signals: avoid random wakeups in sigsuspend()
authorSasha Levin <sasha.levin@oracle.com>
Fri, 5 Feb 2016 23:36:05 +0000 (15:36 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Sat, 6 Feb 2016 02:10:40 +0000 (18:10 -0800)
A random wakeup can get us out of sigsuspend() without TIF_SIGPENDING
being set.

Avoid that by making sure we were signaled, like sys_pause() does.

Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kernel/signal.c

index f3f1f7a972fd40f3d437d6bf3b5db2cacaffe6ca..0508544c8ced0d96913905dc53af68f38b6ee618 100644 (file)
@@ -3508,8 +3508,10 @@ static int sigsuspend(sigset_t *set)
        current->saved_sigmask = current->blocked;
        set_current_blocked(set);
 
-       __set_current_state(TASK_INTERRUPTIBLE);
-       schedule();
+       while (!signal_pending(current)) {
+               __set_current_state(TASK_INTERRUPTIBLE);
+               schedule();
+       }
        set_restore_sigmask();
        return -ERESTARTNOHAND;
 }