[PATCH v5 19/23] arm64: dts: rockchip: rk3568-evb: Enable VOP2 and hdmi

Johan Jonker jbx6244 at gmail.com
Thu Feb 10 13:15:17 UTC 2022



On 2/10/22 12:47, Sascha Hauer wrote:
> On Thu, Feb 10, 2022 at 01:10:32AM +0100, Johan Jonker wrote:
>> Hi Sascha,
>>
>> Something with port and endpoint gives notifications.
>> Somehow with the conversion of rockchip,dw-hdmi.txt to YAML not all SoC
>> options were checked/covered (see rk3328 and rk3568).
>>
>> Allow multiple vop:
>> port or
>> port at 0
>> 1x vop -> endpoint
>>
>> 2x vop -> endpoint at 0
>>        -> endpoint at 1
>>
>> Also allow for connector:
>> (not all existing DT have this yet)
>> port at 1
>>       -> endpoint
>>
>> See also at graph.yaml
> 
> Ok, let me see if I get this right. The HDMI can either have one or two
> ports. It has one when the HDMI connector is not described in the device
> tree and two when it is. Also the first (or only) port can have one or
> two endpoints. It has one endpoint when the SoC has a single VOP and two
> endpoints when the SoC has two VOPs.
> 
> It's been a painful morning because my YAML knowledge is fairly limited,
> but here's what I came up with:
> 
>   ports:
>     $ref: /schemas/graph.yaml#/properties/ports
>     unevaluatedProperties: false
>     patternProperties:
>       "^port(@0)?$":
>         $ref: /schemas/graph.yaml#/properties/port
>         description: Input of the DWC HDMI TX
>         properties:
>           endpoint:
>             description: Connection to the VOP
>           endpoint at 0:
>             description: Connection to the VOPB
>           endpoint at 1:
>             description: Connection to the VOPL
>     properties:
>       port at 1:
>         $ref: /schemas/graph.yaml#/properties/port
>         description: Output of the DWC HDMI TX
> 
> Does this look sane to you? With this a make dtbs_check on this binding
> doesn't generate warnings anymore, and even better it does generate
> warnings when I add some wrong nodes/properties to the dts files.

Hi,

Send a patch to rob+dt and he will let you know... ;)

Also could you test the ARM branch as well just for sure with:
ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- make  dtbs_check
DT_SCHEMA_FILES=Documentation/devicetree/bindings/display/rockchip/rockchip,dw-hdmi.yaml

Could you group your serie a bit:
dt-bindings
drm patches
dts changes

Johan

> 
>> ===
>>
>> dtc -I dtb -O dts rk3568-evb1-v10.dtb
>> <stdout>: Warning (avoid_unnecessary_addr_size):
>> /hdmi at fe0a0000/ports/port at 1: unnecessary #address-cells/#size-cells
>> without "ranges" or child "reg" property
>> <stdout>: Warning (graph_child_address): /hdmi at fe0a0000/ports/port at 0:
>> graph node has single child node 'endpoint at 0',
>> #address-cells/#size-cells are not necessary
>> <stdout>: Warning (graph_child_address): /hdmi at fe0a0000/ports/port at 1:
>> graph node has single child node 'endpoint', #address-cells/#size-cells
>> are not necessary
> 
> That's easy. I dropped the @0 suffix and reg property from the node.
> 
> Sascha
> 


More information about the dri-devel mailing list