06.03.2017 17:23, Denis V. Lunev wrote:
> On 02/25/2017 10:31 PM, Vladimir Sementsov-Ogievskiy wrote:
>> Hi all!
>>
>> Here are some migration related bugs, two about INACTIVE bdses and one
>> use-after-free.
>>
>> I'm absolutely not sure, that these bugs should be fixed like I'm fixing,
>> but problem definitely exists.
>>
>> Reset in stopped state is strange case, may be such usage should be
>> restricted.
>> About INACTIVE - looks like it should be a separate run-state, not only
>> bdrv-flag.
>> Situation with migration state, which is global, but is set/reset/changed
>> in not controlled manner is not very good too..
>>
>> Vladimir Sementsov-Ogievskiy (4):
>> iotests: add migration corner cases test
>> qmp-cont: invalidate on RUN_STATE_PRELAUNCH
>> savevm: fix savevm after migration
>> migration: fix use-after-free of to_dst_file
>>
>> block/snapshot.c | 3 +-
>> migration/savevm.c | 16 +++++++++++
>> qmp.c | 3 +-
>> tests/qemu-iotests/175 | 71 ++++++++++++++++++++++++++++++++++++++++++++++
>> tests/qemu-iotests/175.out | 5 ++++
>> tests/qemu-iotests/group | 1 +
>> 6 files changed, 97 insertions(+), 2 deletions(-)
>> create mode 100644 tests/qemu-iotests/175
>> create mode 100644 tests/qemu-iotests/175.out
>>
> guys, what about patches 1-3?
>
> Den
At least, with 2 and 3, fixing an assert? Do someone going to work with
adding a runstate?
--
Best regards,
Vladimir