Re: Random sig[segv|fpe] on 164LX with 256MB RAM

[Andrew Hobgood <chaos strange net>]

> > > Filling with seed of s=0x420
> > > memtest: memory violation at pc=120000900 rp=1200008c8 (bad address = 00000000)
> > > Segmentation fault (core dumped)
> 
> STOP! STOP!!!

I did, as soon as I recognized what was going on. =P

> Just check what hte malloc is returning, if its returning NULL then its failed;
> turn down that nice friendly SIZE parameter at the top.

As I did... I incrementally tried values until it'd drive me to just the edge 
of swapping out.

> If the memory test fails it will print a pair of values, the value it expected
> and the value it got back with a big furry error about comparison failing.

And as a reference point, I got no furry errors through about, oh 0x100 seeds.
So it *seems* like my memory is just fine.  Any other theories for my random
segfaults and fpe's?

I've got some additional data... an rxvt croaked on me yesterday...

Sep 23 01:49:04 schizo kernel: rxvt: memory violation at pc=155559fc1ec rp=15555
9f9d90 (bad address = 60000000)

I hope that someone can recognize the error and give me a way to fix it.  As
far as I know, the Aspen Systems people wouldn't be giving me broken RAM, and
that's backed up by Dave's memtester.

/Andrew