Patchwork [03/15] block: perform zero-detection during copy-on-read

login
register
mail settings
Submitter Kevin Wolf
Date Feb. 10, 2012, 12:47 p.m.
Message ID <1328878064-4907-4-git-send-email-kwolf@redhat.com>
Download mbox | patch
Permalink /patch/140610/
State New
Headers show

Comments

Kevin Wolf - Feb. 10, 2012, 12:47 p.m.
From: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>

Copy-on-Read populates the image file with data read from a backing
image.  In order to avoid bloating the image file when all zeroes are
read we should scan the buffer and perform an optimized zero write
operation.

Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 block.c |   14 +++++++++++---
 1 files changed, 11 insertions(+), 3 deletions(-)

Patch

diff --git a/block.c b/block.c
index c9fa5c1..ae297bb 100644
--- a/block.c
+++ b/block.c
@@ -1517,6 +1517,7 @@  static int coroutine_fn bdrv_co_do_copy_on_readv(BlockDriverState *bs,
      */
     void *bounce_buffer;
 
+    BlockDriver *drv = bs->drv;
     struct iovec iov;
     QEMUIOVector bounce_qiov;
     int64_t cluster_sector_num;
@@ -1537,14 +1538,21 @@  static int coroutine_fn bdrv_co_do_copy_on_readv(BlockDriverState *bs,
     iov.iov_base = bounce_buffer = qemu_blockalign(bs, iov.iov_len);
     qemu_iovec_init_external(&bounce_qiov, &iov, 1);
 
-    ret = bs->drv->bdrv_co_readv(bs, cluster_sector_num, cluster_nb_sectors,
-                                 &bounce_qiov);
+    ret = drv->bdrv_co_readv(bs, cluster_sector_num, cluster_nb_sectors,
+                             &bounce_qiov);
     if (ret < 0) {
         goto err;
     }
 
-    ret = bs->drv->bdrv_co_writev(bs, cluster_sector_num, cluster_nb_sectors,
+    if (drv->bdrv_co_write_zeroes &&
+        buffer_is_zero(bounce_buffer, iov.iov_len)) {
+        ret = drv->bdrv_co_write_zeroes(bs, cluster_sector_num,
+                                        cluster_nb_sectors);
+    } else {
+        ret = drv->bdrv_co_writev(bs, cluster_sector_num, cluster_nb_sectors,
                                   &bounce_qiov);
+    }
+
     if (ret < 0) {
         /* It might be okay to ignore write errors for guest requests.  If this
          * is a deliberate copy-on-read then we don't want to ignore the error.