Re: [PATCH 2/2] workqueue: defer wake_up_process() outside pool->lock on hot paths

[Hillf Danton]

On Tue, 26 May 2026 14:08:06 -0400 Breno Leitao wrote:
> @@ -2423,6 +2424,15 @@ static void __queue_work(int cpu, struct workqueue_struct *wq,
>  
>  out:
>  	raw_spin_unlock(&pool->lock);
> +	/*
> +	 * Issue the wakeup after dropping pool->lock to shorten the
> +	 * locked region on this hot enqueue path.  kick_pool_pick() did all
> +	 * of the work that required the lock (worker selection and
> +	 * wake_cpu setup); the wake_up_process() itself only needs to
> +	 * take the target rq->lock.
> +	 */
> +	if (wake_p)
> +		wake_up_process(wake_p);
>  	rcu_read_unlock();
>  }
>
I suspect this works without the pool lock held, go and see workers culled
in idle_cull_fn() for example.