Re: [PATCH v3 2/5] ARM: dts: add reference voltage property for MXS LRADC

[Hector Palacios <hector.palacios@digi.com>]

Dear Alexandre,

On 08/22/2013 12:13 AM, Alexandre Belloni wrote:
> Hi Pawel,
>
> On 14/08/2013 16:44, Pawel Moll wrote:
>> On Tue, 2013-08-13 at 22:23 +0100, Jonathan Cameron wrote:
>>> On 07/22/13 15:04, Hector Palacios wrote:
>>>> Some LRADC channels have fixed pre-dividers so they can measure
>>>> different voltages at full scale. The reference voltage allows to
>>>> expose a scaling attribute through the IIO sysfs so that a user can
>>>> compute the real voltage out of a measured sample value.
>>>>
>>>> diff --git a/Documentation/devicetree/bindings/staging/iio/adc/mxs-lradc.txt b/Documentation/devicetree/bindings/staging/iio/adc/mxs-lradc.txt
>>>> index 4688205..6ec485c 100644
>>>> --- a/Documentation/devicetree/bindings/staging/iio/adc/mxs-lradc.txt
>>>> +++ b/Documentation/devicetree/bindings/staging/iio/adc/mxs-lradc.txt
>>>> @@ -1,9 +1,12 @@
>>>>   * Freescale i.MX28 LRADC device driver
>>>>
>>>>   Required properties:
>>>> -- compatible: Should be "fsl,imx28-lradc"
>>>> +- compatible: "fsl,imx28-lradc", "fsl,imx23-lradc"
>>>>   - reg: Address and length of the register set for the device
>>>>   - interrupts: Should contain the LRADC interrupts
>>>> +- fsl,vref: Reference voltage (in mV) for each LRADC channel. This is the
>>>> +	    maximum voltage that can be measured at full scale in each channel
>>>> +	    considering fixed pre-dividers.
>>
>> So, let me try to rephrase what I read above.
>>
>> There's an ADC with X channels. And there's a reference voltage source
>> (one?). Now, each of the ADC channels have a (different?) voltage
>> divider, taking the voltage from the reference source and feeding it to
>> the ADC comparator. How much am I wrong?
>>
>
> You are not so wrong. There is indeed actually only one reference
> voltage (and that is 1.85V). But, before feeding the voltage to the ADC
> channels, you sometimes have a divider. Then, after the channel muxing,
> you can add a by 2 divider.
>
> Mandatory ascii art:
>
>              +-----+
>              |     |
>     +-ch1--->|     |
>              |     |
>              |     |
>              |     |     +-----+
>     +-ch2--->|     |     |     |
>              | MUX |++-->| ADC +----------->
>       ch3    |     | |   |     |
>      +----+  |     | |   +-----+
>      |    |  |     | |      |
>    +-> :4 +->|     | |  +---+--+
>      |    |  |     | |  |      |
>      +----+  |     | +->|  :2  |
>              +-----+    |      |
>                         +------+
>
>
>> If I'm not wrong at all, I'd say that the reference source could be
>> described as a standard fixed regulator
>> (Documentation/devicetree/bindings/regulator/fixed-regulator.txt) and
>> the ADC node should have some king of "reference-supply" phandle to the
>> regulator node. Now, if the dividers factors are *really* fixed, the
>> driver could know about them and calculate the effective reference
>> voltage on its own, couldn't it?
>>
>> Let me repeat the "DT standard disclaimer": the tree, in general, should
>> describe the way components are *wired up*, not much more.
>>
>
> So, from my point of view, the divider that is before the mux (the by 4
> divider on channel 3 on my drawing) is not part of the the ADC, it is
> not fixed by that IP. And indeed, that changed between the i.mx23 and
> i.mx28 while the IP is the same.

The dividers only make sense and affect the ADC, so whether they should be considered 
part of the ADC IP or not is a philosophical question.
In my opinion, the different dividers between the i.mx23 and i.mx28 channels are the 
kind of hardware differences that fit nicely in the DeviceTree, describing the hardware.

> So, the two solutions you suggest are:
> 1/ using a fixed-regulator phandle per channel

Since the dividers only affect and have meaning on the ADC channels, creating a 
regulator for each channel that has a different divider looks to me like an overworked 
solution. These are not real voltage sources. They are just indicators of the maximum 
reference voltage that an ADC channel can measure.

> 2/ hard-coding the dividers in the driver using the compatible string to
> know which divider is on which channel.
>
> I feel that solution 2 is less future proof but at the same time, I
> don't believe we will see that IP in another chip in the future.

This was what I originally submitted but it then looked like it would better fit in 
the DeviceTree. The spear-adc seemed to use a similar approach:

http://permalink.gmane.org/gmane.linux.kernel.iio/7994

Best regards,
--
Hector Palacios