Re: [PATCH] ASoC: snd-soc-dummy: add Device Tree support

From: Judy Hsiao <judyhsiao@chromium.org>
To: Mark Brown <broonie@kernel.org>
Cc: Liam Girdwood <lgirdwood@gmail.com>,
	Rob Herring <robh+dt@kernel.org>,
	Jaroslav Kysela <perex@perex.cz>, Takashi Iwai <tiwai@suse.com>,
	Douglas Anderson <dianders@chromium.org>,
	Dylan Reid <dgreid@chromium.org>,
	Jimmy Cheng-Yi Chiang <cychiang@google.com>,
	Judy Hsiao <judyhsiao@google.com>,
	Tzung-Bi Shih <tzungbi@chromium.org>,
	Stephen Boyd <swboyd@chromium.org>,
	linux-arm-kernel@lists.infradead.org,
	linux-arm-msm@vger.kernel.org, devicetree@vger.kernel.org,
	alsa-devel@alsa-project.org
Subject: Re: [PATCH] ASoC: snd-soc-dummy: add Device Tree support
Date: Wed, 23 Jun 2021 22:29:15 +0800	[thread overview]
Message-ID: <CAJXt+b_15SEFW9ztg=CaYHfVJk-q+u46eLVgdPm+2ugjwajc1g@mail.gmail.com> (raw)
In-Reply-To: <20210622162307.GF4574@sirena.org.uk>

On Wed, Jun 23, 2021 at 12:23 AM Mark Brown <broonie@kernel.org> wrote:
>
> On Wed, Jun 23, 2021 at 12:10:53AM +0800, Judy Hsiao wrote:
>
> > Thanks for your review comment.
> > This patch is used to support multi-channel where we want one codec to
> > control the only GPIO shared by 4 amps.
>
> So you've got 4 instances of the same CODEC?  Then I'd expect to see
> those all individually represented in DT.  Or if there's a single
> physical CODEC then I'm not sure what the dummies are for?
>
> > In snd_soc_runtime_calc_hw(), by creating dummy codecs that share a
> > DAI link with a real codec:
> >   1. The min/ max channel of  CPU DAI will be directly adopted.
> >   2. The formats and sample rates of the DAI link will be determined
> > by the real codec unless the real codec supports the rate
> >       and format that do not intersect with the rate and format of
> > snd-soc-dummy.
> > That is the reason why we don’t specify the format and sample rates of
> > the dummy codec with the real codec determining the properties .
>
> It's not clear to me why you'd not just describe the actual CODECs here
> rather than using a dummy CODEC, the fact that the dummy CODEC is doing
> what you want is just an accident of the implementation rather than a
> description of the hardware.

Thanks for your inputs. Specifying four codes for the multi-channel works fine.
We have not thought of specifying four codes before as we want to avoid loading
the codec driver multiple times, but actually loading the
snd-soc-dummy just has the
similar cost. By specifying four codes, the dtsi file describes the
real hardware schematic.
I will specify four codec in the dtsi file to support the four channel
use case and this
snd-soc-dummy patch is not needed. Thanks for the discussion!