[PATCH v2] arm64: dts: imx8mp: Add Hantro G1, G2 DT nodes

[Lucas Stach]

Am Sonntag, dem 18.12.2022 um 05:20 +0100 schrieb Marek Vasut:
> Add DT nodes for the Hantro VPU found in i.MX8MP SoC.
> 
> Signed-off-by: Marek Vasut <marex at denx.de>
> ---
> Cc: Abel Vesa <abel.vesa at nxp.com>
> Cc: Adam Ford <aford173 at gmail.com>
> Cc: Alexander Stein <alexander.stein at ew.tq-group.com>
> Cc: Ezequiel Garcia <ezequiel at vanguardiasur.com.ar>
> Cc: Fabio Estevam <festevam at gmail.com>
> Cc: Jacky Bai <ping.bai at nxp.com>
> Cc: Krzysztof Kozlowski <krzysztof.kozlowski+dt at linaro.org>
> Cc: Laurent Pinchart <laurent.pinchart at ideasonboard.com>
> Cc: Lucas Stach <l.stach at pengutronix.de>
> Cc: Marek Vasut <marex at denx.de>
> Cc: Markus Niebel <Markus.Niebel at ew.tq-group.com>
> Cc: NXP Linux Team <linux-imx at nxp.com>
> Cc: Peng Fan <peng.fan at nxp.com>
> Cc: Pengutronix Kernel Team <kernel at pengutronix.de>
> Cc: Richard Cochran <richardcochran at gmail.com>
> Cc: Richard Zhu <hongxing.zhu at nxp.com>
> Cc: Rob Herring <robh+dt at kernel.org>
> Cc: Sascha Hauer <s.hauer at pengutronix.de>
> Cc: Shawn Guo <shawnguo at kernel.org>
> To: linux-arm-kernel at lists.infradead.org
> ---
> V2: Drop the VC8000E
> ---
>  arch/arm64/boot/dts/freescale/imx8mp.dtsi | 25 +++++++++++++++++++++++
>  1 file changed, 25 insertions(+)
> 
> diff --git a/arch/arm64/boot/dts/freescale/imx8mp.dtsi b/arch/arm64/boot/dts/freescale/imx8mp.dtsi
> index 34c2c80c47eb5..d472170ee9ff1 100644
> --- a/arch/arm64/boot/dts/freescale/imx8mp.dtsi
> +++ b/arch/arm64/boot/dts/freescale/imx8mp.dtsi
> @@ -1639,6 +1639,28 @@ gpu2d: gpu at 38008000 {
>  			power-domains = <&pgc_gpu2d>;
>  		};
>  
> +		vpu_g1: video-codec at 38300000 {
> +			compatible = "nxp,imx8mm-vpu-g1";
> +			reg = <0x38300000 0x10000>;
> +			interrupts = <GIC_SPI 7 IRQ_TYPE_LEVEL_HIGH>;
> +			clocks = <&clk IMX8MP_CLK_VPU_G1_ROOT>;
> +			assigned-clocks = <&clk IMX8MP_CLK_VPU_G1>;
> +			assigned-clock-parents = <&clk IMX8MP_VPU_PLL_OUT>;
> +			assigned-clock-rates = <700000000>;

Where does this clock rate come from? G1 nominal rate is 600MHz,
overdrive rate is 800MHz. The general convention is to put the nominal
rate into the assigned clocks. Overdrive rates need proper OPP support
in the drivers and is not generally enabled right now.

> +			power-domains = <&vpumix_blk_ctrl IMX8MP_VPUBLK_PD_G1>;
> +		};
> +
> +		vpu_g2: video-codec at 38310000 {
> +			compatible = "nxp,imx8mq-vpu-g2";
> +			reg = <0x38310000 0x10000>;
> +			interrupts = <GIC_SPI 8 IRQ_TYPE_LEVEL_HIGH>;
> +			clocks = <&clk IMX8MP_CLK_VPU_G2_ROOT>;
> +			assigned-clocks = <&clk IMX8MP_CLK_VPU_G2>;
> +			assigned-clock-parents = <&clk IMX8MP_VPU_PLL_OUT>;
> +			assigned-clock-rates = <700000000>;

Same comment as above G2 nominal rate is 500MHz. You might want to
source this from SYS_PLL2, instead of the VPU PLL.

Regards,
Lucas
> +			power-domains = <&vpumix_blk_ctrl IMX8MP_VPUBLK_PD_G2>;
> +		};
> +
>  		vpumix_blk_ctrl: blk-ctrl at 38330000 {
>  			compatible = "fsl,imx8mp-vpu-blk-ctrl", "syscon";
>  			reg = <0x38330000 0x100>;
> @@ -1650,6 +1672,9 @@ vpumix_blk_ctrl: blk-ctrl at 38330000 {
>  				 <&clk IMX8MP_CLK_VPU_G2_ROOT>,
>  				 <&clk IMX8MP_CLK_VPU_VC8KE_ROOT>;
>  			clock-names = "g1", "g2", "vc8000e";
> +			assigned-clocks = <&clk IMX8MP_CLK_VPU_BUS>;
> +			assigned-clock-parents = <&clk IMX8MP_VPU_PLL_OUT>;
> +			assigned-clock-rates = <700000000>;

VPU_BUS is the parent of VPU_ROOT and should be assigned in the
pgc_vpumix node, where it is used. Again nominal clock frequency
600MHz. But then maybe it would make sense to limit this to 500MHz for
now to be able to source it from a system PLL.

Regards,
Lucas
>  			interconnects = <&noc IMX8MP_ICM_VPU_G1 &noc IMX8MP_ICN_VIDEO>,
>  					<&noc IMX8MP_ICM_VPU_G2 &noc IMX8MP_ICN_VIDEO>,
>  					<&noc IMX8MP_ICM_VPU_H1 &noc IMX8MP_ICN_VIDEO>;