Repository navigation
perf(core): drop the per-state-read heap allocations in the witness path - #198
Merged
Merged
Conversation
Deletes the second sources of truth and the dead parameters/accessors the /simplify review found, with no observable change on canonical-block paths. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CHVpMMX9N69sUNbuKgpBVY
Deletes the `writer` parameter threaded through `validate_block`, `validate_block_deriving_updates`, `replay_block` and `verify_and_replay`, and with it EIP-3155 trace output. A feature deletion, not a cleanup. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CHVpMMX9N69sUNbuKgpBVY
…s_used from the executor Both are definitional rewrites of security-critical derivations, argued in the PR body: `trie_hash()` is `keccak256(encoded_2718())`, and mega-evm's `gas_used` is the last receipt's cumulative gas. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CHVpMMX9N69sUNbuKgpBVY
Block integrity verification becomes the RPC retry loop's finalize step and runs on the blocking pool; the advancer's pre-advance hooks and store commit run there too. Panic semantics are preserved and pinned by tests. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CHVpMMX9N69sUNbuKgpBVY
`WitnessDatabase` reads go through salt's `find()` with in-place decoding and stack-allocated key encodings; `LightWitness::metadata` stops cloning the `SaltValue` it only reads. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CHVpMMX9N69sUNbuKgpBVY
This was referenced Sep 3, 2026
Contributor
Claude review status
✅ Review clean Last reviewed: New this round: 0 finding(s), 0 question(s) · Resolved this round: 0 · Open questions: 0 |
Codecov Report❌ Patch coverage is
☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 50741dc320
ℹ️ About Codex in GitHub
Codex has been enabled to automatically 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 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
flyq
marked this pull request as draft
September 4, 2026 00:28
Base automatically changed from
liquan/perf/blocking-work-off-the-runtime
to
main
September 20, 2026 06:18
…tate-read-allocations # Conflicts: # bin/stateless-validator/src/app.rs # bin/stateless-validator/src/chain_sync.rs # bin/stateless-validator/src/metrics.rs # bin/stateless-validator/src/runner.rs # bin/stateless-validator/src/validator_db.rs # bin/stateless-validator/tests/integration.rs # crates/stateless-common/Cargo.toml # crates/stateless-common/src/metrics.rs # crates/stateless-common/src/rpc_client.rs # crates/stateless-core/src/chain_spec.rs # crates/stateless-core/src/db.rs # crates/stateless-core/src/executor.rs # crates/stateless-core/src/pipeline/advancer.rs # crates/stateless-core/src/pipeline/fetcher.rs # crates/stateless-core/src/pipeline/tests.rs # crates/stateless-r2/src/fetch.rs
… test bite The encoder docs justified themselves as "for per-state-read hot paths", which asks the next contributor an unmeasurable question; the decidable one is whether the caller needs to own the key. encode() is now documented as the owned form derived from the borrowing primaries. stack_key_encodings_match_encode compared each encoder against encode(), which delegates to it — true by construction, so it killed no mutation. It now asserts the literal byte layout, which catches a swapped concatenation (verified: the mutation fails the new test and passes the old one). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
flyq
marked this pull request as ready for review
September 20, 2026 07:32
Troublor
approved these changes
Sep 23, 2026
vincent-k2026
approved these changes
Sep 24, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
PR 5/6 of the #170 split. #197 has merged and
mainis merged in as of960a3bd, so this now targetsmaindirectly. Removes the heap allocations on the per-state-read path of block replay. Listed in vincent's review of #170 as "A hot-path optimization" worth its own line.This lookup runs once per unique account and storage slot touched by a block, so it is the hottest thing in
WitnessDatabase.WitnessDatabasereads go through salt'sfind(), which hands back theSaltValue(an inline array) instead ofplain_value()'s freshly heap-allocated copy of the value bytes. Callers decode from.value()in place.PlainKey::account_key_bytes(a borrow of the address) andPlainKey::storage_key_bytes(aFixedBytes<52>) replacePlainKey::{Account,Storage}(..).encode()'sVec.encode()itself delegates to them, so there is one definition of each encoding.bucket_id_for_account/bucket_id_for_slotuse them too — that is one allocation per call removed on another per-read path.LightWitness::metadatastops cloning theSaltValueit only reads.Net: two heap allocations per witnessed state read, plus one per
bucket_id_for_*call.Correctness
The interesting claim is that
find()returns the same bytesplain_value()did. That is pinned by the mainnet fixture replays (validate_block_mainnet_fixtures,validate_block_deriving_updates_mainnet_fixtures): a read that returned different bytes changes execution and fails the post-state-root check on every fixture.stack_key_encodings_match_encodepins the narrower claim: each encoder produces one specific byte layout. It asserts the literal bytes rather than one encoder againstencode()— whileencode()delegates, comparing the two passes by construction and pins nothing, so a drift in the shared primary would go unnoticed. Mutation-checked: swapping the concatenation order instorage_key_bytesfails the test.Follow-ups (deliberately not in this PR)
Two siblings of the same root cause are left alone, both because they reach outside this diff:
FIXMEatstate.rs:232names two copy sources on the read path. This PR removes the second (salt_val.value().to_vec()). The first isStateReader::valuereturningSaltValueby value, which lands here as a 94-byte clone per read; removing it means changing that trait to return a borrow and bumping the tag.LightWitnessExecutor::direct_lookup_tblallocates oneVecper witness entry at build time (light_witness.rs:253) — the same "plain key on the heap" shape, at the construction site. Unifying it with the stack encoders needs an enum key or a raw-entry lookup, since a padded 52-byte inline key cannotBorrow<[u8]>for a 20-byte account key.Testing
cargo fmt --all --check,cargo clippy --workspace --all-targets --all-features(0 warnings),cargo sort --check, full workspace suite 490 passed / 0 failed,cargo test -p stateless-core --no-default-features --lib --no-runclean.A quality pass landed in
cccd964: the encoder docs now state the rule by ownership (own the key →encode(); borrow it for a lookup → the byte encoders) instead of by "hot path", which is the question a contributor can actually answer; and the encoder test asserts literal bytes instead of comparing an encoder against the caller that forwards to it.