Re: [PATCH v3 07/20] dt-bindings: firmware: google,gs101-acpm-ipc: update PMIC examples

[Krzysztof Kozlowski <krzk@kernel.org>]

On 08/11/2025 15:08, André Draszik wrote:
> Hi Krzysztof,
> 
> Thanks for your review!
> 
> On Tue, 2025-11-04 at 09:31 +0100, Krzysztof Kozlowski wrote:
>> On Mon, Nov 03, 2025 at 07:14:46PM +0000, André Draszik wrote:
>>> In a typical system using the Samsung S2MPG10 PMIC, an S2MPG11 is used
>>> as a sub-PMIC.
>>>
>>> The interface for both is the ACPM firmware protocol, so update the
>>> example here to describe the connection for both.
>>>
>>> Signed-off-by: André Draszik <andre.draszik@linaro.org>
>>> ---
>>>  .../bindings/firmware/google,gs101-acpm-ipc.yaml   | 40 ++++++++++++++++++++--
>>>  1 file changed, 37 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/Documentation/devicetree/bindings/firmware/google,gs101-acpm-ipc.yaml
>>> b/Documentation/devicetree/bindings/firmware/google,gs101-acpm-ipc.yaml
>>> index 4a1e3e3c0505aad6669cadf9b7b58aa4c7f284cb..c25e155926e5f44bd74f195cdbff3672c7499f8e 100644
>>> --- a/Documentation/devicetree/bindings/firmware/google,gs101-acpm-ipc.yaml
>>> +++ b/Documentation/devicetree/bindings/firmware/google,gs101-acpm-ipc.yaml
>>> @@ -45,6 +45,15 @@ properties:
>>>        compatible:
>>>          const: samsung,s2mpg10-pmic
>>>  
>>> +  pmic2:
>>
>> pmic-2
>>
>> Are there more pmics? Bindings are supposed to be complete (see writing
>> bindings) and if you did follow this approach earlier, you would nicely
>> call first "pmic-1" (instead of "pmic") and then "pmic-2".
> 
> There aren't any more PMICs on ACPM, no. At the time 'pmic' was added, it wasn't clear
> unfortunately that two nodes would be needed in the end.
> 
> See also https://lore.kernel.org/all/963bbf8db71efc0d729bb9141c133df2c56881fc.camel@linaro.org/
> 
> That said, I believe we can change the existing node name from pmic to pmic-1 without
> any driver breaking. The sysfs path would change, but I don't think anybody cares about
> it at this stage, so I think such a change would be fine. The ACPM driver doesn't care
> about node names and instantiates all children regardless of name.
> 
> I propose to update the binding (and DTS subsequently) to add pmic-1, to allow 'pmic' as
> a legacy fallback (i.e. to not issue errors during validation of existing DTSs until
> they're updated) and to use pmic-2 for the 2nd pmic.
> 
> OK?

So deprecate 'pmic' and add new 'pmic-1'.

Best regards,
Krzysztof