On Thu, Aug 29, 2019 at 04:30:29PM +0400, Marc-André Lureau wrote:
Hi On Tue, Aug 27, 2019 at 11:19 AM Ján Tomko <jtomko@xxxxxxxxxx> wrote:On Fri, Aug 23, 2019 at 12:51:56PM -0400, Cole Robinson wrote: >On 8/23/19 12:21 PM, Cole Robinson wrote: >> From: Marc-André Lureau <marcandre.lureau@xxxxxxxxxx> >> >> Accept a new attribute to specify usage of helper process, ex: >> >> <video> >> <model type='virtio' vhostuser='yes'/> >> </video> >> > >For other devices, we have > ><interface type='vhostuser'> ><interface><driver name='vhost'/> ><hostdev type='scsi_host'><source protocol='vhost'> > >Which is an attempt to make this more generic. >IMO using vhostuser='yes' missing verb? >and is the simplest match to qemu terminology but maybe other people >have stronger opinions. Yes. This does not belong in the model - it's a backend thing. <driver name='vhost'/> sounds best to me (and similarly to what commit 175077fd707db6ad87d6e2a079e82bc290ac2421 did, we can also expose <driver name='qemu'/>)Shouldn't it be <driver name='vhost-user'/>?
Even better :) Jano
('vhost' would be for a kernel backend)Jano >
Attachment:
signature.asc
Description: PGP signature
-- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list