Re: [PULL 14/53] virtio-rng-pci: fix transitional migration compat for vectors

[Dr. David Alan Gilbert]

* Michael Tokarev (mjt@tls.msk.ru) wrote:
> 02.03.2023 11:25, Michael S. Tsirkin wrote:
> > From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
> > 
> > In bad9c5a516 ("virtio-rng-pci: fix migration compat for vectors") I
> > fixed the virtio-rng-pci migration compatibility, but it was discovered
> > that we also need to fix the other aliases of the device for the
> > transitional cases.
> > 
> > Fixes: 9ea02e8f1 ('virtio-rng-pci: Allow setting nvectors, so we can use 
> > MSI-X')
> > bz: https://bugzilla.redhat.com/show_bug.cgi?id=2162569
> > Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
> > Message-Id: <20230207174944.138255-1-dgilbert@redhat.com>
> > Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > ---
> >   hw/core/machine.c | 2 ++
> >   1 file changed, 2 insertions(+)
> > 
> > diff --git a/hw/core/machine.c b/hw/core/machine.c
> > index f29e700ee4..1cf6822e06 100644
> > --- a/hw/core/machine.c
> > +++ b/hw/core/machine.c
> > @@ -47,6 +47,8 @@ const size_t hw_compat_7_2_len = 
> > G_N_ELEMENTS(hw_compat_7_2);
> >   GlobalProperty hw_compat_7_1[] = {
> >       { "virtio-device", "queue_reset", "false" },
> >       { "virtio-rng-pci", "vectors", "0" },
> > +    { "virtio-rng-pci-transitional", "vectors", "0" },
> > +    { "virtio-rng-pci-non-transitional", "vectors", "0" },
> >   };
> 
> If we consider this one for 7.2 stable, the previous change here, which
> added "virtio-rng-pci" right before the lines being added, should also
> be picked up, which is bad9c5a5166fd5e3a892b7b0477cf2f4bd3a959a:
>  From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
>  Date: Mon, 9 Jan 2023 10:58:09 +0000
>  Subject: virtio-rng-pci: fix migration compat for vectors
> 
> Should the two both be included in -stable, or both omitted?

I think both included; however note the warning in the original patch
(that's also true in this one); live migration from an unpatched world
to the patched would will break, but it fixes live migration from older
Qemu; you get to pick your favourite breakage.

Dave

> Thanks,
> 
> /mjt
> 
-- 
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK