[PATCH 3/7] IIO: add bindings for STM32 DFSDM ADC driver

[Arnaud Pouliquen]

Hi Jonathan,

On 01/29/2017 01:42 PM, Jonathan Cameron wrote:
> On 29/01/17 11:58, Jonathan Cameron wrote:
>> On 23/01/17 16:32, Arnaud Pouliquen wrote:
>>> This patch adds documentation of device tree bindings for the
>>> STM32 DFSDM ADC.
>>>
>>> Signed-off-by: Arnaud Pouliquen <arnaud.pouliquen at st.com>
>> Trivial bits inline.
>>> ---
>>>  .../bindings/iio/adc/st,stm32-dfsdm-adc.txt        | 60 ++++++++++++++++++++++
>>>  1 file changed, 60 insertions(+)
>>>  create mode 100644 Documentation/devicetree/bindings/iio/adc/st,stm32-dfsdm-adc.txt 
>>>
>>> diff --git a/Documentation/devicetree/bindings/iio/adc/st,stm32-dfsdm-adc.txt  b/Documentation/devicetree/bindings/iio/adc/st,stm32-dfsdm-adc.txt 
>>> new file mode 100644
>>> index 0000000..c156bcb
>>> --- /dev/null
>>> +++ b/Documentation/devicetree/bindings/iio/adc/st,stm32-dfsdm-adc.txt 	
>>> @@ -0,0 +1,60 @@
>>> +STMicroelectronics STM32 DFSDM ADC device driver
>>> +
>>> +STM32 DFSDM ADC is a sigma delta analog-to-digital converter.
>>> +It has to be declared in device-tree as a subnode of the DFSDM mfd node.
>>> +
>>> +It has several multiplexed input channels. Conversions can be performed
>>> +in single, scan or discontinuous mode. Conversions can be launched in software
>>> +or using hardware triggers.
>>> +Each instance of the driver uses one filter instance handle by the DFSDM mfd
>>> +driver.
>>> +
>>> +DFSDM also offers extra features:
>>> +-The analog watchdog feature allows the application to detect if the
>>> + input voltage goes beyond the user-defined, higher or lower thresholds.
>>> +-The short circuit detection allows allows the application to detect if the
>>> + input is in CC.
>>> +-The clock absence detection allows application to detect if SPI input is clocked.
>>> +
>>> +Required properties:
>>> +- compatible:	Must be "st,stm32-dfsdm-adc".
>>> +- reg:		Specifies the DFSDM filter instance.
>>> +- #io-channel-cells = <1>: See the IIO bindings section "IIO consumers".
>>> +- st,adc-channels:	List of single-ended channels muxed for this ADC.
>>> +- st,adc-channel-names:	List of single-ended channels Name.
>>> +
>>> +Optional properties:
>>> +- st,adc-channel-types:	Single-ended channel input type. Default value is 0.
>>> +			- 0: SPI with data on rising edge (default)
>>> +			- 1: SPI with data on falling edge
>> This is an element that can be described without magic numbers so I'd
>> prefer that you do so.
>>
>> Also spell check edage -> edge.
>>
>>> +			- 2: manchester codec, rising edage = logic 0
>>> +			- 3: manchester codec, rising edage = logic 1
>>> +- st,adc-channel-clk-src: Conversion clock source. default value is 1.
>>> +			  - 0: External SPI clocl (CLKIN x)
>>> +			  - 1: internal SPI clock (CLKOUT) (default)
>>> +			  - 2: internal SPI clock divided by 2 (falling edge).
>>> +			  - 2: internal SPI clock divided by 2 (rising edge).
>> 3?
>>> +- st,adc-alt-channel: 	  must be defined if Two ADCs are connected on same SPI
>>> +			  input.
>>> +			  If not set channel n is connected to SPI input n.
>>> +			  If set channel n is connected to SPI input n + 1.
>> ? Two data inputs with shared clock?
>>> +
>>> +Example:
>>> +	dfsdm: dfsdm at 4400D000 {
> I'd like to actually see the ADCs on the end of these explicitly
> described as well.  It's entirely possible we'll have something
> that isn't simply a dumb sigma delta modulator that will need
> it's own driver.
> 
> In some senses what we actually have here is a data bus so we
> should be able to describe things on it in that fashion
> rather than just describing the data that is in coming.
> Our boundary of knowledge isn't at the edge of this device!
> 
> A quick example google fed me would be the gain control on a
> maxim ds8102.
> 
> Tricky bit here will be defining this nice and generally when we
> currently only have the one piece of hardware to do it with ;)
> 
> That's what makes kernel development fun!
> 

What you describe here seems similar to audio codecs device defined in
ASoC. An Audio codec is an external chip that is connected to SoC
through an audio bus.
ASoC allows to bind this codec with the SOC DAI (digital Audio Interface).

FYI we should have this kind of feature to support on STM32 uin second step:
DSFDM input can also process stm32 ADC output through an hardware pipe
(stm32-adc IIO driver that has been recently integrated in kernel).
Internally, to pipe the ADC with DFSDM, we use an extension of inkern
interface that is very similar with the one mentioned by Lars:

https://github.com/analogdevicesinc/linux/blob/xcomm_zynq/drivers
/iio/buffer/hw_consumer.c

Don't know if this is what Lars has in mind... but a solution could be
to define a new type of IIO that would be something like
"postprocessing" or "filter". This new iio device type could be piped to
a frontend IIO device to export a new one (through hw_consumer.c interface).
In this case for simple device like PDM audio microphone that does not
required control a kind of dummy frontend device could be defined...

Regards
Arnaud

>>> +		iio_dfsdm0: iio-dfsdm at 0 {
>>> +			compatible = "st,stm32-dfsdm-adc";
>>> +			#io-channel-cells = <1>;
>>> +			reg = <0>;
>>> +			st,adc-channels = <1>;
>>> +			st,adc-channel-names = "in0";
>>> +		};
>>> +		iio_dfsdm1: iio-dfsdm at 1 {
>>> +			compatible = "st,stm32-dfsdm-adc";
>>> +			#io-channel-cells = <1>;
>>> +			reg = <1>;
>>> +			st,adc-channels = <1>;
>>> +			st,adc-channel-names = "in1";
>>> +			st,adc-channel-types = <1>;
>>> +			st,adc-alt-channel = <1>;
>>> +		};
>>> +	};
>>>
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-iio" in
>> the body of a message to majordomo at vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>