Re: [RFC PATCH v2 0/2] platform/x86: huawei-wmi: native Turbo profiles with capability detection

From: Eren Özen

Date: Wed Nov 18 2026 - 17:19:47 EST


Hi Fuwei,

I retested v2 on the same HUAWEI MateBook 16 (AMD Ryzen 7 5800H, DMI
product name "CREM-WXX9", BIOS 2.32), this time without any DMI entry.
I applied both patches unmodified on top of c339f4010dea and built the
driver as a module for a 7.2-based kernel.

- The capability check finds \GTUB and \STUB, and the profile device is
registered. Loading the module did not change the firmware mode
(checked by reading \GTUB before and after).
- platform_profile_choices is "balanced performance", and writing either
value switches the firmware mode (read back through \GTUB).
- GNOME's Power Mode menu (tuned-ppd) switches between the two.
- Fn+P updates platform_profile in both directions, and 0x2a0/0x2a1 are
not logged as "Unknown key pressed".
- When the charger is unplugged in performance mode, platform_profile
follows the firmware back to balanced.
- On battery, writing "performance" fails and the mode stays balanced.
- After plugging the charger back in, platform_profile still matches the
firmware.
- The driver logged no warnings during the test.

One small thing: after Fn+P, platform_profile changes to performance,
but GNOME's menu (Fedora's tuned-ppd) keeps showing balanced. It was the
same with v1. I haven't checked whether tuned-ppd listens for this
change.

About your first question, on this laptop the capability check is what
made the difference. v1 needed an extra DMI entry here, and v2 works
without any change.

Tested-by: Eren Özen <erenozen@xxxxxxxxxxxxxx> # MateBook 16 CREM-WXX9, BIOS 2.32

Thanks,
Eren