Re: [PATCH 1/8] dt-bindings: embedded-controller: qcom,hamoa-crd-ec: Add qcom,tsens

[Anvesh Jain P <anvesh.p@oss.qualcomm.com>]

On 8/19/2026 10:08 PM, Konrad Dybcio wrote:
>>>> Each tsens provider has more physical sensors (16) than the count used
>>>> here. The leading N we take exactly matches the number of sensors
>>>> already wired to board thermal-zones for other purposes (CPU clusters,
>>>> GPU, camera, etc.) on that provider — e.g. on hamoa, tsens2 has 11 such
>>>> zones and tsens3 has 15, matching qcom,tsens = <&tsens2 11>, <&tsens3
>>>> 15>;. The remaining sensor IDs aren't referenced by any thermal-zone on
>>>> this platform, so there's no per-purpose reading to include in the SoC
>>>> Tj average.
>>> What you said is true
>>>
>>> It also doesn't answer the question why that isn't the case
>>>
>>> Konrad
>> What I meant is, the driver looks up a thermal zone for each sensor id,
>> and skips the ones that don't have a zone in DT. Sensor ids past 11/15
>> on this board don't have a zone assigned, so there's nothing to read there.
> You can keep repeating what you said and I can keep repeating you're not
> answering the question "why do we only choose to read the data from
> certain sensors and ignore others"
> 
> Konrad

We only read sensors that are physically wired into the board's thermal
zones, those actually measure relevant parts of the SoC. The rest aren't
connected to anything, so we are ignoring those. This is what I am
trying to say.


-- 
Best Regards,
Anvesh