[vfio-users] Help Patching Qemu for Vendor

Alex Williamson alex.williamson at redhat.com
Sun May 8 22:08:59 UTC 2016


On Sun, 8 May 2016 16:02:54 -0500
Spam House <kthxplzdie at gmail.com> wrote:

> Thanks for the fast responses!
> 
> I must be trying to add it the wrong way.
> 
> sudo EDITOR=nano virsh edit win10
> 
> add:
>     <hyperv>
>       <relaxed state='on'/>
>       <vapic state='on'/>
>       <spinlocks state='on' retries='8191'/>
>       <vendor id='none'/>

The format is:

      <vendor_id state='on' value='KeenlyKVM'/>

As documented:

https://libvirt.org/formatdomain.html#elementsFeatures

>     </hyperv>
> 
> Saving results in:
> error: unsupported configuration: unsupported HyperV Enlightenment feature:
> vendor
> 
> Any suggestions? Thanks again!
> 
> 
> 
> On Sun, May 8, 2016 at 3:49 PM, Alex Williamson <alex.l.williamson at gmail.com
> > wrote:  
> 
> > On Sun, May 8, 2016 at 2:30 PM, Spam House <kthxplzdie at gmail.com> wrote:
> >  
> >> Hey All,
> >>
> >> Could anyone give me a quick summary of how to apply this (
> >> http://www.spinics.net/lists/kvm/msg121742.html) patch to Qemu on Arch
> >> Linux.
> >>
> >> I've cloned the Qemu git and have tried various ways of parsing the info
> >> in the link into other files and using 'ed' to try and append/update the
> >> files with no success.
> >>
> >> Example:
> >> git clone qemu
> >> cd qemu/target-i386
> >> nano cpu-qom.b (insert diff changes and save add 'w' at the very end)
> >> ed - cpu-qom.h < cpu-qom.b
> >>
> >> ed returns ? and deletes/blanks out cpu-qom.h
> >>
> >> Complete noob any help appreciated!
> >>
> >> I do have a windows10 vm with my 980ti passthrough running successfully.
> >> The performance has degraded since switching from Unraid to Arch and the
> >> only difference I can find is HyperV related. It seems this is the last
> >> hurtle.
> >>  
> >
> > This has been in QEMU since the 2.5 release, if you're cloning qemu.git,
> > the change is already there.  Surely Arch has QEMU builds including it as
> > well.
> >  




More information about the vfio-users mailing list