[ On Tuesday, February 8, 2000 at 13:42:12 (-0500), Thor Lancelot Simon wrote: ]
> Subject: Re: m88k system call convention
>
> I'm 99% sure Motorola ported their SVR4 to mvme88k, and if we're going
> to go ELF I strongly suggest we use whichever calling convention they
> did.

Yes, they did.

> One would hope that Data General's "SVR4" (DG/UX) also did the same
> thing Motorola did, but gosh, why do I suspect we have one of those
> cases here where one "SVR4" for a given architecture can't run the
> other "SVR4"'s binaries? :-)

I'm 90% sure they followed the spec. and that you're wrong in this case.

I just happen to have right here in front of me (well it was on the
floor behind me until a second ago :-) a copy of the 88open Consortium
"Object Compatibility Standard" (release 1.1, April 1990); and a copy of
the "Binary Compatibility Standard" (also r1.1, Apr.1990); as well as
the "Networking Supplements" and "X Window System Supplements" for each;
and some various errata sheets.  The whole pile (2" of 8.5x11 paper) is
free to anyone working on such compatability for NetBSD if you help pay
for the shipping to your location (or pick it up here in Toronto).  I
don't have, or ever expect to have, any 88k machines except my NCD
xterms or that I'd have the time or need to work on such compatability
if I did run one.

(I was wondering why this particular binder wasn't in the as-yet
un-opened and still in the garage boxes of other binders!  ;-)

The technical chair of these standards was a Motorola person and I'm
certain that DG was a member of 88open too.  In fact I think I even got
these documents from someone at DG.  Looks like they were officially
sanctioned by POSIX too as Hal Jespersen was the editor.

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <gwoods@acm.org>      <robohack!woods>
Planix, Inc. <woods@planix.com>; Secrets of the Weird <woods@weird.com>