Re: [PULL 00/21] vfio queue

[Cédric Le Goater]

On 10/6/23 12:33, Cédric Le Goater wrote:
> On 10/6/23 08:19, Cédric Le Goater wrote:
> > The following changes since commit 2f3913f4b2ad74baeb5a6f1d36efbd9ecdf1057d:
> >
> >    Merge tag 'for_upstream' of https://git.kernel.org/pub/scm/virt/kvm/mst/qemu 
> > into staging (2023-10-05 09:01:01 -0400)
> >
> > are available in the Git repository at:
> >
> >    https://github.com/legoater/qemu/ tags/pull-vfio-20231006
> >
> > for you to fetch changes up to 6e86aaef9ac57066aa923211a164df95b7b3cdf7:
> >
> >    vfio/common: Move legacy VFIO backend code into separate container.c 
> > (2023-10-05 22:04:52 +0200)
> >
> > ----------------------------------------------------------------
> > vfio queue:
> >
> > * Fix for VFIO display when using Intel vGPUs
> > * Support for dynamic MSI-X
> > * Preliminary work for IOMMUFD support
>
> Stefan,
>
> I just did some tests on z with passthough devices (PCI and AP) and
> the series is not bisectable. QEMU crashes at patch  :
>
>    "vfio/pci: Introduce vfio_[attach/detach]_device".
>
> Also, with everything applied, the guest fails to start with :
>
>   vfio: IRQ 0 not available (number of irqs 0)
>
> So, please hold on and sorry for the noise. I will start digging
> on my side.

We found a couple of issues after further testing. Fixes are under way.
Let's drop this PR. I will send a reduced one today.

Thanks,

C.