On 27/09/2024 10:02, Heiko Stuebner wrote: > Hi Krzysztof, > > Am Freitag, 27. September 2024, 09:30:30 CEST schrieb Krzysztof Kozlowski: >> On 27/09/2024 09:01, Frank Wang wrote: >>> Hi Krzysztof, >>> >>> On 2024/9/26 22:19, Krzysztof Kozlowski wrote: >>>> On 26/09/2024 12:32, Frank Wang wrote: >>>>> + - if: >>>>> + properties: >>>>> + compatible: >>>>> + contains: >>>>> + enum: >>>>> + - rockchip,rk3576-usb2phy >>>>> + then: >>>>> + properties: >>>>> + clocks: >>>>> + minItems: 3 >>>>> + maxItems: 3 >>>> Read one more time the example I gave you. Top-level constraints are >>>> saying max one clock. >>>> >>>> Best regards, >>>> Krzysztof >>>> >>> >>> Sorry for overlooking this, I will set both "clocks" and "clock-names" >>> to true, and add the else case below the above codes for the "old" SoCs. >>> Just like the below. >>> >>> - clocks: >>> - maxItems: 1 >>> + clocks: true >>> >>> - clock-names: >>> - const: phyclk >>> + clock-names: true >> >> For the third time, read the code I gave you. Do you see something like >> this there? Why doing all the time something different than existing code? > > On vacation right now so late to the party, and somewhat confused :-) . > > I've tried to find the code you mentioned, but did fail. > In [0] you mention "maybe oneOf". The other replies in that version were > about the ordering needing to stay for the older phy variants. > > [1] in v2 has that NAK thing and [2] from v3 references that example again > > I am probably just blind, but could use a pointer. Oh, maybe I did not provide the link? I apologize. I thought I gave reference to standard example. My bad. Here it goes: https://elixir.bootlin.com/linux/v6.11-rc6/source/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml#L127 Best regards, Krzysztof
