Re: [PATCH v5 7/7] arm64: dts: imx8dxl-evk: add lcdif overlay

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Medium] arm64: dts: imx8dxl-evk: invalid and unparsed lcdif endpoint bus-width
- [Medium] arm64: dts: imx8dxl-evk: disabling acm and codecs breaks audio consumers
--

commit cfee436980bc9bed627f94a41d5e90946202
Author: Frank Li <Frank.Li@nxp.com>

arm64: dts: imx8dxl-evk: add lcdif overlay

This commit adds a device tree overlay to support the waiken,101wx001-wct
LCD panel on the i.MX8DXL EVK board, enabling adma_lcdif and disabling
peripherals that share pins with the LCDIF interface.

> diff --git a/arch/arm64/boot/dts/freescale/imx8dxl-evk-lcdif-101wx001-wct.dtso b/arch/arm64/boot/dts/freescale/imx8dxl-evk-lcdif-101wx001-wct.dtso
> new file mode 100644
> index 0000000000000..39abf9b3884ea
> --- /dev/null
> +++ b/arch/arm64/boot/dts/freescale/imx8dxl-evk-lcdif-101wx001-wct.dtso
[ ... ]
> +&acm {
> +	status = "disabled";
> +};
> +
> +&adma_lcdif {
> +	pinctrl-0 = <&pinctrl_lcdif>, <&lcd_0_fun>, <&lcd_1_fun>, <&lcd_2_fun>, <&lcd_3_fun>;
> +	pinctrl-names = "default";
> +	status = "okay";
> +
> +	port {
> +		lcdif_out: endpoint {
> +			bus-width = <18>;
> +			remote-endpoint = <&panel_in>;
> +		};
> +	};
> +};

[Severity: Medium]
Does setting bus-width = <18> on the lcdif_out endpoint work with the
fsl,lcdif binding and the mxsfb driver?

In Documentation/devicetree/bindings/display/fsl,lcdif.yaml, the port
property is defined as:

  port:
    $ref: /schemas/graph.yaml#/properties/port
    description: The LCDIF output port

Because it references /schemas/graph.yaml#/properties/port directly without
referencing video-interfaces.yaml on the endpoint, endpoint sets
additionalProperties: false and only allows reg, remote-endpoint, and
phandle. Adding bus-width = <18> causes dtbs_check schema validation to fail.

Furthermore, at runtime neither mxsfb, drm_panel_bridge, nor panel-simple
parses bus-width from the endpoint node. When mxsfb_crtc_atomic_enable()
calls mxsfb_set_formats() in drivers/gpu/drm/mxsfb/mxsfb_kms.c, it selects
the bus width strictly from bus_format:

  switch (bus_format) {
  case MEDIA_BUS_FMT_RGB565_1X16:
          ctrl |= CTRL_BUS_WIDTH_16;
          break;
  case MEDIA_BUS_FMT_RGB666_1X18:
          ctrl |= CTRL_BUS_WIDTH_18;
          break;
  case MEDIA_BUS_FMT_RGB888_1X24:
          ctrl |= CTRL_BUS_WIDTH_24;
          break;

Since waiken_101wx001_wct_panel in drivers/gpu/drm/panel/panel-simple.c
advertises MEDIA_BUS_FMT_RGB888_1X24, mxsfb programs CTRL_BUS_WIDTH_24
even though only 18 data lines (ADMA_LCDIF_D00 through ADMA_LCDIF_D17) are
routed on the board.

[ ... ]
> +&sai0 {
> +	status = "disabled";
> +};
> +
> +&sai1 {
> +	status = "disabled";
> +};
> +
> +&wm8960_1 {
> +	status = "disabled";
> +};
> +
> +&wm8960_2 {
> +	status = "disabled";
> +};
> +
> +&wm8960_3 {
> +	status = "disabled";
> +};

[Severity: Medium]
Should the remaining audio nodes and sound cards in imx8dxl-evk.dts and
imx8-ss-audio.dtsi also be disabled when &acm, &sai0, &sai1, and
&wm8960_1..3 are disabled?

In the base device tree, several enabled nodes depend on &acm and the
disabled codecs/SAIs:

1. The audio LPCG clock nodes (sai0_lpcg through sai3_lpcg, spdif0_lpcg,
mclkout0_lpcg, mclkout1_lpcg, and mqs0_lpcg) take parent clocks from &acm.
With &acm disabled, imx8qxp_lpcg_clk_probe() ->
imx_lpcg_parse_clks_from_dt() in drivers/clk/imx/clk-imx8qxp-lpcg.c fails to
resolve parent clock names via of_clk_parent_fill():

  ret = of_clk_parent_fill(np, parent_names, count);
  if (ret != count) {
          dev_err(&pdev->dev, "failed to get clock parent names\n");
          return count;
  }

2. &asrc0 remains enabled in imx8dxl-evk.dts and directly consumes clocks
from &acm, causing fsl_asrc_probe() to fail with -ENOENT.

3. &sai2 and &sai3 remain enabled in imx8dxl-evk.dts and consume clocks from
sai2_lpcg and sai3_lpcg, causing fsl_sai_probe() to fail clock acquisition
and defer probe indefinitely.

4. The root sound card nodes (sound-bt-sco, sound-wm8960-1, sound-wm8960-2,
and sound-wm8960-3) remain enabled in imx8dxl-evk.dts while referencing the
disabled &sai0, &sai1, &wm8960_1..3, and failing &asrc0/&sai2/&sai3 nodes,
leaving their drivers in permanent probe deferral.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261027-dxl_lcdif-v5-0-3a974e4b94b2@nxp.com?part=7