savecore.c re-revisited, pstat problems

[utzoo!decvax!pur-ee!purdue!cak]

I sent the fixes to 4bsdbugs, but thanks for the reminder. I never did
learn to count from 0!

On a similar subject, we have an alternative to renice (called renice)
that lets users move their own jobs around. Converting it to 4.1 it
doesn't work any more -- it can't find processes. And the strange
thing is that pstat fails to find the same processes! BUt ps shows
them. Any ideas on this one?

Also, we have had a problem with mba hangs. We have our root disk (RM03)
and TE16 on the same mba, and under heavy load, the mba occasionally
hangs. I have a core dump to dig through, but haven't had time. We
saw this a lot under 4.0 where our (now departed) systems programmer
had done a local hack (he never really understood mutual exclusion,
it seems), so we blamed it on him, and looked forward to 4.1 where
the code would be right. But it still hangs sometimes. Has anyone
else seen this behaviour? Is there a fix?

On a general note, 4.1 seems slower to us than 4.0....wasn't it
supposed to be the other way around? We get more jobs in the
run q, which I guess it the idea, but general user opinion is
that the system runs slower.  Also, we're seeing jobs get niced
down by system, even when no one else is on....

chris