From mboxrd@z Thu Jan 1 00:00:00 1970 From: Liam Girdwood Subject: Re: soc-dsp programming model for loopbacks Date: Wed, 25 Jan 2012 17:48:57 +0000 Message-ID: <1327513737.11901.47.camel@odin> References: <007801ccdae0$381fcf40$a85f6dc0$@bossart@linux.intel.com> <20120124220036.GA11635@sirena.org.uk> <1327511545.11901.41.camel@odin> <20120125172051.GS3687@opensource.wolfsonmicro.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from na3sys009aog103.obsmtp.com (na3sys009aog103.obsmtp.com [74.125.149.71]) by alsa0.perex.cz (Postfix) with ESMTP id 322D0103BCB for ; Wed, 25 Jan 2012 18:51:23 +0100 (CET) Received: by wibhm4 with SMTP id hm4so3153128wib.15 for ; Wed, 25 Jan 2012 09:51:09 -0800 (PST) In-Reply-To: <20120125172051.GS3687@opensource.wolfsonmicro.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: alsa-devel-bounces@alsa-project.org Errors-To: alsa-devel-bounces@alsa-project.org To: Mark Brown Cc: alsa-devel@alsa-project.org, Pierre-Louis Bossart List-Id: alsa-devel@alsa-project.org On Wed, 2012-01-25 at 17:20 +0000, Mark Brown wrote: > On Wed, Jan 25, 2012 at 05:12:25PM +0000, Liam Girdwood wrote: > > On Tue, 2012-01-24 at 22:00 +0000, Mark Brown wrote: > > > > Note that the use of soc-dsp isn't required for on-SoC DSPs - it really > > > depends on how the hardware looks. If the DSP external format is very > > > strongly tied to the format used to DMA to and from main memory (or > > > other interfaces) then soc-dsp is what you need, if the DSP is able to > > > rewrite formats more flexibly then it should be more direct to model it > > > as a CODEC device with whatever number of DAIs and routing between them. > > > It's also needed if your DSP cant control or configure any of your host > > CPU DAI hardware controllers that it needs to physically use to route > > audio data. > > Depends how they're allocated - the DSP driver can always do the > management on behalf of the DSP firmware if the hardware is structured > in a way that makes that make sense (for example, if the CPU never talks > to the outside world directly). Wouldn't that either mean the DSP driver (re)implements some of the CPU DAI driver operations OR the DSP driver calling the CPU DAI driver ops directly ? Regards Liam