On Sat, Jul 29, 2023 at 05:15:06PM +0500, Nikita Travkin wrote: > Conor Dooley писал(а) 29.07.2023 17:10: > > On Sat, Jul 29, 2023 at 05:06:14PM +0500, Nikita Travkin wrote: > >> Conor Dooley писал(а) 29.07.2023 15:03: > >> > On Fri, Jul 28, 2023 at 10:19:30PM +0500, Nikita Travkin wrote: > > > >> >> + interrupt-names: > >> >> + items: > >> >> + - const: fifo > >> > > >> > Same here, but do you really need a name, when you have only one > >> > interrupt? > >> > > >> > >> Hm, thinking of this more, the hardware actually has more than one > >> interrupt, even though this one seems to be the only really useful > >> one. Would a better way forward be to list all of them > > > > Yes. > > > >> (and fix > >> the driver to get the value by it's name) > > > > It's not a fix to do that, the order of the interrupts is not variable, > > so there's nothing wrong with using the indices. You can do it if you > > like. > > > >> or it would be > >> acceptable to leave the names here and extend the list at a later > >> date when (if ever) other interrupts are needed? > > > > If you know what they are, please describe them now, even if the driver > > does not use them (yet). > > > > Thanks for the clarification! Will make sure both drivers have all > interrupts described in v2 Note that bindings describe hardware, not the driver. The driver need not touch the other interrupts if it does not use them, but make the hardware description (IOW the dt-binding) accurate & as complete as you can. Thanks, Conor.
Attachment:
signature.asc
Description: PGP signature
