Callers don't agree whether bdrv_reopen_queue_child() should be called with the AioContext lock held or not. Standardise on holding the lock (as done by QMP blockdev-reopen and the replication block driver) and fix bdrv_reopen() to do the same.
Signed-off-by: Kevin Wolf <kw...@redhat.com> --- block.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/block.c b/block.c index 214e5055a5..191dfc5d0c 100644 --- a/block.c +++ b/block.c @@ -4153,6 +4153,8 @@ static bool bdrv_recurse_has_child(BlockDriverState *bs, * bs_queue, or the existing bs_queue being used. * * bs must be drained between bdrv_reopen_queue() and bdrv_reopen_multiple(). + * + * To be called with bs->aio_context locked. */ static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue, BlockDriverState *bs, @@ -4311,6 +4313,7 @@ static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue, return bs_queue; } +/* To be called with bs->aio_context locked */ BlockReopenQueue *bdrv_reopen_queue(BlockReopenQueue *bs_queue, BlockDriverState *bs, QDict *options, bool keep_old_opts) @@ -4475,11 +4478,11 @@ int bdrv_reopen(BlockDriverState *bs, QDict *opts, bool keep_old_opts, GLOBAL_STATE_CODE(); bdrv_subtree_drained_begin(bs); + queue = bdrv_reopen_queue(NULL, bs, opts, keep_old_opts); + if (ctx != qemu_get_aio_context()) { aio_context_release(ctx); } - - queue = bdrv_reopen_queue(NULL, bs, opts, keep_old_opts); ret = bdrv_reopen_multiple(queue, errp); if (ctx != qemu_get_aio_context()) { -- 2.38.1