From mboxrd@z Thu Jan 1 00:00:00 1970 From: Pierre-Louis Bossart Subject: Re: [PATCH] ALSA: compress: Add DSP private metadata interface Date: Tue, 23 Feb 2016 08:39:31 -0600 Message-ID: <56CC6F23.6090705@linux.intel.com> References: <1456163671-40987-1-git-send-email-tim.sheridan@imgtec.com> <20160223094223.GK17690@localhost> <98E3058E5505294895CA69F176CAD6626A6D8BE1@LEMAIL01.le.imgtec.org> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; Format="flowed" Content-Transfer-Encoding: 7bit Return-path: Received: from mga02.intel.com (mga02.intel.com [134.134.136.20]) by alsa0.perex.cz (Postfix) with ESMTP id 8CFB4260535 for ; Tue, 23 Feb 2016 15:39:36 +0100 (CET) In-Reply-To: <98E3058E5505294895CA69F176CAD6626A6D8BE1@LEMAIL01.le.imgtec.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: alsa-devel-bounces@alsa-project.org Sender: alsa-devel-bounces@alsa-project.org To: Tim Sheridan , 'Vinod Koul' , Takashi Iwai Cc: "alsa-devel@alsa-project.org" , Neil Jones List-Id: alsa-devel@alsa-project.org > AXD's got various features which require mode setting (e.g. for > selecting between multi-device synchronization mechanisms that it > supports). This would need an enum control, but I guess that this > could be fudged with mutually exclusive switch controls for each of > the enum values. AXD's also got numeric configuration parameters > which it needs to be able to receive (e.g. parameters for stream > synchronization heuristics, audio filter coefficients, etc.). Maybe a > volume control with a domain wide enough could be used for this kind > of configuration parameter. It's rather horrible, but might be fine > for while the AXD driver is out of tree and there aren't more > appropriate ALSA control types (e.g. enum and non-volume integer). > :-/ I am not too sure I understand what you refer to as synchronization or multi-device synchronization mechanisms. Since you mentioned you only need an initial time, I wonder if this is conceptually related to the start_at() functionality discussed last year for PCM? If you have a set of implementation-defined mechanisms related to algorithms, why not use an ALSA binary control where you define you own syntax and content and use the alsa control layer as a 'dumb' pipe between user and kernel. That's what everyone does for DSP control.