## What happened

On a Lenovo Legion 5 15IAX11, neither the display-brightness keys/OSD nor `omarchy brightness display` change the actual screen brightness. Keyboard backlight is unaffected and works fine (separate hardware).

## What was expected

Pressing the brightness keys (or running `omarchy brightness display +5%`) should visibly dim/brighten the panel.

## Root cause (traced end-to-end, not just a guess)

This is **not** a keybinding or script bug — every layer Omarchy controls works correctly:

1. The physical key correctly emits `KEY_BRIGHTNESSUP`/`KEY_BRIGHTNESSDOWN` on the ACPI "Video Bus" evdev device (confirmed with `evtest`).
2. Hyprland correctly receives and dispatches the `XF86MonBrightnessUp/Down` binding to `omarchy-brightness-display` (confirmed by temporarily logging every dispatch while pressing the keys — it fired every time).
3. `omarchy-brightness-display` correctly calls `brightnessctl`, which successfully writes new values and reads them back correctly, for **both** backlight devices present on this machine:
   - `intel_backlight` (the one `omarchy-hw-display` picks — it has no case for NVIDIA WMI EC backlights, so it falls through to this)
   - `nvidia_0` (registered by `nvidia_wmi_ec_backlight`)
4. The screen does not respond to writes to **either** device, even at 10%.

This matches a known, currently-unresolved upstream kernel bug reported against `nvidia_wmi_ec_backlight`: [bugzilla.kernel.org #221430](https://bugzilla.kernel.org/show_bug.cgi?id=221430) (different laptop model, identical symptom). Per that report and the linked linux-acpi mailing list discussion, affected laptops' firmware advertises `nvidia_wmi_ec` as the authoritative backlight, but the WMI method that's supposed to set brightness reads from an undocumented ACPI variable that only Windows populates before calling it — so the EC silently ignores whatever value Linux (or anything running on Linux) writes, on either backlight interface. A "backlight-aggregate" driver to work around this was proposed but is not implemented in mainline as of this writing.

**This means no userspace/compositor-level fix is possible** until the kernel-side aggregate driver lands, or a per-model DMI quirk is added (none exists yet for 15IAX11).

## Ask for Omarchy

Two independent, low-risk things that would help affected users even though they can't fix the underlying EC behavior:

1. `omarchy-hw-display`'s heuristic (`/usr/share/omarchy/bin/omarchy-hw-display`) has no case for `nvidia_wmi_ec`-style backlights (only `gmux_backlight`, `amdgpu_bl*`, `intel_backlight`, `acpi_video*`). Worth adding `nvidia_0`/`nvidia_wmi_ec_backlight*` to the candidate list for correctness on hybrid-graphics laptops, even on models where it happens not to be wired to real hardware.
2. Consider surfacing a hint (in the OSD or `omarchy brightness display` output) when brightness writes appear to have no effect on a hybrid-graphics system, pointing at `acpi_backlight=native` as a known (not guaranteed) workaround, so users aren't stuck thinking it's an Omarchy bug.

## Steps to reproduce

1. On a Lenovo Legion 5 15IAX11 (or similar hybrid Intel+NVIDIA laptop affected by kernel bug #221430), press the brightness keys or run `omarchy brightness display +5%`.
2. Observe the OSD/CLI report a successful change, but the panel brightness does not visibly change.

## System details

- Omarchy version: 4.0.3-1
- Model: Lenovo Legion 5 15IAX11 (board LNVNB161216), BIOS UFCN27WW (2026-06-22)
- Kernel: 7.2.3-arch1-3
- GPU: Intel (xe driver) + NVIDIA (nvidia-open-dkms 610.57.04-1, hybrid/Optimus)
- Backlight devices present: `intel_backlight` (max 192000), `nvidia_0` (max 100, via `nvidia_wmi_ec_backlight`)
- Keyboard backlight: unaffected, works normally

## Diagnostics

```
$ evtest /dev/input/event6   # "Video Bus"
Event: type 1 (EV_KEY), code 224 (KEY_BRIGHTNESSDOWN), value 1   # fires correctly on keypress

$ hyprctl devices | grep -A1 video-bus
    video-bus            # correctly recognized as a keyboard by Hyprland

$ brightnessctl -d intel_backlight set 60%   # succeeds, read-back correct, screen unaffected
$ brightnessctl -d nvidia_0 set 10%          # succeeds, read-back correct, screen unaffected
```

Happy to attach the full `evtest`/`hyprctl` logs if useful.