Re: [PATCH v3] arm64: dts: qcom: x1e78100-thinkpad-t14s: Add thermal zones for PMICs

[Stephan Gerhold <stephan.gerhold@linaro.org>]

On Wed, Oct 14, 2026 at 09:35:24PM +0200, Daniel Lezcano wrote:
> Add thermal zones for the ThinkPad T14s and describe the board
> temperature sensors exposed through the PMK8550 VADC.
> 
> Add a passive thermal zone for the keyboard area, which is the main
> thermal constraint under sustained load. The EC reports a thermal alert
> at 57.85°C, so use this value as the passive trip point and allow the
> thermal framework to throttle all CPUs and the GPU when the threshold is
> reached. Use a 1000 m°C hysteresis and a 500 ms passive polling interval.
> 
> An earlier approach used staged CPU cluster throttling in an attempt to
> start thermal mitigation before reaching the target temperature. Testing
> did not show a significant improvement compared to throttling all CPUs at
> the EC thermal alert threshold. It also introduced a latency imbalance
> between the different CPU clusters, which could negatively affect
> workloads depending on where tasks are scheduled.
> 
> A more optimal solution would require a significantly more complex,
> system-wide approach, accounting for the workload distribution and the
> available power and thermal headroom across the CPU clusters and other
> devices. Such an approach would also require kernel mechanisms which are
> not currently available. Keep the thermal policy simple here and use a
> single passive trip, which at least ensures that the system is actively
> mitigated before reaching temperatures that could ultimately lead to a
> thermal shutdown.
> 
> The mitigation temperature is derived from the platform behaviour when
> using the balanced power profile available under Windows. A separate
> change will adjust this trip point dynamically according to events
> reported by the embedded controller. Name the trip
> "keyboard-ec-alert" so that the EC driver can identify the corresponding
> trip point in the device tree.
> 
> Also add a 73°C critical trip to shut the system down if the keyboard
> temperature reaches an unsafe level.
> 
> Add a thermal zone for the back cover with a 73°C critical shutdown
> trip.
> 
> Describe the SoC, SMB, SDX, east and west board temperature sensors as
> monitoring-only thermal zones.
> 
> Thermal camera measurements were also used to compare the reported
> temperatures with the physical temperature of the chassis. The measured
> values appear to primarily represent temperatures inside the enclosure
> rather than the user-contact surface temperature. No unsafe external
> surface temperature was observed during these tests, although this should
> not be considered a substitute for formal platform thermal qualification.
> 

Thanks to Mark, I was able to get some info on the location/purpose of
the thermal sensors in the T14s. As you noted already here, the sensors
are primarily focused on specific components inside the enclosure rather
than user-contact surface temperatures. Below I added some suggestions
for more accurate names for the thermal zones. Perhaps that helps to
optimize the thermal control further.

> Finally, enable ADC thermal monitoring on the corresponding PMK8550 VADC
> channels used by these thermal zones.
> 
> Signed-off-by: Daniel Lezcano <daniel.lezcano@oss.qualcomm.com>
> ---
>  .../qcom/x1e78100-lenovo-thinkpad-t14s.dtsi   | 105 ++++++++++++++++++
>  1 file changed, 105 insertions(+)
> 
> diff --git a/arch/arm64/boot/dts/qcom/x1e78100-lenovo-thinkpad-t14s.dtsi b/arch/arm64/boot/dts/qcom/x1e78100-lenovo-thinkpad-t14s.dtsi
> index 75ad74bcb042..2d0a434d2248 100644
> --- a/arch/arm64/boot/dts/qcom/x1e78100-lenovo-thinkpad-t14s.dtsi
> +++ b/arch/arm64/boot/dts/qcom/x1e78100-lenovo-thinkpad-t14s.dtsi
> @@ -502,6 +502,81 @@ platform {
>  			};
>  		};
>  	};
> +
> +	thermal-zones {
> +		backcover-thermal {
> +			critical-action = "shutdown";
> +			thermal-sensors = <&pmk8550_vadc ADC5_GEN3_AMUX1_THM_100K_PU(1)>;

dram-thermal or memory-thermal (or maybe keyboard-thermal if you want to
give it a user-surface name, although that purpose seems to be
secondary).

Next to DRAM memory, on the top part of the mainboard, facing upwards to
the keyboard. Unrelated to the backcover.

> +
> +			trips {
> +				backcover-critical {
> +					temperature = <73000>;
> +					hysteresis = <1000>;
> +					type = "critical";
> +				};

Is this critical temperature is appropriate for the DRAM temperature?
Windows seems to have 115°C here as critical trip point which is maybe a
bit too high.

> +			};
> +		};
> +
> +		east-thermal {
> +			thermal-sensors = <&pmk8550_vadc ADC5_GEN3_AMUX4_THM_100K_PU(1)>;

gpu-pmic-thermal (next to GPU-related PMICs close to the SoC)

> +		};
> +
> +		keyboard-thermal {
> +			critical-action = "shutdown";
> +			polling-delay-passive = <500>;
> +
> +			thermal-sensors = <&pmk8550_vadc ADC5_GEN3_AMUX2_GPIO_100K_PU(1)>;

fan-thermal or heatpipe-thermal (or maybe backcover-thermal if you want
to give it a user-surface name, although that purpose seems to be
secondary).

Next to fan/heat pipe, on the bottom part of the mainboard, facing
downwards to the back cover. Unrelated to the keyboard.

> +
> +			trips {
> +				keyboard_ec_alert: keyboard-ec-alert {
> +					temperature = <57850>;
> +					hysteresis = <1000>;
> +					type = "passive";
> +				};
> +
> +				keyboard-critical {
> +					temperature = <73000>;
> +					hysteresis = <1000>;
> +					type = "critical";
> +				};
> +			};
> +
> +			cooling-maps {
> +				map0 {
> +					trip = <&keyboard_ec_alert>;
> +					cooling-device = <&cpu0 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> +							 <&cpu1 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> +							 <&cpu2 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> +							 <&cpu3 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> +							 <&cpu4 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> +							 <&cpu5 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> +							 <&cpu6 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> +							 <&cpu7 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> +							 <&cpu8 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> +							 <&cpu9 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> +							 <&cpu10 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> +							 <&cpu11 THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> +							 <&gpu THERMAL_NO_LIMIT THERMAL_NO_LIMIT>;
> +				};
> +			};
> +		};
> +
> +		sdx-thermal {
> +			thermal-sensors = <&pmk8550_vadc ADC5_GEN3_AMUX3_THM_100K_PU(1)>;

wwan-thermal

> +		};
> +
> +		smb-thermal {
> +			thermal-sensors = <&pmk8550_vadc ADC5_GEN3_AMUX2_THM_100K_PU(1)>;

charger-thermal

> +		};
> +
> +		soc-thermal {
> +			thermal-sensors = <&pmk8550_vadc ADC5_GEN3_AMUX1_GPIO_100K_PU(1)>;

cpu-pmic-thermal (next to CPU-related PMICs close to SoC)

> +		};
> +
> +		west-thermal {
> +			thermal-sensors = <&pmk8550_vadc ADC5_GEN3_AMUX5_THM_100K_PU(1)>;

typec-thermal or usbc-thermal

Next to USB-C ports, bottom part of mainboard, facing towards the back
cover.

> +		};
> +	};

There is also ssd-thermal @ SYS_THERM7
thermal-sensors = <&pmk8550_adc_tm ADC5_GEN3_AMUX6_THM_100K_PU(1)>;
as I mentioned before.

Windows uses the EC sensors for throttling, although given that we don't
have interrupt support for the EC yet, I think it is fine to initially
use the PMIC sensors.

Thanks,
Stephan