1365600 - Crash in shutdownhang | __psynch_cvwait | _pthread_cond_wait | pthread_cond_signal_thread_np | <name omitted> | nsXULWindow::ShowModal [Mac] (engimail/password/expunge enabled on shutdown))

Status: RESOLVED INCOMPLETE

(Thunderbird :: General, defect)

Description

[Wayne Mery (:wsmwk)]

Pretty rare so far for 52.x. But for 45.x it is #1 crash for Mac - https://crash-stats.mozilla.com/topcrashers/?product=Thunderbird&version=45.8.0

Not Sierra-only

report bp-56113d58-b2af-460c-a781-dd94b0170517 (lightning is only addon in this crash report)
0 	libsystem_kernel.dylib	__psynch_cvwait	
1 	libsystem_pthread.dylib	_pthread_cond_wait	
2 	libsystem_pthread.dylib	pthread_cond_signal_thread_np	
3 	libnss3.dylib	PR_WaitCondVar	nsprpub/pr/src/pthreads/ptsynch.c:396
4 	XUL	<name omitted>	/builds/slave/tb-rel-c-esr52-m64_bld-0000000/build/objdir-tb/x86_64/dist/include/mozilla/CondVar.h:79
5 	XUL	nsThread::GetEvent(bool, nsIRunnable**, mozilla::BaseAutoLock<mozilla::Mutex>&)	xpcom/threads/nsThread.cpp:789
6 	XUL	nsThread::ProcessNextEvent(bool, bool*)	xpcom/threads/nsThread.cpp:1206
7 	XUL	NS_ProcessNextEvent(nsIThread*, bool)	xpcom/glue/nsThreadUtils.cpp:361
8 	XUL	nsXULWindow::ShowModal()	xpfe/appshell/nsXULWindow.cpp:408
9 	XUL	nsWindowWatcher::OpenWindowInternal(mozIDOMWindowProxy*, char const*, char const*, char const*, bool, bool, bool, nsIArray*, bool, bool, nsIDocShellLoadInfo*, mozIDOMWindowProxy**)	embedding/components/windowwatcher/nsWindowWatcher.cpp:1323
10 	XUL	nsWindowWatcher::OpenWindow(mozIDOMWindowProxy*, char const*, char const*, char const*, nsISupports*, mozIDOMWindowProxy**)	embedding/components/windowwatcher/nsWindowWatcher.cpp:353
11 	XUL	NS_InvokeByIndex	xpcom/reflect/xptcall/md/unix/xptcinvoke_x86_64_unix.cpp:180
12 	XUL	XPCWrappedNative::CallMethod(XPCCallContext&, XPCWrappedNative::CallMode)	js/xpconnect/src/XPCWrappedNative.cpp:2058
13 	XUL	XPC_WN_CallMethod(JSContext*, unsigned int, JS::Value*)	js/xpconnect/src/XPCWrappedNativeJSOps.cpp:999
14 	XUL	js::InternalCallOrConstruct(JSContext*, JS::CallArgs const&, js::MaybeConstruct)	js/src/jscntxtinlines.h:239
15 	XUL	Interpret	js/src/vm/Interpreter.cpp:510
16 	XUL	js::RunScript(JSContext*, js::RunState&)	js/src/vm/Interpreter.cpp:405
17 	XUL	js::InternalCallOrConstruct(JSContext*, JS::CallArgs const&, js::MaybeConstruct)	js/src/vm/Interpreter.cpp:477
18 	XUL	<name omitted>	js/src/vm/Interpreter.cpp:523
19 	XUL	JS_CallFunctionValue(JSContext*, JS::Handle<JSObject*>, JS::Handle<JS::Value>, JS::HandleValueArray const&, JS::MutableHandle<JS::Value>)	js/src/jsapi.cpp:2769
20 	XUL	nsXPCWrappedJSClass::CallMethod(nsXPCWrappedJS*, unsigned short, XPTMethodDescriptor const*, nsXPTCMiniVariant*)	js/xpconnect/src/XPCWrappedJSClass.cpp:1213 

bp-740c5ac9-3250-489d-a3a8-c6fc00170517 45.8.0


__pthread_kill | abort | abort_message | default_terminate_handler is another big crash in 45.8.0, but no examples yet in version 52. And oddly, lists first date as 2017-05-10  https://crash-stats.mozilla.com/signature/?product=Thunderbird&signature=__pthread_kill%20%7C%20abort%20%7C%20abort_message%20%7C%20default_terminate_handler&date=%3E%3D2017-05-10T13%3A54%3A09.000Z&date=%3C2017-05-17T13%3A54%3A09.000Z&_columns=date&_columns=product&_columns=version&_columns=build_id&_columns=platform&_columns=reason&_columns=address&_columns=install_time&_sort=-version&_sort=-date&page=1

Comment 1

[Wayne Mery (:wsmwk)]

#20 crash for 52.2.1

Comment 2

[Christian Riechers]

One with TB 52.6.0. Doesn't happen in safe mode.
bp-59a06695-977f-44c5-889e-a94a80180303
https://support.mozilla.org/en-US/questions/1195410

Comment 3

[Wayne Mery (:wsmwk)]

Thanks Christian.

So far there are no crashes for 59.0b1 and 59.0b2 - let's hope it continues and people can try https://www.mozilla.org/en-US/thunderbird/channel/

As previously noted this happens for multiple versions of macOS.

Comment 4

[Wayne Mery (:wsmwk)]

#1 crash for Mac beta so definitely not gone. I suspect this another Mac version of bug 1257058 password related bug.  Some users also mention enigmail.  Will need to evaluate after 1257058 is fixed

Comment 5

[Wayne Mery (:wsmwk)]

(In reply to Wayne Mery (:wsmwk) from comment #4 April 2018)

#1 crash for Mac beta so definitely not gone.

Still #1 crash for Mac. So bug 1257058 was not a substantial help

Comment 6

[Wayne Mery (:wsmwk)]

Attachment: TB bug 1365600 crash graph.png

Very strong decline Sept-Dec 2018. But then in mid-Jan a big spike which continues today, as seen in the graph https://crash-stats.mozilla.com/signature/?product=Thunderbird&signature=shutdownhang%20%7C%20__psynch_cvwait%20%7C%20_pthread_cond_wait%20%7C%20pthread_cond_signal_thread_np&date=%3E%3D2018-09-24T13%3A43%3A00.000Z&date=%3C2019-03-24T13%3A43%3A00.000Z#graphs

So what changed?

Comment 7

[Wayne Mery (:wsmwk)]

Still #1 crash and #11 overall for 60.8.0. Approx same for 68.0.
One person has reported imap expunge on shutdown as causing crashes.

Comment 9

[Steven Michaud [:smichaud] (Retired)]

Adding signature for macOS Catalina (10.15).

Comment 10

[Steven Michaud [:smichaud] (Retired)]

This bug might benefit from an analysis using a HookCase hook library (https://github.com/steven-michaud/HookCase).

Comment 11

[Steven Michaud [:smichaud] (Retired)]

These crashes seems to be happening more frequently on macOS Catalina than on earlier versions of macOS.

Comment 12

[Steven Michaud [:smichaud] (Retired)]

Note that these crashes also happen a lot in Firefox, so this bug isn't exclusively a Thunderbird problem.

Comment 13

[Wayne Mery (:wsmwk)]

Can you reproduce the crash using steps of bug 1257058 comment 0?

Comment 14

[Steven Michaud [:smichaud] (Retired)]

I haven't tested that STR. And given how complex the testcase is to set up, I won't be trying it anytime soon.

I did try the simplified STR from bug 1257058 comment 24, and had no hang, or any other problems.

Comment 15

[Wayne Mery (:wsmwk)]

Thanks Steven.

m_kato, are you able to see the crash comments of the past month for this bug and the signature for bug 1524247, and correlate this still happens for passwords? And if not, is there some other correlation?

Comment 16

[Makoto Kato [:m_kato]]

(In reply to Wayne Mery (:wsmwk) from comment #15)

Thanks Steven.

m_kato, are you able to see the crash comments of the past month for this bug and the signature for bug 1524247, and correlate this still happens for passwords? And if not, is there some other correlation?

bp-ff6e7b2b-01a0-481d-8cb7-9f88d0191125 is waiting nsImapProtocol::AuthLogin(char const*, nsTString<char16_t> const&, unsigned long long). Sicne OAuth2ThreadHelper::GetXOAuth2String, I guess that users uses Google's imap or calendar. But I don't know whether this issue is fixed.

So yes, this is waiting password.

Comment 17

[Christian Riechers]

Tried to reproduce with TB 72.0a1 (2019-11-29) (64-bit) on Linux.
Two IMAP accounts, one of them Gmail with Oauth2 authentication. No master password, no stored passwords.
Upon TB start, closed the Gmail Outh2 prompt, and hit Ok for the IMAP password prompt.
No crash upon TB shutdown.

Comment 18

[Wayne Mery (:wsmwk)]

The original signatures end abruptly ~December 3, and these begin:

• ~200 per day - shutdownhang | __psynch_cvwait | _pthread_cond_wait | pthread_cond_signal_thread_np | <name omitted> | nsThread::Shutdown | nsThreadManager::Shutdown
• ~50 per day - shutdownhang | __psynch_cvwait | _pthread_cond_wait | pthread_cond_signal_thread_np | <name omitted> | nsXULWindow::ShowModal

Comment 19

[Steven Michaud [:smichaud] (Retired)]

The original signatures end abruptly ~December 3, and these begin:

That's when pthread_cond_signal_thread_np got added to the prefix list.

Comment 21

[Wayne Mery (:wsmwk)]

Steven, thanks for pointing that out.

So we have new signatures (in order of frequency from the topcrash list of 68.4.1) :

• shutdownhang | __psynch_cvwait | _pthread_cond_wait | pthread_cond_signal_thread_np | <name omitted> | mozilla::detail::MutexImpl::unlock | nsThread::Shutdown | nsThreadManager::Shutdown
• shutdownhang | __psynch_cvwait | _pthread_cond_wait | pthread_cond_signal_thread_np | <name omitted> | DebugDump
• shutdownhang | __psynch_cvwait | _pthread_cond_wait | pthread_cond_signal_thread_np | <name omitted> | nsThread::Shutdown | nsTextFormatter::dosprintf::HEX
• shutdownhang | __psynch_cvwait | _pthread_cond_wait | pthread_cond_signal_thread_np | <name omitted> | js::jit::GetPropIRGenerator::tryAttachWindowProxy
• shutdownhang | __psynch_cvwait | _pthread_cond_wait | pthread_cond_signal_thread_np | <name omitted> | imp_removeBlock
• shutdownhang | __psynch_cvwait | _pthread_cond_wait | mozilla::detail::MutexImpl::unlock | pthread_cond_signal_thread_np | <name omitted> | mozilla::detail::MutexImpl::unlock | nsThread::Shutdown | nsThreadManager::Shutdown

Comment 22

[Anje]

https://crash-stats.mozilla.org/report/index/3feae727-242a-46aa-bde6-0b5640200629

Comment 23

[Wayne Mery (:wsmwk)]

Anje, do you still see this crash?

Unlike bug 1524247, this bug's signature does not exist for version 78.
shutdownhang | __psynch_cvwait | _pthread_cond_wait | pthread_cond_signal_thread_np | <name omitted> | nsThread::Shutdown | nsThreadManager::Shutdown

Comment 24

[Anje]

I did not experience that crash. It was a user in Support forum, but cannot locate that forum question. Not seen any recent evidence or report of it.