[vfio-users] cpu core pinning with multiple cpus

[Blank Field]

No libvirt support means there is only a dirty hack available: passing
arbitrary qemu:arg qemu commandline options. Everything else stays the same.
On Oct 2, 2015 7:27 PM, "Eddie Yen" <missile0407 at gmail.com> wrote:

> Thanks for reply!
> So the suggestion is keep both commands in the XML file, and create the
> script that make XML forwarding to that script.
>
> As libvirt not support yet, I can launch VM by using virsh only, or I can
> launch it on virt-manager either?
>
> 2015-10-03 0:17 GMT+08:00 Alex Williamson <alex.l.williamson at gmail.com>:
>
>> On Fri, Oct 2, 2015 at 10:12 AM, Eddie Yen <missile0407 at gmail.com> wrote:
>>
>>> Thanks for sharing!
>>>
>>> I noticed that it had kvm=off and hv_vendor inside the value.
>>> If I got kvm=off, can I also delete this?
>>>
>>> <kvm>
>>>    <hidden state='on'/>
>>> </kvm>
>>>
>>
>> It's redundant, but the <qemu:args> variant is just temporary until the
>> patch is accepted and we can get libvirt support.  I think the way that
>> I'll choose to enable on my system until then is via a wrapper script.  For
>> instance, create a script with something like:
>>
>> #!/bin/sh
>> exec /usr/local/bin/qemu-system-x86_64 \
>> `echo "\$@" | sed 's|hv_time|hv_time,hv_vendor_id=|g'`
>>
>> Then change the <emulator> tag to point to this script.
>>
>>
>>> And, hv_vendor=<GPU vendor> or <CPU vendor> ?
>>>
>>
>> Free form string or blank, 12 characters.  As the patch indicates, the
>> value is only used for reporting and diagnostic purposes.
>>
>
>
> _______________________________________________
> vfio-users mailing list
> vfio-users at redhat.com
> https://www.redhat.com/mailman/listinfo/vfio-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/vfio-users/attachments/20151002/370291cb/attachment.htm>