Re: patch-2.1.37-2{,-1}

Manuel J. Galan (
Thu, 01 May 1997 23:01:26 +0100

Stephan K. Zitz wrote:
> Even with the most recent patch made available (regarding ipv6 symbols, etc. .
> .) against patch-2.1.37-2, compiling arch/i386/kernel/init_task.c has caused
> mine and one other person's gcc to core.
> Is anyone else having this problem? Is there a workaround? I realize this is
> a development kernel, but it'd be nice to at least get the kernel to compile
> to continue the other development ;)

Me too ( hey this is not an adult check hacker list! :-)

'arch/i386/kernel/init_task.c' is the offender

gcc dies with the (in)famous signal 11 (I was almost going
to replace my 128 Mb. ECC parity RAM) but ... not: the problem
can be reproduced always in the same place...

What is happening with __SMP__? Are we -kind of- stuck? It seems
that we are facing a *major threat* to make it work.

A few days ago I faced something which I never would have expected
before: NT4 running a SMP kernel for several days without a glitch
in a Supermicro dual ppro mb (*not* the PDNH) where Linux 2.0.xx and
2.1.xx failed to boot or work more than a few minutes... ;-((

Manuel J. Galan