Replies: 2 comments 2 replies
-
The I've avoid accepting a generic gpio signal binding as usually a GPIO should be associated with some other dts node. We added support for inferred bindings as a middle ground: https://docs.zephyrproject.org/latest/guides/dts/bindings.html#inferred-bindings |
Beta Was this translation helpful? Give feedback.
-
Our general principle is to conform to upstream dts binding conventions as much as possible. So we utilize concepts from: We don't support all the things defined here, but if what you need is covered here we can implement it. |
Beta Was this translation helpful? Give feedback.
-
I've noticed there are device tree bindings for both gpio-keys and gpio-leds that are identical expect for their description. Should we consolidate these two definitions?
Also I would like to have a list of named gpios that don't fall into the key or led category; therefore, I would like to create a more generic version:
gpio-signals
. Thegpio-signals
binding would be exactly the same as gothgpio-keys
andgpio-leds
.If we have the more generically named version (
gpio-signals
), should we deprecate and remove thegpio-keys
andgpio-leds
compatibility bindings and move everything to thegpio-signals
binding?Related question, but I can create a new discussion if needed: I also noticed that the
GPIO_
* flags indented for use in the device tree only include the lower 8 bits. This does not include whether the gpio is an input or an output. When listing our named gpios (with abovegpio-signals
), it make sense to define in thedts
file whether it is an output and what the initial output state should be.Specifying only the lower 8 bits is not a limit of the device tree, per se, but it looks to be convention. Can or should we change this convention, or what is the motivation behind this convention?
Beta Was this translation helpful? Give feedback.
All reactions