On Thu, Jan 30, 2025 at 04:42:24PM +0400, Marc-André Lureau wrote:
> Hi
>
> On Wed, Jan 29, 2025 at 7:25 PM Daniel P. Berrangé <berrange@xxxxxxxxxx> wrote:
> >
> > On Wed, Jan 29, 2025 at 05:40:40PM +0400, marcandre.lureau@xxxxxxxxxx wrote:
> > > From: Marc-André Lureau <marcandre.lureau@xxxxxxxxxx>
> > >
> > > Wire the external server RDP support with QEMU.
> > >
> > > Check the configuration, allocate a port, start the process
> > > and set the credentials.
> > >
> > > Signed-off-by: Marc-André Lureau <marcandre.lureau@xxxxxxxxxx>
> > > ---
> > > docs/formatdomain.rst | 25 ++++--
> > > src/conf/domain_conf.h | 1 +
> > > src/qemu/qemu_extdevice.c | 46 +++++++++--
> > > src/qemu/qemu_hotplug.c | 49 ++++++++++-
> > > src/qemu/qemu_hotplug.h | 1 +
> > > src/qemu/qemu_process.c | 167 ++++++++++++++++++++++++++++++++++----
> > > 6 files changed, 257 insertions(+), 32 deletions(-)
> >
> >
> > > + if (!virDomainDefHasDBus(vm->def, false)) {
> > > + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
> > > + _("qemu-rdp support requires a D-Bus bus graphics device."));
> > > + return -1;
> > > + }
> >
> > IMHO this is not the correct way to model it.
> >
> > Use of DBus is an internal impl detail of the RDP backend that is
> > being configured. Users shouldn't have to know that - they should
> > just be able to request 'RDP' alone for the <graphics> type.
> >
> > Consider if we wanted to provide a non-integrated VNC server for
> > QEMU. We wouldn't do that by requiring use of a <graphics type=dbus>
> > alongside the <graphics type=vnc>, as that config has the distinct
> > meaning of requiring VNC and DBus in parallel.
> >
> > Where we have a choice of underlying implementations, we would
> > typically have something like a <backend type="qemu|dbus"/>
> > as a child element to select between them.
>
> I wish it would be an implementation detail. Unfortunately, the qemu
> devices needs to be configured to export themself on dbus too
> (chardev/usb/audio), this is also exposed on the domain XML. For the
> most simple case, ie display only, we could probably hide the dbus
> implementation detail. But in general, we probably better be explicit
> about it. Being explicit would also avoid surprises when checking the
> list of available displays (think domdisplay --all, or by inspecting
> the bus). I am not sure it's worth trying to hide it (for ex, should
> <redirdev bus='usb'> be exposed on dbus by default when dbus or rdp is
> enabled?). If people feel it's really best, I can give it a try, but I
> fear it's going to be more struggle and surprises without much
> benefit.
Hmm, yeah, that's awkward really, so guess we can't do much better than
having dual <graphics> tags :-(
With 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 :|
