From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jeremy Fitzhardinge <jeremy@goop.org>
Subject: [PATCH RFC] xen/blkfront: use tagged queuing for
	barriers
Date: Thu, 22 Jul 2010 14:20:31 -0700
Message-ID: <4C48B61F.2030302@goop.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Return-path: <xen-devel-bounces@lists.xensource.com>
List-Unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
List-Post: <mailto:xen-devel@lists.xensource.com>
List-Help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-Subscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
Sender: xen-devel-bounces@lists.xensource.com
Errors-To: xen-devel-bounces@lists.xensource.com
To: Daniel Stodden <daniel.stodden@citrix.com>
Cc: "Xen-devel@lists.xensource.com" <Xen-devel@lists.xensource.com>, Jake Wires <Jake.Wires@citrix.com>
List-Id: xen-devel@lists.xenproject.org

When barriers are supported, then use QUEUE_ORDERED_TAG to tell the block
subsystem that it doesn't need to do anything else with the barriers.
Previously we used ORDERED_DRAIN which caused the block subsystem to
drain all pending IO before submitting the barrier, which would be
very expensive.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index 92091a7..b61a021 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c
@@ -418,10 +418,20 @@ static int xlvbd_init_blk_queue(struct gendisk *gd, u16 sector_size)
 static int xlvbd_barrier(struct blkfront_info *info)
 {
 	int err;
+	unsigned ordered = QUEUE_ORDERED_NONE;
 
-	err = blk_queue_ordered(info->rq,
-				info->feature_barrier ? QUEUE_ORDERED_DRAIN : QUEUE_ORDERED_NONE,
-				NULL);
+	/*
+	 * If we don't have barrier support, then there's really no
+	 * way to guarantee write ordering, so we really just have to
+	 * send writes to the backend and hope for the best.  If
+	 * barriers are supported, we don't really know what the
+	 * flushing semantics of the barrier are, so again, tag the
+	 * requests in order and hope for the best.
+	 */
+	if (info->feature_barrier)
+		ordered = QUEUE_ORDERED_TAG;
+
+	err = blk_queue_ordered(info->rq, ordered, NULL);
 
 	if (err)
 		return err;
@@ -508,8 +518,7 @@ static int xlvbd_alloc_gendisk(blkif_sector_t capacity,
 	info->rq = gd->queue;
 	info->gd = gd;
 
-	if (info->feature_barrier)
-		xlvbd_barrier(info);
+	xlvbd_barrier(info);
 
 	if (vdisk_info & VDISK_READONLY)
 		set_disk_ro(gd, 1);