Re: P_NOLOAD + ULE = high load averages

From: David O'Brien <obrien_at_FreeBSD.org> Date: Thu, 22 Apr 2004 10:50:36 -0700 · This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:52 UTC

On Wed, Apr 21, 2004 at 03:32:30AM +0300, Nikos Ntarmos wrote:
> --- sys/kern/sched_ule.c.orig   Wed Apr 21 03:10:29 2004
> +++ sys/kern/sched_ule.c        Wed Apr 21 03:20:34 2004
> _at__at_ -357,7 +357,7 _at__at_
>         if (class == PRI_TIMESHARE)
>                 kseq->ksq_load_timeshare++;
>         kseq->ksq_load++;
> -       if (class != PRI_ITHD)
> +       if (class != PRI_ITHD && (ke->ke_proc->p_flag & P_NOLOAD) == 0)

This does not apply to 5.2-CURRENT (rev. 1.104).  I've adjusted your
patch locally, but can you please email me an updated testd patch just to
make sure what I'm going to commit is what you want.

-- 
-- David  (obrien_at_FreeBSD.org)