Re: [patch] use cheaper elv_queue_empty when unplug a device

[Jens Axboe <axboe@suse.de>]

On Tue, Mar 29 2005, Nick Piggin wrote:
> @@ -2577,19 +2577,18 @@ static int __make_request(request_queue_
>  	spin_lock_prefetch(q->queue_lock);
>  
>  	barrier = bio_barrier(bio);
> -	if (barrier && (q->ordered == QUEUE_ORDERED_NONE)) {
> +	if (unlikely(barrier) && (q->ordered == QUEUE_ORDERED_NONE)) {
>  		err = -EOPNOTSUPP;
>  		goto end_io;
>  	}
>  
> -again:
>  	spin_lock_irq(q->queue_lock);
>  
>  	if (elv_queue_empty(q)) {
>  		blk_plug_device(q);
>  		goto get_rq;
>  	}

This should just goto get_rq, the plug should happen only at the end
where you did add it:

> @@ -2693,10 +2675,11 @@ get_rq:
>  	req->rq_disk = bio->bi_bdev->bd_disk;
>  	req->start_time = jiffies;
>  
> +	spin_lock_irq(q->queue_lock);
> +	if (elv_queue_empty(q))
> +		blk_plug_device(q);
>  	add_request(q, req);
>  out:
> -	if (freereq)
> -		__blk_put_request(q, freereq);
>  	if (bio_sync(bio))
>  		__generic_unplug_device(q);
>  

-- 
Jens Axboe

-
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/