Hi,

bdrv_replace_child_noperm() drains the child via
bdrv_parent_drained_{begin,end}_single().  When it removes a child, the
bdrv_parent_drained_end_single() at its end will be called on an empty
child, making the BDRV_POLL_WHILE() in it poll the main AioContext
(because c->bs is NULL).

That\ufffd\ufffd\ufffds wrong, though, because it\ufffd\ufffd\ufffds supposed to operate on the parent.
bdrv_parent_drained_end_single_no_poll() will have scheduled any BHs in
the parents\ufffd\ufffd\ufffd AioContext, which may be anything, not necessarily the main
context.  Therefore, we must poll the parent\ufffd\ufffd\ufffds context.

Patch 3 does this for both bdrv_parent_drained_{begin,end}_single().
Patch 1 ensures that we can legally call
bdrv_child_get_parent_aio_context() from those functions (currently
marked as GLOBAL_STATE_CODE(), which I don\ufffd\ufffd\ufffdt think it is), and patch 2
fixes blk_do_set_aio_context() to not cause an assertion failure if it
beginning a drain can end up in blk_get_aio_context() before blk->ctx
has been updated.


Hanna Reitz (3):
  block: bdrv_child_get_parent_aio_context is not GS
  block-backend: Update ctx immediately after root
  block: Start/end drain on correct AioContext

 block.c               | 2 +-
 block/block-backend.c | 4 +++-
 block/io.c            | 6 ++++--
 3 files changed, 8 insertions(+), 4 deletions(-)

-- 
2.36.1