Re: [PATCH] memory: use signed arithmetic

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Wed, 2011-08-03 at 11:26 +0300, Avi Kivity wrote:
> On 08/03/2011 01:15 AM, Richard Henderson wrote:
> > On 08/02/2011 03:06 PM, Avi Kivity wrote:
> > >  I don't think there's any cpu which has a real 64-bit physical
> > >  address space? Don't they all truncate it?
> >
> > I don't know.  You're right that x86_64 does, at 48 bits.
> > The alpha system I'm trying to emulate does, at 50 bits.
> >
> > I guess if IBM agrees wrt p-series and z-series emulation, then
> > I'd be ok, so long as you add a comment above that structure that
> > says no existing hw implementation actually uses 63 address bits.
> 
> Ben, can you confirm that pseries physical addresses are 63 bits wide or 
> smaller?
> 
> IIRC zseries has no mmio, and Zettabyte machines are still rare.

We are fine yes. We might grow to 50 bits but I doubt we'll ever do the
full 64, especially since we have some hard-wired assumptions that in
real mode (MMU off) the top 2 bits are ignored.

Cheers,
Ben.