* [PATCH] arm64: dts: imx8mp: Enable gpu passive throttling
@ 2025-05-08 10:18 Martin Kepplinger-Novaković
2025-05-27 6:04 ` Martin Kepplinger-Novakovic
2025-06-11 9:16 ` Shawn Guo
0 siblings, 2 replies; 3+ messages in thread
From: Martin Kepplinger-Novaković @ 2025-05-08 10:18 UTC (permalink / raw)
To: robh, krzk+dt, conor+dt, shawnguo, s.hauer
Cc: kernel, festevam, devicetree, imx, linux-arm-kernel, linux-kernel,
Martin Kepplinger-Novaković
Hook up the gpu as a passive cooling device to the thermal zones' alert
trip point just like the cpu.
The gpu here consists of 3D GPU, 2D GPU and NPU.
One way to test would be to set one "alert" trip point low enough
and watch the cooling device state increase:
echo 10000 > /sys/class/thermal/thermal_zone0/trip_point_0_temp
watch cat /sys/class/thermal/cooling_device*/cur_state
And of course set the trip point back to its original value and watch
the cooling device states jump to 0 again.
Signed-off-by: Martin Kepplinger-Novaković <martink@posteo.de>
---
arch/arm64/boot/dts/freescale/imx8mp.dtsi | 13 +++++++++++--
1 file changed, 11 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/boot/dts/freescale/imx8mp.dtsi b/arch/arm64/boot/dts/freescale/imx8mp.dtsi
index 50a07c56faffc..dea9342d071c4 100644
--- a/arch/arm64/boot/dts/freescale/imx8mp.dtsi
+++ b/arch/arm64/boot/dts/freescale/imx8mp.dtsi
@@ -320,7 +320,10 @@ map0 {
<&A53_0 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
<&A53_1 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
<&A53_2 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
- <&A53_3 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>;
+ <&A53_3 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
+ <&gpu3d THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
+ <&gpu2d THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
+ <&npu THERMAL_NO_LIMIT THERMAL_NO_LIMIT>;
};
};
};
@@ -350,7 +353,10 @@ map0 {
<&A53_0 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
<&A53_1 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
<&A53_2 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
- <&A53_3 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>;
+ <&A53_3 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
+ <&gpu3d THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
+ <&gpu2d THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
+ <&npu THERMAL_NO_LIMIT THERMAL_NO_LIMIT>;
};
};
};
@@ -2229,6 +2235,7 @@ gpu3d: gpu@38000000 {
<&clk IMX8MP_CLK_GPU_ROOT>,
<&clk IMX8MP_CLK_GPU_AHB>;
clock-names = "core", "shader", "bus", "reg";
+ #cooling-cells = <2>;
assigned-clocks = <&clk IMX8MP_CLK_GPU3D_CORE>,
<&clk IMX8MP_CLK_GPU3D_SHADER_CORE>;
assigned-clock-parents = <&clk IMX8MP_SYS_PLL2_1000M>,
@@ -2245,6 +2252,7 @@ gpu2d: gpu@38008000 {
<&clk IMX8MP_CLK_GPU_ROOT>,
<&clk IMX8MP_CLK_GPU_AHB>;
clock-names = "core", "bus", "reg";
+ #cooling-cells = <2>;
assigned-clocks = <&clk IMX8MP_CLK_GPU2D_CORE>;
assigned-clock-parents = <&clk IMX8MP_SYS_PLL2_1000M>;
assigned-clock-rates = <1000000000>;
@@ -2302,6 +2310,7 @@ npu: npu@38500000 {
<&clk IMX8MP_CLK_ML_AXI>,
<&clk IMX8MP_CLK_ML_AHB>;
clock-names = "core", "shader", "bus", "reg";
+ #cooling-cells = <2>;
power-domains = <&pgc_mlmix>;
};
--
2.39.5
^ permalink raw reply related [flat|nested] 3+ messages in thread* Re: [PATCH] arm64: dts: imx8mp: Enable gpu passive throttling
2025-05-08 10:18 [PATCH] arm64: dts: imx8mp: Enable gpu passive throttling Martin Kepplinger-Novaković
@ 2025-05-27 6:04 ` Martin Kepplinger-Novakovic
2025-06-11 9:16 ` Shawn Guo
1 sibling, 0 replies; 3+ messages in thread
From: Martin Kepplinger-Novakovic @ 2025-05-27 6:04 UTC (permalink / raw)
To: robh, krzk+dt, conor+dt, shawnguo, s.hauer
Cc: kernel, festevam, devicetree, imx, linux-arm-kernel, linux-kernel
Am Donnerstag, dem 08.05.2025 um 10:18 +0000 schrieb Martin Kepplinger-
Novaković:
> Hook up the gpu as a passive cooling device to the thermal zones'
> alert
> trip point just like the cpu.
>
> The gpu here consists of 3D GPU, 2D GPU and NPU.
>
> One way to test would be to set one "alert" trip point low enough
> and watch the cooling device state increase:
>
> echo 10000 > /sys/class/thermal/thermal_zone0/trip_point_0_temp
> watch cat /sys/class/thermal/cooling_device*/cur_state
>
> And of course set the trip point back to its original value and watch
> the cooling device states jump to 0 again.
>
> Signed-off-by: Martin Kepplinger-Novaković <martink@posteo.de>
> ---
> arch/arm64/boot/dts/freescale/imx8mp.dtsi | 13 +++++++++++--
> 1 file changed, 11 insertions(+), 2 deletions(-)
>
> diff --git a/arch/arm64/boot/dts/freescale/imx8mp.dtsi
> b/arch/arm64/boot/dts/freescale/imx8mp.dtsi
> index 50a07c56faffc..dea9342d071c4 100644
> --- a/arch/arm64/boot/dts/freescale/imx8mp.dtsi
> +++ b/arch/arm64/boot/dts/freescale/imx8mp.dtsi
> @@ -320,7 +320,10 @@ map0 {
> <&A53_0
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> <&A53_1
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> <&A53_2
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> - <&A53_3
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>;
> + <&A53_3
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> + <&gpu3d
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> + <&gpu2d
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> + <&npu
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>;
> };
> };
> };
> @@ -350,7 +353,10 @@ map0 {
> <&A53_0
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> <&A53_1
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> <&A53_2
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> - <&A53_3
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>;
> + <&A53_3
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> + <&gpu3d
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> + <&gpu2d
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> + <&npu
> THERMAL_NO_LIMIT THERMAL_NO_LIMIT>;
> };
> };
> };
> @@ -2229,6 +2235,7 @@ gpu3d: gpu@38000000 {
> <&clk IMX8MP_CLK_GPU_ROOT>,
> <&clk IMX8MP_CLK_GPU_AHB>;
> clock-names = "core", "shader", "bus", "reg";
> + #cooling-cells = <2>;
> assigned-clocks = <&clk
> IMX8MP_CLK_GPU3D_CORE>,
> <&clk
> IMX8MP_CLK_GPU3D_SHADER_CORE>;
> assigned-clock-parents = <&clk
> IMX8MP_SYS_PLL2_1000M>,
> @@ -2245,6 +2252,7 @@ gpu2d: gpu@38008000 {
> <&clk IMX8MP_CLK_GPU_ROOT>,
> <&clk IMX8MP_CLK_GPU_AHB>;
> clock-names = "core", "bus", "reg";
> + #cooling-cells = <2>;
> assigned-clocks = <&clk
> IMX8MP_CLK_GPU2D_CORE>;
> assigned-clock-parents = <&clk
> IMX8MP_SYS_PLL2_1000M>;
> assigned-clock-rates = <1000000000>;
> @@ -2302,6 +2310,7 @@ npu: npu@38500000 {
> <&clk IMX8MP_CLK_ML_AXI>,
> <&clk IMX8MP_CLK_ML_AHB>;
> clock-names = "core", "shader", "bus", "reg";
> + #cooling-cells = <2>;
> power-domains = <&pgc_mlmix>;
> };
>
hi Shawn and all,
Do you have any objections or second thoughts about this?
thank you,
martin
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH] arm64: dts: imx8mp: Enable gpu passive throttling
2025-05-08 10:18 [PATCH] arm64: dts: imx8mp: Enable gpu passive throttling Martin Kepplinger-Novaković
2025-05-27 6:04 ` Martin Kepplinger-Novakovic
@ 2025-06-11 9:16 ` Shawn Guo
1 sibling, 0 replies; 3+ messages in thread
From: Shawn Guo @ 2025-06-11 9:16 UTC (permalink / raw)
To: Martin Kepplinger-Novaković
Cc: robh, krzk+dt, conor+dt, shawnguo, s.hauer, kernel, festevam,
devicetree, imx, linux-arm-kernel, linux-kernel
On Thu, May 08, 2025 at 10:18:02AM +0000, Martin Kepplinger-Novaković wrote:
> Hook up the gpu as a passive cooling device to the thermal zones' alert
> trip point just like the cpu.
>
> The gpu here consists of 3D GPU, 2D GPU and NPU.
>
> One way to test would be to set one "alert" trip point low enough
> and watch the cooling device state increase:
>
> echo 10000 > /sys/class/thermal/thermal_zone0/trip_point_0_temp
> watch cat /sys/class/thermal/cooling_device*/cur_state
>
> And of course set the trip point back to its original value and watch
> the cooling device states jump to 0 again.
>
> Signed-off-by: Martin Kepplinger-Novaković <martink@posteo.de>
Applied, thanks!
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2025-06-11 9:17 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-05-08 10:18 [PATCH] arm64: dts: imx8mp: Enable gpu passive throttling Martin Kepplinger-Novaković
2025-05-27 6:04 ` Martin Kepplinger-Novakovic
2025-06-11 9:16 ` Shawn Guo
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox