Subject: kernel panics since 3/1: m_copym errors
To: None <tech-net@netbsd.org>
From: Hisashi T Fujinaka <htodd@twofifty.com>
List: tech-net
Date: 03/04/2005 10:07:26
I've been having repeated crashes on a -current i386 system with what
appears to be mbuf copying code. Since I'm often not physically near the
system and I need it to be up, I just have it rebooted and haven't
looked at a stack trace.

I'm sure this is a bit uninformative without a stack trace, but I see:

Mar  1 12:12:46 chris /netbsd: panic: m_copym: m == 0, len 1 [!COPYALL]
Mar  2 07:32:34 chris /netbsd: panic: m_copym: m == 0, len 1 [!COPYALL]
Mar  2 07:32:34 chris /netbsd: panic: m_copydata: m == 0, len 1
Mar  2 15:58:22 chris /netbsd: panic: m_copydata: m == 0, len 1
Mar  3 20:05:28 chris /netbsd: panic: m_copydata: m == 0, len 1
Mar  4 07:32:29 chris /netbsd: panic: m_copydata: m == 0, len 2

So I have a couple of simple questions, has anything major changed in
the past week? Is m_copym or m_copydata used directly from userland
at all? Is the mbuf copying code used in anything besides the network
stack?

The system that goes down all the time is a mail and web server. I have
two other -current systems that have no problems. I even swapped out the
nic when I had a chance, thinking that was the problem. But now it's
time to look at things more closely.

I have looked through 4.3bsd network code in the past, but it's
been a while.

Thanks.

-- 
Hisashi T Fujinaka - htodd@twofifty.com
BSEE(6/86) + BSChem(3/95) + BAEnglish(8/95) + MSCS(8/03) + $2.50 = latte