From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S964909AbVLQAlz (ORCPT ); Fri, 16 Dec 2005 19:41:55 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S964902AbVLQAlz (ORCPT ); Fri, 16 Dec 2005 19:41:55 -0500 Received: from rwcrmhc14.comcast.net ([216.148.227.89]:6528 "EHLO rwcrmhc12.comcast.net") by vger.kernel.org with ESMTP id S964901AbVLQAly (ORCPT ); Fri, 16 Dec 2005 19:41:54 -0500 From: Jesse Barnes To: "David S. Miller" Subject: Re: [PATCH 1/19] MUTEX: Introduce simple mutex implementation Date: Fri, 16 Dec 2005 16:41:49 -0800 User-Agent: KMail/1.9 Cc: torvalds@osdl.org, dhowells@redhat.com, nickpiggin@yahoo.com.au, arjan@infradead.org, akpm@osdl.org, alan@lxorguk.ukuu.org.uk, cfriesen@nortel.com, hch@infradead.org, matthew@wil.cx, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org References: <20051216.145306.132052494.davem@davemloft.net> In-Reply-To: <20051216.145306.132052494.davem@davemloft.net> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200512161641.49571.jbarnes@virtuousgeek.org> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Friday, December 16, 2005 2:53 pm, David S. Miller wrote: > When the write comes along, the next transaction occurs to kick it > out the other cpu(s) caches and then the local line is placed into > Owned state. > > I'll have to add "put write prefetch in CAS sequences" onto my sparc64 > TODO list :-) Note that under contention prefetching with a write bias can cause a lot more cache line bouncing than a regular load into shared state (assuming you do a load and test before you try the CAS). We actually saw this on large Altix machines, http://lia64.bkbits.net:8080/to-linus-2.5/cset%403f2082b3xCvMG9OSeNu3aWhoe6jnOg?nav=index.html| src/.|src/include|src/include/asm-ia64| related/include/asm-ia64/spinlock.h fixed things up for us. Jesse