On Fri, Jan 17, 2025 at 3:52 PM Andre Przywara <andre.przywara@xxxxxxx> wrote: > That is actually a good argument: At the moment I am happy with my > proposal (the allwinner,pinmux = <number>; property), but that seems like > standard #15 then. > So would biting the bullet and adopting the Apple/STM32 way then be more > sustainable? I suppose... > On the other hand: the allwinner,pinmux solution has the advantage of being > already written and proven working, also it stays very close to the > existing description/binding - so implementations like U-Boot could just > keep on using the "function" string. > > I am a bit torn here... I don't think I will find the solitude to > implement this "Apple" approach in the next few weeks. I think whatever the Allwinner maintainers agree is the best should be what you go for. It is a lot of hobbyist maintainers in this space and for them the bar should be lower. If you have buy-in from the other maintainers, then go for that solution. Yours, Linus Walleij