Re: [PATCH v4 06/10] drm/rockchip: dw_hdmi_qp: Set supported_formats platdata

[Nicolas Frattaroli <nicolas.frattaroli@collabora.com>]

On Tuesday, 18 November 2025 21:14:31 Central European Standard Time Cristian Ciocaltea wrote:
> On 11/17/25 9:11 PM, Nicolas Frattaroli wrote:
> > With the introduction of the supported_formats member in the
> > dw-hdmi-qp platform data struct, drivers that have access to this
> > information should now set it.
> > 
> > Set it in the rockchip dw_hdmi_qp glue driver, where such a bitmask of
> > supported color formats already exists. It just needs to be converted to
> > the appropriate HDMI_COLORSPACE_ mask.
> > 
> > This allows this information to be passed down to the dw-hdmi-qp core,
> > which sets it in the bridge it creates, and consequently will allow the
> > common HDMI bridge code to act on it.
> > 
> > Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
> > ---
> >  drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c | 24 ++++++++++++++++++++++++
> >  1 file changed, 24 insertions(+)
> > 
> > diff --git a/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c b/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
> > index c9fe6aa3e3e3..7c294751de19 100644
> > --- a/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
> > +++ b/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
> > @@ -468,6 +468,28 @@ static const struct of_device_id dw_hdmi_qp_rockchip_dt_ids[] = {
> >  };
> >  MODULE_DEVICE_TABLE(of, dw_hdmi_qp_rockchip_dt_ids);
> >  
> > +static const u32 supported_colorformats = DRM_COLOR_FORMAT_AUTO |
> > +					  DRM_COLOR_FORMAT_RGB444 |
> > +					  DRM_COLOR_FORMAT_YCBCR444;
> > +
> > +static unsigned int __pure drm_to_hdmi_fmts(const u32 fmt)
> > +{
> > +	unsigned int res = 0;
> > +
> > +	if (fmt & DRM_COLOR_FORMAT_AUTO)
> > +		res |= BIT(HDMI_COLORSPACE_RGB);
> > +	if (fmt & DRM_COLOR_FORMAT_RGB444)
> > +		res |= BIT(HDMI_COLORSPACE_RGB);
> > +	if (fmt & DRM_COLOR_FORMAT_YCBCR444)
> > +		res |= BIT(HDMI_COLORSPACE_YUV444);
> > +	if (fmt & DRM_COLOR_FORMAT_YCBCR422)
> > +		res |= BIT(HDMI_COLORSPACE_YUV422);
> > +	if (fmt & DRM_COLOR_FORMAT_YCBCR420)
> > +		res |= BIT(HDMI_COLORSPACE_YUV420);
> > +
> > +	return res;
> > +}
> > +
> >  static int dw_hdmi_qp_rockchip_bind(struct device *dev, struct device *master,
> >  				    void *data)
> >  {
> > @@ -521,6 +543,8 @@ static int dw_hdmi_qp_rockchip_bind(struct device *dev, struct device *master,
> >  	plat_data.phy_data = hdmi;
> >  	plat_data.max_bpc = 10;
> >  
> > +	plat_data.supported_formats = drm_to_hdmi_fmts(supported_colorformats);
> 
> Any reason why this cannot be simply set as
> 
>   BIT(HDMI_COLORSPACE_RGB) | BIT(HDMI_COLORSPACE_YUV444) | BIT(HDMI_COLORSPACE_YUV422)
> 
> and get rid of the unnecessary conversion?

My gut feeling lead me towards trying to have a single source of
truth for the supported color formats, but upon further reflection
this is indeed way too verbose and lead me to move the
supported_colorformats definition into this patch rather than the
one where it's needed for registering the property.

So I agree with you here and will simplify this by just setting
these as you described.

Kind regards,
Nicolas Frattaroli

> 
> > +
> >  	encoder = &hdmi->encoder.encoder;
> >  	encoder->possible_crtcs = drm_of_find_possible_crtcs(drm, dev->of_node);
> >  
> > 
> 
>