I keep getting these. Even once when I tried to list the directory with "ls". "ls" worked directly after that with hardly any pause. I've been trying to compile a 2.6 kernel using gcc-3.4.6 and get segmentation faults at different parts of the build each time.

I finally got drunk and asked someone with more experience than me and he said it could be bad memory. What are the chances of a stick of RAM being bad at certain parts? Is there any way of testing the memory?

It is bad memory.

If gcc craps out at different stages of the kernel build the
chances of bad ram are > 99% (even though there's a remote chance
of it being DMA related as the stuff gets paged out). I've had
a very nasty case on one athlon machine, where the RAM sticks
were fine in other machines but didn't like this one board for
the heck of it. With memtest they only failed ONE out of all the
subtests, but that was enough to upset Linux (winDOHs ME which
the original owner had been using for yonks never had any problems).

As for the testing: memtest86
It's part of many live CDs.


Cheers,
Tink

I am getting the same seg faults while trying to upgrade kernel and I am in the process of testing with memtest 3.2 and got several errors with memory, but not sure if its the cpu cache or the ram. The program is kinda confusing but at lest I know some memory is bad.

Very true that it could be bad memory, but my machine also likes to segfault if I overclock too far, so if you are OC'd, try taking it down and see if the segfaults continue. If they cease, you've gone too far.