> IMHO, I think the solution you proposed is similar in term of complexity
> to the RZN1_ADC_NO_CHANNEL approach. On my side, I would prefer the
> RZN1_ADC_NO_CHANNEL approach to keep the decoupling between IIO chan and
> ADC core chans.
> 
> That's said, I am still open to move in your direction if you still think
> it is more relevant than the RZN1_ADC_NO_CHANNEL approach. Just tell me.

Well, in deed, I like "my" approach a tad better, but I am not demanding
it. It is your driver and you have reasons to do it like you implemented
it - you chose the way, I am fine with both. But maybe add a comment
(mention the decoupling) why it was decided this way.