## What happened

On a laptop with a single Nvidia dGPU (no hybrid iGPU) driving both the internal panel and an external monitor, the internal (`eDP-1`) display goes black and never recovers after **either**:

- the idle screensaver/lock timeout fires and is later dismissed, or
- the system does a real suspend/resume (lid close or manual suspend)

The external monitor keeps working fine in both cases, and the mouse cursor can be moved onto the dead internal panel — the compositor is alive and thinks the output is active, it just isn't putting anything on screen. Only a full reboot recovers the panel.

## System details

- Omarchy version: 4.0.1-1
- Kernel: 7.1.9-arch1-2
- GPU: NVIDIA GeForce RTX 5060 Max-Q (GB206M), single dGPU, no hybrid graphics
- Driver: nvidia-open-dkms / nvidia-utils 610.57.04-1 (GSP path)
- Internal panel: Samsung Display Corp. ATNA60HS03-0 (eDP-1)
- External: LG Ultrawide (HDMI-A-1)
- `~/.config/hypr/monitors.lua` generated by nwg-displays, two `hl.monitor()` entries, no per-monitor DPMS overrides

## Root cause (traced in source)

Both wake paths funnel through the same script: `omarchy-system-wake` -> `omarchy-brightness-display on`. Its `on` handler (`/usr/share/omarchy/bin/omarchy-brightness-display`) contains:

```bash
elif [[ $step == "on" ]]; then
  # Skip the dispatch when every active display is already lit: a redundant
  # DPMS enable right after system resume forces another modeset, which blanks
  # the panel for a beat (visible flash at the unlock screen).
  hyprctl monitors -j 2>/dev/null | jq -e '[.[] | select(.disabled == false)] | length > 0 and all(.dpmsStatus)' >/dev/null 2>&1 && exit 0
  hyprctl dispatch 'hl.dsp.dpms({ action = "enable" })' >/dev/null 2>&1
  exit 0
fi
```

This skips the real `hl.dsp.dpms({ action = "enable" })` call whenever `hyprctl monitors -j` already reports `dpmsStatus: true` on every active output — intended to avoid a double-modeset flash.

On this hardware that flag lies. I captured a live `hyprctl monitors` dump while the internal panel was physically black:

```
Monitor eDP-1 (ID 1):
	[email protected] at 3440x255
	...
	disabled: false
	dpmsStatus: 1
```

`dpmsStatus` reads `1` (on) even though nothing is being scanned out to the panel — almost certainly an Nvidia KMS quirk where the reported DPMS state doesn't reflect whether the modeset actually landed. Because `omarchy-brightness-display on` trusts this flag with no verification and no fallback, the real re-enable dispatch that would fix the panel never fires, and the compositor has no reason to try again.

This explains why it happens after *both* idle-screensaver dismissal and real suspend/resume — both call the exact same `omarchy-system-wake` -> `omarchy-brightness-display on` path (idle-triggered wake via `Service.qml`'s `cancelIdleCycle`, and post-suspend wake via the lock plugin's `runWake()` on password entry).

## Expected

The internal panel should reliably come back after screensaver dismissal and after suspend/resume, without requiring a reboot.

## Suggested fix direction

Don't trust `dpmsStatus` alone as proof the panel is actually lit. Options:
- Always issue the `hl.dsp.dpms({ action = "enable" })` call on wake (accept the occasional flash) rather than skipping based on a driver-reported flag known to desync on Nvidia.
- Or verify actual scanout state (e.g. compare a frame checksum / render timestamp) before trusting `dpmsStatus`.
- Or, as a narrower fix, exclude Nvidia-driven outputs from the "skip if already lit" optimization.

## Secondary finding (real suspend hygiene)

Separately, Omarchy's Nvidia installer (`/usr/share/omarchy/install/hardware/nvidia.sh`) sets `nvidia_drm modeset=1` but never enables `nvidia-suspend.service` / `nvidia-resume.service` / `nvidia-hibernate.service`, nor sets `NVreg_PreserveVideoMemoryAllocations=1`. These are shipped by `nvidia-utils` specifically to preserve video memory allocations across suspend and are generally required for reliable Nvidia suspend/resume on Linux. Worth enabling by default for Nvidia installs regardless of the DPMS bug above.

## Diagnostics

- `omarchy version`: 4.0.1-1
- `lspci -k` GPU line: `01:00.0 VGA compatible controller: NVIDIA Corporation GB206M [GeForce RTX 5060 Max-Q / Mobile] (rev a1)`, driver in use: `nvidia`
- Live `hyprctl monitors` capture showing `dpmsStatus: 1` on a physically black `eDP-1` available on request