From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andi Kleen Subject: Re: [PATCH] mm: Work around Intel SNB GTT bug with some physical pages. Date: Mon, 07 May 2012 16:57:09 -0700 Message-ID: References: <1336432421-17972-1-git-send-email-marcheu@chromium.org> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <1336432421-17972-1-git-send-email-marcheu@chromium.org> (=?utf-8?Q?=22St=C3=A9phane?= Marchesin"'s message of "Mon, 7 May 2012 16:13:41 -0700") Sender: linux-kernel-owner@vger.kernel.org To: =?utf-8?Q?St=C3=A9phane?= Marchesin Cc: linux-kernel@vger.kernel.org, keithp@keithp.com, torvalds@linux-foundation.org, seanpaul@chromium.org, olofj@chromium.org, dri-devel@lists.freedesktop.org List-Id: dri-devel@lists.freedesktop.org St=C3=A9phane Marchesin writes: > While investing some Sandy Bridge rendering corruption, I found out > that all physical memory pages below 1MiB were returning garbage when > read through the GTT. This has been causing graphics corruption (when > it's used for textures, render targets and pixmaps) and GPU hangups > (when it's used for GPU batch buffers). It would be possible to exlude GFP_DMA from the page allocator. That covers the first 16MB. You just need a custom zone list with ZONE_DMA. -Andi --=20 ak@linux.intel.com -- Speaking for myself only