qemu-img rebase: Fix output image corruption
authorKevin Wolf <kwolf@redhat.com>
Thu, 29 Apr 2010 12:47:48 +0000 (14:47 +0200)
committerKevin Wolf <kwolf@redhat.com>
Mon, 3 May 2010 08:07:32 +0000 (10:07 +0200)
qemu-img rebase must always give clusters in the COW file priority over those
in the backing file. As it failed to use number of non-allocated clusters but
assumed the maximum, it was possible that allocated clusters were taken from
the backing file instead, leading to a corrupted output image.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
qemu-img.c

index c21d9990e050cb44f01722a11e1275b204145be9..d3c30a74f371581136903bbba44cf7d9f75e378f 100644 (file)
@@ -1136,7 +1136,7 @@ static int img_rebase(int argc, char **argv)
     if (!unsafe) {
         uint64_t num_sectors;
         uint64_t sector;
-        int n, n1;
+        int n;
         uint8_t * buf_old;
         uint8_t * buf_new;
 
@@ -1155,8 +1155,8 @@ static int img_rebase(int argc, char **argv)
             }
 
             /* If the cluster is allocated, we don't need to take action */
-            if (bdrv_is_allocated(bs, sector, n, &n1)) {
-                n = n1;
+            ret = bdrv_is_allocated(bs, sector, n, &n);
+            if (ret) {
                 continue;
             }