On Tue, May 05, 2015 at 11:47:30AM +1000, David Gibson wrote: > On Fri, Apr 24, 2015 at 12:17:29PM +0530, Bharata B Rao wrote: > > From: Andreas Färber <afaer...@suse.de> > > > > Signed-off-by: Andreas Färber <afaer...@suse.de> > > Signed-off-by: Bharata B Rao <bhar...@linux.vnet.ibm.com> > > So, how to organize this generically is still under discussion. For > now, I don't think this generic outline is really worth it. In any > case I can't really take it through my tree. > > What I'd suggest instead is just implementing the POWER core device in > the ppc specific code. As the generic socket vs. core vs. whatever > stuff clarifies, that POWER core device might become a "virtual > socket" or CM or whatever, but I think we'll be able to keep the > external interface compatible with the right use of aliases. > > In the meantime it should at least give us a draft we can experiment > with on Power without requiring new generic infrastructure.
Makes sense, I will switch to the semantics that I had in v1 where I enabled CPU hotplug for POWER8 using device_add. (qemu) device_add POWER8-powerpc64-cpu,id=XXX Regards, Bharata.
