Re: 68040 buserror patch
On Sat, Oct 07, 2000 at 07:26:25PM -0500, Andrew McPherson wrote:
> On Sat, 7 Oct 2000, Michel Lanners wrote:
> > Hi all,
> > On 6 Oct, this message from Andrew McPherson echoed through cyberspace:
> > >> The patch applies with minimal intervention to 2.2.16 from mac68k CVS
> > >> and it seems to work (for me at least :). Would somebody with write
> > >> access to CVS please commit it?
> > >>
> > >
> > > OK, I applied the patch and it seems to work. ypbind runs now. :) I put up
> > > a kernel at ftp://macduff.dhs.org/pub/cvs/vmlinux-2.2.16-100500.gz. I'll
> > > commit the patch to CVS shortly.
> > Though your kernel does run, it locks up on my Performa 475 after some
> > time (of inactivity?).
interesting.. I had once a problem with an Xserver that crashed always when
I was not watching it. Turned out to be the screen blanking mechanism ;)
Maybe your kernel tries to spin down something?
> Hmmm.... that's not good at all. Question: Does your Performa still have a
> 68LC040 processor, or does it have a full 68040? If it has an LC040, than
> we may have an interesting decision to make: If a patch fixes a nasty bug
> on machines that otherwise work, but severely breaks already broken
> machines, what do we do about it? The patch appears to be mostly assembly,
> and though I'm not proficient at 68k assembly, I'll look to see if it does
> anything that would break an LC040.
the buserror patch should not break anything itself, I am running variations
of it for more than 3 months. The assembly part of it is fairly minimal btw.
Possibly it is something with 2.2.16? I have tried 2.2.17 which exhibits
the classical 68040 fork problem, so I have not really tested anything newer
than 2.2.10 in the 2.2 branch.
Could you test the patch with your previous stable kernel? Or 2.2.16
without this patch?