Re: acer-wmi: Nitro button doesn't produce a WMI event

[Hans de Goede <hdegoede@redhat.com>]

Hi Armin,

On 6-Jan-25 11:29 PM, Armin Wolf wrote:
> Am 06.01.25 um 19:33 schrieb Hans de Goede:
> 
>> Hi all,
>>
>> On 6-Jan-25 5:59 PM, Armin Wolf wrote:
>>> Am 02.01.25 um 07:18 schrieb Hridesh MG:
>>>
>>>> On Thu, Jan 2, 2025 at 1:51 AM Armin Wolf <W_Armin@gmx.de> wrote:
>>>>> Am 01.01.25 um 20:53 schrieb Hridesh MG:
>>>>>
>>>>>>> This ACPI method should trigger the turbo mode button (found inside the DSDT table):
>>>>>>>
>>>>>>>            Method (_Q58, 0, NotSerialized)  // _Qxx: EC Query, xx=0x00-0xFF
>>>>>>>             {
>>>>>>>                    Debug = "=====PROJECT_QUERY_58====="
>>>>>>>                     ^^^WMID.FEBC [Zero] = 0x07
>>>>>>>                     ^^^WMID.FEBC [One] = 0x04
>>>>>>>                     ^^^WMID.FEBC [0x02] = One
>>>>>>>                     Notify (WMID, 0xBC) // Device-Specific
>>>>>>>            }
>>>>>>>
>>>>>> I feel like an idiot right now but I just realized something: So far
>>>>>> I've been assuming that the button I've been calling the "Turbo
>>>>>> Button" worked the same way on both Nitro and Predator laptops, but
>>>>>> that's not the case.
>>>>>>
>>>>>> On Predator laptops, the button directly enables Turbo mode but on the
>>>>>> Nitro it only opens the Nitro Sense app. I had assumed that both
>>>>>> buttons simply opened the app and that directly enabling Turbo mode
>>>>>> was a feature provided by the Linux driver.
>>>>>>
>>>>>> Given this, the ACPI code that you linked earlier is probably for the
>>>>>> Predator's "Turbo Button". Could it be that the button on my laptop
>>>>>> doesn’t use that ACPI code at all, considering its functionality is
>>>>>> completely different? (Though i doubt Acer would leave it in if it
>>>>>> wasn't being used)
>>>>> It is quite common for manufactures to just copy and paste ACPI code snippets, so it is not
>>>>> unusual to have some unused code inside the ACPI tables.
>>>>>
>>>>> Did you receive any input events or dmesg messages when pressing that button?
>>>>>
>>>>> If no then it could also be that this button depends on the Intel THC touch controller to work,
>>>>> take a look at https://lore.kernel.org/all/20241216014127.3722172-1-even.xu@intel.com for details
>>>>> about the Intel THC controller.
>>>>>
>>>>> Thanks,
>>>>> Armin Wolf
>>>>>
>>>> I do receive input events, it shows up with the keycode 433 and symbol
>>>> XF86Presentation. I think you mentioned something regarding an i8042
>>>> filter, can we use that to call the function to cycle between the
>>>> profiles?
>>> Since this is a valid keycode i advise against filtering it out. Instead it would make more
>>> sense to allow userspace to control the turbo state for example trough a sysfs attribute.
>>>
>>> Then you can use a userspace program to react to this key press.
>> Since some laptops handle this completely in the EC, so far drivers
>> which get an event for this have been reacting to this event
>> themselves and calling the platform_profile_cycle() helper to cycle
>> through the various performance profiles.
>>
>> I think it would be best to do this here too.
>>
>> Regards,
>>
>> Hans
> 
> This keycode event is different from the WMI notification used to toggle the turbo status.
> 
> As mentioned above, the turbo button on the Predator series is meant to toggle the turbo mode,
> while the special button found on the Nitro series is meant to only open the OEM application.
> 
> Toggling the turbo mode in response to this special button will confuse users should Acer
> mark the button with a small application icon.
> 
> I rather let applications change the turbo mode from userspace. This would also benifit machines
> which do not have any special buttons for this.

Ah I see.

I agree if the button does not cycle performance-profiles under Windows then it
should just send some KEY_FOO evdev event. 

Regards,

Hans