kernel freeze 2.8

[Flemming Laugaard <Flemming.Laugaard@uni-c.dk>]

Hi all

I wrote to the list a week or 2 ago, about a problem with some deadbeef
message and a system freeze.

It seems that it's not because of a memory err. With ECC memory chips the
messages goes away, but we still experience problems. After a while the 
console is dead and it's not possible to start _any_ TCP conns to the
server (ie SSH), but the kernel will "happily" route packages.

It looks like userspace is gone somewhere faraway.

The configuration is something likes this:

700 MHz AMD 
128 MB RAM
ASUS A7V (BIOS 1004D)
Adaptec SCSI ctrl
1 nic 3com 90x type
1 nic Tulip type

The kernel (2.8) is patched with the latest patches.
IPF is enabled. The problem will appear with a ipf rules set with or
without states. Apparently the freeze only happens under high netload.  

There is a bugreport in http://cvs.openbsd.org/cgi-bin/wwwgnats.pl/full?pr=1504
which seems to be almost the same.

Anybody got any good ideas to what the problem is? It looks like a
kernel/ipf problem.

Anybody? Please? 

-- 
Kind regards
Flemming Laugaard