docs: Silo -level_20 freeze was machine-load flakiness (D117/D134), not a real bug

docs/HANDOFF.md

@@ -304,10 +304,12 @@ behavior exactly.
   (23), `PadLookInvertY`.
 - Verified per commit: build green, `-level_09` crash-free to frame 620 @
   91.6% (unregressed), `-level_20` crash-free.
-- **New bug logged (not fixed — out of scope):** Silo `-level_20` freezes
-  ~3 s in on the silo→Bond camera descent (no crash; capture wedges
-  ~frame 320). Reproduces on a clean pre-M-25 build. User-reported. Goes
-  with the in-level playtest / camera-transition work.
+- **Silo `-level_20` capture freeze — D117/D134 load sensitivity, NOT a
+  real bug and not an M-25 regression.** Full-length captures froze at
+  ~frame 320 (silo→Bond camera descent) twice, but only on a machine still
+  loaded from the sweep (the pre-M-25 baseline froze the same way). On an
+  idle machine the user drove Silo live past ~1800 frames, VI pacemaker
+  healthy, no stall. See `docs/M-25-QOL-REVIEW.md`.
 
 ## Next task (M-24 — Opus 5)
 

docs/LEVEL-STATUS.md

@@ -11,11 +11,13 @@ M-25 re-sweep (2026-08-30, port-QoL build `863f436b`): **20 / 21 PASS**,
 unchanged. Caverns (39) flagged CRASH `0xc000001d` once under sweep load →
 2/2 clean on retry (D117/D134 concurrent-load flakiness, not a regression).
 Cuba (54) still the only real crash (known D129 credits path, parked).
-**Silo (20) caveat:** the sweep reports PASS because it early-exits on the
-first rendered frames, but a full-length capture wedges consistently at
-~frame 320 on the silo→Bond camera descent (user-reported M-25;
-reproduces identically on a pre-M-25 baseline). Real bug, belongs with the
-camera-transition / in-level playtest work — not caught by a boot sweep.
+**Silo (20):** PASS. A full-length `GE_PCDUMP` capture froze at ~frame 320
+(silo→Bond camera descent) twice — but both those runs, *and* the pre-M-25
+baseline comparison, were on a machine still loaded from the sweep. On an
+idle machine Silo runs fine (user drove it live past ~1800 frames, VI
+pacemaker healthy, no heartbeat stall). Classic D117/D134 load-sensitive
+stall, not a deterministic bug and not an M-25 regression. If it recurs on
+an idle box, count N runs before believing it (PORT-LEARNINGS §E).
 
 Last full re-sweep: 2026-08-29 (M-22), after D134 landed (SP/DP task-done
 event no longer dropped into a full sched interruptQ — the "frame-2 hang"

docs/M-25-QOL-REVIEW.md

@@ -171,10 +171,14 @@ Cannot change the `OSContPad` contract.
 
 ## Known pre-existing bugs noted this session (not addressed — out of scope)
 
-- **Silo (`-level_20`) early freeze** — ~3 s in, when the camera pans from
-  the silo down to the player view, the sim wedges (no crash; frame capture
-  freezes ~frame 320). Reproduces on a clean pre-M-25 build. User-reported
-  M-25. Belongs with the in-level playtest / camera-transition work.
+- **Silo (`-level_20`) capture freeze — likely just D117/D134 load
+  sensitivity, NOT a real bug.** A full-length `GE_PCDUMP` capture froze at
+  ~frame 320 (silo→Bond camera descent) twice, and the pre-M-25 baseline
+  comparison froze too — but all of those ran on a machine still loaded from
+  the 21-level sweep. On an idle machine the user drove Silo live past
+  ~1800 frames with a healthy VI pacemaker and no heartbeat stall. Not an
+  M-25 regression; not currently believed deterministic. Recount N runs on
+  an idle box if it recurs.
 
 ---