feat(mpt): Copy-on-hash

[clabby]

Overview

Adjusts the MPT crate to copy-on-hash rather than hash-in-place when computing the root hash of the trie.

This is less efficient, but is needed for completing execution of Optimistic blocks in the interop proof. Doing this enables for us to not rely solely on the L2AccountProof hint for output root computation, which depends on the block the proof is fetched within being canonical.

Benchmarks

The slowdown isn't crazy significant, but more keccak256 operations may not be amazing for users running kona on zkVMs. Native wallclock benchmarks don't directly translate to prover cost, but do give a rough indication on the hit (Benchmark: "Compute root, fully open trie - 65,536 nodes"):

Asterisc

Interestingly, doesn't really affect asterisc cycle count nor wallclock time (but note, we only ran one block, not a range proof):

main (61ed1419) - ~4610000000 cycles.
cl/mpt/copy-when-sealing (37c8dae) - ~4630000000 cycles.

If we care about this hit, we can look into adding some additional caching that preserves the full cache of the trie in-memory, but prevents re-hashing nodes whose cache has not been invalidated by a write.

Meta

closes #981

[clabby]

📚 $\text{Stack Overview}$

Pulls submitted in this stack:

• fix(client): Interop bugfixes #1006
• feat(client): Superchain Consolidation #1004
• feat(host): Reintroduce L2BlockData hint #1003
• feat(mpt): Copy-on-hash #1001 👈
• main

This comment was automatically generated by st.

[codecov]

Codecov Report

Attention: Patch coverage is 93.93939% with 2 lines in your changes missing coverage. Please review.

Project coverage is 89.1%. Comparing base (f0a7bfe) to head (456dd2e).
Report is 2 commits behind head on main.

✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
crates/executor/src/db/mod.rs	85.7%	1 Missing ⚠️
crates/executor/src/executor/mod.rs	0.0%	1 Missing ⚠️

Additional details and impacted files

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[refcell]

IMO this hit is not significant enough to be a blocker for the interop workstream.

[clabby]

IMO this hit is not significant enough to be a blocker for the interop workstream.

Yeah, agree. @ratankaliani & @hashcashier, please report back in the issues if the additional cache is needed for your usecases.