Trace ScriptWrappableVisitor.m_markingDeque by oilpan gc

third_party/WebKit/Source/platform/heap/Heap.cpp

Index: third_party/WebKit/Source/platform/heap/Heap.cpp
diff --git a/third_party/WebKit/Source/platform/heap/Heap.cpp b/third_party/WebKit/Source/platform/heap/Heap.cpp
index a6a75d4b01ffe538a3d2397e8a7e6f7e00de0b13..e5420f86989bcaf0095f8fcf0b01a4d7c920d515 100644
--- a/third_party/WebKit/Source/platform/heap/Heap.cpp
+++ b/third_party/WebKit/Source/platform/heap/Heap.cpp
@@ -32,6 +32,7 @@
 
 #include "base/sys_info.h"
 #include "platform/Histogram.h"
+#include "platform/RuntimeEnabledFeatures.h"
 #include "platform/ScriptForbiddenScope.h"
 #include "platform/TraceEvent.h"
 #include "platform/heap/BlinkGCMemoryDumpProvider.h"
@@ -565,6 +566,10 @@ void ThreadHeap::collectGarbage(BlinkGC::StackState stackState, BlinkGC::GCType
     ThreadHeap::reportMemoryUsageHistogram();
     WTF::Partitions::reportMemoryUsageHistogram();
 
+    if (RuntimeEnabledFeatures::traceWrappablesEnabled()) {
+        state->invalidateDeadObjectsInWrappersMarkingDeque();

[haraken, 2016/06/09 06:28:31]

I begin to think it is not enough to call
invalidateDeadObjectsInWrappersMarkingDeque from the GCing thread. We need to
call it for all other threads that get involved in the GC.

Specifically, you need to move the call to ThreadState::postGC() so that it's
called for all ThreadStates that get involved in the GC.

[Marcel Hlopko, 2016/06/09 09:00:50]

You are right. Done.

+    }
+
     state->heap().postGC(gcType);
     state->heap().decommitCallbackStacks();
 }