SUMMARY: Kernel panic at trap level 2.

From: Thomas Leitner (tom@radar.tu-graz.ac.at)
Date: Thu Jun 22 2000 - 14:51:09 CDT


Hi,

Sorry for the long overdue summary. The kernel panic did not occur again
until I got some good advices from list members here. Basically I did two
things:

- use BSD compression rather than deflate compression with PPP by
  using these statements in /etc/ppp/options:

    bsdcomp 15
    deflate 0

  As my serial line is rather flakey (4-5 errors per day) it seems that
  the BSD compression is more robust against transmission errors.

- increased the kernel stack size with these statements in /etc/system

   *
   * increase kernel stack size to avoid crashes
   *
   set lwp_default_stksize=0x4000
   set rpcmod:svc_run_stksize=0x4000
   *

Thanks to the following people for their replies:

Adi Masputra <Adi.Masputra@eng.sun.com>
Kathie Hibbard <hibbard@millenium.Colorado.EDU>
Roger Fujii <rmf@lookhere.com>
Kevin Sheehan {Consulting Poster Child} <kevin@joltin.com>
"Siepe, Carlos" <CSiepe@uniFON.com.ar>
George DONE <george@romsys.ro>
Sean <shadow6@uswest.net>

Thanks again everybody // Tom

P.S: WARNING: I've also got one suspicious e-mail with the following contents:

--------------------------------------- SNIP
Date: Fri, 5 May 2000 18:15:45 EDT
From: Linuxmindcontrol@aol.com
To: tom@radar.tu-graz.ac.at
Subject: Re: AGAIN: Kernel panic at trap level 2.

do a "strings" on the core file and sound me the output..i will try to
help
in the analysis....
i.e.
{strings "core file" > /tmp/sout}
--------------------------------------- SNIP

Needless to say that I did *not* send him what he wanted. With this
information he probably had enough to easily break into my machine.

P.P.S: Here's my original posting.

-----------------------------------------------------------------------

Hello,

It happened again. The machine crashed this morning again with a
similar crash traceback:

# echo '$c' | adb -k unix.1 vmcore.1
physmem f7e6
complete_panic(0xf,0x10406c00,0x10006900,0x0,0x10006904,0x44) + 24
do_panic(0x1,0x1004d400,0x1004d400,0x104041b0,0x0,0xffffffc0) + 158
panic(0x10404168,0x2,0x0,0x0,0x0,0x60004) + 1c
sys_tl1_panic(0x10704c7c,0x10704c80,0x4049c118,0x0,0x0,0x108d2740) + 8
sfmmu_tteload_array(0x70137fb8,0x4049c118,0x72aec000,0x4049c124,0x1,0x10527a10)+
4c

I am up to the latest recommended patches including 106541-10.

I am running pppd 2.3.11 on the box and frequently see messages like
these in /var/adm/messages

May 5 07:54:42 finwss02 unix: ppp0: bad fcs (len=489)
May 5 07:54:43 finwss02 unix: z_decompress0: bad seq # 5, expected 4

They occur about 2-3 times a day. Can this cause the crash? Maybe the
decompression routine in the ppp driver overwrites some kernel memory?

Any ideas anyone // Tom

P.S: Here's my old posting about the problem.

On Mon, 24 Apr 2000, Thomas Leitner wrote:

> As a followup to my question, here's a traceback of the crash:
>
> /opt/crash> echo '$c' | adb -k unix.0 vmcore.0
> physmem f7e6
> complete_panic(0xf,0x10406c00,0x10006900,0x0,0x10006904,0x44) + 24
> do_panic(0x1,0x1004d400,0x1004d400,0x104041b0,0x0,0xffffffc0) + 158
> panic(0x10404168,0x2,0x0,0x0,0x0,0x0) + 1c
> sys_tl1_panic(0x20001,0x0,0x24,0x4,0x4,0x10411a74) + 8
> page_create_get_something(0x10410f34) + a0
>
> Does this ring a bell for anybody?

-- 
--------------------------------------------------------------------------
Dr. Tom Leitner                             Dept. of Communications
                                            Graz University of Technology, 
e-mail    : tom@radar.tu-graz.ac.at         Inffeldgasse 12
Phone     : +43-316-873-7455                A-8010 Graz / Austria / Europe
Fax       : +43-316-463-697
Home page : http://wiis.tu-graz.ac.at/people/tom.html
PGP public key on : ftp://wiis.tu-graz.ac.at/pgp-keys/tom.asc or send 
mail with subject "get Thomas Leitner" to pgp-public-keys@keys.pgp.net
--------------------------------------------------------------------------
    Before we have the paperless office, we have the paperless toilet!



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:14:10 CDT