On 01/12/2022 17:45, Conor Dooley wrote: > Hey Yanhong, > > On Thu, Dec 01, 2022 at 07:36:29AM -0600, Rob Herring wrote: >> >> On Thu, 01 Dec 2022 17:02:38 +0800, Yanhong Wang wrote: >>> Add bindings for the StarFive dwmac module on the StarFive RISC-V SoCs. >>> >>> Signed-off-by: Yanhong Wang <yanhong.wang@xxxxxxxxxxxxxxxx> >>> --- >>> .../devicetree/bindings/net/snps,dwmac.yaml | 1 + >>> .../bindings/net/starfive,dwmac-plat.yaml | 106 ++++++++++++++++++ >>> MAINTAINERS | 5 + >>> 3 files changed, 112 insertions(+) >>> create mode 100644 Documentation/devicetree/bindings/net/starfive,dwmac-plat.yaml >>> >> >> My bot found errors running 'make DT_CHECKER_FLAGS=-m dt_binding_check' >> on your patch (DT_CHECKER_FLAGS is new in v5.13): >> >> yamllint warnings/errors: >> ./Documentation/devicetree/bindings/net/starfive,dwmac-plat.yaml:30:16: [warning] wrong indentation: expected 14 but found 15 (indentation) >> >> dtschema/dtc warnings/errors: >> ./Documentation/devicetree/bindings/net/starfive,dwmac-plat.yaml: $id: relative path/filename doesn't match actual path or filename >> expected: http://devicetree.org/schemas/net/starfive,dwmac-plat.yaml# >> Documentation/devicetree/bindings/net/starfive,dwmac-plat.example.dts:21:18: fatal error: dt-bindings/clock/starfive-jh7110.h: No such file or directory >> 21 | #include <dt-bindings/clock/starfive-jh7110.h> > > Perhaps, rather than putting a long list of "prerequisite-patch-id" in > your cover letters etc, you drop the need for headers from your bindings > entirely? Otherwise, you need to wait for the clock bindings to be applied > before any of your other peripherals etc can have drivers/bindings upstream. > > AFAIU, and Rob/Krzk please correct me, the example in a dt-binding > really is an *example* and there's no requirement for it to match the > jh7110 dts exactly. Because of that you can drop the header & just do > something like `clocks = <&clk 7>, <&clk 77>;` etc and the example is > still valid. Same goes for all of the other driver patchsets for new > StarFive stuff, like the pmu or crypto, that also have dt-bindings. > > The only person who has to worry then about dependencies is me when I > apply the .dts patches :) It's nice when example looks good and matching final DTS, but that's not a requirement. Pretty often clocks/resets/domains are mocked for the example. Best regards, Krzysztof