On Thu, 20 Apr 2000, Alan Cox wrote:
> The kernel will log bogus network setups. It should definitely _not_ be
> crashing it.
If you're suggesting it's something else, I'm not certain what it
could be, other than perhaps the ATM stuff. I went back and did it again
in single user mode to get some logs without much stuff running.
I went into single user mode with "init 1" and took down all the
network interfaces, include lo. I ran ver_linux:
-- Versions installed: (if some fields are empty or looks
-- unusual then possibly you have very old versions)
Linux moebius.student.cwru.edu 2.2.14 #2 Fri Apr 14 11:48:01 EDT 2000 i686 unknown
Kernel modules 2.3.10
Gnu C 2.95.2
Binutils 2.9.5.0.31
Linux C Library 2.1.3
Dynamic linker ldd: version 1.9.11
Procps .
Mount 2.10f
Net-tools 2.05
Kbd command
Sh-utils 2.0g
Modules Loaded ipx af_packet nls_iso8859-1 nls_cp437
I set ipx_configure --auto_interface=on. ipx_configure settinsg
looked like this (output from ipx_configure):
Auto Primary Select is OFF
Auto Interface Create is ON
I then ran the ATM daemons as I normally do:
atmsigd -b
ilmid -b -l syslog
zeppelin -F syslog -2 &
/etc/atmsigd.conf contains:
sig uni30
debug level fatal
debug log syslog
entity 0.0.5 {
default
}
I did _not_ ifconfig. I immediately starting getting the familar
network collision messages:
Apr 22 10:33:12 moebius kernel: IPX: Network number collision ca008137
Apr 22 10:33:12 moebius kernel: lec0 EtherII and lec0 SNAP
Apr 22 10:33:12 moebius kernel: IPX: Network number collision ca008137
Apr 22 10:33:12 moebius kernel: lec0 EtherII and lec0 802.3
Apr 22 10:33:12 moebius kernel: IPX: Network number collision ca008137
Apr 22 10:33:12 moebius kernel: lec0 EtherII and lec0 802.2
I did _nothing_ after this point but wait. After about 5 minutes,
I got:
Kernel panic: skput:over: c0143fc5:241 put:241 dev:lec0
In swapper task - not syncing
(This did not get logged, I copied it from screen by hand).
As mentioned before, I am on a college LAN with hundreds (probably
thousands) of clients. The IPX frame type is supposed to be EtherII. The
following information was taken after I put up the final ATM daemon
(zeppelin), while the network collision errors were occuring:
"ipx_configure check lec0 EtherII" output:
IPX Address for (EtherII, EtherII) is CA008137:00204808C79B.
"ipx_configure check lec0 802.2" output:
IPX Address for (802.2, 802.2) is 00000000:00204808C79B.
"ipx_configure check lec0 802.3" output:
IPX Address for (802.3, 802.3) is 00000000:00204808C79B.
"ipx_configure check lec0 SNAP" output:
IPX Address for (SNAP, SNAP) is 00000000:00204808C79B.
"ps aux" output:
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 1 2.1 0.7 1020 468 ? S 10:25 0:04 init [
root 2 0.0 0.0 0 0 ? SW 10:25 0:00 [kflushd]
root 3 0.0 0.0 0 0 ? SW 10:25 0:00 [kupdate]
root 4 0.0 0.0 0 0 ? SW 10:25 0:00 [kpiod]
root 5 0.0 0.0 0 0 ? SW 10:25 0:00 [kswapd]
root 228 0.0 0.7 1020 468 tty1 S 10:25 0:00 init [
root 229 0.0 2.0 2120 1308 tty1 S 10:25 0:00 bash
root 334 0.0 0.9 1352 600 ? S 10:28 0:00 /sbin/syslogd
root 336 0.2 1.0 1264 668 ? S 10:28 0:00 /sbin/klogd
root 340 0.0 0.8 1204 548 tty1 S 10:28 0:00 /usr/sbin/atmsigd -b
root 346 0.0 0.8 1164 572 tty1 S 10:28 0:00 /usr/sbin/ilmid -b -l
syslog
root 361 0.0 0.6 1040 416 ? S 10:28 0:00 /usr/sbin/zeppelin -F
syslog -2
root 362 0.0 1.5 2624 960 tty1 R 10:28 0:00 ps aux
Something is certainly going on and it's quite reproducible. I'm
willing to do whatever tests or whatnot to try to resolve this.
Chris Pimlott
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to [EMAIL PROTECTED]
