Hi, this series fixes the bug that I found the other day and described here:
https://lists.gnu.org/archive/html/qemu-block/2019-03/msg00764.html https://lists.gnu.org/archive/html/qemu-block/2019-03/msg00791.html Patch 2 is the important one, patch 3 is a test case, and patch 1 will allow us to detect this kind of errors earlier. Regards, Berto v2: - Patch 2: Change the rollback order [Vladimir] - Patch 3: Use self.wait_until_completed() [Vladimir] v1: https://lists.gnu.org/archive/html/qemu-block/2019-03/msg00851.html - Initial version Output of backport-diff against v1: Key: [----] : patches are identical [####] : number of functional differences between upstream/downstream patch [down] : patch is downstream-only The flags [FC] indicate (F)unctional and (C)ontextual differences, respectively 001/3:[----] [--] 'block: continue until base is found in bdrv_freeze_backing_chain() et al' 002/3:[----] [-C] 'block: freeze the backing chain earlier in stream_start()' 003/3:[0017] [FC] 'block: test block-stream with a base node that is used by block-commit' Alberto Garcia (3): block: continue until base is found in bdrv_freeze_backing_chain() et al block: freeze the backing chain earlier in stream_start() block: test block-stream with a base node that is used by block-commit block.c | 21 ++++++++++++++------- block/stream.c | 13 +++++++------ tests/qemu-iotests/030 | 17 +++++++++++++++++ tests/qemu-iotests/030.out | 4 ++-- 4 files changed, 40 insertions(+), 15 deletions(-) -- 2.11.0
