From a2806e0661a2de36dd203c530134b7fa765868a3 Mon Sep 17 00:00:00 2001 From: igneum-labs <337424239+igneum-labs@users.noreply.github.com> Date: Tue, 6 Oct 2026 20:29:07 +0000 Subject: [PATCH] Counter ASIC 3.0 gates (node): the P5 row, the dependency pass for 0.3.15.1 (6 to 1 vulnerabilities, the one left named) Co-Authored-By: Claude Fable 5.1 --- docs/plans/counter-asic-3-gate/node-gates.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/docs/plans/counter-asic-3-gate/node-gates.md b/docs/plans/counter-asic-3-gate/node-gates.md index 2187f211e..19c261eb6 100644 --- a/docs/plans/counter-asic-3-gate/node-gates.md +++ b/docs/plans/counter-asic-3-gate/node-gates.md @@ -21,4 +21,6 @@ | P4 | the two 0.3.15 node faults found on the live devnet (the canary's version 1026, the pruned hub's sync panic), fixed in the same fork tree | the unit tests; the node-cut compat gate on real nodes (a MINING new node beside a 0.3.14 node on the live object, the old node accepting the new node's blocks; a clean new node joining through the old hub; a clean OLD node joining through the NEW node, which must survive serving a sync from the genesis; the new node restarted and re-synced; every header version the block version; no panic); the ledger rows | GREEN on fork ca3-v4-order-fix 7961c5f1 (= 791ff22c + 713ef876 + 17c60367 the signal gate + 7961c5f1 the retention error; the shipper's release-0.3.15-node). The gate `infra/fast-time/node-compat.mjs` (20:15:5x to 20:19:04Z, suffix 994, the live object plus CPU genesis bits, the 0.3.14 binary `target-0314/release/igneumd` as the old side): SUMMARY PASS: one digest `b0afb2ee93d0d627` on all five nodes; accepted old/new 120/75, rejected 0/0; header versions {0: the genesis, 2: 195}; counts 195/195/195/195/195 at the end (the clean new node through the old hub, the clean old node served from the genesis by the new node, which stayed alive, and the restarted new node all at 195); no `wrong block version`, reject or panic line in any log. The known-failed case on the canary binary 713ef876 beside 0.3.14 (20:00 to 20:02Z): SUMMARY FAIL, `HandleRelayInvsFlow flow error: wrong block version: got 1026 but expected 2, disconnecting` on the old nodes, `P2P, got reject message: wrong block version` on the new, counts 138 against the new node's 101, the restarted new node never re-synced (the fleet's p2-3090-1). Unit tests: `block_template_uses_current_block_version` (2 on the thirteen-field state, 1026 with both v4 fields), `program_class_signal_rule`, `a_sync_request_below_retention_is_an_error_not_a_panic`; the box on 7961c5f1: kaspa-consensus 99, consensus-core 111 + 7, rc 0 (the shipper's six-crate line: exec 20, miner 18, pow 15, p2p-flows 33). Ledger rows N1 and N2 in `docs/fud-ledger.md`. Owed: a truly pruned serving node in a harness (hours of fast-time chain at pruning depth 13,838); an old node whose datadir holds version-1026 blocks relaying them (the shipper's roll-back finding), queued. Summaries `node-compat-20261006-fixed.json`, `node-compat-20261006-canary-failed-case.json` | +| P5 | the dependency pass for 0.3.15.1 (the night battery's audit, mirror branch fbb72e5: the fork's lock file) | `cargo audit` before and after on the box; the six-crate suite | DONE on fork `ca3-v4-deps` f8f0f1df (from 7961c5f1): `cargo audit` on igneum-build-1 before: 6 vulnerabilities (crossbeam-epoch RUSTSEC-2026-0204, h2 2026-0258, quinn-proto 2026-0185, ruint 2026-0220, rustls 2026-0285, tracing-subscriber 2025-0055; the battery's 22 counts advisory paths), 16 warnings (unmaintained, unsound, yanked); the three peer-reachable crates first (h2 0.4.6 to 0.4.20, quinn-proto 0.11.14 to 0.11.19, rustls 0.23.39 to 0.23.45 with rustls-webpki 0.103.15), then crossbeam-epoch 0.9.21 and ruint 1.20.1 (rand_pcg added), every one a patch or minor bump by the audit's own solution line; after: 1 vulnerability, 16 warnings; the one left is tracing-subscriber 0.2.25, pinned by ark-relations 0.5.1 through ark-groth16 and ark-snark (a 0.2 to 0.3 major bump outside the fork's pins, which name 0.3 in bridge/Cargo.toml), named here, not taken. The six-crate suite on the box on the new lock: kaspa-consensus 99, consensus-core 111 + 7, igneum-exec 20, kaspa-pow 15, igneum-miner 18, kaspa-p2p-flows 33, rc 0 (80 s). The 16 warnings are unmaintained or unsound transitive crates (async-std, atty, bincode 1, derivative, instant, mach, paste, proc-macro-error, rustls-pemfile, anyhow, event-listener, faster-hex, lru, chacha20, spin), none with a bump to take, each a dependency choice for 0.3.16 | + Summary files in this directory: `class-v4-20261006-1553Z-cpu.json` (G4 run 1), `class-v4-20261006-never-failed-case.json` (G4 run 2, the failed case), `class-v4-20261006-metal.json` (G4b), `class-v4-20261006-id-rerun.json` (G4 on the program-id fix, with the id assertion), `class-v4-20261006-id-failed-case.json` (the id assertion's failed case).