From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <devicetree-owner@vger.kernel.org>
Received: from mail-db8lp0185.outbound.messaging.microsoft.com ([213.199.154.185]:37029
	"EHLO db8outboundpool.messaging.microsoft.com" rhost-flags-OK-OK-OK-OK)
	by vger.kernel.org with ESMTP id S1752392Ab3HELdf (ORCPT
	<rfc822;devicetree@vger.kernel.org>); Mon, 5 Aug 2013 07:33:35 -0400
Date: Mon, 5 Aug 2013 19:28:05 +0800
From: Nicolin Chen <b42378@freescale.com>
Subject: Re: [PATCH v3 1/2] ASoC: fsl: Add S/PDIF CPU DAI driver
Message-ID: <20130805112804.GA6770@MrMyself>
References: <1375687781-23225-1-git-send-email-b42378@freescale.com>
 <1375687781-23225-2-git-send-email-b42378@freescale.com>
 <1375697222.4000.59.camel@pizza.hi.pengutronix.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Disposition: inline
In-Reply-To: <1375697222.4000.59.camel@pizza.hi.pengutronix.de>
Sender: devicetree-owner@vger.kernel.org
To: Philipp Zabel <p.zabel@pengutronix.de>
Cc: broonie@kernel.org, lars@metafoo.de, linuxppc-dev@lists.ozlabs.org, alsa-devel@alsa-project.org, devicetree@vger.kernel.org, timur@tabi.org, rob.herring@calxeda.com
List-ID: <devicetree@vger.kernel.org>

Hi Philipp,

   Thanks for the comments! I'll revise it and send a v4.
   Here is my reply to some parts of your question.

On Mon, Aug 05, 2013 at 12:07:02PM +0200, Philipp Zabel wrote:
> This looks to me like a case of configuration data in the device tree.
> Couldn't the tx/rx clock source be determined automatically or at least
> the SoC specific clock sources to the mux be known to the driver, so
> that we can use clock phandles here?
> 
> What happens if a different tx-clk-source is needed for 48 kHz than for
> 44.1 kHz?

You are right about the case. I'll try to figure out a nicer way, including
for the different sample-rate.


> Do we already have anything for this in the kernel?
> It could also be done using:
> 
> 	u8 tmp = input;
> 
>         tmp = ((tmp & 0b10101010) >> 1) | ((tmp << 1) & 0b10101010);
>         tmp = ((tmp & 0b11001100) >> 2) | ((tmp << 2) & 0b11001100);
>         tmp = ((tmp & 0b11110000) >> 4) | ((tmp << 4) & 0b11110000);
> 
> 	return tmp;

I was also wondering this question but I couldn't find it.
And obviously your algorithm is better :)


> In principle, it should also be possible to let the SDMA engine only
> take the FIFOs' MSBs for S16_LE. Is this a limitation of the SDMA ROM
> script?

You are right about the idea and the inference.