Re: [PATCH] dt-bindings: input: matrix-keymap: Add common 'linux,no-autorepeat' property

[Dmitry Torokhov <dmitry.torokhov@gmail.com>]

On Mon, May 23, 2022 at 02:45:03PM -0500, Rob Herring wrote:
> On Mon, May 23, 2022 at 11:47:01AM -0700, Dmitry Torokhov wrote:
> > On Mon, May 23, 2022 at 12:04:49PM -0500, Rob Herring wrote:
> > > 'linux,no-autorepeat' is a common property used in multiple bindings,
> > > but doesn't have a common type definition nor description. Add a common
> > > definition and drop the now redundant description from
> > > holtek,ht16k33.yaml.
> > 
> > We have "autorepeat" in the common input binding description, should we
> > not promote it over "no-autorepeat"?
> 
> We're kind of stuck with it I think. We can't just deprecate one and 
> switch existing users as what would neither property present mean?

I agree, we should not change existng bindings.

> 
> Hopefully, documented in input.yaml vs. matrix-keypad.yaml is enough to 
> say which one is preferred for new users. 

So this is what I have issue with, as I think it will introduce
confusion: we have drivers/input/matrix-keymap.c and corresponding
binding Documentation/devicetree/bindings/input/matrix-keymap.yaml that
deals with parsing keymap-related properties for various matrix keypads.
It does not specify autorepeat handling one way or another. Then we have
drivers/input/keyboard/matrix_keypad.c that is one implementation of
matrix keypads, and it does have linux,no-autorepeat, but it does not
mean that other devices resembling matrix keypad are forced to use
linux,no-autorepeat. And that is why I think putting this property into
the generic binding
Documentation/devicetree/bindings/input/matrix-keymap.yaml is not a good
idea.

Thanks.

-- 
Dmitry