Move recover authority logic inside submitted lambda #1827
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
There was a problem hiding this comment.
Pull Request Overview
This PR refactors the eth_call execution flow to improve performance by moving authority recovery logic from the synchronous path into the asynchronous fiber pool execution. The key motivation is to avoid blocking fiber promises from Rust threads, which can cause performance degradation.
Key changes:
- Authority recovery logic moved inside the lambda submitted to the fiber pool
- Replaced parallel
recover_authorities()with sequentialrecover_authority()calls since eth_call transactions have a single authority - Removed unused
execute_block.hppinclude
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
maxkozlovsky
force-pushed
the
max/moverecover
branch
from
f8318c8 to
7e2b658
Compare
Chen-Yifan
previously approved these changes
Oct 24, 2025
maxkozlovsky
force-pushed
the
max/moverecover
branch
6 times, most recently
from
a1c91a7 to
4c2780b
Compare
zander-xyz
previously approved these changes
Nov 18, 2025
Since submit_eth_call_to_pool() is called from rust thread, blocking on fiber promises is not appropriate and can lead to performance issues. Move recovert_authorities() logic inside lambda executed in fiber pool. Since eth call txn has single authority, parallelization logic inside recover_authorities() only intoduces overhead without benefits. Replace parallelization with single call to recover_authority(). Modify recover_senders_and_authorities() to recover senders and authorities directly using recover_sender() and recover_authority() instead of using the pool-based monad::recover_senders() and monad::recover_authorities(). This avoids potential deadlock if all fibers in the pool are occupied, since the function is called from within a fiber pool lambda context." Refactor fiber pools to enable thread sharing between fiber groups. Introduce FiberThreadPool and FiberGroup abstractions that allow multiple fiber groups to share OS threads while maintaining separate task queues and fiber limits. Architecture: PriorityPool = FiberThreadPool + FiberGroup (isolated, dedicated threads) Shared setup = FiberThreadPool + multiple FiberGroups (shared threads) Thread 0 runs a bootstrap fiber that handles fiber creation requests to ensure fibers are created on threads with proper scheduler configuration. Use trace_tx_exec_poo to parallelize trace transaction execution to avoid deadlock because of using same pool.
maxkozlovsky
force-pushed
the
max/moverecover
branch
from
4c2780b to
5dd99a7
Compare
Chen-Yifan
approved these changes
Nov 18, 2025
zander-xyz
approved these changes
Nov 18, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Since submit_eth_call_to_pool() is called from rust thread, blocking on fiber promises is not appropriate and can lead to performance issues. Move recovert_authorities() logic inside lambda executed in fiber pool.
Since eth call txn has single authority, parallelization logic inside recover_authorities() only intoduces overhead without benefits. Replace parallelization with single call to recover_authority().