fix: don't panic when shutdown races an in-flight publish - #2
Merged
Merged
Conversation
The CAN readers have no shutdown path — they block in recv until the process exits — so they keep calling PublishData throughout shutdown. Disconnect closed the publish queue's channel out from under them, and a reader that reached the send first took the process down with "send on closed channel". On a live bus at ~200 frames/s that window is hit reliably, so every stop panicked instead of exiting cleanly. Close under a write lock and hold the read lock across the send, the same way StopDBQueue already guards the database queue against its identical race.
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Seen on the car — every
docker compose downends in:The CAN readers have no shutdown path — they block in
recvuntil the process exits — so they keep publishing throughout shutdown.Disconnectclosed the queue channel out from under them, and a reader that reached the send first killed the process. At ~200 frames/s that window is hit reliably, so the relay panicked on every stop rather than exiting cleanly.Fix mirrors what
StopDBQueuealready does for the identical race on the database queue: close under a write lock, hold the read lock across the send.Reproduced the race directly (8 writers against a concurrent close, 200 iterations,
-race): panics reliably before, clean after. That harness was scratch and isn't committed.Note this is a clean-shutdown bug only — it does not drop frames while running, and
StopDBQueueruns beforeDisconnect, so the final batch still reaches SQLite before the panic could occur.