1905103 - Crash in [@ mozilla::dom::IOUtils::EventQueue::EventQueue]

Status: NEW

(Core :: DOM: Core & HTML, defect)

Description

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

Only one crash per install, no crashes for v127 but 5-58 per beta with v128, all reports for Windows.

Could this be from bug 1895375?

Crash report: https://crash-stats.mozilla.org/report/index/e4f64b50-dcd4-457a-94aa-524b20240627

MOZ_CRASH Reason: MOZ_RELEASE_ASSERT(mBackgroundEventTarget)

Top 10 frames:

0  xul.dll  mozilla::dom::IOUtils::EventQueue::EventQueue()  dom/system/IOUtils.cpp:2334
1  xul.dll  mozilla::dom::IOUtils::GetState()  dom/system/IOUtils.cpp:2315
2  xul.dll  mozilla::dom::IOUtils::WithPromiseAndState(mozilla::dom::GlobalObject&, mozil...  dom/system/IOUtils.cpp:318
2  xul.dll  mozilla::dom::IOUtils::Remove(mozilla::dom::GlobalObject&, nsTSubstring<char1...  dom/system/IOUtils.cpp:694
2  xul.dll  mozilla::dom::IOUtils_Binding::remove(JSContext*, unsigned int, JS::Value*)  dom/bindings/IOUtilsBinding.cpp:1812
3  xul.dll  mozilla::dom::StaticMethodPromiseWrapper(JSContext*, unsigned int, JS::Value*)  dom/bindings/BindingUtils.cpp:3307
4  xul.dll  CallJSNative(JSContext*, bool (*)(JSContext*, unsigned int, JS::Value*), js::...  js/src/vm/Interpreter.cpp:481
4  xul.dll  js::InternalCallOrConstruct(JSContext*, JS::CallArgs const&, js::MaybeConstru...  js/src/vm/Interpreter.cpp:575
4  xul.dll  InternalCall(JSContext*, js::AnyInvokeArgs const&, js::CallReason)  js/src/vm/Interpreter.cpp:642
4  xul.dll  js::CallFromStack(JSContext*, JS::CallArgs const&, js::CallReason)  js/src/vm/Interpreter.cpp:647

Comment 1

[Ray Kraesig [:rkraesig]]

(In reply to Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout) from comment #0)

Could this be from bug 1895375?

I don't understand how it could be. Everything in that patchset is synchronous C++ code.

Comment 2

[Ray Kraesig [:rkraesig]]

Possible clue: it seems all of these crash-reports have ShutdownProgress: xpcom-shutdown-threads and ShutdownReason: AppClose.

Comment 3

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

The Bugbug bot thinks this bug should belong to the 'Core::DOM: Core & HTML' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Comment 4

[Hsin-Yi Tsai (she/her) (REO Fx 129) [:hsinyi]]

(In reply to Ray Kraesig [:rkraesig] from comment #2)

Possible clue: it seems all of these crash-reports have ShutdownProgress: xpcom-shutdown-threads and ShutdownReason: AppClose.

Jens, does this ring any bell? Do you have any ideas of the changes in fx 128 that could cause this?

Comment 5

[Jens Stutte [:jstutte]]

I assume we find the ThreadManager in shutdown after we started xpcom-shutdown-threads indicating that no new threads can be created. Creating a task queue on the background IO pool might want to create a new thread on the IO pool. Whatever is causing that IOUtils usage from Javascript should stop to do so earlier than this shutdown phase (or better: Have an async shutdown blocker for an earlier phase).

Comment 6

[Jens Stutte [:jstutte]]

Jens, does this ring any bell? Do you have any ideas of the changes in fx 128 that could cause this?

I do not think that the thread pool idle grace timeout could make threads go away earlier (which would be a reason for the need to create a new one), and we did not touch the settings of those pools a lot, anyways. So while I cannot 100% exclude it to be related, the underlying cause is different and was probably always there.

Comment 7

[Jens Stutte [:jstutte]]

One more datapoint for at least some of the crashes I clicked on is that they are shutting down right after start (3-5 seconds). And all seem to be instances of a background task, mostly install some uninstall.

Comment 8

[Jens Stutte [:jstutte]]

The only bug for background tasks that landed in 128 seems to be bug 1833735, not sure if it could be related.