Subject: re: Moving scheduler semantics from cpu_switch() to kern_synch.c
To: Jason Thorpe <thorpej@shagadelic.org>
From: matthew green <mrg@eterna.com.au>
List: tech-kern
Date: 09/22/2006 02:50:42 
   
   > OK, why can't we use a single value instead of a per-cpu one?
   
   Because we'd like to have per-CPU run queues (for processor affinity  
   -- actually, I think the Mach "processor set" model is not a bad  
   example to follow here).


it just seems suboptimal to have to set a flag in every cpu_info when
there is a (random) process to run.


on a related issue, i kind of like the idea of real idle threads as
opposed to the method used for current SMP sysetms (where this is all
handled in MD code.)