Re: [Qemu-devel] [PATCH v5 05/14] vl: handle "-device dimm"

[Hu Tao]

On Sun, Jul 14, 2013 at 06:58:23PM +0200, Paolo Bonzini wrote:
> Il 12/07/2013 04:39, Hu Tao ha scritto:
> > > without necessity to specify to be hot-plugged DIMMs at startup, hot-plug
> > > could become more flexible since arbitrarily sized DIMMs with required 
> > > NUMA
> > > mapping could be specified during hot-plug time, for example:
> > > 
> > >   device_add dimm,id=dimm1,bus=membus.0,size=3G,node=1,...
> > 
> > Seems impossible as ACPI memory devices' ranges must be specified at 
> > startup,
> > but I'll be glad if I'm wrong.
> 
> Yeah, I also understood the problem to be the SRAT, not the DSDT...  Of
> course you could reserve a single large part of the address range in
> "-numa mem", and then add one or more DIMMs covering only part of the
> address range.

still needs to specify -dimm at startup, right?

> 
> Perhaps "populated" could be a size instead of a boolean, to do
> something like what Igor suggested.

seems odd that a memory device is partially populated. How can guest OS
and firmware cope with that?

> 
> Paolo