Re: [PATCH RFC 1/2] linux/vnuma: vnuma support for pv guest

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

On Thu, Aug 29, 2013 at 01:12:15AM +0200, Dario Faggioli wrote:
> On mar, 2013-08-27 at 10:33 -0400, Konrad Rzeszutek Wilk wrote:
> > Unfortunatly that won't happen if you boot this under dom0. It will
> > find on some AMD machines the AMD Northbridge and try to scan that.
> > And blow up.
> > 
> Well, perhaps this was not evident enough in Elena's mails and patches,
> but her current target is actually DomU. :-P
> 
> However, that's only because we're proceeding in steps, so yes, once
> DomU will be working, Dom0 will become the target, and we'll have to
> start caring about not making things explode! :-O
> 
> > If you look at the git commit that did the 'numa = 0' you will see
> > the backstory.
> > 
> That's something interesting to look at... I remember having seen that
> discussion but not the details. I'll check when back from vacation.
> 
> > I think part of enablement of ' numa = 1' is to wrap it with
> > 
> > 	if (xen_initial_domain() && xen_supports_this_hypercall())
> > 		numa = 1;
> > 
> > in the #2 patch.
> > 
> Mmm... Did you mean '!xen_initial_domain()' ? Or is it me that did not
> understand your comment about Dom0 above?

For right now '!' would suffice.
> 
> In case you did mean '!', again, this would be fine for now, since the
> target is DomUs. At the point when we will start chasing Dom0, it would
> be interesting to find a solution to this issue... :-/


> 
> Anyway, thanks for pointing this out.
> 
> Regards,
> Dario
> 
> -- 
> <<This happens because I choose it to happen!>> (Raistlin Majere)
> -----------------------------------------------------------------
> Dario Faggioli, Ph.D, http://about.me/dario.faggioli
> Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)
>