On 10/21/2014 10:38 AM, Zhang Haoyu wrote:
If there are still pending i/o while deleting snapshot,
because deleting snapshot is done in non-coroutine context, and
the pending i/o read/write (bdrv_co_do_rw) is done in coroutine context,
so it's possible to cause concurrency problem between above two operations.
Add bdrv_drain_all() to bdrv_snapshot_delete() to avoid this problem.

Signed-off-by: Zhang Haoyu <zhan...@sangfor.com>
---
  block/snapshot.c | 4 ++++
  1 file changed, 4 insertions(+)

diff --git a/block/snapshot.c b/block/snapshot.c
index 85c52ff..ebc386a 100644
--- a/block/snapshot.c
+++ b/block/snapshot.c
@@ -236,6 +236,10 @@ int bdrv_snapshot_delete(BlockDriverState *bs,
          error_setg(errp, "snapshot_id and name are both NULL");
          return -EINVAL;
      }
+
+    /* drain all pending i/o before deleting snapshot */
+    bdrv_drain_all();
+
      if (drv->bdrv_snapshot_delete) {
          return drv->bdrv_snapshot_delete(bs, snapshot_id, name, errp);
      }


Reviewed-by: Paolo Bonzini <pbonz...@redhat.com>

Reply via email to