Re: [PATCH v6 1/2] media: dt-bindings: Add Himax HM1092 NIR sensor

[Rob Herring <robh@kernel.org>]

On Sat, Aug 01, 2026 at 01:01:49PM +0000, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [Medium] The schema fails to restrict the `data-lanes` property to a single lane, allowing physically invalid device tree configurations to silently pass validation.
> --
> 
> --- Patch [1]: [PATCH v6 1/2] media: dt-bindings: Add Himax HM1092 NIR sensor ---
> commit ee5879a0923ec6b56edc74d8a66ff697b111ae09
> Author: Ramshouriesh R <rshouriesh@gmail.com>
> 
> media: dt-bindings: Add Himax HM1092 NIR sensor
> 
> This commit adds a device tree binding schema for the Himax HM1092, a 1
> megapixel monochrome near-infrared CMOS image sensor. The sensor is described
> as outputting 10-bit RAW data over a single MIPI CSI-2 data lane.
> 
> > diff --git a/Documentation/devicetree/bindings/media/i2c/himax,hm1092.yaml b/Documentation/devicetree/bindings/media/i2c/himax,hm1092.yaml
> > new file mode 100644
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/media/i2c/himax,hm1092.yaml
> [ ... ]
> > +  port:
> > +    $ref: /schemas/graph.yaml#/$defs/port-base
> > +    additionalProperties: false
> > +
> > +    properties:
> > +      endpoint:
> > +        $ref: /schemas/media/video-interfaces.yaml#
> > +        unevaluatedProperties: false
> > +
> > +        required:
> > +          - link-frequencies
> 
> [Severity: Medium]
> The commit message mentions the sensor outputs over a single MIPI CSI-2 data
> lane, but does this schema restrict the data-lanes property to a single lane?
> 
> Without explicitly defining a data-lanes property constraint (such as
> maxItems: 1 or items: - const: 1) in this endpoint block, the schema inherits
> the permissive defaults from video-interfaces.yaml. This would allow device
> trees specifying physically impossible multi-lane configurations
> (e.g., data-lanes = <1 2>;) to silently pass schema validation.

data-lanes should not be present, but that's probably an issue on a 
number of bindings. There's not a great way to so this other than using 
'additionalProperties' instead and listing all the properties used from 
video-interfaces.yaml. But then you have to list 'remote-endpoint' 
everywhere which is not great either. I'd rather just allow the property 
even if it doesn't make sense for the h/w.

Rob