Aurora wip4: port Aurora20 enhancement series

.github/workflows/aurora-enhancements.yml

@@ -16,7 +16,7 @@ on:
       profile:
         description: Kernel profile (eGPU remains an unqualified experiment)
         type: choice
-        options: [candidate, egpu, unified72]
+        options: [candidate, egpu, unified72, unified72-egpu]
         default: candidate
 
 permissions:
@@ -34,7 +34,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        profile: ${{ github.event_name == 'workflow_dispatch' && fromJSON(format('["{0}"]', inputs.profile)) || fromJSON('["candidate","unified72"]') }}
+        profile: ${{ github.event_name == 'workflow_dispatch' && fromJSON(format('["{0}"]', inputs.profile)) || fromJSON('["candidate","unified72","unified72-egpu"]') }}
     steps:
       - uses: actions/checkout@v7
       - name: Install native build tools

mseries/OS9-HANDOFF.md

@@ -1,3 +1,62 @@
+## Monitor restore and wip4 deployment gate — 2026-10-26
+
+The WD22TB4 picture returned after the owner moved the cable from rear-left to front-left. The live AOC is DP-2,
+1920×1080/60, and current kernel output shows a successful modeset on the alternate DCP. This confirms the workaround,
+not same-port recovery or repeat plug/unplug reliability. The deployed display-profile watcher had two D-Bus tuple
+signature defects; both are fixed in [PR #601](https://github.com/techmore/ElementaryOS_MSeries/pull/601), the watcher
+was restarted without restarting Wingpanel/compositor, and the AOC profile now matches. Actual post-fix
+`ApplyMonitorsConfig` was then applied manually and succeeded; the AOC profile matches and the live mode remains 1920×1080/60. The automatic physical unplug/replug restore path is still unverified. Keep the working front-left connection in place.
+
+Per owner's sequencing request, WIP4/upstream review is happening before another cable cycle. The refreshed 2026-10-26 scan covers Aurora #186, #163, #9, #154, #157, #170, and issue #130. No new narrow fix is ready to add: #186 overlaps local patch 0002 in `dcp-fabric.c`; #163 is a TIPD USB-preservation fix and conflicts with #186; #9 is a broad older tunnel implementation while the pinned source already has tunnel support; #154 is for simultaneous tunnels; #157 is already patch 0001; #170 and #130 do not match this hotplug trigger. Keep the WIP4 source pin unchanged. Details: `aurora/unified72/upstream-review-2026-10-26.md`.
+
+WIP4 preparation is now concrete: the run 39944207967 package/source/provenance passed artifact/checksum review, and an
+offline trial bundle is staged privately at `/home/sdolbec/aurora-wip4-trial-20261026-v2/`. It includes candidate
+vmlinuz/initrd, the proposed GRUB entry, current `boot.bin` as byte-exact rollback, package/source inputs and provenance.
+The proposed config retains `.20` as persistent default; `grub-script-check` and the initramfs root-unlock/module audit
+passed. Receipt: `aurora/unified72/boot-preflight-wip4-2026-10-26/preparation.json`. No kernel package, bootloader,
+live GRUB config or shared boot payload was changed. Next: review rollback accessibility/fallback compatibility and,
+for a hardware trial, obtain explicit owner approval for the attended boot. Then qualify dock retry/restore with the
+watcher fix. Do not claim permanent recovery from the current picture alone.
+
+## wip4 finalized with Aurora20 ports — 2026-10-22 (Codex)
+
+Owner requested the six Aurora20 enhancements be ported into wip4 and asked for an upstream audit relative to the
+J314S/M1 Pro + WD22TB4 HDMI setup. The integration branch pins the same unmerged Aurora 7.2.11 #158 tree and #157
+display patch, retains the local DCP timeout recovery patch, and adds the six Aurora20 ports. Normal wip4 keeps
+AMDGPU/Nouveau disabled; `unified72-egpu` is a distinct optional profile/package. The M3 thermal fallback is a
+7.2.11-adapted policy helper and host control only: the M3 backend still refuses runtime startup.
+
+- Exact-source `prepare` succeeded for `unified72` and `unified72-egpu`; all selected patches applied in manifest
+  order. `candidate.py check` passed 9 PMP bootargs, 3 PMP power and 20 M3 runtime/thermal policy host controls for
+  each profile, plus the MacSMC production discovery harness (all-readable and 65 failure cases under ASan/UBSan).
+  Aurora Python admission suite passed 83/83.
+- Owner relevance: the sRGB/DCP changes target the WD22TB4 display path. udmabuf and optional SMC-property discovery
+  are generic/conditional. M3 thermal, J613/T8122 cameras and eGPU are not J314S-qualified; camera firmware/m1n1
+  prerequisites are recorded in `aurora/unified72/AURORA20-PORT-AUDIT-2026-10-22.md`.
+- **Built, not booted.** ARM64 candidate, normal wip4 and wip4-eGPU Image/modules/DTBs builds and package-boundary
+  inspections passed in [run 39854228647](https://github.com/techmore/ElementaryOS_MSeries/actions/runs/39854228647).
+  DCP KUnit and hardware checks remain open. wip3 stays installed and running, `.20` remains the persistent default,
+  and no boot payload was staged or armed.
+- Upstream #158/#157 and Aurora #135/#134/#119 remain open; #67 is an open draft. Heads and local hashes are recorded
+  in the audit/series manifests. The review branch is [PR #533](https://github.com/techmore/ElementaryOS_MSeries/pull/533).
+  Next hardware gate is a separate attended wip4 trial with matched boot-chain/rollback review.
+
+## Dock monitor dark on wip3: dcpext1 DPTX stall, wip4 draft patch — 2026-10-22 (Claude Code)
+
+Owner plugged the WD22TB4 on the running wip3 trial kernel; no picture. Dock USB/Thunderbolt fine; `28cc00000` (dcpext1) never
+answered `validate DPTX target` (six attempts), its release then timed out too, and replugging the same port failed again
+(`DP tunnel setup failed: -110`). Moving the dock to another USB-C port routed it to `289c00000` and gave 1080p/60 at once.
+Full timeline, what is and is not shown: `mseries/aurora/hardware-tests/2026-10-22-j314s-wip3-dcpext1-dptx-wedge.md`.
+
+- **Cause (partly shown).** The pinned source keeps a Thunderbolt route owned when its release is refused; a timeout is
+  treated as a refusal, so the port stays on the dead pipeline. Why `28cc00000` was picked first and why it went silent: not shown.
+- **Draft.** wip4 candidate (`unified72/series.json` revision 5, `patches/0002-...`). Build-only: patch applies on the pinned
+  files, routing logic exercised in userspace; kernel build, KUnit and every boot are open. wip3 untouched, nothing staged.
+- **Workaround on wip3.** Use a different USB-C port; reboot clears a pinned port.
+- **Side note.** Audio blips on dock plug/unplug coincide with `mseries-power` re-applying the CPU profile on each AC flip; one logged
+  ALSA underrun at the first plug. Not investigated further.
+- **Next.** CI native build of the PR, then owner decides on unarmed staging; design the same-plug re-request on the Thunderbolt side.
+
 ## Bounded Aurora CI dependency downloads — 2026-10-22
 
 The unified72 check spent 47m17s fetching 144 MB from the Azure Ubuntu
@@ -77,6 +136,20 @@ unqualified until the session package is installed and the next login.**
   card; typing shows dots; Caps Lock shows the warning; your password unlocks. If it misbehaves:
   `Environment=PANTHEON_LOCK_COMMAND=gtklock` drop-in restores the stock look.
 
+## Aurora pinned-source fetch follow-up — 2026-10-22
+
+PR #518's unified72 run [39770268111](https://github.com/techmore/ElementaryOS_MSeries/actions/runs/39770268111) and PR #528's
+first retry run [39836069576](https://github.com/techmore/ElementaryOS_MSeries/actions/runs/39836069576) both stopped at
+pinned-source preparation on repeated codeload HTTP 429 responses. The retry logic honored the bounded attempt/delay
+policy; the candidate profile passed source preparation and entered its isolated CI kernel build, while unified72 did
+not. This follow-up keeps the codeload checksum path, then falls back only for retryable HTTP errors to the explicitly
+pinned Git URL, commit and tree SHA-1 in `unified72/series.json`; it verifies both Git identities before archiving. Bad
+checksums and permanent HTTP errors still fail closed. The preparation receipt records the acquisition method and actual
+archive digest. Local tests cover retry and fallback behavior; they do not
+prove the remote Git fetch or a wip4 build. No source archive was fetched/prepared locally, and no hardware, installed
+kernel, boot payload or wip4 artifact was produced. PR #518 remains conflicted with `dev` and unreviewed. Next: let the
+updated CI validate the fallback and record candidate/unified72 build outcomes separately.
+
 ## Compositor: lock screen shows typing again; lock on idle — 2026-10-22 (Claude Code)
 
 Owner: "the screen saved to login comes up … there is no keyboard feedback", and asked for lock on idle honouring

mseries/ROADMAP.md

@@ -1,3 +1,26 @@
+## Monitor restore + WIP4 deployment gate — 2026-10-26
+
+The WD22TB4 picture returned on front-left after failing on rear-left. The profile watcher D-Bus tuple defects are fixed
+in PR #601; manual profile application succeeds, while automatic unplug/replug recovery remains unverified. Per owner sequencing, finish
+WIP4/upstream preparation before another cable test. The refreshed scan also covered #9, #154, #157, #170 and issue #130. No new narrow fix is ready to add: #186 overlaps the local DCP patch; #163 conflicts with #186 and preserves dock USB rather than recovering a tunneled display; #9 is a broad older tunnel implementation already represented in the pinned source; #154 concerns simultaneous tunnels; #157 is already included; #170 and #130 have different triggers. Keep the source pin unchanged. See [2026-10-26 upstream scan](aurora/unified72/upstream-review-2026-10-26.md). Pinned #158 merged unchanged. Exact ARM64 artifact 39944207967 is now checksum-audited and its private offline trial
+bundle is staged at `/home/sdolbec/aurora-wip4-trial-20261026-v2/`; candidate initrd/vmlinuz, proposed GRUB config,
+package/source/provenance and byte-exact shared-boot rollback are included. `.20` stays the proposed persistent default.
+No kernel package, live GRUB config or boot payload has been changed. Review rollback access/fallback compatibility,
+then get owner approval for an attended trial and repeat recovery qualification. Receipt:
+`aurora/unified72/boot-preflight-wip4-2026-10-26/preparation.json`.
+See [handoff](OS9-HANDOFF.md) and [upstream audit](aurora/unified72/AURORA20-PORT-AUDIT-2026-10-22.md).
+
+## Aurora 7.2.11 wip4 port set built — 2026-10-22
+
+Wip4 revision 6 now carries the J314S-targeted external-display/DCP series plus all six requested Aurora20
+enhancements, checked in order against the exact pinned #158 source. The eGPU modules are limited to a separately
+named opt-in profile; the M3 thermal policy port is dormant while M3 runtime startup is refused. Exact upstream states,
+firmware scope and J314S relevance are in [the port audit](aurora/unified72/AURORA20-PORT-AUDIT-2026-10-22.md).
+Both normal and eGPU source preparation and all production host controls pass; the 83 Aurora admission tests pass.
+ARM64 candidate, unified72 and unified72-eGPU kernel/module/DTB builds and package inspections passed in
+[run 39854228647](https://github.com/techmore/ElementaryOS_MSeries/actions/runs/39854228647). DCP KUnit and hardware
+acceptance remain open. Nothing was installed or staged. Review branch: [PR #533](https://github.com/techmore/ElementaryOS_MSeries/pull/533).
+
 ## Bounded Aurora CI dependency downloads — 2026-10-22
 
 The unified72 check spent 47m17s fetching 144 MB from the Azure Ubuntu
@@ -115,6 +138,17 @@ suspend and controlled power remain open. Read latest handoff before proceeding.
 
 # Roadmap, features and follow-ups (MacBook Pro 14" M1 Pro, elementary OS 9 on Asahi)
 
+
+## Pinned Aurora source fetch follow-up — 2026-10-22
+
+Both #518 run 39770268111 and #528 run 39836069576 received repeated HTTP 429 responses from the pinned codeload URL;
+the retry policy exhausted before unified72 source preparation. The candidate profile passed prep and entered its isolated
+CI build. The current follow-up retains archive SHA-256 verification, retries transient responses, then uses the exact
+Git URL, commit and tree SHA-1 pinned in `aurora/unified72/series.json` only after a retryable HTTP error. Mocked checks
+cover both paths, and the preparation receipt records the acquisition method and archive digest; the remote fallback and
+wip4 build still need CI evidence. No local source fetch or hardware/kernel
+installation occurred. #518 remains conflicted with `dev` and unreviewed; verify updated CI before considering its refresh.
+
 ## Factory speaker protection wired into ARM64 images — 2026-10-21
 
 Ubuntu Resolute `speakersafetyd` 1.1.2-4 now ships in the Apple Silicon overlay,

mseries/aurora/README.md

@@ -24,12 +24,14 @@ replace the historical "not booted" descriptions below for current status.
 The first freeze is **experimental**, with boot/PMP/idle-accounting evidence;
 full hardware and energy acceptance remains incomplete.
 The [energy port plan](ENERGY-PORT-PLAN-2026-10-18.md) audits 52 branches and
-12 open PR heads and introduces a separate [7.2.10 WIP build](unified72/README.md).
+12 open PR heads and introduces the separate [7.2.11 wip4 build](unified72/README.md).
 
 For the newer source-build enhancement candidate (udmabuf, scoped M3 thermal/
 camera fixes and SMC battery-reporting hardening), see
 [enhancements](enhancements/README.md). It has a separate kernel/package identity;
 the fixed `.20` installer and staging instructions below do not deploy it.
+The Aurora20 ports are also audited against the separate [7.2.11 wip4 series](unified72/README.md), with owner-device
+relevance and firmware/eGPU gates in its [port audit](unified72/AURORA20-PORT-AUDIT-2026-10-22.md).
 
 This opt-in lane follows **omacom/linux-aurora release
 `aurora-2026.10.17.20`**, kernel source