8e709b0a24
Snapshot of every non-log artifact under audit-runs/ from audits 003 through 058: findings.md per audit, comparison CSVs, probe diffs, schema docs, register-dump txts, lr-trace JSONL streams, the saved canary patch diffs, etc. ~284 files / ~52 MB total. Excluded (per .gitignore): probe stdout/stderr/log streams (the raw firehose), guest-memory dumps under audit-026/027/029 (4.5 GB of .bin files; *.bin pattern added to .gitignore this commit). Also adds the orphan audit-058-sub825070F0-activation directory that a subagent accidentally created at project-root instead of under xenia-rs/audit-runs/; relocated to its proper home. Purpose: cross-machine continuity. With these summaries committed, a fresh clone gives the next session the full per-audit context (findings + tables + cascade predictions) without dependence on local-only working tree. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2.5 KiB
2.5 KiB
| 1 | angle | name | loc_budget | cascade_d_prediction | risk | prerequisites | expected_evidence_gained |
|---|---|---|---|---|---|---|---|
| 2 | A | XAudioSubmitRenderDriverFrame | ~30 | NO (host-side audio output only) | very low (host-only) | none | minimal — no guest callback chain to bootstrap cluster |
| 3 | B | Long-horizon -n 5B | 0 | NO (counters identical to 500M) | none (read-only run) | none | DONE this audit — bit-identical to 500M; rules out compute-time plateau |
| 4 | C | STFS / XamContent enum impl | ~80-120 | UNLIKELY (game gets coherent NO-MORE-FILES path) | low | canary content_manager.cc reading | medium — fixes a real correctness gap (XamContentCreateEnumerator writes nothing to out-pointers); may unblock content-discovery wait if there is one |
| 5 | D | Plug missing canary-only exports | ~50-100 | UNLIKELY (none are renderer/UI critical) | low | canary impl reading | low — XamUserGetXUID + XamUserReadProfileSettings could be gating user-profile reads but XamUserGetSigninState already returns signed-in |
| 6 | E | M5.5 caller chain for sub_82452DC0 | 0 (read-only DB) | N/A — diagnostic only | none | none | DONE this audit — work-submitter sub_82452DC0 FIRES on tid=13 cycle 8127 from sub_821CB030+0x19C; γ-class wedge is real but not in caller-reachability |
| 7 | F | Relax M11 static-init detector | ~150-300 (analyzer) | N/A — diagnostic only | none | re-run analyzer pipeline | HUGE — finds CRT driver sub_824ACB38 that already runs and invokes the 24+ GamePart factory regs; reframes audit-049 narrative — they DO bootstrap. New runtime probing reveals the actual stall. |
| 8 | G | Long-horizon canary trace via patch | ~50-100 canary | UNLIKELY in Linux Debug (host-presenter block at frame 42) | medium (canary patches already proven safe) | canary build | medium — bypassing intro-video host-presenter block in canary may extend reach without Wine |
| 9 | H (NEW) | Probe and signal handle 0x1288 manually | ~30-60 | YES POSSIBLE (unblocks tid=13 join) | medium (synthetic signal could cause bad state) | audit-049 wait analysis | HIGH — if synthetic NtSetEvent(0x1288) unparks tid=13 and boot advances we learn (a) what tid=13 does next (b) whether the front-end UI flow runs |
| 10 | I (NEW) | Trace what work-submitter callback does and where it diverges | ~0 (probing) | N/A — diagnostic | none | probe sub_82452DC0 + descendants | HIGH — work-submitter fires but never reaches signal-back path; identify the LR within sub_82452DC0 chain that branches to NtSetEvent only in canary; 1-PC fix candidate |
| 11 | J (NEW) | XamContentCreateEnumerator writes proper handle/buffer_size | ~50 | NO (still NO-MORE-FILES from XamEnumerate) | low | xam_content.cc:129 reading | low-medium — closes a stub-correctness gap; may be prerequisite for future content-discovery flow |