debugging / fixing kernel panics on ubuntu

[Kristian Hermansen]

On 4/29/07, Jarod Wilson <jarod at wilsonet.com> wrote:
> You may want to check out kdump. Its what Red Hat ships as its standard kernel
> crash dump collection method in RHEL5. The bulk of it is in upstream as of a
> few kernel releases ago. The only caveat for your situation is that x86_64
> relocatable (within memory) kernel support isn't upstream yet, so I believe
> you'll need a separate kernel to kexec into that is set up to run at a
> different memory offset for panic dump purposes.
>
> nb: i386 and ia64 kernels are relocatable, x86_64 patches exist and are in
> RHEL5, but haven't yet been merged in Linus' tree -- may make 2.6.22, ppc64
> is still a work in progress.

Thanks for the tip!  Apparently Ubuntu included kdump in the previous release...
http://packages.ubuntu.com/edgy/devel/linux-image-kdump

> There's also a link there to the upstream project page which could have
> further pointers.

Well, I couldn't determine the cause of the "Aieee!", until I read this:
http://resource.intel.com/telecom/support/tnotes/tnbyos/2000/tn062.htm

Then, I just left my laptop at home for today, and when I got back, I
had a nice kernel fault trace on my virtual terminal.  I snapped a
pic...
http://kristian-hermansen.com/tmp/feisty-ndiswrapper-amd64-aieee.jpg

Now, this bug is most likely due to the fact that I am running amd64,
with ndiswrapped 64-bit broadcom driver, using a non-repository
version of ndiswrapper (from source).  I should have been more careful
:-)  Of course, I wish I didn't have closed-source drivers, grrrr
BROADCOM!!!

However, I would still like to know the real reason for the panic.  So
maybe I'll break out kdump if I see it happen again using the
repository version of ndiswrapper.  Thanks for the help!
-- 
Kristian Hermansen

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.