Subject: Re: NetBSD on the BeBox?
To: Rob Healey <rhealey@altair.helios.mn.org>
From: Jason Thorpe <thorpej@nas.nasa.gov>
List: netbsd-ports
Date: 12/15/1995 14:45:42
On Fri, 15 Dec 1995 09:14:44 -0600 (CST) 
 rhealey@altair.helios.mn.org (Rob Healey) wrote:

 > 	Does it matter that much at this point? At least we'll FINALLY
 > 	get a PPC port rolling. It probably will have enough in common,
 > 	like the m68k ports, that CHRP, MAC and Power Amiga ports would
 > 	find it to be VERY useful!

We should be careful when embarking on a ppc porting effort to _name_ 
things correctly.  The m68k stuff is done pretty well (even though 
there's a lot of code that could still be shared).

For powerpc ports, I'd suggest a naming scheme like the following:

	sys/arch/ppc		Generic powerpc code, a'la sys/arch/m68k
	sys/arch/be		A "Be" port, uses ppc code
	sys/arch/mvmeppc	Motorola MVME ppc cards
	sys/arch/macppc		Power macs

	... etc ...

It's arguable that current ports should have been named better:

	sys/arch/i386 -> sys/arch/i386pc

One day we might support things like Intel Multibus-II systems, Sequents, 
etc.  (Hey!  It could happen!  :-)

Same with sparc, alpha, pmax:

	sys/arch/sparc -> sys/arch/sunsparc
	sys/arch/alpha -> sys/arch/decalpha
	sys/arch/pmax -> sys/arch/decmips

Of course, changing the names of several ports probably isn't very likely 
to happen, but at least we can learn from the experience :-)

This also suggests that we might want to consider a change in how obj* 
directories are handled.  For example, one day there may actually be a 
common m68k binary distribution.  For that to happen, the hp300 needs to 
change MID numbers (and link page size), and a couple of key programs and 
libraries (mostly, a couple of things in sbin and libkvm) need to be 
worked on.  But, when that day comes, we should be building into 
obj.m68k, rather than obj.sun3, obj.amiga, obj.mvme68k, etc.  This might 
be as simple as making a new set of make environment variables, like 
USROBJMACHINEARCH instead of USROBJMACHINE.  But, then again, I'm not the 
make(1) guru.  :-)

Sorry if this message is a bit to the side of the Real Topic at hand, but 
this has been on my mind :-)

--------------------------------------------------------------------------
Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                               Home: 408.866.1912
NAS: M/S 258-6                                          Work: 415.604.0935
Moffett Field, CA 94035                                Pager: 415.428.6939