Re: port-powerpc/43022 (assembling powerpc64 atomic_add.S fails while building librump)

[Antti Kantee <pooka%Netbsd.org@localhost>]

The following reply was made to PR port-powerpc/43022; it has been noted by 
GNATS.

From: Antti Kantee <pooka%Netbsd.org@localhost>
To: Dennis Ferguson <dennis.c.ferguson%gmail.com@localhost>
Cc: gnats-bugs%NetBSD.org@localhost
Subject: Re: port-powerpc/43022 (assembling powerpc64 atomic_add.S fails while 
building librump)
Date: Sun, 21 Mar 2010 10:55:36 +0200

 On Sun Mar 21 2010 at 16:30:23 +0800, Dennis Ferguson wrote:
 > On 21 Mar 2010, at 13:51 , pooka%NetBSD.org@localhost wrote:
 > > Note that a macppc64 kernel build fails with:
 > > nbmkdep: /objs/tools/bin/powerpc--netbsd-gcc: not found: No such file or 
 > > directo
 > > ry
 > > 
 > > I wouldn't attempt to do anything with powerpc64 before elementary 
 > > things like that are fixed.
 > 
 > Um, I think fixing that isn't so elementary: there is no macppc64
 > kernel to build, that is code which was never written.  I'm
 > considering taking that on since I'd really like to have one.
 > 
 > I'm pretty sure, however, that at one point you could successfully
 > build a 64 bit powerpc userland for the non-existant kernel, and it
 > is bit rot in that which I've been filing bugs against.  If you are
 > saying there is no value in filing bugs against the userland build
 > bit rot until I have a kernel to build too I can live with that,
 > but I'm a little bit surprised by it.  Not only does it make any
 > work I'd like to do slightly harder (more local modifications == more
 > potential merge failures from cvs updates), but if I never finish
 > the kernel it leaves this work for the next person who's interested
 > in a 64 bit powerpc kernel to do all over again.
 > 
 
 No, I'm saying don't be surprised if *kernel code* doesn't build for
 powerpc64 (yes, your patches address that).