Re: [RFC QEMU PATCH] ui: Make the DisplayType enum entries conditional

[Thomas Huth]

On 09/06/2021 14.50, Gerd Hoffmann wrote:
>    Hi,
>
> > > The #if CONFIG_SDL approach will not work because qemu will continue to
> > > report sdl as supported even when the sdl module is not installed any
> > > more.
> >
> > I guess we'd need a separate QMP command to fix that, which tries to load
> > the modules first when being called? Something similar to what is being done
> > in qemu_display_help() ?
>
> That would work, yes.
>
> > That's certainly doable, too, just a little bit more complex...
>
> Alternative idea: turn QemuDisplay into an ObjectClass, then it'll be
> visible in qom introspection.  Likewise a bit more complex ...
>
> > do we want that?  Or is the quick-n-easy way via the schema good
> > enough for most use cases?
>
> Would be better than nothing, but I'd prefer something which works
> properly with modular qemu ...

I'm not sure whether we can even make it 100% rock solid if we introduce a 
dedicated QMP command here. For example imagine that libvirt did its probing 
while a X11 server was running, so it discovered that QEMU could be used 
with SDL. Now the user stops the X11 server, thus the cached information 
that QEMU could be used with SDL becomes useless. I think that's somehow 
similar to the situation whether the module is available or not. The 
information that is shown by "virsh domcapabilities" can only be an 
indication of what can be used in the best case (module available, X11 
server running etc), but it can never be a 100% guarantee that the UI 
interface can really really be used.
Thus I tend to continue with the simple way via the QAPI schema right now, 
unless someone really has an urgent need for a separate QMP command (at 
least for the BZ that I listed in my original mail, the simple way via the 
QAPI schema is enough).

 Thomas