From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.gmx.net (mail.gmx.net [213.165.64.20]) by ozlabs.org (Postfix) with SMTP id 40411DDEE2 for ; Tue, 6 May 2008 17:51:27 +1000 (EST) Content-Type: text/plain; charset="us-ascii" Date: Tue, 06 May 2008 09:51:15 +0200 From: "Gerhard Pircher" In-Reply-To: <1210031058.21644.120.camel@pasglop> Message-ID: <20080506075115.269150@gmx.net> MIME-Version: 1.0 References: <6DA6B1DA-11B5-4B89-88AD-561B4D17B33B@kernel.crashing.org> <20080505195054.172000@gmx.net> <1210031058.21644.120.camel@pasglop> Subject: Re: [PATCH] Sam440ep support To: benh@kernel.crashing.org Cc: Takashi Iwai , linuxppc-dev@ozlabs.org, cjg@cruxppc.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , -------- Original-Nachricht -------- > Datum: Tue, 06 May 2008 09:44:18 +1000 > Von: Benjamin Herrenschmidt > An: Gerhard Pircher > CC: Kumar Gala , cjg@cruxppc.org, linuxppc-dev@ozlabs.org > Betreff: Re: [PATCH] Sam440ep support > > On Mon, 2008-05-05 at 21:50 +0200, Gerhard Pircher wrote: > > This is a (bad) hack that I also use on the AmigaOne to get the ALSA > > sound drivers working with DMA, because ALSA doesn't work with > > dma-noncoherent.c. The problem is the "nopage" mechanism, which fails > > with non coherent DMA allocations due to their own virtual address > > space (correct me, if I'm wrong). > > > > See this thread for more info: > > http://readlist.com/lists/vger.kernel.org/linux-kernel/45/226541.html > > > > This is a general problem that affects all powerpc boards that use > > dma-noncoherent.c with ALSA PCI drivers. > > The link above doesn't provide any useful information on the problem and > it contains itself a non working link... Sorry, I didn't check the embedded link. This one should work: http://www.uwsg.iu.edu/hypermail/linux/kernel/0406.2/0801.html IIRC the problem is the mmaping of non coherent DMA allocations, as you already know. The link above points to a very old (from 2004) and quite long thread where the correct DMA API for mmaping DMA allocations was discussed. > Can somebody explains exactly what's going on ? That shouldn't be hard > to fix. I can't believe the problem has been around for 2 years and > nobody actually bothered fixing it properly. Takashi Iwai posted a preliminary patch a long time ago. I tested it on my machine and it failed with non coherent scatter-gather DMA allocations (I guess almost all ALSA PCI drivers use SG DMA?). Thanks! Gerhard -- GMX startet ShortView.de. Hier findest Du Leute mit Deinen Interessen! Jetzt dabei sein: http://www.shortview.de/?mc=sv_ext_mf@gmx