Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756650AbXJ2Tqj (ORCPT ); Mon, 29 Oct 2007 15:46:39 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754100AbXJ2Tmn (ORCPT ); Mon, 29 Oct 2007 15:42:43 -0400 Received: from caffeine.csclub.uwaterloo.ca ([129.97.134.17]:59363 "EHLO caffeine.csclub.uwaterloo.ca" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755153AbXJ2Tmm (ORCPT ); Mon, 29 Oct 2007 15:42:42 -0400 Date: Mon, 29 Oct 2007 15:42:41 -0400 To: Wijnand Rietman Cc: linux-kernel@vger.kernel.org Subject: Re: CONFIG_IDEDMA_AUTO & 2.4.32 & hdparm -d1 Message-ID: <20071029194241.GA27646@csclub.uwaterloo.ca> References: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.13 (2006-08-11) From: lsorense@csclub.uwaterloo.ca (Lennart Sorensen) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1665 Lines: 38 On Mon, Oct 29, 2007 at 02:05:44PM +0100, Wijnand Rietman wrote: > I am running Linux on a PC104 board from a compact flash card. Because > the physical DMA lines are missing on some of the compact flash > carrier cards, I would like to be able to boot with DMA disabled and > to enable DMA in a later stage if desired (and supported by hardware). > > The problem is that when I set CONFIG_IDEDMA_AUTO to FALSE, I cannot > enable DMA anymore with hdparm, not even with hardware that supports > DMA. If I set CONFIG_IDEDMA_AUTO to TRUE, I can use hdparm to disable > and enable DMA (but of course I get long timeouts with the carrier > cards that do not have the physical DMA lines). > > Q1: Is this how the CONFIG_IDEDMA_AUTO setting is supposed to work in > combination with hdparm? > Q2: Is there a way to boot with DMA disabled and to enable it later on > once the system is up and running? > > Kernel: 2.4.32 > Chipset: CS5530 > > Any help is highly appreciated! I had trouble with DMA and compact flash cards on our board (prior to the new revision) since it did not have the DMA lines (they weren't part of the CF spec when the board was designed). The way I booted was to pass 'ide=nodma' and that worked fine (at least on the sc1200). I could enable dma with hdparm later if I wanted (which of course made the ide driver rather upset when it tried to then use the non existant dma lines). -- Len Sorensen - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/