On Thu, Mar 07, 2019 at 11:46:19AM -0500, Michael S. Tsirkin wrote: > On Thu, Mar 07, 2019 at 04:19:44PM +0000, Daniel P. Berrangé wrote: > > On Thu, Mar 07, 2019 at 11:05:36AM -0500, Michael S. Tsirkin wrote: > > > On Thu, Mar 07, 2019 at 02:51:20PM +0000, Daniel P. Berrangé wrote: > > > > The broadening of vhost-user support is useful with that in much the > > > > same way I imagine. > > > > > > vhost user has more of an impact but is also a bigger maintainance > > > burden as clients are packaged, can be restarted etc individually. > > > > It feels like we're having/accepted that cost already though since > > vhostuser exists today & has been expanding to cover more backends. > > What I am trying to say is that we could eaily add support for > extensions just for in-tree code since these don't create an API that > needs to be maintained. > > So e.g. we do not need feature negotiation.
Ah, I see what you mean now. Having stuff in-tree makes migration saner too since we don't have combinatorial expansion of impls to worry about testnig > But yes, this could be an extension of vhost-user in some way. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|