On Mon, May 18, 2026 at 10:44:29AM +0800, Damon Ding wrote: > Hi Conor, > > On 5/15/2026 5:04 PM, Conor Dooley wrote: > > On Fri, May 15, 2026 at 11:57:58AM +0800, Damon Ding wrote: > > > Hi Conor, > > > > > > On 5/15/2026 2:16 AM, Conor Dooley wrote: > > > > On Thu, May 14, 2026 at 03:01:31PM +0800, Damon Ding wrote: > > > > > Expose the inherited properties from the base analogix-dp schema > > > > > to satisfy unevaluatedProperties constraints. > > > > > > > > > > Signed-off-by: Damon Ding <[email protected]> > > > > > > > > Given it's unevaluatedProperties, not addtionalProperties, this patch > > > > shouldn't be needed? > > > > > > > > > > When I remove both the top-level data-lanes property and those explicit > > > "xxx: true" property entries and run the dtbs check with: > > > > > > make CHECK_DTBS=y CROSS_COMPILE=aarch64-linux-gnu- LT0=none LLVM=1 > > > LLVM_IAS=1 ARCH=arm64 rockchip/rk3588-evb1-v10.dtb > > > rockchip/rk3588s-evb1-v10.dtb rockchip/rk3399-sapphire-excavator.dtb > > > rockchip/rk3576-evb1-v10.dtb -j4 > > > > > > It results in validation errors like these: > > > > > > /home/ding/drm-misc/arch/arm64/boot/dts/rockchip/rk3576-evb1-v10.dtb: > > > edp@27dc0000: ports:port@1:endpoint: Unevaluated properties are not > > > allowed > > > ('data-lanes' was unexpected) > > > from schema $id: > > > http://devicetree.org/schemas/display/rockchip/rockchip,analogix-dp.yaml# > > > /home/ding/drm-misc/arch/arm64/boot/dts/rockchip/rk3576-evb1-v10.dtb: > > > edp@27dc0000: Unevaluated properties are not allowed ('force-hpd', > > > 'interrupts', 'phy-names', 'phys', 'ports', 'reg' were unexpected) > > > from schema $id: > > > http://devicetree.org/schemas/display/rockchip/rockchip,analogix-dp.yaml# > > > > > > I suspect that the properties defined in the child binding are > > > overriding/masking all the inherited properties from the parent Analogix > > > DP > > > schema. > > > > > > Is there a better way to fix this issue without explicitly listing all > > > inherited properties as true? > > > > The example in this file uses most of the properties that you mention > > above: > > dp@ff970000 { > > compatible = "rockchip,rk3288-dp"; > > reg = <0xff970000 0x4000>; > > interrupts = <GIC_SPI 98 IRQ_TYPE_LEVEL_HIGH>; > > clocks = <&cru SCLK_EDP>, <&cru PCLK_EDP_CTRL>; > > clock-names = "dp", "pclk"; > > phys = <&dp_phy>; > > phy-names = "dp"; > > resets = <&cru 111>; > > reset-names = "dp"; > > rockchip,grf = <&grf>; > > pinctrl-0 = <&edp_hpd>; > > pinctrl-names = "default"; > > > > dt_binding_check reports no problems with this node, so I think the > > problem might lie elsewhere? > > There's no edp node in the dts you mention above, so this looks like an > > interaction with something that's not yet upstream. > > Ah, I see. > > I did add extra changes to the edp node on my local rk3576-evb1 board DTS, > which indeed caused this 'asymmetric information' issue and made the > validation fail incorrectly. > > I will try adding the data-lanes property to the example in the binding > document to ensure dt_binding_check passes cleanly. > > > > > If this is required for the rk3576 edp, then you should include this > > patch in the rk3576 edp support series rather than this one anyway where > > it can actually be evaluated alongside the node it apparently causes > > problems with. > > The data-lanes property is actually applicable and useful for all Rockchip > eDP platforms, not just rk3576. > > In the next version, I will modify the series to better separate this common > data-lanes change from the rk3576 specific eDP support patches.
To be clear, I wasn't talking about data-lanes here, I was talking about this patch that addressed the properties inherited from the common analogix-dp.yaml.
signature.asc
Description: PGP signature
