updated kernel panics with dhcp

[Joe Smith]

Andy Green wrote:
> ...  I would hit 'A' at the grub boot
> menu, and append stuff like
> 
> acpi=noirq + pci=biosirq ... Even acpi=off
> 
> might be worth it just as a test.

Ok, tried both: no change.

I didn't report exactly what sequence I tried because the results were 
not entirely consistent.

The first thing I tried was power off the modem and re-boot, then 
re-start the modem, drop to single user and start eth0. That worked.

Then I tried booting to single user and manually starting eth0 with the 
modem running. That panic'ed.

Nothing raises my blood pressure like repeatedly booting a PC while 
making random changes, so I gave up at that point.

Today, after reading the suggestions, I tried the kernel options above 
by booting to single user and then
# mount -a -o ro,remount
# service network start

One try with "acpi=noirq pci=biosirq" failed.

The next try, with "acpi=off" and the RJ45 unplugged until right before 
I started eth0, worked.

I re-booted and tried the same sequence and this time it panic'ed again.

So, it has managed to start eth0 twice, but there's no obvious pattern 
to the successes.

The panic location was the same every time I remembered to check: 
__list_add+0x48/0x5c

Thanks for the suggestions--I'm willing to try almost anything that 
would help diagnose the problem.

Should I file a bug report?

<Joe