On Wed, Oct 16, 2024 at 8:59 PM Rob Herring <robh@xxxxxxxxxx> wrote: > This property has shortcomings if you want to prevent exposing > non-existent GPIOs to userspace. You really need a mask because what > if GPIO0 doesn't exist? That does happen in $OS (DT hat on), a famous OS has such sparse masks for GPIO chips. They have not become DT bindings (yet). As it turns out they are mostly used for cases where firmware/BIOS/secure world want to "steal" some GPIOs out of the pool and as such has nothing to do with hardware description. It has been shown that in practice what HW engineers do is line up as many IP instances they need and just truncate the last one, rendering the upper bits unused. They are practical and least effort-oriented people who have yet to be seen doing something crazy like e.g. connect every second bit to an actual pin. > > It is also reflected in several of the new yaml bindings, a git grep > > ngpios will show you. > > Yes, I know. You can also find push back on using it. > > Anyway, I did my push back and what's one more user (sigh), so: > > Acked-by: Rob Herring (Arm) <robh@xxxxxxxxxx> Thanks Rob, some push back is always good! Yours, Linus Walleij