δΊ 2013-4-1 23:52, Eric Blake ει:
On 04/01/2013 04:01 AM, Wenchao Xia wrote:
Last operaton should be cancelled first.
s/operaton/operation/
[I don't care enough about US vs. UK to say whether canceled or
cancelled looks better]
Signed-off-by: Wenchao Xia <xiaw...@linux.vnet.ibm.com>
---
blockdev.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/blockdev.c b/blockdev.c
index 75416fb..a24d10e 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -954,7 +954,7 @@ void qmp_transaction(BlockdevActionList *dev_list, Error
**errp)
dev_entry = dev_entry->next;
states = g_malloc0(sizeof(BlkTransactionStates));
- QSIMPLEQ_INSERT_TAIL(&snap_bdrv_states, states, entry);
+ QSIMPLEQ_INSERT_HEAD(&snap_bdrv_states, states, entry);
Is this a bug fix that for something that can be triggered by existing
use of the 'transaction' command even without the additions you made in
patches 1 and 2? If so, this probably ought to come first in the
series, and you probably ought to consider enhancing the testsuite to
show why it matters.
Originally it only support backing chain, it will have no
difference about the sequence to delete it, but matters if
other types of operation are added.
--
Best Regards
Wenchao Xia