Hi, > On Fri, Jun 18, 2021 at 09:55:38AM +0200, Geert Uytterhoeven wrote: > > This causes lots of failures like: > > > > arm/boot/dts/r8a7743-iwg20d-q7-dbcm-ca.dt.yaml: hdmi@39: > > 'avdd-supply' is a required property > > > > Should all supplies be required? > > Looking at the driver, missing supplies are automatically replaced by > > dummy regulators by the regulator framework. > > Generally speaking, I like DT bindings to be descriptive of the > hardware, and thus require power supplies that are needed for the device > to function, even if they are fixed supplies. > > This being said, I think there's also room to group some power supplies > together in the bindings, when they are not meant by the device to be > controlled separately. In this specific case, we also need to take into > account that the adv7511 and adv7533 have different supplies. Thanks for the review, guys. Yes, there were some dtbs check warnings to be expected, the consensus in a previous version of the patch was that that shouldn't be a blocker for a binding conversion and that the *.dts definitions should eventually be fixed to comply with the binding, which is, IMO, a more reasonable process to keep the binding conversion effort progressing. Cheers, Ricardo
