Re: [PATCH 2/3] dt-bindings: arm: Document reboot mode magic

[Elliot Berman]

On 11/20/2023 8:41 AM, Krzysztof Kozlowski wrote:
> On 20/11/2023 16:45, Elliot Berman wrote:
>>
>>
>> On 11/20/2023 2:56 AM, Krzysztof Kozlowski wrote:
>>> On 17/11/2023 22:18, Elliot Berman wrote:
>>>> -
>>>> -additionalProperties: false
>>>> +  - if:
>>>> +      properties:
>>>> +        compatible:
>>>> +          contains:
>>>> +            const: arm,psci-1.0
>>>> +    then:
>>>> +      $ref: /schemas/power/reset/reboot-mode.yaml#
>>>> +      properties:
>>>> +        # "mode-normal" is just SYSTEM_RESET
>>>> +        mode-normal: false
>>>> +      patternProperties:
>>>> +        "^mode-.*$":
>>>> +          maxItems: 2
>>>
>>> And if you tested the patch, it would tell you it can be max 1 item.
>>
>> make dt_binding_check DT_SCHEMA_FILES=arm/psci.yaml
> 
> psci.example.dtb: psci: mode-edl: [[0]] is too short
> psci.example.dtb: psci: mode-bootloader: [[1, 2]] is too short
> 
> psci.example.dtb: psci: Unevaluated properties are not allowed
> ('mode-bootloader', 'mode-edl' were unexpected)
> 

Ah, tip of tree for dt-schema doesn't seem to report the error.

Doesn't report the error:
dt-validate --version
2023.10.dev17+g58feadb

Reports the error:
dt-validate --version
2023.9

Looks likely related to generated the processed-schema.json rather
than dt-validate itself. The tip of tree dt-validate does report
the error if processed-schema.json is generated by 2023.9 tool,
but not if the schema was also generated by tip-of-tree mkschema.

I'll try bisecting the error and report back.

>>
>> passes for me. Rob explained why it's working (and why it shouldn't), 
>> so I'll fix it according to his recommendation in v2.
> 
> Then you wanted uint32-matrix.
> 
> 
> Best regards,
> Krzysztof
>