Subject: Need advice how to debug VM problem on port-hp700.
To: None <port-hp700@NetBSD.org, tech-kern@NetBSD.org>
From: Jochen Kunz <jkunz@unixag-kl.fh-kl.de>
List: tech-kern
Date: 10/28/2003 17:27:46
Hi.

On my way reanimating port-hp700 I got to somthing that looks like VM
problem. First symtom was that the kernel froze while trying to do
DHCP/BOOTP. It hung for ever in tsleep(). I traced this down to the
point where cpu_switch() writes to curlwp. (Third assembler instruction
of that function.) I threw in more debug printf()s into the nfs_boot
code and now I get an endless: 
root on ie0
kernel: DTLB miss trap, code=0
kernel: DTLB miss trap, code=0
kernel: DTLB miss trap, code=0
kernel: DTLB miss trap, code=0

So this seams not to be a problem with cpu_switch() but with the VM
system. The kernel crashes somewhere, depending on the actual position
of "somthing" inside the kernel. Leading to an endless sequence of page
faults? Or some bugs in the trap handling code triggered by a page
fault?

Unfortunately I have no clue about all this VM stuff and how to hunt
bugs of this type. So I am looking for some friendly VM guru giving me
some insight.
-- 


tschüß,
       Jochen

Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/