From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Ujfalusi Subject: Re: Wolfson 5102 on Beaglebone Black using ALSA Date: Mon, 16 May 2016 15:33:55 +0300 Message-ID: <47b8f68c-12be-adb5-1e29-7ca19feac581@ti.com> References: <4BEB79E8-7315-48DF-BBB6-E944114621BF@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset="windows-1252" Content-Transfer-Encoding: quoted-printable Return-path: Received: from arroyo.ext.ti.com (arroyo.ext.ti.com [198.47.19.12]) by alsa0.perex.cz (Postfix) with ESMTP id 5D7E9266B4A for ; Mon, 16 May 2016 14:33:59 +0200 (CEST) In-Reply-To: <4BEB79E8-7315-48DF-BBB6-E944114621BF@gmail.com> 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: Clark Dunson , alsa-devel@alsa-project.org List-Id: alsa-devel@alsa-project.org On 05/15/16 01:16, Clark Dunson wrote: > Hey everybody! > = > We have been working on integrating the Wolfson 5102 codec with the > Beaglebone Black. We=92ve made pretty decent progress, got the arizona c= ode > in kernel 4.4 registering the wm5102 as ALSA sound card (both player and > recorder), have McAsp0 as 24.576 MHz clock source, and have 12 IRQs showi= ng > up in /proc/interrupts. > = > We have been working with sound{compatible =3D "simple-audio-card=94}, bu= t now > have hit our first real question. sysclk is hard coded to =910=92 under > simple-audio-card, and we had to hack the kernel and change that to =911= =92 to > achieve ALSA sound card registration. We=92d all like for this work to > proceed without having to hack the kernel, and wondered if one of you mig= ht > recommend another approach or selection for device tree sound{compatible} > that might work out better? For simple card you would need: http://mailman.alsa-project.org/pipermail/alsa-devel/2016-February/104316.h= tml to be able to select the McASP clocks. It is not going to apply on linux-next becasue of: http://mailman.alsa-project.org/pipermail/alsa-devel/2016-May/107783.html but it is not should not be a big issue to resolve (one patch to move the clkid to binding header). The davinci-mcasp driver will remain broken in this sense for a foreseen future as we need to convert the daVinci architecture to CCF first to be ab= le to progress with the clock selection. -- = P=E9ter