Re: [BUG] TASK_DEAD task is able to be woken up in special condition

[Oleg Nesterov <oleg@redhat.com>]

On 01/18, Ingo Molnar wrote:
>
> Well, but does the -rt kernel suffer from the same race?

I think yes.

To remind, the problem is generic, it is not bound to this
particular place. Let me repeat:

Suppose that the task T does

	set_current_state(TASK_INTERRUPTIBLE);

	set_current_state(TASK_UNINTERRUPTIBLE);
	schedule();

try_to_wake_up(TASK_UNINTERRUPTIBLE) in between can observe this task
in TASK_INTERRUPTIBLE state. Then it can set RUNNING/WAKING after T
sets ->state = TASK_UNINTERRUPTIBLE.

For example, this is possibly if T simply does wait_event() twice when
the the 1st wait_event() doesn't sleep.


do_exit() is different because it can not handle the spurious wakeup.
Well, may be we can? we can simply do

		for (;;) {
			tsk->state = TASK_DEAD;
			schedule();
		}

__schedule() can't race with ttwu() once it takes rq->lock. If the
exiting task is deactivated, finish_task_switch() will see EXIT_DEAD.
Unless I missed something, the only problem is preempt_disable(),
but schedule_debug() checks ->exit_state.

OTOH, if we fix this race then probably schedule_debug() should
check state == EXIT_DEAD instead.

> So if there's something in the -rt kernel that fixes this race
> we'd like to have that. If the bug is present in the -rt kernel
> then why didn't it ever get triggered? We caught much more
> narrow races in -rt, and very early on in the project.

I do not know.

Oleg.