Am Montag, den 30.05.2016, 18:39 +0200 schrieb Peter Senna Tschudin:
> Add support to attach a drm_bridge to imx-ldb in addition to
> existing support to attach a LVDS panel.
>
> Signed-off-by: Peter Senna Tschudin <peter.senna at collabora.com>
> ---
> drivers/gpu/drm/imx/imx-ldb.c | 75
> +++++++++++++++++++++++++++++++------------
> 1 file changed, 54 insertions(+), 21 deletions(-)
>
> diff --git a/drivers/gpu/drm/imx/imx-ldb.c b/drivers/gpu/drm/imx/imx-ldb.c
> index a58eee5..7233a81 100644
> --- a/drivers/gpu/drm/imx/imx-ldb.c
> +++ b/drivers/gpu/drm/imx/imx-ldb.c
> @@ -57,7 +57,11 @@ struct imx_ldb_channel {
> struct imx_ldb *ldb;
> struct drm_connector connector;
> struct drm_encoder encoder;
> +
> + /* Defines what is connected to the ldb, only one at a time */
> struct drm_panel *panel;
> + struct drm_bridge *ext_bridge;
> +
Just bridge should be clear enough.
> struct device_node *child;
> int chno;
> void *edid;
> @@ -295,6 +299,10 @@ static void imx_ldb_encoder_mode_set(struct drm_encoder
> *encoder,
> }
> }
>
> +static void imx_ldb_encoder_enable(struct drm_encoder *encoder)
> +{
> +}
> +
Why?
Note that Liu Ying's atomic series touches this also, and after
"drm/imx: atomic phase 3 step 3: Legacy callback fixups" the enable
callback exists. Depending on how either series proceeds, one will have
to be rebased onto the other.
> static void imx_ldb_encoder_disable(struct drm_encoder *encoder)
> {
> struct imx_ldb_channel *imx_ldb_ch = enc_to_imx_ldb_ch(encoder);
> @@ -373,6 +381,7 @@ static const struct drm_encoder_helper_funcs
> imx_ldb_encoder_helper_funcs = {
> .prepare = imx_ldb_encoder_prepare,
> .commit = imx_ldb_encoder_commit,
> .mode_set = imx_ldb_encoder_mode_set,
> + .enable = imx_ldb_encoder_enable,
> .disable = imx_ldb_encoder_disable,
> };
>
> @@ -417,16 +426,28 @@ static int imx_ldb_register(struct drm_device *drm,
> drm_encoder_init(drm, &imx_ldb_ch->encoder, &imx_ldb_encoder_funcs,
> DRM_MODE_ENCODER_LVDS, NULL);
>
> - drm_connector_helper_add(&imx_ldb_ch->connector,
> - &imx_ldb_connector_helper_funcs);
> - drm_connector_init(drm, &imx_ldb_ch->connector,
> - &imx_ldb_connector_funcs, DRM_MODE_CONNECTOR_LVDS);
> -
> - if (imx_ldb_ch->panel)
> + if (imx_ldb_ch->panel) {
if (!imx_ldb_ch->bridge) {
> + drm_connector_helper_add(&imx_ldb_ch->connector,
> + &imx_ldb_connector_helper_funcs);
> + drm_connector_init(drm, &imx_ldb_ch->connector,
> + &imx_ldb_connector_funcs,
> + DRM_MODE_CONNECTOR_LVDS);
> drm_panel_attach(imx_ldb_ch->panel, &imx_ldb_ch->connector);
>
> - drm_mode_connector_attach_encoder(&imx_ldb_ch->connector,
> - &imx_ldb_ch->encoder);
> + drm_mode_connector_attach_encoder(&imx_ldb_ch->connector,
> + &imx_ldb_ch->encoder);
> + }
> +
> + if (imx_ldb_ch->ext_bridge) {
> + imx_ldb_ch->ext_bridge->encoder = &imx_ldb_ch->encoder;
> +
> + imx_ldb_ch->encoder.bridge = imx_ldb_ch->ext_bridge;
> + ret = drm_bridge_attach(drm, imx_ldb_ch->ext_bridge);
> + if (ret) {
> + DRM_ERROR("Failed to initialize bridge with drm\n");
> + return ret;
> + }
> + }
>
> return 0;
> }
> @@ -583,23 +604,35 @@ static int imx_ldb_bind(struct device *dev, struct
> device *master, void *data)
> endpoint = of_get_child_by_name(port, "endpoint");
> if (endpoint) {
> remote =
> of_graph_get_remote_port_parent(endpoint);
> - if (remote)
> - channel->panel =
> of_drm_find_panel(remote);
> - else
> - return -EPROBE_DEFER;
> - if (!channel->panel) {
> - dev_err(dev, "panel not found: %s\n",
> - remote->full_name);
Let's keep this message, at least as dev_dbg.
> - return -EPROBE_DEFER;
> + if (remote) {
> + /* Only one of these two will succeed */
> + channel->panel =
> + of_drm_find_panel(remote);
> +
> + channel->ext_bridge =
> + of_drm_find_bridge(remote);
> +
> + /*
> + * If the bridge is compiled as a
> + * module, it may take some time until
> + * the bridge driver is available.
> + * Defer until the bridge driver is
> + * ready.
> + */
> + if (!channel->panel &&
> + !channel->ext_bridge)
> + return -EPROBE_DEFER;
> }
What happens if (!remote)?
> }
> }
>
> - edidp = of_get_property(child, "edid", &channel->edid_len);
> - if (edidp) {
> - channel->edid = kmemdup(edidp, channel->edid_len,
> - GFP_KERNEL);
> - } else if (!channel->panel) {
> + if (channel->panel) {
> + edidp = of_get_property(child, "edid",
> + &channel->edid_len);
> + if (edidp)
> + channel->edid = kmemdup(edidp,
> + channel->edid_len, GFP_KERNEL);
> + } else {
Any reason to disallow overriding EDID from DT if there is no panel? I'd
leave this part as is.
regards
Philipp
