[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20161209002635.GD5423@codeaurora.org>
Date: Thu, 8 Dec 2016 16:26:35 -0800
From: Stephen Boyd <sboyd@...eaurora.org>
To: Kuninori Morimoto <kuninori.morimoto.gx@...esas.com>
Cc: Russell King - ARM Linux <linux@...linux.org.uk>,
Rob Herring <robh@...nel.org>,
Linux-ALSA <alsa-devel@...a-project.org>,
Linux-DT <devicetree@...r.kernel.org>,
Michael Turquette <mturquette@...libre.com>,
Linux-Kernel <linux-kernel@...r.kernel.org>,
Mark Brown <broonie@...nel.org>, linux-clk@...r.kernel.org,
Linux-ARM <linux-arm-kernel@...ts.infradead.org>
Subject: Re: [PATCH 3/3] ASoC: simple-card-utils: enable
clocks/clock-names/clock-ranges
On 12/09, Kuninori Morimoto wrote:
>
> Hi Stephen
>
> > > From: Kuninori Morimoto <kuninori.morimoto.gx@...esas.com>
> > >
> > > Current simple-card is supporting this style for clocks
> > >
> > > sound {
> > > ...
> > > simple-audio-card,cpu {
> > > sound-dai = <&xxx>;
> > > clocks = <&cpu_clock>;
> > > };
> > > simple-audio-card,codec {
> > > sound-dai = <&xxx>;
> > > clocks = <&codec_clock>;
> > > };
> > > };
> > >
> > > Now, it can support this style too, because we can use
> > > devm_get_clk_from_child() now.
> > >
> > > sound {
> > > ...
> > > clocks = <&cpu_clock>, <&codec_clock>;
> > > clock-names = "cpu", "codec";
> > > clock-ranges;
> > > ...
> > > simple-audio-card,cpu {
> > > sound-dai = <&xxx>;
> > > };
> > > simple-audio-card,codec {
> > > sound-dai = <&xxx>;
> > > };
> > > };
> > >
> > > Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@...esas.com>
> >
> > I don't see any reason why we need this patch though. The binding
> > works as is, so supporting different styles doesn't seem like a
> > good idea to me. Let's just keep what we have? Even if a sub-node
> > like cpu or codec gets more than one element in the clocks list
> > property, we can make that work by passing a clock-name then
> > based on some sort of other knowledge.
>
> OK, thanks. Let's skip this patch.
> But I believe this idea/method itself is not wrong (?)
>
Right it's not wrong, just seems confusing to have two methods.
--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project
Powered by blists - more mailing lists