Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751365AbWIIH05 (ORCPT ); Sat, 9 Sep 2006 03:26:57 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751366AbWIIH04 (ORCPT ); Sat, 9 Sep 2006 03:26:56 -0400 Received: from gate.crashing.org ([63.228.1.57]:27790 "EHLO gate.crashing.org") by vger.kernel.org with ESMTP id S1751365AbWIIH04 (ORCPT ); Sat, 9 Sep 2006 03:26:56 -0400 Subject: Re: Opinion on ordering of writel vs. stores to RAM From: Benjamin Herrenschmidt To: Paul Mackerras Cc: Linus Torvalds , linux-kernel@vger.kernel.org, akpm@osdl.org, segher@kernel.crashing.org, davem@davemloft.net In-Reply-To: <17666.11971.416250.857749@cargo.ozlabs.ibm.com> References: <17666.8433.533221.866510@cargo.ozlabs.ibm.com> <17666.11971.416250.857749@cargo.ozlabs.ibm.com> Content-Type: text/plain Date: Sat, 09 Sep 2006 17:24:41 +1000 Message-Id: <1157786681.31071.168.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Evolution 2.6.1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 707 Lines: 18 > I suspect the best thing at this point is to move the sync in writeX() > before the store, as you suggest, and add an "eieio" before the load > in readX(). That does mean that we are then relying on driver writers > putting in the mmiowb() between a writeX() and a spin_unlock, but at > least that is documented. Well, why keep the sync in writel then ? Isn't it agreed that the driver should use an explicit barrier ? Or did I misunderstand Linus ? Ben. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/