[libvirt] [PATCH 2/9] qemu: process: Don't try to redetect missing qemuCaps on reconnect

[Daniel P. Berrangé]

On Mon, Nov 25, 2019 at 02:07:13PM +0100, Peter Krempa wrote:
> On Mon, Nov 25, 2019 at 12:59:00 +0000, Daniel Berrange wrote:
> > On Mon, Nov 25, 2019 at 01:53:47PM +0100, Peter Krempa wrote:
> > > The redetection was originally added in 43c01d3838 as a way to recover
> > > from libvirtd upgrade from the time when we didn't persist the qemu
> > > capabilities in the status XML. Also this the oldest supported qemu by
> > > more than two years.
> > > 
> > > Even if somebody would have a running VM running at least qemu 1.5 with
> > > such an old libvirt we certainly wouldn't do the right thing by
> > > redetecting the capabilities and then trying to communicate with qemu.
> > > 
> > > For now it will be the best to just stop considering this scenario any
> > > more.
> > > 
> > > Signed-off-by: Peter Krempa <pkrempa at redhat.com>
> > > ---
> > >  src/qemu/qemu_process.c | 7 -------
> > >  1 file changed, 7 deletions(-)
> > > 
> > > diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
> > > index 1b88c471f4..a76a8da841 100644
> > > --- a/src/qemu/qemu_process.c
> > > +++ b/src/qemu/qemu_process.c
> > > @@ -8085,13 +8085,6 @@ qemuProcessReconnect(void *opaque)
> > >          goto error;
> > >      }
> > > 
> > > -    /* If upgrading from old libvirtd we won't have found any
> > > -     * caps in the domain status, so re-query them
> > > -     */
> > > -    if (!priv->qemuCaps &&
> > > -        (qemuDomainUpdateQEMUCaps(obj, driver->qemuCapsCache) < 0))
> > > -        goto error;
> > 
> > Shouldn't we be making missing qemuCaps into a fatal error
> > when loading that VM.
> 
> I thought about it for a bit. If we make it an error the VM will vanish
> after upgrade (and continue running). If we don't make an error it will
> not vanish and you'll be able to at least destroy it.
> 
> Given that such situation is super-unlikely, I opted to remove it.
> Nobody will probably run into this situation and if yes then the VM will
> not vanish at least.
> 
> I can add an error though, but to me it seems it's not worth.

My concern is that pretty much all our later code will assume that
priv->qemuCaps is non-NULL.  So by ignoring the error, the VM might
not vanish initially, but libvirtd may well crash shortly after.

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 :|