Re: [PATCH 1/2] dt-bindings: linflexuart: add clock definitions

[Ciprian Marian Costea <ciprianmarian.costea@oss.nxp.com>]

On 9/24/2024 6:27 PM, Conor Dooley wrote:
> On Tue, Sep 24, 2024 at 06:17:11PM +0300, Ciprian Marian Costea wrote:
>> On 9/24/2024 6:01 PM, Conor Dooley wrote:
>>> On Tue, Sep 24, 2024 at 05:52:13PM +0300, Ciprian Marian Costea wrote:
>>>> On 9/24/2024 5:24 PM, Conor Dooley wrote:
>>>>> On Tue, Sep 24, 2024 at 05:12:46PM +0300, Ciprian Costea wrote:
>>>>>> From: Ciprian Marian Costea <ciprianmarian.costea@oss.nxp.com>
>>>>>>
>>>>>> Add clock definitions for NXP LINFlexD UART bindings
>>>>>> and update the binding examples with S32G2 node.
>>>>>>
>>>>>> Signed-off-by: Ciprian Marian Costea <ciprianmarian.costea@oss.nxp.com>
>>>>>> ---
>>>>>>     .../bindings/serial/fsl,s32-linflexuart.yaml  | 21 +++++++++++++++++++
>>>>>>     1 file changed, 21 insertions(+)
>>>>>>
>>>>>> diff --git a/Documentation/devicetree/bindings/serial/fsl,s32-linflexuart.yaml b/Documentation/devicetree/bindings/serial/fsl,s32-linflexuart.yaml
>>>>>> index 4171f524a928..45fcab9e186d 100644
>>>>>> --- a/Documentation/devicetree/bindings/serial/fsl,s32-linflexuart.yaml
>>>>>> +++ b/Documentation/devicetree/bindings/serial/fsl,s32-linflexuart.yaml
>>>>>> @@ -34,6 +34,14 @@ properties:
>>>>>>       interrupts:
>>>>>>         maxItems: 1
>>>>>> +  clocks:
>>>>>> +    maxItems: 2
>>>>>> +
>>>>>> +  clock-names:
>>>>>> +    items:
>>>>>> +      - const: ipg
>>>>>> +      - const: lin
>>>>>
>>>>> Can all devices have 2 clocks, or just the s32g2?
>>>>>
>>>>
>>>> All devices (S32G2, S32G3 and S32V234) have 2 clocks for LINFlexD module.
>>>
>>> I see. How come the driver is capable of working without them?
>>>
>>
>> The driver was working because the LINFlexD clocks were configured and kept
>> enabled by the downstream bootloader (TF-A and U-Boot). This is not ideal
>> since LINFlexD Linux driver should manage those clocks independently and not
>> rely on a previous bootloader configuration (hence the need for this current
>> patchset).
> 
> I'd also mark them as required in the binding too, but the driver will
> still have to account for them being missing, for backwards
> compatibility reasons. Add a comment explaining the history to the
> commit message when you do that.

Already in the second patch from this patchset the clocking support was 
added in the LINFlexD driver as optional, indeed for backwards 
compatibility.
I presumed that because of this optional clock enablement, I should not 
add the clocks as required in the bindings, but I will do so in V2. Thanks.