Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!rutgers!mit-eddie!uw-beaver!ssc-vax!bcsaic!paula
From: paula@bcsaic.UUCP (Paul Allen)
Newsgroups: comp.os.minix
Subject: Re: RE**N:  Future of Minix
Message-ID: <13267@bcsaic.UUCP>
Date: 20 Jul 89 21:39:44 GMT
References: <20059@louie.udel.EDU>
Reply-To: paula@bcsaic.UUCP (Paul Allen)
Organization: Boeing Computer Services AI Center, Seattle
Lines: 40

In article <20059@louie.udel.EDU> HELMER%SDNET.BITNET@vm1.nodak.edu (Guy Helmer) writes:
[Guy quoted John Nall asking if there's any interest in Minix
386 development.]
>
>Definitely!!!  One thing I plan on starting if no one
>else has is 'as' and 'ld' for the 80386 under Minix.  I just finally got
>Intel's 80386 Programmer's Reference Manual, so I've started looking at
>the somewhat warped instruction set.

Let's not forget about the GNU stuff: gcc, gas, ld, etc.  I've got
a copy of gas configured to assemble 386 code on my Sun.  Here's
how big it is:

stehekin% size a386
text    data    bss     dec     hex
73728   40960   36400   151088  24e30

This is a tad big for the Minix compiler, but might provide a
workable starting point.  The gnu loader looks like this:

stehekin% size ld
text    data    bss     dec     hex
40960   8192    11504   60656   ecf0

This might be useable under Minix unless it assumes the presence
of huge amounts of memory at runtime.  The GNU C compiler is
*huge* and uses a meg or two at runtime.  All the the GNU stuff
assumes 32-bit ints.  Due to the extreme limits Minix imposes,
cross-developing a 32-bit Minix on some more reasonable machine 
may be the only way to go.

Has anybody tried cross-compiling Minix code using one of the
DOS-based 386 environments?  They're all quite expensive, but
they are considerably less than buying a Sun!  :-)

Paul
-- 
------------------------------------------------------------------------
Paul L. Allen                       | pallen@atc.boeing.com
Boeing Advanced Technology Center   | ...!uw-beaver!bcsaic!pallen