On Tue, Jun 03, 2014 at 03:32:07PM +0800, Lai Jiangshan wrote:
> The commit ea1abd6197d5 ("workqueue: reimplement idle worker rebinding")
> used a trick which simply removes all to-be-bound idle workers from the
> idle list and lets them add themselves back after completing rebinding.
>
> And this trick caused the @worker_pool->nr_idle may deviate than the actual
> number of idle workers on @worker_pool->idle_list. More specifically,
> nr_idle may be non-zero while ->idle_list is empty. All users of
> ->nr_idle and ->idle_list are audited. The only affected one is
> too_many_workers() which is updated to check %false if ->idle_list is
> empty regardless of ->nr_idle.
>
> The commit/trick was complicated due to it just tried to simplify an even
> more complicated problem (workers had to rebind itself). But the commit
> a9ab775bcadf ("workqueue: directly restore CPU affinity of workers
> from CPU_ONLINE") fixed all these problems and the mentioned trick was
> useless and is gone.
>
> So, now the @worker_pool->nr_idle is exactly the actual number of workers
> on @worker_pool->idle_list. too_many_workers() should recover as it was
> before the trick. So we remove the empty check.
>
> Signed-off-by: Lai Jiangshan <[email protected]>
Applied to wq/for-3.17.
Thanks.
--
tejun
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
