Re: [PATCH 1/2] dt-bindings: iio: dac: add support for mcp47a1

On Sat, 25 Jul 2026 10:41:30 -0500
David Lechner <[email protected]> wrote:

> On 7/25/26 3:12 AM, Joshua Crofts wrote:
> > On Sat, 25 Jul 2026 00:04:51 +0100
> > Jonathan Cameron <[email protected]> wrote:
> >   
> >> On Tue, 21 Jul 2026 20:24:38 +0200
> >> Joshua Crofts <[email protected]> wrote:
> >>  
> >>> The Microchip MCP47A1 is a 6-bit volatile Digital-to-Analog converted
> >>> which communicates via I2C.
> >>>
> >>> Signed-off-by: Joshua Crofts <[email protected]>
> >>> ---
> >>>  .../bindings/iio/dac/microchip,mcp47a1.yaml        | 56 ++++++++++++++++++++++
> >>>  MAINTAINERS                                        |  6 +++
> >>>  2 files changed, 62 insertions(+)
> >>>
> >>> diff --git a/Documentation/devicetree/bindings/iio/dac/microchip,mcp47a1.yaml b/Documentation/devicetree/bindings/iio/dac/microchip,mcp47a1.yaml
> >>> new file mode 100644
> >>> index 000000000000..3322fca03b3a
> >>> --- /dev/null
> >>> +++ b/Documentation/devicetree/bindings/iio/dac/microchip,mcp47a1.yaml    
> >>  
> >>> +
> >>> +examples:
> >>> +  - |
> >>> +    i2c {
> >>> +          #address-cells = <1>;
> >>> +          #size-cells = <0>;
> >>> +
> >>> +          dac@2e {
> >>> +              compatible = "microchip,mcp47a1";
> >>> +              reg = <0x2e>;
> >>> +              vref-supply = <&vref_regulator>;
> >>> +              vdd-supply = <&vdd_regulator>;
> >>> +          };
> >>> +
> >>> +          dac@3e {
> >>> +              compatible = "microchip,mcp47a1";    
> >> Hi Joshua,
> >>
> >> What is the advantage of the second example?  I'd drop it.
> >>  
> > 
> > Hi Jonathan,
> > 
> > Per the datasheet, the MCP47A1 can have two different I2C slave
> > addresses (0x2e and 0x3e), which allows two DACs to share the
> > same bus, something I thought would be good to illustrate in the
> > example.

Given many i2c devices either come in a couple of variants or have
a pinstrap method to change their address, to me this looks like
an unnecessary level of detail.
> >   
> 
> I know we haven't done this much historically, but we could make
> that more obvious with this instead:
> 
>   reg:
>     items:
>       - enum: [0x2e, 0x3e]
> 
> Then the bindings validator would actually check for a valid
> address.
> 
Sure, there is some precedence. I did a quick grep and immediately
spotted:
Documentation/devicetree/bindings/regulator/adi,max77857.yaml

So, whilst I don't want to see a mass set of 'fixes' adding
limits on reg for available I2C addresses I'm fine with this style
in new bindings. 

Thanks,

Jonathan