sched: Check TASK_DEAD rather than EXIT_DEAD in schedule_debug()

schedule_debug() ignores in_atomic() if prev->exit_state != 0.
This is not what we want, ->exit_state is set by exit_notify()
but we should complain until the task does the last schedule()
in TASK_DEAD.

See also 7407251 "PF_DEAD cleanup", I think this ancient
commit explains why schedule() had to rely on ->exit_state,
until that commit exit_notify() disabled preemption and set
PF_DEAD which was used to detect the exiting task.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: David Laight <David.Laight@ACULAB.COM>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Tejun Heo <tj@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/20131113154538.GB15810@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>

kernel/sched/core.c

@@ -2414,10 +2414,10 @@ static inline void schedule_debug(struct task_struct *prev)
 {
 	/*
 	 * Test if we are atomic. Since do_exit() needs to call into
-	 * schedule() atomically, we ignore that path for now.
-	 * Otherwise, whine if we are scheduling when we should not be.
+	 * schedule() atomically, we ignore that path. Otherwise whine
+	 * if we are scheduling when we should not.
 	 */
-	if (unlikely(in_atomic_preempt_off() && !prev->exit_state))
+	if (unlikely(in_atomic_preempt_off() && prev->state != TASK_DEAD))
 		__schedule_bug(prev);
 	rcu_sleep_check();