Re: [PATCH v3 7/7] HID: asus: Re-send the Zenbook Duo keyboard handshake after probe

[Denis Benato]

19 ott 2026 08:12:47 Rebecca Mara Müller <rebecca.mara@xxxxxxxxx>:

> On Sun, Oct 18, 2026 at 09:48:21PM +0200, Denis Benato wrote:
>> Well you know how this things go... Only who does nothing can claim
>> to have made no mistakes.
>
> Thank you, Denis, that was kind.
>
>> So I am concerned that hardware isn't entering s2idle properly...
>> Are you sure it is?
>
> I checked it this morning on the UX8407AA (Panther Lake), keyboard
> docked, with 30 s of s2idle via rtcwake -m freeze -s 30, reading
> /sys/power/suspend_stats/last_hw_sleep and
> /sys/devices/system/cpu/cpuidle/low_power_idle_system_residency_us:
>
>   without hid-asus (hid-generic):  27.6 s hardware sleep, S0ix +27.6 s
>   with v3 loaded:                  28.6 s hardware sleep, S0ix +28.6 s
>
> So it does reach S0ix, with and without the series. With v3 loaded and
> the backlight set to level 3 before suspending, the keyboard backlight
> stayed on unchanged for the whole time, no change in brightness or
> pattern. So I guess it is the second case you mention: nothing tells
> the keyboard to switch the backlight off on suspend.
>

Very good to see it sleeps just fine. So yeah a command is missing... On top of my mind there are a few possibilities:

- is has a tuf leds power devid in asus-wmi (please check in /sys/class/platform/asus* what's available
- as the previous one, but firmware is bugged and the query of the capability returns not supported while the write goes through (there is at least another model doing that)
- it supports ROG-style power controls of zones and asusctl is lacking support
- it has some other devid for that

First thing I would do is to follow instructions in gitlab.com/asus-linux/reverse-engineering to grab the acpi and search for devid that does things (I suggest an llm to quickly find unmapped/missing IDS) and if that fails I would head to windows to see what the asus software sends over USB... On what I deal with (ROGs mainly) the application is armoury crate, but perhaps on zenbooks and other line-ups it's not available and you only have myasus.

If that fails, and windows gets leds off, with "fails" meaning there is no command to tell the firmware what to do on s2idle.... Fire up a windows VM, pass the hid device to it, setup the vm for s2idle and run a capture with wireshark on linux.... That should tell you if an explicit command is needed.

If that's the case setup userspace to turn leds off, don't do it in kernel or TUFs will break lighting while in s2idle.

> And if you ever need something tested, on this machine or anything
> else, feel free to ping me. I enjoyed this and learned a lot, and I
> would like to keep learning.
>
> Thanks,
> Rebecca