[PATCH 2/4] conf: allow to map sound device to host device

Daniel P. Berrangé berrange at redhat.com
Wed Jul 22 14:42:53 UTC 2020

On Wed, Jul 22, 2020 at 08:55:02AM +0200, Gerd Hoffmann wrote:
>   Hi,
> > IIUC, QEMU can expose multiple sound devices to the guest too.
> > 
> > I think this means that we can have a M:N relationship between
> > a sound device, and an audio backend, not just 1:1.
> It's 1:N.  Sound devices have a single backend, but a backends can
> service multiple sound devices.

Sigh yes, of course, wasn't thinking right.

> > Assuming I'm right about the M:N relationship, I assume that
> > of multiple cards all do playback concurrently, something
> > will have todo mixing of the streams ?
> In general it is a good idea to go with 1:1 if possible.  With
> pulseaudio this works fine.  You'll have two streams to pulseaudio
> then, pulseaudio does the mixing, and you'll see both streams in
> mixer apps.
> OSS devices tend to not like being opened multiple times, so you
> have to go with 1:N if you want multiple sound devices.  qemu mixes
> the playback streams then.  Not fully sure what happens with
> recording, probably all sound devices see the very same stream.

Ok, so it sounds like we need the XML schema I illustrated. Even though
we'll recommend 1:1 for general use, we'll want the option to supoort
1:N setups.

|: 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 :|

More information about the libvir-list mailing list