Re: Is it bad to have lots of sleeping tasks?

[Alan Cox]

> > scheduler with 100 lines of patch, and that fitted in the existing Linux
> > runqueue framework).  What's the resistence to scheduler changes?
> 
> Expect Larry to jump on you.

If you can implement an O(1) scheduler that is faster than Linus scheduler
for the normal case and for scalable cases I expect Larry would probably
jump up and down happily instead.

The current scheduler does make some very bad decisions