Looking for diagnostic ideas....

Hello, all.

I have a trusty, albeit older PC (AMD K6-2/500) running an old Linux
distro (2.2.10) serving as my firewall/router through my cable modem
(from in-house 100mb network). This setup has worked superbly for going
on six years. Just in the last few months, however, I've started
noticing kernel panics indicating overflows across the eth0 interface.

Kernel panic: skput: over <address>:28740 put:28740 dev:eth0

The "28740" value is the same each and every time the kernel panics.
Now, this looks to me like an overflow - tried to write x bytes in an
y<x byte buffer. Is that a reasonable interpretation?

The card in question is a TULIP-based card (brand name LiteOn), using
the via-rhine driver. As noted, its worked just about flawlessly since
I built the thing and slapped Linux on it five or six years ago.

Now, given all other variables are the same, am I looking at the
possibility of the network card for eth0 in the early stages of
failing, or could there be some really unusual activity on my network
that I need to investigate further that's exposing a problem in the
driver (seems awfully unlikely at this point)? Given that its a fairly
small machine (only 64MB on this AMD K6-2/500), any possibility I'm
sending it more data than it can handle? Hardly seems likely, either,
but...

Admittedly, I'm grasping at straws, but the first thing I think of when
mysterious failures occur is deteriorating/failing hardware. If there
are other possibilities to pursue, I'd appreciate them.

Many thanks,
-intrepid

On 16 Jan 2006 11:31:03 -0800, (E-Mail Removed) wrote:

>The "28740" value is the same each and every time the kernel panics.
>Now, this looks to me like an overflow - tried to write x bytes in an
>y<x byte buffer. Is that a reasonable interpretation?
No.

>The card in question is a TULIP-based card (brand name LiteOn), using
>the via-rhine driver. As noted, its worked just about flawlessly since
>I built the thing and slapped Linux on it five or six years ago.

Why would you not run tulip driver with a tulip card?

lspci ? dmesg |grep eth[01] ?? uname -a ??


Dust bunnies in the CPU fan heatsink, reseat memory, check for bulging
or leaky filter capacitors on mobo, also inside power supply -- lots
can go wrong with a box that age.

Grant.
--
Cats are smarter than dogs. You can't make eight cats pull
a sled through the snow.

> No.

Okay, why? I thought the very essence of "skputver" in the kernel
panic was to show a buffer overflow condition....please help me
understand why that's not right...?

Thanks,


Grant wrote:
> On 16 Jan 2006 11:31:03 -0800, (E-Mail Removed) wrote:
>
> >The "28740" value is the same each and every time the kernel panics.
> >Now, this looks to me like an overflow - tried to write x bytes in an
> >y<x byte buffer. Is that a reasonable interpretation?
>
> >The card in question is a TULIP-based card (brand name LiteOn), using
> >the via-rhine driver. As noted, its worked just about flawlessly since
> >I built the thing and slapped Linux on it five or six years ago.
>
> Why would you not run tulip driver with a tulip card?
>
> lspci ? dmesg |grep eth[01] ?? uname -a ??
>
>
> Dust bunnies in the CPU fan heatsink, reseat memory, check for bulging
> or leaky filter capacitors on mobo, also inside power supply -- lots
> can go wrong with a box that age.
>
> Grant.
> --
> Cats are smarter than dogs. You can't make eight cats pull
> a sled through the snow.