Re: [PATCH v1 2/7] drm/mediatek: fix to get right bridge for dpi encoder

From: CK Hu
Date: Mon May 28 2018 - 04:30:27 EST


Hi, Bibby:

Some inline comment.

On Mon, 2018-05-14 at 15:52 +0800, Bibby Hsieh wrote:
> From: chunhui dai <chunhui.dai@xxxxxxxxxxxx>
>
> 1, dpi is an encoder, there is an bridge in the struct
> of decoder, we could use it.

The encoder.bridge is assigned in drm_bridge_attach(), so I think the
design is to assign this member in drm_bridge_attach().
Any assignment outside this function may conflict with this design.

> 2, using of_graph_get_remote_port_parent to get right
> bridge in device tree.
>
> Signed-off-by: chunhui dai <chunhui.dai@xxxxxxxxxxxx>
> ---
> drivers/gpu/drm/mediatek/mtk_dpi.c | 21 +++++++++++++--------
> 1 file changed, 13 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/gpu/drm/mediatek/mtk_dpi.c b/drivers/gpu/drm/mediatek/mtk_dpi.c
> index 0a44ab175422..2b8b34c72697 100644
> --- a/drivers/gpu/drm/mediatek/mtk_dpi.c
> +++ b/drivers/gpu/drm/mediatek/mtk_dpi.c
> @@ -63,7 +63,6 @@ enum mtk_dpi_out_color_format {
> struct mtk_dpi {
> struct mtk_ddp_comp ddp_comp;
> struct drm_encoder encoder;
> - struct drm_bridge *bridge;
> void __iomem *regs;
> struct device *dev;
> struct clk *engine_clk;
> @@ -643,8 +642,8 @@ static int mtk_dpi_bind(struct device *dev, struct device *master, void *data)
>
> /* Currently DPI0 is fixed to be driven by OVL1 */
> dpi->encoder.possible_crtcs = BIT(1);
> -
> - ret = drm_bridge_attach(&dpi->encoder, dpi->bridge, NULL);
> + dpi->encoder.bridge->encoder = &dpi->encoder;

This is done inside drm_bridge_attach().

> + ret = drm_bridge_attach(&dpi->encoder, dpi->encoder.bridge, NULL);
> if (ret) {
> dev_err(dev, "Failed to attach bridge: %d\n", ret);
> goto err_cleanup;
> @@ -709,7 +708,7 @@ static int mtk_dpi_probe(struct platform_device *pdev)
> struct device *dev = &pdev->dev;
> struct mtk_dpi *dpi;
> struct resource *mem;
> - struct device_node *bridge_node;
> + struct device_node *ep, *bridge_node;
> int comp_id;
> const struct of_device_id *match;
> int ret;
> @@ -759,15 +758,21 @@ static int mtk_dpi_probe(struct platform_device *pdev)
> return -EINVAL;
> }
>
> - bridge_node = of_graph_get_remote_node(dev->of_node, 0, 0);
> - if (!bridge_node)
> + ep = of_graph_get_next_endpoint(dev->of_node, NULL);
> + if (ep) {
> + bridge_node = of_graph_get_remote_port_parent(ep);
> + of_node_put(ep);
> + }
> + if (!bridge_node) {
> + dev_err(dev, "Failed to find bridge node\n");
> return -ENODEV;
> + }

This looks like mt8173 has the same bug, so send this patch
independently and it would be more quick to merge.
For mtk_dsi, it use drm_of_find_panel_or_bridge(), so I think you could
use the same function to get bridge.

Regards,
CK

>
> dev_info(dev, "Found bridge node: %pOF\n", bridge_node);
>
> - dpi->bridge = of_drm_find_bridge(bridge_node);
> + dpi->encoder.bridge = of_drm_find_bridge(bridge_node);
> of_node_put(bridge_node);
> - if (!dpi->bridge)
> + if (!dpi->encoder.bridge)
> return -EPROBE_DEFER;
>
> comp_id = mtk_ddp_comp_get_id(dev->of_node, MTK_DPI);