From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752868AbZBNLo1 (ORCPT ); Sat, 14 Feb 2009 06:44:27 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751507AbZBNLoT (ORCPT ); Sat, 14 Feb 2009 06:44:19 -0500 Received: from nox.protox.org ([88.191.38.29]:54498 "EHLO nox.protox.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751227AbZBNLoT (ORCPT ); Sat, 14 Feb 2009 06:44:19 -0500 X-Greylist: delayed 534 seconds by postgrey-1.27 at vger.kernel.org; Sat, 14 Feb 2009 06:44:18 EST Subject: Re: [PATCH]: drm: radeon: Use surface for PCI GART table. From: Jerome Glisse Reply-To: glisse@freedesktop.org To: David Miller Cc: michel@daenzer.net, benh@kernel.crashing.org, airlied@linux.ie, dri-devel@lists.sourceforge.net, linux-kernel@vger.kernel.org In-Reply-To: <20090214.020431.193689843.davem@davemloft.net> References: <20090214.015107.127190039.davem@davemloft.net> <1234605599.4823.177.camel@thor> <20090214.020431.193689843.davem@davemloft.net> Content-Type: text/plain; charset="UTF-8" Date: Sat, 14 Feb 2009 12:35:11 +0100 Message-Id: <1234611311.2860.2.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Evolution 2.24.3 (2.24.3-1.fc10) Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Sat, 2009-02-14 at 02:04 -0800, David Miller wrote: > From: Michel Dänzer > Date: Sat, 14 Feb 2009 10:59:59 +0100 > > > On Sat, 2009-02-14 at 01:51 -0800, David Miller wrote: > > > This allocates a physical surface for the PCI GART table, this way no > > > matter what other surface configurations exist the GART table will > > > always be seen by the hardware properly. > > > > BTW, I don't think the swapping settings affect GPU access to the table, > > only CPU access, but this is a good solution anyway. > > They absolutely and positively do effect GPU access to the table. > I've proven it with many conclusive tests over the past 3 days. > > > This is only really necessary on big endian, but other than that: > > > > Acked-By: Michel Dänzer > > Thanks. This will need to be revisited with KMS as i think we should not allow userspace to play with surface. Maybe we could setup surface on BO mapping and disable them on bo unmapping (we would to register some callback on vmclose). Cheers, Jerome Glisse