fix(db): return error when MDBX commit is aborted

MDBX's transaction commit returns MDBX_RESULT_TRUE (-1) when the
transaction was aborted due to previous errors. The Rust wrapper
converts this to Ok(true), but reth's DbTx::commit() implementation
was passing this through as a successful result.

Since commit() is typically used as `provider_rw.commit()?` throughout
the codebase, which only checks for Err variants, these silent failures
were going unnoticed - potentially leading to data loss.

This commit:
- Changes DbTx::commit() return type from Result<bool, _> to Result<(), _>
- Maps MDBX_RESULT_TRUE to a proper error in the mdbx tx implementation
- Updates all trait implementations to match the new signature

The bool return value was not useful anyway since it only indicated
whether the commit was a no-op (MDBX_RESULT_TRUE = aborted) vs real
success (MDBX_SUCCESS/false), and callers never checked it.

See: https://libmdbx.dqdkfa.ru/group__c__transactions.html#ga3146a36cf1d0603cc6e3d52e97ec38a9
Amp-Thread-ID: https://ampcode.com/threads/T-019bbe8c-b7b0-722a-8f37-7d4c19682b06
Co-authored-by: Amp <amp@ampcode.com>

.github/workflows/docker-nightly.yml

@@ -28,10 +28,14 @@ jobs:
         build:
           - name: 'Build and push the nightly reth image'
             command: 'make PROFILE=maxperf docker-build-push-nightly'
+          - name: 'Build and push the nightly edge profiling reth image'
+            command: 'make PROFILE=profiling docker-build-push-nightly-edge-profiling'
           - name: 'Build and push the nightly profiling reth image'
             command: 'make PROFILE=profiling docker-build-push-nightly-profiling'
           - name: 'Build and push the nightly op-reth image'
             command: 'make IMAGE_NAME=$OP_IMAGE_NAME DOCKER_IMAGE_NAME=$OP_DOCKER_IMAGE_NAME PROFILE=maxperf op-docker-build-push-nightly'
+          - name: 'Build and push the nightly edge profiling op-reth image'
+            command: 'make IMAGE_NAME=$OP_IMAGE_NAME DOCKER_IMAGE_NAME=$OP_DOCKER_IMAGE_NAME PROFILE=profiling op-docker-build-push-nightly-edge-profiling'
           - name: 'Build and push the nightly profiling op-reth image'
             command: 'make IMAGE_NAME=$OP_IMAGE_NAME DOCKER_IMAGE_NAME=$OP_DOCKER_IMAGE_NAME PROFILE=profiling op-docker-build-push-nightly-profiling'
     steps:

.github/workflows/hive.yml

@@ -15,16 +15,27 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
-  prepare-reth:
+  prepare-reth-stable:
     uses: ./.github/workflows/prepare-reth.yml
     with:
       image_tag: ghcr.io/paradigmxyz/reth:latest
       binary_name: reth
+      cargo_features: "asm-keccak"
+      artifact_name: "reth-stable"
+
+  prepare-reth-edge:
+    uses: ./.github/workflows/prepare-reth.yml
+    with:
+      image_tag: ghcr.io/paradigmxyz/reth:latest
+      binary_name: reth
+      cargo_features: "asm-keccak edge"
+      artifact_name: "reth-edge"
 
   prepare-hive:
     if: github.repository == 'paradigmxyz/reth'
     timeout-minutes: 45
-    runs-on: depot-ubuntu-latest-16
+    runs-on:
+      group: Reth
     steps:
       - uses: actions/checkout@v6
       - name: Checkout hive tests
@@ -76,6 +87,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
+        storage: [stable, edge]
         # ethereum/rpc to be deprecated:
         # https://github.com/ethereum/hive/pull/1117
         scenario:
@@ -85,7 +97,7 @@ jobs:
           - sim: devp2p
             limit: discv4
           # started failing after https://github.com/ethereum/go-ethereum/pull/31843, no
-          # action on our side, remove from here when we get unxpected passes on these tests
+          # action on our side, remove from here when we get unexpected passes on these tests
           # - sim: devp2p
           #  limit: eth
           #  include:
@@ -175,10 +187,12 @@ jobs:
           - sim: ethereum/eels/consume-rlp
             limit: .*tests/paris.*
     needs:
-      - prepare-reth
+      - prepare-reth-stable
+      - prepare-reth-edge
       - prepare-hive
-    name: run ${{ matrix.scenario.sim }}${{ matrix.scenario.limit && format(' - {0}', matrix.scenario.limit) }}
-    runs-on: depot-ubuntu-latest-16
+    name: ${{ matrix.storage }} / ${{ matrix.scenario.sim }}${{ matrix.scenario.limit && format(' - {0}', matrix.scenario.limit) }}
+    runs-on:
+      group: Reth
     permissions:
       issues: write
     steps:
@@ -195,7 +209,7 @@ jobs:
       - name: Download reth image
         uses: actions/download-artifact@v7
         with:
-          name: artifacts
+          name: reth-${{ matrix.storage }}
           path: /tmp
 
       - name: Load Docker images

.github/workflows/lint.yml

@@ -277,7 +277,7 @@ jobs:
       - uses: dtolnay/rust-toolchain@stable
       - uses: mozilla-actions/sccache-action@v0.0.9
       - uses: rui314/setup-mold@v1
-      - uses: taiki-e/cache-cargo-install-action@v2
+      - uses: taiki-e/cache-cargo-install-action@v3
         with:
           tool: zepter
       - name: Eagerly pull dependencies

.github/workflows/prepare-reth.yml

@@ -21,6 +21,11 @@ on:
         required: false
         type: string
         description: "Optional cargo package path"
+      artifact_name:
+        required: false
+        type: string
+        default: "artifacts"
+        description: "Name for the uploaded artifact"
 
 jobs:
   prepare-reth:
@@ -52,5 +57,5 @@ jobs:
         id: upload
         uses: actions/upload-artifact@v6
         with:
-          name: artifacts
+          name: ${{ inputs.artifact_name }}
           path: ./artifacts

.github/workflows/unit.yml

@@ -19,29 +19,22 @@ concurrency:
 
 jobs:
   test:
-    name: test / ${{ matrix.type }} (${{ matrix.partition }}/${{ matrix.total_partitions }})
+    name: test / ${{ matrix.type }} / ${{ matrix.storage }}
     runs-on: depot-ubuntu-latest-4
     env:
       RUST_BACKTRACE: 1
+      EDGE_FEATURES: ${{ matrix.storage == 'edge' && 'edge' || '' }}
     strategy:
       matrix:
+        type: [ethereum, optimism]
+        storage: [stable, edge]
         include:
           - type: ethereum
-            args: --features "asm-keccak ethereum" --locked
-            partition: 1
-            total_partitions: 2
-          - type: ethereum
-            args: --features "asm-keccak ethereum" --locked
-            partition: 2
-            total_partitions: 2
+            features: asm-keccak ethereum
+            exclude_args: ""
           - type: optimism
-            args: --features "asm-keccak" --locked --exclude reth --exclude reth-bench --exclude "example-*" --exclude "reth-ethereum-*" --exclude "*-ethereum"
-            partition: 1
-            total_partitions: 2
-          - type: optimism
-            args: --features "asm-keccak" --locked --exclude reth --exclude reth-bench --exclude "example-*" --exclude "reth-ethereum-*" --exclude "*-ethereum"
-            partition: 2
-            total_partitions: 2
+            features: asm-keccak
+            exclude_args: --exclude reth --exclude reth-bench --exclude "example-*" --exclude "reth-ethereum-*" --exclude "*-ethereum"
     timeout-minutes: 30
     steps:
       - uses: actions/checkout@v6
@@ -59,9 +52,9 @@ jobs:
       - name: Run tests
         run: |
           cargo nextest run \
-            ${{ matrix.args }} --workspace \
+            --features "${{ matrix.features }} $EDGE_FEATURES" --locked \
+            ${{ matrix.exclude_args }} --workspace \
             --exclude ef-tests --no-tests=warn \
-            --partition hash:${{ matrix.partition }}/2 \
             -E "!kind(test) and not binary(e2e_testsuite)"
 
   state:

.gitignore

@@ -12,6 +12,9 @@ target/
 # Generated by Intellij-based IDEs.
 .idea
 
+# ck-search metadata
+.ck
+
 # Generated by MacOS
 .DS_Store
 

CLAUDE.md

@@ -18,7 +18,7 @@ Reth is a high-performance Ethereum execution client written in Rust, focusing o
 6. **Pipeline (`crates/stages/`)**: Staged sync architecture for blockchain synchronization
 7. **Trie (`crates/trie/`)**: Merkle Patricia Trie implementation with parallel state root computation
 8. **Node Builder (`crates/node/`)**: High-level node orchestration and configuration
-9  **The Consensus Engine (`crates/engine/`)**: Handles processing blocks received from the consensus layer with the Engine API (newPayload, forkchoiceUpdated)
+9. **The Consensus Engine (`crates/engine/`)**: Handles processing blocks received from the consensus layer with the Engine API (newPayload, forkchoiceUpdated)
 
 ### Key Design Principles
 

CONTRIBUTING.md

@@ -51,9 +51,7 @@ elsewhere.
 <!-- -   **Asking in the support Telegram:** The [Foundry Support Telegram][support-tg] is a fast and easy way to ask questions. -->
 <!-- -   **Opening a discussion:** This repository comes with a discussions board where you can also ask for help. Click the "Discussions" tab at the top. -->
 
-If you have reviewed existing documentation and still have questions, or you are having problems, you can get help by *
-*opening a discussion**. This repository comes with a discussions board where you can also ask for help. Click the "
-Discussions" tab at the top.
+If you have reviewed existing documentation and still have questions, or you are having problems, you can get help by **opening a discussion**. This repository comes with a discussions board where you can also ask for help. Click the "Discussions" tab at the top.
 
 As Reth is still in heavy development, the documentation can be a bit scattered. The [Reth Docs][reth-docs] is our
 current best-effort attempt at keeping up-to-date information.

Cargo.toml

@@ -487,7 +487,7 @@ revm-inspectors = "0.33.2"
 alloy-chains = { version = "0.2.5", default-features = false }
 alloy-dyn-abi = "1.4.1"
 alloy-eip2124 = { version = "0.2.0", default-features = false }
-alloy-eip7928 = { version = "0.1.0" }
+alloy-eip7928 = { version = "0.1.0", default-features = false }
 alloy-evm = { version = "0.25.1", default-features = false }
 alloy-primitives = { version = "1.5.0", default-features = false, features = ["map-foldhash"] }
 alloy-rlp = { version = "0.3.10", default-features = false, features = ["core-net"] }
@@ -497,33 +497,33 @@ alloy-trie = { version = "0.9.1", default-features = false }
 
 alloy-hardforks = "0.4.5"
 
-alloy-consensus = { version = "1.1.3", default-features = false }
-alloy-contract = { version = "1.1.3", default-features = false }
-alloy-eips = { version = "1.1.3", default-features = false }
-alloy-genesis = { version = "1.1.3", default-features = false }
-alloy-json-rpc = { version = "1.1.3", default-features = false }
-alloy-network = { version = "1.1.3", default-features = false }
-alloy-network-primitives = { version = "1.1.3", default-features = false }
-alloy-provider = { version = "1.1.3", features = ["reqwest", "debug-api"], default-features = false }
-alloy-pubsub = { version = "1.1.3", default-features = false }
-alloy-rpc-client = { version = "1.1.3", default-features = false }
-alloy-rpc-types = { version = "1.1.3", features = ["eth"], default-features = false }
-alloy-rpc-types-admin = { version = "1.1.3", default-features = false }
-alloy-rpc-types-anvil = { version = "1.1.3", default-features = false }
-alloy-rpc-types-beacon = { version = "1.1.3", default-features = false }
-alloy-rpc-types-debug = { version = "1.1.3", default-features = false }
-alloy-rpc-types-engine = { version = "1.1.3", default-features = false }
-alloy-rpc-types-eth = { version = "1.1.3", default-features = false }
-alloy-rpc-types-mev = { version = "1.1.3", default-features = false }
-alloy-rpc-types-trace = { version = "1.1.3", default-features = false }
-alloy-rpc-types-txpool = { version = "1.1.3", default-features = false }
-alloy-serde = { version = "1.1.3", default-features = false }
-alloy-signer = { version = "1.1.3", default-features = false }
-alloy-signer-local = { version = "1.1.3", default-features = false }
-alloy-transport = { version = "1.1.3" }
-alloy-transport-http = { version = "1.1.3", features = ["reqwest-rustls-tls"], default-features = false }
-alloy-transport-ipc = { version = "1.1.3", default-features = false }
-alloy-transport-ws = { version = "1.1.3", default-features = false }
+alloy-consensus = { version = "1.2.1", default-features = false }
+alloy-contract = { version = "1.2.1", default-features = false }
+alloy-eips = { version = "1.2.1", default-features = false }
+alloy-genesis = { version = "1.2.1", default-features = false }
+alloy-json-rpc = { version = "1.2.1", default-features = false }
+alloy-network = { version = "1.2.1", default-features = false }
+alloy-network-primitives = { version = "1.2.1", default-features = false }
+alloy-provider = { version = "1.2.1", features = ["reqwest", "debug-api"], default-features = false }
+alloy-pubsub = { version = "1.2.1", default-features = false }
+alloy-rpc-client = { version = "1.2.1", default-features = false }
+alloy-rpc-types = { version = "1.2.1", features = ["eth"], default-features = false }
+alloy-rpc-types-admin = { version = "1.2.1", default-features = false }
+alloy-rpc-types-anvil = { version = "1.2.1", default-features = false }
+alloy-rpc-types-beacon = { version = "1.2.1", default-features = false }
+alloy-rpc-types-debug = { version = "1.2.1", default-features = false }
+alloy-rpc-types-engine = { version = "1.2.1", default-features = false }
+alloy-rpc-types-eth = { version = "1.2.1", default-features = false }
+alloy-rpc-types-mev = { version = "1.2.1", default-features = false }
+alloy-rpc-types-trace = { version = "1.2.1", default-features = false }
+alloy-rpc-types-txpool = { version = "1.2.1", default-features = false }
+alloy-serde = { version = "1.2.1", default-features = false }
+alloy-signer = { version = "1.2.1", default-features = false }
+alloy-signer-local = { version = "1.2.1", default-features = false }
+alloy-transport = { version = "1.2.1" }
+alloy-transport-http = { version = "1.2.1", features = ["reqwest-rustls-tls"], default-features = false }
+alloy-transport-ipc = { version = "1.2.1", default-features = false }
+alloy-transport-ws = { version = "1.2.1", default-features = false }
 
 # op
 alloy-op-evm = { version = "0.25.0", default-features = false }
@@ -548,6 +548,7 @@ bytes = { version = "1.5", default-features = false }
 brotli = "8"
 cfg-if = "1.0"
 clap = "4"
+color-eyre = "0.6"
 dashmap = "6.0"
 derive_more = { version = "2", default-features = false, features = ["full"] }
 dirs-next = "2.0.0"
@@ -587,6 +588,7 @@ url = { version = "2.3", default-features = false }
 zstd = "0.13"
 byteorder = "1"
 mini-moka = "0.10"
+moka = "0.12"
 tar-no-std = { version = "0.3.2", default-features = false }
 miniz_oxide = { version = "0.8.4", default-features = false }
 chrono = "0.4.41"
@@ -682,6 +684,7 @@ ethereum_ssz = "0.9.0"
 ethereum_ssz_derive = "0.9.0"
 
 # allocators
+jemalloc_pprof = { version = "0.8", default-features = false }
 tikv-jemalloc-ctl = "0.6"
 tikv-jemallocator = "0.6"
 tracy-client = "0.18.0"
@@ -692,7 +695,7 @@ ahash = "0.8"
 anyhow = "1.0"
 bindgen = { version = "0.71", default-features = false }
 block-padding = "0.3.2"
-cc = "=1.2.15"
+cc = "1.2.15"
 cipher = "0.4.3"
 comfy-table = "7.0"
 concat-kdf = "0.1.0"
@@ -729,6 +732,7 @@ socket2 = { version = "0.5", default-features = false }
 sysinfo = { version = "0.33", default-features = false }
 tracing-journald = "0.3"
 tracing-logfmt = "0.3.3"
+tracing-samply = "0.1"
 tracing-subscriber = { version = "0.3", default-features = false }
 triehash = "0.8"
 typenum = "1.15.0"

Dockerfile

@@ -18,7 +18,7 @@ FROM chef AS builder
 COPY --from=planner /app/recipe.json recipe.json
 
 # Build profile, release by default
-ARG BUILD_PROFILE=release
+ARG BUILD_PROFILE=maxperf
 ENV BUILD_PROFILE=$BUILD_PROFILE
 
 # Extra Cargo flags

DockerfileOp

@@ -14,7 +14,7 @@ RUN cargo chef prepare --recipe-path recipe.json
 FROM chef AS builder
 COPY --from=planner /app/recipe.json recipe.json
 
-ARG BUILD_PROFILE=release
+ARG BUILD_PROFILE=maxperf
 ENV BUILD_PROFILE=$BUILD_PROFILE
 
 ARG RUSTFLAGS=""

LICENSE-APACHE

@@ -186,7 +186,7 @@ APPENDIX: How to apply the Apache License to your work.
    same "printed page" as the copyright notice for easier
    identification within third-party archives.
 
-Copyright 2022-2025 Reth Contributors
+Copyright 2022-2026 Reth Contributors
 
 Licensed under the Apache License, Version 2.0 (the "License");
 you may not use this file except in compliance with the License.

LICENSE-MIT

@@ -1,6 +1,6 @@
 The MIT License (MIT)
 
-Copyright (c) 2022-2025 Reth Contributors
+Copyright (c) 2022-2026 Reth Contributors
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

Makefile

@@ -276,6 +276,11 @@ docker-build-push-latest: ## Build and push a cross-arch Docker image tagged wit
 docker-build-push-nightly: ## Build and push cross-arch Docker image tagged with the latest git tag with a `-nightly` suffix, and `latest-nightly`.
 	$(call docker_build_push,nightly,nightly)
 
+.PHONY: docker-build-push-nightly-edge-profiling
+docker-build-push-nightly-edge-profiling: FEATURES := $(FEATURES) edge
+docker-build-push-nightly-edge-profiling: ## Build and push cross-arch Docker image with edge features tagged with `nightly-edge-profiling`.
+	$(call docker_build_push,nightly-edge-profiling,nightly-edge-profiling)
+
 # Create a cross-arch Docker image with the given tags and push it
 define docker_build_push
 	$(MAKE) build-x86_64-unknown-linux-gnu
@@ -328,6 +333,11 @@ op-docker-build-push-latest: ## Build and push a cross-arch Docker image tagged
 op-docker-build-push-nightly: ## Build and push cross-arch Docker image tagged with the latest git tag with a `-nightly` suffix, and `latest-nightly`.
 	$(call op_docker_build_push,nightly,nightly)
 
+.PHONY: op-docker-build-push-nightly-edge-profiling
+op-docker-build-push-nightly-edge-profiling: FEATURES := $(FEATURES) edge
+op-docker-build-push-nightly-edge-profiling: ## Build and push cross-arch Docker image with edge features tagged with `nightly-edge-profiling`.
+	$(call op_docker_build_push,nightly-edge-profiling,nightly-edge-profiling)
+
 # Note: This requires a buildx builder with emulation support. For example:
 #
 # `docker run --privileged --rm tonistiigi/binfmt --install amd64,arm64`

README.md

@@ -44,7 +44,7 @@ More historical context below:
 -   We released 1.0 "production-ready" stable Reth in June 2024.
     -   Reth completed an audit with [Sigma Prime](https://sigmaprime.io/), the developers of [Lighthouse](https://github.com/sigp/lighthouse), the Rust Consensus Layer implementation. Find it [here](./audit/sigma_prime_audit_v2.pdf).
     -   Revm (the EVM used in Reth) underwent an audit with [Guido Vranken](https://x.com/guidovranken) (#1 [Ethereum Bug Bounty](https://ethereum.org/en/bug-bounty)). We will publish the results soon.
--   We released multiple iterative beta versions, up to [beta.9](https://github.com/paradigmxyz/reth/releases/tag/v0.2.0-beta.9) on Monday June 3, 2024,the last beta release.
+-   We released multiple iterative beta versions, up to [beta.9](https://github.com/paradigmxyz/reth/releases/tag/v0.2.0-beta.9) on Monday June 3, 2024, the last beta release.
 -   We released [beta](https://github.com/paradigmxyz/reth/releases/tag/v0.2.0-beta.1) on Monday March 4, 2024, our first breaking change to the database model, providing faster query speed, smaller database footprint, and allowing "history" to be mounted on separate drives.
 -   We shipped iterative improvements until the last alpha release on February 28, 2024, [0.1.0-alpha.21](https://github.com/paradigmxyz/reth/releases/tag/v0.1.0-alpha.21).
 -   We [initially announced](https://www.paradigm.xyz/2023/06/reth-alpha) [0.1.0-alpha.1](https://github.com/paradigmxyz/reth/releases/tag/v0.1.0-alpha.1) on June 20, 2023.

bin/reth-bench-compare/Cargo.toml

@@ -25,7 +25,9 @@ reth-chainspec.workspace = true
 
 # alloy
 alloy-provider = { workspace = true, features = ["reqwest-rustls-tls"], default-features = false }
+alloy-rpc-client = { workspace = true, features = ["pubsub"] }
 alloy-rpc-types-eth.workspace = true
+alloy-transport-ws.workspace = true
 alloy-primitives.workspace = true
 
 # CLI and argument parsing

bin/reth-bench-compare/README.md

@@ -0,0 +1,50 @@
+# reth-bench-compare
+
+Compare reth performance between two git references.
+
+## Usage
+
+```bash
+reth-bench-compare \
+  --baseline-ref main \
+  --feature-ref my-feature \
+  --blocks 100 \
+  --wait-for-persistence
+```
+
+## Arguments
+
+| Argument | Description | Default | Required |
+|----------|-------------|---------|----------|
+| `--baseline-ref <REF>` | Git reference for baseline | - | Yes |
+| `--feature-ref <REF>` | Git reference to compare | - | Yes |
+| `--blocks <N>` | Number of blocks to benchmark | `100` | No |
+| `--chain <CHAIN>` | Chain to benchmark | `mainnet` | No |
+| `--datadir <PATH>` | Data directory path | OS-specific | No |
+| `--rpc-url <URL>` | RPC endpoint for block data | Chain default | No |
+| `--output-dir <PATH>` | Output directory | `./reth-bench-compare` | No |
+| `--wait-for-persistence` | Wait for block persistence | `false` | No |
+| `--persistence-threshold <N>` | Wait after every N+1 blocks | `2` | No |
+| `--wait-time <DURATION>` | Fixed delay (legacy) | - | No |
+| `--warmup-blocks <N>` | Cache warmup blocks | Same as `--blocks` | No |
+| `--draw` | Generate charts (needs Python/uv) | `false` | No |
+| `--profile` | Enable CPU profiling (needs samply) | `false` | No |
+| `-vvvv` | Debug logging | Info | No |
+| `--features <FEATURES>` | Rust features for both builds | `jemalloc,asm-keccak` | No |
+| `--rustflags <FLAGS>` | RUSTFLAGS for both builds | `-C target-cpu=native` | No |
+| `--baseline-features <FEATURES>` | Features for baseline only | Inherits `--features` | No |
+| `--feature-features <FEATURES>` | Features for feature only | Inherits `--features` | No |
+| `--baseline-rustflags <FLAGS>` | RUSTFLAGS for baseline only | Inherits `--rustflags` | No |
+| `--feature-rustflags <FLAGS>` | RUSTFLAGS for feature only | Inherits `--rustflags` | No |
+| `--baseline-args <ARGS>` | Extra args for baseline node | - | No |
+| `--feature-args <ARGS>` | Extra args for feature node | - | No |
+| `--metrics-port <PORT>` | Metrics endpoint port | `5005` | No |
+| `--sudo` | Run with elevated privileges | `false` | No |
+
+## Output
+
+Results in `./reth-bench-compare/results/<timestamp>/`:
+- `comparison_report.json` - Metrics comparison
+- `per_block_comparison.csv` - Per-block statistics
+- `baseline/` and `feature/` - Individual run results
+- `latency_comparison.png` - Chart (if `--draw` used)

bin/reth-bench-compare/src/benchmark.rs

@@ -18,6 +18,8 @@ pub(crate) struct BenchmarkRunner {
     rpc_url: String,
     jwt_secret: String,
     wait_time: Option<String>,
+    wait_for_persistence: bool,
+    persistence_threshold: Option<u64>,
     warmup_blocks: u64,
 }
 
@@ -28,6 +30,8 @@ impl BenchmarkRunner {
             rpc_url: args.get_rpc_url(),
             jwt_secret: args.jwt_secret_path().to_string_lossy().to_string(),
             wait_time: args.wait_time.clone(),
+            wait_for_persistence: args.wait_for_persistence,
+            persistence_threshold: args.persistence_threshold,
             warmup_blocks: args.get_warmup_blocks(),
         }
     }
@@ -96,13 +100,9 @@ impl BenchmarkRunner {
             &from_block.to_string(),
             "--to",
             &to_block.to_string(),
+            "--wait-time=0ms", // Warmup should avoid persistence waits.
         ]);
 
-        // Add wait-time argument if provided
-        if let Some(ref wait_time) = self.wait_time {
-            cmd.args(["--wait-time", wait_time]);
-        }
-
         cmd.env("RUST_LOG_STYLE", "never")
             .stdout(std::process::Stdio::piped())
             .stderr(std::process::Stdio::piped())
@@ -186,9 +186,16 @@ impl BenchmarkRunner {
             &output_dir.to_string_lossy(),
         ]);
 
-        // Add wait-time argument if provided
+        // Configure wait mode: wait-time takes precedence over persistence-based flow
         if let Some(ref wait_time) = self.wait_time {
             cmd.args(["--wait-time", wait_time]);
+        } else if self.wait_for_persistence {
+            cmd.arg("--wait-for-persistence");
+
+            // Add persistence threshold if specified
+            if let Some(threshold) = self.persistence_threshold {
+                cmd.args(["--persistence-threshold", &threshold.to_string()]);
+            }
         }
 
         cmd.env("RUST_LOG_STYLE", "never")

bin/reth-bench-compare/src/cli.rs

@@ -114,10 +114,29 @@ pub(crate) struct Args {
     #[arg(long)]
     pub profile: bool,
 
-    /// Wait time between engine API calls (passed to reth-bench)
-    #[arg(long, value_name = "DURATION")]
+    /// Optional fixed delay between engine API calls (passed to reth-bench).
+    ///
+    /// When set, reth-bench uses wait-time mode and disables persistence-based flow.
+    /// This flag remains for compatibility with older scripts.
+    #[arg(long, value_name = "DURATION", hide = true)]
     pub wait_time: Option<String>,
 
+    /// Wait for blocks to be persisted before sending the next batch (passed to reth-bench).
+    ///
+    /// When enabled, waits for every Nth block to be persisted using the
+    /// `reth_subscribePersistedBlock` subscription. This ensures the benchmark
+    /// doesn't outpace persistence.
+    #[arg(long)]
+    pub wait_for_persistence: bool,
+
+    /// Engine persistence threshold (passed to reth-bench).
+    ///
+    /// The benchmark waits after every `(threshold + 1)` blocks. By default this
+    /// matches the engine's default persistence threshold (2), so waits occur
+    /// at blocks 3, 6, 9, etc.
+    #[arg(long, value_name = "PERSISTENCE_THRESHOLD")]
+    pub persistence_threshold: Option<u64>,
+
     /// Number of blocks to run for cache warmup after clearing caches.
     /// If not specified, defaults to the same as --blocks
     #[arg(long, value_name = "N")]
@@ -164,12 +183,42 @@ pub(crate) struct Args {
     #[arg(trailing_var_arg = true, allow_hyphen_values = true)]
     pub reth_args: Vec<String>,
 
-    /// Comma-separated list of features to enable during reth compilation
+    /// Comma-separated list of features to enable during reth compilation (applied to both builds)
     ///
     /// Example: `jemalloc,asm-keccak`
     #[arg(long, value_name = "FEATURES", default_value = "jemalloc,asm-keccak")]
     pub features: String,
 
+    /// Comma-separated list of features to enable only for baseline build (overrides --features)
+    ///
+    /// Example: `--baseline-features jemalloc`
+    #[arg(long, value_name = "FEATURES")]
+    pub baseline_features: Option<String>,
+
+    /// Comma-separated list of features to enable only for feature build (overrides --features)
+    ///
+    /// Example: `--feature-features jemalloc,asm-keccak`
+    #[arg(long, value_name = "FEATURES")]
+    pub feature_features: Option<String>,
+
+    /// RUSTFLAGS to use for both baseline and feature builds
+    ///
+    /// Example: `--rustflags "-C target-cpu=native"`
+    #[arg(long, value_name = "FLAGS", default_value = "-C target-cpu=native")]
+    pub rustflags: String,
+
+    /// RUSTFLAGS to use only for baseline build (overrides --rustflags)
+    ///
+    /// Example: `--baseline-rustflags "-C target-cpu=native -C lto"`
+    #[arg(long, value_name = "FLAGS")]
+    pub baseline_rustflags: Option<String>,
+
+    /// RUSTFLAGS to use only for feature build (overrides --rustflags)
+    ///
+    /// Example: `--feature-rustflags "-C target-cpu=native -C lto"`
+    #[arg(long, value_name = "FLAGS")]
+    pub feature_rustflags: Option<String>,
+
     /// Disable automatic --debug.startup-sync-state-idle flag for specific runs.
     /// Can be "baseline", "feature", or "all".
     /// By default, the flag is passed to warmup, baseline, and feature runs.
@@ -328,7 +377,6 @@ pub(crate) async fn run_comparison(args: Args, _ctx: CliContext) -> Result<()> {
         git_manager.repo_root().to_string(),
         output_dir.clone(),
         git_manager.clone(),
-        args.features.clone(),
     )?;
     // Initialize node manager
     let mut node_manager = NodeManager::new(&args);
@@ -448,6 +496,18 @@ async fn run_compilation_phase(
         let ref_type = ref_types[i];
         let commit = &ref_commits[git_ref];
 
+        // Get per-build features and rustflags
+        let features = match ref_type {
+            "baseline" => args.baseline_features.as_ref().unwrap_or(&args.features),
+            "feature" => args.feature_features.as_ref().unwrap_or(&args.features),
+            _ => &args.features,
+        };
+        let rustflags = match ref_type {
+            "baseline" => args.baseline_rustflags.as_ref().unwrap_or(&args.rustflags),
+            "feature" => args.feature_rustflags.as_ref().unwrap_or(&args.rustflags),
+            _ => &args.rustflags,
+        };
+
         info!(
             "Compiling {} binary for reference: {} (commit: {})",
             ref_type,
@@ -459,7 +519,7 @@ async fn run_compilation_phase(
         git_manager.switch_ref(git_ref)?;
 
         // Compile reth (with caching)
-        compilation_manager.compile_reth(commit, is_optimism)?;
+        compilation_manager.compile_reth(commit, is_optimism, features, rustflags)?;
 
         info!("Completed compilation for {} reference", ref_type);
     }
@@ -471,6 +531,7 @@ async fn run_compilation_phase(
     Ok((baseline_commit, feature_commit))
 }
 
+#[allow(clippy::too_many_arguments)]
 /// Run warmup phase to warm up caches before benchmarking
 async fn run_warmup_phase(
     git_manager: &GitManager,
@@ -480,9 +541,15 @@ async fn run_warmup_phase(
     args: &Args,
     is_optimism: bool,
     baseline_commit: &str,
+    starting_tip: u64,
 ) -> Result<()> {
     info!("=== Running warmup phase ===");
 
+    // Unwind to starting block minus warmup blocks, so we end up back at starting_tip
+    let warmup_blocks = args.get_warmup_blocks();
+    let unwind_target = starting_tip.saturating_sub(warmup_blocks);
+    node_manager.unwind_to_block(unwind_target).await?;
+
     // Use baseline for warmup
     let warmup_ref = &args.baseline_ref;
 
@@ -511,12 +578,9 @@ async fn run_warmup_phase(
         node_manager.start_node(&binary_path, warmup_ref, "warmup", &additional_args).await?;
 
     // Wait for node to be ready and get its current tip
-    let current_tip = node_manager.wait_for_node_ready_and_get_tip().await?;
+    let current_tip = node_manager.wait_for_node_ready_and_get_tip(&mut node_process).await?;
     info!("Warmup node is ready at tip: {}", current_tip);
 
-    // Store the tip we'll unwind back to
-    let original_tip = current_tip;
-
     // Clear filesystem caches before warmup run only (unless disabled)
     if args.no_clear_cache {
         info!("Skipping filesystem cache clearing (--no-clear-cache flag set)");
@@ -527,12 +591,9 @@ async fn run_warmup_phase(
     // Run warmup to warm up caches
     benchmark_runner.run_warmup(current_tip).await?;
 
-    // Stop node before unwinding (node must be stopped to release database lock)
+    // Stop node after warmup
     node_manager.stop_node(&mut node_process).await?;
 
-    // Unwind back to starting block after warmup
-    node_manager.unwind_to_block(original_tip).await?;
-
     info!("Warmup phase completed");
     Ok(())
 }
@@ -554,6 +615,27 @@ async fn run_benchmark_workflow(
     let (baseline_commit, feature_commit) =
         run_compilation_phase(git_manager, compilation_manager, args, is_optimism).await?;
 
+    // Switch to baseline reference and get the starting tip
+    git_manager.switch_ref(&args.baseline_ref)?;
+    let binary_path =
+        compilation_manager.get_cached_binary_path_for_commit(&baseline_commit, is_optimism);
+    if !binary_path.exists() {
+        return Err(eyre!(
+            "Cached baseline binary not found at {:?}. Compilation phase should have created it.",
+            binary_path
+        ));
+    }
+
+    // Start node briefly to get the current tip, then stop it
+    info!("=== Determining initial block height ===");
+    let additional_args = args.build_additional_args("baseline", args.baseline_args.as_ref());
+    let (mut node_process, _) = node_manager
+        .start_node(&binary_path, &args.baseline_ref, "baseline", &additional_args)
+        .await?;
+    let starting_tip = node_manager.wait_for_node_ready_and_get_tip(&mut node_process).await?;
+    info!("Node starting tip: {}", starting_tip);
+    node_manager.stop_node(&mut node_process).await?;
+
     // Run warmup phase before benchmarking (skip if warmup_blocks is 0)
     if args.get_warmup_blocks() > 0 {
         run_warmup_phase(
@@ -564,6 +646,7 @@ async fn run_benchmark_workflow(
             args,
             is_optimism,
             &baseline_commit,
+            starting_tip,
         )
         .await?;
     } else {
@@ -579,6 +662,10 @@ async fn run_benchmark_workflow(
         let commit = commits[i];
         info!("=== Processing {} reference: {} ===", ref_type, git_ref);
 
+        // Unwind to starting block minus benchmark blocks, so we end up back at starting_tip
+        let unwind_target = starting_tip.saturating_sub(args.blocks);
+        node_manager.unwind_to_block(unwind_target).await?;
+
         // Switch to target reference
         git_manager.switch_ref(git_ref)?;
 
@@ -612,17 +699,14 @@ async fn run_benchmark_workflow(
             node_manager.start_node(&binary_path, git_ref, ref_type, &additional_args).await?;
 
         // Wait for node to be ready and get its current tip (wherever it is)
-        let current_tip = node_manager.wait_for_node_ready_and_get_tip().await?;
+        let current_tip = node_manager.wait_for_node_ready_and_get_tip(&mut node_process).await?;
         info!("Node is ready at tip: {}", current_tip);
 
-        // Store the tip we'll unwind back to
-        let original_tip = current_tip;
-
         // Calculate benchmark range
         // Note: reth-bench has an off-by-one error where it consumes the first block
         // of the range, so we add 1 to compensate and get exactly args.blocks blocks
-        let from_block = original_tip;
-        let to_block = original_tip + args.blocks;
+        let from_block = current_tip;
+        let to_block = current_tip + args.blocks;
 
         // Run benchmark
         let output_dir = comparison_generator.get_ref_output_dir(ref_type);
@@ -639,9 +723,6 @@ async fn run_benchmark_workflow(
         // Stop node
         node_manager.stop_node(&mut node_process).await?;
 
-        // Unwind back to original tip
-        node_manager.unwind_to_block(original_tip).await?;
-
         // Store results for comparison
         comparison_generator.add_ref_results(ref_type, &output_dir)?;
 

bin/reth-bench-compare/src/comparison.rs

@@ -39,7 +39,8 @@ pub(crate) struct BenchmarkResults {
 #[derive(Debug, Clone, Deserialize, Serialize)]
 pub(crate) struct CombinedLatencyRow {
     pub block_number: u64,
-    pub transaction_count: u64,
+    #[serde(default)]
+    pub transaction_count: Option<u64>,
     pub gas_used: u64,
     pub new_payload_latency: u128,
 }
@@ -48,7 +49,8 @@ pub(crate) struct CombinedLatencyRow {
 #[derive(Debug, Clone, Deserialize, Serialize)]
 pub(crate) struct TotalGasRow {
     pub block_number: u64,
-    pub transaction_count: u64,
+    #[serde(default)]
+    pub transaction_count: Option<u64>,
     pub gas_used: u64,
     pub time: u128,
 }
@@ -97,6 +99,7 @@ pub(crate) struct RefInfo {
 /// Summary of the comparison between references.
 ///
 /// Percent deltas are `(feature - baseline) / baseline * 100`:
+/// - `new_payload_latency_mean_change_percent`: percent changes of the per-block means.
 /// - `new_payload_latency_p50_change_percent` / p90 / p99: percent changes of the respective
 ///   per-block percentiles.
 /// - `per_block_latency_change_mean_percent` / `per_block_latency_change_median_percent` are the
@@ -114,6 +117,7 @@ pub(crate) struct ComparisonSummary {
     pub per_block_latency_change_median_percent: f64,
     pub per_block_latency_change_std_dev_percent: f64,
     pub new_payload_total_latency_change_percent: f64,
+    pub new_payload_latency_mean_change_percent: f64,
     pub new_payload_latency_p50_change_percent: f64,
     pub new_payload_latency_p90_change_percent: f64,
     pub new_payload_latency_p99_change_percent: f64,
@@ -125,7 +129,8 @@ pub(crate) struct ComparisonSummary {
 #[derive(Debug, Serialize)]
 pub(crate) struct BlockComparison {
     pub block_number: u64,
-    pub transaction_count: u64,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub transaction_count: Option<u64>,
     pub gas_used: u64,
     pub baseline_new_payload_latency: u128,
     pub feature_new_payload_latency: u128,
@@ -442,6 +447,10 @@ impl ComparisonGenerator {
             per_block_latency_change_median_percent,
             per_block_latency_change_std_dev_percent,
             new_payload_total_latency_change_percent,
+            new_payload_latency_mean_change_percent: calc_percent_change(
+                baseline.mean_new_payload_latency_ms,
+                feature.mean_new_payload_latency_ms,
+            ),
             new_payload_latency_p50_change_percent: calc_percent_change(
                 baseline.median_new_payload_latency_ms,
                 feature.median_new_payload_latency_ms,
@@ -572,6 +581,10 @@ impl ComparisonGenerator {
             "  Total newPayload time change:                {:+.2}%",
             summary.new_payload_total_latency_change_percent
         );
+        println!(
+            "  NewPayload Latency mean:           {:+.2}%",
+            summary.new_payload_latency_mean_change_percent
+        );
         println!(
             "  NewPayload Latency p50:           {:+.2}%",
             summary.new_payload_latency_p50_change_percent

bin/reth-bench-compare/src/compilation.rs

@@ -13,7 +13,6 @@ pub(crate) struct CompilationManager {
     repo_root: String,
     output_dir: PathBuf,
     git_manager: GitManager,
-    features: String,
 }
 
 impl CompilationManager {
@@ -22,9 +21,8 @@ impl CompilationManager {
         repo_root: String,
         output_dir: PathBuf,
         git_manager: GitManager,
-        features: String,
     ) -> Result<Self> {
-        Ok(Self { repo_root, output_dir, git_manager, features })
+        Ok(Self { repo_root, output_dir, git_manager })
     }
 
     /// Detect if the RPC endpoint is an Optimism chain
@@ -68,7 +66,13 @@ impl CompilationManager {
     }
 
     /// Compile reth using cargo build and cache the binary
-    pub(crate) fn compile_reth(&self, commit: &str, is_optimism: bool) -> Result<()> {
+    pub(crate) fn compile_reth(
+        &self,
+        commit: &str,
+        is_optimism: bool,
+        features: &str,
+        rustflags: &str,
+    ) -> Result<()> {
         // Validate that current git commit matches the expected commit
         let current_commit = self.git_manager.get_current_commit()?;
         if current_commit != commit {
@@ -100,9 +104,8 @@ impl CompilationManager {
         let mut cmd = Command::new("cargo");
         cmd.arg("build").arg("--profile").arg("profiling");
 
-        // Add features
-        cmd.arg("--features").arg(&self.features);
-        info!("Using features: {}", self.features);
+        cmd.arg("--features").arg(features);
+        info!("Using features: {features}");
 
         // Add bin-specific arguments for optimism
         if is_optimism {
@@ -114,11 +117,11 @@ impl CompilationManager {
 
         cmd.current_dir(&self.repo_root);
 
-        // Set RUSTFLAGS for native CPU optimization
-        cmd.env("RUSTFLAGS", "-C target-cpu=native");
+        // Set RUSTFLAGS
+        cmd.env("RUSTFLAGS", rustflags);
+        info!("Using RUSTFLAGS: {rustflags}");
 
-        // Debug log the command
-        debug!("Executing cargo command: {:?}", cmd);
+        info!("Compiling {binary_name} with {cmd:?}");
 
         let output = cmd.output().wrap_err("Failed to execute cargo build command")?;
 
@@ -227,8 +230,7 @@ impl CompilationManager {
         let mut cmd = Command::new("cargo");
         cmd.args(["install", "--locked", "samply"]);
 
-        // Debug log the command
-        debug!("Executing cargo command: {:?}", cmd);
+        info!("Installing samply with {cmd:?}");
 
         let output = cmd.output().wrap_err("Failed to execute cargo install samply command")?;
 
@@ -303,8 +305,7 @@ impl CompilationManager {
         let mut cmd = Command::new("make");
         cmd.arg("install-reth-bench").current_dir(&self.repo_root);
 
-        // Debug log the command
-        debug!("Executing make command: {:?}", cmd);
+        info!("Compiling reth-bench with {cmd:?}");
 
         let output = cmd.output().wrap_err("Failed to execute make install-reth-bench command")?;
 

bin/reth-bench-compare/src/node.rs

@@ -2,7 +2,9 @@
 
 use crate::cli::Args;
 use alloy_provider::{Provider, ProviderBuilder};
+use alloy_rpc_client::RpcClient;
 use alloy_rpc_types_eth::SyncStatus;
+use alloy_transport_ws::WsConnect;
 use eyre::{eyre, OptionExt, Result, WrapErr};
 #[cfg(unix)]
 use nix::sys::signal::{killpg, Signal};
@@ -18,6 +20,9 @@ use tokio::{
 };
 use tracing::{debug, info, warn};
 
+/// Default websocket RPC port used by reth
+const DEFAULT_WS_RPC_PORT: u16 = 8546;
+
 /// Manages reth node lifecycle and operations
 pub(crate) struct NodeManager {
     datadir: Option<String>,
@@ -152,7 +157,10 @@ impl NodeManager {
             metrics_arg,
             "--http".to_string(),
             "--http.api".to_string(),
-            "eth".to_string(),
+            "eth,reth".to_string(),
+            "--ws".to_string(),
+            "--ws.api".to_string(),
+            "eth,reth".to_string(),
             "--disable-discovery".to_string(),
             "--trusted-only".to_string(),
         ]);
@@ -211,6 +219,11 @@ impl NodeManager {
         cmd.arg("--");
         cmd.args(reth_args);
 
+        // Enable tracing-samply
+        if supports_samply_flags(&reth_args[0]) {
+            cmd.arg("--log.samply");
+        }
+
         // Set environment variable to disable log styling
         cmd.env("RUST_LOG_STYLE", "never");
 
@@ -354,8 +367,13 @@ impl NodeManager {
         Ok((child, reth_command))
     }
 
-    /// Wait for the node to be ready and return its current tip
-    pub(crate) async fn wait_for_node_ready_and_get_tip(&self) -> Result<u64> {
+    /// Wait for the node to be ready and return its current tip.
+    ///
+    /// Fails early if the node process exits before becoming ready.
+    pub(crate) async fn wait_for_node_ready_and_get_tip(
+        &self,
+        child: &mut tokio::process::Child,
+    ) -> Result<u64> {
         info!("Waiting for node to be ready and synced...");
 
         let max_wait = Duration::from_secs(120); // 2 minutes to allow for sync
@@ -366,8 +384,23 @@ impl NodeManager {
         let url = rpc_url.parse().map_err(|e| eyre!("Invalid RPC URL '{}': {}", rpc_url, e))?;
         let provider = ProviderBuilder::new().connect_http(url);
 
+        let start_time = tokio::time::Instant::now();
+        let mut iteration = 0;
+
         timeout(max_wait, async {
             loop {
+                iteration += 1;
+                debug!(
+                    "Readiness check iteration {} (elapsed: {:?})",
+                    iteration,
+                    start_time.elapsed()
+                );
+
+                // Check if the node process has exited.
+                if let Some(status) = child.try_wait()? {
+                    return Err(eyre!("Node process exited unexpectedly with {status}"));
+                }
+
                 // First check if RPC is up and node is not syncing
                 match provider.syncing().await {
                     Ok(sync_result) => {
@@ -376,24 +409,48 @@ impl NodeManager {
                                 debug!("Node is still syncing {sync_info:?}, waiting...");
                             }
                             _ => {
+                                debug!("HTTP RPC is up and node is not syncing, checking block number...");
                                 // Node is not syncing, now get the tip
                                 match provider.get_block_number().await {
                                     Ok(tip) => {
-                                        info!("Node is ready and not syncing at block: {}", tip);
-                                        return Ok(tip);
+                                        debug!("HTTP RPC ready at block: {}, checking WebSocket...", tip);
+
+                                        // Verify WebSocket RPC is ready (public endpoint, no JWT required)
+                                        let ws_url = format!("ws://localhost:{}", DEFAULT_WS_RPC_PORT);
+                                        debug!("Attempting WebSocket connection to {} (public endpoint)", ws_url);
+                                        let ws_connect = WsConnect::new(&ws_url);
+
+                                        match RpcClient::connect_pubsub(ws_connect).await
+                                        {
+                                            Ok(_) => {
+                                                info!(
+                                                    "Node is ready (HTTP and WebSocket) at block: {} (took {:?}, {} iterations)",
+                                                    tip, start_time.elapsed(), iteration
+                                                );
+                                                return Ok(tip);
+                                            }
+                                            Err(e) => {
+                                                debug!(
+                                                    "HTTP RPC ready but WebSocket not ready yet (iteration {}): {:?}",
+                                                    iteration, e
+                                                );
+                                                debug!("WebSocket error details: {}", e);
+                                            }
+                                        }
                                     }
                                     Err(e) => {
-                                        debug!("Failed to get block number: {}", e);
+                                        debug!("Failed to get block number (iteration {}): {:?}", iteration, e);
                                     }
                                 }
                             }
                         }
                     }
                     Err(e) => {
-                        debug!("Node RPC not ready yet or failed to check sync status: {}", e);
+                        debug!("Node RPC not ready yet or failed to check sync status (iteration {}): {:?}", iteration, e);
                     }
                 }
 
+                debug!("Sleeping for {:?} before next check", check_interval);
                 sleep(check_interval).await;
             }
         })
@@ -403,7 +460,7 @@ impl NodeManager {
 
     /// Stop the reth node gracefully
     pub(crate) async fn stop_node(&self, child: &mut tokio::process::Child) -> Result<()> {
-        let pid = child.id().expect("Child process ID should be available");
+        let pid = child.id().ok_or_eyre("Child process ID should be available")?;
 
         // Check if the process has already exited
         match child.try_wait() {
@@ -552,3 +609,16 @@ impl NodeManager {
         Ok(())
     }
 }
+
+fn supports_samply_flags(bin: &str) -> bool {
+    let mut cmd = std::process::Command::new(bin);
+    // NOTE: The flag to check must come before --help.
+    // We pass --help as a shortcut to not execute any command.
+    cmd.args(["--log.samply", "--help"]);
+    debug!(?cmd, "Checking samply flags support");
+    let Ok(output) = cmd.output() else {
+        return false;
+    };
+    debug!(?output, "Samply flags support check");
+    output.status.success()
+}

bin/reth-bench/Cargo.toml

@@ -16,6 +16,7 @@ workspace = true
 # reth
 reth-cli-runner.workspace = true
 reth-cli-util.workspace = true
+reth-engine-primitives.workspace = true
 reth-fs-util.workspace = true
 reth-node-api.workspace = true
 reth-node-core.workspace = true
@@ -25,10 +26,11 @@ reth-tracing.workspace = true
 # alloy
 alloy-eips.workspace = true
 alloy-json-rpc.workspace = true
+alloy-network.workspace = true
 alloy-primitives.workspace = true
-alloy-provider = { workspace = true, features = ["engine-api", "reqwest-rustls-tls"], default-features = false }
+alloy-provider = { workspace = true, features = ["engine-api", "pubsub", "reqwest-rustls-tls"], default-features = false }
 alloy-pubsub.workspace = true
-alloy-rpc-client.workspace = true
+alloy-rpc-client = { workspace = true, features = ["pubsub"] }
 alloy-rpc-types-engine.workspace = true
 alloy-transport-http.workspace = true
 alloy-transport-ipc.workspace = true
@@ -50,6 +52,9 @@ tracing.workspace = true
 serde.workspace = true
 serde_json.workspace = true
 
+# url parsing
+url.workspace = true
+
 # async
 async-trait.workspace = true
 futures.workspace = true
@@ -58,6 +63,7 @@ tokio = { workspace = true, features = ["sync", "macros", "time", "rt-multi-thre
 # misc
 clap = { workspace = true, features = ["derive", "env"] }
 eyre.workspace = true
+color-eyre.workspace = true
 thiserror.workspace = true
 humantime.workspace = true
 

bin/reth-bench/README.md

@@ -31,6 +31,14 @@ Otherwise, running `make maxperf` at the root of the repo should be sufficient f
 `reth-bench` contains different commands to benchmark different patterns of engine API calls.
 The `reth-bench new-payload-fcu` command is the most representative of ethereum mainnet live sync, alternating between sending `engine_newPayload` calls and `engine_forkchoiceUpdated` calls.
 
+The `new-payload-fcu` command supports two optional waiting modes that can be used together or independently:
+- `--wait-time <duration>`: Fixed sleep interval between blocks (e.g., `--wait-time 100ms`)
+- `--wait-for-persistence`: Waits for blocks to be persisted using the `reth_subscribePersistedBlock` subscription
+
+When using `--wait-for-persistence`, the benchmark waits after every `(threshold + 1)` blocks, where the threshold defaults to the engine's persistence threshold (2). This can be customized with `--persistence-threshold <N>`.
+
+By default, the WebSocket URL for persistence subscriptions is derived from `--engine-rpc-url` (converting to ws:// on port 8546). Use `--ws-rpc-url` to override this.
+
 Below is an overview of how to run a benchmark:
 
 ### Setup

bin/reth-bench/src/authenticated_transport.rs

@@ -163,7 +163,7 @@ impl AuthenticatedTransport {
 
             // shift the iat forward by one second so there is some buffer time
             let mut shifted_claims = inner_and_claims.1;
-            shifted_claims.iat -= 1;
+            shifted_claims.iat -= 30;
 
             // if the claims are out of date, reset the inner transport
             if !shifted_claims.is_within_time_window() {

bin/reth-bench/src/bench/context.rs

@@ -103,14 +103,20 @@ impl BenchContext {
             (bench_args.from, bench_args.to)
         };
 
-        // If neither `--from` nor `--to` are provided, we will run the benchmark continuously,
+        // If `--to` are not provided, we will run the benchmark continuously,
         // starting at the latest block.
-        let mut benchmark_mode = BenchMode::new(from, to)?;
+        let latest_block = block_provider
+            .get_block_by_number(BlockNumberOrTag::Latest)
+            .full()
+            .await?
+            .ok_or_else(|| eyre::eyre!("Failed to fetch latest block from RPC"))?;
+        let mut benchmark_mode = BenchMode::new(from, to, latest_block.into_inner().number())?;
 
         let first_block = match benchmark_mode {
-            BenchMode::Continuous => {
-                // fetch Latest block
-                block_provider.get_block_by_number(BlockNumberOrTag::Latest).full().await?.unwrap()
+            BenchMode::Continuous(start) => {
+                block_provider.get_block_by_number(start.into()).full().await?.ok_or_else(|| {
+                    eyre::eyre!("Failed to fetch block {} from RPC for continuous mode", start)
+                })?
             }
             BenchMode::Range(ref mut range) => {
                 match range.next() {
@@ -120,7 +126,9 @@ impl BenchContext {
                             .get_block_by_number(block_number.into())
                             .full()
                             .await?
-                            .unwrap()
+                            .ok_or_else(|| {
+                                eyre::eyre!("Failed to fetch block {} from RPC", block_number)
+                            })?
                     }
                     None => {
                         return Err(eyre::eyre!(

bin/reth-bench/src/bench/new_payload_fcu.rs

@@ -1,5 +1,13 @@
 //! Runs the `reth bench` command, calling first newPayload for each block, then calling
 //! forkchoiceUpdated.
+//!
+//! Supports configurable waiting behavior:
+//! - **`--wait-time`**: Fixed sleep interval between blocks.
+//! - **`--wait-for-persistence`**: Waits for every Nth block to be persisted using the
+//!   `reth_subscribePersistedBlock` subscription, where N matches the engine's persistence
+//!   threshold. This ensures the benchmark doesn't outpace persistence.
+//!
+//! Both options can be used together or independently.
 
 use crate::{
     bench::{
@@ -11,16 +19,26 @@ use crate::{
     },
     valid_payload::{block_to_new_payload, call_forkchoice_updated, call_new_payload},
 };
-use alloy_provider::Provider;
+use alloy_eips::BlockNumHash;
+use alloy_network::Ethereum;
+use alloy_provider::{Provider, RootProvider};
+use alloy_pubsub::SubscriptionStream;
+use alloy_rpc_client::RpcClient;
 use alloy_rpc_types_engine::ForkchoiceState;
+use alloy_transport_ws::WsConnect;
 use clap::Parser;
 use csv::Writer;
 use eyre::{Context, OptionExt};
+use futures::StreamExt;
 use humantime::parse_duration;
 use reth_cli_runner::CliContext;
+use reth_engine_primitives::config::DEFAULT_PERSISTENCE_THRESHOLD;
 use reth_node_core::args::BenchmarkArgs;
 use std::time::{Duration, Instant};
 use tracing::{debug, info};
+use url::Url;
+
+const PERSISTENCE_CHECKPOINT_TIMEOUT: Duration = Duration::from_secs(60);
 
 /// `reth benchmark new-payload-fcu` command
 #[derive(Debug, Parser)]
@@ -30,8 +48,31 @@ pub struct Command {
     rpc_url: String,
 
     /// How long to wait after a forkchoice update before sending the next payload.
-    #[arg(long, value_name = "WAIT_TIME", value_parser = parse_duration, default_value = "250ms", verbatim_doc_comment)]
-    wait_time: Duration,
+    #[arg(long, value_name = "WAIT_TIME", value_parser = parse_duration, verbatim_doc_comment)]
+    wait_time: Option<Duration>,
+
+    /// Wait for blocks to be persisted before sending the next batch.
+    ///
+    /// When enabled, waits for every Nth block to be persisted using the
+    /// `reth_subscribePersistedBlock` subscription. This ensures the benchmark
+    /// doesn't outpace persistence.
+    ///
+    /// The subscription uses the regular RPC websocket endpoint (no JWT required).
+    #[arg(long, default_value = "false", verbatim_doc_comment)]
+    wait_for_persistence: bool,
+
+    /// Engine persistence threshold used for deciding when to wait for persistence.
+    ///
+    /// The benchmark waits after every `(threshold + 1)` blocks. By default this
+    /// matches the engine's `DEFAULT_PERSISTENCE_THRESHOLD` (2), so waits occur
+    /// at blocks 3, 6, 9, etc.
+    #[arg(
+        long = "persistence-threshold",
+        value_name = "PERSISTENCE_THRESHOLD",
+        default_value_t = DEFAULT_PERSISTENCE_THRESHOLD,
+        verbatim_doc_comment
+    )]
+    persistence_threshold: u64,
 
     /// The size of the block buffer (channel capacity) for prefetching blocks from the RPC
     /// endpoint.
@@ -50,6 +91,32 @@ pub struct Command {
 impl Command {
     /// Execute `benchmark new-payload-fcu` command
     pub async fn execute(self, _ctx: CliContext) -> eyre::Result<()> {
+        // Log mode configuration
+        if let Some(duration) = self.wait_time {
+            info!("Using wait-time mode with {}ms delay between blocks", duration.as_millis());
+        }
+        if self.wait_for_persistence {
+            info!(
+                "Persistence waiting enabled (waits after every {} blocks to match engine gap > {} behavior)",
+                self.persistence_threshold + 1,
+                self.persistence_threshold
+            );
+        }
+
+        // Set up waiter based on configured options (duration takes precedence)
+        let mut waiter = match (self.wait_time, self.wait_for_persistence) {
+            (Some(duration), _) => Some(PersistenceWaiter::with_duration(duration)),
+            (None, true) => {
+                let sub = self.setup_persistence_subscription().await?;
+                Some(PersistenceWaiter::with_subscription(
+                    sub,
+                    self.persistence_threshold,
+                    PERSISTENCE_CHECKPOINT_TIMEOUT,
+                ))
+            }
+            (None, false) => None,
+        };
+
         let BenchContext {
             benchmark_mode,
             block_provider,
@@ -110,7 +177,6 @@ impl Command {
             }
         });
 
-        // put results in a summary vec so they can be printed at the end
         let mut results = Vec::new();
         let total_benchmark_duration = Instant::now();
         let mut total_wait_time = Duration::ZERO;
@@ -121,14 +187,12 @@ impl Command {
             total_wait_time += wait_start.elapsed();
             result
         } {
-            // just put gas used here
             let gas_used = block.header.gas_used;
             let block_number = block.header.number;
             let transaction_count = block.transactions.len() as u64;
 
-            debug!(target: "reth-bench", ?block_number, "Sending payload",);
+            debug!(target: "reth-bench", ?block_number, "Sending payload");
 
-            // construct fcu to call
             let forkchoice_state = ForkchoiceState {
                 head_block_hash: head,
                 safe_block_hash: safe,
@@ -143,7 +207,6 @@ impl Command {
 
             call_forkchoice_updated(&auth_provider, version, forkchoice_state, None).await?;
 
-            // calculate the total duration and the fcu latency, record
             let total_latency = start.elapsed();
             let fcu_latency = total_latency - new_payload_result.latency;
             let combined_result = CombinedResult {
@@ -154,17 +217,15 @@ impl Command {
                 total_latency,
             };
 
-            // current duration since the start of the benchmark minus the time
-            // waiting for blocks
+            // Exclude time spent waiting on the block prefetch channel from the benchmark duration.
+            // We want to measure engine throughput, not RPC fetch latency.
             let current_duration = total_benchmark_duration.elapsed() - total_wait_time;
-
-            // convert gas used to gigagas, then compute gigagas per second
             info!(%combined_result);
 
-            // wait before sending the next payload
-            tokio::time::sleep(self.wait_time).await;
+            if let Some(w) = &mut waiter {
+                w.on_block(block_number).await?;
+            }
 
-            // record the current result
             let gas_row =
                 TotalGasRow { block_number, transaction_count, gas_used, time: current_duration };
             results.push((gas_row, combined_result));
@@ -175,24 +236,26 @@ impl Command {
             return Err(error);
         }
 
+        // Drop waiter - we don't need to wait for final blocks to persist
+        // since the benchmark goal is measuring Ggas/s of newPayload/FCU, not persistence.
+        drop(waiter);
+
         let (gas_output_results, combined_results): (_, Vec<CombinedResult>) =
             results.into_iter().unzip();
 
-        // write the csv output to files
-        if let Some(path) = self.benchmark.output {
-            // first write the combined results to a file
+        // Write CSV output files
+        if let Some(ref path) = self.benchmark.output {
             let output_path = path.join(COMBINED_OUTPUT_SUFFIX);
             info!("Writing engine api call latency output to file: {:?}", output_path);
-            let mut writer = Writer::from_path(output_path)?;
+            let mut writer = Writer::from_path(&output_path)?;
             for result in combined_results {
                 writer.serialize(result)?;
             }
             writer.flush()?;
 
-            // now write the gas output to a file
             let output_path = path.join(GAS_OUTPUT_SUFFIX);
             info!("Writing total gas output to file: {:?}", output_path);
-            let mut writer = Writer::from_path(output_path)?;
+            let mut writer = Writer::from_path(&output_path)?;
             for row in &gas_output_results {
                 writer.serialize(row)?;
             }
@@ -201,8 +264,8 @@ impl Command {
             info!("Finished writing benchmark output files to {:?}.", path);
         }
 
-        // accumulate the results and calculate the overall Ggas/s
         let gas_output = TotalGasOutput::new(gas_output_results)?;
+
         info!(
             total_duration=?gas_output.total_duration,
             total_gas_used=?gas_output.total_gas_used,
@@ -213,4 +276,278 @@ impl Command {
 
         Ok(())
     }
+
+    /// Returns the websocket RPC URL used for the persistence subscription.
+    ///
+    /// Preference:
+    /// - If `--ws-rpc-url` is provided, use it directly.
+    /// - Otherwise, derive a WS RPC URL from `--engine-rpc-url`.
+    ///
+    /// The persistence subscription endpoint (`reth_subscribePersistedBlock`) is exposed on
+    /// the regular RPC server (WS port, usually 8546), not on the engine API port (usually 8551).
+    /// Since `BenchmarkArgs` only has the engine URL by default, we convert the scheme
+    /// (http→ws, https→wss) and force the port to 8546.
+    fn derive_ws_rpc_url(&self) -> eyre::Result<Url> {
+        if let Some(ref ws_url) = self.benchmark.ws_rpc_url {
+            let parsed: Url = ws_url
+                .parse()
+                .wrap_err_with(|| format!("Failed to parse WebSocket RPC URL: {ws_url}"))?;
+            info!(target: "reth-bench", ws_url = %parsed, "Using provided WebSocket RPC URL");
+            Ok(parsed)
+        } else {
+            let derived = engine_url_to_ws_url(&self.benchmark.engine_rpc_url)?;
+            debug!(
+                target: "reth-bench",
+                engine_url = %self.benchmark.engine_rpc_url,
+                %derived,
+                "Derived WebSocket RPC URL from engine RPC URL"
+            );
+            Ok(derived)
+        }
+    }
+
+    /// Establishes a websocket connection and subscribes to `reth_subscribePersistedBlock`.
+    async fn setup_persistence_subscription(&self) -> eyre::Result<PersistenceSubscription> {
+        let ws_url = self.derive_ws_rpc_url()?;
+
+        info!("Connecting to WebSocket at {} for persistence subscription", ws_url);
+
+        let ws_connect = WsConnect::new(ws_url.to_string());
+        let client = RpcClient::connect_pubsub(ws_connect)
+            .await
+            .wrap_err("Failed to connect to WebSocket RPC endpoint")?;
+        let provider: RootProvider<Ethereum> = RootProvider::new(client);
+
+        let subscription = provider
+            .subscribe_to::<BlockNumHash>("reth_subscribePersistedBlock")
+            .await
+            .wrap_err("Failed to subscribe to persistence notifications")?;
+
+        info!("Subscribed to persistence notifications");
+
+        Ok(PersistenceSubscription::new(provider, subscription.into_stream()))
+    }
+}
+
+/// Converts an engine API URL to the default RPC websocket URL.
+///
+/// Transformations:
+/// - `http`  → `ws`
+/// - `https` → `wss`
+/// - `ws` / `wss` keep their scheme
+/// - Port is always set to `8546`, reth's default RPC websocket port.
+///
+/// This is used when we only know the engine API URL (typically `:8551`) but
+/// need to connect to the node's WS RPC endpoint for persistence events.
+fn engine_url_to_ws_url(engine_url: &str) -> eyre::Result<Url> {
+    let url: Url = engine_url
+        .parse()
+        .wrap_err_with(|| format!("Failed to parse engine RPC URL: {engine_url}"))?;
+
+    let mut ws_url = url.clone();
+
+    match ws_url.scheme() {
+        "http" => ws_url
+            .set_scheme("ws")
+            .map_err(|_| eyre::eyre!("Failed to set WS scheme for URL: {url}"))?,
+        "https" => ws_url
+            .set_scheme("wss")
+            .map_err(|_| eyre::eyre!("Failed to set WSS scheme for URL: {url}"))?,
+        "ws" | "wss" => {}
+        scheme => {
+            return Err(eyre::eyre!(
+            "Unsupported URL scheme '{scheme}' for URL: {url}. Expected http, https, ws, or wss."
+        ))
+        }
+    }
+
+    ws_url.set_port(Some(8546)).map_err(|_| eyre::eyre!("Failed to set port for URL: {url}"))?;
+
+    Ok(ws_url)
+}
+
+/// Waits until the persistence subscription reports that `target` has been persisted.
+///
+/// Consumes subscription events until `last_persisted >= target`, or returns an error if:
+/// - the subscription stream ends unexpectedly, or
+/// - `timeout` elapses before `target` is observed.
+async fn wait_for_persistence(
+    stream: &mut SubscriptionStream<BlockNumHash>,
+    target: u64,
+    last_persisted: &mut u64,
+    timeout: Duration,
+) -> eyre::Result<()> {
+    tokio::time::timeout(timeout, async {
+        while *last_persisted < target {
+            match stream.next().await {
+                Some(persisted) => {
+                    *last_persisted = persisted.number;
+                    debug!(
+                        target: "reth-bench",
+                        persisted_block = ?last_persisted,
+                        "Received persistence notification"
+                    );
+                }
+                None => {
+                    return Err(eyre::eyre!("Persistence subscription closed unexpectedly"));
+                }
+            }
+        }
+        Ok(())
+    })
+    .await
+    .map_err(|_| {
+        eyre::eyre!(
+            "Persistence timeout: target block {} not persisted within {:?}. Last persisted: {}",
+            target,
+            timeout,
+            last_persisted
+        )
+    })?
+}
+
+/// Wrapper that keeps both the subscription stream and the underlying provider alive.
+/// The provider must be kept alive for the subscription to continue receiving events.
+struct PersistenceSubscription {
+    _provider: RootProvider<Ethereum>,
+    stream: SubscriptionStream<BlockNumHash>,
+}
+
+impl PersistenceSubscription {
+    const fn new(
+        provider: RootProvider<Ethereum>,
+        stream: SubscriptionStream<BlockNumHash>,
+    ) -> Self {
+        Self { _provider: provider, stream }
+    }
+
+    const fn stream_mut(&mut self) -> &mut SubscriptionStream<BlockNumHash> {
+        &mut self.stream
+    }
+}
+
+/// Encapsulates the block waiting logic.
+///
+/// Provides a simple `on_block()` interface that handles both:
+/// - Fixed duration waits (when `wait_time` is set)
+/// - Persistence-based waits (when `subscription` is set)
+///
+/// For persistence mode, waits after every `(threshold + 1)` blocks.
+struct PersistenceWaiter {
+    wait_time: Option<Duration>,
+    subscription: Option<PersistenceSubscription>,
+    blocks_sent: u64,
+    last_persisted: u64,
+    threshold: u64,
+    timeout: Duration,
+}
+
+impl PersistenceWaiter {
+    const fn with_duration(wait_time: Duration) -> Self {
+        Self {
+            wait_time: Some(wait_time),
+            subscription: None,
+            blocks_sent: 0,
+            last_persisted: 0,
+            threshold: 0,
+            timeout: Duration::ZERO,
+        }
+    }
+
+    const fn with_subscription(
+        subscription: PersistenceSubscription,
+        threshold: u64,
+        timeout: Duration,
+    ) -> Self {
+        Self {
+            wait_time: None,
+            subscription: Some(subscription),
+            blocks_sent: 0,
+            last_persisted: 0,
+            threshold,
+            timeout,
+        }
+    }
+
+    /// Called once per block. Waits based on the configured mode.
+    #[allow(clippy::manual_is_multiple_of)]
+    async fn on_block(&mut self, block_number: u64) -> eyre::Result<()> {
+        if let Some(wait_time) = self.wait_time {
+            tokio::time::sleep(wait_time).await;
+            return Ok(());
+        }
+
+        let Some(ref mut subscription) = self.subscription else {
+            return Ok(());
+        };
+
+        self.blocks_sent += 1;
+
+        if self.blocks_sent % (self.threshold + 1) == 0 {
+            debug!(
+                target: "reth-bench",
+                target_block = ?block_number,
+                last_persisted = self.last_persisted,
+                blocks_sent = self.blocks_sent,
+                "Waiting for persistence"
+            );
+
+            wait_for_persistence(
+                subscription.stream_mut(),
+                block_number,
+                &mut self.last_persisted,
+                self.timeout,
+            )
+            .await?;
+
+            debug!(
+                target: "reth-bench",
+                persisted = self.last_persisted,
+                "Persistence caught up"
+            );
+        }
+
+        Ok(())
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_engine_url_to_ws_url() {
+        // http -> ws, always uses port 8546
+        let result = engine_url_to_ws_url("http://localhost:8551").unwrap();
+        assert_eq!(result.as_str(), "ws://localhost:8546/");
+
+        // https -> wss
+        let result = engine_url_to_ws_url("https://localhost:8551").unwrap();
+        assert_eq!(result.as_str(), "wss://localhost:8546/");
+
+        // Custom engine port still maps to 8546
+        let result = engine_url_to_ws_url("http://localhost:9551").unwrap();
+        assert_eq!(result.port(), Some(8546));
+
+        // Already ws passthrough
+        let result = engine_url_to_ws_url("ws://localhost:8546").unwrap();
+        assert_eq!(result.scheme(), "ws");
+
+        // Invalid inputs
+        assert!(engine_url_to_ws_url("ftp://localhost:8551").is_err());
+        assert!(engine_url_to_ws_url("not a valid url").is_err());
+    }
+
+    #[tokio::test]
+    async fn test_waiter_with_duration() {
+        let mut waiter = PersistenceWaiter::with_duration(Duration::from_millis(1));
+
+        let start = Instant::now();
+        waiter.on_block(1).await.unwrap();
+        waiter.on_block(2).await.unwrap();
+        waiter.on_block(3).await.unwrap();
+
+        // Should have waited ~3ms total
+        assert!(start.elapsed() >= Duration::from_millis(3));
+    }
 }

bin/reth-bench/src/bench_mode.rs

@@ -5,9 +5,8 @@ use std::ops::RangeInclusive;
 /// Whether or not the benchmark should run as a continuous stream of payloads.
 #[derive(Debug, PartialEq, Eq)]
 pub enum BenchMode {
-    // TODO: just include the start block in `Continuous`
     /// Run the benchmark as a continuous stream of payloads, until the benchmark is interrupted.
-    Continuous,
+    Continuous(u64),
     /// Run the benchmark for a specific range of blocks.
     Range(RangeInclusive<u64>),
 }
@@ -16,18 +15,19 @@ impl BenchMode {
     /// Check if the block number is in the range
     pub fn contains(&self, block_number: u64) -> bool {
         match self {
-            Self::Continuous => true,
+            Self::Continuous(start) => block_number >= *start,
             Self::Range(range) => range.contains(&block_number),
         }
     }
 
     /// Create a [`BenchMode`] from optional `from` and `to` fields.
-    pub fn new(from: Option<u64>, to: Option<u64>) -> Result<Self, eyre::Error> {
+    pub fn new(from: Option<u64>, to: Option<u64>, latest_block: u64) -> Result<Self, eyre::Error> {
         // If neither `--from` nor `--to` are provided, we will run the benchmark continuously,
         // starting at the latest block.
         match (from, to) {
             (Some(from), Some(to)) => Ok(Self::Range(from..=to)),
-            (None, None) => Ok(Self::Continuous),
+            (None, None) => Ok(Self::Continuous(latest_block)),
+            (Some(start), None) => Ok(Self::Continuous(start)),
             _ => {
                 // both or neither are allowed, everything else is ambiguous
                 Err(eyre::eyre!("`from` and `to` must be provided together, or not at all."))

bin/reth-bench/src/main.rs

@@ -23,7 +23,7 @@ use bench::BenchmarkCommand;
 use clap::Parser;
 use reth_cli_runner::CliRunner;
 
-fn main() {
+fn main() -> eyre::Result<()> {
     // Enable backtraces unless a RUST_BACKTRACE value has already been explicitly provided.
     if std::env::var_os("RUST_BACKTRACE").is_none() {
         unsafe {
@@ -31,12 +31,11 @@ fn main() {
         }
     }
 
+    color_eyre::install()?;
+
     // Run until either exit or sigint or sigterm
-    let runner = CliRunner::try_default_runtime().unwrap();
-    runner
-        .run_command_until_exit(|ctx| {
-            let command = BenchmarkCommand::parse();
-            command.execute(ctx)
-        })
-        .unwrap();
+    let runner = CliRunner::try_default_runtime()?;
+    runner.run_command_until_exit(|ctx| BenchmarkCommand::parse().execute(ctx))?;
+
+    Ok(())
 }

bin/reth/Cargo.toml

@@ -81,7 +81,7 @@ backon.workspace = true
 tempfile.workspace = true
 
 [features]
-default = ["jemalloc", "otlp", "reth-revm/portable", "js-tracer", "keccak-cache-global"]
+default = ["jemalloc", "otlp", "reth-revm/portable", "js-tracer", "keccak-cache-global", "asm-keccak"]
 
 otlp = [
     "reth-ethereum-cli/otlp",
@@ -103,6 +103,7 @@ asm-keccak = [
     "reth-node-ethereum/asm-keccak",
 ]
 keccak-cache-global = [
+    "reth-node-core/keccak-cache-global",
     "reth-node-ethereum/keccak-cache-global",
 ]
 jemalloc = [
@@ -115,6 +116,11 @@ jemalloc-prof = [
     "reth-cli-util/jemalloc",
     "reth-cli-util/jemalloc-prof",
     "reth-ethereum-cli/jemalloc-prof",
+    "reth-node-metrics/jemalloc-prof",
+]
+jemalloc-symbols = [
+    "jemalloc-prof",
+    "reth-ethereum-cli/jemalloc-symbols",
 ]
 jemalloc-unprefixed = [
     "reth-cli-util/jemalloc-unprefixed",
@@ -165,6 +171,8 @@ min-trace-logs = [
     "reth-node-core/min-trace-logs",
 ]
 
+edge = ["reth-ethereum-cli/edge"]
+
 [[bin]]
 name = "reth"
 path = "src/main.rs"

bin/reth/src/lib.rs

@@ -2,22 +2,46 @@
 //!
 //! ## Feature Flags
 //!
+//! ### Default Features
+//!
 //! - `jemalloc`: Uses [jemallocator](https://github.com/tikv/jemallocator) as the global allocator.
 //!   This is **not recommended on Windows**. See [here](https://rust-lang.github.io/rfcs/1974-global-allocators.html#jemalloc)
 //!   for more info.
+//! - `otlp`: Enables [OpenTelemetry](https://opentelemetry.io/) metrics export to a configured OTLP
+//!   collector endpoint.
+//! - `js-tracer`: Enables the `JavaScript` tracer for the `debug_trace` endpoints, allowing custom
+//!   `JavaScript`-based transaction tracing.
+//! - `keccak-cache-global`: Enables global caching for Keccak256 hashes to improve performance.
+//! - `asm-keccak`: Replaces the default, pure-Rust implementation of Keccak256 with one implemented
+//!   in assembly; see [the `keccak-asm` crate](https://github.com/DaniPopes/keccak-asm) for more
+//!   details and supported targets.
+//!
+//! ### Allocator Features
+//!
 //! - `jemalloc-prof`: Enables [jemallocator's](https://github.com/tikv/jemallocator) heap profiling
 //!   and leak detection functionality. See [jemalloc's opt.prof](https://jemalloc.net/jemalloc.3.html#opt.prof)
-//!   documentation for usage details. This is **not recommended on Windows**. See [here](https://rust-lang.github.io/rfcs/1974-global-allocators.html#jemalloc)
-//!   for more info.
-//! - `asm-keccak`: replaces the default, pure-Rust implementation of Keccak256 with one implemented
-//!   in assembly; see [the `keccak-asm` crate](https://github.com/DaniPopes/keccak-asm) for more
-//!   details and supported targets
+//!   documentation for usage details. This is **not recommended on Windows**.
+//! - `jemalloc-symbols`: Enables jemalloc symbols for profiling. Includes `jemalloc-prof`.
+//! - `jemalloc-unprefixed`: Uses unprefixed jemalloc symbols.
+//! - `tracy-allocator`: Enables [Tracy](https://github.com/wolfpld/tracy) profiler allocator
+//!   integration for memory profiling.
+//! - `snmalloc`: Uses [snmalloc](https://github.com/snmalloc/snmalloc) as the global allocator. Use
+//!   `--no-default-features` when enabling this, as jemalloc takes precedence.
+//! - `snmalloc-native`: Uses snmalloc with native CPU optimizations. Use `--no-default-features`
+//!   when enabling this.
+//!
+//! ### Log Level Features
+//!
 //! - `min-error-logs`: Disables all logs below `error` level.
 //! - `min-warn-logs`: Disables all logs below `warn` level.
 //! - `min-info-logs`: Disables all logs below `info` level. This can speed up the node, since fewer
 //!   calls to the logging component are made.
 //! - `min-debug-logs`: Disables all logs below `debug` level.
 //! - `min-trace-logs`: Disables all logs below `trace` level.
+//!
+//! ### Development Features
+//!
+//! - `dev`: Enables development mode features, including test vector generation commands.
 
 #![doc(
     html_logo_url = "https://raw.githubusercontent.com/paradigmxyz/reth/main/assets/reth-docs.png",
@@ -170,7 +194,7 @@ pub mod rpc {
         pub use reth_rpc::eth::*;
     }
 
-    /// Re-exported from `reth_rpc::rpc`.
+    /// Re-exported from `reth_rpc_server_types::result`.
     pub mod result {
         pub use reth_rpc_server_types::result::*;
     }

bin/reth/src/main.rs

@@ -3,6 +3,10 @@
 #[global_allocator]
 static ALLOC: reth_cli_util::allocator::Allocator = reth_cli_util::allocator::new_allocator();
 
+#[cfg(all(feature = "jemalloc-prof", unix))]
+#[unsafe(export_name = "_rjem_malloc_conf")]
+static MALLOC_CONF: &[u8] = b"prof:true,prof_active:true,lg_prof_sample:19\0";
+
 use clap::Parser;
 use reth::{args::RessArgs, cli::Cli, ress::install_ress_subprotocol};
 use reth_ethereum_cli::chainspec::EthereumChainSpecParser;

crates/chain-state/src/chain_info.rs

@@ -33,6 +33,7 @@ where
     ) -> Self {
         let (finalized_block, _) = watch::channel(finalized);
         let (safe_block, _) = watch::channel(safe);
+        let (persisted_block, _) = watch::channel(None);
 
         Self {
             inner: Arc::new(ChainInfoInner {
@@ -42,6 +43,7 @@ where
                 canonical_head: RwLock::new(head),
                 safe_block,
                 finalized_block,
+                persisted_block,
             }),
         }
     }
@@ -97,6 +99,11 @@ where
         self.inner.finalized_block.borrow().as_ref().map(SealedHeader::num_hash)
     }
 
+    /// Returns the `BlockNumHash` of the persisted block.
+    pub fn get_persisted_num_hash(&self) -> Option<BlockNumHash> {
+        *self.inner.persisted_block.borrow()
+    }
+
     /// Sets the canonical head of the chain.
     pub fn set_canonical_head(&self, header: SealedHeader<N::BlockHeader>) {
         let number = header.number();
@@ -130,6 +137,18 @@ where
         });
     }
 
+    /// Sets the persisted block of the chain.
+    pub fn set_persisted(&self, num_hash: BlockNumHash) {
+        self.inner.persisted_block.send_if_modified(|current| {
+            if current.map(|b| b.hash) != Some(num_hash.hash) {
+                let _ = current.replace(num_hash);
+                return true
+            }
+
+            false
+        });
+    }
+
     /// Subscribe to the finalized block.
     pub fn subscribe_finalized_block(
         &self,
@@ -141,6 +160,11 @@ where
     pub fn subscribe_safe_block(&self) -> watch::Receiver<Option<SealedHeader<N::BlockHeader>>> {
         self.inner.safe_block.subscribe()
     }
+
+    /// Subscribe to the persisted block.
+    pub fn subscribe_persisted_block(&self) -> watch::Receiver<Option<BlockNumHash>> {
+        self.inner.persisted_block.subscribe()
+    }
 }
 
 /// Container type for all chain info fields
@@ -159,11 +183,14 @@ struct ChainInfoInner<N: NodePrimitives = reth_ethereum_primitives::EthPrimitive
     safe_block: watch::Sender<Option<SealedHeader<N::BlockHeader>>>,
     /// The block that the beacon node considers finalized.
     finalized_block: watch::Sender<Option<SealedHeader<N::BlockHeader>>>,
+    /// The last block that was persisted to disk.
+    persisted_block: watch::Sender<Option<BlockNumHash>>,
 }
 
 #[cfg(test)]
 mod tests {
     use super::*;
+    use alloy_primitives::B256;
     use reth_ethereum_primitives::EthPrimitives;
     use reth_testing_utils::{generators, generators::random_header};
 
@@ -338,4 +365,28 @@ mod tests {
         // Assert that the BlockNumHash returned matches the safe header
         assert_eq!(tracker.get_safe_num_hash(), Some(safe_header.num_hash()));
     }
+
+    #[test]
+    fn test_set_persisted() {
+        let mut rng = generators::rng();
+        let header = random_header(&mut rng, 10, None);
+        let tracker: ChainInfoTracker<EthPrimitives> = ChainInfoTracker::new(header, None, None);
+
+        // Initial state: persisted block should be None
+        assert!(tracker.get_persisted_num_hash().is_none());
+
+        // Set a persisted block
+        let num_hash1 = BlockNumHash::new(10, B256::random());
+        tracker.set_persisted(num_hash1);
+        assert_eq!(tracker.get_persisted_num_hash(), Some(num_hash1));
+
+        // Setting the same block again should not change anything
+        tracker.set_persisted(num_hash1);
+        assert_eq!(tracker.get_persisted_num_hash(), Some(num_hash1));
+
+        // Set a different block
+        let num_hash2 = BlockNumHash::new(20, B256::random());
+        tracker.set_persisted(num_hash2);
+        assert_eq!(tracker.get_persisted_num_hash(), Some(num_hash2));
+    }
 }

crates/chain-state/src/deferred_trie.rs

@@ -37,12 +37,19 @@ pub struct ComputedTrieData {
 
 /// Trie input bundled with its anchor hash.
 ///
-/// This is used to store the trie input and anchor hash for a block together.
+/// The `trie_input` contains the **cumulative** overlay of all in-memory ancestor blocks,
+/// not just this block's changes. Child blocks reuse the parent's overlay in O(1) by
+/// cloning the Arc-wrapped data.
+///
+/// The `anchor_hash` is metadata indicating which persisted base state this overlay
+/// sits on top of. It is CRITICAL for overlay reuse decisions: an overlay built on top
+/// of Anchor A cannot be reused for a block anchored to Anchor B, as it would result
+/// in an incorrect state.
 #[derive(Clone, Debug)]
 pub struct AnchoredTrieInput {
     /// The persisted ancestor hash this trie input is anchored to.
     pub anchor_hash: B256,
-    /// Trie input constructed from in-memory overlays.
+    /// Cumulative trie input overlay from all in-memory ancestors.
     pub trie_input: Arc<TrieInputSorted>,
 }
 
@@ -62,7 +69,8 @@ static DEFERRED_TRIE_METRICS: LazyLock<DeferredTrieMetrics> =
 /// Internal state for deferred trie data.
 enum DeferredState {
     /// Data is not yet available; raw inputs stored for fallback computation.
-    Pending(PendingInputs),
+    /// Wrapped in `Option` to allow taking ownership during computation.
+    Pending(Option<PendingInputs>),
     /// Data has been computed and is ready.
     Ready(ComputedTrieData),
 }
@@ -112,12 +120,12 @@ impl DeferredTrieData {
         ancestors: Vec<Self>,
     ) -> Self {
         Self {
-            state: Arc::new(Mutex::new(DeferredState::Pending(PendingInputs {
+            state: Arc::new(Mutex::new(DeferredState::Pending(Some(PendingInputs {
                 hashed_state,
                 trie_updates,
                 anchor_hash,
                 ancestors,
-            }))),
+            })))),
         }
     }
 
@@ -138,8 +146,9 @@ impl DeferredTrieData {
     ///
     /// # Process
     /// 1. Sort the current block's hashed state and trie updates
-    /// 2. Merge ancestor overlays (oldest -> newest, so later state takes precedence)
-    /// 3. Extend the merged overlay with this block's sorted data
+    /// 2. Reuse parent's cached overlay if available (O(1) - the common case)
+    /// 3. Otherwise, rebuild overlay from ancestors (rare fallback)
+    /// 4. Extend the overlay with this block's sorted data
     ///
     /// Used by both the async background task and the synchronous fallback path.
     ///
@@ -147,49 +156,103 @@ impl DeferredTrieData {
     /// * `hashed_state` - Unsorted hashed post-state (account/storage changes) from execution
     /// * `trie_updates` - Unsorted trie node updates from state root computation
     /// * `anchor_hash` - The persisted ancestor hash this trie input is anchored to
-    /// * `ancestors` - Deferred trie data from ancestor blocks for merging
+    /// * `ancestors` - Deferred trie data from ancestor blocks for merging (oldest -> newest)
     pub fn sort_and_build_trie_input(
-        hashed_state: &HashedPostState,
-        trie_updates: &TrieUpdates,
+        hashed_state: Arc<HashedPostState>,
+        trie_updates: Arc<TrieUpdates>,
         anchor_hash: B256,
         ancestors: &[Self],
     ) -> ComputedTrieData {
-        // Sort the current block's hashed state and trie updates
-        let sorted_hashed_state = Arc::new(hashed_state.clone_into_sorted());
-        let sorted_trie_updates = Arc::new(trie_updates.clone().into_sorted());
+        let sorted_hashed_state = match Arc::try_unwrap(hashed_state) {
+            Ok(state) => state.into_sorted(),
+            Err(arc) => arc.clone_into_sorted(),
+        };
+        let sorted_trie_updates = match Arc::try_unwrap(trie_updates) {
+            Ok(updates) => updates.into_sorted(),
+            Err(arc) => arc.clone_into_sorted(),
+        };
 
-        // Merge trie data from ancestors (oldest -> newest so later state takes precedence)
-        let mut overlay = TrieInputSorted::default();
-        for ancestor in ancestors {
-            let ancestor_data = ancestor.wait_cloned();
-            {
-                let state_mut = Arc::make_mut(&mut overlay.state);
-                state_mut.extend_ref(ancestor_data.hashed_state.as_ref());
-            }
-            {
-                let nodes_mut = Arc::make_mut(&mut overlay.nodes);
-                nodes_mut.extend_ref(ancestor_data.trie_updates.as_ref());
-            }
-        }
+        // Reuse parent's overlay if available and anchors match.
+        // We can only reuse the parent's overlay if it was built on top of the same
+        // persisted anchor. If the anchor has changed (e.g., due to persistence),
+        // the parent's overlay is relative to an old state and cannot be used.
+        let overlay = if let Some(parent) = ancestors.last() {
+            let parent_data = parent.wait_cloned();
 
-        // Extend overlay with current block's sorted data
-        {
-            let state_mut = Arc::make_mut(&mut overlay.state);
-            state_mut.extend_ref(sorted_hashed_state.as_ref());
-        }
-        {
-            let nodes_mut = Arc::make_mut(&mut overlay.nodes);
-            nodes_mut.extend_ref(sorted_trie_updates.as_ref());
-        }
+            match &parent_data.anchored_trie_input {
+                // Case 1: Parent has cached overlay AND anchors match.
+                Some(AnchoredTrieInput { anchor_hash: parent_anchor, trie_input })
+                    if *parent_anchor == anchor_hash =>
+                {
+                    // O(1): Reuse parent's overlay, extend with current block's data.
+                    let mut overlay = TrieInputSorted::new(
+                        Arc::clone(&trie_input.nodes),
+                        Arc::clone(&trie_input.state),
+                        Default::default(), // prefix_sets are per-block, not cumulative
+                    );
+                    // Only trigger COW clone if there's actually data to add.
+                    if !sorted_hashed_state.is_empty() {
+                        Arc::make_mut(&mut overlay.state).extend_ref(&sorted_hashed_state);
+                    }
+                    if !sorted_trie_updates.is_empty() {
+                        Arc::make_mut(&mut overlay.nodes).extend_ref(&sorted_trie_updates);
+                    }
+                    overlay
+                }
+                // Case 2: Parent exists but anchor mismatch or no cached overlay.
+                // We must rebuild from the ancestors list (which only contains unpersisted blocks).
+                _ => Self::merge_ancestors_into_overlay(
+                    ancestors,
+                    &sorted_hashed_state,
+                    &sorted_trie_updates,
+                ),
+            }
+        } else {
+            // Case 3: No in-memory ancestors (first block after persisted anchor).
+            // Build overlay with just this block's data.
+            Self::merge_ancestors_into_overlay(&[], &sorted_hashed_state, &sorted_trie_updates)
+        };
 
         ComputedTrieData::with_trie_input(
-            sorted_hashed_state,
-            sorted_trie_updates,
+            Arc::new(sorted_hashed_state),
+            Arc::new(sorted_trie_updates),
             anchor_hash,
             Arc::new(overlay),
         )
     }
 
+    /// Merge all ancestors and current block's data into a single overlay.
+    ///
+    /// This is a rare fallback path, only used when no ancestor has a cached
+    /// `anchored_trie_input` (e.g., blocks created via alternative constructors).
+    /// In normal operation, the parent always has a cached overlay and this
+    /// function is never called.
+    ///
+    /// Iterates ancestors oldest -> newest, then extends with current block's data,
+    /// so later state takes precedence.
+    fn merge_ancestors_into_overlay(
+        ancestors: &[Self],
+        sorted_hashed_state: &HashedPostStateSorted,
+        sorted_trie_updates: &TrieUpdatesSorted,
+    ) -> TrieInputSorted {
+        let mut overlay = TrieInputSorted::default();
+
+        let state_mut = Arc::make_mut(&mut overlay.state);
+        let nodes_mut = Arc::make_mut(&mut overlay.nodes);
+
+        for ancestor in ancestors {
+            let ancestor_data = ancestor.wait_cloned();
+            state_mut.extend_ref(ancestor_data.hashed_state.as_ref());
+            nodes_mut.extend_ref(ancestor_data.trie_updates.as_ref());
+        }
+
+        // Extend with current block's sorted data last (takes precedence)
+        state_mut.extend_ref(sorted_hashed_state);
+        nodes_mut.extend_ref(sorted_trie_updates);
+
+        overlay
+    }
+
     /// Returns trie data, computing synchronously if the async task hasn't completed.
     ///
     /// - If the async task has completed (`Ready`), returns the cached result.
@@ -204,7 +267,7 @@ impl DeferredTrieData {
     #[instrument(level = "debug", target = "engine::tree::deferred_trie", skip_all)]
     pub fn wait_cloned(&self) -> ComputedTrieData {
         let mut state = self.state.lock();
-        match &*state {
+        match &mut *state {
             // If the deferred trie data is ready, return the cached result.
             DeferredState::Ready(bundle) => {
                 DEFERRED_TRIE_METRICS.deferred_trie_async_ready.increment(1);
@@ -212,11 +275,14 @@ impl DeferredTrieData {
             }
             // If the deferred trie data is pending, compute the trie data synchronously and return
             // the result. This is the fallback path if the async task hasn't completed.
-            DeferredState::Pending(inputs) => {
+            DeferredState::Pending(maybe_inputs) => {
                 DEFERRED_TRIE_METRICS.deferred_trie_sync_fallback.increment(1);
+
+                let inputs = maybe_inputs.take().expect("inputs must be present in Pending state");
+
                 let computed = Self::sort_and_build_trie_input(
-                    &inputs.hashed_state,
-                    &inputs.trie_updates,
+                    inputs.hashed_state,
+                    inputs.trie_updates,
                     inputs.anchor_hash,
                     &inputs.ancestors,
                 );
@@ -441,4 +507,365 @@ mod tests {
         let (_, account) = &overlay_state[0];
         assert_eq!(account.unwrap().nonce, 2);
     }
+
+    /// Helper to create a ready block with anchored trie input containing specific state.
+    fn ready_block_with_state(
+        anchor_hash: B256,
+        accounts: Vec<(B256, Option<Account>)>,
+    ) -> DeferredTrieData {
+        let hashed_state = Arc::new(HashedPostStateSorted::new(accounts, B256Map::default()));
+        let trie_updates = Arc::default();
+        let mut overlay = TrieInputSorted::default();
+        Arc::make_mut(&mut overlay.state).extend_ref(hashed_state.as_ref());
+
+        DeferredTrieData::ready(ComputedTrieData {
+            hashed_state,
+            trie_updates,
+            anchored_trie_input: Some(AnchoredTrieInput {
+                anchor_hash,
+                trie_input: Arc::new(overlay),
+            }),
+        })
+    }
+
+    /// Verifies that first block after anchor (no ancestors) creates empty base overlay.
+    #[test]
+    fn first_block_after_anchor_creates_empty_base() {
+        let anchor = B256::with_last_byte(1);
+        let key = B256::with_last_byte(42);
+        let account = Account { nonce: 1, balance: U256::ZERO, bytecode_hash: None };
+
+        // First block after anchor - no ancestors
+        let first_block = DeferredTrieData::pending(
+            Arc::new(HashedPostState::default().with_accounts([(key, Some(account))])),
+            Arc::new(TrieUpdates::default()),
+            anchor,
+            vec![], // No ancestors
+        );
+
+        let result = first_block.wait_cloned();
+
+        // Should have overlay with just this block's data
+        let overlay = result.anchored_trie_input.as_ref().unwrap();
+        assert_eq!(overlay.anchor_hash, anchor);
+        assert_eq!(overlay.trie_input.state.accounts.len(), 1);
+        let (found_key, found_account) = &overlay.trie_input.state.accounts[0];
+        assert_eq!(*found_key, key);
+        assert_eq!(found_account.unwrap().nonce, 1);
+    }
+
+    /// Verifies that parent's overlay is reused regardless of anchor.
+    #[test]
+    fn reuses_parent_overlay() {
+        let anchor = B256::with_last_byte(1);
+        let key = B256::with_last_byte(42);
+        let account = Account { nonce: 100, balance: U256::ZERO, bytecode_hash: None };
+
+        // Create parent with anchored trie input
+        let parent = ready_block_with_state(anchor, vec![(key, Some(account))]);
+
+        // Create child - should reuse parent's overlay
+        let child = DeferredTrieData::pending(
+            Arc::new(HashedPostState::default()),
+            Arc::new(TrieUpdates::default()),
+            anchor,
+            vec![parent],
+        );
+
+        let result = child.wait_cloned();
+
+        // Verify parent's account is in the overlay
+        let overlay = result.anchored_trie_input.as_ref().unwrap();
+        assert_eq!(overlay.anchor_hash, anchor);
+        assert_eq!(overlay.trie_input.state.accounts.len(), 1);
+        let (found_key, found_account) = &overlay.trie_input.state.accounts[0];
+        assert_eq!(*found_key, key);
+        assert_eq!(found_account.unwrap().nonce, 100);
+    }
+
+    /// Verifies that parent's overlay is NOT reused when anchor changes (after persist).
+    /// The overlay data is dependent on the anchor, so it must be rebuilt from the
+    /// remaining ancestors.
+    #[test]
+    fn rebuilds_overlay_when_anchor_changes() {
+        let old_anchor = B256::with_last_byte(1);
+        let new_anchor = B256::with_last_byte(2);
+        let key = B256::with_last_byte(42);
+        let account = Account { nonce: 50, balance: U256::ZERO, bytecode_hash: None };
+
+        // Create parent with OLD anchor
+        let parent = ready_block_with_state(old_anchor, vec![(key, Some(account))]);
+
+        // Create child with NEW anchor (simulates after persist)
+        // Should NOT reuse parent's overlay because anchor changed
+        let child = DeferredTrieData::pending(
+            Arc::new(HashedPostState::default()),
+            Arc::new(TrieUpdates::default()),
+            new_anchor,
+            vec![parent],
+        );
+
+        let result = child.wait_cloned();
+
+        // Verify result uses new anchor
+        let overlay = result.anchored_trie_input.as_ref().unwrap();
+        assert_eq!(overlay.anchor_hash, new_anchor);
+
+        // Crucially, since we provided `parent` in ancestors but it has a different anchor,
+        // the code falls back to `merge_ancestors_into_overlay`.
+        // `merge_ancestors_into_overlay` reads `parent.hashed_state` (which has the account).
+        // So the account IS present, but it was obtained via REBUILD, not REUSE.
+        // We can check `DEFERRED_TRIE_METRICS` if we want to be sure, but functionally:
+        assert_eq!(overlay.trie_input.state.accounts.len(), 1);
+        let (found_key, found_account) = &overlay.trie_input.state.accounts[0];
+        assert_eq!(*found_key, key);
+        assert_eq!(found_account.unwrap().nonce, 50);
+    }
+
+    /// Verifies that parent without `anchored_trie_input` triggers rebuild path.
+    #[test]
+    fn rebuilds_when_parent_has_no_anchored_input() {
+        let anchor = B256::with_last_byte(1);
+        let key = B256::with_last_byte(42);
+        let account = Account { nonce: 25, balance: U256::ZERO, bytecode_hash: None };
+
+        // Create parent WITHOUT anchored trie input (e.g., from without_trie_input constructor)
+        let parent_state =
+            HashedPostStateSorted::new(vec![(key, Some(account))], B256Map::default());
+        let parent = DeferredTrieData::ready(ComputedTrieData {
+            hashed_state: Arc::new(parent_state),
+            trie_updates: Arc::default(),
+            anchored_trie_input: None, // No anchored input
+        });
+
+        // Create child - should rebuild from parent's hashed_state
+        let child = DeferredTrieData::pending(
+            Arc::new(HashedPostState::default()),
+            Arc::new(TrieUpdates::default()),
+            anchor,
+            vec![parent],
+        );
+
+        let result = child.wait_cloned();
+
+        // Verify overlay is built and contains parent's data
+        let overlay = result.anchored_trie_input.as_ref().unwrap();
+        assert_eq!(overlay.anchor_hash, anchor);
+        assert_eq!(overlay.trie_input.state.accounts.len(), 1);
+    }
+
+    /// Verifies that a chain of blocks with matching anchors builds correct cumulative overlay.
+    #[test]
+    fn chain_of_blocks_builds_cumulative_overlay() {
+        let anchor = B256::with_last_byte(1);
+        let key1 = B256::with_last_byte(1);
+        let key2 = B256::with_last_byte(2);
+        let key3 = B256::with_last_byte(3);
+
+        // Block 1: sets account at key1
+        let block1 = ready_block_with_state(
+            anchor,
+            vec![(key1, Some(Account { nonce: 1, balance: U256::ZERO, bytecode_hash: None }))],
+        );
+
+        // Block 2: adds account at key2, ancestor is block1
+        let block2_hashed = HashedPostState::default().with_accounts([(
+            key2,
+            Some(Account { nonce: 2, balance: U256::ZERO, bytecode_hash: None }),
+        )]);
+        let block2 = DeferredTrieData::pending(
+            Arc::new(block2_hashed),
+            Arc::new(TrieUpdates::default()),
+            anchor,
+            vec![block1.clone()],
+        );
+        // Compute block2's trie data
+        let block2_computed = block2.wait_cloned();
+        let block2_ready = DeferredTrieData::ready(block2_computed);
+
+        // Block 3: adds account at key3, ancestor is block2 (which includes block1)
+        let block3_hashed = HashedPostState::default().with_accounts([(
+            key3,
+            Some(Account { nonce: 3, balance: U256::ZERO, bytecode_hash: None }),
+        )]);
+        let block3 = DeferredTrieData::pending(
+            Arc::new(block3_hashed),
+            Arc::new(TrieUpdates::default()),
+            anchor,
+            vec![block1, block2_ready],
+        );
+
+        let result = block3.wait_cloned();
+
+        // Verify all three accounts are in the cumulative overlay
+        let overlay = result.anchored_trie_input.as_ref().unwrap();
+        assert_eq!(overlay.trie_input.state.accounts.len(), 3);
+
+        // Accounts should be sorted by key (B256 ordering)
+        let accounts = &overlay.trie_input.state.accounts;
+        assert!(accounts.iter().any(|(k, a)| *k == key1 && a.unwrap().nonce == 1));
+        assert!(accounts.iter().any(|(k, a)| *k == key2 && a.unwrap().nonce == 2));
+        assert!(accounts.iter().any(|(k, a)| *k == key3 && a.unwrap().nonce == 3));
+    }
+
+    /// Verifies that child block's state overwrites parent's state for the same key.
+    #[test]
+    fn child_state_overwrites_parent() {
+        let anchor = B256::with_last_byte(1);
+        let key = B256::with_last_byte(42);
+
+        // Parent sets nonce to 10
+        let parent = ready_block_with_state(
+            anchor,
+            vec![(key, Some(Account { nonce: 10, balance: U256::ZERO, bytecode_hash: None }))],
+        );
+
+        // Child overwrites nonce to 99
+        let child_hashed = HashedPostState::default().with_accounts([(
+            key,
+            Some(Account { nonce: 99, balance: U256::ZERO, bytecode_hash: None }),
+        )]);
+        let child = DeferredTrieData::pending(
+            Arc::new(child_hashed),
+            Arc::new(TrieUpdates::default()),
+            anchor,
+            vec![parent],
+        );
+
+        let result = child.wait_cloned();
+
+        // Verify child's value wins (extend_ref uses later value)
+        let overlay = result.anchored_trie_input.as_ref().unwrap();
+        // Note: extend_ref may result in duplicate keys; check the last occurrence
+        let accounts = &overlay.trie_input.state.accounts;
+        let last_account = accounts.iter().rfind(|(k, _)| *k == key).unwrap();
+        assert_eq!(last_account.1.unwrap().nonce, 99);
+    }
+
+    /// Stress test: verify O(N) behavior by building a chain of many blocks.
+    /// This test ensures the fix doesn't regress - previously this would be O(N²).
+    #[test]
+    fn long_chain_builds_in_linear_time() {
+        let anchor = B256::with_last_byte(1);
+        let num_blocks = 50; // Enough to notice O(N²) vs O(N) difference
+
+        let mut ancestors: Vec<DeferredTrieData> = Vec::new();
+
+        let start = Instant::now();
+
+        for i in 0..num_blocks {
+            let key = B256::with_last_byte(i as u8);
+            let account = Account { nonce: i as u64, balance: U256::ZERO, bytecode_hash: None };
+            let hashed = HashedPostState::default().with_accounts([(key, Some(account))]);
+
+            let block = DeferredTrieData::pending(
+                Arc::new(hashed),
+                Arc::new(TrieUpdates::default()),
+                anchor,
+                ancestors.clone(),
+            );
+
+            // Compute and add to ancestors for next iteration
+            let computed = block.wait_cloned();
+            ancestors.push(DeferredTrieData::ready(computed));
+        }
+
+        let elapsed = start.elapsed();
+
+        // With O(N) fix, 50 blocks should complete quickly (< 1 second)
+        // With O(N²), this would take significantly longer
+        assert!(
+            elapsed < Duration::from_secs(2),
+            "Chain of {num_blocks} blocks took {:?}, possible O(N²) regression",
+            elapsed
+        );
+
+        // Verify final overlay has all accounts
+        let final_result = ancestors.last().unwrap().wait_cloned();
+        let overlay = final_result.anchored_trie_input.as_ref().unwrap();
+        assert_eq!(overlay.trie_input.state.accounts.len(), num_blocks);
+    }
+
+    /// Verifies that a multi-ancestor overlay is rebuilt when anchor changes.
+    /// This simulates the "persist prefix then keep building" scenario where:
+    /// 1. A chain of blocks is built with anchor A
+    /// 2. Some blocks are persisted, changing anchor to B
+    /// 3. New blocks must rebuild the overlay from the remaining ancestors
+    #[test]
+    fn multi_ancestor_overlay_rebuilt_after_anchor_change() {
+        let old_anchor = B256::with_last_byte(1);
+        let new_anchor = B256::with_last_byte(2);
+        let key1 = B256::with_last_byte(1);
+        let key2 = B256::with_last_byte(2);
+        let key3 = B256::with_last_byte(3);
+        let key4 = B256::with_last_byte(4);
+
+        // Build a chain of 3 blocks with old_anchor
+        let block1 = ready_block_with_state(
+            old_anchor,
+            vec![(key1, Some(Account { nonce: 1, balance: U256::ZERO, bytecode_hash: None }))],
+        );
+
+        let block2_hashed = HashedPostState::default().with_accounts([(
+            key2,
+            Some(Account { nonce: 2, balance: U256::ZERO, bytecode_hash: None }),
+        )]);
+        let block2 = DeferredTrieData::pending(
+            Arc::new(block2_hashed),
+            Arc::new(TrieUpdates::default()),
+            old_anchor,
+            vec![block1.clone()],
+        );
+        let block2_ready = DeferredTrieData::ready(block2.wait_cloned());
+
+        let block3_hashed = HashedPostState::default().with_accounts([(
+            key3,
+            Some(Account { nonce: 3, balance: U256::ZERO, bytecode_hash: None }),
+        )]);
+        let block3 = DeferredTrieData::pending(
+            Arc::new(block3_hashed),
+            Arc::new(TrieUpdates::default()),
+            old_anchor,
+            vec![block1.clone(), block2_ready.clone()],
+        );
+        let block3_ready = DeferredTrieData::ready(block3.wait_cloned());
+
+        // Verify block3's overlay has all 3 accounts with old_anchor
+        let block3_overlay = block3_ready.wait_cloned().anchored_trie_input.unwrap();
+        assert_eq!(block3_overlay.anchor_hash, old_anchor);
+        assert_eq!(block3_overlay.trie_input.state.accounts.len(), 3);
+
+        // Now simulate persist: create block4 with NEW anchor but same ancestors.
+        // To verify correct rebuilding, we must provide ALL unpersisted ancestors.
+        // If we only provided block3, the rebuild would only see block3's state.
+        // We pass block1, block2, block3 to simulate that they are all still in memory
+        // but the anchor check forces a rebuild (e.g. artificial anchor change).
+        let block4_hashed = HashedPostState::default().with_accounts([(
+            key4,
+            Some(Account { nonce: 4, balance: U256::ZERO, bytecode_hash: None }),
+        )]);
+        let block4 = DeferredTrieData::pending(
+            Arc::new(block4_hashed),
+            Arc::new(TrieUpdates::default()),
+            new_anchor, // Different anchor - simulates post-persist
+            vec![block1, block2_ready, block3_ready],
+        );
+
+        let result = block4.wait_cloned();
+
+        // Verify:
+        // 1. New anchor is used in result
+        assert_eq!(result.anchor_hash(), Some(new_anchor));
+
+        // 2. All 4 accounts are in the overlay (rebuilt from ancestors + extended)
+        let overlay = result.anchored_trie_input.as_ref().unwrap();
+        assert_eq!(overlay.trie_input.state.accounts.len(), 4);
+
+        // 3. All accounts have correct values
+        let accounts = &overlay.trie_input.state.accounts;
+        assert!(accounts.iter().any(|(k, a)| *k == key1 && a.unwrap().nonce == 1));
+        assert!(accounts.iter().any(|(k, a)| *k == key2 && a.unwrap().nonce == 2));
+        assert!(accounts.iter().any(|(k, a)| *k == key3 && a.unwrap().nonce == 3));
+        assert!(accounts.iter().any(|(k, a)| *k == key4 && a.unwrap().nonce == 4));
+    }
 }

crates/chain-state/src/in_memory.rs

@@ -86,14 +86,20 @@ impl<N: NodePrimitives> InMemoryState<N> {
     ///
     /// This tries to acquire a read lock. Drop any write locks before calling this.
     pub(crate) fn update_metrics(&self) {
-        let numbers = self.numbers.read();
-        if let Some((earliest_block_number, _)) = numbers.first_key_value() {
-            self.metrics.earliest_block.set(*earliest_block_number as f64);
+        let (count, earliest, latest) = {
+            let numbers = self.numbers.read();
+            let count = numbers.len();
+            let earliest = numbers.first_key_value().map(|(number, _)| *number);
+            let latest = numbers.last_key_value().map(|(number, _)| *number);
+            (count, earliest, latest)
+        };
+        if let Some(earliest_block_number) = earliest {
+            self.metrics.earliest_block.set(earliest_block_number as f64);
         }
-        if let Some((latest_block_number, _)) = numbers.last_key_value() {
-            self.metrics.latest_block.set(*latest_block_number as f64);
+        if let Some(latest_block_number) = latest {
+            self.metrics.latest_block.set(latest_block_number as f64);
         }
-        self.metrics.num_blocks.set(numbers.len() as f64);
+        self.metrics.num_blocks.set(count as f64);
     }
 
     /// Returns the state for a given block hash.
@@ -311,6 +317,7 @@ impl<N: NodePrimitives> CanonicalInMemoryState<N> {
     /// This will update the links between blocks and remove all blocks that are [..
     /// `persisted_height`].
     pub fn remove_persisted_blocks(&self, persisted_num_hash: BlockNumHash) {
+        self.set_persisted(persisted_num_hash);
         // if the persisted hash is not in the canonical in memory state, do nothing, because it
         // means canonical blocks were not actually persisted.
         //
@@ -438,6 +445,11 @@ impl<N: NodePrimitives> CanonicalInMemoryState<N> {
         self.inner.chain_info_tracker.set_finalized(header);
     }
 
+    /// Persisted block setter.
+    pub fn set_persisted(&self, num_hash: BlockNumHash) {
+        self.inner.chain_info_tracker.set_persisted(num_hash);
+    }
+
     /// Canonical head getter.
     pub fn get_canonical_head(&self) -> SealedHeader<N::BlockHeader> {
         self.inner.chain_info_tracker.get_canonical_head()
@@ -453,6 +465,11 @@ impl<N: NodePrimitives> CanonicalInMemoryState<N> {
         self.inner.chain_info_tracker.get_safe_header()
     }
 
+    /// Persisted block `BlockNumHash` getter.
+    pub fn get_persisted_num_hash(&self) -> Option<BlockNumHash> {
+        self.inner.chain_info_tracker.get_persisted_num_hash()
+    }
+
     /// Returns the `SealedHeader` corresponding to the pending state.
     pub fn pending_sealed_header(&self) -> Option<SealedHeader<N::BlockHeader>> {
         self.pending_state().map(|h| h.block_ref().recovered_block().clone_sealed_header())
@@ -505,6 +522,11 @@ impl<N: NodePrimitives> CanonicalInMemoryState<N> {
         self.inner.chain_info_tracker.subscribe_finalized_block()
     }
 
+    /// Subscribe to new persisted block events.
+    pub fn subscribe_persisted_block(&self) -> watch::Receiver<Option<BlockNumHash>> {
+        self.inner.chain_info_tracker.subscribe_persisted_block()
+    }
+
     /// Attempts to send a new [`CanonStateNotification`] to all active Receiver handles.
     pub fn notify_canon_state(&self, event: CanonStateNotification<N>) {
         self.inner.canon_state_notification_sender.send(event).ok();
@@ -664,22 +686,14 @@ impl<N: NodePrimitives> BlockState<N> {
         receipts.first().map(|receipts| receipts.deref()).unwrap_or_default()
     }
 
-    /// Returns a vector of __parent__ `BlockStates`.
+    /// Returns an iterator over __parent__ `BlockStates`.
     ///
-    /// The block state order in the output vector is newest to oldest (highest to lowest):
+    /// The block state order is newest to oldest (highest to lowest):
     /// `[5,4,3,2,1]`
     ///
     /// Note: This does not include self.
-    pub fn parent_state_chain(&self) -> Vec<&Self> {
-        let mut parents = Vec::new();
-        let mut current = self.parent.as_deref();
-
-        while let Some(parent) = current {
-            parents.push(parent);
-            current = parent.parent.as_deref();
-        }
-
-        parents
+    pub fn parent_state_chain(&self) -> impl Iterator<Item = &Self> + '_ {
+        std::iter::successors(self.parent.as_deref(), |state| state.parent.as_deref())
     }
 
     /// Returns a vector of `BlockStates` representing the entire in memory chain.
@@ -690,6 +704,11 @@ impl<N: NodePrimitives> BlockState<N> {
     }
 
     /// Appends the parent chain of this [`BlockState`] to the given vector.
+    ///
+    /// Parents are appended in order from newest to oldest (highest to lowest).
+    /// This does not include self, only the parent states.
+    ///
+    /// This is a convenience method equivalent to `chain.extend(self.parent_state_chain())`.
     pub fn append_parent_chain<'a>(&'a self, chain: &mut Vec<&'a Self>) {
         chain.extend(self.parent_state_chain());
     }
@@ -927,6 +946,8 @@ impl<N: NodePrimitives<SignedTx: SignedTransaction>> NewCanonicalChain<N> {
                     chain.append_block(
                         exec.recovered_block().clone(),
                         exec.execution_outcome().clone(),
+                        exec.trie_updates(),
+                        exec.hashed_state(),
                     );
                     chain
                 }));
@@ -937,6 +958,8 @@ impl<N: NodePrimitives<SignedTx: SignedTransaction>> NewCanonicalChain<N> {
                     chain.append_block(
                         exec.recovered_block().clone(),
                         exec.execution_outcome().clone(),
+                        exec.trie_updates(),
+                        exec.hashed_state(),
                     );
                     chain
                 }));
@@ -944,6 +967,8 @@ impl<N: NodePrimitives<SignedTx: SignedTransaction>> NewCanonicalChain<N> {
                     chain.append_block(
                         exec.recovered_block().clone(),
                         exec.execution_outcome().clone(),
+                        exec.trie_updates(),
+                        exec.hashed_state(),
                     );
                     chain
                 }));
@@ -1453,19 +1478,18 @@ mod tests {
         let mut test_block_builder: TestBlockBuilder = TestBlockBuilder::default();
         let chain = create_mock_state_chain(&mut test_block_builder, 4);
 
-        let parents = chain[3].parent_state_chain();
+        let parents: Vec<_> = chain[3].parent_state_chain().collect();
         assert_eq!(parents.len(), 3);
         assert_eq!(parents[0].block().recovered_block().number, 3);
         assert_eq!(parents[1].block().recovered_block().number, 2);
         assert_eq!(parents[2].block().recovered_block().number, 1);
 
-        let parents = chain[2].parent_state_chain();
+        let parents: Vec<_> = chain[2].parent_state_chain().collect();
         assert_eq!(parents.len(), 2);
         assert_eq!(parents[0].block().recovered_block().number, 2);
         assert_eq!(parents[1].block().recovered_block().number, 1);
 
-        let parents = chain[0].parent_state_chain();
-        assert_eq!(parents.len(), 0);
+        assert_eq!(chain[0].parent_state_chain().count(), 0);
     }
 
     #[test]
@@ -1476,8 +1500,7 @@ mod tests {
             create_mock_state(&mut test_block_builder, single_block_number, B256::random());
         let single_block_hash = single_block.block().recovered_block().hash();
 
-        let parents = single_block.parent_state_chain();
-        assert_eq!(parents.len(), 0);
+        assert_eq!(single_block.parent_state_chain().count(), 0);
 
         let block_state_chain = single_block.chain().collect::<Vec<_>>();
         assert_eq!(block_state_chain.len(), 1);
@@ -1529,13 +1552,24 @@ mod tests {
         // Test commit notification
         let chain_commit = NewCanonicalChain::Commit { new: vec![block0.clone(), block1.clone()] };
 
+        // Build expected trie updates map
+        let mut expected_trie_updates = BTreeMap::new();
+        expected_trie_updates.insert(0, block0.trie_updates());
+        expected_trie_updates.insert(1, block1.trie_updates());
+
+        // Build expected hashed state map
+        let mut expected_hashed_state = BTreeMap::new();
+        expected_hashed_state.insert(0, block0.hashed_state());
+        expected_hashed_state.insert(1, block1.hashed_state());
+
         assert_eq!(
             chain_commit.to_chain_notification(),
             CanonStateNotification::Commit {
                 new: Arc::new(Chain::new(
                     vec![block0.recovered_block().clone(), block1.recovered_block().clone()],
                     sample_execution_outcome.clone(),
-                    None
+                    expected_trie_updates,
+                    expected_hashed_state
                 ))
             }
         );
@@ -1546,18 +1580,40 @@ mod tests {
             old: vec![block1.clone(), block2.clone()],
         };
 
+        // Build expected trie updates for old chain
+        let mut old_trie_updates = BTreeMap::new();
+        old_trie_updates.insert(1, block1.trie_updates());
+        old_trie_updates.insert(2, block2.trie_updates());
+
+        // Build expected trie updates for new chain
+        let mut new_trie_updates = BTreeMap::new();
+        new_trie_updates.insert(1, block1a.trie_updates());
+        new_trie_updates.insert(2, block2a.trie_updates());
+
+        // Build expected hashed state for old chain
+        let mut old_hashed_state = BTreeMap::new();
+        old_hashed_state.insert(1, block1.hashed_state());
+        old_hashed_state.insert(2, block2.hashed_state());
+
+        // Build expected hashed state for new chain
+        let mut new_hashed_state = BTreeMap::new();
+        new_hashed_state.insert(1, block1a.hashed_state());
+        new_hashed_state.insert(2, block2a.hashed_state());
+
         assert_eq!(
             chain_reorg.to_chain_notification(),
             CanonStateNotification::Reorg {
                 old: Arc::new(Chain::new(
                     vec![block1.recovered_block().clone(), block2.recovered_block().clone()],
                     sample_execution_outcome.clone(),
-                    None
+                    old_trie_updates,
+                    old_hashed_state
                 )),
                 new: Arc::new(Chain::new(
                     vec![block1a.recovered_block().clone(), block2a.recovered_block().clone()],
                     sample_execution_outcome,
-                    None
+                    new_trie_updates,
+                    new_hashed_state
                 ))
             }
         );

crates/chain-state/src/lib.rs

@@ -23,7 +23,8 @@ mod notifications;
 pub use notifications::{
     CanonStateNotification, CanonStateNotificationSender, CanonStateNotificationStream,
     CanonStateNotifications, CanonStateSubscriptions, ForkChoiceNotifications, ForkChoiceStream,
-    ForkChoiceSubscriptions,
+    ForkChoiceSubscriptions, PersistedBlockNotifications, PersistedBlockSubscriptions,
+    WatchValueStream,
 };
 
 mod memory_overlay;

crates/chain-state/src/memory_overlay.rs

@@ -5,14 +5,14 @@ use reth_errors::ProviderResult;
 use reth_primitives_traits::{Account, Bytecode, NodePrimitives};
 use reth_storage_api::{
     AccountReader, BlockHashReader, BytecodeReader, HashedPostStateProvider, StateProofProvider,
-    StateProvider, StateRootProvider, StorageRootProvider,
+    StateProvider, StateProviderBox, StateRootProvider, StorageRootProvider,
 };
 use reth_trie::{
     updates::TrieUpdates, AccountProof, HashedPostState, HashedStorage, MultiProof,
     MultiProofTargets, StorageMultiProof, TrieInput,
 };
 use revm_database::BundleState;
-use std::sync::OnceLock;
+use std::{borrow::Cow, sync::OnceLock};
 
 /// A state provider that stores references to in-memory blocks along with their state as well as a
 /// reference of the historical state provider for fallback lookups.
@@ -24,15 +24,11 @@ pub struct MemoryOverlayStateProviderRef<
     /// Historical state provider for state lookups that are not found in memory blocks.
     pub(crate) historical: Box<dyn StateProvider + 'a>,
     /// The collection of executed parent blocks. Expected order is newest to oldest.
-    pub(crate) in_memory: Vec<ExecutedBlock<N>>,
+    pub(crate) in_memory: Cow<'a, [ExecutedBlock<N>]>,
     /// Lazy-loaded in-memory trie data.
     pub(crate) trie_input: OnceLock<TrieInput>,
 }
 
-/// A state provider that stores references to in-memory blocks along with their state as well as
-/// the historical state provider for fallback lookups.
-pub type MemoryOverlayStateProvider<N> = MemoryOverlayStateProviderRef<'static, N>;
-
 impl<'a, N: NodePrimitives> MemoryOverlayStateProviderRef<'a, N> {
     /// Create new memory overlay state provider.
     ///
@@ -42,7 +38,7 @@ impl<'a, N: NodePrimitives> MemoryOverlayStateProviderRef<'a, N> {
     /// - `historical` - a historical state provider for the latest ancestor block stored in the
     ///   database.
     pub fn new(historical: Box<dyn StateProvider + 'a>, in_memory: Vec<ExecutedBlock<N>>) -> Self {
-        Self { historical, in_memory, trie_input: OnceLock::new() }
+        Self { historical, in_memory: Cow::Owned(in_memory), trie_input: OnceLock::new() }
     }
 
     /// Turn this state provider into a state provider
@@ -53,11 +49,14 @@ impl<'a, N: NodePrimitives> MemoryOverlayStateProviderRef<'a, N> {
     /// Return lazy-loaded trie state aggregated from in-memory blocks.
     fn trie_input(&self) -> &TrieInput {
         self.trie_input.get_or_init(|| {
-            let bundles: Vec<_> =
-                self.in_memory.iter().rev().map(|block| block.trie_data()).collect();
-            TrieInput::from_blocks_sorted(
-                bundles.iter().map(|data| (data.hashed_state.as_ref(), data.trie_updates.as_ref())),
-            )
+            let mut input = TrieInput::default();
+            // Iterate from oldest to newest
+            for block in self.in_memory.iter().rev() {
+                let data = block.trie_data();
+                input.nodes.extend_from_sorted(&data.trie_updates);
+                input.state.extend_from_sorted(&data.hashed_state);
+            }
+            input
         })
     }
 
@@ -71,7 +70,7 @@ impl<'a, N: NodePrimitives> MemoryOverlayStateProviderRef<'a, N> {
 
 impl<N: NodePrimitives> BlockHashReader for MemoryOverlayStateProviderRef<'_, N> {
     fn block_hash(&self, number: BlockNumber) -> ProviderResult<Option<B256>> {
-        for block in &self.in_memory {
+        for block in self.in_memory.iter() {
             if block.recovered_block().number() == number {
                 return Ok(Some(block.recovered_block().hash()));
             }
@@ -90,7 +89,7 @@ impl<N: NodePrimitives> BlockHashReader for MemoryOverlayStateProviderRef<'_, N>
         let mut in_memory_hashes = Vec::with_capacity(range.size_hint().0);
 
         // iterate in ascending order (oldest to newest = low to high)
-        for block in &self.in_memory {
+        for block in self.in_memory.iter() {
             let block_num = block.recovered_block().number();
             if range.contains(&block_num) {
                 in_memory_hashes.push(block.recovered_block().hash());
@@ -112,7 +111,7 @@ impl<N: NodePrimitives> BlockHashReader for MemoryOverlayStateProviderRef<'_, N>
 
 impl<N: NodePrimitives> AccountReader for MemoryOverlayStateProviderRef<'_, N> {
     fn basic_account(&self, address: &Address) -> ProviderResult<Option<Account>> {
-        for block in &self.in_memory {
+        for block in self.in_memory.iter() {
             if let Some(account) = block.execution_output.account(address) {
                 return Ok(account);
             }
@@ -216,7 +215,7 @@ impl<N: NodePrimitives> StateProvider for MemoryOverlayStateProviderRef<'_, N> {
         address: Address,
         storage_key: StorageKey,
     ) -> ProviderResult<Option<StorageValue>> {
-        for block in &self.in_memory {
+        for block in self.in_memory.iter() {
             if let Some(value) = block.execution_output.storage(&address, storage_key.into()) {
                 return Ok(Some(value));
             }
@@ -228,7 +227,7 @@ impl<N: NodePrimitives> StateProvider for MemoryOverlayStateProviderRef<'_, N> {
 
 impl<N: NodePrimitives> BytecodeReader for MemoryOverlayStateProviderRef<'_, N> {
     fn bytecode_by_hash(&self, code_hash: &B256) -> ProviderResult<Option<Bytecode>> {
-        for block in &self.in_memory {
+        for block in self.in_memory.iter() {
             if let Some(contract) = block.execution_output.bytecode(code_hash) {
                 return Ok(Some(contract));
             }
@@ -237,3 +236,46 @@ impl<N: NodePrimitives> BytecodeReader for MemoryOverlayStateProviderRef<'_, N>
         self.historical.bytecode_by_hash(code_hash)
     }
 }
+
+/// An owned state provider that stores references to in-memory blocks along with their state as
+/// well as a reference of the historical state provider for fallback lookups.
+#[expect(missing_debug_implementations)]
+pub struct MemoryOverlayStateProvider<N: NodePrimitives = reth_ethereum_primitives::EthPrimitives> {
+    /// Historical state provider for state lookups that are not found in memory blocks.
+    pub(crate) historical: StateProviderBox,
+    /// The collection of executed parent blocks. Expected order is newest to oldest.
+    pub(crate) in_memory: Vec<ExecutedBlock<N>>,
+    /// Lazy-loaded in-memory trie data.
+    pub(crate) trie_input: OnceLock<TrieInput>,
+}
+
+impl<N: NodePrimitives> MemoryOverlayStateProvider<N> {
+    /// Create new memory overlay state provider.
+    ///
+    /// ## Arguments
+    ///
+    /// - `in_memory` - the collection of executed ancestor blocks in reverse.
+    /// - `historical` - a historical state provider for the latest ancestor block stored in the
+    ///   database.
+    pub fn new(historical: StateProviderBox, in_memory: Vec<ExecutedBlock<N>>) -> Self {
+        Self { historical, in_memory, trie_input: OnceLock::new() }
+    }
+
+    /// Returns a new provider that takes the `TX` as reference
+    #[inline(always)]
+    fn as_ref(&self) -> MemoryOverlayStateProviderRef<'_, N> {
+        MemoryOverlayStateProviderRef {
+            historical: Box::new(self.historical.as_ref()),
+            in_memory: Cow::Borrowed(&self.in_memory),
+            trie_input: self.trie_input.clone(),
+        }
+    }
+
+    /// Wraps the [`Self`] in a `Box`.
+    pub fn boxed(self) -> StateProviderBox {
+        Box::new(self)
+    }
+}
+
+// Delegates all provider impls to [`MemoryOverlayStateProviderRef`]
+reth_storage_api::macros::delegate_provider_impls!(MemoryOverlayStateProvider<N> where [N: NodePrimitives]);

crates/chain-state/src/noop.rs

@@ -2,7 +2,7 @@
 
 use crate::{
     CanonStateNotifications, CanonStateSubscriptions, ForkChoiceNotifications,
-    ForkChoiceSubscriptions,
+    ForkChoiceSubscriptions, PersistedBlockNotifications, PersistedBlockSubscriptions,
 };
 use reth_primitives_traits::NodePrimitives;
 use reth_storage_api::noop::NoopProvider;
@@ -27,3 +27,10 @@ impl<C: Send + Sync, N: NodePrimitives> ForkChoiceSubscriptions for NoopProvider
         ForkChoiceNotifications(rx)
     }
 }
+
+impl<C: Send + Sync, N: NodePrimitives> PersistedBlockSubscriptions for NoopProvider<C, N> {
+    fn subscribe_persisted_block(&self) -> PersistedBlockNotifications {
+        let (_, rx) = watch::channel(None);
+        PersistedBlockNotifications(rx)
+    }
+}

crates/chain-state/src/notifications.rs

@@ -1,6 +1,6 @@
 //! Canonical chain state notification trait and types.
 
-use alloy_eips::eip2718::Encodable2718;
+use alloy_eips::{eip2718::Encodable2718, BlockNumHash};
 use derive_more::{Deref, DerefMut};
 use reth_execution_types::{BlockReceipts, Chain};
 use reth_primitives_traits::{NodePrimitives, RecoveredBlock, SealedHeader};
@@ -205,22 +205,22 @@ pub trait ForkChoiceSubscriptions: Send + Sync {
     }
 }
 
-/// A stream for fork choice watch channels (pending, safe or finalized watchers)
+/// A stream that yields values from a `watch::Receiver<Option<T>>`, filtering out `None` values.
 #[derive(Debug)]
 #[pin_project::pin_project]
-pub struct ForkChoiceStream<T> {
+pub struct WatchValueStream<T> {
     #[pin]
     st: WatchStream<Option<T>>,
 }
 
-impl<T: Clone + Sync + Send + 'static> ForkChoiceStream<T> {
-    /// Creates a new `ForkChoiceStream`
+impl<T: Clone + Sync + Send + 'static> WatchValueStream<T> {
+    /// Creates a new [`WatchValueStream`]
     pub fn new(rx: watch::Receiver<Option<T>>) -> Self {
         Self { st: WatchStream::from_changes(rx) }
     }
 }
 
-impl<T: Clone + Sync + Send + 'static> Stream for ForkChoiceStream<T> {
+impl<T: Clone + Sync + Send + 'static> Stream for WatchValueStream<T> {
     type Item = T;
 
     fn poll_next(mut self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Option<Self::Item>> {
@@ -234,6 +234,24 @@ impl<T: Clone + Sync + Send + 'static> Stream for ForkChoiceStream<T> {
     }
 }
 
+/// Alias for [`WatchValueStream`] for fork choice watch channels.
+pub type ForkChoiceStream<T> = WatchValueStream<T>;
+
+/// Wrapper around a watch receiver that receives persisted block notifications.
+#[derive(Debug, Deref, DerefMut)]
+pub struct PersistedBlockNotifications(pub watch::Receiver<Option<BlockNumHash>>);
+
+/// A trait that allows subscribing to persisted block events.
+pub trait PersistedBlockSubscriptions: Send + Sync {
+    /// Get notified when a new block is persisted to disk.
+    fn subscribe_persisted_block(&self) -> PersistedBlockNotifications;
+
+    /// Convenience method to get a stream of the persisted blocks.
+    fn persisted_block_stream(&self) -> WatchValueStream<BlockNumHash> {
+        WatchValueStream::new(self.subscribe_persisted_block().0)
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;
@@ -242,6 +260,7 @@ mod tests {
     use reth_ethereum_primitives::{Receipt, TransactionSigned, TxType};
     use reth_execution_types::ExecutionOutcome;
     use reth_primitives_traits::SealedBlock;
+    use std::collections::BTreeMap;
 
     #[test]
     fn test_commit_notification() {
@@ -260,7 +279,8 @@ mod tests {
         let chain: Arc<Chain> = Arc::new(Chain::new(
             vec![block1.clone(), block2.clone()],
             ExecutionOutcome::default(),
-            None,
+            BTreeMap::new(),
+            BTreeMap::new(),
         ));
 
         // Create a commit notification
@@ -295,12 +315,17 @@ mod tests {
         block3.set_block_number(3);
         block3.set_hash(block3_hash);
 
-        let old_chain: Arc<Chain> =
-            Arc::new(Chain::new(vec![block1.clone()], ExecutionOutcome::default(), None));
+        let old_chain: Arc<Chain> = Arc::new(Chain::new(
+            vec![block1.clone()],
+            ExecutionOutcome::default(),
+            BTreeMap::new(),
+            BTreeMap::new(),
+        ));
         let new_chain = Arc::new(Chain::new(
             vec![block2.clone(), block3.clone()],
             ExecutionOutcome::default(),
-            None,
+            BTreeMap::new(),
+            BTreeMap::new(),
         ));
 
         // Create a reorg notification
@@ -362,8 +387,12 @@ mod tests {
         let execution_outcome = ExecutionOutcome { receipts, ..Default::default() };
 
         // Create a new chain segment with `block1` and `block2` and the execution outcome.
-        let new_chain: Arc<Chain> =
-            Arc::new(Chain::new(vec![block1.clone(), block2.clone()], execution_outcome, None));
+        let new_chain: Arc<Chain> = Arc::new(Chain::new(
+            vec![block1.clone(), block2.clone()],
+            execution_outcome,
+            BTreeMap::new(),
+            BTreeMap::new(),
+        ));
 
         // Create a commit notification containing the new chain segment.
         let notification = CanonStateNotification::Commit { new: new_chain };
@@ -420,8 +449,12 @@ mod tests {
             ExecutionOutcome { receipts: old_receipts, ..Default::default() };
 
         // Create an old chain segment to be reverted, containing `old_block1`.
-        let old_chain: Arc<Chain> =
-            Arc::new(Chain::new(vec![old_block1.clone()], old_execution_outcome, None));
+        let old_chain: Arc<Chain> = Arc::new(Chain::new(
+            vec![old_block1.clone()],
+            old_execution_outcome,
+            BTreeMap::new(),
+            BTreeMap::new(),
+        ));
 
         // Define block2 for the new chain segment, which will be committed.
         let mut body = BlockBody::<TransactionSigned>::default();
@@ -449,7 +482,12 @@ mod tests {
             ExecutionOutcome { receipts: new_receipts, ..Default::default() };
 
         // Create a new chain segment to be committed, containing `new_block1`.
-        let new_chain = Arc::new(Chain::new(vec![new_block1.clone()], new_execution_outcome, None));
+        let new_chain = Arc::new(Chain::new(
+            vec![new_block1.clone()],
+            new_execution_outcome,
+            BTreeMap::new(),
+            BTreeMap::new(),
+        ));
 
         // Create a reorg notification with both reverted (old) and committed (new) chain segments.
         let notification = CanonStateNotification::Reorg { old: old_chain, new: new_chain };

crates/chain-state/src/test_utils.rs

@@ -117,7 +117,7 @@ impl<N: NodePrimitives> TestBlockBuilder<N> {
             .map(|_| {
                 let tx = mock_tx(self.signer_build_account_info.nonce);
                 self.signer_build_account_info.nonce += 1;
-                self.signer_build_account_info.balance -= signer_balance_decrease;
+                self.signer_build_account_info.balance -= Self::single_tx_cost();
                 tx
             })
             .collect();

crates/chainspec/src/spec.rs

@@ -80,6 +80,8 @@ pub fn make_genesis_header(genesis: &Genesis, hardforks: &ChainHardforks) -> Hea
         .then_some(EMPTY_REQUESTS_HASH);
 
     Header {
+        number: genesis.number.unwrap_or_default(),
+        parent_hash: genesis.parent_hash.unwrap_or_default(),
         gas_limit: genesis.gas_limit,
         difficulty: genesis.difficulty,
         nonce: genesis.nonce.into(),
@@ -968,7 +970,7 @@ impl<H: BlockHeader> EthereumHardforks for ChainSpec<H> {
 
 /// A trait for reading the current chainspec.
 #[auto_impl::auto_impl(&, Arc)]
-pub trait ChainSpecProvider: Debug + Send + Sync {
+pub trait ChainSpecProvider: Debug + Send {
     /// The chain spec type.
     type ChainSpec: EthChainSpec + 'static;
 

crates/cli/commands/Cargo.toml

@@ -129,3 +129,5 @@ arbitrary = [
     "reth-primitives-traits/arbitrary",
     "reth-ethereum-primitives/arbitrary",
 ]
+
+edge = ["reth-db-common/edge", "reth-stages/rocksdb"]

crates/cli/commands/src/common.rs

@@ -23,7 +23,10 @@ use reth_node_core::{
     dirs::{ChainPath, DataDirPath},
 };
 use reth_provider::{
-    providers::{BlockchainProvider, NodeTypesForProvider, RocksDBProvider, StaticFileProvider},
+    providers::{
+        BlockchainProvider, NodeTypesForProvider, RocksDBProvider, StaticFileProvider,
+        StaticFileProviderBuilder,
+    },
     ProviderFactory, StaticFileProviderFactory,
 };
 use reth_stages::{sets::DefaultStages, Pipeline, PipelineTarget};
@@ -100,15 +103,23 @@ impl<C: ChainSpecParser> EnvironmentArgs<C> {
         }
 
         info!(target: "reth::cli", ?db_path, ?sf_path, "Opening storage");
+        let genesis_block_number = self.chain.genesis().number.unwrap_or_default();
         let (db, sfp) = match access {
             AccessRights::RW => (
                 Arc::new(init_db(db_path, self.db.database_args())?),
-                StaticFileProvider::read_write(sf_path)?,
-            ),
-            AccessRights::RO | AccessRights::RoInconsistent => (
-                Arc::new(open_db_read_only(&db_path, self.db.database_args())?),
-                StaticFileProvider::read_only(sf_path, false)?,
+                StaticFileProviderBuilder::read_write(sf_path)
+                    .with_genesis_block_number(genesis_block_number)
+                    .build()?,
             ),
+            AccessRights::RO | AccessRights::RoInconsistent => {
+                (Arc::new(open_db_read_only(&db_path, self.db.database_args())?), {
+                    let provider = StaticFileProviderBuilder::read_only(sf_path)
+                        .with_genesis_block_number(genesis_block_number)
+                        .build()?;
+                    provider.watch_directory();
+                    provider
+                })
+            }
         };
         // TransactionDB only support read-write mode
         let rocksdb_provider = RocksDBProvider::builder(data_dir.rocksdb())

crates/cli/commands/src/db/get.rs

@@ -2,8 +2,8 @@ use alloy_primitives::{hex, BlockHash};
 use clap::Parser;
 use reth_db::{
     static_file::{
-        ColumnSelectorOne, ColumnSelectorTwo, HeaderWithHashMask, ReceiptMask, TransactionMask,
-        TransactionSenderMask,
+        AccountChangesetMask, ColumnSelectorOne, ColumnSelectorTwo, HeaderWithHashMask,
+        ReceiptMask, TransactionMask, TransactionSenderMask,
     },
     RawDupSort,
 };
@@ -19,7 +19,7 @@ use reth_db_common::DbTool;
 use reth_node_api::{HeaderTy, ReceiptTy, TxTy};
 use reth_node_builder::NodeTypesWithDB;
 use reth_primitives_traits::ValueWithSubKey;
-use reth_provider::{providers::ProviderNodeTypes, StaticFileProviderFactory};
+use reth_provider::{providers::ProviderNodeTypes, ChangeSetReader, StaticFileProviderFactory};
 use reth_static_file_types::StaticFileSegment;
 use tracing::error;
 
@@ -64,6 +64,10 @@ enum Subcommand {
         #[arg(value_parser = maybe_json_value_parser)]
         key: String,
 
+        /// The subkey to get content for, for example address in changeset
+        #[arg(value_parser = maybe_json_value_parser)]
+        subkey: Option<String>,
+
         /// Output bytes instead of human-readable decoded value
         #[arg(long)]
         raw: bool,
@@ -77,33 +81,77 @@ impl Command {
             Subcommand::Mdbx { table, key, subkey, end_key, end_subkey, raw } => {
                 table.view(&GetValueViewer { tool, key, subkey, end_key, end_subkey, raw })?
             }
-            Subcommand::StaticFile { segment, key, raw } => {
-                let (key, mask): (u64, _) = match segment {
+            Subcommand::StaticFile { segment, key, subkey, raw } => {
+                let (key, subkey, mask): (u64, _, _) = match segment {
                     StaticFileSegment::Headers => (
                         table_key::<tables::Headers>(&key)?,
+                        None,
                         <HeaderWithHashMask<HeaderTy<N>>>::MASK,
                     ),
-                    StaticFileSegment::Transactions => {
-                        (table_key::<tables::Transactions>(&key)?, <TransactionMask<TxTy<N>>>::MASK)
-                    }
-                    StaticFileSegment::Receipts => {
-                        (table_key::<tables::Receipts>(&key)?, <ReceiptMask<ReceiptTy<N>>>::MASK)
-                    }
+                    StaticFileSegment::Transactions => (
+                        table_key::<tables::Transactions>(&key)?,
+                        None,
+                        <TransactionMask<TxTy<N>>>::MASK,
+                    ),
+                    StaticFileSegment::Receipts => (
+                        table_key::<tables::Receipts>(&key)?,
+                        None,
+                        <ReceiptMask<ReceiptTy<N>>>::MASK,
+                    ),
                     StaticFileSegment::TransactionSenders => (
                         table_key::<tables::TransactionSenders>(&key)?,
-                        <TransactionSenderMask>::MASK,
+                        None,
+                        TransactionSenderMask::MASK,
                     ),
+                    StaticFileSegment::AccountChangeSets => {
+                        let subkey =
+                            table_subkey::<tables::AccountChangeSets>(subkey.as_deref()).ok();
+                        (
+                            table_key::<tables::AccountChangeSets>(&key)?,
+                            subkey,
+                            AccountChangesetMask::MASK,
+                        )
+                    }
                 };
 
-                let content = tool
-                    .provider_factory
-                    .static_file_provider()
-                    .get_segment_provider(segment, key)?
-                    .cursor()?
-                    .get(key.into(), mask)
-                    .map(|result| {
-                        result.map(|vec| vec.iter().map(|slice| slice.to_vec()).collect::<Vec<_>>())
-                    })?;
+                // handle account changesets differently if a subkey is provided.
+                if let StaticFileSegment::AccountChangeSets = segment {
+                    let Some(subkey) = subkey else {
+                        // get all changesets for the block
+                        let changesets = tool
+                            .provider_factory
+                            .static_file_provider()
+                            .account_block_changeset(key)?;
+
+                        println!("{}", serde_json::to_string_pretty(&changesets)?);
+                        return Ok(())
+                    };
+
+                    let account = tool
+                        .provider_factory
+                        .static_file_provider()
+                        .get_account_before_block(key, subkey)?;
+
+                    if let Some(account) = account {
+                        println!("{}", serde_json::to_string_pretty(&account)?);
+                    } else {
+                        error!(target: "reth::cli", "No content for the given table key.");
+                    }
+
+                    return Ok(())
+                }
+
+                let content = tool.provider_factory.static_file_provider().find_static_file(
+                    segment,
+                    |provider| {
+                        let mut cursor = provider.cursor()?;
+                        cursor.get(key.into(), mask).map(|result| {
+                            result.map(|vec| {
+                                vec.iter().map(|slice| slice.to_vec()).collect::<Vec<_>>()
+                            })
+                        })
+                    },
+                )?;
 
                 match content {
                     Some(content) => {
@@ -139,6 +187,9 @@ impl Command {
                                         )?;
                                     println!("{}", serde_json::to_string_pretty(&sender)?);
                                 }
+                                StaticFileSegment::AccountChangeSets => {
+                                    unreachable!("account changeset static files are special cased before this match")
+                                }
                             }
                         }
                     }

crates/cli/commands/src/db/mod.rs

@@ -162,7 +162,7 @@ impl<C: ChainSpecParser<ChainSpec: EthChainSpec + EthereumHardforks>> Command<C>
                 let access_rights =
                     if command.dry_run { AccessRights::RO } else { AccessRights::RW };
                 db_exec!(self.env, tool, N, access_rights, {
-                    command.execute(&tool, ctx.task_executor.clone())?;
+                    command.execute(&tool, ctx.task_executor.clone(), &data_dir)?;
                 });
             }
             Subcommands::StaticFileHeader(command) => {

crates/cli/commands/src/db/repair_trie.rs

@@ -9,7 +9,10 @@ use reth_db_api::{
     transaction::{DbTx, DbTxMut},
 };
 use reth_db_common::DbTool;
-use reth_node_core::version::version_metadata;
+use reth_node_core::{
+    dirs::{ChainPath, DataDirPath},
+    version::version_metadata,
+};
 use reth_node_metrics::{
     chain::ChainSpecInfo,
     hooks::Hooks,
@@ -53,11 +56,13 @@ impl Command {
         self,
         tool: &DbTool<N>,
         task_executor: TaskExecutor,
+        data_dir: &ChainPath<DataDirPath>,
     ) -> eyre::Result<()> {
         // Set up metrics server if requested
         let _metrics_handle = if let Some(listen_addr) = self.metrics {
             let chain_name = tool.provider_factory.chain_spec().chain().to_string();
             let executor = task_executor.clone();
+            let pprof_dump_dir = data_dir.pprof_dumps();
 
             let handle = task_executor.spawn_critical("metrics server", async move {
                 let config = MetricServerConfig::new(
@@ -73,6 +78,7 @@ impl Command {
                     ChainSpecInfo { name: chain_name },
                     executor,
                     Hooks::builder().build(),
+                    pprof_dump_dir,
                 );
 
                 // Spawn the metrics server
@@ -301,8 +307,8 @@ fn verify_and_repair<N: ProviderNodeTypes>(tool: &DbTool<N>) -> eyre::Result<()>
     if inconsistent_nodes == 0 {
         info!("No inconsistencies found");
     } else {
-        info!("Repaired {} inconsistencies, committing changes", inconsistent_nodes);
         provider_rw.commit()?;
+        info!("Repaired {} inconsistencies and committed changes", inconsistent_nodes);
     }
 
     Ok(())

crates/cli/commands/src/db/settings.rs

@@ -40,12 +40,17 @@ enum Subcommands {
 #[clap(rename_all = "snake_case")]
 pub enum SetCommand {
     /// Store receipts in static files instead of the database
-    ReceiptsInStaticFiles {
+    Receipts {
         #[clap(action(ArgAction::Set))]
         value: bool,
     },
     /// Store transaction senders in static files instead of the database
-    TransactionSendersInStaticFiles {
+    TransactionSenders {
+        #[clap(action(ArgAction::Set))]
+        value: bool,
+    },
+    /// Store account changesets in static files instead of the database
+    AccountChangesets {
         #[clap(action(ArgAction::Set))]
         value: bool,
     },
@@ -94,11 +99,12 @@ impl Command {
             storages_history_in_rocksdb: _,
             transaction_hash_numbers_in_rocksdb: _,
             account_history_in_rocksdb: _,
+            account_changesets_in_static_files: _,
         } = settings.unwrap_or_else(StorageSettings::legacy);
 
         // Update the setting based on the key
         match cmd {
-            SetCommand::ReceiptsInStaticFiles { value } => {
+            SetCommand::Receipts { value } => {
                 if settings.receipts_in_static_files == value {
                     println!("receipts_in_static_files is already set to {}", value);
                     return Ok(());
@@ -106,7 +112,7 @@ impl Command {
                 settings.receipts_in_static_files = value;
                 println!("Set receipts_in_static_files = {}", value);
             }
-            SetCommand::TransactionSendersInStaticFiles { value } => {
+            SetCommand::TransactionSenders { value } => {
                 if settings.transaction_senders_in_static_files == value {
                     println!("transaction_senders_in_static_files is already set to {}", value);
                     return Ok(());
@@ -114,6 +120,14 @@ impl Command {
                 settings.transaction_senders_in_static_files = value;
                 println!("Set transaction_senders_in_static_files = {}", value);
             }
+            SetCommand::AccountChangesets { value } => {
+                if settings.account_changesets_in_static_files == value {
+                    println!("account_changesets_in_static_files is already set to {}", value);
+                    return Ok(());
+                }
+                settings.account_changesets_in_static_files = value;
+                println!("Set account_changesets_in_static_files = {}", value);
+            }
         }
 
         // Write updated settings

crates/cli/commands/src/import_core.rs

@@ -69,9 +69,7 @@ pub async fn import_blocks_from_file<N>(
     provider_factory: ProviderFactory<N>,
     config: &Config,
     executor: impl ConfigureEvm<Primitives = N::Primitives> + 'static,
-    consensus: Arc<
-        impl FullConsensus<N::Primitives, Error = reth_consensus::ConsensusError> + 'static,
-    >,
+    consensus: Arc<impl FullConsensus<N::Primitives> + 'static>,
 ) -> eyre::Result<ImportResult>
 where
     N: ProviderNodeTypes,
@@ -198,7 +196,7 @@ pub fn build_import_pipeline_impl<N, C, E>(
 ) -> eyre::Result<(Pipeline<N>, impl futures::Stream<Item = NodeEvent<N::Primitives>> + use<N, C, E>)>
 where
     N: ProviderNodeTypes,
-    C: FullConsensus<N::Primitives, Error = reth_consensus::ConsensusError> + 'static,
+    C: FullConsensus<N::Primitives> + 'static,
     E: ConfigureEvm<Primitives = N::Primitives> + 'static,
 {
     if !file_client.has_canonical_blocks() {

crates/cli/commands/src/init_cmd.rs

@@ -1,8 +1,9 @@
 //! Command that initializes the node from a genesis file.
 
 use crate::common::{AccessRights, CliNodeTypes, Environment, EnvironmentArgs};
+use alloy_consensus::BlockHeader;
 use clap::Parser;
-use reth_chainspec::{EthChainSpec, EthereumHardforks};
+use reth_chainspec::{ChainSpecProvider, EthChainSpec, EthereumHardforks};
 use reth_cli::chainspec::ChainSpecParser;
 use reth_provider::BlockHashReader;
 use std::sync::Arc;
@@ -22,8 +23,9 @@ impl<C: ChainSpecParser<ChainSpec: EthChainSpec + EthereumHardforks>> InitComman
 
         let Environment { provider_factory, .. } = self.env.init::<N>(AccessRights::RW)?;
 
+        let genesis_block_number = provider_factory.chain_spec().genesis_header().number();
         let hash = provider_factory
-            .block_hash(0)?
+            .block_hash(genesis_block_number)?
             .ok_or_else(|| eyre::eyre!("Genesis hash not found."))?;
 
         info!(target: "reth::cli", hash = ?hash, "Genesis block written");

crates/cli/commands/src/init_state/without_evm.rs

@@ -79,7 +79,7 @@ where
         + StaticFileProviderFactory<Primitives: NodePrimitives<BlockHeader: Compact>>,
 {
     provider_rw.insert_block(
-        SealedBlock::<<Provider::Primitives as NodePrimitives>::Block>::from_sealed_parts(
+        &SealedBlock::<<Provider::Primitives as NodePrimitives>::Block>::from_sealed_parts(
             header.clone(),
             Default::default(),
         )
@@ -99,6 +99,7 @@ where
 /// * Headers: It will push an empty block.
 /// * Transactions: It will not push any tx, only increments the end block range.
 /// * Receipts: It will not push any receipt, only increments the end block range.
+/// * TransactionSenders: If the segment exists, increments the end block range.
 fn append_dummy_chain<N, F>(
     sf_provider: &StaticFileProvider<N>,
     target_height: BlockNumber,
@@ -110,8 +111,15 @@ where
 {
     let (tx, rx) = std::sync::mpsc::channel();
 
-    // Spawn jobs for incrementing the block end range of transactions and receipts
-    for segment in [StaticFileSegment::Transactions, StaticFileSegment::Receipts] {
+    // Spawn jobs for incrementing the block end range of transactions, receipts, and senders.
+    for segment in [
+        StaticFileSegment::Transactions,
+        StaticFileSegment::Receipts,
+        StaticFileSegment::TransactionSenders,
+    ] {
+        if sf_provider.get_highest_static_file_block(segment).is_none() {
+            continue
+        }
         let tx_clone = tx.clone();
         let provider = sf_provider.clone();
         std::thread::spawn(move || {
@@ -151,9 +159,15 @@ where
 
     // If, for any reason, rayon crashes this verifies if all segments are at the same
     // target_height.
-    for segment in
-        [StaticFileSegment::Headers, StaticFileSegment::Receipts, StaticFileSegment::Transactions]
-    {
+    for segment in [
+        StaticFileSegment::Headers,
+        StaticFileSegment::Receipts,
+        StaticFileSegment::Transactions,
+        StaticFileSegment::TransactionSenders,
+    ] {
+        if sf_provider.get_highest_static_file_block(segment).is_none() {
+            continue
+        }
         assert_eq!(
             sf_provider.latest_writer(segment)?.user_header().block_end(),
             Some(target_height),

crates/cli/commands/src/p2p/mod.rs

@@ -72,7 +72,7 @@ impl<C: ChainSpecParser<ChainSpec: EthChainSpec + Hardforks + EthereumHardforks>
                 .split();
                 if result.len() != 1 {
                     eyre::bail!(
-                        "Invalid number of headers received. Expected: 1. Received: {}",
+                        "Invalid number of bodies received. Expected: 1. Received: {}",
                         result.len()
                     )
                 }

crates/cli/commands/src/stage/drop.rs

@@ -87,6 +87,9 @@ impl<C: ChainSpecParser> Command<C> {
                             .unwrap_or_default();
                         writer.prune_transaction_senders(to_delete, 0)?;
                     }
+                    StaticFileSegment::AccountChangeSets => {
+                        writer.prune_account_changesets(highest_block)?;
+                    }
                 }
             }
         }

crates/cli/commands/src/stage/dump/execution.rs

@@ -1,5 +1,5 @@
 use super::setup;
-use reth_consensus::{noop::NoopConsensus, ConsensusError, FullConsensus};
+use reth_consensus::{noop::NoopConsensus, FullConsensus};
 use reth_db::DatabaseEnv;
 use reth_db_api::{
     cursor::DbCursorRO, database::Database, table::TableImporter, tables, transaction::DbTx,
@@ -28,7 +28,7 @@ pub(crate) async fn dump_execution_stage<N, E, C>(
 where
     N: ProviderNodeTypes<DB = Arc<DatabaseEnv>>,
     E: ConfigureEvm<Primitives = N::Primitives> + 'static,
-    C: FullConsensus<E::Primitives, Error = ConsensusError> + 'static,
+    C: FullConsensus<E::Primitives> + 'static,
 {
     let (output_db, tip_block_number) = setup(from, to, &output_datadir.db(), db_tool)?;
 
@@ -169,7 +169,7 @@ fn dry_run<N, E, C>(
 where
     N: ProviderNodeTypes,
     E: ConfigureEvm<Primitives = N::Primitives> + 'static,
-    C: FullConsensus<E::Primitives, Error = ConsensusError> + 'static,
+    C: FullConsensus<E::Primitives> + 'static,
 {
     info!(target: "reth::cli", "Executing stage. [dry-run]");
 

crates/cli/commands/src/stage/dump/merkle.rs

@@ -4,7 +4,7 @@ use super::setup;
 use alloy_primitives::{Address, BlockNumber};
 use eyre::Result;
 use reth_config::config::EtlConfig;
-use reth_consensus::{ConsensusError, FullConsensus};
+use reth_consensus::FullConsensus;
 use reth_db::DatabaseEnv;
 use reth_db_api::{database::Database, models::BlockNumberAddress, table::TableImporter, tables};
 use reth_db_common::DbTool;
@@ -31,7 +31,7 @@ pub(crate) async fn dump_merkle_stage<N>(
     output_datadir: ChainPath<DataDirPath>,
     should_run: bool,
     evm_config: impl ConfigureEvm<Primitives = N::Primitives>,
-    consensus: impl FullConsensus<N::Primitives, Error = ConsensusError> + 'static,
+    consensus: impl FullConsensus<N::Primitives> + 'static,
 ) -> Result<()>
 where
     N: ProviderNodeTypes<DB = Arc<DatabaseEnv>>,
@@ -79,7 +79,7 @@ fn unwind_and_copy<N: ProviderNodeTypes>(
     tip_block_number: u64,
     output_db: &DatabaseEnv,
     evm_config: impl ConfigureEvm<Primitives = N::Primitives>,
-    consensus: impl FullConsensus<N::Primitives, Error = ConsensusError> + 'static,
+    consensus: impl FullConsensus<N::Primitives> + 'static,
 ) -> eyre::Result<()> {
     let (from, to) = range;
     let provider = db_tool.provider_factory.database_provider_rw()?;

crates/cli/commands/src/stage/run.rs

@@ -153,6 +153,7 @@ impl<C: ChainSpecParser<ChainSpec: EthChainSpec + Hardforks + EthereumHardforks>
                         }
                     })
                     .build(),
+                data_dir.pprof_dumps(),
             );
 
             MetricServer::new(config).serve().await?;

crates/cli/runner/src/lib.rs

@@ -18,8 +18,8 @@ use tracing::{debug, error, trace};
 ///
 /// Provides utilities for running a cli command to completion.
 #[derive(Debug)]
-#[non_exhaustive]
 pub struct CliRunner {
+    config: CliRunnerConfig,
     tokio_runtime: tokio::runtime::Runtime,
 }
 
@@ -29,12 +29,18 @@ impl CliRunner {
     ///
     /// The default tokio runtime is multi-threaded, with both I/O and time drivers enabled.
     pub fn try_default_runtime() -> Result<Self, std::io::Error> {
-        Ok(Self { tokio_runtime: tokio_runtime()? })
+        Ok(Self { config: CliRunnerConfig::default(), tokio_runtime: tokio_runtime()? })
     }
 
     /// Create a new [`CliRunner`] from a provided tokio [`Runtime`](tokio::runtime::Runtime).
     pub const fn from_runtime(tokio_runtime: tokio::runtime::Runtime) -> Self {
-        Self { tokio_runtime }
+        Self { config: CliRunnerConfig::new(), tokio_runtime }
+    }
+
+    /// Sets the [`CliRunnerConfig`] for this runner.
+    pub const fn with_config(mut self, config: CliRunnerConfig) -> Self {
+        self.config = config;
+        self
     }
 
     /// Executes an async block on the runtime and blocks until completion.
@@ -74,7 +80,7 @@ impl CliRunner {
             // after the command has finished or exit signal was received we shutdown the task
             // manager which fires the shutdown signal to all tasks spawned via the task
             // executor and awaiting on tasks spawned with graceful shutdown
-            task_manager.graceful_shutdown_with_timeout(Duration::from_secs(5));
+            task_manager.graceful_shutdown_with_timeout(self.config.graceful_shutdown_timeout);
         }
 
         // `drop(tokio_runtime)` would block the current thread until its pools
@@ -128,7 +134,7 @@ impl CliRunner {
             error!(target: "reth::cli", "shutting down due to error");
         } else {
             debug!(target: "reth::cli", "shutting down gracefully");
-            task_manager.graceful_shutdown_with_timeout(Duration::from_secs(5));
+            task_manager.graceful_shutdown_with_timeout(self.config.graceful_shutdown_timeout);
         }
 
         // Shutdown the runtime on a separate thread
@@ -211,6 +217,38 @@ pub struct CliContext {
     pub task_executor: TaskExecutor,
 }
 
+/// Default timeout for graceful shutdown of tasks.
+const DEFAULT_GRACEFUL_SHUTDOWN_TIMEOUT: Duration = Duration::from_secs(5);
+
+/// Configuration for [`CliRunner`].
+#[derive(Debug, Clone)]
+pub struct CliRunnerConfig {
+    /// Timeout for graceful shutdown of tasks.
+    ///
+    /// After the command completes, this is the maximum time to wait for spawned tasks
+    /// to finish before forcefully terminating them.
+    pub graceful_shutdown_timeout: Duration,
+}
+
+impl Default for CliRunnerConfig {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+impl CliRunnerConfig {
+    /// Creates a new config with default values.
+    pub const fn new() -> Self {
+        Self { graceful_shutdown_timeout: DEFAULT_GRACEFUL_SHUTDOWN_TIMEOUT }
+    }
+
+    /// Sets the graceful shutdown timeout.
+    pub const fn with_graceful_shutdown_timeout(mut self, timeout: Duration) -> Self {
+        self.graceful_shutdown_timeout = timeout;
+        self
+    }
+}
+
 /// Creates a new default tokio multi-thread [Runtime](tokio::runtime::Runtime) with all features
 /// enabled
 pub fn tokio_runtime() -> Result<tokio::runtime::Runtime, std::io::Error> {

crates/config/src/config.rs

@@ -22,7 +22,6 @@ pub const DEFAULT_BLOCK_INTERVAL: usize = 5;
 #[cfg_attr(feature = "serde", serde(default))]
 pub struct Config {
     /// Configuration for each stage in the pipeline.
-    // TODO(onbjerg): Can we make this easier to maintain when we add/remove stages?
     pub stages: StageConfig,
     /// Configuration for pruning.
     #[cfg_attr(feature = "serde", serde(default))]
@@ -438,6 +437,8 @@ pub struct BlocksPerFileConfig {
     pub receipts: Option<u64>,
     /// Number of blocks per file for the transaction senders segment.
     pub transaction_senders: Option<u64>,
+    /// Number of blocks per file for the account changesets segment.
+    pub account_change_sets: Option<u64>,
 }
 
 impl StaticFilesConfig {
@@ -445,8 +446,13 @@ impl StaticFilesConfig {
     ///
     /// Returns an error if any blocks per file value is zero.
     pub fn validate(&self) -> eyre::Result<()> {
-        let BlocksPerFileConfig { headers, transactions, receipts, transaction_senders } =
-            self.blocks_per_file;
+        let BlocksPerFileConfig {
+            headers,
+            transactions,
+            receipts,
+            transaction_senders,
+            account_change_sets,
+        } = self.blocks_per_file;
         eyre::ensure!(headers != Some(0), "Headers segment blocks per file must be greater than 0");
         eyre::ensure!(
             transactions != Some(0),
@@ -460,13 +466,22 @@ impl StaticFilesConfig {
             transaction_senders != Some(0),
             "Transaction senders segment blocks per file must be greater than 0"
         );
+        eyre::ensure!(
+            account_change_sets != Some(0),
+            "Account changesets segment blocks per file must be greater than 0"
+        );
         Ok(())
     }
 
     /// Converts the blocks per file configuration into a [`HashMap`] per segment.
     pub fn as_blocks_per_file_map(&self) -> HashMap<StaticFileSegment, u64> {
-        let BlocksPerFileConfig { headers, transactions, receipts, transaction_senders } =
-            self.blocks_per_file;
+        let BlocksPerFileConfig {
+            headers,
+            transactions,
+            receipts,
+            transaction_senders,
+            account_change_sets,
+        } = self.blocks_per_file;
 
         let mut map = HashMap::new();
         // Iterating over all possible segments allows us to do an exhaustive match here,
@@ -477,6 +492,7 @@ impl StaticFilesConfig {
                 StaticFileSegment::Transactions => transactions,
                 StaticFileSegment::Receipts => receipts,
                 StaticFileSegment::TransactionSenders => transaction_senders,
+                StaticFileSegment::AccountChangeSets => account_change_sets,
             };
 
             if let Some(blocks_per_file) = blocks_per_file {
@@ -528,7 +544,7 @@ impl PruneConfig {
 
     /// Returns whether there is any kind of receipt pruning configuration.
     pub fn has_receipts_pruning(&self) -> bool {
-        self.segments.receipts.is_some() || !self.segments.receipts_log_filter.is_empty()
+        self.segments.has_receipts_pruning()
     }
 
     /// Merges values from `other` into `self`.

crates/consensus/common/src/validation.rs

@@ -500,13 +500,11 @@ mod tests {
         let expected_blob_gas_used = 10 * DATA_GAS_PER_BLOB;
 
         // validate blob, it should fail blob gas used validation
-        assert_eq!(
-            validate_block_pre_execution(&block, &chain_spec),
-            Err(ConsensusError::BlobGasUsedDiff(GotExpected {
-                got: 1,
-                expected: expected_blob_gas_used
-            }))
-        );
+        assert!(matches!(
+            validate_block_pre_execution(&block, &chain_spec).unwrap_err(),
+            ConsensusError::BlobGasUsedDiff(diff)
+                if diff.got == 1 && diff.expected == expected_blob_gas_used
+        ));
     }
 
     #[test]
@@ -517,10 +515,10 @@ mod tests {
 
         // Test exceeding default - should fail
         let header_33 = Header { extra_data: Bytes::from(vec![0; 33]), ..Default::default() };
-        assert_eq!(
-            validate_header_extra_data(&header_33, 32),
-            Err(ConsensusError::ExtraDataExceedsMax { len: 33 })
-        );
+        assert!(matches!(
+            validate_header_extra_data(&header_33, 32).unwrap_err(),
+            ConsensusError::ExtraDataExceedsMax { len } if len == 33
+        ));
 
         // Test with custom larger limit - should pass
         assert!(validate_header_extra_data(&header_33, 64).is_ok());

crates/consensus/consensus/src/lib.rs

@@ -11,9 +11,10 @@
 
 extern crate alloc;
 
-use alloc::{boxed::Box, fmt::Debug, string::String, vec::Vec};
+use alloc::{boxed::Box, fmt::Debug, string::String, sync::Arc, vec::Vec};
 use alloy_consensus::Header;
 use alloy_primitives::{BlockHash, BlockNumber, Bloom, B256};
+use core::error::Error;
 use reth_execution_types::BlockExecutionResult;
 use reth_primitives_traits::{
     constants::{GAS_LIMIT_BOUND_DIVISOR, MAXIMUM_GAS_LIMIT_BLOCK, MINIMUM_GAS_LIMIT},
@@ -49,15 +50,12 @@ pub trait FullConsensus<N: NodePrimitives>: Consensus<N::Block> {
 /// Consensus is a protocol that chooses canonical chain.
 #[auto_impl::auto_impl(&, Arc)]
 pub trait Consensus<B: Block>: HeaderValidator<B::Header> {
-    /// The error type related to consensus.
-    type Error;
-
     /// Ensures that body field values match the header.
     fn validate_body_against_header(
         &self,
         body: &B::Body,
         header: &SealedHeader<B::Header>,
-    ) -> Result<(), Self::Error>;
+    ) -> Result<(), ConsensusError>;
 
     /// Validate a block disregarding world state, i.e. things that can be checked before sender
     /// recovery and execution.
@@ -69,7 +67,7 @@ pub trait Consensus<B: Block>: HeaderValidator<B::Header> {
     /// **This should not be called for the genesis block**.
     ///
     /// Note: validating blocks does not include other validations of the Consensus
-    fn validate_block_pre_execution(&self, block: &SealedBlock<B>) -> Result<(), Self::Error>;
+    fn validate_block_pre_execution(&self, block: &SealedBlock<B>) -> Result<(), ConsensusError>;
 }
 
 /// `HeaderValidator` is a protocol that validates headers and their relationships.
@@ -125,7 +123,7 @@ pub trait HeaderValidator<H = Header>: Debug + Send + Sync {
 }
 
 /// Consensus Errors
-#[derive(Debug, PartialEq, Eq, Clone, thiserror::Error)]
+#[derive(Debug, Clone, thiserror::Error)]
 pub enum ConsensusError {
     /// Error when the gas used in the header exceeds the gas limit.
     #[error("block used gas ({gas_used}) is greater than gas limit ({gas_limit})")]
@@ -135,7 +133,7 @@ pub enum ConsensusError {
         /// The gas limit in the block header.
         gas_limit: u64,
     },
-    /// Error when the gas the gas limit is more than the maximum allowed.
+    /// Error when the gas limit is more than the maximum allowed.
     #[error(
         "header gas limit ({gas_limit}) exceed the maximum allowed gas limit ({MAXIMUM_GAS_LIMIT_BLOCK})"
     )]
@@ -410,6 +408,9 @@ pub enum ConsensusError {
     /// Other, likely an injected L2 error.
     #[error("{0}")]
     Other(String),
+    /// Other unspecified error.
+    #[error(transparent)]
+    Custom(#[from] Arc<dyn Error + Send + Sync>),
 }
 
 impl ConsensusError {
@@ -447,3 +448,34 @@ pub struct TxGasLimitTooHighErr {
     /// The maximum allowed gas limit
     pub max_allowed: u64,
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[derive(thiserror::Error, Debug)]
+    #[error("Custom L2 consensus error")]
+    struct CustomL2Error;
+
+    #[test]
+    fn test_custom_error_conversion() {
+        // Test conversion from custom error to ConsensusError
+        let custom_err = CustomL2Error;
+        let arc_err: Arc<dyn Error + Send + Sync> = Arc::new(custom_err);
+        let consensus_err: ConsensusError = arc_err.into();
+
+        // Verify it's the Custom variant
+        assert!(matches!(consensus_err, ConsensusError::Custom(_)));
+    }
+
+    #[test]
+    fn test_custom_error_display() {
+        let custom_err = CustomL2Error;
+        let arc_err: Arc<dyn Error + Send + Sync> = Arc::new(custom_err);
+        let consensus_err: ConsensusError = arc_err.into();
+
+        // Verify the error message is preserved through transparent attribute
+        let error_message = format!("{}", consensus_err);
+        assert_eq!(error_message, "Custom L2 consensus error");
+    }
+}

crates/consensus/consensus/src/noop.rs

@@ -55,19 +55,17 @@ impl<H> HeaderValidator<H> for NoopConsensus {
 }
 
 impl<B: Block> Consensus<B> for NoopConsensus {
-    type Error = ConsensusError;
-
     /// Validates body against header (no-op implementation).
     fn validate_body_against_header(
         &self,
         _body: &B::Body,
         _header: &SealedHeader<B::Header>,
-    ) -> Result<(), Self::Error> {
+    ) -> Result<(), ConsensusError> {
         Ok(())
     }
 
     /// Validates block before execution (no-op implementation).
-    fn validate_block_pre_execution(&self, _block: &SealedBlock<B>) -> Result<(), Self::Error> {
+    fn validate_block_pre_execution(&self, _block: &SealedBlock<B>) -> Result<(), ConsensusError> {
         Ok(())
     }
 }

crates/consensus/consensus/src/test_utils.rs

@@ -61,13 +61,11 @@ impl<N: NodePrimitives> FullConsensus<N> for TestConsensus {
 }
 
 impl<B: Block> Consensus<B> for TestConsensus {
-    type Error = ConsensusError;
-
     fn validate_body_against_header(
         &self,
         _body: &B::Body,
         _header: &SealedHeader<B::Header>,
-    ) -> Result<(), Self::Error> {
+    ) -> Result<(), ConsensusError> {
         if self.fail_body_against_header() {
             Err(ConsensusError::BaseFeeMissing)
         } else {
@@ -75,7 +73,7 @@ impl<B: Block> Consensus<B> for TestConsensus {
         }
     }
 
-    fn validate_block_pre_execution(&self, _block: &SealedBlock<B>) -> Result<(), Self::Error> {
+    fn validate_block_pre_execution(&self, _block: &SealedBlock<B>) -> Result<(), ConsensusError> {
         if self.fail_validation() {
             Err(ConsensusError::BaseFeeMissing)
         } else {

crates/consensus/debug-client/src/providers/rpc.rs

@@ -89,8 +89,8 @@ where
                 match res {
                     Ok(block) => {
                         if tx.send((self.convert)(block)).await.is_err() {
-                            // Channel closed.
-                            break;
+                            // Channel closed - receiver dropped, exit completely.
+                            return;
                         }
                     }
                     Err(err) => {
@@ -107,7 +107,7 @@ where
             debug!(
                 target: "consensus::debug-client",
                 url=%self.url,
-                "Re-estbalishing block subscription",
+                "Re-establishing block subscription",
             );
         }
     }

crates/e2e-test-utils/src/setup_builder.rs

@@ -11,6 +11,7 @@ use reth_node_builder::{
     PayloadTypes,
 };
 use reth_node_core::args::{DiscoveryArgs, NetworkArgs, RpcServerArgs};
+use reth_primitives_traits::AlloyBlockHeader;
 use reth_provider::providers::BlockchainProvider;
 use reth_rpc_server_types::RpcModuleSelection;
 use reth_tasks::TaskManager;
@@ -157,8 +158,8 @@ where
                     .await?;
 
                 let node = NodeTestContext::new(node, self.attributes_generator).await?;
-
-                let genesis = node.block_hash(0);
+                let genesis_number = self.chain_spec.genesis_header().number();
+                let genesis = node.block_hash(genesis_number);
                 node.update_forkchoice(genesis, genesis).await?;
 
                 eyre::Ok(node)

crates/e2e-test-utils/src/setup_import.rs

@@ -113,7 +113,6 @@ pub async fn setup_engine_with_chain_import(
         let rocksdb_dir_path = datadir.join("rocksdb");
 
         // Initialize the database using init_db (same as CLI import command)
-        // Use the same database arguments as the node will use
         let db_args = reth_node_core::args::DatabaseArgs::default().database_args();
         let db_env = reth_db::init_db(&db_path, db_args)?;
         let db = Arc::new(db_env);
@@ -317,7 +316,8 @@ mod tests {
 
         // Import the chain
         {
-            let db_env = reth_db::init_db(&db_path, DatabaseArguments::default()).unwrap();
+            let db_args = reth_node_core::args::DatabaseArgs::default().database_args();
+            let db_env = reth_db::init_db(&db_path, db_args).unwrap();
             let db = Arc::new(db_env);
 
             let provider_factory: ProviderFactory<
@@ -475,7 +475,8 @@ mod tests {
         let datadir = temp_dir.path().join("datadir");
         std::fs::create_dir_all(&datadir).unwrap();
         let db_path = datadir.join("db");
-        let db_env = reth_db::init_db(&db_path, DatabaseArguments::default()).unwrap();
+        let db_args = reth_node_core::args::DatabaseArgs::default().database_args();
+        let db_env = reth_db::init_db(&db_path, db_args).unwrap();
         let db = Arc::new(reth_db::test_utils::TempDatabase::new(db_env, db_path));
 
         // Create static files path

crates/e2e-test-utils/src/testsuite/actions/engine_api.rs

@@ -2,9 +2,7 @@
 
 use crate::testsuite::{Action, Environment};
 use alloy_primitives::B256;
-use alloy_rpc_types_engine::{
-    ExecutionPayloadV1, ExecutionPayloadV2, ExecutionPayloadV3, PayloadStatusEnum,
-};
+use alloy_rpc_types_engine::{ExecutionPayloadV3, PayloadStatusEnum};
 use alloy_rpc_types_eth::{Block, Header, Receipt, Transaction, TransactionRequest};
 use eyre::Result;
 use futures_util::future::BoxFuture;
@@ -131,7 +129,10 @@ where
             })?;
 
             // Convert block to ExecutionPayloadV3
-            let payload = block_to_payload_v3(block.clone());
+            let payload = ExecutionPayloadV3::from_block_unchecked(
+                block.hash(),
+                &block.map_transactions(|tx| tx.inner).into_consensus(),
+            );
 
             // Send the payload to the target node
             let target_engine = env.node_clients[self.node_idx].engine.http_client();
@@ -327,32 +328,3 @@ where
         })
     }
 }
-
-/// Helper function to convert a block to `ExecutionPayloadV3`
-fn block_to_payload_v3(block: Block) -> ExecutionPayloadV3 {
-    use alloy_primitives::U256;
-
-    ExecutionPayloadV3 {
-        payload_inner: ExecutionPayloadV2 {
-            payload_inner: ExecutionPayloadV1 {
-                parent_hash: block.header.inner.parent_hash,
-                fee_recipient: block.header.inner.beneficiary,
-                state_root: block.header.inner.state_root,
-                receipts_root: block.header.inner.receipts_root,
-                logs_bloom: block.header.inner.logs_bloom,
-                prev_randao: block.header.inner.mix_hash,
-                block_number: block.header.inner.number,
-                gas_limit: block.header.inner.gas_limit,
-                gas_used: block.header.inner.gas_used,
-                timestamp: block.header.inner.timestamp,
-                extra_data: block.header.inner.extra_data.clone(),
-                base_fee_per_gas: U256::from(block.header.inner.base_fee_per_gas.unwrap_or(0)),
-                block_hash: block.header.hash,
-                transactions: vec![], // No transactions needed for buffering tests
-            },
-            withdrawals: block.withdrawals.unwrap_or_default().to_vec(),
-        },
-        blob_gas_used: block.header.inner.blob_gas_used.unwrap_or(0),
-        excess_blob_gas: block.header.inner.excess_blob_gas.unwrap_or(0),
-    }
-}

crates/engine/invalid-block-hooks/src/witness.rs

@@ -5,7 +5,7 @@ use pretty_assertions::Comparison;
 use reth_engine_primitives::InvalidBlockHook;
 use reth_evm::{execute::Executor, ConfigureEvm};
 use reth_primitives_traits::{NodePrimitives, RecoveredBlock, SealedHeader};
-use reth_provider::{BlockExecutionOutput, StateProvider, StateProviderFactory};
+use reth_provider::{BlockExecutionOutput, StateProvider, StateProviderBox, StateProviderFactory};
 use reth_revm::{
     database::StateProviderDatabase,
     db::{BundleState, State},
@@ -80,13 +80,13 @@ fn sort_bundle_state_for_comparison(bundle_state: &BundleState) -> BundleStateSo
                     BundleAccountSorted {
                         info: acc.info.clone(),
                         original_info: acc.original_info.clone(),
-                        storage: BTreeMap::from_iter(acc.storage.clone()),
+                        storage: acc.storage.iter().map(|(k, v)| (*k, *v)).collect(),
                         status: acc.status,
                     },
                 )
             })
             .collect(),
-        contracts: BTreeMap::from_iter(bundle_state.contracts.clone()),
+        contracts: bundle_state.contracts.iter().map(|(k, v)| (*k, v.clone())).collect(),
         reverts: bundle_state
             .reverts
             .iter()
@@ -98,7 +98,7 @@ fn sort_bundle_state_for_comparison(bundle_state: &BundleState) -> BundleStateSo
                             *addr,
                             AccountRevertSorted {
                                 account: rev.account.clone(),
-                                storage: BTreeMap::from_iter(rev.storage.clone()),
+                                storage: rev.storage.iter().map(|(k, v)| (*k, *v)).collect(),
                                 previous_status: rev.previous_status,
                                 wipe_storage: rev.wipe_storage,
                             },
@@ -114,7 +114,7 @@ fn sort_bundle_state_for_comparison(bundle_state: &BundleState) -> BundleStateSo
 
 /// Extracts execution data including codes, preimages, and hashed state from database
 fn collect_execution_data(
-    mut db: State<StateProviderDatabase<Box<dyn StateProvider>>>,
+    mut db: State<StateProviderDatabase<StateProviderBox>>,
 ) -> eyre::Result<CollectionResult> {
     let bundle_state = db.take_bundle();
     let mut codes = BTreeMap::new();
@@ -530,9 +530,7 @@ mod tests {
         // Create a State with StateProviderTest
         let state_provider = StateProviderTest::default();
         let mut state = State::builder()
-            .with_database(StateProviderDatabase::new(
-                Box::new(state_provider) as Box<dyn StateProvider>
-            ))
+            .with_database(StateProviderDatabase::new(Box::new(state_provider) as StateProviderBox))
             .with_bundle_update()
             .build();
 

crates/engine/primitives/src/config.rs

@@ -135,6 +135,8 @@ pub struct TreeConfig {
     storage_worker_count: usize,
     /// Number of account proof worker threads.
     account_worker_count: usize,
+    /// Whether to enable V2 storage proofs.
+    enable_proof_v2: bool,
 }
 
 impl Default for TreeConfig {
@@ -163,6 +165,7 @@ impl Default for TreeConfig {
             allow_unwind_canonical_header: false,
             storage_worker_count: default_storage_worker_count(),
             account_worker_count: default_account_worker_count(),
+            enable_proof_v2: false,
         }
     }
 }
@@ -194,6 +197,7 @@ impl TreeConfig {
         allow_unwind_canonical_header: bool,
         storage_worker_count: usize,
         account_worker_count: usize,
+        enable_proof_v2: bool,
     ) -> Self {
         Self {
             persistence_threshold,
@@ -219,6 +223,7 @@ impl TreeConfig {
             allow_unwind_canonical_header,
             storage_worker_count,
             account_worker_count,
+            enable_proof_v2,
         }
     }
 
@@ -500,4 +505,15 @@ impl TreeConfig {
         self.account_worker_count = account_worker_count.max(MIN_WORKER_COUNT);
         self
     }
+
+    /// Return whether V2 storage proofs are enabled.
+    pub const fn enable_proof_v2(&self) -> bool {
+        self.enable_proof_v2
+    }
+
+    /// Setter for whether to enable V2 storage proofs.
+    pub const fn with_enable_proof_v2(mut self, enable_proof_v2: bool) -> Self {
+        self.enable_proof_v2 = enable_proof_v2;
+        self
+    }
 }

crates/engine/service/src/service.rs

@@ -1,7 +1,7 @@
 use futures::{Stream, StreamExt};
 use pin_project::pin_project;
 use reth_chainspec::EthChainSpec;
-use reth_consensus::{ConsensusError, FullConsensus};
+use reth_consensus::FullConsensus;
 use reth_engine_primitives::{BeaconEngineMessage, ConsensusEngineEvent};
 use reth_engine_tree::{
     backfill::PipelineSync,
@@ -70,7 +70,7 @@ where
     /// Constructor for `EngineService`.
     #[expect(clippy::too_many_arguments)]
     pub fn new<V, C>(
-        consensus: Arc<dyn FullConsensus<N::Primitives, Error = ConsensusError>>,
+        consensus: Arc<dyn FullConsensus<N::Primitives>>,
         chain_spec: Arc<N::ChainSpec>,
         client: Client,
         incoming_requests: EngineMessageStream<N::Payload>,

crates/engine/tree/Cargo.toml

@@ -52,6 +52,7 @@ futures.workspace = true
 thiserror.workspace = true
 tokio = { workspace = true, features = ["rt", "rt-multi-thread", "sync", "macros"] }
 mini-moka = { workspace = true, features = ["sync"] }
+moka = { workspace = true, features = ["sync"] }
 smallvec.workspace = true
 
 # metrics

crates/engine/tree/docs/root.md

@@ -128,12 +128,12 @@ we send them along with the state updates to the [Sparse Trie Task](#sparse-trie
 
 ### Finishing the calculation
 
-Once all transactions are executed, the [Engine](#engine) sends a `StateRootMessage::FinishStateUpdates` message
+Once all transactions are executed, the [Engine](#engine) sends a `StateRootMessage::FinishedStateUpdates` message
 to the State Root Task, marking the end of receiving state updates.
 
 Every time we receive a new proof from the [MultiProof Manager](#multiproof-manager), we also check
 the following conditions:
-1. Are all updates received? (`StateRootMessage::FinishStateUpdates` was sent)
+1. Are all updates received? (`StateRootMessage::FinishedStateUpdates` was sent)
 2. Is `ProofSequencer` empty? (no proofs are pending for sequencing)
 3. Are all proofs that were sent to the [`MultiProofManager::spawn_or_queue`](#multiproof-manager) finished
 calculating and were sent to the [Sparse Trie Task](#sparse-trie-task)?

crates/engine/tree/src/backfill.rs

@@ -47,7 +47,7 @@ impl BackfillSyncState {
 }
 
 /// Backfill sync mode functionality.
-pub trait BackfillSync: Send + Sync {
+pub trait BackfillSync: Send {
     /// Performs a backfill action.
     fn on_action(&mut self, action: BackfillAction);
 

crates/engine/tree/src/chain.rs

@@ -219,10 +219,19 @@ pub enum HandlerEvent<T> {
 }
 
 /// Internal events issued by the [`ChainOrchestrator`].
-#[derive(Clone, Debug)]
+#[derive(Debug)]
 pub enum FromOrchestrator {
     /// Invoked when backfill sync finished
     BackfillSyncFinished(ControlFlow),
     /// Invoked when backfill sync started
     BackfillSyncStarted,
+    /// Gracefully terminate the engine service.
+    ///
+    /// When this variant is received, the engine will persist all remaining in-memory blocks
+    /// to disk before shutting down. Once persistence is complete, a signal is sent through
+    /// the oneshot channel to notify the caller.
+    Terminate {
+        /// Channel to signal termination completion.
+        tx: tokio::sync::oneshot::Sender<()>,
+    },
 }

crates/engine/tree/src/download.rs

@@ -4,7 +4,7 @@ use crate::{engine::DownloadRequest, metrics::BlockDownloaderMetrics};
 use alloy_consensus::BlockHeader;
 use alloy_primitives::B256;
 use futures::FutureExt;
-use reth_consensus::{Consensus, ConsensusError};
+use reth_consensus::Consensus;
 use reth_network_p2p::{
     full_block::{FetchFullBlockFuture, FetchFullBlockRangeFuture, FullBlockClient},
     BlockClient,
@@ -81,7 +81,7 @@ where
     B: Block,
 {
     /// Create a new instance
-    pub fn new(client: Client, consensus: Arc<dyn Consensus<B, Error = ConsensusError>>) -> Self {
+    pub fn new(client: Client, consensus: Arc<dyn Consensus<B>>) -> Self {
         Self {
             full_block_client: FullBlockClient::new(client, consensus),
             inflight_full_block_requests: Vec::new(),

crates/engine/tree/src/engine.rs

@@ -6,6 +6,7 @@ use crate::{
     download::{BlockDownloader, DownloadAction, DownloadOutcome},
 };
 use alloy_primitives::B256;
+use crossbeam_channel::Sender;
 use futures::{Stream, StreamExt};
 use reth_chain_state::ExecutedBlock;
 use reth_engine_primitives::{BeaconEngineMessage, ConsensusEngineEvent};
@@ -15,7 +16,6 @@ use reth_primitives_traits::{Block, NodePrimitives, SealedBlock};
 use std::{
     collections::HashSet,
     fmt::Display,
-    sync::mpsc::Sender,
     task::{ready, Context, Poll},
 };
 use tokio::sync::mpsc::UnboundedReceiver;

crates/engine/tree/src/metrics.rs

@@ -19,6 +19,8 @@ pub(crate) struct PersistenceMetrics {
     pub(crate) remove_blocks_above_duration_seconds: Histogram,
     /// How long it took for blocks to be saved
     pub(crate) save_blocks_duration_seconds: Histogram,
+    /// How many blocks we persist at once.
+    pub(crate) save_blocks_block_count: Histogram,
     /// How long it took for blocks to be pruned
     pub(crate) prune_before_duration_seconds: Histogram,
 }

crates/engine/tree/src/persistence.rs

@@ -1,6 +1,6 @@
 use crate::metrics::PersistenceMetrics;
-use alloy_consensus::BlockHeader;
 use alloy_eips::BlockNumHash;
+use crossbeam_channel::Sender as CrossbeamSender;
 use reth_chain_state::ExecutedBlock;
 use reth_errors::ProviderError;
 use reth_ethereum_primitives::EthPrimitives;
@@ -16,7 +16,6 @@ use std::{
     time::Instant,
 };
 use thiserror::Error;
-use tokio::sync::oneshot;
 use tracing::{debug, error};
 
 /// Writes parts of reth's in memory tree state to the database and static files.
@@ -142,27 +141,25 @@ where
         &self,
         blocks: Vec<ExecutedBlock<N::Primitives>>,
     ) -> Result<Option<BlockNumHash>, PersistenceError> {
-        let first_block_hash = blocks.first().map(|b| b.recovered_block.num_hash());
-        let last_block_hash = blocks.last().map(|b| b.recovered_block.num_hash());
-        debug!(target: "engine::persistence", first=?first_block_hash, last=?last_block_hash, "Saving range of blocks");
+        let first_block = blocks.first().map(|b| b.recovered_block.num_hash());
+        let last_block = blocks.last().map(|b| b.recovered_block.num_hash());
+        let block_count = blocks.len();
+        debug!(target: "engine::persistence", ?block_count, first=?first_block, last=?last_block, "Saving range of blocks");
 
         let start_time = Instant::now();
-        let last_block_hash_num = blocks.last().map(|block| BlockNumHash {
-            hash: block.recovered_block().hash(),
-            number: block.recovered_block().header().number(),
-        });
 
-        if last_block_hash_num.is_some() {
+        if last_block.is_some() {
             let provider_rw = self.provider.database_provider_rw()?;
 
             provider_rw.save_blocks(blocks)?;
             provider_rw.commit()?;
         }
 
-        debug!(target: "engine::persistence", first=?first_block_hash, last=?last_block_hash, "Saved range of blocks");
+        debug!(target: "engine::persistence", first=?first_block, last=?last_block, "Saved range of blocks");
 
+        self.metrics.save_blocks_block_count.record(block_count as f64);
         self.metrics.save_blocks_duration_seconds.record(start_time.elapsed());
-        Ok(last_block_hash_num)
+        Ok(last_block)
     }
 }
 
@@ -186,13 +183,13 @@ pub enum PersistenceAction<N: NodePrimitives = EthPrimitives> {
     ///
     /// First, header, transaction, and receipt-related data should be written to static files.
     /// Then the execution history-related data will be written to the database.
-    SaveBlocks(Vec<ExecutedBlock<N>>, oneshot::Sender<Option<BlockNumHash>>),
+    SaveBlocks(Vec<ExecutedBlock<N>>, CrossbeamSender<Option<BlockNumHash>>),
 
     /// Removes block data above the given block number from the database.
     ///
     /// This will first update checkpoints from the database, then remove actual block data from
     /// static files.
-    RemoveBlocksAbove(u64, oneshot::Sender<Option<BlockNumHash>>),
+    RemoveBlocksAbove(u64, CrossbeamSender<Option<BlockNumHash>>),
 
     /// Update the persisted finalized block on disk
     SaveFinalizedBlock(u64),
@@ -264,7 +261,7 @@ impl<T: NodePrimitives> PersistenceHandle<T> {
     pub fn save_blocks(
         &self,
         blocks: Vec<ExecutedBlock<T>>,
-        tx: oneshot::Sender<Option<BlockNumHash>>,
+        tx: CrossbeamSender<Option<BlockNumHash>>,
     ) -> Result<(), SendError<PersistenceAction<T>>> {
         self.send_action(PersistenceAction::SaveBlocks(blocks, tx))
     }
@@ -293,7 +290,7 @@ impl<T: NodePrimitives> PersistenceHandle<T> {
     pub fn remove_blocks_above(
         &self,
         block_num: u64,
-        tx: oneshot::Sender<Option<BlockNumHash>>,
+        tx: CrossbeamSender<Option<BlockNumHash>>,
     ) -> Result<(), SendError<PersistenceAction<T>>> {
         self.send_action(PersistenceAction::RemoveBlocksAbove(block_num, tx))
     }
@@ -322,22 +319,22 @@ mod tests {
         PersistenceHandle::<EthPrimitives>::spawn_service(provider, pruner, sync_metrics_tx)
     }
 
-    #[tokio::test]
-    async fn test_save_blocks_empty() {
+    #[test]
+    fn test_save_blocks_empty() {
         reth_tracing::init_test_tracing();
         let persistence_handle = default_persistence_handle();
 
         let blocks = vec![];
-        let (tx, rx) = oneshot::channel();
+        let (tx, rx) = crossbeam_channel::bounded(1);
 
         persistence_handle.save_blocks(blocks, tx).unwrap();
 
-        let hash = rx.await.unwrap();
+        let hash = rx.recv().unwrap();
         assert_eq!(hash, None);
     }
 
-    #[tokio::test]
-    async fn test_save_blocks_single_block() {
+    #[test]
+    fn test_save_blocks_single_block() {
         reth_tracing::init_test_tracing();
         let persistence_handle = default_persistence_handle();
         let block_number = 0;
@@ -347,37 +344,35 @@ mod tests {
         let block_hash = executed.recovered_block().hash();
 
         let blocks = vec![executed];
-        let (tx, rx) = oneshot::channel();
+        let (tx, rx) = crossbeam_channel::bounded(1);
 
         persistence_handle.save_blocks(blocks, tx).unwrap();
 
-        let BlockNumHash { hash: actual_hash, number: _ } =
-            tokio::time::timeout(std::time::Duration::from_secs(10), rx)
-                .await
-                .expect("test timed out")
-                .expect("channel closed unexpectedly")
-                .expect("no hash returned");
+        let BlockNumHash { hash: actual_hash, number: _ } = rx
+            .recv_timeout(std::time::Duration::from_secs(10))
+            .expect("test timed out")
+            .expect("no hash returned");
 
         assert_eq!(block_hash, actual_hash);
     }
 
-    #[tokio::test]
-    async fn test_save_blocks_multiple_blocks() {
+    #[test]
+    fn test_save_blocks_multiple_blocks() {
         reth_tracing::init_test_tracing();
         let persistence_handle = default_persistence_handle();
 
         let mut test_block_builder = TestBlockBuilder::eth();
         let blocks = test_block_builder.get_executed_blocks(0..5).collect::<Vec<_>>();
         let last_hash = blocks.last().unwrap().recovered_block().hash();
-        let (tx, rx) = oneshot::channel();
+        let (tx, rx) = crossbeam_channel::bounded(1);
 
         persistence_handle.save_blocks(blocks, tx).unwrap();
-        let BlockNumHash { hash: actual_hash, number: _ } = rx.await.unwrap().unwrap();
+        let BlockNumHash { hash: actual_hash, number: _ } = rx.recv().unwrap().unwrap();
         assert_eq!(last_hash, actual_hash);
     }
 
-    #[tokio::test]
-    async fn test_save_blocks_multiple_calls() {
+    #[test]
+    fn test_save_blocks_multiple_calls() {
         reth_tracing::init_test_tracing();
         let persistence_handle = default_persistence_handle();
 
@@ -386,11 +381,11 @@ mod tests {
         for range in ranges {
             let blocks = test_block_builder.get_executed_blocks(range).collect::<Vec<_>>();
             let last_hash = blocks.last().unwrap().recovered_block().hash();
-            let (tx, rx) = oneshot::channel();
+            let (tx, rx) = crossbeam_channel::bounded(1);
 
             persistence_handle.save_blocks(blocks, tx).unwrap();
 
-            let BlockNumHash { hash: actual_hash, number: _ } = rx.await.unwrap().unwrap();
+            let BlockNumHash { hash: actual_hash, number: _ } = rx.recv().unwrap().unwrap();
             assert_eq!(last_hash, actual_hash);
         }
     }

crates/engine/tree/src/tree/block_buffer.rs

@@ -14,7 +14,7 @@ use std::collections::{BTreeMap, HashMap, HashSet, VecDeque};
 /// * [`BlockBuffer::remove_old_blocks`] to remove old blocks that precede the finalized number.
 ///
 /// Note: Buffer is limited by number of blocks that it can contain and eviction of the block
-/// is done by last recently used block.
+/// is done in FIFO order (oldest inserted block is evicted first).
 #[derive(Debug)]
 pub struct BlockBuffer<B: Block> {
     /// All blocks in the buffer stored by their block hash.
@@ -66,9 +66,14 @@ impl<B: Block> BlockBuffer<B> {
     pub fn insert_block(&mut self, block: SealedBlock<B>) {
         let hash = block.hash();
 
-        self.parent_to_child.entry(block.parent_hash()).or_default().insert(hash);
-        self.earliest_blocks.entry(block.number()).or_default().insert(hash);
-        self.blocks.insert(hash, block);
+        match self.blocks.entry(hash) {
+            std::collections::hash_map::Entry::Occupied(_) => return,
+            std::collections::hash_map::Entry::Vacant(entry) => {
+                self.parent_to_child.entry(block.parent_hash()).or_default().insert(hash);
+                self.earliest_blocks.entry(block.number()).or_default().insert(hash);
+                entry.insert(block);
+            }
+        };
 
         // Add block to FIFO queue and handle eviction if needed
         if self.block_queue.len() >= self.max_blocks {

crates/engine/tree/src/tree/cached_state.rs

@@ -31,6 +31,9 @@ pub(crate) struct CachedStateProvider<S> {
 
     /// Metrics for the cached state provider
     metrics: CachedStateMetrics,
+
+    /// If prewarm enabled we populate every cache miss
+    prewarm: bool,
 }
 
 impl<S> CachedStateProvider<S>
@@ -39,12 +42,32 @@ where
 {
     /// Creates a new [`CachedStateProvider`] from an [`ExecutionCache`], state provider, and
     /// [`CachedStateMetrics`].
-    pub(crate) const fn new_with_caches(
+    pub(crate) const fn new(
         state_provider: S,
         caches: ExecutionCache,
         metrics: CachedStateMetrics,
     ) -> Self {
-        Self { state_provider, caches, metrics }
+        Self { state_provider, caches, metrics, prewarm: false }
+    }
+}
+
+impl<S> CachedStateProvider<S> {
+    /// Enables pre-warm mode so that every cache miss is populated.
+    ///
+    /// This is only relevant for pre-warm transaction execution with the intention to pre-populate
+    /// the cache with data for regular block execution. During regular block execution the
+    /// cache doesn't need to be populated because the actual EVM database
+    /// [`State`](revm::database::State) also caches internally during block execution and the cache
+    /// is then updated after the block with the entire [`BundleState`] output of that block which
+    /// contains all accessed accounts,code,storage. See also [`ExecutionCache::insert_state`].
+    pub(crate) const fn prewarm(mut self) -> Self {
+        self.prewarm = true;
+        self
+    }
+
+    /// Returns whether this provider should pre-warm cache misses.
+    const fn is_prewarm(&self) -> bool {
+        self.prewarm
     }
 }
 
@@ -123,7 +146,10 @@ impl<S: AccountReader> AccountReader for CachedStateProvider<S> {
         self.metrics.account_cache_misses.increment(1);
 
         let res = self.state_provider.basic_account(address)?;
-        self.caches.account_cache.insert(*address, res);
+
+        if self.is_prewarm() {
+            self.caches.account_cache.insert(*address, res);
+        }
         Ok(res)
     }
 }
@@ -148,15 +174,19 @@ impl<S: StateProvider> StateProvider for CachedStateProvider<S> {
         match self.caches.get_storage(&account, &storage_key) {
             (SlotStatus::NotCached, maybe_cache) => {
                 let final_res = self.state_provider.storage(account, storage_key)?;
-                let account_cache = maybe_cache.unwrap_or_default();
-                account_cache.insert_storage(storage_key, final_res);
-                // we always need to insert the value to update the weights.
-                // Note: there exists a race when the storage cache did not exist yet and two
-                // consumers looking up the a storage value for this account for the first time,
-                // however we can assume that this will only happen for the very first (mostlikely
-                // the same) value, and don't expect that this will accidentally
-                // replace an account storage cache with additional values.
-                self.caches.insert_storage_cache(account, account_cache);
+
+                if self.is_prewarm() {
+                    let account_cache = maybe_cache.unwrap_or_default();
+                    account_cache.insert_storage(storage_key, final_res);
+                    // we always need to insert the value to update the weights.
+                    // Note: there exists a race when the storage cache did not exist yet and two
+                    // consumers looking up the a storage value for this account for the first time,
+                    // however we can assume that this will only happen for the very first
+                    // (mostlikely the same) value, and don't expect that this
+                    // will accidentally replace an account storage cache with
+                    // additional values.
+                    self.caches.insert_storage_cache(account, account_cache);
+                }
 
                 self.metrics.storage_cache_misses.increment(1);
                 Ok(final_res)
@@ -183,7 +213,11 @@ impl<S: BytecodeReader> BytecodeReader for CachedStateProvider<S> {
         self.metrics.code_cache_misses.increment(1);
 
         let final_res = self.state_provider.bytecode_by_hash(code_hash)?;
-        self.caches.code_cache.insert(*code_hash, final_res.clone());
+
+        if self.is_prewarm() {
+            self.caches.code_cache.insert(*code_hash, final_res.clone());
+        }
+
         Ok(final_res)
     }
 }
@@ -595,6 +629,11 @@ impl SavedCache {
         Arc::strong_count(&self.usage_guard) == 1
     }
 
+    /// Returns the current strong count of the usage guard.
+    pub(crate) fn usage_count(&self) -> usize {
+        Arc::strong_count(&self.usage_guard)
+    }
+
     /// Returns the [`ExecutionCache`] belonging to the tracked hash.
     pub(crate) const fn cache(&self) -> &ExecutionCache {
         &self.caches
@@ -785,7 +824,7 @@ mod tests {
 
         let caches = ExecutionCacheBuilder::default().build_caches(1000);
         let state_provider =
-            CachedStateProvider::new_with_caches(provider, caches, CachedStateMetrics::zeroed());
+            CachedStateProvider::new(provider, caches, CachedStateMetrics::zeroed());
 
         // check that the storage is empty
         let res = state_provider.storage(address, storage_key);
@@ -808,7 +847,7 @@ mod tests {
 
         let caches = ExecutionCacheBuilder::default().build_caches(1000);
         let state_provider =
-            CachedStateProvider::new_with_caches(provider, caches, CachedStateMetrics::zeroed());
+            CachedStateProvider::new(provider, caches, CachedStateMetrics::zeroed());
 
         // check that the storage returns the expected value
         let res = state_provider.storage(address, storage_key);

crates/engine/tree/src/tree/error.rs

@@ -6,15 +6,13 @@ use reth_errors::{BlockExecutionError, BlockValidationError, ProviderError};
 use reth_evm::execute::InternalBlockExecutionError;
 use reth_payload_primitives::NewPayloadError;
 use reth_primitives_traits::{Block, BlockBody, SealedBlock};
-use tokio::sync::oneshot::error::TryRecvError;
 
-/// This is an error that can come from advancing persistence. Either this can be a
-/// [`TryRecvError`], or this can be a [`ProviderError`]
+/// This is an error that can come from advancing persistence.
 #[derive(Debug, thiserror::Error)]
 pub enum AdvancePersistenceError {
-    /// An error that can be from failing to receive a value from persistence
-    #[error(transparent)]
-    RecvError(#[from] TryRecvError),
+    /// The persistence channel was closed unexpectedly
+    #[error("persistence channel closed")]
+    ChannelClosed,
     /// A provider error
     #[error(transparent)]
     Provider(#[from] ProviderError),

crates/engine/tree/src/tree/instrumented_state.rs

@@ -83,7 +83,7 @@ where
 {
     /// Creates a new [`InstrumentedStateProvider`] from a state provider with the provided label
     /// for metrics.
-    pub fn from_state_provider(state_provider: S, source: &'static str) -> Self {
+    pub fn new(state_provider: S, source: &'static str) -> Self {
         Self {
             state_provider,
             metrics: StateProviderMetrics::new_with_labels(&[("source", source)]),

crates/engine/tree/src/tree/invalid_headers.rs

@@ -48,6 +48,7 @@ impl InvalidHeaderCache {
         // if we get here, the entry has been hit too many times, so we evict it
         self.headers.remove(hash);
         self.metrics.hit_evictions.increment(1);
+        self.metrics.count.set(self.headers.len() as f64);
         None
     }
 

crates/engine/tree/src/tree/metrics.rs

@@ -64,6 +64,7 @@ impl EngineApiMetrics {
         &self,
         executor: E,
         mut transactions: impl Iterator<Item = Result<impl ExecutableTx<E>, BlockExecutionError>>,
+        transaction_count: usize,
         state_hook: Box<dyn OnStateHook>,
     ) -> Result<(BlockExecutionOutput<E::Receipt>, Vec<Address>), BlockExecutionError>
     where
@@ -75,7 +76,7 @@ impl EngineApiMetrics {
         // be accessible.
         let wrapper = MeteredStateHook { metrics: self.executor.clone(), inner_hook: state_hook };
 
-        let mut senders = Vec::new();
+        let mut senders = Vec::with_capacity(transaction_count);
         let mut executor = executor.with_state_hook(Some(Box::new(wrapper)));
 
         let f = || {
@@ -320,7 +321,7 @@ impl NewPayloadStatusMetrics {
 }
 
 /// Metrics for non-execution related block validation.
-#[derive(Metrics)]
+#[derive(Metrics, Clone)]
 #[metrics(scope = "sync.block_validation")]
 pub(crate) struct BlockValidationMetrics {
     /// Total number of storage tries updated in the state root calculation
@@ -347,6 +348,14 @@ pub(crate) struct BlockValidationMetrics {
     pub(crate) post_execution_validation_duration: Histogram,
     /// Total duration of the new payload call
     pub(crate) total_duration: Histogram,
+    /// Size of `HashedPostStateSorted` (`total_len`)
+    pub(crate) hashed_post_state_size: Histogram,
+    /// Size of `TrieUpdatesSorted` (`total_len`)
+    pub(crate) trie_updates_sorted_size: Histogram,
+    /// Size of `AnchoredTrieInput` overlay `TrieUpdatesSorted` (`total_len`)
+    pub(crate) anchored_overlay_trie_updates_size: Histogram,
+    /// Size of `AnchoredTrieInput` overlay `HashedPostStateSorted` (`total_len`)
+    pub(crate) anchored_overlay_hashed_state_size: Histogram,
 }
 
 impl BlockValidationMetrics {
@@ -529,6 +538,7 @@ mod tests {
         let _result = metrics.execute_metered::<_, EmptyDB>(
             executor,
             input.clone_transactions_recovered().map(Ok::<_, BlockExecutionError>),
+            input.transaction_count(),
             state_hook,
         );
 
@@ -585,6 +595,7 @@ mod tests {
         let _result = metrics.execute_metered::<_, EmptyDB>(
             executor,
             input.clone_transactions_recovered().map(Ok::<_, BlockExecutionError>),
+            input.transaction_count(),
             state_hook,
         );
 

crates/engine/tree/src/tree/mod.rs

@@ -37,17 +37,12 @@ use reth_revm::database::StateProviderDatabase;
 use reth_stages_api::ControlFlow;
 use revm::state::EvmState;
 use state::TreeState;
-use std::{
-    fmt::Debug,
-    sync::{
-        mpsc::{Receiver, RecvError, RecvTimeoutError, Sender},
-        Arc,
-    },
-    time::Instant,
-};
+use std::{fmt::Debug, ops, sync::Arc, time::Instant};
+
+use crossbeam_channel::{Receiver, Sender};
 use tokio::sync::{
     mpsc::{unbounded_channel, UnboundedReceiver, UnboundedSender},
-    oneshot::{self, error::TryRecvError},
+    oneshot,
 };
 use tracing::*;
 
@@ -63,7 +58,6 @@ mod persistence_state;
 pub mod precompile_cache;
 #[cfg(test)]
 mod tests;
-// TODO(alexey): compare trie updates in `insert_block_inner`
 #[expect(unused)]
 mod trie_updates;
 
@@ -240,7 +234,7 @@ where
     C: ConfigureEvm<Primitives = N> + 'static,
 {
     provider: P,
-    consensus: Arc<dyn FullConsensus<N, Error = ConsensusError>>,
+    consensus: Arc<dyn FullConsensus<N>>,
     payload_validator: V,
     /// Keeps track of internals such as executed and buffered blocks.
     state: EngineApiTreeState<N>,
@@ -326,7 +320,7 @@ where
     #[expect(clippy::too_many_arguments)]
     pub fn new(
         provider: P,
-        consensus: Arc<dyn FullConsensus<N, Error = ConsensusError>>,
+        consensus: Arc<dyn FullConsensus<N>>,
         payload_validator: V,
         outgoing: UnboundedSender<EngineApiEvent<N>>,
         state: EngineApiTreeState<N>,
@@ -338,7 +332,7 @@ where
         engine_kind: EngineApiKind,
         evm_config: C,
     ) -> Self {
-        let (incoming_tx, incoming) = std::sync::mpsc::channel();
+        let (incoming_tx, incoming) = crossbeam_channel::unbounded();
 
         Self {
             provider,
@@ -368,7 +362,7 @@ where
     #[expect(clippy::complexity)]
     pub fn spawn_new(
         provider: P,
-        consensus: Arc<dyn FullConsensus<N, Error = ConsensusError>>,
+        consensus: Arc<dyn FullConsensus<N>>,
         payload_validator: V,
         persistence: PersistenceHandle<N>,
         payload_builder: PayloadBuilderHandle<T>,
@@ -423,23 +417,34 @@ where
     /// This will block the current thread and process incoming messages.
     pub fn run(mut self) {
         loop {
-            match self.try_recv_engine_message() {
-                Ok(Some(msg)) => {
+            match self.wait_for_event() {
+                LoopEvent::EngineMessage(msg) => {
                     debug!(target: "engine::tree", %msg, "received new engine message");
-                    if let Err(fatal) = self.on_engine_message(msg) {
-                        error!(target: "engine::tree", %fatal, "insert block fatal error");
+                    match self.on_engine_message(msg) {
+                        Ok(ops::ControlFlow::Break(())) => return,
+                        Ok(ops::ControlFlow::Continue(())) => {}
+                        Err(fatal) => {
+                            error!(target: "engine::tree", %fatal, "insert block fatal error");
+                            return
+                        }
+                    }
+                }
+                LoopEvent::PersistenceComplete { result, start_time } => {
+                    if let Err(err) = self.on_persistence_complete(result, start_time) {
+                        error!(target: "engine::tree", %err, "Persistence complete handling failed");
                         return
                     }
                 }
-                Ok(None) => {
-                    debug!(target: "engine::tree", "received no engine message for some time, while waiting for persistence task to complete");
-                }
-                Err(_err) => {
-                    error!(target: "engine::tree", "Engine channel disconnected");
+                LoopEvent::Disconnected => {
+                    error!(target: "engine::tree", "Channel disconnected");
                     return
                 }
             }
 
+            // Always check if we need to trigger new persistence after any event:
+            // - After engine messages: new blocks may have been inserted that exceed the
+            //   persistence threshold
+            // - After persistence completion: we can now persist more blocks if needed
             if let Err(err) = self.advance_persistence() {
                 error!(target: "engine::tree", %err, "Advancing persistence failed");
                 return
@@ -447,6 +452,47 @@ where
         }
     }
 
+    /// Blocks until the next event is ready: either an incoming engine message or a persistence
+    /// completion (if one is in progress).
+    ///
+    /// Uses biased selection to prioritize persistence completion to update in-memory state and
+    /// unblock further writes.
+    fn wait_for_event(&mut self) -> LoopEvent<T, N> {
+        // Take ownership of persistence rx if present
+        let maybe_persistence = self.persistence_state.rx.take();
+
+        if let Some((persistence_rx, start_time, action)) = maybe_persistence {
+            // Biased select prioritizes persistence completion to update in memory state and
+            // unblock further writes
+            crossbeam_channel::select_biased! {
+                recv(persistence_rx) -> result => {
+                    // Don't put it back - consumed (oneshot-like behavior)
+                    match result {
+                        Ok(value) => LoopEvent::PersistenceComplete {
+                            result: value,
+                            start_time,
+                        },
+                        Err(_) => LoopEvent::Disconnected,
+                    }
+                },
+                recv(self.incoming) -> msg => {
+                    // Put the persistence rx back - we didn't consume it
+                    self.persistence_state.rx = Some((persistence_rx, start_time, action));
+                    match msg {
+                        Ok(m) => LoopEvent::EngineMessage(m),
+                        Err(_) => LoopEvent::Disconnected,
+                    }
+                },
+            }
+        } else {
+            // No persistence in progress - just wait on incoming
+            match self.incoming.recv() {
+                Ok(m) => LoopEvent::EngineMessage(m),
+                Err(_) => LoopEvent::Disconnected,
+            }
+        }
+    }
+
     /// Invoked when previously requested blocks were downloaded.
     ///
     /// If the block count exceeds the configured batch size we're allowed to execute at once, this
@@ -821,7 +867,8 @@ where
         new_head_number: u64,
         current_head_number: u64,
     ) -> Vec<ExecutedBlock<N>> {
-        let mut old_blocks = Vec::new();
+        let mut old_blocks =
+            Vec::with_capacity((current_head_number.saturating_sub(new_head_number)) as usize);
 
         for block_num in (new_head_number + 1)..=current_head_number {
             if let Some(block_state) = self.canonical_in_memory_state.state_by_number(block_num) {
@@ -926,48 +973,6 @@ where
         Ok(())
     }
 
-    /// Determines if the given block is part of a fork by checking that these
-    /// conditions are true:
-    /// * walking back from the target hash to verify that the target hash is not part of an
-    ///   extension of the canonical chain.
-    /// * walking back from the current head to verify that the target hash is not already part of
-    ///   the canonical chain.
-    ///
-    /// The header is required as an arg, because we might be checking that the header is a fork
-    /// block before it's in the tree state and before it's in the database.
-    fn is_fork(&self, target: BlockWithParent) -> ProviderResult<bool> {
-        let target_hash = target.block.hash;
-        // verify that the given hash is not part of an extension of the canon chain.
-        let canonical_head = self.state.tree_state.canonical_head();
-        let mut current_hash;
-        let mut current_block = target;
-        loop {
-            if current_block.block.hash == canonical_head.hash {
-                return Ok(false)
-            }
-            // We already passed the canonical head
-            if current_block.block.number <= canonical_head.number {
-                break
-            }
-            current_hash = current_block.parent;
-
-            let Some(next_block) = self.sealed_header_by_hash(current_hash)? else { break };
-            current_block = next_block.block_with_parent();
-        }
-
-        // verify that the given hash is not already part of canonical chain stored in memory
-        if self.canonical_in_memory_state.header_by_hash(target_hash).is_some() {
-            return Ok(false)
-        }
-
-        // verify that the given hash is not already part of persisted canonical chain
-        if self.provider.block_number(target_hash)?.is_some() {
-            return Ok(false)
-        }
-
-        Ok(true)
-    }
-
     /// Invoked when we receive a new forkchoice update message. Calls into the blockchain tree
     /// to resolve chain forks and ensure that the Execution Layer is working with the latest valid
     /// chain.
@@ -1228,39 +1233,13 @@ where
         .with_event(TreeEvent::Download(DownloadRequest::single_block(target))))
     }
 
-    /// Attempts to receive the next engine request.
-    ///
-    /// If there's currently no persistence action in progress, this will block until a new request
-    /// is received. If there's a persistence action in progress, this will try to receive the
-    /// next request with a timeout to not block indefinitely and return `Ok(None)` if no request is
-    /// received in time.
-    ///
-    /// Returns an error if the engine channel is disconnected.
-    #[expect(clippy::type_complexity)]
-    fn try_recv_engine_message(
-        &self,
-    ) -> Result<Option<FromEngine<EngineApiRequest<T, N>, N::Block>>, RecvError> {
-        if self.persistence_state.in_progress() {
-            // try to receive the next request with a timeout to not block indefinitely
-            match self.incoming.recv_timeout(std::time::Duration::from_millis(500)) {
-                Ok(msg) => Ok(Some(msg)),
-                Err(err) => match err {
-                    RecvTimeoutError::Timeout => Ok(None),
-                    RecvTimeoutError::Disconnected => Err(RecvError),
-                },
-            }
-        } else {
-            self.incoming.recv().map(Some)
-        }
-    }
-
     /// Helper method to remove blocks and set the persistence state. This ensures we keep track of
     /// the current persistence action while we're removing blocks.
     fn remove_blocks(&mut self, new_tip_num: u64) {
         debug!(target: "engine::tree", ?new_tip_num, last_persisted_block_number=?self.persistence_state.last_persisted_block.number, "Removing blocks using persistence task");
         if new_tip_num < self.persistence_state.last_persisted_block.number {
             debug!(target: "engine::tree", ?new_tip_num, "Starting remove blocks job");
-            let (tx, rx) = oneshot::channel();
+            let (tx, rx) = crossbeam_channel::bounded(1);
             let _ = self.persistence.remove_blocks_above(new_tip_num, tx);
             self.persistence_state.start_remove(new_tip_num, rx);
         }
@@ -1282,55 +1261,23 @@ where
             .expect("Checked non-empty persisting blocks");
 
         debug!(target: "engine::tree", count=blocks_to_persist.len(), blocks = ?blocks_to_persist.iter().map(|block| block.recovered_block().num_hash()).collect::<Vec<_>>(), "Persisting blocks");
-        let (tx, rx) = oneshot::channel();
+        let (tx, rx) = crossbeam_channel::bounded(1);
         let _ = self.persistence.save_blocks(blocks_to_persist, tx);
 
         self.persistence_state.start_save(highest_num_hash, rx);
     }
 
-    /// Attempts to advance the persistence state.
+    /// Triggers new persistence actions if no persistence task is currently in progress.
     ///
-    /// If we're currently awaiting a response this will try to receive the response (non-blocking)
-    /// or send a new persistence action if necessary.
+    /// This checks if we need to remove blocks (disk reorg) or save new blocks to disk.
+    /// Persistence completion is handled separately via the `wait_for_event` method.
     fn advance_persistence(&mut self) -> Result<(), AdvancePersistenceError> {
-        if self.persistence_state.in_progress() {
-            let (mut rx, start_time, current_action) = self
-                .persistence_state
-                .rx
-                .take()
-                .expect("if a persistence task is in progress Receiver must be Some");
-            // Check if persistence has complete
-            match rx.try_recv() {
-                Ok(last_persisted_hash_num) => {
-                    self.metrics.engine.persistence_duration.record(start_time.elapsed());
-                    let Some(BlockNumHash {
-                        hash: last_persisted_block_hash,
-                        number: last_persisted_block_number,
-                    }) = last_persisted_hash_num
-                    else {
-                        // if this happened, then we persisted no blocks because we sent an
-                        // empty vec of blocks
-                        warn!(target: "engine::tree", "Persistence task completed but did not persist any blocks");
-                        return Ok(())
-                    };
-
-                    debug!(target: "engine::tree", ?last_persisted_block_hash, ?last_persisted_block_number, elapsed=?start_time.elapsed(), "Finished persisting, calling finish");
-                    self.persistence_state
-                        .finish(last_persisted_block_hash, last_persisted_block_number);
-                    self.on_new_persisted_block()?;
-                }
-                Err(TryRecvError::Closed) => return Err(TryRecvError::Closed.into()),
-                Err(TryRecvError::Empty) => {
-                    self.persistence_state.rx = Some((rx, start_time, current_action))
-                }
-            }
-        }
-
         if !self.persistence_state.in_progress() {
             if let Some(new_tip_num) = self.find_disk_reorg()? {
                 self.remove_blocks(new_tip_num)
             } else if self.should_persist() {
-                let blocks_to_persist = self.get_canonical_blocks_to_persist()?;
+                let blocks_to_persist =
+                    self.get_canonical_blocks_to_persist(PersistTarget::Threshold)?;
                 self.persist_blocks(blocks_to_persist);
             }
         }
@@ -1338,11 +1285,97 @@ where
         Ok(())
     }
 
+    /// Finishes termination by persisting all remaining blocks and signaling completion.
+    ///
+    /// This blocks until all persistence is complete. Always signals completion,
+    /// even if an error occurs.
+    fn finish_termination(
+        &mut self,
+        pending_termination: oneshot::Sender<()>,
+    ) -> Result<(), AdvancePersistenceError> {
+        trace!(target: "engine::tree", "finishing termination, persisting remaining blocks");
+        let result = self.persist_until_complete();
+        let _ = pending_termination.send(());
+        result
+    }
+
+    /// Persists all remaining blocks until none are left.
+    fn persist_until_complete(&mut self) -> Result<(), AdvancePersistenceError> {
+        loop {
+            // Wait for any in-progress persistence to complete (blocking)
+            if let Some((rx, start_time, _action)) = self.persistence_state.rx.take() {
+                let result = rx.recv().map_err(|_| AdvancePersistenceError::ChannelClosed)?;
+                self.on_persistence_complete(result, start_time)?;
+            }
+
+            let blocks_to_persist = self.get_canonical_blocks_to_persist(PersistTarget::Head)?;
+
+            if blocks_to_persist.is_empty() {
+                debug!(target: "engine::tree", "persistence complete, signaling termination");
+                return Ok(())
+            }
+
+            debug!(target: "engine::tree", count = blocks_to_persist.len(), "persisting remaining blocks before shutdown");
+            self.persist_blocks(blocks_to_persist);
+        }
+    }
+
+    /// Tries to poll for a completed persistence task (non-blocking).
+    ///
+    /// Returns `true` if a persistence task was completed, `false` otherwise.
+    #[cfg(test)]
+    pub fn try_poll_persistence(&mut self) -> Result<bool, AdvancePersistenceError> {
+        let Some((rx, start_time, action)) = self.persistence_state.rx.take() else {
+            return Ok(false);
+        };
+
+        match rx.try_recv() {
+            Ok(result) => {
+                self.on_persistence_complete(result, start_time)?;
+                Ok(true)
+            }
+            Err(crossbeam_channel::TryRecvError::Empty) => {
+                // Not ready yet, put it back
+                self.persistence_state.rx = Some((rx, start_time, action));
+                Ok(false)
+            }
+            Err(crossbeam_channel::TryRecvError::Disconnected) => {
+                Err(AdvancePersistenceError::ChannelClosed)
+            }
+        }
+    }
+
+    /// Handles a completed persistence task.
+    fn on_persistence_complete(
+        &mut self,
+        last_persisted_hash_num: Option<BlockNumHash>,
+        start_time: Instant,
+    ) -> Result<(), AdvancePersistenceError> {
+        self.metrics.engine.persistence_duration.record(start_time.elapsed());
+
+        let Some(BlockNumHash {
+            hash: last_persisted_block_hash,
+            number: last_persisted_block_number,
+        }) = last_persisted_hash_num
+        else {
+            // if this happened, then we persisted no blocks because we sent an empty vec of blocks
+            warn!(target: "engine::tree", "Persistence task completed but did not persist any blocks");
+            return Ok(())
+        };
+
+        debug!(target: "engine::tree", ?last_persisted_block_hash, ?last_persisted_block_number, elapsed=?start_time.elapsed(), "Finished persisting, calling finish");
+        self.persistence_state.finish(last_persisted_block_hash, last_persisted_block_number);
+        self.on_new_persisted_block()?;
+        Ok(())
+    }
+
     /// Handles a message from the engine.
+    ///
+    /// Returns `ControlFlow::Break(())` if the engine should terminate.
     fn on_engine_message(
         &mut self,
         msg: FromEngine<EngineApiRequest<T, N>, N::Block>,
-    ) -> Result<(), InsertBlockFatalError> {
+    ) -> Result<ops::ControlFlow<()>, InsertBlockFatalError> {
         match msg {
             FromEngine::Event(event) => match event {
                 FromOrchestrator::BackfillSyncStarted => {
@@ -1352,6 +1385,13 @@ where
                 FromOrchestrator::BackfillSyncFinished(ctrl) => {
                     self.on_backfill_sync_finished(ctrl)?;
                 }
+                FromOrchestrator::Terminate { tx } => {
+                    debug!(target: "engine::tree", "received terminate request");
+                    if let Err(err) = self.finish_termination(tx) {
+                        error!(target: "engine::tree", %err, "Termination failed");
+                    }
+                    return Ok(ops::ControlFlow::Break(()))
+                }
             },
             FromEngine::Request(request) => {
                 match request {
@@ -1359,7 +1399,7 @@ where
                         let block_num_hash = block.recovered_block().num_hash();
                         if block_num_hash.number <= self.state.tree_state.canonical_block_number() {
                             // outdated block that can be skipped
-                            return Ok(())
+                            return Ok(ops::ControlFlow::Continue(()))
                         }
 
                         debug!(target: "engine::tree", block=?block_num_hash, "inserting already executed block");
@@ -1467,7 +1507,7 @@ where
                 }
             }
         }
-        Ok(())
+        Ok(ops::ControlFlow::Continue(()))
     }
 
     /// Invoked if the backfill sync has finished to target.
@@ -1701,10 +1741,10 @@ where
     }
 
     /// Returns a batch of consecutive canonical blocks to persist in the range
-    /// `(last_persisted_number .. canonical_head - threshold]`. The expected
-    /// order is oldest -> newest.
+    /// `(last_persisted_number .. target]`. The expected order is oldest -> newest.
     fn get_canonical_blocks_to_persist(
         &self,
+        target: PersistTarget,
     ) -> Result<Vec<ExecutedBlock<N>>, AdvancePersistenceError> {
         // We will calculate the state root using the database, so we need to be sure there are no
         // changes
@@ -1715,9 +1755,12 @@ where
         let last_persisted_number = self.persistence_state.last_persisted_block.number;
         let canonical_head_number = self.state.tree_state.canonical_block_number();
 
-        // Persist only up to block buffer target
-        let target_number =
-            canonical_head_number.saturating_sub(self.config.memory_block_buffer_target());
+        let target_number = match target {
+            PersistTarget::Head => canonical_head_number,
+            PersistTarget::Threshold => {
+                canonical_head_number.saturating_sub(self.config.memory_block_buffer_target())
+            }
+        };
 
         debug!(
             target: "engine::tree",
@@ -2507,14 +2550,11 @@ where
             Ok(Some(_)) => {}
         }
 
-        // determine whether we are on a fork chain
-        let is_fork = match self.is_fork(block_id) {
-            Err(err) => {
-                let block = convert_to_block(self, input)?;
-                return Err(InsertBlockError::new(block, err.into()).into());
-            }
-            Ok(is_fork) => is_fork,
-        };
+        // determine whether we are on a fork chain by comparing the block number with the
+        // canonical head. This is a simple check that is sufficient for the event emission below.
+        // A block is considered a fork if its number is less than or equal to the canonical head,
+        // as this indicates there's already a canonical block at that height.
+        let is_fork = block_id.block.number <= self.state.tree_state.current_canonical_head.number;
 
         let ctx = TreeCtx::new(&mut self.state, &self.canonical_in_memory_state);
 
@@ -2831,6 +2871,26 @@ where
     }
 }
 
+/// Events received in the main engine loop.
+#[derive(Debug)]
+enum LoopEvent<T, N>
+where
+    N: NodePrimitives,
+    T: PayloadTypes,
+{
+    /// An engine API message was received.
+    EngineMessage(FromEngine<EngineApiRequest<T, N>, N::Block>),
+    /// A persistence task completed.
+    PersistenceComplete {
+        /// The result of the persistence operation.
+        result: Option<BlockNumHash>,
+        /// When the persistence operation started.
+        start_time: Instant,
+    },
+    /// A channel was disconnected.
+    Disconnected,
+}
+
 /// Block inclusion can be valid, accepted, or invalid. Invalid blocks are returned as an error
 /// variant.
 ///
@@ -2860,3 +2920,12 @@ pub enum InsertPayloadOk {
     /// The payload was valid and inserted into the tree.
     Inserted(BlockStatus),
 }
+
+/// Target for block persistence.
+#[derive(Debug, Clone, Copy)]
+enum PersistTarget {
+    /// Persist up to `canonical_head - memory_block_buffer_target`.
+    Threshold,
+    /// Persist all blocks up to and including the canonical head.
+    Head,
+}

crates/engine/tree/src/tree/payload_processor/bal.rs

@@ -2,16 +2,122 @@
 
 use alloy_consensus::constants::KECCAK_EMPTY;
 use alloy_eip7928::BlockAccessList;
-use alloy_primitives::{keccak256, U256};
+use alloy_primitives::{keccak256, Address, StorageKey, U256};
 use reth_primitives_traits::Account;
 use reth_provider::{AccountReader, ProviderError};
 use reth_trie::{HashedPostState, HashedStorage};
+use std::ops::Range;
+
+/// Returns the total number of storage slots (both changed and read-only) across all accounts in
+/// the BAL.
+pub fn total_slots(bal: &BlockAccessList) -> usize {
+    bal.iter().map(|account| account.storage_changes.len() + account.storage_reads.len()).sum()
+}
+
+/// Iterator over storage slots in a [`BlockAccessList`], with range-based filtering.
+///
+/// Iterates over all `(Address, StorageKey)` pairs representing both changed and read-only
+/// storage slots across all accounts in the BAL. For each account, changed slots are iterated
+/// first, followed by read-only slots. The iterator intelligently skips accounts and slots
+/// outside the specified range for efficient traversal.
+#[derive(Debug)]
+pub(crate) struct BALSlotIter<'a> {
+    bal: &'a BlockAccessList,
+    range: Range<usize>,
+    current_index: usize,
+    account_idx: usize,
+    /// Index within the current account's combined slots (changed + read-only).
+    /// If `slot_idx < storage_changes.len()`, we're in changed slots.
+    /// Otherwise, we're in read-only slots at index `slot_idx - storage_changes.len()`.
+    slot_idx: usize,
+}
+
+impl<'a> BALSlotIter<'a> {
+    /// Creates a new iterator over storage slots within the specified range.
+    pub(crate) fn new(bal: &'a BlockAccessList, range: Range<usize>) -> Self {
+        let mut iter = Self { bal, range, current_index: 0, account_idx: 0, slot_idx: 0 };
+        iter.skip_to_range_start();
+        iter
+    }
+
+    /// Skips to the first item within the range.
+    fn skip_to_range_start(&mut self) {
+        while self.account_idx < self.bal.len() {
+            let account = &self.bal[self.account_idx];
+            let slots_in_account = account.storage_changes.len() + account.storage_reads.len();
+
+            // Check if this account contains items in our range
+            let account_end = self.current_index + slots_in_account;
+
+            if account_end <= self.range.start {
+                // Entire account is before range, skip it
+                self.current_index = account_end;
+                self.account_idx += 1;
+                self.slot_idx = 0;
+            } else if self.current_index < self.range.start {
+                // Range starts somewhere in this account
+                let skip_slots = self.range.start - self.current_index;
+                self.slot_idx = skip_slots;
+                self.current_index = self.range.start;
+                break;
+            } else {
+                // We're at or past range start
+                break;
+            }
+        }
+    }
+}
+
+impl<'a> Iterator for BALSlotIter<'a> {
+    type Item = (Address, StorageKey);
+
+    fn next(&mut self) -> Option<Self::Item> {
+        // Check if we've exceeded the range
+        if self.current_index >= self.range.end {
+            return None;
+        }
+
+        // Find the next valid slot
+        while self.account_idx < self.bal.len() {
+            let account = &self.bal[self.account_idx];
+            let changed_len = account.storage_changes.len();
+            let total_len = changed_len + account.storage_reads.len();
+
+            if self.slot_idx < total_len {
+                let address = account.address;
+                let slot = if self.slot_idx < changed_len {
+                    // We're in changed slots
+                    account.storage_changes[self.slot_idx].slot
+                } else {
+                    // We're in read-only slots
+                    account.storage_reads[self.slot_idx - changed_len]
+                };
+
+                self.slot_idx += 1;
+                self.current_index += 1;
+
+                // Check if we've reached the end of range
+                if self.current_index > self.range.end {
+                    return None;
+                }
+
+                return Some((address, slot));
+            }
+
+            // Move to next account
+            self.account_idx += 1;
+            self.slot_idx = 0;
+        }
+
+        None
+    }
+}
 
 /// Converts a Block Access List into a [`HashedPostState`] by extracting the final state
 /// of modified accounts and storage slots.
-pub fn bal_to_hashed_post_state<P>(
+pub(crate) fn bal_to_hashed_post_state<P>(
     bal: &BlockAccessList,
-    provider: &P,
+    provider: P,
 ) -> Result<HashedPostState, ProviderError>
 where
     P: AccountReader,
@@ -20,7 +126,10 @@ where
 
     for account_changes in bal {
         let address = account_changes.address;
-        let hashed_address = keccak256(address);
+
+        // Always fetch the account; even if we don't need the db account to construct the final
+        // `Account`, doing this fills the cache.
+        let existing_account = provider.basic_account(&address)?;
 
         // Get the latest balance (last balance change if any)
         let balance = account_changes.balance_changes.last().map(|change| change.post_balance);
@@ -39,12 +148,14 @@ where
             None
         };
 
-        // Only fetch account from provider if we're missing any field
-        let existing_account = if balance.is_none() || nonce.is_none() || code_hash.is_none() {
-            provider.basic_account(&address)?
-        } else {
-            None
-        };
+        // If the account was only read then don't add it to the HashedPostState
+        if balance.is_none() &&
+            nonce.is_none() &&
+            code_hash.is_none() &&
+            account_changes.storage_changes.is_empty()
+        {
+            continue
+        }
 
         // Build the final account state
         let account = Account {
@@ -58,6 +169,7 @@ where
             }),
         };
 
+        let hashed_address = keccak256(address);
         hashed_state.accounts.insert(hashed_address, Some(account));
 
         // Process storage changes
@@ -75,9 +187,7 @@ where
                 }
             }
 
-            if !storage_map.storage.is_empty() {
-                hashed_state.storages.insert(hashed_address, storage_map);
-            }
+            hashed_state.storages.insert(hashed_address, storage_map);
         }
     }
 
@@ -315,4 +425,117 @@ mod tests {
         // Should have the last value
         assert_eq!(*stored_value, U256::from(300));
     }
+
+    #[test]
+    fn test_bal_slot_iter() {
+        // Create test data with multiple accounts and slots (both changed and read-only)
+        let addr1 = Address::repeat_byte(0x01);
+        let addr2 = Address::repeat_byte(0x02);
+        let addr3 = Address::repeat_byte(0x03);
+
+        // Account 1: 2 changed slots + 1 read-only = 3 total slots (indices 0, 1, 2)
+        let account1 = AccountChanges {
+            address: addr1,
+            storage_changes: vec![
+                SlotChanges {
+                    slot: StorageKey::from(U256::from(100)),
+                    changes: vec![StorageChange::new(0, B256::ZERO)],
+                },
+                SlotChanges {
+                    slot: StorageKey::from(U256::from(101)),
+                    changes: vec![StorageChange::new(0, B256::ZERO)],
+                },
+            ],
+            storage_reads: vec![StorageKey::from(U256::from(102))],
+            balance_changes: vec![],
+            nonce_changes: vec![],
+            code_changes: vec![],
+        };
+
+        // Account 2: 1 changed slot + 1 read-only = 2 total slots (indices 3, 4)
+        let account2 = AccountChanges {
+            address: addr2,
+            storage_changes: vec![SlotChanges {
+                slot: StorageKey::from(U256::from(200)),
+                changes: vec![StorageChange::new(0, B256::ZERO)],
+            }],
+            storage_reads: vec![StorageKey::from(U256::from(201))],
+            balance_changes: vec![],
+            nonce_changes: vec![],
+            code_changes: vec![],
+        };
+
+        // Account 3: 2 changed slots + 1 read-only = 3 total slots (indices 5, 6, 7)
+        let account3 = AccountChanges {
+            address: addr3,
+            storage_changes: vec![
+                SlotChanges {
+                    slot: StorageKey::from(U256::from(300)),
+                    changes: vec![StorageChange::new(0, B256::ZERO)],
+                },
+                SlotChanges {
+                    slot: StorageKey::from(U256::from(301)),
+                    changes: vec![StorageChange::new(0, B256::ZERO)],
+                },
+            ],
+            storage_reads: vec![StorageKey::from(U256::from(302))],
+            balance_changes: vec![],
+            nonce_changes: vec![],
+            code_changes: vec![],
+        };
+
+        let bal = vec![account1, account2, account3];
+
+        // Test 1: Iterate over all slots (range 0..8)
+        let items: Vec<_> = BALSlotIter::new(&bal, 0..8).collect();
+        assert_eq!(items.len(), 8);
+        // Account 1: changed slots first (100, 101), then read-only (102)
+        assert_eq!(items[0], (addr1, StorageKey::from(U256::from(100))));
+        assert_eq!(items[1], (addr1, StorageKey::from(U256::from(101))));
+        assert_eq!(items[2], (addr1, StorageKey::from(U256::from(102))));
+        // Account 2: changed slot (200), then read-only (201)
+        assert_eq!(items[3], (addr2, StorageKey::from(U256::from(200))));
+        assert_eq!(items[4], (addr2, StorageKey::from(U256::from(201))));
+        // Account 3: changed slots (300, 301), then read-only (302)
+        assert_eq!(items[5], (addr3, StorageKey::from(U256::from(300))));
+        assert_eq!(items[6], (addr3, StorageKey::from(U256::from(301))));
+        assert_eq!(items[7], (addr3, StorageKey::from(U256::from(302))));
+
+        // Test 2: Range that skips first account (range 3..6)
+        let items: Vec<_> = BALSlotIter::new(&bal, 3..6).collect();
+        assert_eq!(items.len(), 3);
+        assert_eq!(items[0], (addr2, StorageKey::from(U256::from(200))));
+        assert_eq!(items[1], (addr2, StorageKey::from(U256::from(201))));
+        assert_eq!(items[2], (addr3, StorageKey::from(U256::from(300))));
+
+        // Test 3: Range within first account (range 1..2)
+        let items: Vec<_> = BALSlotIter::new(&bal, 1..2).collect();
+        assert_eq!(items.len(), 1);
+        assert_eq!(items[0], (addr1, StorageKey::from(U256::from(101))));
+
+        // Test 4: Range spanning multiple accounts (range 2..5)
+        let items: Vec<_> = BALSlotIter::new(&bal, 2..5).collect();
+        assert_eq!(items.len(), 3);
+        // Last slot from account 1 (read-only)
+        assert_eq!(items[0], (addr1, StorageKey::from(U256::from(102))));
+        // Account 2 (changed + read-only)
+        assert_eq!(items[1], (addr2, StorageKey::from(U256::from(200))));
+        assert_eq!(items[2], (addr2, StorageKey::from(U256::from(201))));
+
+        // Test 5: Empty range
+        let items: Vec<_> = BALSlotIter::new(&bal, 5..5).collect();
+        assert_eq!(items.len(), 0);
+
+        // Test 6: Range beyond end (starts at index 6)
+        let items: Vec<_> = BALSlotIter::new(&bal, 6..100).collect();
+        assert_eq!(items.len(), 2);
+        assert_eq!(items[0], (addr3, StorageKey::from(U256::from(301))));
+        assert_eq!(items[1], (addr3, StorageKey::from(U256::from(302))));
+
+        // Test 7: Range that starts in read-only slots (index 2 is the read-only slot of account 1)
+        let items: Vec<_> = BALSlotIter::new(&bal, 2..4).collect();
+        assert_eq!(items.len(), 2);
+        assert_eq!(items[0], (addr1, StorageKey::from(U256::from(102))));
+        assert_eq!(items[1], (addr2, StorageKey::from(U256::from(200))));
+    }
 }

crates/engine/tree/src/tree/payload_processor/configured_sparse_trie.rs

@@ -1,7 +1,7 @@
 //! Configured sparse trie enum for switching between serial and parallel implementations.
 
 use alloy_primitives::B256;
-use reth_trie::{Nibbles, ProofTrieNode, TrieMasks, TrieNode};
+use reth_trie::{BranchNodeMasks, Nibbles, ProofTrieNode, TrieNode};
 use reth_trie_sparse::{
     errors::SparseTrieResult, provider::TrieNodeProvider, LeafLookup, LeafLookupError,
     SerialSparseTrie, SparseTrieInterface, SparseTrieUpdates,
@@ -44,7 +44,7 @@ impl SparseTrieInterface for ConfiguredSparseTrie {
     fn with_root(
         self,
         root: TrieNode,
-        masks: TrieMasks,
+        masks: Option<BranchNodeMasks>,
         retain_updates: bool,
     ) -> SparseTrieResult<Self> {
         match self {
@@ -75,7 +75,7 @@ impl SparseTrieInterface for ConfiguredSparseTrie {
         &mut self,
         path: Nibbles,
         node: TrieNode,
-        masks: TrieMasks,
+        masks: Option<BranchNodeMasks>,
     ) -> SparseTrieResult<()> {
         match self {
             Self::Serial(trie) => trie.reveal_node(path, node, masks),

crates/engine/tree/src/tree/payload_processor/mod.rs

@@ -3,11 +3,11 @@
 use super::precompile_cache::PrecompileCacheMap;
 use crate::tree::{
     cached_state::{
-        CachedStateMetrics, ExecutionCache as StateExecutionCache, ExecutionCacheBuilder,
-        SavedCache,
+        CachedStateMetrics, CachedStateProvider, ExecutionCache as StateExecutionCache,
+        ExecutionCacheBuilder, SavedCache,
     },
     payload_processor::{
-        prewarm::{PrewarmCacheTask, PrewarmContext, PrewarmTaskEvent},
+        prewarm::{PrewarmCacheTask, PrewarmContext, PrewarmMode, PrewarmTaskEvent},
         sparse_trie::StateRootComputeOutcome,
     },
     sparse_trie::SparseTrieTask,
@@ -23,13 +23,16 @@ use multiproof::{SparseTrieUpdate, *};
 use parking_lot::RwLock;
 use prewarm::PrewarmMetrics;
 use rayon::prelude::*;
-use reth_engine_primitives::ExecutableTxIterator;
 use reth_evm::{
     execute::{ExecutableTxFor, WithTxEnv},
-    ConfigureEvm, EvmEnvFor, OnStateHook, SpecFor, TxEnvFor,
+    ConfigureEvm, EvmEnvFor, ExecutableTxIterator, ExecutableTxTuple, OnStateHook, SpecFor,
+    TxEnvFor,
 };
+use reth_execution_types::ExecutionOutcome;
 use reth_primitives_traits::NodePrimitives;
-use reth_provider::{BlockReader, DatabaseProviderROFactory, StateProviderFactory, StateReader};
+use reth_provider::{
+    BlockReader, DatabaseProviderROFactory, StateProvider, StateProviderFactory, StateReader,
+};
 use reth_revm::{db::BundleState, state::EvmState};
 use reth_trie::{hashed_cursor::HashedCursorFactory, trie_cursor::TrieCursorFactory};
 use reth_trie_parallel::{
@@ -43,6 +46,7 @@ use reth_trie_sparse::{
 use reth_trie_sparse_parallel::{ParallelSparseTrie, ParallelismThresholds};
 use std::{
     collections::BTreeMap,
+    ops::Not,
     sync::{
         atomic::AtomicBool,
         mpsc::{self, channel},
@@ -50,7 +54,7 @@ use std::{
     },
     time::Instant,
 };
-use tracing::{debug, debug_span, error, instrument, warn, Span};
+use tracing::{debug, debug_span, instrument, warn, Span};
 
 pub mod bal;
 mod configured_sparse_trie;
@@ -92,6 +96,13 @@ pub const SPARSE_TRIE_MAX_NODES_SHRINK_CAPACITY: usize = 1_000_000;
 /// 144MB.
 pub const SPARSE_TRIE_MAX_VALUES_SHRINK_CAPACITY: usize = 1_000_000;
 
+/// Type alias for [`PayloadHandle`] returned by payload processor spawn methods.
+type IteratorPayloadHandle<Evm, I, N> = PayloadHandle<
+    WithTxEnv<TxEnvFor<Evm>, <I as ExecutableTxTuple>::Tx>,
+    <I as ExecutableTxTuple>::Error,
+    <N as NodePrimitives>::Receipt,
+>;
+
 /// Entrypoint for executing the payload.
 #[derive(Debug)]
 pub struct PayloadProcessor<Evm>
@@ -200,7 +211,6 @@ where
     ///
     /// This returns a handle to await the final state root and to interact with the tasks (e.g.
     /// canceling)
-    #[allow(clippy::type_complexity)]
     #[instrument(
         level = "debug",
         target = "engine::tree::payload_processor",
@@ -215,7 +225,7 @@ where
         multiproof_provider_factory: F,
         config: &TreeConfig,
         bal: Option<Arc<BlockAccessList>>,
-    ) -> PayloadHandle<WithTxEnv<TxEnvFor<Evm>, I::Tx>, I::Error>
+    ) -> IteratorPayloadHandle<Evm, I, N>
     where
         P: BlockReader + StateProviderFactory + StateReader + Clone + 'static,
         F: DatabaseProviderROFactory<Provider: TrieCursorFactory + HashedCursorFactory>
@@ -229,70 +239,72 @@ where
 
         let span = Span::current();
         let (to_sparse_trie, sparse_trie_rx) = channel();
-
-        // We rely on the cursor factory to provide whatever DB overlay is necessary to see a
-        // consistent view of the database, including the trie tables. Because of this there is no
-        // need for an overarching prefix set to invalidate any section of the trie tables, and so
-        // we use an empty prefix set.
-
-        // Create and spawn the storage proof task
-        let task_ctx = ProofTaskCtx::new(multiproof_provider_factory);
-        let storage_worker_count = config.storage_worker_count();
-        let account_worker_count = config.account_worker_count();
-        let proof_handle = ProofWorkerHandle::new(
-            self.executor.handle().clone(),
-            task_ctx,
-            storage_worker_count,
-            account_worker_count,
-        );
-
-        let multi_proof_task = MultiProofTask::new(
-            proof_handle.clone(),
-            to_sparse_trie,
-            config.multiproof_chunking_enabled().then_some(config.multiproof_chunk_size()),
-        );
-
-        // wire the multiproof task to the prewarm task
-        let to_multi_proof = Some(multi_proof_task.state_root_message_sender());
+        let (to_multi_proof, from_multi_proof) = crossbeam_channel::unbounded();
 
         // Handle BAL-based optimization if available
         let prewarm_handle = if let Some(bal) = bal {
-            // When BAL is present, skip spawning prewarm tasks entirely and send BAL to multiproof
-            debug!(target: "engine::tree::payload_processor", "BAL present, skipping prewarm tasks");
+            // When BAL is present, use BAL prewarming and send BAL to multiproof
+            debug!(target: "engine::tree::payload_processor", "BAL present, using BAL prewarming");
 
             // Send BAL message immediately to MultiProofTask
-            if let Some(ref sender) = to_multi_proof &&
-                let Err(err) = sender.send(MultiProofMessage::BlockAccessList(bal))
-            {
-                // In this case state root validation will simply fail
-                error!(target: "engine::tree::payload_processor", ?err, "Failed to send BAL to MultiProofTask");
-            }
+            let _ = to_multi_proof.send(MultiProofMessage::BlockAccessList(Arc::clone(&bal)));
 
-            // Spawn minimal cache-only task without prewarming
+            // Spawn with BAL prewarming
             self.spawn_caching_with(
                 env,
                 prewarm_rx,
                 transaction_count_hint,
                 provider_builder.clone(),
                 None, // Don't send proof targets when BAL is present
+                Some(bal),
             )
         } else {
-            // Normal path: spawn with full prewarming
+            // Normal path: spawn with transaction prewarming
             self.spawn_caching_with(
                 env,
                 prewarm_rx,
                 transaction_count_hint,
                 provider_builder.clone(),
-                to_multi_proof.clone(),
+                Some(to_multi_proof.clone()),
+                None,
             )
         };
 
+        // Create and spawn the storage proof task
+        let task_ctx = ProofTaskCtx::new(multiproof_provider_factory);
+        let storage_worker_count = config.storage_worker_count();
+        let account_worker_count = config.account_worker_count();
+        let v2_proofs_enabled = config.enable_proof_v2();
+        let proof_handle = ProofWorkerHandle::new(
+            self.executor.handle().clone(),
+            task_ctx,
+            storage_worker_count,
+            account_worker_count,
+            v2_proofs_enabled,
+        );
+
+        let multi_proof_task = MultiProofTask::new(
+            proof_handle.clone(),
+            to_sparse_trie,
+            config.multiproof_chunking_enabled().then_some(config.multiproof_chunk_size()),
+            to_multi_proof.clone(),
+            from_multi_proof,
+        );
+
         // spawn multi-proof task
         let parent_span = span.clone();
+        let saved_cache = prewarm_handle.saved_cache.clone();
         self.executor.spawn_blocking(move || {
             let _enter = parent_span.entered();
             // Build a state provider for the multiproof task
             let provider = provider_builder.build().expect("failed to build provider");
+            let provider = if let Some(saved_cache) = saved_cache {
+                let (cache, metrics) = saved_cache.split();
+                Box::new(CachedStateProvider::new(provider, cache, metrics))
+                    as Box<dyn StateProvider>
+            } else {
+                Box::new(provider)
+            };
             multi_proof_task.run(provider);
         });
 
@@ -303,7 +315,7 @@ where
         self.spawn_sparse_trie_task(sparse_trie_rx, proof_handle, state_root_tx);
 
         PayloadHandle {
-            to_multi_proof,
+            to_multi_proof: Some(to_multi_proof),
             prewarm_handle,
             state_root: Some(state_root_rx),
             transactions: execution_rx,
@@ -320,13 +332,14 @@ where
         env: ExecutionEnv<Evm>,
         transactions: I,
         provider_builder: StateProviderBuilder<N, P>,
-    ) -> PayloadHandle<WithTxEnv<TxEnvFor<Evm>, I::Tx>, I::Error>
+        bal: Option<Arc<BlockAccessList>>,
+    ) -> IteratorPayloadHandle<Evm, I, N>
     where
         P: BlockReader + StateProviderFactory + StateReader + Clone + 'static,
     {
         let (prewarm_rx, execution_rx, size_hint) = self.spawn_tx_iterator(transactions);
         let prewarm_handle =
-            self.spawn_caching_with(env, prewarm_rx, size_hint, provider_builder, None);
+            self.spawn_caching_with(env, prewarm_rx, size_hint, provider_builder, None, bal);
         PayloadHandle {
             to_multi_proof: None,
             prewarm_handle,
@@ -400,7 +413,8 @@ where
         transaction_count_hint: usize,
         provider_builder: StateProviderBuilder<N, P>,
         to_multi_proof: Option<CrossbeamSender<MultiProofMessage>>,
-    ) -> CacheTaskHandle
+        bal: Option<Arc<BlockAccessList>>,
+    ) -> CacheTaskHandle<N::Receipt>
     where
         P: BlockReader + StateProviderFactory + StateReader + Clone + 'static,
     {
@@ -410,20 +424,13 @@ where
             transactions = mpsc::channel().1;
         }
 
-        let (saved_cache, cache, cache_metrics) = if self.disable_state_cache {
-            (None, None, None)
-        } else {
-            let saved_cache = self.cache_for(env.parent_hash);
-            let cache = saved_cache.cache().clone();
-            let cache_metrics = saved_cache.metrics().clone();
-            (Some(saved_cache), Some(cache), Some(cache_metrics))
-        };
+        let saved_cache = self.disable_state_cache.not().then(|| self.cache_for(env.parent_hash));
 
         // configure prewarming
         let prewarm_ctx = PrewarmContext {
             env,
             evm_config: self.evm_config.clone(),
-            saved_cache,
+            saved_cache: saved_cache.clone(),
             provider: provider_builder,
             metrics: PrewarmMetrics::default(),
             terminate_execution: Arc::new(AtomicBool::new(false)),
@@ -444,11 +451,16 @@ where
         {
             let to_prewarm_task = to_prewarm_task.clone();
             self.executor.spawn_blocking(move || {
-                prewarm_task.run(transactions, to_prewarm_task);
+                let mode = if let Some(bal) = bal {
+                    PrewarmMode::BlockAccessList(bal)
+                } else {
+                    PrewarmMode::Transactions(transactions)
+                };
+                prewarm_task.run(mode, to_prewarm_task);
             });
         }
 
-        CacheTaskHandle { cache, to_prewarm_task: Some(to_prewarm_task), cache_metrics }
+        CacheTaskHandle { saved_cache, to_prewarm_task: Some(to_prewarm_task) }
     }
 
     /// Returns the cache for the given parent hash.
@@ -479,38 +491,40 @@ where
         BPF::AccountNodeProvider: TrieNodeProvider + Send + Sync,
         BPF::StorageNodeProvider: TrieNodeProvider + Send + Sync,
     {
-        // Reuse a stored SparseStateTrie, or create a new one using the desired configuration if
-        // there's none to reuse.
         let cleared_sparse_trie = Arc::clone(&self.sparse_state_trie);
-        let sparse_state_trie = cleared_sparse_trie.lock().take().unwrap_or_else(|| {
-            let default_trie = SparseTrie::blind_from(if self.disable_parallel_sparse_trie {
-                ConfiguredSparseTrie::Serial(Default::default())
-            } else {
-                ConfiguredSparseTrie::Parallel(Box::new(
-                    ParallelSparseTrie::default()
-                        .with_parallelism_thresholds(PARALLEL_SPARSE_TRIE_PARALLELISM_THRESHOLDS),
-                ))
-            });
-            ClearedSparseStateTrie::from_state_trie(
-                SparseStateTrie::new()
-                    .with_accounts_trie(default_trie.clone())
-                    .with_default_storage_trie(default_trie)
-                    .with_updates(true),
-            )
-        });
-
-        let task =
-            SparseTrieTask::<_, ConfiguredSparseTrie, ConfiguredSparseTrie>::new_with_cleared_trie(
-                sparse_trie_rx,
-                proof_worker_handle,
-                self.trie_metrics.clone(),
-                sparse_state_trie,
-            );
-
+        let disable_parallel_sparse_trie = self.disable_parallel_sparse_trie;
+        let trie_metrics = self.trie_metrics.clone();
         let span = Span::current();
+
         self.executor.spawn_blocking(move || {
             let _enter = span.entered();
 
+            // Reuse a stored SparseStateTrie, or create a new one using the desired configuration
+            // if there's none to reuse.
+            let sparse_state_trie = cleared_sparse_trie.lock().take().unwrap_or_else(|| {
+                let default_trie = SparseTrie::blind_from(if disable_parallel_sparse_trie {
+                    ConfiguredSparseTrie::Serial(Default::default())
+                } else {
+                    ConfiguredSparseTrie::Parallel(Box::new(
+                        ParallelSparseTrie::default()
+                            .with_parallelism_thresholds(PARALLEL_SPARSE_TRIE_PARALLELISM_THRESHOLDS),
+                    ))
+                });
+                ClearedSparseStateTrie::from_state_trie(
+                    SparseStateTrie::new()
+                        .with_accounts_trie(default_trie.clone())
+                        .with_default_storage_trie(default_trie)
+                        .with_updates(true),
+                )
+            });
+
+            let task = SparseTrieTask::<_, ConfiguredSparseTrie, ConfiguredSparseTrie>::new_with_cleared_trie(
+                sparse_trie_rx,
+                proof_worker_handle,
+                trie_metrics,
+                sparse_state_trie,
+            );
+
             let (result, trie) = task.run();
             // Send state root computation result
             let _ = state_root_tx.send(result);
@@ -581,12 +595,15 @@ where
 }
 
 /// Handle to all the spawned tasks.
+///
+/// Generic over `R` (receipt type) to allow sharing `Arc<ExecutionOutcome<R>>` with the
+/// caching task without cloning the expensive `BundleState`.
 #[derive(Debug)]
-pub struct PayloadHandle<Tx, Err> {
+pub struct PayloadHandle<Tx, Err, R> {
     /// Channel for evm state updates
     to_multi_proof: Option<CrossbeamSender<MultiProofMessage>>,
     // must include the receiver of the state root wired to the sparse trie
-    prewarm_handle: CacheTaskHandle,
+    prewarm_handle: CacheTaskHandle<R>,
     /// Stream of block transactions
     transactions: mpsc::Receiver<Result<Tx, Err>>,
     /// Receiver for the state root
@@ -595,7 +612,7 @@ pub struct PayloadHandle<Tx, Err> {
     _span: Span,
 }
 
-impl<Tx, Err> PayloadHandle<Tx, Err> {
+impl<Tx, Err, R: Send + Sync + 'static> PayloadHandle<Tx, Err, R> {
     /// Awaits the state root
     ///
     /// # Panics
@@ -631,12 +648,12 @@ impl<Tx, Err> PayloadHandle<Tx, Err> {
 
     /// Returns a clone of the caches used by prewarming
     pub(super) fn caches(&self) -> Option<StateExecutionCache> {
-        self.prewarm_handle.cache.clone()
+        self.prewarm_handle.saved_cache.as_ref().map(|cache| cache.cache().clone())
     }
 
     /// Returns a clone of the cache metrics used by prewarming
     pub(super) fn cache_metrics(&self) -> Option<CachedStateMetrics> {
-        self.prewarm_handle.cache_metrics.clone()
+        self.prewarm_handle.saved_cache.as_ref().map(|cache| cache.metrics().clone())
     }
 
     /// Terminates the pre-warming transaction processing.
@@ -648,9 +665,14 @@ impl<Tx, Err> PayloadHandle<Tx, Err> {
 
     /// Terminates the entire caching task.
     ///
-    /// If the [`BundleState`] is provided it will update the shared cache.
-    pub(super) fn terminate_caching(&mut self, block_output: Option<&BundleState>) {
-        self.prewarm_handle.terminate_caching(block_output)
+    /// If the [`ExecutionOutcome`] is provided it will update the shared cache using its
+    /// bundle state. Using `Arc<ExecutionOutcome>` allows sharing with the main execution
+    /// path without cloning the expensive `BundleState`.
+    pub(super) fn terminate_caching(
+        &mut self,
+        execution_outcome: Option<Arc<ExecutionOutcome<R>>>,
+    ) {
+        self.prewarm_handle.terminate_caching(execution_outcome)
     }
 
     /// Returns iterator yielding transactions from the stream.
@@ -662,17 +684,18 @@ impl<Tx, Err> PayloadHandle<Tx, Err> {
 }
 
 /// Access to the spawned [`PrewarmCacheTask`].
+///
+/// Generic over `R` (receipt type) to allow sharing `Arc<ExecutionOutcome<R>>` with the
+/// prewarm task without cloning the expensive `BundleState`.
 #[derive(Debug)]
-pub(crate) struct CacheTaskHandle {
+pub(crate) struct CacheTaskHandle<R> {
     /// The shared cache the task operates with.
-    cache: Option<StateExecutionCache>,
-    /// Metrics for the caches
-    cache_metrics: Option<CachedStateMetrics>,
+    saved_cache: Option<SavedCache>,
     /// Channel to the spawned prewarm task if any
-    to_prewarm_task: Option<std::sync::mpsc::Sender<PrewarmTaskEvent>>,
+    to_prewarm_task: Option<std::sync::mpsc::Sender<PrewarmTaskEvent<R>>>,
 }
 
-impl CacheTaskHandle {
+impl<R: Send + Sync + 'static> CacheTaskHandle<R> {
     /// Terminates the pre-warming transaction processing.
     ///
     /// Note: This does not terminate the task yet.
@@ -684,20 +707,25 @@ impl CacheTaskHandle {
 
     /// Terminates the entire pre-warming task.
     ///
-    /// If the [`BundleState`] is provided it will update the shared cache.
-    pub(super) fn terminate_caching(&mut self, block_output: Option<&BundleState>) {
+    /// If the [`ExecutionOutcome`] is provided it will update the shared cache using its
+    /// bundle state. Using `Arc<ExecutionOutcome>` avoids cloning the expensive `BundleState`.
+    pub(super) fn terminate_caching(
+        &mut self,
+        execution_outcome: Option<Arc<ExecutionOutcome<R>>>,
+    ) {
         if let Some(tx) = self.to_prewarm_task.take() {
-            // Only clone when we have an active task and a state to send
-            let event = PrewarmTaskEvent::Terminate { block_output: block_output.cloned() };
+            let event = PrewarmTaskEvent::Terminate { execution_outcome };
             let _ = tx.send(event);
         }
     }
 }
 
-impl Drop for CacheTaskHandle {
+impl<R> Drop for CacheTaskHandle<R> {
     fn drop(&mut self) {
-        // Ensure we always terminate on drop
-        self.terminate_caching(None);
+        // Ensure we always terminate on drop - send None without needing Send + Sync bounds
+        if let Some(tx) = self.to_prewarm_task.take() {
+            let _ = tx.send(PrewarmTaskEvent::Terminate { execution_outcome: None });
+        }
     }
 }
 
@@ -748,10 +776,34 @@ impl ExecutionCache {
             warn!(blocked_for=?elapsed, "Blocked waiting for execution cache mutex");
         }
 
-        cache
-            .as_ref()
-            .filter(|c| c.executed_block_hash() == parent_hash && c.is_available())
-            .cloned()
+        if let Some(c) = cache.as_ref() {
+            let cached_hash = c.executed_block_hash();
+            // Check that the cache hash matches the parent hash of the current block. It won't
+            // match in case it's a fork block.
+            let hash_matches = cached_hash == parent_hash;
+            // Check `is_available()` to ensure no other tasks (e.g., prewarming) currently hold
+            // a reference to this cache. We can only reuse it when we have exclusive access.
+            let available = c.is_available();
+            let usage_count = c.usage_count();
+
+            debug!(
+                target: "engine::caching",
+                %cached_hash,
+                %parent_hash,
+                hash_matches,
+                available,
+                usage_count,
+                "Existing cache found"
+            );
+
+            if hash_matches && available {
+                return Some(c.clone());
+            }
+        } else {
+            debug!(target: "engine::caching", %parent_hash, "No cache found");
+        }
+
+        None
     }
 
     /// Clears the tracked cache

crates/engine/tree/src/tree/payload_processor/prewarm.rs

@@ -14,25 +14,30 @@
 use crate::tree::{
     cached_state::{CachedStateProvider, SavedCache},
     payload_processor::{
-        executor::WorkloadExecutor, multiproof::MultiProofMessage,
+        bal::{total_slots, BALSlotIter},
+        executor::WorkloadExecutor,
+        multiproof::MultiProofMessage,
         ExecutionCache as PayloadExecutionCache,
     },
     precompile_cache::{CachedPrecompile, PrecompileCacheMap},
     ExecutionEnv, StateProviderBuilder,
 };
 use alloy_consensus::transaction::TxHashRef;
+use alloy_eip7928::BlockAccessList;
 use alloy_eips::Typed2718;
 use alloy_evm::Database;
 use alloy_primitives::{keccak256, map::B256Set, B256};
 use crossbeam_channel::Sender as CrossbeamSender;
 use metrics::{Counter, Gauge, Histogram};
 use reth_evm::{execute::ExecutableTxFor, ConfigureEvm, Evm, EvmFor, SpecFor};
+use reth_execution_types::ExecutionOutcome;
 use reth_metrics::Metrics;
 use reth_primitives_traits::NodePrimitives;
-use reth_provider::{BlockReader, StateProviderBox, StateProviderFactory, StateReader};
-use reth_revm::{database::StateProviderDatabase, db::BundleState, state::EvmState};
+use reth_provider::{AccountReader, BlockReader, StateProvider, StateProviderFactory, StateReader};
+use reth_revm::{database::StateProviderDatabase, state::EvmState};
 use reth_trie::MultiProofTargets;
 use std::{
+    ops::Range,
     sync::{
         atomic::{AtomicBool, Ordering},
         mpsc::{self, channel, Receiver, Sender},
@@ -42,6 +47,14 @@ use std::{
 };
 use tracing::{debug, debug_span, instrument, trace, warn, Span};
 
+/// Determines the prewarming mode: transaction-based or BAL-based.
+pub(super) enum PrewarmMode<Tx> {
+    /// Prewarm by executing transactions from a stream.
+    Transactions(Receiver<Tx>),
+    /// Prewarm by prefetching slots from a Block Access List.
+    BlockAccessList(Arc<BlockAccessList>),
+}
+
 /// A wrapper for transactions that includes their index in the block.
 #[derive(Clone)]
 struct IndexedTransaction<Tx> {
@@ -86,7 +99,7 @@ where
     /// Sender to emit evm state outcome messages, if any.
     to_multi_proof: Option<CrossbeamSender<MultiProofMessage>>,
     /// Receiver for events produced by tx execution
-    actions_rx: Receiver<PrewarmTaskEvent>,
+    actions_rx: Receiver<PrewarmTaskEvent<N::Receipt>>,
     /// Parent span for tracing
     parent_span: Span,
 }
@@ -105,7 +118,7 @@ where
         to_multi_proof: Option<CrossbeamSender<MultiProofMessage>>,
         transaction_count_hint: usize,
         max_concurrency: usize,
-    ) -> (Self, Sender<PrewarmTaskEvent>) {
+    ) -> (Self, Sender<PrewarmTaskEvent<N::Receipt>>) {
         let (actions_tx, actions_rx) = channel();
 
         trace!(
@@ -135,8 +148,11 @@ where
     /// For Optimism chains, special handling is applied to the first transaction if it's a
     /// deposit transaction (type 0x7E/126) which sets critical metadata that affects all
     /// subsequent transactions in the block.
-    fn spawn_all<Tx>(&self, pending: mpsc::Receiver<Tx>, actions_tx: Sender<PrewarmTaskEvent>)
-    where
+    fn spawn_all<Tx>(
+        &self,
+        pending: mpsc::Receiver<Tx>,
+        actions_tx: Sender<PrewarmTaskEvent<N::Receipt>>,
+    ) where
         Tx: ExecutableTxFor<Evm> + Clone + Send + 'static,
     {
         let executor = self.executor.clone();
@@ -160,12 +176,7 @@ where
             };
 
             // Initialize worker handles container
-            let mut handles = Vec::with_capacity(workers_needed);
-
-            // Only spawn initial workers as needed
-            for i in 0..workers_needed {
-                handles.push(ctx.spawn_worker(i, &executor, actions_tx.clone(), done_tx.clone()));
-            }
+            let handles = ctx.clone().spawn_workers(workers_needed, &executor, actions_tx.clone(), done_tx.clone());
 
             // Distribute transactions to workers
             let mut tx_index = 0usize;
@@ -248,7 +259,7 @@ where
     ///
     /// This method is called from `run()` only after all execution tasks are complete.
     #[instrument(level = "debug", target = "engine::tree::payload_processor::prewarm", skip_all)]
-    fn save_cache(self, state: BundleState) {
+    fn save_cache(self, execution_outcome: Arc<ExecutionOutcome<N::Receipt>>) {
         let start = Instant::now();
 
         let Self { execution_cache, ctx: PrewarmContext { env, metrics, saved_cache, .. }, .. } =
@@ -265,7 +276,8 @@ where
                 let new_cache = SavedCache::new(hash, caches, cache_metrics);
 
                 // Insert state into cache while holding the lock
-                if new_cache.cache().insert_state(&state).is_err() {
+                // Access the BundleState through the shared ExecutionOutcome
+                if new_cache.cache().insert_state(execution_outcome.state()).is_err() {
                     // Clear the cache on error to prevent having a polluted cache
                     *cached = None;
                     debug!(target: "engine::caching", "cleared execution cache on update error");
@@ -286,6 +298,86 @@ where
         }
     }
 
+    /// Runs BAL-based prewarming by spawning workers to prefetch storage slots.
+    ///
+    /// Divides the total slots across `max_concurrency` workers, each responsible for
+    /// prefetching a range of slots from the BAL.
+    #[instrument(level = "debug", target = "engine::tree::payload_processor::prewarm", skip_all)]
+    fn run_bal_prewarm(
+        &self,
+        bal: Arc<BlockAccessList>,
+        actions_tx: Sender<PrewarmTaskEvent<N::Receipt>>,
+    ) {
+        // Only prefetch if we have a cache to populate
+        if self.ctx.saved_cache.is_none() {
+            trace!(
+                target: "engine::tree::payload_processor::prewarm",
+                "Skipping BAL prewarm - no cache available"
+            );
+            let _ =
+                actions_tx.send(PrewarmTaskEvent::FinishedTxExecution { executed_transactions: 0 });
+            return;
+        }
+
+        let total_slots = total_slots(&bal);
+
+        trace!(
+            target: "engine::tree::payload_processor::prewarm",
+            total_slots,
+            max_concurrency = self.max_concurrency,
+            "Starting BAL prewarm"
+        );
+
+        if total_slots == 0 {
+            // No slots to prefetch, signal completion immediately
+            let _ =
+                actions_tx.send(PrewarmTaskEvent::FinishedTxExecution { executed_transactions: 0 });
+            return;
+        }
+
+        let (done_tx, done_rx) = mpsc::channel();
+
+        // Calculate number of workers needed (at most max_concurrency)
+        let workers_needed = total_slots.min(self.max_concurrency);
+
+        // Calculate slots per worker
+        let slots_per_worker = total_slots / workers_needed;
+        let remainder = total_slots % workers_needed;
+
+        // Spawn workers with their assigned ranges
+        for i in 0..workers_needed {
+            let start = i * slots_per_worker + i.min(remainder);
+            let extra = if i < remainder { 1 } else { 0 };
+            let end = start + slots_per_worker + extra;
+
+            self.ctx.spawn_bal_worker(
+                i,
+                &self.executor,
+                Arc::clone(&bal),
+                start..end,
+                done_tx.clone(),
+            );
+        }
+
+        // Drop our handle to done_tx so we can detect completion
+        drop(done_tx);
+
+        // Wait for all workers to complete
+        let mut completed_workers = 0;
+        while done_rx.recv().is_ok() {
+            completed_workers += 1;
+        }
+
+        trace!(
+            target: "engine::tree::payload_processor::prewarm",
+            completed_workers,
+            "All BAL prewarm workers completed"
+        );
+
+        // Signal that execution has finished
+        let _ = actions_tx.send(PrewarmTaskEvent::FinishedTxExecution { executed_transactions: 0 });
+    }
+
     /// Executes the task.
     ///
     /// This will execute the transactions until all transactions have been processed or the task
@@ -297,15 +389,24 @@ where
         name = "prewarm and caching",
         skip_all
     )]
-    pub(super) fn run(
+    pub(super) fn run<Tx>(
         self,
-        pending: mpsc::Receiver<impl ExecutableTxFor<Evm> + Clone + Send + 'static>,
-        actions_tx: Sender<PrewarmTaskEvent>,
-    ) {
-        // spawn execution tasks.
-        self.spawn_all(pending, actions_tx);
+        mode: PrewarmMode<Tx>,
+        actions_tx: Sender<PrewarmTaskEvent<N::Receipt>>,
+    ) where
+        Tx: ExecutableTxFor<Evm> + Clone + Send + 'static,
+    {
+        // Spawn execution tasks based on mode
+        match mode {
+            PrewarmMode::Transactions(pending) => {
+                self.spawn_all(pending, actions_tx);
+            }
+            PrewarmMode::BlockAccessList(bal) => {
+                self.run_bal_prewarm(bal, actions_tx);
+            }
+        }
 
-        let mut final_block_output = None;
+        let mut final_execution_outcome = None;
         let mut finished_execution = false;
         while let Ok(event) = self.actions_rx.recv() {
             match event {
@@ -318,9 +419,9 @@ where
                     // completed executing a set of transactions
                     self.send_multi_proof_targets(proof_targets);
                 }
-                PrewarmTaskEvent::Terminate { block_output } => {
+                PrewarmTaskEvent::Terminate { execution_outcome } => {
                     trace!(target: "engine::tree::payload_processor::prewarm", "Received termination signal");
-                    final_block_output = Some(block_output);
+                    final_execution_outcome = Some(execution_outcome);
 
                     if finished_execution {
                         // all tasks are done, we can exit, which will save caches and exit
@@ -334,7 +435,7 @@ where
 
                     finished_execution = true;
 
-                    if final_block_output.is_some() {
+                    if final_execution_outcome.is_some() {
                         // all tasks are done, we can exit, which will save caches and exit
                         break
                     }
@@ -344,9 +445,9 @@ where
 
         debug!(target: "engine::tree::payload_processor::prewarm", "Completed prewarm execution");
 
-        // save caches and finish
-        if let Some(Some(state)) = final_block_output {
-            self.save_cache(state);
+        // save caches and finish using the shared ExecutionOutcome
+        if let Some(Some(execution_outcome)) = final_execution_outcome {
+            self.save_cache(execution_outcome);
         }
     }
 }
@@ -388,10 +489,10 @@ where
             metrics,
             terminate_execution,
             precompile_cache_disabled,
-            mut precompile_cache_map,
+            precompile_cache_map,
         } = self;
 
-        let state_provider = match provider.build() {
+        let mut state_provider = match provider.build() {
             Ok(provider) => provider,
             Err(err) => {
                 trace!(
@@ -404,13 +505,15 @@ where
         };
 
         // Use the caches to create a new provider with caching
-        let state_provider: StateProviderBox = if let Some(saved_cache) = saved_cache {
+        if let Some(saved_cache) = saved_cache {
             let caches = saved_cache.cache().clone();
             let cache_metrics = saved_cache.metrics().clone();
-            Box::new(CachedStateProvider::new_with_caches(state_provider, caches, cache_metrics))
-        } else {
-            state_provider
-        };
+            state_provider = Box::new(
+                CachedStateProvider::new(state_provider, caches, cache_metrics)
+                    // ensure we pre-warm the cache
+                    .prewarm(),
+            );
+        }
 
         let state_provider = StateProviderDatabase::new(state_provider);
 
@@ -452,7 +555,7 @@ where
     fn transact_batch<Tx>(
         self,
         txs: mpsc::Receiver<IndexedTransaction<Tx>>,
-        sender: Sender<PrewarmTaskEvent>,
+        sender: Sender<PrewarmTaskEvent<N::Receipt>>,
         done_tx: Sender<()>,
     ) where
         Tx: ExecutableTxFor<Evm>,
@@ -529,27 +632,134 @@ where
     }
 
     /// Spawns a worker task for transaction execution and returns its sender channel.
-    fn spawn_worker<Tx>(
-        &self,
-        idx: usize,
-        executor: &WorkloadExecutor,
-        actions_tx: Sender<PrewarmTaskEvent>,
+    fn spawn_workers<Tx>(
+        self,
+        workers_needed: usize,
+        task_executor: &WorkloadExecutor,
+        actions_tx: Sender<PrewarmTaskEvent<N::Receipt>>,
         done_tx: Sender<()>,
-    ) -> mpsc::Sender<IndexedTransaction<Tx>>
+    ) -> Vec<mpsc::Sender<IndexedTransaction<Tx>>>
     where
         Tx: ExecutableTxFor<Evm> + Send + 'static,
     {
-        let (tx, rx) = mpsc::channel();
+        let mut handles = Vec::with_capacity(workers_needed);
+        let mut receivers = Vec::with_capacity(workers_needed);
+
+        for _ in 0..workers_needed {
+            let (tx, rx) = mpsc::channel();
+            handles.push(tx);
+            receivers.push(rx);
+        }
+
+        // Spawn a separate task spawning workers in parallel.
+        let executor = task_executor.clone();
+        let span = Span::current();
+        task_executor.spawn_blocking(move || {
+            let _enter = span.entered();
+            for (idx, rx) in receivers.into_iter().enumerate() {
+                let ctx = self.clone();
+                let actions_tx = actions_tx.clone();
+                let done_tx = done_tx.clone();
+                let span = debug_span!(target: "engine::tree::payload_processor::prewarm", "prewarm worker", idx);
+                executor.spawn_blocking(move || {
+                    let _enter = span.entered();
+                    ctx.transact_batch(rx, actions_tx, done_tx);
+                });
+            }
+        });
+
+        handles
+    }
+
+    /// Spawns a worker task for BAL slot prefetching.
+    ///
+    /// The worker iterates over the specified range of slots in the BAL and ensures
+    /// each slot is loaded into the cache by accessing it through the state provider.
+    fn spawn_bal_worker(
+        &self,
+        idx: usize,
+        executor: &WorkloadExecutor,
+        bal: Arc<BlockAccessList>,
+        range: Range<usize>,
+        done_tx: Sender<()>,
+    ) {
         let ctx = self.clone();
-        let span =
-            debug_span!(target: "engine::tree::payload_processor::prewarm", "prewarm worker", idx);
+        let span = debug_span!(
+            target: "engine::tree::payload_processor::prewarm",
+            "bal prewarm worker",
+            idx,
+            range_start = range.start,
+            range_end = range.end
+        );
 
         executor.spawn_blocking(move || {
             let _enter = span.entered();
-            ctx.transact_batch(rx, actions_tx, done_tx);
+            ctx.prefetch_bal_slots(bal, range, done_tx);
         });
+    }
 
-        tx
+    /// Prefetches storage slots from a BAL range into the cache.
+    ///
+    /// This iterates through the specified range of slots and accesses them via the state
+    /// provider to populate the cache.
+    #[instrument(level = "debug", target = "engine::tree::payload_processor::prewarm", skip_all)]
+    fn prefetch_bal_slots(
+        self,
+        bal: Arc<BlockAccessList>,
+        range: Range<usize>,
+        done_tx: Sender<()>,
+    ) {
+        let Self { saved_cache, provider, metrics, .. } = self;
+
+        // Build state provider
+        let state_provider = match provider.build() {
+            Ok(provider) => provider,
+            Err(err) => {
+                trace!(
+                    target: "engine::tree::payload_processor::prewarm",
+                    %err,
+                    "Failed to build state provider in BAL prewarm thread"
+                );
+                let _ = done_tx.send(());
+                return;
+            }
+        };
+
+        // Wrap with cache (guaranteed to be Some since run_bal_prewarm checks)
+        let saved_cache = saved_cache.expect("BAL prewarm should only run with cache");
+        let caches = saved_cache.cache().clone();
+        let cache_metrics = saved_cache.metrics().clone();
+        let state_provider = CachedStateProvider::new(state_provider, caches, cache_metrics);
+
+        let start = Instant::now();
+
+        // Track last seen address to avoid fetching the same account multiple times.
+        let mut last_address = None;
+
+        // Iterate through the assigned range of slots
+        for (address, slot) in BALSlotIter::new(&bal, range.clone()) {
+            // Fetch the account if this is a different address than the last one
+            if last_address != Some(address) {
+                let _ = state_provider.basic_account(&address);
+                last_address = Some(address);
+            }
+
+            // Access the slot to populate the cache
+            let _ = state_provider.storage(address, slot);
+        }
+
+        let elapsed = start.elapsed();
+
+        trace!(
+            target: "engine::tree::payload_processor::prewarm",
+            ?range,
+            elapsed_ms = elapsed.as_millis(),
+            "BAL prewarm worker completed"
+        );
+
+        // Signal completion
+        let _ = done_tx.send(());
+        metrics.bal_slot_iteration_duration.record(elapsed.as_secs_f64());
     }
 }
 
@@ -589,14 +799,18 @@ fn multiproof_targets_from_state(state: EvmState) -> (MultiProofTargets, usize)
 }
 
 /// The events the pre-warm task can handle.
-pub(super) enum PrewarmTaskEvent {
+///
+/// Generic over `R` (receipt type) to allow sharing `Arc<ExecutionOutcome<R>>` with the main
+/// execution path without cloning the expensive `BundleState`.
+pub(super) enum PrewarmTaskEvent<R> {
     /// Forcefully terminate all remaining transaction execution.
     TerminateTransactionExecution,
     /// Forcefully terminate the task on demand and update the shared cache with the given output
     /// before exiting.
     Terminate {
-        /// The final block state output.
-        block_output: Option<BundleState>,
+        /// The final execution outcome. Using `Arc` allows sharing with the main execution
+        /// path without cloning the expensive `BundleState`.
+        execution_outcome: Option<Arc<ExecutionOutcome<R>>>,
     },
     /// The outcome of a pre-warm task
     Outcome {
@@ -628,4 +842,6 @@ pub(crate) struct PrewarmMetrics {
     pub(crate) cache_saving_duration: Gauge,
     /// Counter for transaction execution errors during prewarming
     pub(crate) transaction_errors: Counter,
+    /// A histogram of BAL slot iteration duration during prefetching
+    pub(crate) bal_slot_iteration_duration: Histogram,
 }

crates/engine/tree/src/tree/payload_processor/sparse_trie.rs

@@ -121,8 +121,9 @@ where
                 ParallelStateRootError::Other(format!("could not calculate state root: {e:?}"))
             })?;
 
-        self.metrics.sparse_trie_final_update_duration_histogram.record(start.elapsed());
-        self.metrics.sparse_trie_total_duration_histogram.record(now.elapsed());
+        let end = Instant::now();
+        self.metrics.sparse_trie_final_update_duration_histogram.record(end.duration_since(start));
+        self.metrics.sparse_trie_total_duration_histogram.record(end.duration_since(now));
 
         Ok(StateRootComputeOutcome { state_root, trie_updates })
     }
@@ -166,15 +167,14 @@ where
 
     // Update storage slots with new values and calculate storage roots.
     let span = tracing::Span::current();
-    let (tx, rx) = mpsc::channel();
-    state
+    let results: Vec<_> = state
         .storages
         .into_iter()
         .map(|(address, storage)| (address, storage, trie.take_storage_trie(&address)))
         .par_bridge()
         .map(|(address, storage, storage_trie)| {
             let _enter =
-                debug_span!(target: "engine::tree::payload_processor::sparse_trie", parent: span.clone(), "storage trie", ?address)
+                debug_span!(target: "engine::tree::payload_processor::sparse_trie", parent: &span, "storage trie", ?address)
                     .entered();
 
             trace!(target: "engine::tree::payload_processor::sparse_trie", "Updating storage");
@@ -217,13 +217,7 @@ where
 
             SparseStateTrieResult::Ok((address, storage_trie))
         })
-        .for_each_init(
-            || tx.clone(),
-            |tx, result| {
-                let _ = tx.send(result);
-            },
-        );
-    drop(tx);
+        .collect();
 
     // Defer leaf removals until after updates/additions, so that we don't delete an intermediate
     // branch node during a removal and then re-add that branch back during a later leaf addition.
@@ -235,7 +229,7 @@ where
     let _enter =
         tracing::debug_span!(target: "engine::tree::payload_processor::sparse_trie", "account trie")
             .entered();
-    for result in rx {
+    for result in results {
         let (address, storage_trie) = result?;
         trie.insert_storage_trie(address, storage_trie);
 

crates/engine/tree/src/tree/payload_validator.rs

@@ -34,13 +34,13 @@ use reth_primitives_traits::{
     SealedHeader, SignerRecoverable,
 };
 use reth_provider::{
-    providers::OverlayStateProviderFactory, BlockExecutionOutput, BlockReader,
-    DatabaseProviderFactory, ExecutionOutcome, HashedPostStateProvider, ProviderError,
-    PruneCheckpointReader, StageCheckpointReader, StateProvider, StateProviderFactory, StateReader,
-    StateRootProvider, TrieReader,
+    providers::OverlayStateProviderFactory, BlockExecutionOutput, BlockNumReader, BlockReader,
+    ChangeSetReader, DatabaseProviderFactory, DatabaseProviderROFactory, ExecutionOutcome,
+    HashedPostStateProvider, ProviderError, PruneCheckpointReader, StageCheckpointReader,
+    StateProvider, StateProviderFactory, StateReader, TrieReader,
 };
 use reth_revm::db::State;
-use reth_trie::{updates::TrieUpdates, HashedPostState, TrieInputSorted};
+use reth_trie::{updates::TrieUpdates, HashedPostState, StateRoot, TrieInputSorted};
 use reth_trie_parallel::root::{ParallelStateRoot, ParallelStateRootError};
 use revm_primitives::Address;
 use std::{
@@ -111,7 +111,7 @@ where
     /// Provider for database access.
     provider: P,
     /// Consensus implementation for validation.
-    consensus: Arc<dyn FullConsensus<Evm::Primitives, Error = ConsensusError>>,
+    consensus: Arc<dyn FullConsensus<Evm::Primitives>>,
     /// EVM configuration.
     evm_config: Evm,
     /// Configuration for the tree.
@@ -135,8 +135,15 @@ impl<N, P, Evm, V> BasicEngineValidator<P, Evm, V>
 where
     N: NodePrimitives,
     P: DatabaseProviderFactory<
-            Provider: BlockReader + TrieReader + StageCheckpointReader + PruneCheckpointReader,
+            Provider: BlockReader
+                          + TrieReader
+                          + StageCheckpointReader
+                          + PruneCheckpointReader
+                          + ChangeSetReader
+                          + BlockNumReader,
         > + BlockReader<Header = N::BlockHeader>
+        + ChangeSetReader
+        + BlockNumReader
         + StateProviderFactory
         + StateReader
         + HashedPostStateProvider
@@ -148,7 +155,7 @@ where
     #[allow(clippy::too_many_arguments)]
     pub fn new(
         provider: P,
-        consensus: Arc<dyn FullConsensus<N, Error = ConsensusError>>,
+        consensus: Arc<dyn FullConsensus<N>>,
         evm_config: Evm,
         validator: V,
         config: TreeConfig,
@@ -374,7 +381,8 @@ where
         let mut state_provider = ensure_ok!(provider_builder.build());
         drop(_enter);
 
-        // fetch parent block
+        // Fetch parent block. This goes to memory most of the time unless the parent block is
+        // beyond the in-memory buffer.
         let Some(parent_block) = ensure_ok!(self.sealed_header_by_hash(parent_hash, ctx.state()))
         else {
             return Err(InsertBlockError::new(
@@ -399,7 +407,7 @@ where
             "Decided which state root algorithm to run"
         );
 
-        // use prewarming background task
+        // Get an iterator over the transactions in the payload
         let txs = self.tx_iterator_for(&input)?;
 
         // Extract the BAL, if valid and available
@@ -424,21 +432,16 @@ where
         // Use cached state provider before executing, used in execution after prewarming threads
         // complete
         if let Some((caches, cache_metrics)) = handle.caches().zip(handle.cache_metrics()) {
-            state_provider = Box::new(CachedStateProvider::new_with_caches(
-                state_provider,
-                caches,
-                cache_metrics,
-            ));
+            state_provider =
+                Box::new(CachedStateProvider::new(state_provider, caches, cache_metrics));
         };
 
+        if self.config.state_provider_metrics() {
+            state_provider = Box::new(InstrumentedStateProvider::new(state_provider, "engine"));
+        }
+
         // Execute the block and handle any execution errors
-        let (output, senders) = match if self.config.state_provider_metrics() {
-            let state_provider =
-                InstrumentedStateProvider::from_state_provider(&state_provider, "engine");
-            self.execute_block(&state_provider, env, &input, &mut handle)
-        } else {
-            self.execute_block(&state_provider, env, &input, &mut handle)
-        } {
+        let (output, senders) = match self.execute_block(state_provider, env, &input, &mut handle) {
             Ok(output) => output,
             Err(err) => return self.handle_execution_error(input, err, &parent_block),
         };
@@ -522,7 +525,7 @@ where
             }
 
             let (root, updates) = ensure_ok_post_block!(
-                state_provider.state_root_with_updates(hashed_state.clone()),
+                self.compute_state_root_serial(block.parent_hash(), &hashed_state, ctx.state()),
                 block
             );
             (root, updates, root_time.elapsed())
@@ -552,17 +555,14 @@ where
             .into())
         }
 
-        // terminate prewarming task with good state output
-        handle.terminate_caching(Some(&output.state));
+        // Create ExecutionOutcome and wrap in Arc for sharing with both the caching task
+        // and the deferred trie task. This avoids cloning the expensive BundleState.
+        let execution_outcome = Arc::new(ExecutionOutcome::from((output, block_num_hash.number)));
 
-        Ok(self.spawn_deferred_trie_task(
-            block,
-            output,
-            block_num_hash.number,
-            &ctx,
-            hashed_state,
-            trie_output,
-        ))
+        // Terminate prewarming task with the shared execution outcome
+        handle.terminate_caching(Some(Arc::clone(&execution_outcome)));
+
+        Ok(self.spawn_deferred_trie_task(block, execution_outcome, &ctx, hashed_state, trie_output))
     }
 
     /// Return sealed block header from database or in-memory state by hash.
@@ -605,10 +605,10 @@ where
         state_provider: S,
         env: ExecutionEnv<Evm>,
         input: &BlockOrPayload<T>,
-        handle: &mut PayloadHandle<impl ExecutableTxFor<Evm>, Err>,
+        handle: &mut PayloadHandle<impl ExecutableTxFor<Evm>, Err, N::Receipt>,
     ) -> Result<(BlockExecutionOutput<N::Receipt>, Vec<Address>), InsertBlockErrorKind>
     where
-        S: StateProvider,
+        S: StateProvider + Send,
         Err: core::error::Error + Send + Sync + 'static,
         V: PayloadValidator<T, Block = N::Block>,
         T: PayloadTypes<BuiltPayload: BuiltPayload<Primitives = N>>,
@@ -617,12 +617,13 @@ where
         debug!(target: "engine::tree::payload_validator", "Executing block");
 
         let mut db = State::builder()
-            .with_database(StateProviderDatabase::new(&state_provider))
+            .with_database(StateProviderDatabase::new(state_provider))
             .with_bundle_update()
             .without_state_clear()
             .build();
 
-        let evm = self.evm_config.evm_with_env(&mut db, env.evm_env.clone());
+        let spec_id = *env.evm_env.spec_id();
+        let evm = self.evm_config.evm_with_env(&mut db, env.evm_env);
         let ctx =
             self.execution_ctx_for(input).map_err(|e| InsertBlockErrorKind::Other(Box::new(e)))?;
         let mut executor = self.evm_config.create_executor(evm, ctx);
@@ -638,7 +639,7 @@ where
                 CachedPrecompile::wrap(
                     precompile,
                     self.precompile_cache_map.cache_for_address(*address),
-                    *env.evm_env.spec_id(),
+                    spec_id,
                     Some(metrics),
                 )
             });
@@ -649,6 +650,7 @@ where
         let (output, senders) = self.metrics.execute_metered(
             executor,
             handle.iter_transactions().map(|res| res.map_err(BlockExecutionError::other)),
+            input.transaction_count(),
             state_hook,
         )?;
         let execution_finish = Instant::now();
@@ -663,8 +665,6 @@ where
     ///
     /// Returns `Ok(_)` if computed successfully.
     /// Returns `Err(_)` if error was encountered during computation.
-    /// `Err(ProviderError::ConsistentView(_))` can be safely ignored and fallback computation
-    /// should be used instead.
     #[instrument(level = "debug", target = "engine::tree::payload_validator", skip_all)]
     fn compute_state_root_parallel(
         &self,
@@ -694,6 +694,35 @@ where
         ParallelStateRoot::new(factory, prefix_sets).incremental_root_with_updates()
     }
 
+    /// Compute state root for the given hashed post state in serial.
+    fn compute_state_root_serial(
+        &self,
+        parent_hash: B256,
+        hashed_state: &HashedPostState,
+        state: &EngineApiTreeState<N>,
+    ) -> ProviderResult<(B256, TrieUpdates)> {
+        let (mut input, block_hash) = self.compute_trie_input(parent_hash, state)?;
+
+        // Extend state overlay with current block's sorted state.
+        input.prefix_sets.extend(hashed_state.construct_prefix_sets());
+        let sorted_hashed_state = hashed_state.clone_into_sorted();
+        Arc::make_mut(&mut input.state).extend_ref(&sorted_hashed_state);
+
+        let TrieInputSorted { nodes, state, .. } = input;
+        let prefix_sets = hashed_state.construct_prefix_sets();
+
+        let factory = OverlayStateProviderFactory::new(self.provider.clone())
+            .with_block_hash(Some(block_hash))
+            .with_trie_overlay(Some(nodes))
+            .with_hashed_state_overlay(Some(state));
+
+        let provider = factory.database_provider_ro()?;
+
+        Ok(StateRoot::new(&provider, &provider)
+            .with_prefix_sets(prefix_sets.freeze())
+            .root_with_updates()?)
+    }
+
     /// Validates the block after execution.
     ///
     /// This performs:
@@ -793,6 +822,7 @@ where
         PayloadHandle<
             impl ExecutableTxFor<Evm> + use<N, P, Evm, V, T>,
             impl core::error::Error + Send + Sync + 'static + use<N, P, Evm, V, T>,
+            N::Receipt,
         >,
         InsertBlockErrorKind,
     > {
@@ -838,8 +868,12 @@ where
             }
             StateRootStrategy::Parallel | StateRootStrategy::Synchronous => {
                 let start = Instant::now();
-                let handle =
-                    self.payload_processor.spawn_cache_exclusive(env, txs, provider_builder);
+                let handle = self.payload_processor.spawn_cache_exclusive(
+                    env,
+                    txs,
+                    provider_builder,
+                    block_access_list,
+                );
 
                 // Record prewarming initialization duration
                 self.metrics
@@ -1027,8 +1061,7 @@ where
     fn spawn_deferred_trie_task(
         &self,
         block: RecoveredBlock<N::Block>,
-        output: BlockExecutionOutput<N::Receipt>,
-        block_number: u64,
+        execution_outcome: Arc<ExecutionOutcome<N::Receipt>>,
         ctx: &TreeCtx<'_, N>,
         hashed_state: HashedPostState,
         trie_output: TrieUpdates,
@@ -1053,16 +1086,33 @@ where
             ancestors,
         );
         let deferred_handle_task = deferred_trie_data.clone();
-        let deferred_compute_duration =
-            self.metrics.block_validation.deferred_trie_compute_duration.clone();
+        let block_validation_metrics = self.metrics.block_validation.clone();
 
         // Spawn background task to compute trie data. Calling `wait_cloned` will compute from
         // the stored inputs and cache the result, so subsequent calls return immediately.
         let compute_trie_input_task = move || {
             let result = panic::catch_unwind(AssertUnwindSafe(|| {
                 let compute_start = Instant::now();
-                let _ = deferred_handle_task.wait_cloned();
-                deferred_compute_duration.record(compute_start.elapsed().as_secs_f64());
+                let computed = deferred_handle_task.wait_cloned();
+                block_validation_metrics
+                    .deferred_trie_compute_duration
+                    .record(compute_start.elapsed().as_secs_f64());
+
+                // Record sizes of the computed trie data
+                block_validation_metrics
+                    .hashed_post_state_size
+                    .record(computed.hashed_state.total_len() as f64);
+                block_validation_metrics
+                    .trie_updates_sorted_size
+                    .record(computed.trie_updates.total_len() as f64);
+                if let Some(anchored) = &computed.anchored_trie_input {
+                    block_validation_metrics
+                        .anchored_overlay_trie_updates_size
+                        .record(anchored.trie_input.nodes.total_len() as f64);
+                    block_validation_metrics
+                        .anchored_overlay_hashed_state_size
+                        .record(anchored.trie_input.state.total_len() as f64);
+                }
             }));
 
             if result.is_err() {
@@ -1078,7 +1128,7 @@ where
 
         ExecutedBlock::with_deferred_trie_data(
             Arc::new(block),
-            Arc::new(ExecutionOutcome::from((output, block_number))),
+            execution_outcome,
             deferred_trie_data,
         )
     }
@@ -1158,10 +1208,17 @@ pub trait EngineValidator<
 impl<N, Types, P, Evm, V> EngineValidator<Types> for BasicEngineValidator<P, Evm, V>
 where
     P: DatabaseProviderFactory<
-            Provider: BlockReader + TrieReader + StageCheckpointReader + PruneCheckpointReader,
+            Provider: BlockReader
+                          + TrieReader
+                          + StageCheckpointReader
+                          + PruneCheckpointReader
+                          + ChangeSetReader
+                          + BlockNumReader,
         > + BlockReader<Header = N::BlockHeader>
         + StateProviderFactory
         + StateReader
+        + ChangeSetReader
+        + BlockNumReader
         + HashedPostStateProvider
         + Clone
         + 'static,
@@ -1265,4 +1322,15 @@ impl<T: PayloadTypes> BlockOrPayload<T> {
         // TODO decode and return `BlockAccessList`
         None
     }
+
+    /// Returns the number of transactions in the payload or block.
+    pub fn transaction_count(&self) -> usize
+    where
+        T::ExecutionData: ExecutionPayload,
+    {
+        match self {
+            Self::Payload(payload) => payload.transaction_count(),
+            Self::Block(block) => block.transaction_count(),
+        }
+    }
 }

crates/engine/tree/src/tree/persistence_state.rs

@@ -22,12 +22,12 @@
 
 use alloy_eips::BlockNumHash;
 use alloy_primitives::B256;
+use crossbeam_channel::Receiver as CrossbeamReceiver;
 use std::time::Instant;
-use tokio::sync::oneshot;
 use tracing::trace;
 
 /// The state of the persistence task.
-#[derive(Default, Debug)]
+#[derive(Debug)]
 pub struct PersistenceState {
     /// Hash and number of the last block persisted.
     ///
@@ -36,7 +36,7 @@ pub struct PersistenceState {
     /// Receiver end of channel where the result of the persistence task will be
     /// sent when done. A None value means there's no persistence task in progress.
     pub(crate) rx:
-        Option<(oneshot::Receiver<Option<BlockNumHash>>, Instant, CurrentPersistenceAction)>,
+        Option<(CrossbeamReceiver<Option<BlockNumHash>>, Instant, CurrentPersistenceAction)>,
 }
 
 impl PersistenceState {
@@ -50,7 +50,7 @@ impl PersistenceState {
     pub(crate) fn start_remove(
         &mut self,
         new_tip_num: u64,
-        rx: oneshot::Receiver<Option<BlockNumHash>>,
+        rx: CrossbeamReceiver<Option<BlockNumHash>>,
     ) {
         self.rx =
             Some((rx, Instant::now(), CurrentPersistenceAction::RemovingBlocks { new_tip_num }));
@@ -60,7 +60,7 @@ impl PersistenceState {
     pub(crate) fn start_save(
         &mut self,
         highest: BlockNumHash,
-        rx: oneshot::Receiver<Option<BlockNumHash>>,
+        rx: CrossbeamReceiver<Option<BlockNumHash>>,
     ) {
         self.rx = Some((rx, Instant::now(), CurrentPersistenceAction::SavingBlocks { highest }));
     }

crates/engine/tree/src/tree/precompile_cache.rs

@@ -1,50 +1,58 @@
 //! Contains a precompile cache backed by `schnellru::LruMap` (LRU by length).
 
 use alloy_primitives::Bytes;
-use parking_lot::Mutex;
+use dashmap::DashMap;
+use moka::policy::EvictionPolicy;
 use reth_evm::precompiles::{DynPrecompile, Precompile, PrecompileInput};
 use revm::precompile::{PrecompileId, PrecompileOutput, PrecompileResult};
 use revm_primitives::Address;
-use schnellru::LruMap;
-use std::{
-    collections::HashMap,
-    hash::{Hash, Hasher},
-    sync::Arc,
-};
+use std::{hash::Hash, sync::Arc};
 
 /// Default max cache size for [`PrecompileCache`]
 const MAX_CACHE_SIZE: u32 = 10_000;
 
 /// Stores caches for each precompile.
 #[derive(Debug, Clone, Default)]
-pub struct PrecompileCacheMap<S>(HashMap<Address, PrecompileCache<S>>)
+pub struct PrecompileCacheMap<S>(Arc<DashMap<Address, PrecompileCache<S>>>)
 where
-    S: Eq + Hash + std::fmt::Debug + Send + Sync + Clone;
+    S: Eq + Hash + std::fmt::Debug + Send + Sync + Clone + 'static;
 
 impl<S> PrecompileCacheMap<S>
 where
     S: Eq + Hash + std::fmt::Debug + Send + Sync + Clone + 'static,
 {
-    pub(crate) fn cache_for_address(&mut self, address: Address) -> PrecompileCache<S> {
+    pub(crate) fn cache_for_address(&self, address: Address) -> PrecompileCache<S> {
+        // Try just using `.get` first to avoid acquiring a write lock.
+        if let Some(cache) = self.0.get(&address) {
+            return cache.clone();
+        }
+        // Otherwise, fallback to `.entry` and initialize the cache.
+        //
+        // This should be very rare as caches for all precompiles will be initialized as soon as
+        // first EVM is created.
         self.0.entry(address).or_default().clone()
     }
 }
 
 /// Cache for precompiles, for each input stores the result.
-///
-/// [`LruMap`] requires a mutable reference on `get` since it updates the LRU order,
-/// so we use a [`Mutex`] instead of an `RwLock`.
 #[derive(Debug, Clone)]
-pub struct PrecompileCache<S>(Arc<Mutex<LruMap<CacheKey<S>, CacheEntry>>>)
+pub struct PrecompileCache<S>(
+    moka::sync::Cache<Bytes, CacheEntry<S>, alloy_primitives::map::DefaultHashBuilder>,
+)
 where
-    S: Eq + Hash + std::fmt::Debug + Send + Sync + Clone;
+    S: Eq + Hash + std::fmt::Debug + Send + Sync + Clone + 'static;
 
 impl<S> Default for PrecompileCache<S>
 where
     S: Eq + Hash + std::fmt::Debug + Send + Sync + Clone + 'static,
 {
     fn default() -> Self {
-        Self(Arc::new(Mutex::new(LruMap::new(schnellru::ByLength::new(MAX_CACHE_SIZE)))))
+        Self(
+            moka::sync::CacheBuilder::new(MAX_CACHE_SIZE as u64)
+                .initial_capacity(MAX_CACHE_SIZE as usize)
+                .eviction_policy(EvictionPolicy::lru())
+                .build_with_hasher(Default::default()),
+        )
     }
 }
 
@@ -52,63 +60,31 @@ impl<S> PrecompileCache<S>
 where
     S: Eq + Hash + std::fmt::Debug + Send + Sync + Clone + 'static,
 {
-    fn get(&self, key: &CacheKeyRef<'_, S>) -> Option<CacheEntry> {
-        self.0.lock().get(key).cloned()
+    fn get(&self, input: &[u8], spec: S) -> Option<CacheEntry<S>> {
+        self.0.get(input).filter(|e| e.spec == spec)
     }
 
     /// Inserts the given key and value into the cache, returning the new cache size.
-    fn insert(&self, key: CacheKey<S>, value: CacheEntry) -> usize {
-        let mut cache = self.0.lock();
-        cache.insert(key, value);
-        cache.len()
-    }
-}
-
-/// Cache key, spec id and precompile call input. spec id is included in the key to account for
-/// precompile repricing across fork activations.
-#[derive(Debug, Clone, PartialEq, Eq, Hash)]
-pub struct CacheKey<S>((S, Bytes));
-
-impl<S> CacheKey<S> {
-    const fn new(spec_id: S, input: Bytes) -> Self {
-        Self((spec_id, input))
-    }
-}
-
-/// Cache key reference, used to avoid cloning the input bytes when looking up using a [`CacheKey`].
-#[derive(Debug, Clone, PartialEq, Eq)]
-pub struct CacheKeyRef<'a, S>((S, &'a [u8]));
-
-impl<'a, S> CacheKeyRef<'a, S> {
-    const fn new(spec_id: S, input: &'a [u8]) -> Self {
-        Self((spec_id, input))
-    }
-}
-
-impl<S: PartialEq> PartialEq<CacheKey<S>> for CacheKeyRef<'_, S> {
-    fn eq(&self, other: &CacheKey<S>) -> bool {
-        self.0 .0 == other.0 .0 && self.0 .1 == other.0 .1.as_ref()
-    }
-}
-
-impl<'a, S: Hash> Hash for CacheKeyRef<'a, S> {
-    fn hash<H: Hasher>(&self, state: &mut H) {
-        self.0 .0.hash(state);
-        self.0 .1.hash(state);
+    fn insert(&self, input: Bytes, value: CacheEntry<S>) -> usize {
+        self.0.insert(input, value);
+        self.0.entry_count() as usize
     }
 }
 
 /// Cache entry, precompile successful output.
 #[derive(Debug, Clone, PartialEq, Eq)]
-pub struct CacheEntry(PrecompileOutput);
+pub struct CacheEntry<S> {
+    output: PrecompileOutput,
+    spec: S,
+}
 
-impl CacheEntry {
+impl<S> CacheEntry<S> {
     const fn gas_used(&self) -> u64 {
-        self.0.gas_used
+        self.output.gas_used
     }
 
     fn to_precompile_result(&self) -> PrecompileResult {
-        Ok(self.0.clone())
+        Ok(self.output.clone())
     }
 }
 
@@ -190,9 +166,7 @@ where
     }
 
     fn call(&self, input: PrecompileInput<'_>) -> PrecompileResult {
-        let key = CacheKeyRef::new(self.spec_id.clone(), input.data);
-
-        if let Some(entry) = &self.cache.get(&key) {
+        if let Some(entry) = &self.cache.get(input.data, self.spec_id.clone()) {
             self.increment_by_one_precompile_cache_hits();
             if input.gas >= entry.gas_used() {
                 return entry.to_precompile_result()
@@ -204,8 +178,10 @@ where
 
         match &result {
             Ok(output) => {
-                let key = CacheKey::new(self.spec_id.clone(), Bytes::copy_from_slice(calldata));
-                let size = self.cache.insert(key, CacheEntry(output.clone()));
+                let size = self.cache.insert(
+                    Bytes::copy_from_slice(calldata),
+                    CacheEntry { output: output.clone(), spec: self.spec_id.clone() },
+                );
                 self.set_precompile_cache_size_metric(size as f64);
                 self.increment_by_one_precompile_cache_misses();
             }
@@ -246,31 +222,12 @@ impl CachedPrecompileMetrics {
 
 #[cfg(test)]
 mod tests {
-    use std::hash::DefaultHasher;
-
     use super::*;
     use reth_evm::{EthEvmFactory, Evm, EvmEnv, EvmFactory};
     use reth_revm::db::EmptyDB;
     use revm::{context::TxEnv, precompile::PrecompileOutput};
     use revm_primitives::hardfork::SpecId;
 
-    #[test]
-    fn test_cache_key_ref_hash() {
-        let key1 = CacheKey::new(SpecId::PRAGUE, b"test_input".into());
-        let key2 = CacheKeyRef::new(SpecId::PRAGUE, b"test_input");
-        assert!(PartialEq::eq(&key2, &key1));
-
-        let mut hasher = DefaultHasher::new();
-        key1.hash(&mut hasher);
-        let hash1 = hasher.finish();
-
-        let mut hasher = DefaultHasher::new();
-        key2.hash(&mut hasher);
-        let hash2 = hasher.finish();
-
-        assert_eq!(hash1, hash2);
-    }
-
     #[test]
     fn test_precompile_cache_basic() {
         let dyn_precompile: DynPrecompile = (|_input: PrecompileInput<'_>| -> PrecompileResult {
@@ -293,12 +250,11 @@ mod tests {
             reverted: false,
         };
 
-        let key = CacheKey::new(SpecId::PRAGUE, b"test_input".into());
-        let expected = CacheEntry(output);
-        cache.cache.insert(key, expected.clone());
+        let input = b"test_input";
+        let expected = CacheEntry { output, spec: SpecId::PRAGUE };
+        cache.cache.insert(input.into(), expected.clone());
 
-        let key = CacheKeyRef::new(SpecId::PRAGUE, b"test_input");
-        let actual = cache.cache.get(&key).unwrap();
+        let actual = cache.cache.get(input, SpecId::PRAGUE).unwrap();
 
         assert_eq!(actual, expected);
     }
@@ -312,7 +268,7 @@ mod tests {
         let address1 = Address::repeat_byte(1);
         let address2 = Address::repeat_byte(2);
 
-        let mut cache_map = PrecompileCacheMap::default();
+        let cache_map = PrecompileCacheMap::default();
 
         // create the first precompile with a specific output
         let precompile1: DynPrecompile = (PrecompileId::custom("custom"), {

crates/engine/tree/src/tree/tests.rs

@@ -4,7 +4,7 @@ use crate::{
     tree::{
         payload_validator::{BasicEngineValidator, TreeCtx, ValidationOutcome},
         persistence_state::CurrentPersistenceAction,
-        TreeConfig,
+        PersistTarget, TreeConfig,
     },
 };
 
@@ -31,7 +31,7 @@ use std::{
     collections::BTreeMap,
     str::FromStr,
     sync::{
-        mpsc::{channel, Receiver, Sender},
+        mpsc::{Receiver, Sender},
         Arc,
     },
 };
@@ -97,6 +97,7 @@ struct TestChannel<T> {
 impl<T: Send + 'static> TestChannel<T> {
     /// Creates a new test channel
     fn spawn_channel() -> (Sender<T>, Receiver<T>, TestChannelHandle) {
+        use std::sync::mpsc::channel;
         let (original_tx, original_rx) = channel();
         let (wrapped_tx, wrapped_rx) = channel();
         let (release_tx, release_rx) = channel();
@@ -143,7 +144,9 @@ struct TestHarness {
         BasicEngineValidator<MockEthProvider, MockEvmConfig, MockEngineValidator>,
         MockEvmConfig,
     >,
-    to_tree_tx: Sender<FromEngine<EngineApiRequest<EthEngineTypes, EthPrimitives>, Block>>,
+    to_tree_tx: crossbeam_channel::Sender<
+        FromEngine<EngineApiRequest<EthEngineTypes, EthPrimitives>, Block>,
+    >,
     from_tree_rx: UnboundedReceiver<EngineApiEvent>,
     blocks: Vec<ExecutedBlock>,
     action_rx: Receiver<PersistenceAction>,
@@ -153,6 +156,7 @@ struct TestHarness {
 
 impl TestHarness {
     fn new(chain_spec: Arc<ChainSpec>) -> Self {
+        use std::sync::mpsc::channel;
         let (action_tx, action_rx) = channel();
         Self::with_persistence_channel(chain_spec, action_tx, action_rx)
     }
@@ -205,7 +209,7 @@ impl TestHarness {
             engine_api_tree_state,
             canonical_in_memory_state,
             persistence_handle,
-            PersistenceState::default(),
+            PersistenceState { last_persisted_block: BlockNumHash::default(), rx: None },
             payload_builder,
             // always assume enough parallelism for tests
             TreeConfig::default().with_legacy_state_root(false).with_has_enough_parallelism(true),
@@ -285,7 +289,8 @@ impl TestHarness {
         let fcu_state = self.fcu_state(block_hash);
 
         let (tx, rx) = oneshot::channel();
-        self.tree
+        let _ = self
+            .tree
             .on_engine_message(FromEngine::Request(
                 BeaconEngineMessage::ForkchoiceUpdated {
                     state: fcu_state,
@@ -398,10 +403,8 @@ impl ValidatorTestHarness {
 
     /// Configure `PersistenceState` for specific persistence scenarios
     fn start_persistence_operation(&mut self, action: CurrentPersistenceAction) {
-        use tokio::sync::oneshot;
-
         // Create a dummy receiver for testing - it will never receive a value
-        let (_tx, rx) = oneshot::channel();
+        let (_tx, rx) = crossbeam_channel::bounded(1);
 
         match action {
             CurrentPersistenceAction::SavingBlocks { highest } => {
@@ -497,11 +500,17 @@ fn test_tree_persist_block_batch() {
     test_harness.to_tree_tx.send(FromEngine::DownloadedBlocks(blocks)).unwrap();
 
     // process the message
-    let msg = test_harness.tree.try_recv_engine_message().unwrap().unwrap();
-    test_harness.tree.on_engine_message(msg).unwrap();
+    let msg = match test_harness.tree.wait_for_event() {
+        super::LoopEvent::EngineMessage(msg) => msg,
+        other => panic!("unexpected event: {other:?}"),
+    };
+    let _ = test_harness.tree.on_engine_message(msg).unwrap();
 
     // we now should receive the other batch
-    let msg = test_harness.tree.try_recv_engine_message().unwrap().unwrap();
+    let msg = match test_harness.tree.wait_for_event() {
+        super::LoopEvent::EngineMessage(msg) => msg,
+        other => panic!("unexpected event: {other:?}"),
+    };
     match msg {
         FromEngine::DownloadedBlocks(blocks) => {
             assert_eq!(blocks.len(), tree_config.max_execute_block_batch_size());
@@ -577,7 +586,7 @@ async fn test_engine_request_during_backfill() {
         .with_backfill_state(BackfillSyncState::Active);
 
     let (tx, rx) = oneshot::channel();
-    test_harness
+    let _ = test_harness
         .tree
         .on_engine_message(FromEngine::Request(
             BeaconEngineMessage::ForkchoiceUpdated {
@@ -658,7 +667,7 @@ async fn test_holesky_payload() {
         TestHarness::new(HOLESKY.clone()).with_backfill_state(BackfillSyncState::Active);
 
     let (tx, rx) = oneshot::channel();
-    test_harness
+    let _ = test_harness
         .tree
         .on_engine_message(FromEngine::Request(
             BeaconEngineMessage::NewPayload {
@@ -752,8 +761,8 @@ async fn test_tree_state_on_new_head_reorg() {
         })
     );
 
-    // after advancing persistence, we should be at `None` for the next action
-    test_harness.tree.advance_persistence().unwrap();
+    // after polling persistence completion, we should be at `None` for the next action
+    test_harness.tree.try_poll_persistence().unwrap();
     let current_action = test_harness.tree.persistence_state.current_action().cloned();
     assert_eq!(current_action, None);
 
@@ -883,7 +892,8 @@ async fn test_get_canonical_blocks_to_persist() {
         .with_persistence_threshold(persistence_threshold)
         .with_memory_block_buffer_target(memory_block_buffer_target);
 
-    let blocks_to_persist = test_harness.tree.get_canonical_blocks_to_persist().unwrap();
+    let blocks_to_persist =
+        test_harness.tree.get_canonical_blocks_to_persist(PersistTarget::Threshold).unwrap();
 
     let expected_blocks_to_persist_length: usize =
         (canonical_head_number - memory_block_buffer_target - last_persisted_block_number)
@@ -902,7 +912,8 @@ async fn test_get_canonical_blocks_to_persist() {
 
     assert!(test_harness.tree.state.tree_state.sealed_header_by_hash(&fork_block_hash).is_some());
 
-    let blocks_to_persist = test_harness.tree.get_canonical_blocks_to_persist().unwrap();
+    let blocks_to_persist =
+        test_harness.tree.get_canonical_blocks_to_persist(PersistTarget::Threshold).unwrap();
     assert_eq!(blocks_to_persist.len(), expected_blocks_to_persist_length);
 
     // check that the fork block is not included in the blocks to persist
@@ -981,7 +992,7 @@ async fn test_engine_tree_live_sync_transition_required_blocks_requested() {
     let backfill_tip_block = main_chain[(backfill_finished_block_number - 1) as usize].clone();
     // add block to mock provider to enable persistence clean up.
     test_harness.provider.add_block(backfill_tip_block.hash(), backfill_tip_block.into_block());
-    test_harness.tree.on_engine_message(FromEngine::Event(backfill_finished)).unwrap();
+    let _ = test_harness.tree.on_engine_message(FromEngine::Event(backfill_finished)).unwrap();
 
     let event = test_harness.from_tree_rx.recv().await.unwrap();
     match event {
@@ -991,7 +1002,7 @@ async fn test_engine_tree_live_sync_transition_required_blocks_requested() {
         _ => panic!("Unexpected event: {event:#?}"),
     }
 
-    test_harness
+    let _ = test_harness
         .tree
         .on_engine_message(FromEngine::DownloadedBlocks(vec![main_chain
             .last()
@@ -1047,7 +1058,7 @@ async fn test_fcu_with_canonical_ancestor_updates_latest_block() {
 
     // Send FCU to the canonical ancestor
     let (tx, rx) = oneshot::channel();
-    test_harness
+    let _ = test_harness
         .tree
         .on_engine_message(FromEngine::Request(
             BeaconEngineMessage::ForkchoiceUpdated {
@@ -1943,4 +1954,53 @@ mod forkchoice_updated_tests {
             .unwrap();
         assert!(result.is_some(), "OpStack should handle canonical head");
     }
+
+    /// Test that engine termination persists all blocks and signals completion.
+    #[test]
+    fn test_engine_termination_with_everything_persisted() {
+        let chain_spec = MAINNET.clone();
+        let mut test_block_builder = TestBlockBuilder::eth().with_chain_spec((*chain_spec).clone());
+
+        // Create 10 blocks to persist
+        let blocks: Vec<_> = test_block_builder.get_executed_blocks(1..11).collect();
+        let canonical_tip = blocks.last().unwrap().recovered_block().number;
+        let test_harness = TestHarness::new(chain_spec).with_blocks(blocks);
+
+        // Create termination channel
+        let (terminate_tx, mut terminate_rx) = oneshot::channel();
+
+        let to_tree_tx = test_harness.to_tree_tx.clone();
+        let action_rx = test_harness.action_rx;
+
+        // Spawn tree in background thread
+        std::thread::Builder::new()
+            .name("Engine Task".to_string())
+            .spawn(|| test_harness.tree.run())
+            .unwrap();
+
+        // Send terminate request
+        to_tree_tx
+            .send(FromEngine::Event(FromOrchestrator::Terminate { tx: terminate_tx }))
+            .unwrap();
+
+        // Handle persistence actions until termination completes
+        let mut last_persisted_number = 0;
+        loop {
+            if terminate_rx.try_recv().is_ok() {
+                break;
+            }
+
+            if let Ok(PersistenceAction::SaveBlocks(saved_blocks, sender)) =
+                action_rx.recv_timeout(std::time::Duration::from_millis(100))
+            {
+                if let Some(last) = saved_blocks.last() {
+                    last_persisted_number = last.recovered_block().number;
+                }
+                sender.send(saved_blocks.last().map(|b| b.recovered_block().num_hash())).unwrap();
+            }
+        }
+
+        // Ensure we persisted right to the tip
+        assert_eq!(last_persisted_number, canonical_tip);
+    }
 }

crates/era-utils/src/history.rs

@@ -116,7 +116,7 @@ where
 /// these stages that this work has already been done. Otherwise, there might be some conflict with
 /// database integrity.
 pub fn save_stage_checkpoints<P>(
-    provider: &P,
+    provider: P,
     from: BlockNumber,
     to: BlockNumber,
     processed: u64,
@@ -170,18 +170,14 @@ where
     <P as NodePrimitivesProvider>::Primitives: NodePrimitives<BlockHeader = BH, BlockBody = BB>,
 {
     let reader = open(meta)?;
-    let iter =
-        reader
-            .iter()
-            .map(Box::new(decode)
-                as Box<dyn Fn(Result<BlockTuple, E2sError>) -> eyre::Result<(BH, BB)>>);
+    let iter = reader.iter().map(decode as fn(_) -> _);
     let iter = ProcessIter { iter, era: meta };
 
     process_iter(iter, writer, provider, hash_collector, block_numbers)
 }
 
 type ProcessInnerIter<R, BH, BB> =
-    Map<BlockTupleIterator<R>, Box<dyn Fn(Result<BlockTuple, E2sError>) -> eyre::Result<(BH, BB)>>>;
+    Map<BlockTupleIterator<R>, fn(Result<BlockTuple, E2sError>) -> eyre::Result<(BH, BB)>>;
 
 /// An iterator that wraps era file extraction. After the final item [`EraMeta::mark_as_processed`]
 /// is called to ensure proper cleanup.
@@ -309,7 +305,7 @@ where
         writer.append_header(&header, &hash)?;
 
         // Write bodies to database.
-        provider.append_block_bodies(vec![(header.number(), Some(body))])?;
+        provider.append_block_bodies(vec![(header.number(), Some(&body))])?;
 
         hash_collector.insert(hash, number)?;
     }

crates/era/src/era/types/consensus.rs

@@ -59,11 +59,36 @@
 //!     Ok(())
 //! }
 //! ```
-
 use crate::e2s::{error::E2sError, types::Entry};
 use snap::{read::FrameDecoder, write::FrameEncoder};
 use std::io::{Read, Write};
 
+/// Maximum allowed decompressed size for a signed beacon block SSZ payload.
+const MAX_DECOMPRESSED_SIGNED_BEACON_BLOCK_BYTES: usize = 256 * 1024 * 1024; // 256 MiB
+
+/// Maximum allowed decompressed size for a beacon state SSZ payload.
+const MAX_DECOMPRESSED_BEACON_STATE_BYTES: usize = 2 * 1024 * 1024 * 1024; // 2 GiB
+
+fn decompress_snappy_bounded(
+    compressed: &[u8],
+    max_decompressed_bytes: usize,
+    what: &str,
+) -> Result<Vec<u8>, E2sError> {
+    let mut decoder = FrameDecoder::new(compressed).take(max_decompressed_bytes as u64);
+    let mut decompressed = Vec::new();
+
+    Read::read_to_end(&mut decoder, &mut decompressed)
+        .map_err(|e| E2sError::SnappyDecompression(format!("Failed to decompress {what}: {e}")))?;
+
+    if decompressed.len() >= max_decompressed_bytes {
+        return Err(E2sError::SnappyDecompression(format!(
+            "Failed to decompress {what}: decompressed data exceeded limit of {max_decompressed_bytes} bytes"
+        )));
+    }
+
+    Ok(decompressed)
+}
+
 /// `CompressedSignedBeaconBlock` record type: [0x01, 0x00]
 pub const COMPRESSED_SIGNED_BEACON_BLOCK: [u8; 2] = [0x01, 0x00];
 
@@ -104,13 +129,11 @@ impl CompressedSignedBeaconBlock {
 
     /// Decompress to get the original ssz-encoded signed beacon block
     pub fn decompress(&self) -> Result<Vec<u8>, E2sError> {
-        let mut decoder = FrameDecoder::new(self.data.as_slice());
-        let mut decompressed = Vec::new();
-        Read::read_to_end(&mut decoder, &mut decompressed).map_err(|e| {
-            E2sError::SnappyDecompression(format!("Failed to decompress signed beacon block: {e}"))
-        })?;
-
-        Ok(decompressed)
+        decompress_snappy_bounded(
+            self.data.as_slice(),
+            MAX_DECOMPRESSED_SIGNED_BEACON_BLOCK_BYTES,
+            "signed beacon block",
+        )
     }
 
     /// Convert to an [`Entry`]
@@ -168,13 +191,11 @@ impl CompressedBeaconState {
 
     /// Decompress to get the original ssz-encoded beacon state
     pub fn decompress(&self) -> Result<Vec<u8>, E2sError> {
-        let mut decoder = FrameDecoder::new(self.data.as_slice());
-        let mut decompressed = Vec::new();
-        Read::read_to_end(&mut decoder, &mut decompressed).map_err(|e| {
-            E2sError::SnappyDecompression(format!("Failed to decompress beacon state: {e}"))
-        })?;
-
-        Ok(decompressed)
+        decompress_snappy_bounded(
+            self.data.as_slice(),
+            MAX_DECOMPRESSED_BEACON_STATE_BYTES,
+            "beacon state",
+        )
     }
 
     /// Convert to an [`Entry`]
@@ -260,4 +281,15 @@ mod tests {
         let result = CompressedBeaconState::from_entry(&invalid_entry);
         assert!(result.is_err());
     }
+
+    #[test]
+    fn test_bounded_decompression_rejects_oversized_output() {
+        let ssz_data = vec![42u8; 1024];
+        let compressed = CompressedBeaconState::from_ssz(&ssz_data).unwrap();
+
+        let err =
+            decompress_snappy_bounded(compressed.data.as_slice(), 100, "beacon state").unwrap_err();
+
+        assert!(format!("{err:?}").contains("exceeded limit"));
+    }
 }

crates/era/src/era1/types/execution.rs

@@ -252,7 +252,7 @@ impl CompressedBody {
             let mut encoder = FrameEncoder::new(&mut compressed);
 
             Write::write_all(&mut encoder, rlp_data).map_err(|e| {
-                E2sError::SnappyCompression(format!("Failed to compress header: {e}"))
+                E2sError::SnappyCompression(format!("Failed to compress body: {e}"))
             })?;
 
             encoder.flush().map_err(|e| {
@@ -339,7 +339,7 @@ impl CompressedReceipts {
             let mut encoder = FrameEncoder::new(&mut compressed);
 
             Write::write_all(&mut encoder, rlp_data).map_err(|e| {
-                E2sError::SnappyCompression(format!("Failed to compress header: {e}"))
+                E2sError::SnappyCompression(format!("Failed to compress receipts: {e}"))
             })?;
 
             encoder.flush().map_err(|e| {

crates/ethereum/cli/Cargo.toml

@@ -51,7 +51,12 @@ jemalloc = [
     "reth-node-metrics/jemalloc",
 ]
 jemalloc-prof = [
-    "reth-node-core/jemalloc",
+    "jemalloc",
+    "reth-node-metrics/jemalloc-prof",
+]
+jemalloc-symbols = [
+    "jemalloc-prof",
+    "reth-node-metrics/jemalloc-symbols",
 ]
 tracy-allocator = []
 
@@ -81,3 +86,5 @@ min-trace-logs = [
     "tracing/release_max_level_trace",
     "reth-node-core/min-trace-logs",
 ]
+
+edge = ["reth-cli-commands/edge"]

crates/ethereum/cli/src/app.rs

@@ -1,4 +1,8 @@
-use crate::{interface::Commands, Cli};
+use crate::{
+    interface::{Commands, NoSubCmd},
+    Cli,
+};
+use clap::Subcommand;
 use eyre::{eyre, Result};
 use reth_chainspec::{ChainSpec, EthChainSpec, Hardforks};
 use reth_cli::chainspec::ChainSpecParser;
@@ -18,20 +22,26 @@ use std::{fmt, sync::Arc};
 
 /// A wrapper around a parsed CLI that handles command execution.
 #[derive(Debug)]
-pub struct CliApp<Spec: ChainSpecParser, Ext: clap::Args + fmt::Debug, Rpc: RpcModuleValidator> {
-    cli: Cli<Spec, Ext, Rpc>,
+pub struct CliApp<
+    Spec: ChainSpecParser,
+    Ext: clap::Args + fmt::Debug,
+    Rpc: RpcModuleValidator,
+    SubCmd: Subcommand + fmt::Debug = NoSubCmd,
+> {
+    cli: Cli<Spec, Ext, Rpc, SubCmd>,
     runner: Option<CliRunner>,
     layers: Option<Layers>,
     guard: Option<FileWorkerGuard>,
 }
 
-impl<C, Ext, Rpc> CliApp<C, Ext, Rpc>
+impl<C, Ext, Rpc, SubCmd> CliApp<C, Ext, Rpc, SubCmd>
 where
     C: ChainSpecParser,
     Ext: clap::Args + fmt::Debug,
     Rpc: RpcModuleValidator,
+    SubCmd: ExtendedCommand + Subcommand + fmt::Debug,
 {
-    pub(crate) fn new(cli: Cli<C, Ext, Rpc>) -> Self {
+    pub(crate) fn new(cli: Cli<C, Ext, Rpc, SubCmd>) -> Self {
         Self { cli, runner: None, layers: Some(Layers::new()), guard: None }
     }
 
@@ -98,9 +108,9 @@ where
         self.init_tracing(&runner)?;
 
         // Install the prometheus recorder to be sure to record all metrics
-        let _ = install_prometheus_recorder();
+        install_prometheus_recorder();
 
-        run_commands_with::<C, Ext, Rpc, N>(self.cli, runner, components, launcher)
+        run_commands_with::<C, Ext, Rpc, N, SubCmd>(self.cli, runner, components, launcher)
     }
 
     /// Initializes tracing with the configured options.
@@ -117,8 +127,8 @@ where
 
 /// Run CLI commands with the provided runner, components and launcher.
 /// This is the shared implementation used by both `CliApp` and Cli methods.
-pub(crate) fn run_commands_with<C, Ext, Rpc, N>(
-    cli: Cli<C, Ext, Rpc>,
+pub(crate) fn run_commands_with<C, Ext, Rpc, N, SubCmd>(
+    cli: Cli<C, Ext, Rpc, SubCmd>,
     runner: CliRunner,
     components: impl CliComponentsBuilder<N>,
     launcher: impl AsyncFnOnce(
@@ -131,6 +141,7 @@ where
     Ext: clap::Args + fmt::Debug,
     Rpc: RpcModuleValidator,
     N: CliNodeTypes<Primitives: NodePrimitives<BlockHeader: HeaderMut>, ChainSpec: Hardforks>,
+    SubCmd: ExtendedCommand + Subcommand + fmt::Debug,
 {
     match cli.command {
         Commands::Node(command) => {
@@ -167,9 +178,19 @@ where
         #[cfg(feature = "dev")]
         Commands::TestVectors(command) => runner.run_until_ctrl_c(command.execute()),
         Commands::ReExecute(command) => runner.run_until_ctrl_c(command.execute::<N>(components)),
+        Commands::Ext(command) => command.execute(runner),
     }
 }
 
+/// A trait for extension subcommands that can be added to the CLI.
+///
+/// Consumers implement this trait for their custom subcommands to define
+/// how they should be executed.
+pub trait ExtendedCommand {
+    /// Execute the extension command with the provided CLI runner.
+    fn execute(self, runner: CliRunner) -> Result<()>;
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;

crates/ethereum/cli/src/interface.rs

@@ -37,10 +37,11 @@ pub struct Cli<
     C: ChainSpecParser = EthereumChainSpecParser,
     Ext: clap::Args + fmt::Debug = NoArgs,
     Rpc: RpcModuleValidator = DefaultRpcModuleValidator,
+    SubCmd: Subcommand + fmt::Debug = NoSubCmd,
 > {
     /// The command to run
     #[command(subcommand)]
-    pub command: Commands<C, Ext>,
+    pub command: Commands<C, Ext, SubCmd>,
 
     /// The logging configuration for the CLI.
     #[command(flatten)]
@@ -71,15 +72,18 @@ impl Cli {
     }
 }
 
-impl<C: ChainSpecParser, Ext: clap::Args + fmt::Debug, Rpc: RpcModuleValidator> Cli<C, Ext, Rpc> {
+impl<
+        C: ChainSpecParser,
+        Ext: clap::Args + fmt::Debug,
+        Rpc: RpcModuleValidator,
+        SubCmd: crate::app::ExtendedCommand + Subcommand + fmt::Debug,
+    > Cli<C, Ext, Rpc, SubCmd>
+{
     /// Configures the CLI and returns a [`CliApp`] instance.
     ///
     /// This method is used to prepare the CLI for execution by wrapping it in a
     /// [`CliApp`] that can be further configured before running.
-    pub fn configure(self) -> CliApp<C, Ext, Rpc>
-    where
-        C: ChainSpecParser<ChainSpec = ChainSpec>,
-    {
+    pub fn configure(self) -> CliApp<C, Ext, Rpc, SubCmd> {
         CliApp::new(self)
     }
 
@@ -208,10 +212,10 @@ impl<C: ChainSpecParser, Ext: clap::Args + fmt::Debug, Rpc: RpcModuleValidator>
         let _guard = self.init_tracing(&runner, Layers::new())?;
 
         // Install the prometheus recorder to be sure to record all metrics
-        let _ = install_prometheus_recorder();
+        install_prometheus_recorder();
 
         // Use the shared standalone function to avoid duplication
-        run_commands_with::<C, Ext, Rpc, N>(self, runner, components, launcher)
+        run_commands_with::<C, Ext, Rpc, N, SubCmd>(self, runner, components, launcher)
     }
 
     /// Initializes tracing with the configured options.
@@ -245,7 +249,11 @@ impl<C: ChainSpecParser, Ext: clap::Args + fmt::Debug, Rpc: RpcModuleValidator>
 
 /// Commands to be executed
 #[derive(Debug, Subcommand)]
-pub enum Commands<C: ChainSpecParser, Ext: clap::Args + fmt::Debug> {
+pub enum Commands<
+    C: ChainSpecParser,
+    Ext: clap::Args + fmt::Debug,
+    SubCmd: Subcommand + fmt::Debug = NoSubCmd,
+> {
     /// Start the node
     #[command(name = "node")]
     Node(Box<node::NodeCommand<C, Ext>>),
@@ -291,9 +299,27 @@ pub enum Commands<C: ChainSpecParser, Ext: clap::Args + fmt::Debug> {
     /// Re-execute blocks in parallel to verify historical sync correctness.
     #[command(name = "re-execute")]
     ReExecute(re_execute::Command<C>),
+    /// Extension subcommands provided by consumers.
+    #[command(flatten)]
+    Ext(SubCmd),
 }
 
-impl<C: ChainSpecParser, Ext: clap::Args + fmt::Debug> Commands<C, Ext> {
+/// A no-op subcommand type for when no extension subcommands are needed.
+///
+/// This is the default type parameter for `Commands` when consumers don't need
+/// to add custom subcommands.
+#[derive(Debug, Subcommand)]
+pub enum NoSubCmd {}
+
+impl crate::app::ExtendedCommand for NoSubCmd {
+    fn execute(self, _runner: CliRunner) -> eyre::Result<()> {
+        match self {}
+    }
+}
+
+impl<C: ChainSpecParser, Ext: clap::Args + fmt::Debug, SubCmd: Subcommand + fmt::Debug>
+    Commands<C, Ext, SubCmd>
+{
     /// Returns the underlying chain being used for commands
     pub fn chain_spec(&self) -> Option<&Arc<C::ChainSpec>> {
         match self {
@@ -313,6 +339,7 @@ impl<C: ChainSpecParser, Ext: clap::Args + fmt::Debug> Commands<C, Ext> {
             Self::Config(_) => None,
             Self::Prune(cmd) => cmd.chain_spec(),
             Self::ReExecute(cmd) => cmd.chain_spec(),
+            Self::Ext(_) => None,
         }
     }
 }
@@ -536,4 +563,100 @@ mod tests {
             _ => panic!("Expected Stage command"),
         };
     }
+
+    #[test]
+    fn test_extensible_subcommands() {
+        use crate::app::ExtendedCommand;
+        use reth_cli_runner::CliRunner;
+        use reth_rpc_server_types::DefaultRpcModuleValidator;
+        use std::sync::atomic::{AtomicBool, Ordering};
+
+        #[derive(Debug, Subcommand)]
+        enum CustomCommands {
+            /// A custom hello command
+            #[command(name = "hello")]
+            Hello {
+                /// Name to greet
+                #[arg(long)]
+                name: String,
+            },
+            /// Another custom command
+            #[command(name = "goodbye")]
+            Goodbye,
+        }
+
+        static EXECUTED: AtomicBool = AtomicBool::new(false);
+
+        impl ExtendedCommand for CustomCommands {
+            fn execute(self, _runner: CliRunner) -> eyre::Result<()> {
+                match self {
+                    Self::Hello { name } => {
+                        assert_eq!(name, "world");
+                        EXECUTED.store(true, Ordering::SeqCst);
+                        Ok(())
+                    }
+                    Self::Goodbye => Ok(()),
+                }
+            }
+        }
+
+        // Test parsing the custom "hello" command
+        let cli = Cli::<
+            EthereumChainSpecParser,
+            NoArgs,
+            DefaultRpcModuleValidator,
+            CustomCommands,
+        >::try_parse_from(["reth", "hello", "--name", "world"])
+        .unwrap();
+
+        match &cli.command {
+            Commands::Ext(CustomCommands::Hello { name }) => {
+                assert_eq!(name, "world");
+            }
+            _ => panic!("Expected Ext(Hello) command"),
+        }
+
+        // Test parsing the custom "goodbye" command
+        let cli = Cli::<
+            EthereumChainSpecParser,
+            NoArgs,
+            DefaultRpcModuleValidator,
+            CustomCommands,
+        >::try_parse_from(["reth", "goodbye"])
+        .unwrap();
+
+        match &cli.command {
+            Commands::Ext(CustomCommands::Goodbye) => {}
+            _ => panic!("Expected Ext(Goodbye) command"),
+        }
+
+        // Test that built-in commands still work alongside custom ones
+        let cli = Cli::<
+            EthereumChainSpecParser,
+            NoArgs,
+            DefaultRpcModuleValidator,
+            CustomCommands,
+        >::try_parse_from(["reth", "node"])
+        .unwrap();
+
+        match &cli.command {
+            Commands::Node(_) => {}
+            _ => panic!("Expected Node command"),
+        }
+
+        // Test executing the custom command
+        let cli = Cli::<
+            EthereumChainSpecParser,
+            NoArgs,
+            DefaultRpcModuleValidator,
+            CustomCommands,
+        >::try_parse_from(["reth", "hello", "--name", "world"])
+        .unwrap();
+
+        if let Commands::Ext(cmd) = cli.command {
+            let runner = CliRunner::try_default_runtime().unwrap();
+            cmd.execute(runner).unwrap();
+            assert!(EXECUTED.load(Ordering::SeqCst), "Custom command should have been executed");
+        }
+    }
 }