On Thu, Jun 29, 2017 at 10:39:26 +0200, Marek Marczykowski-Górecki wrote:
> On Thu, Jun 29, 2017 at 10:04:42AM +0200, Jiri Denemark wrote:
> > On Thu, Jun 29, 2017 at 03:11:42 +0200, Marek Marczykowski-Górecki wrote:
> > > Set CPU features in appropriate libxl structure.
> > > Use old "xend" syntax, because it allow control over any bit and libvirt
> > > already have API for translating features to appropriate cpuid bits. And
> > > also features naming in libxl do not match the one of libvirt in
> > > multiple cases.
> >
> > How does the new syntax look like?
>
> "host,tm=0,sse3=0"
Is "host" fixed there or are other CPU models supported too? If only
"host" is allowed, libxl driver should support host-passthrough mode
only and it would be pretty easy to map it to the libxl's syntax. The
following CPU XML
<cpu mode='host-passthrough'>
<feature name='avx' policy='disable'/>
<feature name='aes' policy='require'/>
</cpu>
could be directly translated into "host,avx=0,aes=1".
> > And would it be actually better?
>
> Maybe? But the new syntax use different names for many features (like
> pni vs sse3, or sep vs sysenter), so it would require some ugly
> translation table...
I don't a translation table would be ugly.
> What I would really like here, is to get list of bits specified with
> <feature ...>, _without_ those defined by <model>. And maybe even
> without requiring <model> to be set at all.
That's what host-passthrough mode is for.
Jirka
--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list
