Re: [PATCH v3 2/2] arm64: dts: qcom: sm8650: setup gpu thermal with higher temperatures

From: Konrad Dybcio
Date: Mon Feb 03 2025 - 08:08:09 EST


On 3.02.2025 9:23 AM, neil.armstrong@xxxxxxxxxx wrote:
> On 01/02/2025 16:37, Konrad Dybcio wrote:
>> On 29.01.2025 3:41 PM, Neil Armstrong wrote:
>>> On 29/01/2025 10:43, Neil Armstrong wrote:
>>>> On the SM8650, the dynamic clock and voltage scaling (DCVS) for the GPU
>>>> is done from the HLOS, but the GPU can achieve a much higher temperature
>>>> before failing according the the reference downstream implementation.
>>>>
>>>> Set higher temperatures in the GPU trip points corresponding to
>>>> the temperatures provided by Qualcomm in the dowstream source, much
>>>> closer to the junction temperature and with a higher critical
>>>> temperature trip in the case the HLOS DCVS cannot handle the
>>>> temperature surge.
>>>
>>> Since the tsens MAX_THRESHOLD which leads to a system
>>> monitor thermal shutdown is set at 120C, I need to lower
>>> the critical and hot trip point, so please ignore this patchset.
>>
>> Should we make the "critical" trip point something like 110 or so? If
>> LMH triggers a hard shutdown at 120, the OS will not have any time to
>> take action. And 120 sounds like we're pushing it quite hard anyway.
>
>
> My plan is to harmonize and use 110 for hot and 115 for critical, and
> if available any passive cooling devices is available at 95C

sounds good!

Konrad