Bill Holler wrote: > Very cool! > > The idle loop needs to know an estimate of when the next cyclic > will fire. We currently look at the cy_expire field of the top > cyclic on the cpu's cyclic heap to see when the next cyclic will > fire. Can we still use the cy_expire field of the 2 cyclics used > for callout's for this? (e.g. when the interval is set to infinity, > will cy_expire also be set to infinity?) > Certainly. > I would like to see the design doc and webrev if possilble. :-) At the end of this week, I plan to generate the final design doc and webrev. I will send a pointer. Also, I will point you to the project gate in case Sasha or anyone else would like to experiment with it. I hear that the thread accounting work you guys are doing is based on callouts. Might want to check it out. Of course, this code has not been thoroughly tested yet.
Madhavan
