Find the source of the kernel lockups > kernel 2.4.4
Anthony Liu
anthony@nexus-online.com
Tue, 31 Jul 2001 20:31:06 +0800
On Tue, Jul 31, 2001 at 11:52:17AM +0200, Holger Brueckner wrote:
> hi since a lot of people seem to have similar problems we should
> somehow join to find out the problem. since i'm not a kernel hacker i
> won't be very helpfull in digging through the sources, but i'm sure
> willing to provide a lot of tracing information if you tell me what
> you'd like to have.
I am not sure if the problem I experienced was triggered by the same
bug as the original poster.
>
> ok first steps:
>
> 1) find out what problems everybody has, and which setup they're
> using. fill in your setup here:
>
> Holger: asus/celeron 633, on board 3c905tornado, 4 port dlink dec
> chipset kernel pppoe, pppd 2.4.0, ip source routing, masquerading on
> ppp connected via german telekom t-dsl. netfilter with conntracking
> symptom: unreproducable random lockups. only hard resets work.
>
> asus motherboard, pentium 200mmx, intel eepro100,
> teles 16/0 isdn card ipppd (unknown version). netfilter with
> conntracking symptom: unreproducable random lockups. only hard resets
> work. mostly on low load after high load. (freeing memory bug ??)
All RTL8139c nics, tested on 2.4.4 and 2.4.7 (2.4.5-6 skipped). Debian
pppd 2.4.1-3, pppoe 3.0-4, iptables 1.2.2-2, ADSL connection. Problem:
When starting adsl connection, the login server on the telecom side
either not responding or offline, in about 60 seconds or so kernel
panic and system hang. The problem IS REPEATABLE. The kernel exception
message suggested that the pppd process is causing the problem. Kernel
2.2.19 did not experience the kernel exception (all with the same set
of software except the kernel), though the adsl connection remain
unreachable.
>
> 2) find out what has changed from 2.4.3 (which seems to work finde for
> most of us) to 2.4.4 and above
In my case, ppp may have triggered the problem, but under 2.2 kernel
it did not happen with the same set of software. So the real bug is
likely in the 2.4 kernel.
>
> 3) find a solution :)
Forward the message to LKM?
--
Linux 2.4.4fw #3 Thu Jun 21 21:29:13 CST 2001 i586 unknown
19:30:01 up 14:57, 1 user, load average: 0.00, 0.01, 0.00