Re: [PATCH v3 3/5] dt-bindings: media: Add CAMSS device for Hawi and Maili

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Medium] dt-bindings: media: hawi-camss missing required data-lanes in endpoint
- [Low] dt-bindings: media: hawi-camss example unit-address mismatches first reg
--

commit 97b880fd3cf2adf2e5bdc34dc34d8524f1dd635b
Author: Hangxiang Ma <hangxiang.ma@oss.qualcomm.com>

dt-bindings: media: Add CAMSS device for Hawi and Maili

This commit adds the Device Tree binding schema for the Camera Subsystem
(CAMSS) on the Qualcomm Hawi and Maili platforms.

> diff --git a/Documentation/devicetree/bindings/media/qcom,hawi-camss.yaml b/Documentation/devicetree/bindings/media/qcom,hawi-camss.yaml
> new file mode 100644
> index 0000000000000..32c1cc59e8767
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/media/qcom,hawi-camss.yaml
> @@ -0,0 +1,314 @@

[ ... ]

> +    patternProperties:
> +      "^port@[0-5]$":
> +        $ref: /schemas/graph.yaml#/$defs/port-base
> +        unevaluatedProperties: false
> +        description:
> +          Input ports for receiving CSI data on CSIPHY 0-5.
> +
> +        properties:
> +          endpoint:
> +            $ref: video-interfaces.yaml#
> +            unevaluatedProperties: false
> +
> +            properties:
> +              data-lanes:
> +                minItems: 1
> +                maxItems: 4
> +
> +              bus-type:
> +                enum:
> +                  - 1 # MEDIA_BUS_TYPE_CSI2_CPHY
> +                  - 4 # MEDIA_BUS_TYPE_CSI2_DPHY

[Severity: Medium]
As reported by Sashiko on v2 [1], should the endpoint schema under
patternProperties ^port@[0-5]$ include a required list specifying data-lanes,
matching sibling bindings such as qcom,kaanapali-camss.yaml and
qcom,x1e80100-camss.yaml?

During driver probe, camss_parse_endpoint_node() in
drivers/media/platform/qcom/camss/camss.c calls v4l2_fwnode_endpoint_parse()
on each port@[0-5] endpoint. When bus-type is omitted and data-lanes is
absent, v4l2_fwnode_endpoint_parse() does not detect a CSI-2 D-PHY bus and
camss_parse_endpoint_node() fails with -EINVAL. Even when bus-type is
explicitly set to D-PHY (4), omitting data-lanes leaves lncfg->num_data as 0,
causing csiphy_stream_on() in drivers/media/platform/qcom/camss/camss-csiphy.c
to fail phy_mipi_dphy_get_default_config_for_hsclk() with -EINVAL.

[ ... ]

> +    soc {
> +        #address-cells = <2>;
> +        #size-cells = <2>;
> +
> +        isp@9253000 {
> +            compatible = "qcom,hawi-camss";
> +
> +            reg = <0x0 0x0900b000 0x0 0x1000>,
> +                  <0x0 0x09253000 0x0 0x5e80>,

[Severity: Low]
Should the node unit-address in the example be isp@900b000 instead of
isp@9253000?

The first reg entry (top) is at 0x0900b000, while 0x09253000 is the second
reg entry (csid0). Running make dt_binding_check triggers a dtc
simple_bus_reg warning when the node unit-address does not match the first
reg address.

[1] https://lore.kernel.org/r/20261013-hawi-maili-camss-v2-3-fc93222da89c@oss.qualcomm.com

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261030-hawi-maili-camss-v3-0-9742bca8cef2@oss.qualcomm.com?part=3

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy