On 18/09/2017 11:37, Pavel Butsykin wrote: >>> >> >> To understand the patch better this doesn't fix _all_ stopped states, >> only migration, right? That said it's a valid bugfix even independent >> of the effects for stopped runstate. > Yes, the bug only appears on the migration. Actually, to protect memory > during the migration, this approach is already used for other virtio > devices, for example net_vhost, see virtio_net_vhost_status().
Cool, can you rebase? Paolo