fixes memleak and refactors submission pipe a bit#425
Conversation
![devin-ai-integration[bot]](https://avatars.githubusercontent.com/in/811515?s=60&v=4){kind=small}
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
| block_number: payload.block_number, | ||
| parent_hash: payload.parent_hash, | ||
| fee_recipient: payload.fee_recipient, | ||
| bid_slot: bid_trace.slot, |
There was a problem hiding this comment.
nit: maybe just call it slot? bid.bid_slot
crates/types/src/hydration.rs
Outdated
| None => return s.can_hydrate_inner( | ||
| &FxHashMap::default(), | ||
| &FxHashMap::default(), | ||
| max_blobs_per_block, | ||
| ), |
There was a problem hiding this comment.
very much a nit: is this worth an early return?
| let (submission, maybe_tx_root) = match self.hydrate(submission_data.submission) { | ||
| Ok(v) => v, | ||
| Err(e) => { | ||
| send_submission_result(producers, &self.future_results, submission_ref, Err(e)); | ||
| error!(?e, "hydration failed after pre-check passed"); | ||
| return; | ||
| } | ||
| }; |
There was a problem hiding this comment.
🟡 Optimistic bid left in sorter when hydration fails post-sort in handle_submission
In handle_submission (crates/relay/src/auctioneer/submit_block.rs), the optimistic bid is sorted into bid_sorter at line 71 and the success result is sent to the builder at line 79. The sim request is sent at line 95. Then at line 97-101, if self.hydrate() fails, the function returns early without storing the payload or cleaning up the bid from the sorter. The bid sorter now references a block_hash with no corresponding entry in self.payloads. When get_header is called (crates/relay/src/auctioneer/get_header.rs:41-43), it finds the block_hash from the sorter but fails to look up the payload, triggering the "this should never happen" error path. While the can_hydrate pre-check in validate_submission (crates/relay/src/auctioneer/validation.rs:35-38) makes this path practically unreachable, the lack of cleanup on the error path means a subtle cache inconsistency or future code change could expose this.
Was this helpful? React with 👍 or 👎 to provide feedback.
Co-authored-by: devin-ai-integration[bot] <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: devin-ai-integration[bot] <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: devin-ai-integration[bot] <158243242+devin-ai-integration[bot]@users.noreply.github.com>
0w3n-d
force-pushed
the
od/fix_memleak_plus_refactor
branch
from
bdd30d1 to
3aa518d
Compare
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
2 participants
Uh oh!
There was an error while loading. Please reload this page.