1993910 - Crash in [@ mozilla::dom::indexedDB::(anonymous namespace)::ConnectionPool::CloseDatabase(mozilla::dom::indexedDB::(anonymous namespace)::ConnectionPool::DatabaseInfo&)]

Status: NEW

(Core :: Storage: IndexedDB, defect)

Description

[Gabriele Svelto [:gsvelto]]

Crash report: https://crash-stats.mozilla.org/report/index/77aa81ec-cffd-4a76-8087-d73ae0251012

MOZ_CRASH Reason:

MOZ_CRASH(NS_SUCCEEDED(aDatabaseInfo.Dispatch( MakeAndAddRef<CloseConnectionRunnable>(aDatabaseInfo))))

Top 10 frames:

0  xul.dll  mozilla::dom::indexedDB::(anonymous namespace)::ConnectionPool::CloseDatabase...  dom/indexedDB/ActorsParent.cpp:8506
1  xul.dll  mozilla::dom::indexedDB::(anonymous namespace)::ConnectionPool::IdleTimerCall...  dom/indexedDB/ActorsParent.cpp:7857
1  xul.dll  std::find_if(mozilla::ArrayIterator<mozilla::dom::indexedDB::(anonymous names...  /builds/worker/fetches/vs/VC/Tools/MSVC/14.39.33519/include/xutility:6278
1  xul.dll  mozilla::dom::indexedDB::(anonymous namespace)::ConnectionPool::IdleTimerCall...  dom/indexedDB/ActorsParent.cpp:7848
1  xul.dll  mozilla::dom::indexedDB::(anonymous namespace)::ConnectionPool::IdleTimerCall...  dom/indexedDB/ActorsParent.cpp:7852
2  xul.dll  nsTimerImpl::Fire::<lambda_14>::operator()(nsTimerImpl::FuncCallback const&) ...  xpcom/threads/nsTimerImpl.cpp:750
2  xul.dll  mozilla::detail::VariantImplementation<unsigned char, 3, nsTimerImpl::FuncCal...  mfbt/Variant.h:310
2  xul.dll  mozilla::detail::VariantImplementation<unsigned char, 2, nsCOMPtr<nsIObserver...  mfbt/Variant.h:319
2  xul.dll  mozilla::detail::VariantImplementation<unsigned char, 1, nsCOMPtr<nsITimerCal...  mfbt/Variant.h:319
2  xul.dll  mozilla::detail::VariantImplementation<unsigned char, 0, nsTimerImpl::Unknown...  mfbt/Variant.h:319

I found this during nightly crash triage but this does not appear to be a new crash. I was wondering what could be triggering it and I suspect it might be an OOM condition in disguise. The crashes are only on Windows and several of them are happening on machine with very little commit space left available.

Comment 1

[Jan Varga [:janv]]

That's basically a failure to dispatch a runnable to the task queue on top of the thread pool. It might be OOM related, but it could also be an edge case where the code tries to dispatch a runnable to a thread pool that's already in shutdown mode. Maybe we could add some QM_TRY telemetry to reveal what kind of error code is returned from Dispatch.