[PATCH v4 11/12] ASoC: add bindings for stm32 DFSDM filter

[Arnaud Pouliquen]

Hello Rob,

Thanks for the Review,


On 11/15/2017 04:43 PM, Rob Herring wrote:
> On Thu, Nov 09, 2017 at 11:12:33AM +0100, Arnaud Pouliquen wrote:
>> Add bindings that describes audio settings to support
>> Digital Filter for pulse density modulation(PDM) microphone.
>> 
>> Signed-off-by: Arnaud Pouliquen <arnaud.pouliquen at st.com>
>> ---
>> V3 -> V4 changes:
>>        - Update to move on of_graph description.
>>        - Link to DFSDM IIO bindings.
>> 
>>  .../devicetree/bindings/sound/st,stm32-adfsdm.txt  | 63 ++++++++++++++++++++++
>>  1 file changed, 63 insertions(+)
>>  create mode 100644 Documentation/devicetree/bindings/sound/st,stm32-adfsdm.txt
>> 
>> diff --git a/Documentation/devicetree/bindings/sound/st,stm32-adfsdm.txt b/Documentation/devicetree/bindings/sound/st,stm32-adfsdm.txt
>> new file mode 100644
>> index 0000000..75e298b
>> --- /dev/null
>> +++ b/Documentation/devicetree/bindings/sound/st,stm32-adfsdm.txt
>> @@ -0,0 +1,63 @@
>> +STMicroelectronics audio DFSDM DT bindings
>> +
>> +This driver supports audio PDM microphone capture through Digital Filter format
> 
> Bindings aren't drivers.
Yes, sorry very bad wording, i will change sentence.
> 
>> +Sigma Delta modulators (DFSDM).
>> +
>> +Required properties:
>> +  - compatible: "st,stm32h7-dfsdm-dai".
>> +
>> +  - #sound-dai-cells : Must be equal to 0
>> +
>> +  - io-channels : phandle to iio dfsdm instance node.
>> +             [See: ../iio/adc/st,stm32-dfsdm-adc.txt for DFSDM options]
>> +
>> +Example of a sound card using audio DFSDM node.
>> +
>> +     sound_card {
>> +             compatible = "audio-graph-card";
>> +
>> +             dais = <&cpu_port>;
>> +     };
>> +
>> +     dfsdm: dfsdm at 40017000 {
>> +             compatible = "st,stm32h7-dfsdm";
>> +             reg = <0x40017000 0x400>;
>> +             clocks = <&rcc DFSDM1_CK>;
>> +             clock-names = "dfsdm";
>> +             #address-cells = <1>;
>> +             #size-cells = <0>;
>> +
>> +             dfsdm_adc0: dfsdm-adc at 0 {
>> +                     compatible = "st,stm32-dfsdm-dmic";
>> +                     reg = <0>;
>> +                     interrupts = <110>;
>> +                     dmas = <&dmamux1 101 0x400 0x00>;
>> +                     dma-names = "rx";
>> +                     st,adc-channels = <1>;
>> +                     st,adc-channel-names = "dmic0";
>> +                     st,adc-channel-types = "SPI_R";
>> +                     st,adc-channel-clk-src = "CLKOUT";
>> +                     st,filter-order = <5>;
>> +             };
>> +     }
>> +
>> +     dfsdm_dai0:dfsdm_dai at 0 {
> 
> Unit address without reg property is not valid. Building your dtb with
> W=2 will tell you this.
> 
> Need a space after the ':'.
> 
> Should be a child of dfsdm?

> 
>> +             compatible = "st,stm32h7-dfsdm-dai";
>> +             #sound-dai-cells = <0>;
>> +             io-channels = <&dfsdm_adc0 0>;
> 
> It doesn't seem like this is an actual h/w block. Why can't 'dais' point
> directly to the ADC?DFSDM is not simple to describe in DT for audio.
Assumption is that  audio feature should be part of the ASoc framework.
This means that we expect to  use of_graph based on a DT description.
of_graph should point to the DAI device. In this case we need to
describe a DAI Alsa device ( associated compatible driver in ASoC):
dfsdm_dai0.
This DAI device is a client of the IIO device dfsdm_adc0 using in-kernel
iio consumer interface, so points to it using io-channels property.

Now as you propose above, i can describe DAI device as a child of the
DFSDM ADC device like this:


	dfsdm: dfsdm at 40017000 {
		compatible = "st,stm32h7-dfsdm";
		reg = <0x40017000 0x400>;
		clocks = <&rcc DFSDM1_CK>;
		clock-names = "dfsdm";
		#address-cells = <1>;
		#size-cells = <0>;

		dfsdm_adc0: dfsdm-adc at 0 {
			compatible = "st,stm32-dfsdm-dmic";
			reg = <0>;
			interrupts = <110>;
			dmas = <&dmamux1 101 0x400 0x00>;
			dma-names = "rx";
			st,adc-channels = <1>;
			st,adc-channel-names = "dmic0";
			st,adc-channel-types = "SPI_R";
			st,adc-channel-clk-src = "CLKOUT";
			st,filter-order = <5>;
			
			dfsdm_dai0: dfsdm-dai {
				compatible = "st,stm32h7-dfsdm-dai";
				#sound-dai-cells = <0>;
				io-channels = <&dfsdm_adc0 0>;
				cpu_port: port {
				dfsdm_endpoint: endpoint {
					remote-endpoint = <&dmic0_endpoint>;
				};
			};
		};
	};

Is is something that would sound better for you?

Regards,

Arnaud
> 
>> +             cpu_port: port {
>> +                     dfsdm_endpoint: endpoint {
>> +                             remote-endpoint = <&dmic0_endpoint>;
>> +                     };
>> +             };
>> +     };
>> +
>> +     dmic0: dmic at 0 {
>> +             compatible = "dmic-codec";
>> +             #sound-dai-cells = <0>;
>> +             port {
>> +                     dmic0_endpoint: endpoint {
>> +                             remote-endpoint = <&dfsdm_endpoint>;
>> +                     };
>> +             };
>> +     };
>> -- 
>> 2.7.4
>>