On Mon, 29 Jun 2026 at 23:11, Denis Benato [off-list ref] wrote:

On 6/29/26 22:21, Antheas Kapenekakis wrote:

quoted

On Mon, 29 Jun 2026 at 22:12, Denis Benato [off-list ref] wrote:

quoted

On 6/29/26 11:21, Jiri Kosina wrote:

quoted

On Fri, 19 Jun 2026, Denis Benato wrote:

quoted

Hi all,

I have added support for controlling the (way too bright) XG mobile
LEDs in hid-asus and added the i2c version of already supported
hardware that was probed only when it's a USB: these are two separate
features changes and are the only two that are not fixes for
pre-existing issue (see below).

Auto-review bot has spotted a bunch of pre-existing problems alongside
problems in my own code therefore at this point I am going to fix the
more problems I can and including those fixes and improvements in this
patchset.

For v4 I decided to follow Antheas' suggestion of reusing the existing
workqueue and by making it more generic I solved a good bunch of issues.

The v5 iteration is simply me fixing a bunch of bugs in new code spotted
by the bot. Thanks for providing such a useful tool!

On a side node this patchset has a few more warnings: specifically
"WARNING: Prefer kzalloc_obj over kzalloc with sizeof" but it's a false
positive as that would introduce sleeping calls in atomic contexts.

Denis,

thanks. Could you please flag which patches you'd prefer to go in still
for 7.1 and which ones are not critical and could wait for 7.2? The whole
lot is quite big.

Hi Jiri!

Patch 1 contains multiple fixes, mainly use of sleeping calls in atomic contexts:
I'd say slap it in 7.1 fixes.

Patch 2 is a minor improvement removing an OOM message to align the driver
to newer code where that would have been a flagged warning: probably a fix for
7.1 but there's no rush for this one.

Patch 3 is a logic fix, condition is very rare (not sure if even possible) but belongs
to 7.1 almost certainly.

I would like to take my time on reviewing patch 4 in the next cycle.
The patch 5 does the backlight for the station? That looked _fine_ to
me but a second pair of eyes would be good because it is big.

Mmhh I think you may have a mistake here? Patch 4 has your reviewed-by
and it's for the xgm led.

Patch 5 adds backlight for TUFs 2025 and 2026 models: asus is using
the (almost) same firmware as recent ROGs just i2c instead of usb,
so patch 5 add i2c to the list.

I want to review for the next cycle:
HID: asus: refactor the two workqueues and init sequence

Antheas

quoted

Antheas

quoted

Patch 4 new feature support: goes to 7.2.

Patch 5 unsure. It adds support to the backlight, but to properly do it I'll need
to wait for lamparray to be merged since autonomous mode must be disabled
to take effect. I'd say 7.2 just because of the missing lamparray. Merging in 7.1
has no downsides, yet IDK if putting in 7.1 violates some rule (probably yes?).

Please give a (very) quick read to the following link:
https://github.com/OpenGamingCollective/asusctl/issues/119#issuecomment-4723566635

This is something I am full of tickets for, probably we have reached the thousand
between discord, github and gitlab: very prominent issue and I'm eagerly waiting
for lamparray support to appear upstream.

Thanks,
Denis

quoted

Thanks,