Re: [PATCH v2 1/2] dt-bindings: mfd: Add ROHM BD71879

[Matti Vaittinen <mazziesaccount@gmail.com>]

On 4/10/24 13:03, Conor Dooley wrote:
> On Wed, Apr 10, 2024 at 10:57:08AM +0300, Matti Vaittinen wrote:
>> On 4/4/24 22:54, Andreas Kemnade wrote:
>>> As this chip was seen in several devices in the wild, add it.
>>>
>>> Signed-off-by: Andreas Kemnade <andreas@kemnade.info>
>>> Suggested-by: Matti Vaittinen <mazziesaccount@gmail.com>
>>> ---
>>>    .../devicetree/bindings/mfd/rohm,bd71828-pmic.yaml         | 7 ++++++-
>>>    1 file changed, 6 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/Documentation/devicetree/bindings/mfd/rohm,bd71828-pmic.yaml b/Documentation/devicetree/bindings/mfd/rohm,bd71828-pmic.yaml
>>> index 0b62f854bf6b..07f99738fcf6 100644
>>> --- a/Documentation/devicetree/bindings/mfd/rohm,bd71828-pmic.yaml
>>> +++ b/Documentation/devicetree/bindings/mfd/rohm,bd71828-pmic.yaml
>>> @@ -17,7 +17,12 @@ description: |
>>>    properties:
>>>      compatible:
>>> -    const: rohm,bd71828
>>> +    oneOf:
>>> +      - const: rohm,bd71828
>>> +
>>> +      - items:
>>> +          - const: rohm,bd71879
>>> +          - const: rohm,bd71828
>>>      reg:
>>>        description:
>>
>> Am I correct, this reads as:
>>
>> Either
>> 	compatible = rohm,bd71828
>> or
>> 	compatible = rohm,bd71879, rohm,bd71828
>>
>> but not compatible = rohm,bd71879?
> 
> Correct.

Thanks Conor!

I guess that if we later find out that there is a 'gizmo' in BD71828 
which is not present on BD71879 - and if we write driver supporting 
this, then we need to have handling for both the BD71879 and BD71828 in 
this driver to ensure the fall-back wont happen. It's a bit ugly as I 
think the fall-back compatible should be used only in case where the 
"fall-back" HW can be guaranteed to have exactly same, or a subset of, 
features of the "full board".

Well, I've been told these are similar by HW colleagues. I have my 
doubts if 'similar' really is '_similar_', or if it's 'similar, except 
of course...' - but let's trust the HW guys on this. We have the above 
mentioned workaround if we should not have trusted...

Acked-By: Matti Vaittinen <mazziesaccount@gmail.com>

-- 
Matti Vaittinen
Linux kernel developer at ROHM Semiconductors
Oulu Finland

~~ When things go utterly wrong vim users can always type :help! ~~