> On Mon, Mar 26, 2007 at 10:31:36AM +0900, YAMAMOTO Takashi wrote:
> > > Any comments/suggestions?
> > 
> > autonice is scheduler-independent as well.
> > (thus the "runtm" argument for sched_pstats_hook is unnecessary.)
> 
> Mmh, I tend to disagree. autonice is not a generic mechanism
> needed by all schedulers but a specific mechanism of the 4BSD
> scheduler to prevent starvation. So, I think it should go into the
> scheduler-dependent hook.

do you mean 4BSD scheduler has its specific reason to need autonice?

YAMAMOTO Takashi