Re: Kernel Oops

[Jason Boxman <jasonb@edseek.com>]

On Thursday 13 September 2001 01:40 pm, Lazar Fleysher wrote:
> Hi everybody

Hey!

> I have noticed in /var/log/kern.log
> the following message: ( system slink kernel 2.0.38)
>
> Sep  9 12:03:39 zoroport kernel: Serial driver version 4.13p1 with no
> serial opt
> ions enabled
> Sep  9 12:03:40 zoroport kernel: ttyS00 at 0x03f8 (irq = 4) is a 16450
> Sep  9 12:03:40 zoroport kernel: ttyS01 at 0x02f8 (irq = 3) is a 16450
> Sep  9 16:48:13 zoroport kernel: Unable to handle kernel NULL pointer
> dereferenc

I just read someone yesterday when I was reading up on the BadRAM kernel 
patch in the docs that his is very often a sign of bad RAM.  (The NULL 
pointer)

You might try removing some of your modules, if you can, to see if the 
problem goes away.

Also, you can get a problem called "memtester" which I believe has a Debian 
package available.  You can then boot it with LILO and test 99.9% of your 
RAM, not counting what few K it uses for itself.  It will tell you with 
extreme accuracy whether nor not you have some bad RAM modules.

>From there, if you want, you can fiddle with the BadMEM patch that's 
available for Linux 2.4.x.  If your chip has some non fluxuating areas that 
are damaged you may be able to map them out of usuable RAM and continue along 
happily.  This is pretty experimental though.  I wouldn't use it on a 
production machine.  Just buy a new stick if need be, it's cheap.

> e at virtual address c0000000
> Sep  9 16:48:13 zoroport kernel: current->tss.cr3 = 00101000, %cr3 =
> 00101000
> Sep  9 16:48:13 zoroport kernel: *pde = 00102067
> Sep  9 16:48:13 zoroport kernel: *pte = 00000000
> Sep  9 16:48:13 zoroport kernel: Oops: 0000
> Sep  9 16:48:13 zoroport kernel: CPU:    0
>
>
<snip logs>
>
> Could someone tell what this is and how to respond to such things...
> Is it a hardware problem or software?
>
> Thank you
>
> Lazar
>
> PS There is a "hole" in time because the system had contacted a time
> server and adjusted its time...