]> xenbits.xensource.com Git - people/liuw/libxenctrl-split/qemu-xen-traditional.git/commitdiff
xen_disk: support empty write barriers
authorIan Jackson <ian.jackson@eu.citrix.com>
Tue, 14 Dec 2010 18:42:20 +0000 (18:42 +0000)
committerIan Jackson <Ian.Jackson@eu.citrix.com>
Tue, 14 Dec 2010 18:42:20 +0000 (18:42 +0000)
this patch can be applied to both qemu-xen and qemu and adds support
for empty write barriers to xen_disk.

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Acked-by: Gerd Hoffmann <kraxel@redhat.com>
hw/xen_disk.c

index 38b5fbf814744954c38abccccf1597a6bcec7952..94af0014560155fb1bce608d5dc76e779678deff 100644 (file)
@@ -182,6 +182,10 @@ static int ioreq_parse(struct ioreq *ioreq)
        ioreq->prot = PROT_WRITE; /* to memory */
        break;
     case BLKIF_OP_WRITE_BARRIER:
+        if (!ioreq->req.nr_segments) {
+            ioreq->presync = 1;
+            return 0;
+        }
        if (!syncwrite)
            ioreq->presync = ioreq->postsync = 1;
        /* fall through */
@@ -306,7 +310,7 @@ static int ioreq_runio_qemu_sync(struct ioreq *ioreq)
     int i, rc, len = 0;
     off_t pos;
 
-    if (ioreq_map(ioreq) == -1)
+    if (ioreq->req.nr_segments && ioreq_map(ioreq) == -1)
        goto err;
     if (ioreq->presync)
        bdrv_flush(blkdev->bs);
@@ -330,6 +334,8 @@ static int ioreq_runio_qemu_sync(struct ioreq *ioreq)
        break;
     case BLKIF_OP_WRITE:
     case BLKIF_OP_WRITE_BARRIER:
+        if (!ioreq->req.nr_segments)
+            break;
        pos = ioreq->start;
        for (i = 0; i < ioreq->v.niov; i++) {
            rc = bdrv_write(blkdev->bs, pos / BLOCK_SIZE,
@@ -387,7 +393,7 @@ static int ioreq_runio_qemu_aio(struct ioreq *ioreq)
 {
     struct XenBlkDev *blkdev = ioreq->blkdev;
 
-    if (ioreq_map(ioreq) == -1)
+    if (ioreq->req.nr_segments && ioreq_map(ioreq) == -1)
        goto err;
 
     ioreq->aio_inflight++;
@@ -404,6 +410,8 @@ static int ioreq_runio_qemu_aio(struct ioreq *ioreq)
     case BLKIF_OP_WRITE:
     case BLKIF_OP_WRITE_BARRIER:
         ioreq->aio_inflight++;
+        if (!ioreq->req.nr_segments)
+            break;
         bdrv_aio_writev(blkdev->bs, ioreq->start / BLOCK_SIZE,
                         &ioreq->v, ioreq->v.size / BLOCK_SIZE,
                         qemu_aio_complete, ioreq);