Re: [linux-mac68k] Re: Results of trying 2.6.7 and 2.6.8 on Macintosh Quadra 650
On Thu, 26 Aug 2004, Christian T. Steigies wrote:
> On Thu, Aug 26, 2004 at 05:02:59PM +0200, Geert Uytterhoeven wrote:
> > >
> > > I just checked out the amiga and mac kernel images and they do need fixing
> > > (according to objdump -p, I don't have Amiboot). So, I guess Roman's patch
> > > wasn't applied?
> >
> > Yep, same result here. The kernel I did manage to boot with Amiga-Lilo did have
> > the stack program header... Strange...
> >
> > > It would be good to see the patch tested more widely if that would help
> > > get it into CVS. I've only tested the patch with binutils-2.15 (gcc 3.3.4
> > > and 3.4.1) and it worked fine. It would be nice if it also fixed kernels
> > > built with binutils-2.14.
> >
> > I can confirm it fixed my uClinux kernel linked with 2.14.90.0.7.
>
> The famous "Roman's" patch is for binutils, right? I can't find it in the
> list right now, but maybe I am not looking hard enough. If it is a patch for
> the linux source, I do not see it in CVS yet, there where no chances since I
> built the kernel-images. Can somebody point me to "the patch" so I can try
> to build a new cross compiler with it? Or maybe send some mac patches along
> or whatever you want in the kernel? Or should I just upload the current
> images to the debian archive?
It is this patch for linux 2.6.8,
http://lists.debian.org/debian-68k/2004/08/msg00322.html
It is not in the linux-m68k CVS yet. It fixes the "unsupported backward
seek" error from a variety bootloaders (inluding penguin and amiboot) that
results from booting an m68k kernel built using binutils 2.14 or 2.15. It
should make my crude perl hack obsolete.
I've personally tested it with FSF binutils-2.15, and just now with Debian
testing binutils-2.14.90.0.7-8. With earlier binutils, it would ideally
have no effect on the resulting image, but I don't know if anyone has
tried that.
-F
Reply to: