[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <a58956a3-b9cd-408f-98af-7ec31f2277dd@oss.qualcomm.com>
Date: Thu, 23 Oct 2025 16:26:36 +0200
From: Konrad Dybcio <konrad.dybcio@....qualcomm.com>
To: Manikanta Mylavarapu <quic_mmanikan@...cinc.com>, andersson@...nel.org,
konradybcio@...nel.org, robh@...nel.org, krzk+dt@...nel.org,
conor+dt@...nel.org, linux-arm-msm@...r.kernel.org,
devicetree@...r.kernel.org, linux-kernel@...r.kernel.org
Cc: quic_srichara@...cinc.com, quic_varada@...cinc.com,
kathiravan.thirumoorthy@....qualcomm.com
Subject: Re: [PATCH v1] arm64: dts: qcom: ipq5424: add cooling maps for CPU
thermal zones
On 10/23/25 2:21 PM, Manikanta Mylavarapu wrote:
>
>
> On 10/23/2025 2:09 PM, Konrad Dybcio wrote:
>> On 10/23/25 6:38 AM, Manikanta Mylavarapu wrote:
>>> Add cooling-maps to the cpu1, cpu2, and cpu3 thermal zones to associate
>>> passive trip points with CPU cooling devices. This enables proper
>>> thermal mitigation by allowing the thermal framework to throttle CPUs
>>> based on temperature thresholds. Also, label the trip points to allow
>>> referencing them in the cooling maps.
>>>
>>> Signed-off-by: Manikanta Mylavarapu <quic_mmanikan@...cinc.com>
>>> ---
>>
>> [...]
>>
>>> cpu1-thermal {
>>> @@ -1254,18 +1269,28 @@ cpu1-thermal {
>>> thermal-sensors = <&tsens 12>;
>>>
>>> trips {
>>> - cpu-critical {
>>> + cpu1_crit: cpu-critical {
>>> temperature = <120000>;
>>> hysteresis = <9000>;
>>> type = "critical";
>>> };
>>>
>>> - cpu-passive {
>>> + cpu1_alert: cpu-passive {
>>> temperature = <110000>;
>>> hysteresis = <9000>;
>>> type = "passive";
>>> };
>>
>> This means "software will start throttling the cpufreq at 110C"
>>
>> Is this what you want?
>>
>
> Hi Konrad,
>
> Thank you for reviewing the patch.
>
> Yes, the intent is to initiate software-based cpufreq throttling at 110°C.
>
>> Also, is there nothing like LMH on this platform that throttles the
>> device way earlier, without OS intervention?
>>
>
> IPQ5424 platform does not support LMH.
Thanks for confirming
Usually it seems like the trips are ordered by temperature but I don't
think there's a strong preference
The hysteresis value is a little high, but I'm assuming that came out of
some internal tuning
Reviewed-by: Konrad Dybcio <konrad.dybcio@....qualcomm.com>
Konrad
Powered by blists - more mailing lists