Repository navigation
docs(quic): the BBR3 ledger names its reference commits and every deliberate divergence - #8899
Conversation
Quest-Claim: fce51c9188330d91
…iberate divergence Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
|
Warning Review limit reachedYou've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Next included review available in 13 minutes. View limit details
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. |
There was a problem hiding this comment.
✅ No new issues found.
Reviewed changes — docs-only PR plus one code comment, read in full.
- BBR3 references ledger:
rs/moq-quic/README.mdadds aBBR3 referencesentry naming draft-07, Linux90210de4and QUICHE51a950db, and lists the deliberate divergences. The loss-samples entry now cites draft-07. - Vendor-sync pointer:
quest/m3/vendor-sync/quinn.mdpoints at the ledger instead of the finished epic. - Epic removal:
quest/m1/bbr3-references/README.mdis deleted. Its stated goal and plan items are covered by the new ledger entry, and no references to it remain in the checkout. - Comment:
maybe_go_downinrs/moq-congestion/src/bbr3/mod.rsnotes that the ACK's bytes still count as in flight, as Linux'sprior_in_flightdoes. The code is unchanged.
claude-sonnet-5-5 | 𝕏
kixelated
left a comment
There was a problem hiding this comment.
Automated review by review (OpenAI)
Reviewed commit: 43171bc0caeec8f969ed813036404e11953ac84b.
No actionable defect found in this documentation-focused static review.
Read all four changed files and spot-checked the ledger against the controller implementation: application-limited bandwidth aging and plateau handling, ProbeRTT skip/lowest-sample adoption and interrupted-probe aging, recovery snapshots, Drain reset, Reno round bounds/floors, and the pre-ACK precautionary-stop calculation. Located corresponding regression tests, including idle gaps, per-packet loss counts, ProbeRTT skip/undo, Drain signals and small-BDP Reno timing. The source change is only a comment; no behavior changes.
Also read the pinned Linux tcp_bbr.c and verified its any-ACK growth reset, empty-window bandwidth-filter retention, inflight target, and precautionary stop using prior_in_flight. The current v3 copy of that file is byte-for-byte identical to the pinned copy. The pinned QUICHE source is accessible. The ledger preserves the distinction between intentional departures and unmeasured comparison questions, and the vendor-sync instructions now point to the retained ledger rather than the deleted epic.
Limits: GitHub-only inspection, not a complete independent draft-07 conformance audit or verification of every upstream branch/directory head. No builds, simulations, benchmark runs, or tests executed. Test presence is not a claim that every listed regression runs through the seeded network-simulation path. COMMENT only, not approval.

Finishes the BBR3 references epic: its children (#8805, #8806, #8807, #8809, #8810, #8848) have merged, so what was left was its own Goal, the ledger in
rs/moq-quic/README.md.Summary
rs/moq-congestion/src/bbr3/mod.rsorrs/moq-quic/src/tests;Lost::new_burstis gone. Nothing in the Plan was left undone.rs/moq-quic/README.md: a newBBR3 referencesentry names draft-07, Linux BBRv390210de4, and QUICHE BBR351a950db, rechecked today (the Linuxv3head and QUICHE'scongestion_controlare unchanged, draft-07 is still the newest), and lists every deliberate divergence. The loss-samples entry now cites draft-07.quest/m3/vendor-sync/quinn.mdpoints at the ledger entry for the reference heads instead of the finished epic.maybe_go_downrecords that the precautionary stop reads inflight with the ACK's bytes counted, as Linux'sprior_in_flight, a divergence found while writing the ledger.Decisions
C.inflightis read after the ACK; and the Reno round bound is picked from 62 or 63 (draft-07's prose) rather than its pseudocode's 63. Recommendation: keep both, measurement-only.Public API / wire
None. Docs and one comment.
Validation
Docs change; no behavior to exercise.
cargo checkandcargo fmt --checkon moq-congestion pass, andquest checkreports 539 documents ok. Reference heads checked through the GitHub API; draft-07 text read in full for the sections above.Follow-ups
None.
Models
Size: S
Implementer: Claude, claude-sonnet-5-5, high effort (arm B)
🤖 Generated with Claude Code
Review
Rewritten by
quest reviewafter each round.Head: 43171bc
Base: origin/main at 3860774
Risk: medium,
the base has no RISK.md to name what is always high or customer-facing, so every change is mediumClassified: medium by default
Reviewer: Codex, gpt-6.1-sol, medium effort
Session:
01a1e3ea-7ee7-7310-946c-e536edf2131b, round 1Score: 10/10,
The changes consolidate reference tracking through documentation and one explanatory comment without changing runtime behavior or introducing actionable defects.Rounds: 1; first round 0; findings 0: fixed 0, refuted 0, open 0
Findings: none