Re: Dom0 PAE panic when starting xend

[Christoph Egger <Christoph_Egger%gmx.de@localhost>]

On Tuesday 10 March 2009 17:40:31 Christoph Egger wrote:
> On Tuesday 10 March 2009 17:21:02 Manuel Bouyer wrote:
> > On Tue, Mar 10, 2009 at 05:14:03PM +0100, Christoph Egger wrote:
> > > > Good catch; x86_btop() needs to be fixed as it could break other
> > > > things. For the problem to show up you need to have 4G or more RAM
> > > > though (which might also explain why it doesn't cause problems for
> > > > some).
> > > >
> > > > The problem is that x86_btop() is used for both physical and virtual
> > > > addresses, which doesn't fit very well PAE. Maybe just casting to
> > > > paddr_t would be good enough ...
> > >
> > > My machine has exactly 4GB RAM.
> >
> > OK, so that's most likely it. With a PCI hole of 768M or more (that's
> > what I see on hardware I have) you have about 3/4G or ram above the 4GB
> > barrier. I'll also add that Xen likes to allocate memory
> > top-down, so in such a configuration dom0 is likely getting most of its
> > RAM (if not all) above the 4GB barrier.
>
> I just tried with casting to paddr_t. Unfortunately, this doesn't fix
> the mmu_update failure. xenevtmmap() runs twice before the
> panic happens.

xen_start_info.store_mfn has the value 0x10088a.

That's not large enough that casting to paddr_t has an effect.

Christoph