"Charles M. Hannum" writes:
> 
> I have a few things to note about this patch:
> 
> 1) It would be less of a performance lose to simply make p_priority
> and p_nice be ints; they wouldn't require extra locking when written.
> 
> 2) It's not necessary to call splstatclock() when reading a snapshot
> of p_nice, p_priority, or p_estcpu (such as in retrunnable() or
> proc_compare()), providing they're marked as volatile.  You're not
> even consistent about this anyway.

yea, I thought some of the spl regions that I added were
somewhat gratuitous, but I was just trying to be more consistent
than the original code.

> 3) You made exactly the mistake that the big comment in userret(),
> right above the code you modified, mentions.

hmm, yea, it did seem like that wasn't quite right, but I couldn't
put my finger on it.

on the other hand, I've figured out a much better and less intrusive
fix for the zs overrun problem that all this is trying to solve,
which I'll post over on port-sparc shortly.

so, ah, "never mind..."  :-)

-Chuck