Path: utzoo!mnetor!uunet!husc6!think!bloom-beacon!mit-eddie!uw-beaver!ssc-vax!bcsaic!paula
From: paula@bcsaic.UUCP (Paul Allen)
Newsgroups: comp.os.minix
Subject: Re: Minix on the 80386, NSC32016
Message-ID: <4825@bcsaic.UUCP>
Date: 12 Apr 88 20:41:33 GMT
References: <749@devon.UUCP>
Reply-To: paula@bcsaic.UUCP (Paul Allen)
Followup-To: comp.os.minix
Organization: Boeing Computer Services AI Center, Seattle
Lines: 43
Keywords: i wish i was in the land of demand-paged virtual memory

In article <749@devon.UUCP> hermit@chessene.UUCP (Mark Buda) writes:
[...]
>
>And, how much work has anybody done in getting Minix to run on the 80386?
>I don't mean just run like and 8086/8, I mean like adding virtual memory,
>and such. Has anybody ported GCC to the 386 and gotten it to run under
>Minix? Would anybody like GCC ported to the 386 if nobody's done it?

Included in my long-range plans is the acquisition of a 386 box to
replace my aging Z-100.  The idea of porting the GNU c compiler to run
on a 386 under minix is really attractive.  From the 15 minutes I spent
perusing the gcc sources, it looks like it would be at least possible to
add a 386 code generator.  The machine details needed by the code
generator appear to all be contained in a header file.  (Note that I
have never written a compiler, and I haven't written any 386 code!)

Assuming that one has stock 8086 minix running on a 386 box, the steps that 
I see are something like this:

1 - Fix minix so it can handle an a.out with more than 64K of code and
	64K of data.  I've no idea how hard this would be.

2 - Add 386 code-generation capability to gcc using some reasonable
	machine like a Sun or a VAX (or your MesSDOS 386!) as a
	development environment.

3 - Move the compiler executable to your 386 minix environment for testing.
	Compile the compiler with itself under minix.

4 - Repeat steps 2 and 3 as needed.  :-)


So, I'll ask the question too.  Is anybody working on getting minix to
take full advantage of the 386 and/or porting gcc to the 386/minix?  I
would love to work on something like this, but realistically, I won't
have time to do it by myself.  I might be game for some sort of
collaboration, however.

Paul Allen
-- 
------------------------------------------------------------------------
Paul L. Allen                       | paula@boeing.com
Boeing Advanced Technology Center   | ...!uw-beaver!ssc-vax!bcsaic!paula