Ben Harris wrote:

> machines run).  I've alrady got the Technical Reference Manual, and I've
> managed to extract the data sheets for the most of the old
> ARM/MEMC/IOC/VIDC chipsets from ARM Ltd (once they'd found them).  I
> think I understand the machine pretty well, and I have half a clue what
> goes on in the depths of an operating system.  In short, I think I'm in a
> position to start porting NetBSD to this silly thing.  What now?

Another good source is a copy of the old VLSI reference manual on the
VIDC/IOMD/MEMC chipset.

> In case you're curious, these machines can't run NetBSD/arm32, because the
> ARM2 and ARM3 don't have 32-bit mode.  They're restricted to a 26-bit
> address bus, and have a slightly different register model.  The MMU in the
> older machines is different too, being a separate chip (or several) (the
> MEMC).

Some things are still there in common though, I think. Some of the VIDC
stuff is still similarish, the podule bus is very similar (minus some
changes) etc. Some of the bootloader stuff could move over too. So it
might be best to move with NetBSD/arm32 as a start and somehow
regurgitate stuff? Maybe putting MMU/page table/CPU ops through function
pointers instead of the current way (loads of ifdefs) might be a good
place to start, and work back from there.

> Anyway, can anyone tell me if anyone else is working on this, and if not,
> whether there's anything I should do before I get out my Emacs and start
> hacking?

Nobody else is working on it AFAIK; this was just some random ramblings.

	Neil

-- 
Neil A. Carson   Business development, Causality and Chalice Technology
Causality:       WWW: http://www.causality.com email: neil@causality.com
Chalice Tech:    WWW: http://www.chaltech.com  email: neil@chaltech.com
Office contact:  Tel/Fax: +44 (0)181 930 7408  Mobile: +44 (0)370 593183
SA110+1100, 21285/PCI, ARM, 3D, RTOSs, NetBSD, Image Processing, NC OSs