> From: "Chris G. Demetriou" <cgd@postgres.berkeley.edu>
> To: mw@eunet.ch
> Cc: glass@sun-lamp.cs.berkeley.edu (Adam Glass), m68k@sun-lamp.cs.berkeley.edu
> Subject: Re: 8k 
> Date: Mon, 14 Feb 94 16:13:05 -0800

> > While I have your attention, m68k programmers.. I saw that in some
> > spots (very few actually) there are now "#ifdef m68k". I guess this is
> > a good idea, even a necessary step. BUT I'd strongly vote for using
> > "#ifdef mc68000" instead. [ ... ]
> 
> we currently define "-Dm68k -Dmc68000 -Dmc68020", so "external"
> sources should have no problems.  This was the way it was originally
> for the hp300, in gcc1.* from UCB.
> 
> mc68000 is a CPU.  m68k is a family of CPUs.  In our case, "m68k"
> actually is a bit more generic than the CPU: it's the set of machines
> which use stuff from /sys/arch/m68k.
> 

I think a lot of this is historical.  Some of the original 68k kernel (and
other) sources came from MIT in the early 80s.  They used "mc68000" as I
recall.  When the 68020 came out, "mc68020" was added.  Don't know where
"m68k" came from.  So all really refered to CPUs.

Me?  I ifdef'ed 68k code under "hp300" :-)

------------------------------------------------------------------------------