Re: [PATCH v3 4/7] arm64: dts: imx8-ss-dma: Add lcdif support for i.MX8QXP and i.MX8DXL

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] arm64: dts: imx8: missing adma_lcdif and lcd_clk_lpcg deletion on imx8qm
--

commit 90895a7ff82b0903e1aaa5301959b704913e49e1
Author: Frank Li <Frank.Li@nxp.com>

arm64: dts: imx8-ss-dma: Add lcdif support for i.MX8QXP and i.MX8DXL

This commit adds the adma_lcdif and lcd_clk_lpcg device tree nodes to
imx8-ss-dma.dtsi for i.MX8QXP and overrides the compatible string, clock
rates, and interrupt in imx8dxl-ss-adma.dtsi for i.MX8DXL.

> diff --git a/arch/arm64/boot/dts/freescale/imx8-ss-dma.dtsi b/arch/arm64/boot/dts/freescale/imx8-ss-dma.dtsi
> index 4de78f870c05c..87082c3fa9338 100644
> --- a/arch/arm64/boot/dts/freescale/imx8-ss-dma.dtsi
> +++ b/arch/arm64/boot/dts/freescale/imx8-ss-dma.dtsi
> @@ -149,6 +149,24 @@ lpuart3: serial@5a090000 {
>  		status = "disabled";
>  	};
>  
> +	adma_lcdif: lcdif@5a180000 {
> +		compatible = "fsl,imx8qxp-lcdif", "fsl,imx6sx-lcdif";
> +		reg = <0x5a180000 0x10000>;
> +		clocks = <&lcd_clk_lpcg IMX_LPCG_CLK_0>,
> +			 <&lcd_clk_lpcg IMX_LPCG_CLK_4>,
> +			 <&clk IMX_SC_R_LCD_0 IMX_SC_PM_CLK_MISC0>;
> +		clock-names = "pix", "axi", "disp_axi";
> +		assigned-clocks = <&clk IMX_SC_R_LCD_0 IMX_SC_PM_CLK_PER>,
> +				  <&clk IMX_SC_R_LCD_0 IMX_SC_PM_CLK_MISC0>,
> +				  <&clk IMX_SC_R_ELCDIF_PLL IMX_SC_PM_CLK_PLL>;
> +		assigned-clock-parents = <&clk IMX_SC_R_ELCDIF_PLL IMX_SC_PM_CLK_PLL>,
> +					 <&clk IMX_SC_R_LCD_0 IMX_SC_PM_CLK_BYPASS>;
> +		assigned-clock-rates = <0>, <24000000>, <804000000>;
> +		interrupts = <GIC_SPI 62 IRQ_TYPE_LEVEL_HIGH>;
> +		power-domains = <&pd IMX_SC_R_LCD_0>;
> +		status = "disabled";
> +	};
> +
>  	adma_pwm: pwm@5a190000 {
>  		compatible = "fsl,imx8qxp-pwm", "fsl,imx27-pwm";
>  		reg = <0x5a190000 0x1000>;
> @@ -298,6 +316,17 @@ uart3_lpcg: clock-controller@5a490000 {
>  		power-domains = <&pd IMX_SC_R_UART_3>;
>  	};
>  
> +	lcd_clk_lpcg: clock-controller@5a580000 {
> +		compatible = "fsl,imx8qxp-lpcg";
> +		reg = <0x5a580000 0x4>;
> +		#clock-cells = <1>;
> +		clocks = <&clk IMX_SC_R_LCD_0 IMX_SC_PM_CLK_PER>,
> +			 <&dma_ipg_clk>;
> +		clock-indices = <IMX_LPCG_CLK_0>, <IMX_LPCG_CLK_4>;
> +		clock-output-names = "lcd_clk_lpcg", "lcd_ipg_clk";
> +		power-domains = <&pd IMX_SC_R_LCD_0>;
> +	};

[Severity: Medium]
Should arch/arm64/boot/dts/freescale/imx8qm-ss-dma.dtsi also delete
&adma_lcdif and &lcd_clk_lpcg?

arch/arm64/boot/dts/freescale/imx8-ss-dma.dtsi is included by
arch/arm64/boot/dts/freescale/imx8qm.dtsi:

	#include "imx8-ss-img.dtsi"
	#include "imx8-ss-dma.dtsi"
	#include "imx8-ss-conn.dtsi"
	...
};

#include "imx8qm-ss-img.dtsi"
#include "imx8qm-ss-dma.dtsi"

i.MX8QM does not have the adma_lcdif peripheral or the IMX_SC_R_LCD_0 and
IMX_SC_R_ELCDIF_PLL resources (which are omitted from
imx8qm_clk_scu_rsrc_table[] in drivers/clk/imx/clk-imx8qm-rsrc.c).

For the companion adma_pwm and adma_pwm_lpcg nodes,
arch/arm64/boot/dts/freescale/imx8qm-ss-dma.dtsi explicitly removes them
(added in commit e07c4c0a5972 ("arm64: dts: imx8: delete pwm nodes in dma
subsys for imx8qm")):

/delete-node/ &adma_pwm;
/delete-node/ &adma_pwm_lpcg;

Without deleting &adma_lcdif and &lcd_clk_lpcg in imx8qm-ss-dma.dtsi, all
i.MX8QM and i.MX8QP device trees inherit both nodes, and lcd_clk_lpcg
(clock-controller@5a580000) is enabled by default.

On i.MX8QM, does this cause U-Boot resource partition ownership warnings
and cause Linux probing of 5a580000.clock-controller to fail power domain
attachment in imx_scu_add_pm_domain() -> imx_scu_pd_xlate() (-ENOENT) and
clock lookup in imx_clk_scu_alloc_dev() -> imx_scu_of_clk_src_get()
(-ENODEV) for IMX_SC_R_LCD_0?

> +
>  	adma_pwm_lpcg: clock-controller@5a590000 {
>  		compatible = "fsl,imx8qxp-lpcg";
>  		reg = <0x5a590000 0x10000>;

[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261020-dxl_lcdif-v3-0-adc944659e14@nxp.com?part=4