Re: [PATCH v4 2/3] locking/mutex: Enable optimistic spinning of woken task in wait queue

From: Peter Zijlstra
Date: Mon Aug 08 2016 - 13:29:55 EST


On Mon, Jul 18, 2016 at 04:39:25PM -0400, Waiman Long wrote:
> Ding Tianhong reported a live-lock situation where a constant stream
> of incoming optimistic spinners blocked a task in the wait list from
> getting the mutex.
>
> This patch attempts to fix this live-lock condition by enabling the
> woken task in the wait queue to enter into an optimistic spinning
> loop itself in parallel with the regular spinners in the OSQ. This
> should prevent the live-lock condition from happening.

No, two spinners are not in fact starvation proof. It makes the reported
life-lock scenario much less likely, but it does not guarantee anything.

> + /*
> + * Optimistically spinning on the mutex without the wait lock

There should either be a '.' at the end of that line, or the next line
should not start with a capital.

Also, I don't see how the two sentences are related, should they be in
the same paragraph?

> + * The state has to be set to running to avoid another waker
> + * spinning on the on_cpu flag while the woken waiter is
> + * spinning on the mutex.
> + */
> + acquired = mutex_optimistic_spin(lock, ww_ctx, use_ww_ctx,
> + true);
> spin_lock_mutex(&lock->wait_lock, flags);
> }
> __set_task_state(task, TASK_RUNNING);