[ltp] Kernel panicking a bit too often on a TP 600E + other symptoms - SOLVED, thanks...

Jakub Hegenbart linux-thinkpad@linux-thinkpad.org
Sun, 23 Nov 2003 15:56:15 +0100

I've finally solved the problem. I managed to run memtest from LILO and 
discovered some 600 000 errors within 4 minutes. The errors started 
appearing after 64 MBs of physical adress space and it was the LSB that 
was permanently a logical zero. Since I have 32M (onboard) + 32M 
(SO-DIMM) + 128MB (SO-DIMM), I thought the 128M DIMM is guilty. I 
removed the modules entirely and memtest passed OK. I put them both 
inside and guess what? Still A-OK.

Most probably the module pins were somehow out of alignment. It crashed 
twice again when I pressed the memory cover plate. The memory housing is 
a bit too shallow and the 128 M DIMM is seated under the (most 
deformable) center part of the plate so when I pressed the bottom of the 
NB on that place, I worsened the pin contact. Somebody should hang for 
the SO-DIMM slot construction (and for the 600E mem seating) :/ Well, it 
doesn't crash anymore even with both modules, but just to be sure, I put 
only the 128M module to the upper (0) slot which is less vulnerable this 
way and half the modules means half the chance of inserting them 
incorrectly. 160M is just fine for me :) 192M makes hardly any 
performance difference.

And the X vs. console stability? Well, it sooner got into the utterly 
malfunctional RAM area under X. I'm just glad that everything is as 
deterministic here as it should be...

Thank you all very much for your patience.

Jakub Hegenbart