On 30/10/2024 09:14, Stanislav Jakubek wrote: > On Wed, Oct 30, 2024 at 08:48:25AM +0100, Krzysztof Kozlowski wrote: >> On 30/10/2024 08:42, Stanislav Jakubek wrote: >>>> >>>>> + >>>>> + '#address-cells': >>>>> + const: 1 >>>>> + >>>>> + '#interrupt-cells': >>>>> + const: 1 >>>>> + >>>>> + '#size-cells': >>>>> + const: 0 >>>>> + >>>>> + regulators: >>>>> + type: object >>>>> + $ref: /schemas/regulator/sprd,sc2731-regulator.yaml# >>>>> + >>>>> +patternProperties: >>>>> + "^adc@[0-9a-f]+$": >>>>> + type: object >>>>> + $ref: /schemas/iio/adc/sprd,sc2720-adc.yaml# >>>>> + >>>>> + "^charger@[0-9a-f]+$": >>>>> + type: object >>>>> + $ref: /schemas/power/supply/sc2731-charger.yaml# >>>>> + >>>>> + "^efuse@[0-9a-f]+$": >>>>> + type: object >>>>> + $ref: /schemas/nvmem/sprd,sc2731-efuse.yaml# >>>> >>>> I don't think this was merged. You still have dependency. >>> >>> This is in next-20241029, which this patch is based on. >> >> Try what I wrote below and see if this works... > > I assume you meant the MFD maintainers' tree here. > Yes, that tree doesn't have the nvmem patch this depends on. > > Would the approach with listing the compatibles and additionalProperties: > true be considered a temporary workaround? Not really, it's a correct approach. The node will be validated anyway by efuse/child schema. Best regards, Krzysztof