[PATCH v7 05/12] dt-bindings: display/msm: move common MDSS properties to mdss-common.yaml
Dmitry Baryshkov
dmitry.baryshkov at linaro.org
Fri Sep 23 20:32:50 UTC 2022
On 22/09/2022 15:28, Krzysztof Kozlowski wrote:
> On 22/09/2022 13:46, Dmitry Baryshkov wrote:
>>>> - ranges: true
>>>> + maxItems: 2
>>>>
>>>> interconnects:
>>>> - items:
>>>> - - description: Interconnect path from mdp0 port to the data bus
>>>> - - description: Interconnect path from mdp1 port to the data bus
>>>> + maxItems: 2
>>>
>>> I think this is not equivalent now, because you have in total minItems:1
>>> and maxItems:2, while in past minItems was 2.
>>
>> This means that I should have minItems:2, maxItems:2, which, if I got it
>> right, is frowned upon. Let me doublecheck though if it works as expected.
>
> It is frowned upon only if it is alone, because for missing minItems,
> maxItems implies minItems. Here you have minItems in other schema, so
> there is no such case
Well, I just checked, the schema will throw an error if I put a single
interconnects or iommus entry. If I understand correctly these two
clauses are evaluated separately. So, the dpu-common's clause tells
minItems:1, maxItems:2. The platform schema file contains just
maxItems:2, which implicitly adds minItems:2 to _this_ clause.
Thus I think I'll leave this part as is.
For the reference (with single-entry iommus and interconnects properties):
/home/lumag/Projects/Qcomm/build-64/Documentation/devicetree/bindings/display/msm/dpu-sdm845.example.dtb:
display-subsystem at ae00000: iommus: [[4294967295, 2176, 8]] is too short
From schema:
/home/lumag/Projects/Qcomm/kernel/Documentation/devicetree/bindings/display/msm/dpu-sdm845.yaml
/home/lumag/Projects/Qcomm/build-64/Documentation/devicetree/bindings/display/msm/dpu-sdm845.example.dtb:
display-subsystem at ae00000: interconnects: [[4294967295, 1, 0,
4294967295, 1, 0]] is too short
From schema:
/home/lumag/Projects/Qcomm/kernel/Documentation/devicetree/bindings/display/msm/dpu-sdm845.yaml
--
With best wishes
Dmitry
More information about the dri-devel
mailing list