On Wed, 7 Nov 2007 00:39:48 +0000
David Laight <david@l8s.co.uk> wrote:

> On Tue, Nov 06, 2007 at 06:39:38PM -0500, Greg Troxel wrote:
> > "Steven M. Bellovin" <smb@cs.columbia.edu> writes:
> > 
> > > Some time back, a change was made (by Greg Troxel, I thought) so
> > > that rbus was calculated automatically.  It doesn't seem to
> > > happen on amd64.  I have 3GB machine; here's what happened:
> > 
> > That's what I'd expect - I put the heuristic in the i386 machdep
> > code. If you want to add it for amd64, or a different one that
> > matches those platforms, that sounds sensible.  We could also have
> > a common ruotine somewhere, since i386 and amd64 seem to share a
> > lot.
> > 
> > But, with amd64 having more address space, the whole notion of
> > using 1G or 2G as the rbus place seems odd.
> 
> It may have more address space, but there is still a physical address
> boundary at 4GB.
> 
That's why I didn't just copy the code over -- I know I don't know the
architectural details.  (Btw, the i386 version isn't working right,
either, but I haven't experimented with it at much.  The comments note
that it isn't tested....  I'll try to get more details on that in a
few days.)


		--Steve Bellovin, http://www.cs.columbia.edu/~smb