On 12/02/24 22:09, Josua Mayer wrote: > Hi Vignesh, > > Am 12.02.24 um 05:10 schrieb Vignesh Raghavendra: >> Hi Josua, >> >> On 11/02/24 20:37, Josua Mayer wrote: >>> This series adds DT bindings and dts descriptions for SolidRun AM642 >>> based SoM and Hummingboard EVB. >>> >>> Additionally a commit from downstream vendor kernel are included, >>> enhancing support for pru based ethernet. >>> I wasn't sure how to properly annotate it in commit description / >>> signed-off area ...: >>> >>> 1. add description for "Industrial Ethernet Peripherals" (IEP) to am64 >>> https://git.ti.com/cgit/ti-linux-kernel/ti-linux-kernel/commit/arch/arm64/boot/dts/ti/k3-am64-main.dtsi?h=ti-linux-6.1.y-cicd&id=5afb73d82a014b59462162d960b350b8c58e5ae6 >>> IEP is already supported in-tree by a driver, and used in >>> k3-am65-main.dtsi. >>> >>> Unfortunately dtbs_check reported many problems, I put some remarks: >>> >>> - 'mux-controller' does not match any of the regexes >>> The expectation seems to be that a mux-controller at minimum has an >>> address, something to put behind an @. However this is a gpio mux, not >>> sure how to name it better. >>> >> I don't see this warning locally. Are you using updated dt-schema? > pip3 install dtschema --upgrade > Defaulting to user installation because normal site-packages is not writeable > Requirement already satisfied: dtschema in ~/.local/lib/python3.11/site-packages (2023.11) > > Re-Tested on 6.8-rc1 > >> reg >> is not necessary gpio-mux as per gpio-mux.yaml > The error is not about reg property, it is about the node name: > > mux-controller { > compatible = "gpio-mux"; > ... > }; > > DTC_CHK arch/arm64/boot/dts/ti/k3-am642-hummingboard-t.dtb > .../arch/arm64/boot/dts/ti/k3-am642-hummingboard-t.dtb: syscon@43000000: 'mux-controller' does not match any of the regexes: '^chipid@[0-9a-f]+$', '^clock-controller@[0-9a-f]+$', '^mux-controller@[0-9a-f]+$', 'phy@[0-9a-f]+$', 'pinctrl-[0-9]+' > from schema $id: http://devicetree.org/schemas/mfd/ti,j721e-system-controller.yaml# This warning exists not because of gpio-mux but because of SoC level (k3-am64-main.dtsi) dtsi issue. There is a separate series to fix the same. So this should be okay to ignore for now. >> >>> - unevaluated properties: interrupts, interrupt-parent >>> sensors and flash yaml are missing interrupt descriptions, but these >>> parts definitely have an interrupt signal in this solidrun board. >>> >> Please add them to appropriate schema as necessary > Okay. > Looks like it is only two: > .../arch/arm64/boot/dts/ti/k3-am642-hummingboard-t.dtb: humidity-sensor@41: 'interrupt-parent', 'interrupts' do not match any of the regexes: 'pinctrl-[0-9]+' > from schema $id: http://devicetree.org/schemas/iio/humidity/ti,hdc2010.yaml# > .../arch/arm64/boot/dts/ti/k3-am642-hummingboard-t.dtb: flash@0: Unevaluated properties are not allowed ('interrupt-parent', 'interrupts' were unexpected) > from schema $id: http://devicetree.org/schemas/mtd/jedec,spi-nor.yaml# Yup, thanks! you could post a cut down version of the dts with all the problematic nodes removed until yaml updates are merged, so as to get the basic support in. >> >>> - wrong names for pinctrl nodes >>> Other TI DTSs consistently end with *-pins-default. Should a different >>> naming convention be used? >>> >> No, pinctrl nodes need to end in -pins. All TI boards have been updated >> to new schema [0] and sysconfig tool on dev.ti.com/sysconfig generates >> appropriately. Please fix > Okay, will do ... >> >> >>> - cdns,phy-type required property >>> inherited from k3-am64-main.dtsi >>> there is a PHY_NONE value in dt-bindings/phy/phy.h, >>> but not allowed in phy-cadence-torrent.yaml >>> >> Sorry, I didnt get what's the issue wrt cdns,phy-type ? > There were two issues, but they both disappeared as per 6.8-rc1 :) >> That's nice! -- Regards Vignesh
