[Qemu-devel] [PATCH 03/20] qemu-img: do not zero-pad the compressed write buffer

[Kevin Wolf <kwolf@redhat.com>]

From: Stefan Hajnoczi <stefanha@redhat.com>

bdrv_write_compressed() does not allow requests that span the end of the
device.  Therefore it is useless to zero-pad the last cluster and
thereby exceed the end of the device.

Let image formats handle zero-padding the final compressed cluster, if
necessary.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 qemu-img.c | 8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)

diff --git a/qemu-img.c b/qemu-img.c
index 31627b0..cd096a1 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -1425,12 +1425,8 @@ static int img_convert(int argc, char **argv)
             }
             assert (remainder == 0);
 
-            if (n < cluster_sectors) {
-                memset(buf + n * 512, 0, cluster_size - n * 512);
-            }
-            if (!buffer_is_zero(buf, cluster_size)) {
-                ret = bdrv_write_compressed(out_bs, sector_num, buf,
-                                            cluster_sectors);
+            if (!buffer_is_zero(buf, n * BDRV_SECTOR_SIZE)) {
+                ret = bdrv_write_compressed(out_bs, sector_num, buf, n);
                 if (ret != 0) {
                     error_report("error while compressing sector %" PRId64
                                  ": %s", sector_num, strerror(-ret));
-- 
1.8.1.4