[GH-ISSUE #4386] store: UdsAcceptor::listen_fd_ is accessed without synchronization across threads #1101

New issue

Closed

opened 2026-10-29 21:22:53 +00:00 by bsakaguchi · 1 comment

bsakaguchi commented

2026-10-29 21:22:53 +00:00

Owner

Originally created by @quantz8a on GitHub (Sep 29, 2026).
Original GitHub issue: https://github.com/kvcache-ai/Mooncake/issues/4386

Background

UdsAcceptor::acceptLoop() runs on a dedicated std::jthread. It reads listen_fd_ for every accept() call and, when the loop exits, closes the descriptor and sets listen_fd_ = -1.

UdsAcceptor::stop() runs on another thread: it calls wakeAccept(), joins the accept thread, reads listen_fd_, closes it, and resets it to -1.

Concurrent unsynchronized access to the non-atomic int listen_fd_ from those two threads is a data race and undefined behavior.

Scope

mooncake-store/src/uds_transport.cpp

Proposed fix

#4384 removes the trailing close/write from acceptLoop() so the accept thread only reads listen_fd_, while stop() closes the descriptor after join().

Originally created by @quantz8a on GitHub (Sep 29, 2026). Original GitHub issue: https://github.com/kvcache-ai/Mooncake/issues/4386 ## Background `UdsAcceptor::acceptLoop()` runs on a dedicated `std::jthread`. It reads `listen_fd_` for every `accept()` call and, when the loop exits, closes the descriptor and sets `listen_fd_ = -1`. `UdsAcceptor::stop()` runs on another thread: it calls `wakeAccept()`, joins the accept thread, reads `listen_fd_`, closes it, and resets it to `-1`. Concurrent unsynchronized access to the non-atomic `int listen_fd_` from those two threads is a data race and undefined behavior. ## Scope - `mooncake-store/src/uds_transport.cpp` ## Proposed fix - #4384 removes the trailing close/write from `acceptLoop()` so the accept thread only reads `listen_fd_`, while `stop()` closes the descriptor after `join()`.

bsakaguchi closed this issue

2026-10-29 21:22:53 +00:00

bsakaguchi commented

2026-10-29 21:22:54 +00:00

Author

Owner

@github-actions[bot] commented on GitHub (Sep 29, 2026):

Thanks for opening this issue, @quantz8a!

Field	Value
Issue	#4386
GitHub user ID	`3152505`
Reporter	@quantz8a

A maintainer will triage this when possible. To help us respond faster, please include:

Mooncake version or commit SHA
Environment (OS, CUDA/driver, RDMA stack if relevant)
Steps to reproduce and expected vs. actual behavior

Useful links: Documentation · Contributing guide

This message was posted automatically by the issue bot.

@github-actions[bot] commented on GitHub (Sep 29, 2026): Thanks for opening this issue, @quantz8a! | Field | Value | |-------|-------| | **Issue** | #4386 | | **GitHub user ID** | `3152505` | | **Reporter** | @quantz8a | A maintainer will triage this when possible. To help us respond faster, please include: - Mooncake version or commit SHA - Environment (OS, CUDA/driver, RDMA stack if relevant) - Steps to reproduce and expected vs. actual behavior Useful links: [Documentation](https://kvcache-ai.github.io/Mooncake/) · [Contributing guide](https://github.com/kvcache-ai/Mooncake/blob/main/CONTRIBUTING.md) > This message was posted automatically by the issue bot.