[vfio-users] Processor Root Ports appeared in iommu_groups - why?

[Alex Williamson]

On Mon, Oct 19, 2015 at 5:45 PM, Blank Field <ihatethisfield at gmail.com>
wrote:

> I've migrated through the whole 4.1 kernel branch and currently i am using
> 4.2.3-200.fc22-x86_64...
> And i don't remember my pci-e root ports being in the same group as my
> GPUs.
>
> "qemu-system-x86_64: -device
> vfio-pci,host=01:00.0,id=hostdev0,bus=pci.0,addr=0x3,rombar=1,romfile=/mnt/hdd/qemu/hybridmagic.rom:
> vfio: error, group 1 is not viable, please ensure all devices within the
> iommu_group are bound to their vfio bus driver."
> (and the same for the second GPU)
> which leads us to:
> ./1/devices:
> 0000:00:02.0  0000:01:00.0  0000:01:00.1
> and
> ./2/devices:
> 0000:00:04.0  0000:02:00.0  0000:02:00.1
> where 01:00.0 and 02:00.0 are GPUs, and 04.0 and 02.0 are Processor Root
> Ports.
>
> ...should i bind them too?
>

No, only endpoints can be bound to vfio-pci.  Is pcieport the driver for
them, if so, leave them alone.


> What was changed and why?
>

Dunno, first I've heard of it.


> This is an AMD FM2 system, i've thought ACS problems are specific to intel
> systems.
>

Um, no.  Send `sudo lspci -vvv` and `find /sys/kernel/iommu_groups`
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/vfio-users/attachments/20151019/f53c24cf/attachment.htm>