On 23/02/2024 15:56, Xu Yang wrote: > Hi Rob, > >> >> On Wed, Feb 21, 2024 at 10:58:41PM +0800, Xu Yang wrote: >>> As more and more NXP i.MX chips come out, it becomes harder to maintain >>> ci-hdrc-usb2.yaml if more stuffs like property restrictions are added to >>> this file. This will separate i.MX parts out of ci-hdrc-usb2.yaml and add >>> a new schema for NXP ChipIdea USB2 Controller. >>> >>> Signed-off-by: Xu Yang <xu.yang_2@xxxxxxx> >>> >>> --- >>> Changes in v6: >>> - new patch >>> --- >>> .../bindings/usb/ci-hdrc-usb2-imx.yaml | 75 +++++++++++++++++++ >>> 1 file changed, 75 insertions(+) >>> create mode 100644 Documentation/devicetree/bindings/usb/ci-hdrc-usb2-imx.yaml >>> >>> diff --git a/Documentation/devicetree/bindings/usb/ci-hdrc-usb2-imx.yaml b/Documentation/devicetree/bindings/usb/ci- >> hdrc-usb2-imx.yaml >>> new file mode 100644 >>> index 000000000000..2ec62f564bf5 >>> --- /dev/null >>> +++ b/Documentation/devicetree/bindings/usb/ci-hdrc-usb2-imx.yaml >>> @@ -0,0 +1,75 @@ >>> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) >>> +%YAML 1.2 >>> +--- >>> +$id: http://devicetree.org/schemas/usb/ci-hdrc- >> usb2- >> imx.yaml%23&data=05%7C02%7Cxu.yang_2%40nxp.com%7C4ac0c60cd4b4433f0f9f08dc34782572%7C686ea1d3bc2b4c6fa92c >> d99c5c301635%7C0%7C0%7C638442937830606824%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIi >> LCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&sdata=7p1DzvYmBsTgN44jypH7lc56z9hVBsFBYXUwsblk9z8%3D&reserv >> ed=0 >>> +$schema: http://devicetree.org/meta- >> schemas%2Fcore.yaml%23&data=05%7C02%7Cxu.yang_2%40nxp.com%7C4ac0c60cd4b4433f0f9f08dc34782572%7C686ea1d3 >> bc2b4c6fa92cd99c5c301635%7C0%7C0%7C638442937830615622%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLC >> JQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&sdata=fWWy9enbGK5yeKiovday7go3Gss5L%2F%2Fe%2F >> OZcANny0QA%3D&reserved=0 >>> + >>> +title: NXP USB2 ChipIdea USB controller >>> + >>> +maintainers: >>> + - Xu Yang <xu.yang_2@xxxxxxx> >>> + >>> +properties: >>> + compatible: >>> + oneOf: >>> + - enum: >>> + - fsl,imx27-usb >>> + - items: >>> + - enum: >>> + - fsl,imx23-usb >>> + - fsl,imx25-usb >>> + - fsl,imx28-usb >>> + - fsl,imx35-usb >>> + - fsl,imx50-usb >>> + - fsl,imx51-usb >>> + - fsl,imx53-usb >>> + - fsl,imx6q-usb >>> + - fsl,imx6sl-usb >>> + - fsl,imx6sx-usb >>> + - fsl,imx6ul-usb >>> + - fsl,imx7d-usb >>> + - fsl,vf610-usb >>> + - const: fsl,imx27-usb >>> + - items: >>> + - enum: >>> + - fsl,imx8dxl-usb >>> + - fsl,imx8ulp-usb >>> + - const: fsl,imx7ulp-usb >>> + - const: fsl,imx6ul-usb >>> + - items: >>> + - enum: >>> + - fsl,imx8mm-usb >>> + - fsl,imx8mn-usb >>> + - const: fsl,imx7d-usb >>> + - const: fsl,imx27-usb >>> + - items: >>> + - enum: >>> + - fsl,imx6sll-usb >>> + - fsl,imx7ulp-usb >>> + - const: fsl,imx6ul-usb >>> + - const: fsl,imx27-usb >> >> Now you just duplicated all the compatibles, and now any new compatibles >> have to be added in 2 places. For this to work, you have to split >> ci-hdrc-usb2.yaml into 2 files. One with all the common properties and >> one with compatibles (minus imx). This is also needed if imx has any >> extra properties the other don't. >> >> Didn't I say this already? >> > > Yes, I know. > > But according to your words, I need to split ci-hdrc-usb2.yaml into 1 common > file and more than 1 vendor specific files (imx, nvidia, qcom, nuvoton and > others). In this patchset, I only focus on imx part and KK said he or someone > will take over other parts, therefore I just duplicated all the imx compatibles. Someone will take over their bits, but it does not mean you can duplicate IMX. Why you cannot use the approach I asked - move IMX out of that schema? https://lore.kernel.org/all/a0134089-a283-488b-8d7f-3f59dd938b60@xxxxxxxxxx/ Look: "... and move IMX to own file." Move is not "copy". Please run: `man mv` > If I only create imx specific yaml file and remove all compatilbles from common Why would you remove ALL of others? What is the point of this. > file, nvidia, qcom, nuvoton and others compatible info will be lost, is this > feasible? Or should I create mutiple vendor specific files at the same time? You can take a look how it is done in my recent Qualcomm PCI dt-bindings series. Best regards, Krzysztof