On Fri, Apr 25, 2014 at 8:24 AM, Stephen Warren <swarren@xxxxxxxxxxxxx> wrote: > On 04/24/2014 12:22 PM, Thierry Reding wrote: > ... >> The downside of not allowing the gpiod API to support the -gpio suffix >> is that we'll never be able to convert drivers that use such a binding >> and will forever have a hodgepodge of GPIO APIs that we need to support. > > Perhaps rather than making the existing gpiod API automatically search > for both -gpios and -gpio, we could make a new API for the other suffix, > so that driver indicate explicitly which property name they want. That > way, someone can't accidentally write -gpio in the DT and have it still > work. Or, add a parameter to the existing API, but that's probably a lot > more churn. Hm, that is possible, I just worry that this will lead the DT and ACPI semantics to diverge even more, and the present patch make things more coherent from the framework side of things instead of even more elaborate per-HW-info-method :-/ Yours, Linus Walleij -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
