From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261785AbTESJEA (ORCPT ); Mon, 19 May 2003 05:04:00 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261787AbTESJD7 (ORCPT ); Mon, 19 May 2003 05:03:59 -0400 Received: from web11802.mail.yahoo.com ([216.136.172.156]:159 "HELO web11802.mail.yahoo.com") by vger.kernel.org with SMTP id S261785AbTESJD7 (ORCPT ); Mon, 19 May 2003 05:03:59 -0400 Message-ID: <20030519091656.55599.qmail@web11802.mail.yahoo.com> Date: Mon, 19 May 2003 11:16:56 +0200 (CEST) From: =?iso-8859-1?q?Etienne=20Lorrain?= Subject: Re: [PATCH] Use MTRRs by default for vesafb on x86-64 To: linux-kernel@vger.kernel.org MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Whatever the processor, you cannot use write combining MTRR on 16 colors / 4 BPP modes. In this mode you can copy 8 pixels at a time by doing a simple movb, but it cannot handle 32 pixel copy at the same time by a movd, for instance. In 16 color modes, you never have a linear memory mapping of pixels. Etienne. ___________________________________________________________ Do You Yahoo!? -- Une adresse @yahoo.fr gratuite et en français ! Yahoo! Mail : http://fr.mail.yahoo.com