Re: Unable to handle kernel NULL pointer dereference
On Tue, Jul 16, 2002 at 11:59:30AM +0200, Nicos Gollan wrote:
> On Tuesday 16 July 2002 01:01, nate wrote:
>
> Time for another useless anecdote.
>
> I had ordered a PC2100 Micron memory module. When it arrived, it was
> apparently dead, so I sent it back. Got a new one with the same result.
> I tested it on five different boards with three different chipsets
> which all failed to boot up or even POST. The fun thing about it was,
> however, that when used as secondary module, the system ran fine until
> something tried to access memory beyond the first module. A slow POST
> failed on accessing the second module. When booting to a memory test
> program under DOS, the whole memory range was accessible and it showed
> no errors whatever I did. To this day, I don't know what might have
> been wrong with the two modules.
>
> So, never trust your memory ;-)
Somebody should have mentioned memtest86. For a PC, at least, most
memory testing programs are just good at testing the cache, or don't
detect adjacent joined bits or anything other than stuck bits.
memtest86, on the other hand, is an absolute bit nazi. It goes through
the standard tests (all 00, FF, address low byte), but also tacks on
bit walking ( 01, 02, 04, 08, ...), adjacent complements, and a couple
schemes I don't recognize. It also properly controls (and tests!) the
cache memory, unlike most other programs.
You can add memtest86 to your lilo (or grub most likely) and boot the
machine into a dedicated memory testing mode, let that run for a few
hours, and feel pretty damned good about your memory.
--
To UNSUBSCRIBE, email to debian-user-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: