From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753722Ab0IILUl (ORCPT ); Thu, 9 Sep 2010 07:20:41 -0400 Received: from mail.klingt.org ([86.59.21.178]:50425 "EHLO klingt.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753288Ab0IILUi (ORCPT ); Thu, 9 Sep 2010 07:20:38 -0400 X-Greylist: delayed 425 seconds by postgrey-1.27 at vger.kernel.org; Thu, 09 Sep 2010 07:20:38 EDT From: Tim Blechmann To: Peter Zijlstra Subject: Re: [process scheduler] Possible bug in context_swich()? Date: Thu, 9 Sep 2010 13:12:18 +0200 User-Agent: KMail/1.13.5 (Linux/2.6.35.4; KDE/4.5.1; x86_64; ; ) Cc: Mike Galbraith , fabio de francesco , LKML , Ingo Molnar References: <201009081728.18227.fabio@metanix.org> <1283999539.9103.54.camel@marge.simson.net> <1284028746.23762.54.camel@laptop> In-Reply-To: <1284028746.23762.54.camel@laptop> MIME-Version: 1.0 Content-Type: multipart/signed; boundary="nextPart9804333.zPxQLRbZyt"; protocol="application/pgp-signature"; micalg=pgp-sha1 Content-Transfer-Encoding: 7bit Message-Id: <201009091312.25424.tim@klingt.org> X-Greylist: Sender succeeded SMTP AUTH, not delayed by milter-greylist-4.3.7 (klingt.org [86.59.21.178]); Thu, 09 Sep 2010 13:12:27 +0200 (CEST) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org --nextPart9804333.zPxQLRbZyt Content-Type: Text/Plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable On Thursday, September 09, 2010 12:39:06 pm Peter Zijlstra wrote: > On Thu, 2010-09-09 at 04:32 +0200, Mike Galbraith wrote: > > On Wed, 2010-09-08 at 17:54 +0200, Peter Zijlstra wrote: > > > On Wed, 2010-09-08 at 17:28 +0200, fabio de francesco wrote: > > > > In context_switch() (in linux/kernel/sched.c), starting with release > > > > 2.6.33, two "unlikely" macro have been changed to "likely". I think > > > > the previous logic was right while the latter is wrong. > > > >=20 > > > > In case I am missing something I, please, ask someone to explain the > > > > above mentioned inversion of logic through releases. > > >=20 > > > It helps if you CC people, LKML alone is a bit of a gamble. > > >=20 > > > git blame kernel/sched.c, will tell you that the change you refer to > > > comes from: > > >=20 > > > commit 710390d90f143a9ebb87a475215140f426792efd > > > Author: Tim Blechmann > > > Date: Tue Nov 24 11:55:27 2009 +0100 > > >=20 > > > sched: Optimize branch hint in context_switch() > > > =20 > > > Branch hint profiling on my nehalem machine showed over 90% > >=20 > > > incorrect branch hints: > > That change never made any sense to me, seems Tim must have been > > measuring a kthread load. I benched at the time, and saw absolutely > > zero difference one way or the other wrt max ctx rate on my Q6600. >=20 > One option is to simply remove the whole branch hint.. But lets ask Tim > what kind of workload he used.. i was using a standard desktop workload, nothing special ... =2D-=20 tim@klingt.org http://tim.klingt.org Wherever we are, what we hear is mostly noise. When we ignore it, it disturbs us. When we listen to it, we find it fascinating. John Cage --nextPart9804333.zPxQLRbZyt Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) iEYEABECAAYFAkyIwRIACgkQg0On+TFEWzwI+ACffNgSl3oi9ze7w/1J5TWg61sw 8oEAniuDVGYeyjzi+NTv5KhlDXbEPg2V =+DWx -----END PGP SIGNATURE----- --nextPart9804333.zPxQLRbZyt--