Hi Marek,
On Tue, May 25, 2021 at 12:38:47PM +0200, Marek Vasut wrote:
> On 5/18/21 1:03 AM, Laurent Pinchart wrote:
>
> Hi,
>
> [...]
>
> >> @@ -69,10 +70,33 @@ static void lvds_codec_disable(struct drm_bridge *bridge)
> >> "Failed to disable regulator \"vcc\": %d\n", ret);
> >> }
> >>
> >> +static bool lvds_codec_mode_fixup(struct drm_bridge *bridge,
> >> + const struct drm_display_mode *mode,
> >> + struct drm_display_mode *adj)
> >> +{
> >> + struct lvds_codec *lvds_codec = to_lvds_codec(bridge);
> >> + struct drm_encoder *encoder = bridge->encoder;
> >> + struct drm_device *ddev = encoder->dev;
> >> + struct drm_connector *connector;
> >> +
> >> + /* If 'data-mapping' was not specified, do nothing. */
> >> + if (!lvds_codec->bus_format)
> >> + return true;
> >> +
> >> + /* Patch in the LVDS format */
> >> + list_for_each_entry(connector, &ddev->mode_config.connector_list, head) {
> >> + drm_display_info_set_bus_formats(&connector->display_info,
> >> + &lvds_codec->bus_format, 1);
> >> + }
> >
> > This part bothers me, as the format at the input of the LVDS decoder
> > doesn't match the format on the connector. Shouldn't you implement
> > .atomic_get_output_bus_fmts() instead ?
>
> No, I tried that option before this solution and that didn't work. The
> atomic stuff seems to be separate from what I need to pass here, i.e.
> without this, e.g. the mxsfb scanout engine still receives the wrong format.
I fear this needs to be investigated then, as the connected format isn't
the same as the LVDS decoder input format, so the above isn't right.
--
Regards,
Laurent Pinchart
