Subject: Re: supervisor trap page fault in lfs_putpages
To: None <current-users@NetBSD.org>
From: Paul Ripke <stix@stix.id.au>
List: current-users
Date: 12/01/2006 07:53:53
On Thu, Nov 30, 2006 at 09:46:17PM +1100, Paul Ripke wrote:
> Running -current with sources from about 24hrs ago, I'm getting a 100%
> reproducible crash by running /etc/daily. ddb says:
> 
> kernel: supervisor trap page fault, code=0
> 
> <trace deleted>
> 
> System is an old Dell dual PIII, kernel is GENERIC.MP with heaps stripped
> out. I have netbsd.gdb and dump if anyone is interested.
> 
> I can't find any matching PRs, but just thought I'd ask here before
> send-pr'ing.

Just a followup - it was suggested this might be a hardware issue -
I've now run memtest86 for 9+ hours and 20+ passes with no errors.
The problem appeared when I upgraded from the old netbsd-4 branch
to -current just before the hackathon.

I'm guessing it's due to the RAM constrained nature of this system
(only 128 MB) that's bringing out this issue. I also just booted
a kernel without MULTIPROCESSOR, and was able to quickly crash the
system with the same backtrace.

-- 
stix