On Wed, Oct 12, 2022 at 12:34:48PM -0400, Eric Auger wrote:
> In theory the virtio-iommu-pci could be plugged anywhere in the PCIe
> topology and as long as the dt/acpi info are properly built this should
> work. However at the moment we fail to do that because the
> virtio-iommu-pci BDF is not computed at plug time and in that case
> vms->virtio_iommu_bdf gets an incorrect value.
>
> For instance if the virtio-iommu-pci is plugged onto a pcie root port
> and the virtio-iommu protects a virtio-block-pci device the guest does
> not boot.
>
> So let's do not pretend we do support this case and fail the initialize()
> if we detect the virtio-iommu-pci is plugged anywhere else than on the
> root bus. Anyway this ability is not needed.
>
> Signed-off-by: Eric Auger <eric.au...@redhat.com>
> ---
>  hw/virtio/virtio-iommu-pci.c | 12 ++++++++++--
>  1 file changed, 10 insertions(+), 2 deletions(-)

FYI libvirt will already reject any attempts to place the device
anywhere but directly on pcie.0, so from our point of view merging
this patch is perfectly fine.

-- 
Andrea Bolognani / Red Hat / Virtualization


Reply via email to