Le jeu. 10 oct. 2024 à 20:22, Jonathan Cameron <jic23@xxxxxxxxxx> a écrit : > > On Tue, 8 Oct 2024 09:52:50 +0200 > Krzysztof Kozlowski <krzk@xxxxxxxxxx> wrote: > > > On Mon, Oct 07, 2024 at 05:45:45PM +0200, Julien Stephan wrote: > > > ad7380-4 is the only device from ad738x family that doesn't have an > > > internal reference. Moreover its external reference is called REFIN in > > > the datasheet while all other use REFIO as an optional external > > > reference. If refio-supply is omitted the internal reference is > > > used. > > > > > > Fix the binding by adding refin-supply and makes it required for > > > ad7380-4 only. > > > > Maybe let's just use refio as refin? Reference-IO fits here well. > > Otherwise you have two supplies for the same. > Whilst it is ugly, the effort this series is going to in order > to paper over a naming mismatch makes me agree with Krzysztof. > > I think adding a comment to the dt-binding would be sensible > though as people might fall into this hole. > Hi Jonathan and Krzysztof, I am currently adding support for another chip to this family (ADAQ4380-4) and it also uses REFIN.. but in another way ad7380-4 does.. So: - ad7380-4 does not have any internal reference and use a mandatory refin supply as external reference - adaq4380-4 does not have external reference but uses a 3V internal reference derived from a 5V mandatory refin supply - all others (AFAIK) use an optional refio external reference. If omitted, use an internal 2.5V reference. I am not sure using a single refio-supply for all will make things clearer.. What do you think? Should I also send the adaq series now to bring more context? (I wanted feedback on this series first). Cheers Julien > Other than the missing ret =, rest of series looks fine to me > > Jonathan > > > > > Best regards, > > Krzysztof > > >
