Skip to content

The M4 Pro's and M4 Max's sleep drivers, read - #91

Merged
CoreyH merged 2 commits into
mainfrom
docs/m4-drivers
Nov 12, 2026
Merged

CoreyH merged 2 commits into
mainfrom
docs/m4-drivers

Conversation

@CoreyH

@CoreyH CoreyH commented Nov 12, 2026

Copy link
Copy Markdown
Owner

Docs only. Corey OK'd reading the two M4 sleep drivers (aurora-silicon/linux#1477 t6040-sleep.c by fbalazsy, #1664 t6041-sleep.c by AxxzyWasTaken; GPL-2.0-only OR MIT) on 2026-11-12. aura wasn't read; nothing was posted there.

  • RESEARCH.md, "The M4 Pro's and M4 Max's sleep drivers, read": their engine has B2's opening op for op; what each of their stalls was (bit 11 as a latch on a power word, a domain raised under a gated parent, the first gate held by a core or a wake source, the DCS's power-gate mode), and what on T6000 looks like the same thing, marked as judgment. One check made today: none of the look's 51 power words has bit 11 set after noirq.
  • PLAN.md: the next step is now a desk step with no run (B2's ops up to 224 read against those four classes in the recipe's state); pre=1 follows unless the desk step outranks it.
  • CLAUDE.md: the OK recorded beside the one for Naeem's driver.

Also on 2026-11-12, with Corey's OK: a reply posted on omacom/linux-aurora#351 (thanks for the Stowe test; #454 stays its author's PR).

🤖 Generated with Claude Code

…our classes of stall to read B2 against

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>

@pullfrog pullfrog Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ℹ️ Minor research-evidence qualification inline.

Reviewed changes Reviewed the M4 driver research and its implications for the next T6000 investigation.

  • Research findings. Adds revision-pinned driver observations and T6000 comparisons; the cited 51-word bit-11 check matches the committed snapshot.
  • Experiment sequencing. Puts a no-build, no-run desk analysis ahead of the next experiment.
  • Source boundaries. Records permission to read the Linux drivers while keeping aura off limits.

Pullfrog  | Fix all ➔ | Fix 👍s ➔ | View workflow run | Using gpt-6.1-sol | 𝕏

Comment thread RESEARCH.md Resolved
…anism, not a stall it reports (Pullfrog on #91)

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>

@pullfrog pullfrog Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

✅ No new issues found.

Reviewed changes Reviewed the evidence qualification since the previous review, checking the wording against the pinned M4 Pro driver.

  • Qualified the research. Distinguished the proposed wake-source mechanism from observed M4 stalls in RESEARCH.md.
  • Aligned the plan. Preserved pending wake sources as a diagnostic question in both PLAN.md summaries.

Pullfrog  | View workflow run | Using gpt-6.1-sol | 𝕏

@CoreyH
CoreyH merged commit 787ed9f into main Nov 12, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant