[CoreyH]

Docs only: a read of the related projects on GitHub (PR descriptions, DeepIdle's docs, commit headlines; no m1n1 or Aura code, no driver source; nothing posted).

• RESEARCH.md, "What the other projects have since 2026-11-10": the reply on t600x: PMU RTC wake alarm (SPMI notify at 192, Maverick event clear, DT) omacom/linux-aurora#351 (the alarm works on a third PMU, Stowe on a J413), M4 Pro and Max suspend to RAM under aura (J616s (M4 Pro) at EL2: hibernation and suspend-to-RAM aurora-silicon/linux#1477, #1664), DeepIdle's runs R25 to R38B on T6020 (the same reset after the commit; five always-on words that step at the commit; pages that hang a reader), nothing new public from Naeem or maralcbr.
• PLAN.md: two notes for the next build (pages never to read; four words worth reading at rest), and that reading the M4 drivers is Corey's call. The next step is unchanged.

🤖 Generated with Claude Code

[pullfrog]

ℹ️ One minor factual correction inline; no blocking issues found.

Reviewed changes — Reviewed PLAN.md and RESEARCH.md against the cited public reports; no executable code changed.

• Research sweep: Records Stowe alarm validation, M4 Pro/Max sleep results, DeepIdle's latest experiments, and related upstream activity.
• Experiment planning: Adds an unreadable-page warning and four candidate register reads while leaving the next experiment unchanged.

  ^{｜ Fix all ➔ ｜ Fix 👍s ➔ ｜ View workflow run ｜ Using gpt-6.1-sol ｜ 𝕏}

[pullfrog]

✅ No new issues found.

Reviewed changes — Reviewed the correction since fce1130 against DeepIdle's public worklog; no executable changes.

• Corrected register-block attribution: Separated PMGR state counters at 0x29e2c8000 from MPMGR_AON and its unreadable pages.

  ^{｜ View workflow run ｜ Using gpt-6.1-sol ｜ 𝕏}