1421277 - Crash in mozilla::dom::ClientManager::~ClientManager

Status: RESOLVED WORKSFORME

(Core :: DOM: Service Workers, defect, P2)

Description

[Calixte Denizet (:calixte)]

This bug was filed from the Socorro interface and is
report bp-e8c23667-16ef-45ef-8eb2-e9c460171128.
=============================================================

Top 10 frames of crashing thread:

0 libxul.so mozilla::dom::ClientManager::~ClientManager [clone .cold.72] 
1 libxul.so mozilla::dom::ClientManager::Release dom/clients/manager/ClientManager.h:95
2 libxul.so mozilla::dom::ClientSource::~ClientSource dom/clients/manager/ClientSource.cpp:122
3 libxul.so nsGlobalWindowInner::FreeInnerObjects 
4 libxul.so nsGlobalWindowOuter::DetachFromDocShell 
5 libxul.so nsDocShell::Destroy 
6 libxul.so nsFrameLoader::DestroyDocShell 
7 libxul.so nsFrameLoaderDestroyRunnable::Run dom/base/nsFrameLoader.cpp:2249
8 libxul.so nsDocument::MaybeInitializeFinalizeFrameLoaders 
9 libxul.so mozilla::detail::RunnableMethodImpl<nsDocument*, void 0u>::Run xpcom/threads/nsThreadUtils.h:1142

=============================================================

There are 24 crashes in nightly 59 starting with buildid 20171123100420.
:bkelly could you investigate please ?

Comment 1

[Ben Kelly [:bkelly, not reviewing]]

This one is really weird.  Its like the thread-local stopped working properly.

Comment 2

[Ben Kelly [:bkelly, not reviewing]]

The crashes so far are all AMD family 6 processors running linux.

Comment 3

[Andrew Overholt [:overholt]]

I'm going to mark this as P3 due to the fact that all crashes are with quite old CPUs that are going to be hard to reproduce.

Comment 4

[Ben Kelly [:bkelly, not reviewing]]

Andrew Sutherland tells me these are actually intel processors that we call "amd64".  Family 6 model 70=0x46 is probably "haswell ULT":

https://github.com/intel/thermal_daemon/blob/master/src/thd_engine.cpp#L599

Comment 5

[Andrew Overholt [:overholt]]

Given comment 4: P2.

Comment 6

[Ben Kelly [:bkelly, not reviewing]]

Its still extremely low crash rate and specific to a single processor.  Not sure what we can do here.

Comment 7

[Ben Kelly [:bkelly, not reviewing]]

This stopped over 10 days ago.  I'm going to mark WFM and assume one of my other crash fixes for clients fixed this as well.