Re: [PATCH V3 2/5] platform: arm64: Add driver for EC found on Qualcomm reference devices

[Konrad Dybcio]

On 3/10/26 5:58 AM, Sibi Sankar wrote:
> 
> On 3/9/2026 5:40 PM, Konrad Dybcio wrote:
>> On 3/9/26 12:55 PM, Stephan Gerhold wrote:
>>> On Mon, Mar 09, 2026 at 12:47:33PM +0100, Konrad Dybcio wrote:
>>>> On 3/9/26 11:04 AM, Sibi Sankar wrote:
>>>>> On 3/9/2026 2:33 PM, Stephan Gerhold wrote:
>>>>>> On Mon, Mar 09, 2026 at 05:06:43AM +0530, Sibi Sankar wrote:
>>>>>>> Add Embedded controller driver support for Hamoa/Purwa/Glymur qualcomm
>>>>>>> reference boards. It handles fan control, temperature sensors, access
>>>>>>> to EC state changes and supports reporting suspend entry/exit to the
>>>>>>> EC.

[...]

>>> The question is if you really want your device to hit 120°C. :-)
>> And whether the firmware running on *your* laptop actually configures
>> these limits.. I would imagine that to be the case for Windows products
>> where the TZ comes straight from qcom, but I think someone in some thread
>> mentioned LMH is not properly configured on Chrome/TFA.
>>
>> In any case, let's see if we can establish what/whether the EC does in
>> that case
> 
> 
> The "debug mode" which is used to control fan is paused during the
> following conditions:
> 
> 1) When the EC receives EC_MODERN_STANDY_CMD enter cmd
> 2) When SoC enters deep sleep fan pwm is turned off
> 3) When external processors like soccp put fan constraints on the EC
> 
> So when you do set the fan to 0 and put it in your backpack :P
> It would enter suspend, the EC would take back fan control.
> i.e. the LUT based flow that it follows by default. That would
> imply it would be the same behavior as seen on Windows?
> 
> Either way I still think we all are in agreement here. This series just
> exposes fan control knobs to the kernel and doesn't set any fan
> speed by default. The LUT based EC fan control would still be the
> default, the patches to improve this are in-flight i.e. ways to send
> avg SoC temp, update LUT and select available profiles. With all
> of these in place EC would be in good shape in the Qualcomm
> reference devices.

Right, and I suppose we're in agreement that treating the

* user sets fan to 0
* failed to suspend
* 1) didn't happen
* fan still off

chain as an oddball case is OK

Konrad