On 02/04/20 07:37, Markus Armbruster wrote: > Paolo Bonzini <[email protected]> writes: > >> On 01/04/20 17:09, Stefan Hajnoczi wrote: >>>> What do you think about it? >>> >>> Maybe aio_disable_external() is needed to postpone device emulation >>> until after realize has finished? >>> >>> Virtqueue kick ioeventfds are marked "external" and won't be processed >>> while external events are disabled. See also >>> virtio_queue_aio_set_host_notifier_handler() -> >>> aio_set_event_notifier(). >> >> Yes, I think Stefan is right. > > Is this issue limited to SCSI devices?
Yes, because they (and more specifically virtio-scsi) are the only ones that can be processed 1) outside the BQL 2) by a device other than themselves (in this case, by the adapter). Paolo
