On Wed 13 Mar 2019 03:32:09 PM CET, Alberto Garcia wrote: >>> > 2. And therefore, why we just don't call bdrv_replace_node(top, >>> > base, errp) from bdrv_drop_intermediate? >>> >>> I think that would not call role->update_filename(). >> >> And role->update_filename() involves I/O, so you can't roll back >> across it anyway. I think this was the reason why we didn't roll back >> the in-memory state either, it would become inconsistent with what is >> on the disk. > > But I think at least for the bdrv_check_update_perm() case we should > call bdrv_abort_perm_update() on all nodes if things fail, and we're > currently not doing it.
Forget this, bdrv_replace_child() already sets the new permissions after every iteration of the loop, so the only place that needs to call bdrv_abort_perm_update() is where we are doing it already. Berto
