One of my systems is exhibiting random coredumps accompanied by
messages in /var/log/messages like
Feb 24 00:51:31 /netbsd: NMI: system interrupts:
40000000<VME=0,SBUS=0,ME>
Feb 24 00:55:03 /netbsd: trap type 0x21: pc=0x100c2d0c
npc=0x100c2d10 psr=41400085<S>
Feb 24 00:55:04 /netbsd: pid 12425 (xdpyinfo), uid 101: exited on
signal 4 (core dumped)
ME here is "Module Error (async)", according to the comments in
intreg.h. The hardware is an LX.
Does this mean bad CPU, bad RAM, what? The code responsible for
generating these messages looks basically identical all the way from
1.4T through -current (the oldest and newest source trees I have handy
access to); comments in that code lead me to suspect a bad CPU, but
I'd
rather check that with people with more clue before I assume the CPU's
gone bad (for an LX it's soldered in)....