[PATCH V2] drm/bridge: lvds-codec: Add support for pixel data sampling edge select

[Laurent Pinchart]

Hi Rob,

There's a question for you below.

On Mon, Mar 22, 2021 at 12:37:47PM +0200, Laurent Pinchart wrote:
> Hi Marek,
> 
> (CC'ing Ron and the DT mailing list for the DT discussion)
> 
> On Mon, Mar 22, 2021 at 11:29:04AM +0100, Marek Vasut wrote:
> > On 3/22/21 2:14 AM, Laurent Pinchart wrote:
> > > Hi Marek,
> > 
> > Hi,
> > 
> > [...]
> > 
> > >> diff --git a/Documentation/devicetree/bindings/display/bridge/lvds-codec.yaml b/Documentation/devicetree/bindings/display/bridge/lvds-codec.yaml
> > >> index e5e3c72630cf..399a6528780a 100644
> > >> --- a/Documentation/devicetree/bindings/display/bridge/lvds-codec.yaml
> > >> +++ b/Documentation/devicetree/bindings/display/bridge/lvds-codec.yaml
> > >> @@ -74,6 +74,13 @@ properties:
> > >>   
> > >>       additionalProperties: false
> > >>   
> > >> +  pixelclk-active:
> > >> +    description: |
> > >> +      Data sampling on rising or falling edge.
> > >> +      Use 0 to sample pixel data on rising edge and
> > >> +      Use 1 to sample pixel data on falling edge and
> > >> +    enum: [0, 1]
> > > 
> > > The idea is good, but instead of adding a custom property, how about
> > > reusing the pclk-sample property defined in
> > > ../../media/video-interfaces.yaml ?
> > 
> > Repeating myself from V1 discussion ... Either is fine by me, but I 
> > think pixelclk-active, which comes from panel-timings.yaml is closer to 
> > the video than multimedia bindings.
> 
> That's a good point. The part that bothers me a bit is that it would be
> nice to define the property in a single YAML schema, referenced by
> individual bindings. video-interfaces.yaml is there for that purpose. We
> could do something similar on the display side, or consider the
> pixelclk-active usage in panel-timings.yaml an exception that we can't
> switch to video-interfaces.yaml as backward compatibility must be
> preserved.
> 
> I don't have a too strong preference, whatever Rob prefers would be fine
> with me.
> 
> > > The property is only valid for encoders, so I would at least mention
> > > that in the description, or, better, handle this based on the compatible
> > > string to allow validation.
> > 
> > How does that work in the Yaml file ?
> 
> Something along the lines of
> 
> if:
>   not:
>     properties:
>       compatible:
>         contains:
>           const: lvds-encoder
> then:
>   properties:
>     pixelclk-active: false
> 
> My YAML schema foo is a bit rusty though, I apologize if this doesn't
> work as-is. There are lots of similar examples in DT bindings that
> should hopefully be right :-)
> 
> > >> +
> > >>     powerdown-gpios:
> > >>       description:
> > >>         The GPIO used to control the power down line of this device.
> > >> diff --git a/drivers/gpu/drm/bridge/lvds-codec.c b/drivers/gpu/drm/bridge/lvds-codec.c
> > >> index dcf579a4cf83..cab81ccd895d 100644
> > >> --- a/drivers/gpu/drm/bridge/lvds-codec.c
> > >> +++ b/drivers/gpu/drm/bridge/lvds-codec.c
> > 
> > [...]
> > 
> > >> @@ -126,6 +146,7 @@ static int lvds_codec_probe(struct platform_device *pdev)
> > >>   	 */
> > >>   	lvds_codec->bridge.of_node = dev->of_node;
> > >>   	lvds_codec->bridge.funcs = &funcs;
> > >> +	lvds_codec->bridge.timings = &lvds_codec->timings;
> > >>   	drm_bridge_add(&lvds_codec->bridge);
> > >>   
> > >>   	platform_set_drvdata(pdev, lvds_codec);
> > >> @@ -142,19 +163,20 @@ static int lvds_codec_remove(struct platform_device *pdev)
> > >>   	return 0;
> > >>   }
> > >>   
> > >> +static const struct lvds_codec_data decoder_data = {
> > >> +	.connector_type = DRM_MODE_CONNECTOR_DPI,
> > >> +	.is_encoder = false,
> > > 
> > > The two fields are a bit redundant, as the decoder is always
> > > LVDS-to-DPI, and the encoder DPI-to-LVDS. I don't mind too much, but
> > > maybe we could drop the connector_type field, and derive the connector
> > > type from is_encoder ?
> > 
> > Or the other way around instead ? That is, if the connector_type is 
> > LVDS, then it is encoder , otherwise its decoder ?
> 
> Either way works for me.
> 
> > > One may then say that we could drop the lvds_codec_data structure as it
> > > contains a single field, but I foresee a need to have device-specific
> > > timings at some point, so I think it's a good addition.
> > 
> > [...]

-- 
Regards,

Laurent Pinchart