Subject: cross-compiling on i386 for m68k
To: None <m68k@netbsd.org>
From: Thomas J. Trebisky <tom@as.arizona.edu>
List: m68k
Date: 11/15/1994 10:21:02
Hello,

I have pretty much finished up getting NetBSD 1.0 installed on a
i386 box I have recently acquired. Now I am getting down to
business with what I really want to do, namely use this
as a development system for a m68k port.  The target system is
a motorola vm04 (a versabus monster, the heart of which is
a 68030 running at 20Mhz, but I spare you all the details).

At any rate, I am posting to ask for hints and help in getting
a cross-development platform set up, expecting and hoping that
some of you have covered this ground already and can help me and
save me a lot of time and effort.  Here is what I see as the
task before me, please correct me where you see me going astray.

1) Buid gcc as a cross compiler.  Apparently I have 2.4.5 on my
system, and I can go into the source directory and config this as
a cross-compiler.  But I noticed yesterday in the 2.6.1 announcement
that NetBSD-m68k is a newly supported configuration, perhaps this
is the way to go.  Either way, I expect this step to be easy.

2) Build the binutils for cross-compile use.  From prior experience
I expect gas to be pretty easy.  gnu-ld may or may not be easy
depending on progress made with the binutils since my experience
with them a year or so ago when I was cross-compiling from a sparc.

3) Now I want to think about making boot floppies.  This may or may
not be easy, but I expect due to little/big endian problems that
I will have to set up a special mkfs to build an exportable filesystem.
I expect the bulk of this work to consist of inserting htonl() and
htons() macros thru the mkfs source code, but maybe there is a better
way.

4) If the above scenario is true, as I believe it is, I will also
need some kind of utility to load files into a new filesystem, since
I won't just be able to mount and copy.  The old BSD 4.3 utility
mkproto would be just the thing, and perhaps I will need to reverse
engineer this for my purposes.

Once I get to this point I should be off to the races, ... at least for now,
any helpful comments or discussion will be appreciated!

	Tom
-- 
	Tom Trebisky			Steward Observatory
	ttrebisky@as.arizona.edu	University of Arizona
	(602) 621-5135			Tucson, Arizona 85721