueue twice or corrupt the pending->list.

Quoted 8 lines

Note that even sys_exit_group() can trigger this race, not only sys_timer_delete().

Move the callsite of flush_sigqueue(tsk->pending) under ->siglock.

This patch doesn't touch flush_sigqueue(->shared_pending) below, it is called when there are no other threads which can play with signals, and sigqueue_free() can't be used outside of our thread group.

afacit this is the only needed-in-2.6.26 patch from these two three-patch series, yes?

#2 lore raw
Quoted 10 lines · Andrew Morton
Quoted 7 lines · On Sat, 17 May 2008 19:14:18 +0400 Oleg Nesterov <[email protected]> wrote:

__exit_signal() does flush_sigqueue(tsk->pending) outside of ->siglock. This can race with another thread doing sigqueue_free(), we can free the same SIGQUEUE_PREALLOC sigqueue twice or corrupt the pending->list.

Note that even sys_exit_group() can trigger this race, not only sys_timer_delete().

Quoted 8 lines
Quoted 5 lines

Move the callsite of flush_sigqueue(tsk->pending) under ->siglock.

This patch doesn't touch flush_sigqueue(->shared_pending) below, it is called when there are no other threads which can play with signals, and sigqueue_free() can't be used outside of our thread group.

afacit this is the only needed-in-2.6.26 patch from these two three-patch series, yes?

Yes, other patches are not bugfixes.

Probably this fix is not "urgent" too, the race is very old and nobody complained.

Oleg.