Re: [PATCH RFC 08/12] vfio/migration: Fix incorrect reporting for VFIO pending data

[Peter Xu <peterx@redhat.com>]

On Mon, Apr 06, 2026 at 03:34:20PM +0300, Avihai Horon wrote:
> 
> On 4/2/2026 6:55 PM, Peter Xu wrote:
> > External email: Use caution opening links or attachments
> > 
> > 
> > On Thu, Apr 02, 2026 at 11:28:09AM -0400, Peter Xu wrote:
> > > > Now that migration->stopcopy_size holds only the stopcopy size (and not
> > > > stopcopy+precopy), we should remove these lines from
> > > > vfio_update_estimated_pending_data():
> > > > 
> > > >      /* The total size remaining requires separate accounting */
> > > >      migration->stopcopy_size -= data_size;
> > > Good point, I missed this part when changing the definition, I'll fix.
> > Instead of changing the definition, I found that it might be cleaner to
> > keep the definition of stopcopy cache variable for VFIO (considering the
> > other invokation during save setup), then just do one more step for math.
> > 
> > The new version should look like this:
> 
> Yes, that's better. Just small comment below.
> 
> > 
> > ===8<===
> > 
> > diff --git a/hw/vfio/migration.c b/hw/vfio/migration.c
> > index 655ff7712a..5aff8de844 100644
> > --- a/hw/vfio/migration.c
> > +++ b/hw/vfio/migration.c
> > @@ -585,19 +585,27 @@ static void vfio_state_pending(void *opaque, MigPendingData *pending)
> >   {
> >       VFIODevice *vbasedev = opaque;
> >       VFIOMigration *migration = vbasedev->migration;
> > -    uint64_t remain;
> > +    uint64_t precopy_size, stopcopy_size;
> > 
> >       if (pending->exact) {
> >           vfio_state_pending_sync(vbasedev);
> > -        remain = migration->stopcopy_size;
> >       } else {
> >           if (!vfio_device_state_is_precopy(vbasedev)) {
> >               return;
> 
> Now in the estimate path we shouldn't bail early if not in precopy state, so
> estimate/exact reports are consistent also for VFIO devices that don't
> support precopy.

Will do.

-- 
Peter Xu