On 11/13/2002 17:53:07 Charles M. Hannum <abuse@spamalicious.com> wrote:
> >> This is actually a kernel bug.  One page is not being synced correctly
> >> in the I$ on some systems.
> > But the new kernel worked fine with the old ld.elf_so.
> `You were lucky before.'

I get this too - luckily I've got an old ld.elf_so that works stashed away.

> > What should I do now? Wait for a new kernel or a new ld.elf_so?
> > Btw, my source tree is from the weekend, maybe it's already fixed? 
> Well, it's definitely a kernel bug, so I'm not going to do anything to
> ld.elf_so to fix it.  As for the kernel, to my knowledge it hasn't
> been fixed.  It only seems to affect MP machines, so I have no way to
> test it myself.

No, it's not just a problem on MP machines - it happens on my iBook too (I'd
post a dmesg, but the machine is at home). I don't have any MP specific stuff
in my kernel config, in fact it's just GENERIC with some stuff disabled
(using 'include' and 'no options XXX')

If there are tests that people want me to try, let me know.

Ian.