summary refs log tree commit diff stats
path: root/qemu-io-cmds.c
diff options
context:
space:
mode:
authorLukas Straub <lukasstraub2@web.de>2021-05-18 13:42:14 +0200
committerKevin Wolf <kwolf@redhat.com>2021-06-02 14:23:20 +0200
commit5529b02da2dcd1ef6bc6cd42d4fbfb537fe2276f (patch)
treec5252dcc187ab20c8d68a6f61ceb2db48890a3fd /qemu-io-cmds.c
parentdd2db39d78431ab5a0b78777afaab3d61e94533e (diff)
downloadfocaccia-qemu-5529b02da2dcd1ef6bc6cd42d4fbfb537fe2276f.tar.gz
focaccia-qemu-5529b02da2dcd1ef6bc6cd42d4fbfb537fe2276f.zip
block/quorum: Provide .bdrv_co_flush instead of .bdrv_co_flush_to_disk
The quorum block driver uses a custom flush callback to handle the
case when some children return io errors. In that case it still
returns success if enough children are healthy.
However, it provides it as the .bdrv_co_flush_to_disk callback, not
as .bdrv_co_flush. This causes the block layer to do it's own
generic flushing for the children instead, which doesn't handle
errors properly.

Fix this by providing .bdrv_co_flush instead of
.bdrv_co_flush_to_disk so the block layer uses the custom flush
callback.

Signed-off-by: Lukas Straub <lukasstraub2@web.de>
Reported-by: Minghao Yuan <meeho@qq.com>
Message-Id: <20210518134214.11ccf05f@gecko.fritz.box>
Tested-by: Zhang Chen <chen.zhang@intel.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Diffstat (limited to 'qemu-io-cmds.c')
0 files changed, 0 insertions, 0 deletions