> I think the real fix is to make sure that preemption never hits while
> current->state == TASK_ZOMBIE.
>
> In other words, I think the _correct_ fix is to just make
> preempt_schedule() check for something like
>
> if (current->state != TASK_RUNNING)
> return;
>
> and just getting rid of the current "unconditionally set state to
> running".
>
> (Side note: if the state isn't running, we're most likely going to
> reschedule in a controlled manner soon anyway. Sure, there are some loops
> which set state to non-runnable early for race avoidance, but is it a
> _problem_?)
Eh, maybe - what about all the code that sets non-running before putting
itself on a wait queue?
Robert Love
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/