Unable to handle kernel paging request linked to network problem

Marc (marcl@magic.metawire.com)
Mon, 31 Mar 1997 16:20:04 -0800 (PST)


[Please CC me a copy of your answer as I've been reading the list on the
newsgroup side, and since the gateway(s) seem to be broken (I get all kinds
of messages, but none kernel related), I subscribed to the mailing list, but
it may take a while before I receive messages]

This message looks like me asking for help for a network setup problem, but
in fact, those are a sequence of events that I hope will shed light on the
kernel problem that seems to be occuring.

I've never had any problem so far with the kernel I'm currently running
(2.0.29 with dmsdos patch, and a boomerang network card in Vortex
compatibility mode added about 10 days ago, both running as modules).

This morning, I played with my second network card (ne2000) and I was trying
to setup a point to point link with another machine (I wanted to route to it
without having to use 4 IPs to create a special 2 IP net).
So, I did the following:
ifconfig eth0 198.147.96.93 pointopoint 198.147.96.79 # (#1 ptp #2)
route add 198.147.96.79 gw 198.147.96.93
arp -s 198.147.96.79 00:60:97:7F:4E:CA netmask 255.255.255.255 pub -i eth1
(eth1 is the lan side and the ethernet address is eth1's)

(On the other linux machine I setup a similar config without worring about
arp but with the following routes:
route add 198.147.96.93
route add -net default gw 198.147.96.93)

I know this violates a few rules (ethernet was not designed to be point to
point, but it seems to work anyway), but I tried it and it works.

The problem is that a little while after playing with all this, I got the
following problem:

moremagic:/var/tmp# /etc/rc.d/init.d/network stop
/etc/sysconfig/network-scripts/ifdown-eth: line 6: 4593 Segmentation fault ifconfig ${DEVICE} down
/etc/sysconfig/network-scripts/ifdown-eth: line 6: 4602 Segmentation fault ifconfig ${DEVICE} down
/etc/sysconfig/network-scripts/ifdown-lo: line 6: 4611 Segmentation fault ifconfig ${DEVICE} down
[device is successively eth1, eth0, and loopback]

Mar 31 14:25:47 moremagic kernel: Unable to handle kernel paging request at virtual address c2acc6d8
Mar 31 14:25:47 moremagic kernel: current->tss.cr3 = 01748000, ^Pr3 = 01748000
Mar 31 14:25:47 moremagic kernel: *pde = 0009e067
Mar 31 14:25:47 moremagic kernel: *pte = 00000000
Mar 31 14:25:48 moremagic kernel: Unable to handle kernel paging request at virtual address c2acc6d8
Mar 31 14:25:48 moremagic kernel: current->tss.cr3 = 0116e000, ^Pr3 = 0116e000
Mar 31 14:25:48 moremagic kernel: *pde = 0009e067
Mar 31 14:25:48 moremagic kernel: *pte = 00000000
Mar 31 14:25:48 moremagic kernel: Unable to handle kernel paging request at virtual address c2acc6d8
Mar 31 14:25:48 moremagic kernel: current->tss.cr3 = 00992000, ^Pr3 = 00992000
Mar 31 14:25:48 moremagic kernel: *pde = 0009e067
Mar 31 14:25:48 moremagic kernel: *pte = 0000000

moremagic:/var/tmp# /etc/rc.d/init.d/network start
/etc/sysconfig/network-scripts/ifup-eth: line 12: 4651 Segmentation fault ifconfig ${DEVICE} ${IPADDR} netmask ${NETMASK} broadcast ${BROADCAST}
/etc/sysconfig/network-scripts/ifup-eth: line 12: 4667 Segmentation fault ifconfig ${DEVICE} ${IPADDR} netmask ${NETMASK} broadcast ${BROADCAST}
/etc/sysconfig/network-scripts/ifup-lo: line 11: 4682 Segmentation fault ifconfig ${DEVICE} ${IPADDR}

Mar 31 14:27:23 moremagic kernel: Unable to handle kernel paging request at virtual address c2acc6d8
Mar 31 14:27:23 moremagic kernel: current->tss.cr3 = 01896000, ^Pr3 = 01896000
Mar 31 14:27:23 moremagic kernel: *pde = 0009e067
Mar 31 14:27:23 moremagic kernel: *pte = 00000000
Mar 31 14:27:24 moremagic kernel: Unable to handle kernel paging request at virtual address c2acc6d8
Mar 31 14:27:24 moremagic kernel: current->tss.cr3 = 014e5000, ^Pr3 = 014e5000
Mar 31 14:27:24 moremagic kernel: *pde = 0009e067
Mar 31 14:27:24 moremagic kernel: *pte = 00000000
Mar 31 14:27:25 moremagic kernel: Unable to handle kernel paging request at virtual address c2acc6d8
Mar 31 14:27:25 moremagic kernel: current->tss.cr3 = 01595000, ^Pr3 = 01595000
Mar 31 14:27:25 moremagic kernel: *pde = 0009e067
Mar 31 14:27:25 moremagic kernel: *pte = 0000000

moremagic:/var/tmp# tcpdump
Segmentation fault
Mar 31 15:26:40 moremagic kernel: Unable to handle kernel paging request at virtual address c2acc6e0
Mar 31 15:26:40 moremagic kernel: current->tss.cr3 = 010f2000, ^Pr3 = 010f2000
Mar 31 15:26:40 moremagic kernel: *pde = 0009e067
Mar 31 15:26:40 moremagic kernel: *pte = 0000000
etc...

Routing still works, adding and deleting routes doesn't do any harm, but
each ifconfig command, or network peeking program (I assume they do
internally a ifconfig DEV promisc) causes an "Unable to handle kernel paging
request").

Right now the machine is still up, and I'll happily create account(s) for
whoever wants to have a closer look while the machine is up in this state (I
assume that once I reboot, I'll most likely won't be able to reproduce the
problem).

Thanks,
Marc