[vfio-users] kernel v4.2 VM failing with passthrough

[globalgorrilla at fastmail.fm]

Same with rc7. Presumably you have 4.2 working with passthrough of GPUs 
without changing anything?

On 13 Aug 2015, at 16:11, globalgorrilla at fastmail.fm wrote:

> Nope. Just starts, and stops immediately:
>
> 2015-08-13 23:05:33.068+0000: starting up
> [snip qemu command which is same as before]
> 2015-08-13 23:05:37.071+0000: shutting down
>
> On 08/13/2015 12:26 PM, Alex Williamson wrote:
>> On Thu, 2015-08-13 at 12:22 -0700, globalgorrilla at fastmail.fm wrote:
>>> Hi,
>>>
>>> VMs are terminating shortly after launch when passing through any 
>>> Quadro
>>> K2200 or R9 260X on 4.2.0-rc6, using 4.1.4 it's working perfectly.
>>>
>>> The VM will go to running and then terminate without ever seeing the
>>> OVMF boot screen. I'm not seeing anything in dmesg output.
>>>
>>> It appears specific to the GPUs it seems as the I350-T4 VFs work 
>>> fine on
>>> a separate VM that uses only them in passthrough.
>>>
>>> I'm using libvirt 1.2.1, qemu 2.4, and OVMF.fd from edk2 (5c02fc) 
>>> Aug
>>> 10.
>>>
>>> /etc/default/grub:
>>> intel_iommu=on iommu=pt iommu=1 pci=assign-busses
>>>
>>> /etc/modprobe.d/vfio-pci.conf
>>> options vfio-pci
>>> ids=1002:ffffffff:ffffffff:ffffffff:00030000:ffff00ff,1002:ffffffff:ffffffff:ffffffff:00040300:ffffffff,10de:ffffffff:ffffffff:ffffffff:00030000:ffff00ff,10de:ffffffff:ffffffff:ffffffff:00040300:ffffffff
>>>
>>> How is 4.2 working for others?
>>>
>>> I saw there was some refactoring in mtrr.c that got fixed in rc6.
>>>
>>> Anything changing that might be related to the issue I'm seeing? IRQ
>>> handling?
>>>
>>
>> Do you see anything in the libvirt
>> log, /var/log/libvirt/qemu/$DOMAIN.log?
>>