## Summary

On a ThinkPad P1 Gen 8 docked to two Apple Studio Displays over Thunderbolt, suspend/resume with the lid closed produced: (a) an unresponsive wake lasting ~1 minute (no reaction to keyboard until the lid was opened), during which the Hyprland session restarted, and (b) the second Studio Display wedged in a state where Hyprland lists it as connected and enabled but with mode `0x0` — black screen — from which no runtime recovery succeeds.

## Expected

Resume from suspend restores the session promptly and both external displays return to their configured modes.

## Actual

1. Wake was unresponsive with the lid closed; after opening the lid and waiting about a minute, the lock screen's password UI appeared. The journal shows the GPU re-initialising (`i915 GuC firmware reload`) and a fresh Hyprland start (new instance's xkbcomp warnings) at resume time, i.e. the compositor/session restarted around resume.
2. After unlocking, the second Studio Display stayed black. `hyprctl monitors` shows it enabled but with no mode:

```
Monitor DP-4 (ID 2):
	[email protected] at 2560x0
	description: Apple Computer Inc Studio XDR 0x9ED85B95
```

The first display (DP-2) came up normally at 5120x2880@60.

## Timeline (journal, this boot)

- 19:10:55 — brief suspend entry/exit (`PM: suspend entry (s2idle)` … `PM: suspend exit`, 5 s wall)
- 19:11:21 — suspend again (s2idle) — the long sleep
- 19:14:16 — resume: `i915 ... GuC firmware i915/mtl_guc_70.bin` reload, GT0/GT1 re-init
- 19:14:30 — new Hyprland instance logs xkbcomp warnings (session restart)
- 19:17:59 — replugging the Thunderbolt cable: pciehp re-enumeration plus:

```
i915 0000:00:02.0: [drm] drm_WARN_ON(!(intel_dp->psr.sink_support && intel_dp->psr.source_support) && ...)
WARNING: drivers/gpu/drm/i915/display/intel_psr.c:2465 at intel_psr_disable [i915]
Workqueue: display_unordered intel_encoder_link_check_work_fn [i915]
Call Trace: intel_ddi_disable+0x146/0x280 [i915] ...
```

## Recovery attempts (all failed while the session was live)

- `hyprctl eval 'hl.monitor({ output = "DP-4", mode = "preferred", position = "auto", scale = 2 })'` → `ok`, no change
- Same with explicit `mode = "5120x2880@60"` → `ok`, no change
- Disable/re-enable cycle via `hl.monitor({ output = "DP-4", disabled = true/false })` → accepted, state never changes
- `hyprctl reload` → `ok`, no config errors, no change
- Physically replugging the Thunderbolt cable several times → Thunderbolt re-enumerates (pciehp slots), the PSR WARN above fires once, display never recovers

Throughout all of this the connector's EDID remains valid (`/sys/class/drm/card0-DP-4/edid` decodes to serial `0x9ED85B95`), so the panel and cable are healthy; the output is wedged below or inside the compositor's bookkeeping.

## System details

- Omarchy 4.0.4-1
- Kernel 7.2.3-arch1-3 (x86_64)
- Hyprland 0.56.2, SDDM, UWSM
- ThinkPad P1 Gen 8 (21Q8A007CD), Intel Core Ultra 9 285H (Arrow Lake), i915 (`mtl_guc_70` firmware); NVIDIA dGPU present (nvidia_drm loaded)
- 2× Apple Studio Display (5120×2880) via Thunderbolt, lid closed (eDP-1 disabled)
- Monitor config: generic rule — preferred mode, auto position, scale 2 (`~/.config/hypr/monitors.lua`)

Full `omarchy-debug --no-sudo --print` output available on request (2086 lines).