ad7124 adc driver refactoring

[trik]

Hi everyone. Writing some code to read a PT100 thermocouple via an AD7124 ADC, I was digging in the driver code and made some discoveries:

• The zephyr,differential property on channels is ignored. Bipolarity can be set using the bipolar-mask property on the device node
  

  zephyr/drivers/adc/adc_ad7124.c

  Line 418 in 74056a7

  new_cfg->props.bipolar = config->bipolar_mask & BIT(cfg->channel_id);

  
  The bipolar-mask property could be dropped in favor of the channel configuration be me more compliant to the ADC api

• Since AD7124 has two programmable excitation current sources, we could add two properties to set the current sources values (can be different) and enable the ADC_CONFIGURABLE_EXCITATION_CURRENT_SOURCE_PIN config

• AD7124 provides two GPIO output pins, it could be useful to have a GPIO driver.

If it sounds reasonable, i could take in charge.

[ttmut]

cc @yasinustunerg

[yasinustunerg]

About your first point — when you suggest dropping the bipolar-mask property, could you clarify how you plan to handle bipolar mode instead? Since zephyr,differential typically refers to input configuration rather than voltage encoding, do you see them as related, or would you propose a separate way to handle bipolar settings?

[trik]

ok, forget the number 1, i misunderstood the unipolar/bipolar configuration of ad7124. what about number 2? in addition, we could add also a post filter configuration

[yasinustunerg]

Could you open a PR for the remaining parts? These points can be discussed through the PR.

[trik]

here it is #91136
it's just for the current sources, the post filter settings could be a little bit complicated cause just a bit for channel won't be enough. if have any suggestion on this..

one more thing. at the moment we can set the ADC mode but if we use something diffrent from the default AD7124_CONTINUOUS we won't get any sample. maybe we could check the ADC mode at the beginning of the read function and temporarily switch to AD7124_SINGLE

[ttmut]

Hi @trik, since #91136 has been merged, can you close this discussion as resolved, unless you are planning for further additions?

[trik]

Sure