Re: [Qemu-devel] [PATCH 3/9] vfio: unplug failover primary device before migration

[Cornelia Huck <cohuck@redhat.com>]

On Tue, 13 Aug 2019 08:45:49 +0200
Jens Freimann <jfreimann@redhat.com> wrote:

> On Mon, Aug 12, 2019 at 03:22:52PM -0600, Alex Williamson wrote:
> >On Mon, 12 Aug 2019 17:18:54 +0200
> >Cornelia Huck <cohuck@redhat.com> wrote:
> >  
> >> On Fri,  2 Aug 2019 17:05:59 +0200
> >> Jens Freimann <jfreimann@redhat.com> wrote:
> >>  
> >> > As usual block all vfio-pci devices from being migrated, but make an
> >> > exception for failover primary devices. This is achieved by setting
> >> > unmigratable to 0 but also add a migration blocker for all vfio-pci
> >> > devices except failover primary devices. These will be unplugged before
> >> > migration happens by the migration handler of the corresponding
> >> > virtio-net standby device.
> >> >
> >> > Signed-off-by: Jens Freimann <jfreimann@redhat.com>
> >> > ---
> >> >  hw/vfio/pci.c | 24 +++++++++++++++++++++++-
> >> >  hw/vfio/pci.h |  1 +
> >> >  2 files changed, 24 insertions(+), 1 deletion(-)

> >> This patch interacts with support for vfio migration (last posted in
> >> <1562665760-26158-1-git-send-email-kwankhede@nvidia.com>, I've not seen
> >> a later version yet.)
> >>
> >> With that, we'd have three cases to consider:
> >> 1) device is a failover primary
> >> 2) device has a migration region
> >> 3) none of the above
> >>
> >> Can 1) and 2) happen simultaneously? If yes, what should take
> >> precedence?  
> >
> >Great questions.  I would assume that a user specifying this option
> >intends the behavior here regardless of the device's support for
> >migration, which could be made more clear and easier to test by adding
> >this option to other, otherwise migratable, QEMU NICs.  
> 
> I agree and think it makes sense that if a user intentionally marks a
> device as a primary device of a failover pair then it should override
> the use of an existing migration region of the device.

Yes, that makes sense to me as well.