From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262635AbVCXQ4Q (ORCPT ); Thu, 24 Mar 2005 11:56:16 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262638AbVCXQ4Q (ORCPT ); Thu, 24 Mar 2005 11:56:16 -0500 Received: from omx2-ext.sgi.com ([192.48.171.19]:39631 "EHLO omx2.sgi.com") by vger.kernel.org with ESMTP id S262635AbVCXQ4O (ORCPT ); Thu, 24 Mar 2005 11:56:14 -0500 From: Jesse Barnes To: Benjamin Herrenschmidt Subject: Re: [PATCH] ppc32/64: Map prefetchable PCI without guarded bit Date: Thu, 24 Mar 2005 08:54:45 -0800 User-Agent: KMail/1.7.2 Cc: Andrew Morton , linuxppc-dev list , Linux Kernel list References: <1111645464.5569.15.camel@gaston> In-Reply-To: <1111645464.5569.15.camel@gaston> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200503240854.45741.jbarnes@engr.sgi.com> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wednesday, March 23, 2005 10:24 pm, Benjamin Herrenschmidt wrote: > While experimenting with framebuffer access performances, we noticed a > very significant improvement in write access to it when not setting > the "guarded" bit on the MMU mappings. This bit basically says that > reads and writes won't have side effects (it allows speculation). It > appears that it also disables write combining. Doesn't pgprot_writecombine imply non-guarded, so can't you use it instead? Either way, you'll probably want to fix fbmem.c as well and turn off _PAGE_GUARDED? Maybe it's time for a more generic call to support this stuff, both for in-kernel mappings and ones that we export to userspace. Jesse