Re: [PATCH 1/1] dt-bindings: ata: ahci-fsl-qoriq: add fsl,ls1046a-ahci and fsl,ls1012a-ahci

From: Damien Le Moal
Date: Thu Jul 04 2024 - 20:21:03 EST


On 7/3/24 05:45, Frank Li wrote:
> On Wed, Jun 26, 2024 at 10:17:55AM +0200, Krzysztof Kozlowski wrote:
>> On 25/06/2024 22:57, Frank Li wrote:
>>> Add compatible string 'fsl,ls1046a-ahci' and 'fsl,ls1012a-ahci' compatible
>>> string. Allow 'fsl,ls1012a-ahci' fallback to 'fsl,ls1043a-ahci'.
>>>
>>> ls1046a ahci ecc disable bit is difference with other chips.
>>>
>>> Signed-off-by: Frank Li <Frank.Li@xxxxxxx>
>>> ---
>>> .../devicetree/bindings/ata/fsl,ahci.yaml | 19 ++++++++++++-------
>>> 1 file changed, 12 insertions(+), 7 deletions(-)
>>>
>>> diff --git a/Documentation/devicetree/bindings/ata/fsl,ahci.yaml b/Documentation/devicetree/bindings/ata/fsl,ahci.yaml
>>> index 162b3bb5427ed..a244bc603549d 100644
>>> --- a/Documentation/devicetree/bindings/ata/fsl,ahci.yaml
>>> +++ b/Documentation/devicetree/bindings/ata/fsl,ahci.yaml
>>> @@ -11,13 +11,18 @@ maintainers:
>>>
>>> properties:
>>> compatible:
>>> - enum:
>>> - - fsl,ls1021a-ahci
>>> - - fsl,ls1043a-ahci
>>> - - fsl,ls1028a-ahci
>>> - - fsl,ls1088a-ahci
>>> - - fsl,ls2080a-ahci
>>> - - fsl,lx2160a-ahci
>>> + oneOf:
>>> + - items:
>>> + - const: fsl,ls1012a-ahci
>>> + - const: fsl,ls1043a-ahci
>>> + - enum:
>>> + - fsl,ls1021a-ahci
>>> + - fsl,ls1043a-ahci
>>> + - fsl,ls1046a-ahci
>>
>> Where is the driver change for this?
>>
>> Your commit does not explain why you are doing it and without driver
>> change adding new support it is not obvious. This probably applies to
>> all your patches.
>
> I think I missed ls1012a and ls1021a. Commit message is wrong. This is
> for legancy platorm.
>
> Basic try to eliminate the CHECK_DTBS warning. ls1012a use
>
> "fsl,ls1012a-ahci", "fsl,ls1043a-ahci". There are two methods,
> 1. change binding doc to allow "fsl,ls1012a-ahci", "fsl,ls1043a-ahci"

But then shouldn't you also change the drivers/ata/ahci_qoriq.c to add ls1012a
as a compatible ?

> 2. remove "fsl,ls1012a-ahci".

I am not sure if that is acceptable since there is one device tree using it out
there already.

I am no DT expert, but I think (1) with the driver change is the right approach.
Krzysztof ?

--
Damien Le Moal
Western Digital Research