From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263002AbTK3T6Y (ORCPT ); Sun, 30 Nov 2003 14:58:24 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263008AbTK3T6X (ORCPT ); Sun, 30 Nov 2003 14:58:23 -0500 Received: from 204.Red-213-96-224.pooles.rima-tde.net ([213.96.224.204]:19725 "EHLO betawl.net") by vger.kernel.org with ESMTP id S263002AbTK3T6W (ORCPT ); Sun, 30 Nov 2003 14:58:22 -0500 Date: Sun, 30 Nov 2003 20:58:15 +0100 From: Santiago Garcia Mantinan To: linux-kernel@vger.kernel.org Subject: IDE DMA setting not available on 2.4.23 as a module Message-ID: <20031130195815.GA2409@man.beta.es> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.4i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi! Yesterday I started to upgrade my systems to 2.4.23, some of them were already running pre or rc versions, but when I tried my Pentium MMX wich boots out of SCSI and on which I like to have IDE driver as a module, I found that the DMA setting was not working, hdparm was saying: setting using_dma to 1 (on) HDIO_SET_DMA failed: Operation not permitted So I changed the driver option to set DMA by default to ON, but nothing changed, still the same problem, then I tried to compile the IDE driver into the kernel instead of having it as a module and then the dma support started to work, either having the driver enable it by default or by setting it with hdparm. The motherboard uses a 430TX chipset, thus with a PIIX4 IDE controller. Is this a bug or is this a known and accepted limitation of having the IDE driver as a module? If you want me to test any patch to see if we can fix this, or need any more info, don't hesitate to contact me. Regards... -- Manty/BestiaTester -> http://manty.net