Re: [PATCH 0/2] platform: Enable ACPI WMI and ASUS WMI driver for ARM64

[chipang_kao]

Hi Bryan, Hans,

Thanks for the review and valuable feedback.

> On Thu, 17 Sep 2026 15:30:00 +0100, Bryan O'Donoghue <bryan.odonoghue@xxxxxxxxxx> wrote:
> Thanks, but there's already a series to this effect under discussion:
>   https://lore.kernel.org/all/20260724025959.142826-1-W_Armin@xxxxxx/
> 
> I only now notice the second patch is not covered by Armin's series.
> But this duplicates Kconfig entries.
> 
> I suppose the best approach would be to split drivers/platform/x86/Kconfig 
> into two and source the generic one here.

Thanks for pointing out Armin's series and the Kconfig duplication issue.
I had review Armin's series and make sure not to duplicate work. Regarding 
the driver location/Kconfig split, I agree that avoiding Kconfig duplication 
is necessary, and moving/refactoring the driver directly under platform/wmi 
(as Hans suggested below) seems to be the cleanest path forward.

> On Fri, 18 Sep 2026 09:38:04 +0200, Hans de Goede <hansg@xxxxxxxxxx> wrote:
> Yes ACPI / WMI is used on some ARM64 laptops, but AFAIK all of
> those are currently still booting through Devicetree since
> the current generation ARM64 laptops is missing various info Linux
> needs in the ACPI tables. So for a fully functional system 
> Devicetree is necessary.
> ...
> Questions:
> 1. Has this been tested on an actual Asus ARM64 laptop or other
>    ARM64 device? 
> 2. On which laptop/device model ?
> 3. How was this tested, how was the laptop booted; and does booting
>    it this way lead to loosing any other functionality
> 4. Were any other kernel patches used during testing?

Regarding Hans's questions:

1. Yes, this has been tested on an actual ASUS ARM64 laptop.
2. This specific model has not been publicly released yet, so I
   cannot disclose the exact model name or marketing designation
   at this stage. However, it is an upcoming ARM64 platform 
   featuring ACPI firmware.
3. The laptop was booted natively via ACPI firmware (UEFI). To
   verify the WMI functionality without modifying the base
   distribution kernel, we built and tested the modules (`wmi.ko`,
   `asus-wmi.ko`, and `asus-nb-wmi.ko`) out-of-tree and loaded
   them using `insmod`. The driver successfully bound to the WMI
   devices exposed by the ACPI tables and functioned as expected.
4. No additional kernel patches were required. Note that we have
   dropped patch 1/2 from this series because Armin's patch series
   already removes the X86 dependency from drivers/platform/wmi/Kconfig.
   Since Armin's series is currently under discussion and not yet
   merged, our v2 will focus solely on moving the ASUS drivers to
   drivers/platform/wmi/asus/ while depending on the WMI updates.

> As for patch 2/2 I think that if we want this the driver should
> simply be moved to platform/wmi .

I agree with moving the driver to drivers/platform/wmi instead of 
duplicating it under drivers/platform/arm64. I will address this in v2 
patch.

Best regards,
Chipang