>> both have killed off long-running bc jobs with a "Profiling
>> timer expired" message.
>
> I've seen this too. Didn't reproduce very well, but
> happened a couple of times.
> (Sometimes after half an hour, an rpc daemon once
> died after a couple of days.)

I've experienced processes were killed by "Profiling timer expired"
messages several times before.  Kernels made from -current tree
around last August and September brought me a series of inconsistent
crashes/bogusdoms.  I'm not sure whether they were related with
so-called "siganl lossage" which happend at recent -current or not.

My working development kernel built with 10/5 -current seems run well
for last 7 days as below;

	$ uptime
	 8:05AM  up 6 days, 19:10, 1 user, load averages: 0.15, 0.11, 0.09

Tohru Nishimura
Nara Institute of Science and Technology