Research and Issue Index

Research and issue index

This page maps DeckDoc's repository issues and upstream evidence to shipped behavior. Upstream issue reports are field evidence, not automatically confirmed root causes.

DeckDoc issue-backed modules

DeckDoc issue	Topic	Implemented	Wiki route
#1	SOF DSP panic / IPC -22	audio_sof.sh, rem_audio_sof.sh	Audio
#2	systemd core-dump analysis	coredump_analysis.sh	Crashes/report reading
#3	Wi-Fi firmware after resume	wifi_firmware.sh	Network/resume
#4	Gamescope session health	gamescope_session.sh	Display/Gamescope
#5	memory/swap/OOM	memory_swap.sh	Crashes/GPU/memory
#6	Steam client logs	steam_client_logs.sh	Evidence
#7	microSD/mmc errors	mmc_sd_card.sh	Storage/microSD
#8	suspend/resume state	acpi_pm_state.sh	Network/resume
#9	DXVK/VKD3D GPU fault differentiation	dxvk_page_fault.sh	Crashes/GPU/memory

Issues #2–#9 may still be open in GitHub even though their initial diagnostic modules shipped in PR #10. Treat them as implementation/follow-up trackers, not as evidence that the modules are absent.

The diagnostic-center follow-up closes their remaining acceptance gaps with time-bounded crash-family classification, broader wireless-driver discovery and coupled resume evidence, user-service restart counts, live-versus-cumulative memory pressure, bounded Steam helper crash rates, severity-aware mmc/ext4 signals, live fan/temperature and charge-limit context, and neutral GPU-fault attribution. Each branch has a healthy/triggered regression fixture; the issues should close when that change merges.

PR #14 added display-blackout correlation, guarded forced composition, current-versus-historical crash cleanup, session-user routing, and the original incident runbooks.

The long-off startup blackout is a community-reported research case added on 2026-07-21. It remains unverified and intentionally separate from PR #14's live-render, multi-plane LCD signature.

Full-platform follow-up issues

Issue	Scope	Current branch
#15	Model and capability manifest	Priority foundation; not implemented yet
#16	Reproducible, signed DeckDoc Rescue	Alpha collector and builder added
#17	Unified incident timeline and evidence-access ledger	Source data exists; normalization remains
#18	Continuous-probe production hardening	Opt-in bounded prototype added
#19	Privileged authorization security review	Exact-command broker prototype added
#20	Dock/USB-C/PD validation	Read-only module and fixtures added
#21	Network stages, Bluetooth/input, and update health	Wiki routes exist; module gaps remain
#22	Redacted bundle and storage-risk gate	Safety requirements documented; implementation remains
#23	DeckMD GitHub Pages symptom checker	Static checker and schema validation added

The deep research catalog covers 33 failure families, decision boundaries, detector candidates, and primary sources behind these priorities.

Upstream symptom evidence

SteamOS and Steam

• Reviewing log information documents /tmp/dumps/, steam_stdout.txt, and journal-based log review.
• SteamOS #1376 reports loss of audio after sleep.
• SteamOS #2313 reports SOF Vangogh IPC -22 and a rarer wireless failure after resume.
• SteamOS #2475 reports a fan-resume problem when sleeping while charging at a configured charge limit.
• SteamOS #2037 reports SD/ext4 corruption on a SteamOS handheld; it informs detection patterns but is not Deck-specific proof.
• SteamOS #1324, #2632, and #1015 provide comparison black-screen cases with different scopes and GPU evidence.

Gamescope and graphics

• Gamescope describes the compositor/direct-flip role in the SteamOS presentation path.
• Gamescope #1368 documents direct-scanout/ plane-transition artifacts improved by forced composition on multiple AMD devices.
• Linux AMD Display Core debugging emphasizes dmesg and pre/post display state.
• Linux AMDGPU debugfs documents low-level debug interfaces, including why DeckDoc does not blindly trigger a GPU reset.

Linux data sources

• journalctl defines boot, kernel, unit, and priority filtering.
• coredumpctl defines crash matching, metadata, storage, and access limitations.
• BTRFS device stats defines persistent read/write/flush/corruption/generation counters.

Research rules for new signatures

Before adding a signature or fix:

1. Link a primary source or attach a sanitized real incident.
2. State affected model/version and whether the issue is open, fixed, or unknown.
3. Separate reporter hypothesis from maintainer-confirmed cause.
4. Define exact time-scoped detection and likely false positives.
5. Add a fixture for healthy, triggered, unavailable, and stale-history states.
6. For remediation, document precheck, backup, action, verify, report, and rollback.
7. Never turn one anecdote into a universal Deck rule.