> Does anybody recall the "kernel size" problem? I was getting problems with
> this a long time ago. Perhaps now, the "voodoo zone" has entered the SCSI
> driver code. In that case, it's definitely a pmap problem.

Not necessarily.  It could just be a stray pointer of some description.
This doesn't quite seem like the same problem, but it is similar in that
it's a flaky problem only affecting some people (although this seems to
be more universal than the "kernel size problem").

> Perhaps uvm is excercising the machdep memory code in a way it isn't expecting?

Scott says that he's been over the pmap code with a fine-tooth comb and
verified that it seems to be the same as the other hp300-derived ports.
It's possible that there's still a bug in there, but (as far as I know)
it's not affecting anyone but mac68k.  Similarly, there could be more
machine-independent bug that we're tripping on that no one else is, for
whatever reason.

-allen

-- 
                    Allen Briggs - briggs@ninthwonder.com
       Try free *nix: http://www.netbsd.org/, http://www.freebsd.org/,
		      http://www.linux.org/, http://www.openbsd.org/