[ This is my first ever kernel problem report. Please let me know if I'm
reporting it to the wrong place or using the wrong format. ]
Full Description:
~~~~~~~~~~~~~~~~
I'm suffering from intermittent kernel oopses on my main machine. As far
as I can dechipher, the kernel oopses happen in the ipchains.o module.
The kernel oopses identifiy themselves as "tainted", which I don't quite
understand: The only modules *ever* loaded are modules produces from the
standard kernel sources as downloaded from ftp.kernel.org. No patches
have been applied. I have even taken the precaution of clearing out
/lib/modules and re-compiling the modules from the kernel (make modules
&& make modules_install). As far as I can see during boot-up, the kernel
warns about being tainted as soon as ipchains.o is loaded.
/proc/version:
~~~~~~~~~~~~~~
Linux version 2.4.18 (root@hawking) (gcc version 2.95.4 (Debian prerelease)) #4 SMP Mon Mar 25 23:19:15 GMT 2002
Keywords:
~~~~~~~~~
ipchains
2.4.18
netfilter
Output from oops:
~~~~~~~~~~~~~~~~~
See kernel-panic[123].txt attached. I only have the last 24 lines of
each, as they scroll off the screen. Let me know if I need to get
the preceding lines - I presume "vga=<something>" in /etc/lilo.conf
should do the trick.
I ran ksymoops on it (with same modules loaded in the same order),
and it consistently points to the ipchains module and diald. The
stack traces are identical too, although there are slight variations in
the register contents.
The ksymoops output and kernel .config (named config-2.4.18) is attached.
Somewhat reproduceable:
~~~~~~~~~~~~~~~~~~~~~~~
Unfortunately, I cannot reproduce the problem "on demand". But since
it seems diald/ipchains related, I have attached the output of
"ipchains -L -n -v" (packet counts from a later boot of course), so
you can see the firewalling rules in effect.
I have not tried modifying with the firewalling rules (yet);
they have served me quite well under kernel 2.2.
Environment:
~~~~~~~~~~~~
(not sure this matters to much for your diagnostics of this
specific problem, but who knows?)
Debian GNU/Linux - woody
Motherboard: Abit VP6
Single 800MHz Pentium Coppermine (in socket *2*, not 1. The previous
occupant of socket 1 died - cause of death was presumably
overheating: Some of the thermal paste stuff was missing. Worked
13 months though...)
512 Mb memory
NOTE: I am not 110% sure that my hardware is OK. It is not always
willing to boot after a power on/off cycle, and doesn't always get
as far as lilo before hanging - that has happened twice in the last
month.
However, as the oopses consistently point to ipchains/diald, and it
is rock-stable under kernel 2.2.20 (once it manages to start), I
don't think this is a hardware related problem.
Software:
~~~~~~~~~
Debian GNU/Linux - woody.
scripts/ver_linux output:
Linux hawking 2.4.18 #4 SMP Mon Mar 25 23:19:15 GMT 2002 i686 unknown
Gnu C 2.95.4
Gnu make 3.79.1
util-linux 2.11n
mount 2.11n
modutils 2.4.13
e2fsprogs 1.26
reiserfsprogs 3.x.1a
PPP 2.4.1
Linux C Library 2.2.5
Dynamic linker (ldd) 2.2.5
Procps 2.0.7
Net-tools 1.60
Console-tools 0.2.3
Sh-utils 2.0.11
Modules Loaded lp tap0 dummy0 isofs inflate_fs loop ipchains
parport_pc imm parport
Debian package versions:
diald 0.99.4-5 dial on demand daemon for PPP and SLIP.
ppp 2.4.1.uus-1 Point-to-Point Protocol (PPP) daemon.
Processor Information:
~~~~~~~~~~~~~~~~~~~~~~
Single pentium 800 Coppermine:
/proc/cpuinfo:
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 8
model name : Pentium III (Coppermine)
stepping : 6
cpu MHz : 798.714
cache size : 256 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
mca cmov pat pse36 mmx fxsr sse
bogomips : 1592.52
Module Information:
~~~~~~~~~~~~~~~~~~~
/proc/modules report:
lp 5536 0 (autoclean)
tap0 2560 0 (autoclean)
dummy0 960 1 (autoclean)
isofs 24448 3 (autoclean)
inflate_fs 17760 0 (autoclean) [isofs]
loop 8304 9 (autoclean)
ipchains 43848 31
parport_pc 15688 2 (autoclean)
imm 8800 0 (unused)
parport 12768 2 [lp parport_pc imm]
Loaded driver & hardware info:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Not sure it's relevant for this problem, but here goes:
Contents of /proc/ioports & /proc/meminfo attached.
(these bit is from the 2.2.20 kernel)
PCI Information:
~~~~~~~~~~~~~~~~
Not sure it's relevant for this problem, but here goes:
Output of "lspci -vvv" attached.
(this bit is from the 2.2.20 kernel)
SCSI Information:
~~~~~~~~~~~~~~~~~
Not sure it's relevant for this problem, but here goes:
Contents of /proc/scsi/scsi is attached
(this bit is from the 2.2.20 kernel)
Other Information:
~~~~~~~~~~~~~~~~~~
I'm not comfortable with the fact that the ipchains module "taints"
the kernel upon load. I'm aware that ipchains is deprecated in favour of
iptables, and that's the way I'm going. But to get there I only want to
change one thing at a time. Undoubtedly you know the drill already.
Currently, I'm using kernel 2.2.20, and the machine is overdue for a
kernel update. So I decided to go straight for 2.4.18.
Perhaps I would be better off upgrading to 2.4.x (x < 18), switching
over to iptables, and then going to 2.4.18? If so, which version of the
kernel would you suggest? My main criteria is that it should support
both ipchains and iptables, so I can perform a smooth transition.
Regards
-- Karl E. Jørgensen karl@jorgensen.com www.karl.jorgensen.com I'm currently out trying to find myself. If I should get back before I return, please keep me here.
This archive was generated by hypermail 2b29 : Sun Apr 07 2002 - 22:00:11 EST