From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758433AbZDQVwT (ORCPT ); Fri, 17 Apr 2009 17:52:19 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751180AbZDQVwE (ORCPT ); Fri, 17 Apr 2009 17:52:04 -0400 Received: from smtp-vbr3.xs4all.nl ([194.109.24.23]:2086 "EHLO smtp-vbr3.xs4all.nl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751051AbZDQVwD (ORCPT ); Fri, 17 Apr 2009 17:52:03 -0400 X-Greylist: delayed 541 seconds by postgrey-1.27 at vger.kernel.org; Fri, 17 Apr 2009 17:52:03 EDT Date: Fri, 17 Apr 2009 23:42:12 +0200 From: Folkert van Heusden To: Heiko Carstens Cc: Ingo Molnar , Peter Zijlstra , Martin Schwidefsky , Christian Borntraeger , linux-kernel@vger.kernel.org Subject: Re: [PATCH] mutex: have non-spinning mutexes on s390 by default Message-ID: <20090417214212.GG10554@vanheusden.com> References: <20090409174758.74abec87@osiris.boeblingen.de.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20090409174758.74abec87@osiris.boeblingen.de.ibm.com> Organization: www.unixexpert.nl X-Chameleon-Return-To: folkert@vanheusden.com X-Xfmail-Return-To: folkert@vanheusden.com X-Phonenumber: +31-6-41278122 X-URL: http://www.vanheusden.com/ X-PGP-KeyID: 1F28D8AE X-GPG-fingerprint: AC89 09CE 41F2 00B4 FCF2 B174 3019 0E8C 1F28 D8AE X-Key: http://pgp.surfnet.nl:11371/pks/lookup?op=get&search=0x1F28D8AE Read-Receipt-To: Reply-By: Sat Apr 18 21:15:00 CEST 2009 X-Message-Flag: Want to extend your PGP web-of-trust? Coordinate a key-signing at www.biglumber.com User-Agent: Mutt/1.5.18 (2008-05-17) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > The adaptive spinning mutexes will not always do what one would expect on > virtualized architectures like s390. Especially the cpu_relax() loop in > mutex_spin_on_owner might hurt if the mutex holding cpu has been scheduled > away by the hypervisor. > We would end up in a cpu_relax() loop when there is no chance that the > state of the mutex changes until the target cpu has been scheduled again by > the hypervisor. > For that reason we should change the default behaviour to no-spin on s390. Hmmm. Wouldn't this be a change that applies to ibm system p as wel? E.g. with linux in an lpar. Folkert van Heusden -- ---------------------------------------------------------------------- Phone: +31-6-41278122, PGP-key: 1F28D8AE, www.vanheusden.com