Re: virtio_balloon regression in 5.19-rc3

[Jason Wang <jasowang@redhat.com>]

On Fri, Jun 24, 2022 at 5:14 PM Michael S. Tsirkin <mst@redhat.com> wrote:
>
> On Tue, Jun 21, 2022 at 06:10:00PM +0200, Ben Hutchings wrote:
> > On Tue, 2022-06-21 at 17:34 +0800, Jason Wang wrote:
> > > On Tue, Jun 21, 2022 at 5:24 PM David Hildenbrand <david@redhat.com> wrote:
> > > >
> > > > On 20.06.22 20:49, Ben Hutchings wrote:
> > > > > I've tested a 5.19-rc3 kernel on top of QEMU/KVM with machine type
> > > > > pc-q35-5.2.  It has a virtio balloon device defined in libvirt as:
> > > > >
> > > > >     <memballoon model="virtio">
> > > > >       <address type="pci" domain="0x0000" bus="0x05" slot="0x00" function="0x0"/>
> > > > >     </memballoon>
> > > > >
> > > > > but the virtio_balloon driver fails to bind to it:
> > > > >
> > > > >     virtio_balloon virtio4: init_vqs: add stat_vq failed
> > > > >     virtio_balloon: probe of virtio4 failed with error -5
> > > > >
> > > >
> > > > Hmm, I don't see any recent changes to drivers/virtio/virtio_balloon.c
> > > >
> > > > virtqueue_add_outbuf() fails with -EIO if I'm not wrong. That's the
> > > > first call of virtqueue_add_outbuf() when virtio_balloon initializes.
> > > >
> > > >
> > > > Maybe something in generic virtio code changed?
> > >
> > > Yes, we introduced the IRQ hardening. That could be the root cause and
> > > we've received lots of reports so we decide to disable it by default.
> > >
> > > Ben, could you please try this patch: (and make sure
> > > CONFIG_VIRTIO_HARDEN_NOTIFICATION is not set)
> > >
> > > https://lore.kernel.org/lkml/20220620024158.2505-1-jasowang@redhat.com/T/
> >
> > Yes, that patch fixes the regression for me.
> >
> > Ben.
>
>
> Jason are you going to fix balloon to call device_ready before
> registering device with linux?

I'm working on the fix, (spot various bugs during review).

Thanks

> > --
> > Ben Hutchings
> > Any smoothly functioning technology is indistinguishable
> > from a rigged demo.
>
>