2021-08-31 13:43:05

by Michael Walle

[permalink] [raw]
Subject: [PATCH 2/7] arm64: dts: ls1028a: move Mali DP500 node into /soc

Move it inside the /soc subnode because it is part of the CCSR space.

Signed-off-by: Michael Walle <[email protected]>
---
.../arm64/boot/dts/freescale/fsl-ls1028a.dtsi | 41 ++++++++++---------
1 file changed, 21 insertions(+), 20 deletions(-)

diff --git a/arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi b/arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi
index 9a65a7118faa..92e4f004c1c2 100644
--- a/arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi
+++ b/arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi
@@ -793,6 +793,27 @@ QORIQ_CLK_PLL_DIV(16)>,
clock-names = "wdog_clk", "apb_pclk";
};

+ malidp0: display@f080000 {
+ compatible = "arm,mali-dp500";
+ reg = <0x0 0xf080000 0x0 0x10000>;
+ interrupts = <0 222 IRQ_TYPE_LEVEL_HIGH>,
+ <0 223 IRQ_TYPE_LEVEL_HIGH>;
+ interrupt-names = "DE", "SE";
+ clocks = <&dpclk>,
+ <&clockgen QORIQ_CLK_HWACCEL 2>,
+ <&clockgen QORIQ_CLK_HWACCEL 2>,
+ <&clockgen QORIQ_CLK_HWACCEL 2>;
+ clock-names = "pxlclk", "mclk", "aclk", "pclk";
+ arm,malidp-output-port-lines = /bits/ 8 <8 8 8>;
+ arm,malidp-arqos-value = <0xd000d000>;
+
+ port {
+ dpi0_out: endpoint {
+
+ };
+ };
+ };
+
sai1: audio-controller@f100000 {
#sound-dai-cells = <0>;
compatible = "fsl,vf610-sai";
@@ -1139,24 +1160,4 @@ ftm_alarm0: timer@2800000 {
};
};

- malidp0: display@f080000 {
- compatible = "arm,mali-dp500";
- reg = <0x0 0xf080000 0x0 0x10000>;
- interrupts = <0 222 IRQ_TYPE_LEVEL_HIGH>,
- <0 223 IRQ_TYPE_LEVEL_HIGH>;
- interrupt-names = "DE", "SE";
- clocks = <&dpclk>,
- <&clockgen QORIQ_CLK_HWACCEL 2>,
- <&clockgen QORIQ_CLK_HWACCEL 2>,
- <&clockgen QORIQ_CLK_HWACCEL 2>;
- clock-names = "pxlclk", "mclk", "aclk", "pclk";
- arm,malidp-output-port-lines = /bits/ 8 <8 8 8>;
- arm,malidp-arqos-value = <0xd000d000>;
-
- port {
- dp0_out: endpoint {
-
- };
- };
- };
};
--
2.30.2


2022-01-21 21:09:31

by Leo Li

[permalink] [raw]
Subject: RE: [PATCH 2/7] arm64: dts: ls1028a: move Mali DP500 node into /soc



> -----Original Message-----
> From: Michael Walle <[email protected]>
> Sent: Tuesday, August 31, 2021 9:40 PM
> To: [email protected]; [email protected]; linux-
> [email protected]
> Cc: Shawn Guo <[email protected]>; Leo Li <[email protected]>; Rob
> Herring <[email protected]>; Vladimir Oltean <[email protected]>;
> Michael Walle <[email protected]>
> Subject: [PATCH 2/7] arm64: dts: ls1028a: move Mali DP500 node into /soc
>
> Move it inside the /soc subnode because it is part of the CCSR space.

I just noticed that the dp0_out label has been changed to dpi0_out besides the move. Is this an intentional change or a typo? If intentional we probably should mention it, otherwise we should change it back as it is breaking build for off-tree patch that uses the label.

>
> Signed-off-by: Michael Walle <[email protected]>
> ---
> .../arm64/boot/dts/freescale/fsl-ls1028a.dtsi | 41 ++++++++++---------
> 1 file changed, 21 insertions(+), 20 deletions(-)
>
> diff --git a/arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi
> b/arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi
> index 9a65a7118faa..92e4f004c1c2 100644
> --- a/arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi
> +++ b/arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi
> @@ -793,6 +793,27 @@ QORIQ_CLK_PLL_DIV(16)>,
> clock-names = "wdog_clk", "apb_pclk";
> };
>
> + malidp0: display@f080000 {
> + compatible = "arm,mali-dp500";
> + reg = <0x0 0xf080000 0x0 0x10000>;
> + interrupts = <0 222 IRQ_TYPE_LEVEL_HIGH>,
> + <0 223 IRQ_TYPE_LEVEL_HIGH>;
> + interrupt-names = "DE", "SE";
> + clocks = <&dpclk>,
> + <&clockgen QORIQ_CLK_HWACCEL 2>,
> + <&clockgen QORIQ_CLK_HWACCEL 2>,
> + <&clockgen QORIQ_CLK_HWACCEL 2>;
> + clock-names = "pxlclk", "mclk", "aclk", "pclk";
> + arm,malidp-output-port-lines = /bits/ 8 <8 8 8>;
> + arm,malidp-arqos-value = <0xd000d000>;
> +
> + port {
> + dpi0_out: endpoint {
> +
> + };
> + };
> + };
> +
> sai1: audio-controller@f100000 {
> #sound-dai-cells = <0>;
> compatible = "fsl,vf610-sai";
> @@ -1139,24 +1160,4 @@ ftm_alarm0: timer@2800000 {
> };
> };
>
> - malidp0: display@f080000 {
> - compatible = "arm,mali-dp500";
> - reg = <0x0 0xf080000 0x0 0x10000>;
> - interrupts = <0 222 IRQ_TYPE_LEVEL_HIGH>,
> - <0 223 IRQ_TYPE_LEVEL_HIGH>;
> - interrupt-names = "DE", "SE";
> - clocks = <&dpclk>,
> - <&clockgen QORIQ_CLK_HWACCEL 2>,
> - <&clockgen QORIQ_CLK_HWACCEL 2>,
> - <&clockgen QORIQ_CLK_HWACCEL 2>;
> - clock-names = "pxlclk", "mclk", "aclk", "pclk";
> - arm,malidp-output-port-lines = /bits/ 8 <8 8 8>;
> - arm,malidp-arqos-value = <0xd000d000>;
> -
> - port {
> - dp0_out: endpoint {
> -
> - };
> - };
> - };
> };
> --
> 2.30.2

2022-01-21 21:09:53

by Michael Walle

[permalink] [raw]
Subject: Re: [PATCH 2/7] arm64: dts: ls1028a: move Mali DP500 node into /soc

Am 2022-01-20 09:06, schrieb Leo Li:
>> -----Original Message-----
>> From: Michael Walle <[email protected]>
>> Sent: Tuesday, August 31, 2021 9:40 PM
>> To: [email protected]; [email protected];
>> linux-
>> [email protected]
>> Cc: Shawn Guo <[email protected]>; Leo Li <[email protected]>; Rob
>> Herring <[email protected]>; Vladimir Oltean
>> <[email protected]>;
>> Michael Walle <[email protected]>
>> Subject: [PATCH 2/7] arm64: dts: ls1028a: move Mali DP500 node into
>> /soc
>>
>> Move it inside the /soc subnode because it is part of the CCSR space.
>
> I just noticed that the dp0_out label has been changed to dpi0_out
> besides the move. Is this an intentional change or a typo? If
> intentional we probably should mention it, otherwise we should change
> it back as it is breaking build for off-tree patch that uses the
> label.

It's intentional, because dp0_out might sound like displayport. And
from what I've found the output is a display pixel interface, hence dpi.

I agree, that this should have been two patches, must have slipped.
But this series was merged months ago, so we can't change anything
anymore.

Besides, given the fact that there is no support for the DisplayPort
PHY in upstream (yet, I'm working on that), I doubt there are any
out-of-tree device trees, which don't use the proprietary NXP
driver.

Also, if there is an out-of-tree device tree, it should be easy
enough for NXP to change that :)

-michael