On tirsdag, okt 8, 2002, at 02:04 Europe/Copenhagen, Keith Mastin wrote:

>> On Mon, Oct 07, 2002 at 02:30:26AM -0400, Thomas Mueller wrote:
>>> Daniel Eggert <danieleggert@mac.com> writes:
>>>> ld also sometimes gets a signal 11. Signal 4 is SIGILL: 'illegal
>>>> instruction' and signal 11 is SIGSEGV: 'segmentation violation'.
>>>
>>> : Signal 11s are a symptom of hardware memory failures.
>>>
>>> : Perry
>>>
>>> I think NMI is likely to result from hardware memory failure.
>>>
>>> But segment violations can result from software errors, and I don't 
>>> think this
>>
>> It can, but it this case it would not happen at random place, with 
>> different
>> signals.
>
> Sounds like running memtest86 on the box might be in order to verify 
> the
> memory is bad....

This has been suggested, but it only runs on the i386 platform, so I'll 
have to trial and error -- replacing the RAM blocks one by one. I've 
some RAM in an unused (spare part) machine.

Daniel