Re: [PATCH v7 5/9] platform/x86: asus-wmi: Add support for multiple kbd led handlers

From: Antheas Kapenekakis

Date: Fri Oct 31 2025 - 08:14:03 EST


On Fri, 31 Oct 2025 at 09:27, Jiri Kosina <jikos@xxxxxxxxxx> wrote:
>
> On Thu, 23 Oct 2025, Antheas Kapenekakis wrote:
>
> > > 1589
> > > 1590 static void kbd_led_update_all(struct work_struct *work)
> > > 1591 {
> > > 1592 enum led_brightness value;
> > > 1593 struct asus_wmi *asus;
> > > 1594 bool registered, notify;
> > > 1595 int ret;
> > /\ value should have been an int and
> > placed here. It can take the value -1 hence the check
>
> Thanks, that needs to be fixed before the final merge.
>
> > Are there any other comments on the series?
> >
> > The only issue I am aware of is that Denis identified a bug in asusd
> > (asusctl userspace program daemon) in certain Asus G14/G16 laptops
> > that cause laptop keys to become sticky, I have had users also report
> > that bug in previous versions of the series. WIthout asusd running,
> > keyboards work fine incl. with brightness control (did not work
> > before). Given it will take two months for this to reach mainline, I
> > think it is a fair amount of time to address the bug.
>
> One thing that is not clear to me about this -- is this causing a visible
> user-space behavior regression before vs. after the patchset with asusctl?
>
> If so, I am afraid this needs to be root-caused and fixed before the set
> can be considered for inclusion.

Commit 591ba2074337 ("HID: asus: prevent binding to all HID devices on
ROG") adds HID_QUIRK_INPUT_PER_APP and the extra devices seem to
confuse asusd. Since the devices are the same as with hid-asus not
loaded, it is specific to that program.

We can delay that patch until Denis who took over maintenance of the
program can have a deeper look. I will still keep the last part of
that patch that skips the input check, because that causes errors in
devices that do not create an input device (e.g., lightbar).

Antheas

> Thanks,
>
> --
> Jiri Kosina
> SUSE Labs
>
>