On Mon, Oct 20, 2014 at 8:27 AM, Alexandre Courbot <gnurou@xxxxxxxxx> wrote: > On Mon, Oct 20, 2014 at 2:31 PM, John Crispin <blogic@xxxxxxxxxxx> wrote: >> i am not sure i have seen another instance of dts using a count index >> for the following properties array/table/... do you have an example of >> a driver doing so ? > > gpio-bcm-kona does something close: the number of interrupts specified > indicate how many banks the chip has. I would not rely too much on > existing bindings to decide what a "good" practice is. Overall honestly I think too much responsibility of DT bindings (and now also ACPI!) is pushed to the subsystem maintainers, we are Linux GPIO experts, not really hardware description experts. We do a best effort but the current practice of letting the subsystem maintainers have the final word on all device tree bindings is not optimal, I would like to have a DT/OF persons Reviewed-by tag on any substantially new binding that is not just a copy of what every other GPIO driver is doing. Yours, Linus Walleij
