[libvirt] RFC: New CPU hot(un)plug API and XML

[Daniel P. Berrange]

On Thu, Jun 16, 2016 at 09:14:12AM +0200, Peter Krempa wrote:
> On Wed, Jun 15, 2016 at 18:10:01 +0100, Daniel Berrange wrote:
> > On Mon, Jun 13, 2016 at 02:48:51PM +0200, Peter Krempa wrote:
> 
> TL;DR: I mostly agree with your suggestions altough there's a slight
> problem with the suggested XML format. See [1] at the bottom
> 
> > > <domain>
> > >   [...]
> > >   <vcpu current='1'>3</vcpu>
> > >   <vcpus>
> > >     <vcpu id='0' state='enabled'/> <-- option 1, no extra data
> > >     <vcpu id='1' state='disabled' cell='1'/> <--- option 2, just numa node,
> > >                                                   since it's non-obvious
> > >     <vcpu id='2' state='disabled' cell='1' pin='1-2' scheduler='...'/>
> > >      <!-- option 3 all the data duplicated -->
> > >   </vcpus>
> > >   [...]
> > >   <cputune>
> > >     <cpupin ... />
> > >   </cputune>
> > >   [...]
> > >   <cpu>
> > >     <numa>
> > >       <cell id='0' cpus='0' memory='102400' unit='KiB/>
> > >       <cell id='1' cpus='1-2' memory='102400' unit='KiB/>
> > >     </numa>
> > 
> > In the 'virsh capabilities' XML, the CPUs are listed underneath the
> > <cell>. So rather than adding <vcpu> as children of <vcpus>, I think
> > perhaps we should just put then under the <cell> here too. That would
> > avoid the need to add a cell=NN attribute.eg
> > 
> >    <cpu>
> >      <numa>
> >        <cell id='0' cpus='0' memory='102400' unit='KiB>
> >          <cpus>
> >            <cpu id="0" state="enabled"/>
> > 	 </cpus>
> >        </cell>
> >        <cell id='1' cpus='1-2' memory='102400' unit='KiB>
> >          <cpus>
> >            <cpu id="1" state="disabled"/>
> >            <cpu id="2" state="enabled"/>
> >          </cpus>
> >        </cell>
> >      </numa>
> 
> [1]
> 
> I was thinking doing the same at first but there's one caveat. For
> non-numa guests you don't get the <numa> element at all. In such case
> the online/offline state would need to be kept somewhere else and I
> wanted to avoid having two places where to put the data.
> 
> I've opted for adding a new element which would satisfy even non-numa
> guests at the expense that you'd need to collate from two places in case
> of numa guests.

Hmm, yes, I forgot about that, oh well.

Regards,
Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|