Hi Krzysztof, On Mon, 2025-03-31 at 09:34 +0200, Krzysztof Kozlowski wrote: > On Fri, Mar 28, 2025 at 01:28:47PM +0000, André Draszik wrote: > > allOf: > > + - if: > > + properties: > > + compatible: > > + contains: > > + const: samsung,s2mpg10-pmic > > + then: > > + properties: > > + reg: false > > + samsung,s2mps11-acokb-ground: false > > + samsung,s2mps11-wrstbi-ground: false > > + > > + oneOf: > > + - required: [interrupts] > > + - required: [interrupts-extended] > > Drop, you should require only interrupts. As mentioned in the commit message comments, it doesn't work with just interrupts. It appears that dtschema's fixups.py doesn't handle this case. With just interrupts, DT validation will fail if the DT uses interrupts-extended. There was at least one other binding that specified interrupts in the same way, so I went with the same approach. > OTOH, why regulators subnode is not needed? Commit msg mentions they > exist, so they should be required. Binding does not change because you > added or did not add yet some driver support. I wanted to avoid DT validation errors, because we haven't started working on regulators yet, and it might take a little while until everything is in place. I'll make it required in the next version. Thanks! Andre'
