04.02.2022 14:10, Hanna Reitz wrote:
Hi,

The original RFC is here:
https://lists.nongnu.org/archive/html/qemu-block/2022-01/msg00765.html

And v1 (whose cover letter has some exposition) is here:
https://lists.nongnu.org/archive/html/qemu-block/2022-02/msg00124.html


In v2, I’ve fixed the QemuStorageDaemon class added in patch 4 as
suggested by Vladimir:
- Have `__init__()`’s `instance_id` parameter have its actual default
   value instead of some weird `if instance_id is None` construct
- Assert that the QSD class users won’t use --pidfile

(And added Vladimir’s R-b to all patches.)

git 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

001/7:[----] [--] 'block/nbd: Delete reconnect delay timer when done'
002/7:[----] [--] 'block/nbd: Delete open timer when done'
003/7:[----] [--] 'block/nbd: Assert there are no timers when closed'
004/7:[0006] [FC] 'iotests.py: Add QemuStorageDaemon class'
005/7:[----] [--] 'iotests/281: Test lingering timers'
006/7:[----] [--] 'block/nbd: Move s->ioc on AioContext change'
007/7:[----] [--] 'iotests/281: Let NBD connection yield in iothread'


Hanna Reitz (7):
   block/nbd: Delete reconnect delay timer when done
   block/nbd: Delete open timer when done
   block/nbd: Assert there are no timers when closed
   iotests.py: Add QemuStorageDaemon class
   iotests/281: Test lingering timers
   block/nbd: Move s->ioc on AioContext change
   iotests/281: Let NBD connection yield in iothread

  block/nbd.c                   |  64 +++++++++++++++++++++
  tests/qemu-iotests/281        | 101 +++++++++++++++++++++++++++++++++-
  tests/qemu-iotests/281.out    |   4 +-
  tests/qemu-iotests/iotests.py |  40 ++++++++++++++
  4 files changed, 205 insertions(+), 4 deletions(-)


Thanks a lot for fixing this.

Applied to my nbd branch. I'll wait for additional comments until Tuesday 
before preparing a pull req.

--
Best regards,
Vladimir

Reply via email to