on 8/14/00 12:19 PM, David A. Gatwood at dgatwood@deepspace.mklinux.org
wrote something like:

>> 1.) as you mentioned, the 601 CPU support has not been added. I don't recall
>> exactly why it hasn't been added, the info is out there, just no one has
>> done it yet. [or taken it from MkLinux/LinuxPPC]
> 
> I'm assuming the big problem here are the BATs.  That and the clock stuff
> if you're using it.  Can't think of anything else really major different
> between the 601 and the rest of the 6xx family, but I'm probably missing
> something.  Unless I'm missing something pretty big, though, it doesn't
> seem like that much work.  So... what am I missing?  ;-)

    I just checked /sys/arch/powerpc/include/bat.h (I think), and the 601
specific BAT's appear to already be there. If I didn't know any better
(which I really don't) I'd think that a current kernel would at least load
and partially boot on a 601 if we had a boot loader for it.

    Last I remember, the hang up with BootX is that a mechanism had to be
worked out for passing down a device tree into the NetBSD kernel. (Which
would likely be a non-OFW ppc solution. But couldn't one just hardcode
enough information into the kernel so that when (OFW Version==NULL) or some
such, we assume a NuBus based system and default ROM locations, etc.; enough
to pull the Gestault ID and go from there? We're not talking about a whole
lot of info to add, it's just the 61xx, 71xx, 81xx and a few variations of
those models that we're talking about. And with any luck, they'll probably
use most of the same parameters.

    It would be nice though, to get an MI NuBus framework going, though not
necessary at this point. That way 68k and PPC could get NuBus things working
with minimum fuss. I also think that the Data General AViiON [Motorola 88k
based] systems use NuBus, so that would be a kick start for the 88k port as
well (though they may be proprietary like everything else in them)


    Wow that was longer than I wanted. Oh well.


    Chris