From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754751AbZHYJUd (ORCPT ); Tue, 25 Aug 2009 05:20:33 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754687AbZHYJUb (ORCPT ); Tue, 25 Aug 2009 05:20:31 -0400 Received: from tx2ehsobe001.messaging.microsoft.com ([65.55.88.11]:7201 "EHLO TX2EHSOBE001.bigfish.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754156AbZHYJUa convert rfc822-to-8bit (ORCPT ); Tue, 25 Aug 2009 05:20:30 -0400 X-SpamScore: -20 X-BigFish: VPS-20(z34a4jz154dM1432R98dN936eMa594izz1202hzzz32i6bh203h43j61h) X-Spam-TCS-SCL: 0:0 X-FB-SS: 5, X-WSS-ID: 0KOXDXW-01-4FB-02 X-M-MSG: Date: Tue, 25 Aug 2009 11:19:41 +0200 From: Andreas Herrmann To: Peter Zijlstra CC: Ingo Molnar , linux-kernel@vger.kernel.org Subject: Re: [PATCH 14/15] sched: Conditionally limit __cpu_power when child sched domain has type NODE Message-ID: <20090825091941.GI20811@alberich.amd.com> References: <20090820131243.GO29327@alberich.amd.com> <20090820134506.GC29327@alberich.amd.com> <1251128132.7538.300.camel@twins> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline In-Reply-To: <1251128132.7538.300.camel@twins> User-Agent: Mutt/1.5.16 (2007-06-09) X-OriginalArrivalTime: 25 Aug 2009 09:19:41.0906 (UTC) FILETIME=[2D847320:01CA2565] Content-Transfer-Encoding: 8BIT Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, Aug 24, 2009 at 05:35:32PM +0200, Peter Zijlstra wrote: > On Thu, 2009-08-20 at 15:45 +0200, Andreas Herrmann wrote: > > We need this in case of performance policy. All sched_groups in > > child's parent domain (MN in this case) should be limited such that > > tasks are balanced among these sched_groups. > > /me fails at correlating the above changelog and the below patch. > So here we go mess up cpu_power again in order to invluence the > placement policy? Yep, to restrict the capacity of sched_groups in MN domain. But you pointed already out that messing up cpu_power is unwanted. So I have to look for an alternative to influence placement policy such that cpu_power is kept intact and always properly normalized. > > Signed-off-by: Andreas Herrmann > > --- > > kernel/sched.c | 4 ++-- > > 1 files changed, 2 insertions(+), 2 deletions(-) > > > > diff --git a/kernel/sched.c b/kernel/sched.c > > index 0c950dc..ab88d88 100644 > > --- a/kernel/sched.c > > +++ b/kernel/sched.c > > @@ -8555,11 +8555,11 @@ static void init_sched_groups_power(int cpu, struct sched_domain *sd) > > */ > > if (!(sd->flags & SD_POWERSAVINGS_BALANCE) && > > ((child->flags & > > - (SD_SHARE_CPUPOWER | SD_SHARE_PKG_RESOURCES)))) { > > + (SD_SHARE_CPUPOWER | SD_SHARE_PKG_RESOURCES)) || > > + (child->level == SD_LV_NODE))) { > > sd->groups->__cpu_power = 0; > > sg_inc_cpu_power(sd->groups, SCHED_LOAD_SCALE); > > } > > - > > } > > > > /* > Andreas -- Operating | Advanced Micro Devices GmbH System | Karl-Hammerschmidt-Str. 34, 85609 Dornach b. München, Germany Research | Geschäftsführer: Thomas M. McCoy, Giuliano Meroni Center | Sitz: Dornach, Gemeinde Aschheim, Landkreis München (OSRC) | Registergericht München, HRB Nr. 43632