Re: [RFC PATCH v3 04/15] dt-bindings: leds: ROHM BD71282 PMIC LED driver

[Vaittinen, Matti]

Hello Dan,

Thanks for the check once again!

On Tue, 2019-11-05 at 13:14 -0600, Dan Murphy wrote:
> Matti
> 
> On 11/1/19 6:32 AM, Matti Vaittinen wrote:
> > Document ROHM BD71828 PMIC LED driver device tree bindings.
> > 
> > Signed-off-by: Matti Vaittinen <matti.vaittinen@xxxxxxxxxxxxxxxxx>
> > ---
> > 
> > Changes from v2 - new patch
> > 
> >   .../bindings/leds/rohm,leds-bd71828.yaml      | 46
> > +++++++++++++++++++
> >   1 file changed, 46 insertions(+)
> >   create mode 100644
> > Documentation/devicetree/bindings/leds/rohm,leds-bd71828.yaml
> > 
> > diff --git a/Documentation/devicetree/bindings/leds/rohm,leds-
> > bd71828.yaml b/Documentation/devicetree/bindings/leds/rohm,leds-
> > bd71828.yaml
> > new file mode 100644
> > index 000000000000..d8aeac9911ef
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/leds/rohm,leds-bd71828.yaml
> > @@ -0,0 +1,46 @@
> > +# SPDX-License-Identifier: GPL-2.0
> > +%YAML 1.2
> > +---
> > +$id: http://devicetree.org/schemas/leds/rohm,leds-bd71828.yaml#
> > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > +
> > +title: ROHM BD71828 Power Management Integrated Circuit LED driver
> > +
> > +maintainers:
> > +  - Jacek Anaszewski <jacek.anaszewski@xxxxxxxxx>
> > +  - Pavel Machek <pavel@xxxxxx>
> > +  - Dan Murphy <dmurphy@xxxxxx>
> > +  - Rob Herring <robh+dt@xxxxxxxxxx>
> > +  - Mark Rutland <mark.rutland@xxxxxxx>
> I believe you are the maintainer of this driver not the maintainers

Right. Thanks for pointing that out.

> > +
> > +description: |
> > +  This module is part of the ROHM BD71828 MFD device. For more
> > details
> > +  see Documentation/devicetree/bindings/mfd/rohm,bd71828-
> > pmic.yaml.
> > +
> > +  The LED controller is represented as a sub-node of the PMIC node
> > on the device
> > +  tree.
> > +
> > +  The device has two LED outputs referred as GRNLED and AMBLED in
> > data-sheet.
> > +
> > +properties:
> > +  compatible:
> > +    const: rohm,bd71828-led
> > +
> > +patternProperties:
> > +  "^led-[1-2]$":
> > +    type: object
> > +    description:
> > +      Properties for a single LED. Nodes must be named as led-1
> > and led-2.
> 
> Why is this required?  Can't we use the reg as the number and then
> we 
> can use standard node labels

This was related to my idea of using the node-names as unique keys.
Please see:
https://lore.kernel.org/lkml/cover.1572351774.git.matti.vaittinen@xxxxxxxxxxxxxxxxx/

What would you expect the reg = <>; to describe from HW?

> like led@<reg value>.  Then we can check in the code to make sure
> that 
> the output is not out of bounds.
> 
> > +    properties:
> > +      #$ref: "common.yaml#"
> > +      function:
> > +        description:
> > +          Purpose of LED as defined in dt-bindings/leds/common.h
> > +        $ref: "/schemas/types.yaml#/definitions/string"
> > +      color:
> > +        description:
> > +          LED colour as defined in dt-bindings/leds/common.h
> 
> s/colour/color

That depends on your location :)

> But again I believe it is indicated above that the LEDs are either
> going 
> to be green or amber.  Unless they can be any color.

This was my original reason for omitting the DT for BD71828 LEDs
altogether. LEDs are expected to be green and amber - but it is true
that PMIC can not ensure there will be no other colours.

> Are there plans to make sure that the color is either green or amber
> in 
> the code?  I don't see a patch for the code in this series

Yes. As I wrote in cover-letter, the LED driver is pending until I see
how the RFC for adding LED node finding and some more common property
parsing to LED core is received by others. (Although I do understand if
you didn't read the cover-letter. It's quite a bunch of text and
reading it over and over again is no fun).

> 
> > +        $ref: "/schemas/types.yaml#/definitions/uint32"
> > +
> > +required:
> > +  - compatible
> 
> Is there an example of the node and properties?

Yes, in MFD doc.

Br,
	Matti Vaittinen