[PATCH] Move the second call to get_request to the end of the loop

[PATCH] Move the second call to get_request to the end of the loop

[Zhang, Yanmin]

In function get_request_wait, the second call to get_request could be moved to the end of the while loop,
because if the first call to get_request fails, the second call will fail without sleep.

Signed-off-by: Zhang Yanmin <yanmin.zhang@intel.com>

---

--- linux-2.6.26-rc3/block/blk-core.c	2008-05-19 07:17:18.000000000 +0800
+++ linux-2.6.26-rc3_block/block/blk-core.c	2008-05-22 07:00:15.000000000 +0800
@@ -806,35 +806,32 @@ static struct request *get_request_wait(
 	rq = get_request(q, rw_flags, bio, GFP_NOIO);
 	while (!rq) {
 		DEFINE_WAIT(wait);
+		struct io_context *ioc;
 		struct request_list *rl = &q->rq;
 
 		prepare_to_wait_exclusive(&rl->wait[rw], &wait,
 				TASK_UNINTERRUPTIBLE);
 
-		rq = get_request(q, rw_flags, bio, GFP_NOIO);
-
-		if (!rq) {
-			struct io_context *ioc;
+		blk_add_trace_generic(q, bio, rw, BLK_TA_SLEEPRQ);
 
-			blk_add_trace_generic(q, bio, rw, BLK_TA_SLEEPRQ);
-
-			__generic_unplug_device(q);
-			spin_unlock_irq(q->queue_lock);
-			io_schedule();
+		__generic_unplug_device(q);
+		spin_unlock_irq(q->queue_lock);
+		io_schedule();
 
-			/*
-			 * After sleeping, we become a "batching" process and
-			 * will be able to allocate at least one request, and
-			 * up to a big batch of them for a small period time.
-			 * See ioc_batching, ioc_set_batching
-			 */
-			ioc = current_io_context(GFP_NOIO, q->node);
-			ioc_set_batching(q, ioc);
+		/*
+		 * After sleeping, we become a "batching" process and
+		 * will be able to allocate at least one request, and
+		 * up to a big batch of them for a small period time.
+		 * See ioc_batching, ioc_set_batching
+		 */
+		ioc = current_io_context(GFP_NOIO, q->node);
+		ioc_set_batching(q, ioc);
 
-			spin_lock_irq(q->queue_lock);
-		}
+		spin_lock_irq(q->queue_lock);
 		finish_wait(&rl->wait[rw], &wait);
-	}
+
+		rq = get_request(q, rw_flags, bio, GFP_NOIO);
+	};
 
 	return rq;
 }

Re: [PATCH] Move the second call to get_request to the end of the loop

[Jens Axboe]

On Thu, May 22 2008, Zhang, Yanmin wrote:
> In function get_request_wait, the second call to get_request could be
> moved to the end of the while loop, because if the first call to
> get_request fails, the second call will fail without sleep.

Looks safe enough, since we don't drop the lock in between getting the
request and adding to the wait queue. I'll add it to the testing mix,
thanks!


-- 
Jens Axboe