On Thu, Oct 29, 1998 at 10:54:38PM +0100, Stefan Grefen wrote:
> In message <19981029111246.E6028@cs.uni-bonn.de>  Ignatios Souvatzis wrote:
> > On Thu, Oct 29, 1998 at 09:50:43AM +0100, Stefan Grefen wrote:
> > > 
> > > It think it depends which functionality he wants to port.
> > > The process stuff will be the worst (at least on non-segmented CPU's),
> > > and fork() will be a royal pain (at expensive like hell). 
> > 
> > Yeah. But OTOH, if your compiler always produces PC-relative code for the
> > shared code segment, and data accesses relative to either the stack or an
> > static data area pointer register, its easy.
> > 
> > OS/9 level 1 on m6809 worked that way.
> > OS/9 on 68000, too, I think.
> 
> Yes it did (on my Atari :-))).
> But it was 'pseudo'-segmented as there was this 64k limit (all data
> was relative to a a register with max. 16bit offset.

Uhm, but nowadays you would use an 68020 at least, with 32bit relative offsets
if you need them.
	-is