1811047 - Intermittent headless-spi SUMMARY: ThreadSanitizer: data race /builds/worker/workspace/obj-build/dist/include/mozilla/cxxalloc.h:51:10 in operator delete

Status: RESOLVED FIXED

(Core :: Networking: WebSockets, defect, P2)

Description

[Treeherder Bug Filer]

Filed by: rvandermeulen [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer?job_id=402780403&repo=try
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/MYnxQUkDT0iOEDVIkuUPAA/runs/0/artifacts/public/logs/live_backing.log

---

Comment 1

[Ryan VanderMeulen [:RyanVM]]

This is encountered with pretty low-frequency (~5% of the time) when running mochitest-plain in headless-spi (socketprocess) mode (--headless --setpref=network.process.enabled=true --setpref=network.http.network_access_on_socket_process.enabled=true) under TSAN. Seems to hit consistently when running dom/websocket/tests/test_websocket_no_duplicate_packet.html. This isn't currently enabled by default in CI, but I've been running Try pushes in the hopes of getting it green enough to do so.

Comment 2

[Ryan VanderMeulen [:RyanVM]]

Attachment: tsan log

Comment 3

[Daniel Veditz [:dveditz]]

Not clear this could be exploited, but WebSockets are clearly not as Threadsafe as comments in the code claim it is.

Comment 4

[Kershaw Chang [:kershaw]]

Attachment: Bug 1811047, r=#necko

When WebSocketConnectionParent::RecvOnError is called, we call WebSocketChannel::OnError and WebSocketChannel::CleanupConnection could be called.
In WebSocketChannel::CleanupConnection, WebSocketConnectionParent::Close will be called and the WebSocketChannel will be released (set mListener to null).
To avoid WebSocketChannel being released in WebSocketConnectionParent::Close, we don't need to set mListener to null.

Comment 5

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/604ac36df407

Comment 6

[BugBot [:suhaib / :marco/ :calixte]]

The patch landed in nightly and beta is affected.
:kershaw, is this bug important enough to require an uplift?

• If yes, please nominate the patch for beta approval.
• If no, please set status-firefox111 to wontfix.

For more information, please visit auto_nag documentation.

Comment 7

[Kershaw Chang [:kershaw]]

socket process feature is not enabled, so we don't need to uplift.