[jlucaso1]

Cold registry fallback initializes async-lock's event and leaves 72 extra bytes per client in the connected-idle retention check. This follows #1883, which was merged while the retention correction was being validated.

Registry operations now share an Arc-backed async-lock mutex through the complete holder/waiter lifetime. The final lease releases its strong reference and clears the weak slot under the admission lock, allowing both the mutex event and its allocation header to retire when idle. The actual guard drops before its lease. This adds temporary allocations at the start of an isolated operation; concurrent operations share them.

Four lifecycle regressions cover queued ordering, cancellation after notification, cancellation with another waiter, and simultaneous final lease retirement.

Validation:

• All 100 ADV/recovery and mutex lifecycle cases passed on the pre-rebase tree; this file is byte-for-byte identical after applying the patch to main fa8c640c.
• Current-main all-feature library Clippy passes with warnings denied; the no-default-feature wasm32 library check passes. Formatting and diff checks pass.
• Exact-head connected retention, CPU/allocation comparisons and SDK CI are queued. Full SDK tests on the updated base have not completed.

The existing Cargo Deny failure for RUSTSEC-2026-0459 is inherited from main's bundled SQLite dependency. Published Diesel 2.3.14 excludes the fixed libsqlite3-sys 0.39 line; the audit remains enabled.

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

• X
• Mastodon
• Reddit
• LinkedIn

[greptile-apps]

Confidence Score: 5/5

[Medium impact] The PR appears safe to merge.

Summary

Releases the registry mutex allocation when the last holder or waiting operation finishes.

• Registry operations share a mutex that disappears when the client is idle.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  A[Registry operation starts] --> B[Lock weak slot]
  B --> C[Reuse or create shared mutex]
  C --> D[Keep lease and release slot lock]
  D --> E[Wait for registry mutex]
  E --> F[Hold registry guard]
  F --> G[Release guard before lease]
  E --> H[Cancellation drops acquisition before lease]
  G --> I[Drop lease under slot lock]
  H --> I
  I --> J{Strong references remain?}
  J -->|Yes| K[Keep weak slot]
  J -->|No| L[Clear weak slot]

Loading

_{Reviews (1) · Last reviewed commit: "fix(client): retire idle registry mutex ..." · Reviewed by Greptile}

[jlucaso1]

@codex review

[chatgpt-codex-connector]

Codex Review Summary: ✅ Completed

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-10T22:50:47.066178Z	e3a67b9	Draft marked ready
🔒 Security Review	✅ Completed 2026-11-10T22:54:29.613660Z	e3a67b9	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. What shall we delve into next?

Reviewed commit: e3a67b9c37

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[codspeed]

Merging this PR will not alter performance

1 warning, 1 note

⚠️ 141 benchmarks are dominated by system calls

System calls cannot be simulated consistently and are left out of the measure. The reported value understates the real cost. Switch to the Walltime instrument to measure them.

Learn more about system calls

See affected benchmarks

ℹ️ 24 benchmarks have no flamegraph

The flamegraph wasn't generated: most of the time was spent in runtime internals, which would leave nothing useful to analyze.

See affected benchmarks

✅ 852 untouched benchmarks
🆕 472 new benchmarks
⏩ 372 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
🆕	Memory	protocol_address_per_device[128_round0]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[128_round1]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[128_round2]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[128_round3]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[32_round0]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[32_round1]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[32_round2]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[32_round3]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[512_round0]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[512_round1]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[512_round2]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[512_round3]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[128_round0]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[128_round1]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[128_round2]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[128_round3]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[32_round0]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[32_round1]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[32_round2]	N/A	0 B	N/A
🆕	Memory	protocol_address_per_device[32_round3]	N/A	0 B	N/A
...	...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

---

_{Comparing fix/registry-mutex-retention-1883 (e3a67b9) with main (92e64c4)²}

Footnotes

1. 372 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

2. No successful run was found on main (fa8c640) during the generation of this report, so 92e64c4 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩