Refactor the Heap into ThreadHeap to prepare for per thread heaps

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

 /*
  * Copyright (C) 2013 Google Inc. All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are
  * met:
  *
  *     * Redistributions of source code must retain the above copyright
  * notice, this list of conditions and the following disclaimer.
  *     * Redistributions in binary form must reproduce the above
@@ skipping 128 matching lines @@
     {
         // Only cleanup if we parked all threads in which case the GC happened
         // and we need to resume the other threads.
         if (m_resumeThreads)
             ThreadState::resumeThreads();
     }
 private:
     bool m_resumeThreads;
 };
 
-void Heap::flushHeapDoesNotContainCache()
-{
-    s_heapDoesNotContainCache->flush();
-}
-
 void Heap::init()
 {
     ThreadState::init();
-    s_markingStack = new CallbackStack();
-    s_postMarkingCallbackStack = new CallbackStack();
-    s_globalWeakCallbackStack = new CallbackStack();
-    s_ephemeronStack = new CallbackStack();
-    s_heapDoesNotContainCache = new HeapDoesNotContainCache();
-    s_freePagePool = new FreePagePool();
-    s_orphanedPagePool = new OrphanedPagePool();
-    s_allocatedSpace = 0;
-    s_allocatedObjectSize = 0;
-    s_objectSizeAtLastGC = 0;
-    s_markedObjectSize = 0;
-    s_markedObjectSizeAtLastCompleteSweep = 0;
-    s_wrapperCount = 0;
-    s_wrapperCountAtLastGC = 0;
-    s_collectedWrapperCount = 0;
-    s_partitionAllocSizeAtLastGC = WTF::Partitions::totalSizeOfCommittedPages();
-    s_estimatedMarkingTimePerByte = 0.0;
-#if ENABLE(ASSERT)
-    s_gcGeneration = 1;
-#endif
 
     GCInfoTable::init();
 
     if (Platform::current() && Platform::current()->currentThread())
         Platform::current()->registerMemoryDumpProvider(BlinkGCMemoryDumpProvider::instance(), "BlinkGC");
 }
 
 void Heap::shutdown()
 {
     if (Platform::current() && Platform::current()->currentThread())
         Platform::current()->unregisterMemoryDumpProvider(BlinkGCMemoryDumpProvider::instance());
-    s_shutdownCalled = true;
     ThreadState::shutdownHeapIfNecessary();
 }
 
 void Heap::doShutdown()
 {
-    // We don't want to call doShutdown() twice.
-    if (!s_markingStack)
-        return;
-
     ASSERT(!ThreadState::attachedThreads().size());
-    delete s_heapDoesNotContainCache;
-    s_heapDoesNotContainCache = nullptr;
-    delete s_freePagePool;
-    s_freePagePool = nullptr;
-    delete s_orphanedPagePool;
-    s_orphanedPagePool = nullptr;
-    delete s_globalWeakCallbackStack;
-    s_globalWeakCallbackStack = nullptr;
-    delete s_postMarkingCallbackStack;
-    s_postMarkingCallbackStack = nullptr;
-    delete s_markingStack;
-    s_markingStack = nullptr;
-    delete s_ephemeronStack;
-    s_ephemeronStack = nullptr;
-    delete s_regionTree;
-    s_regionTree = nullptr;
     GCInfoTable::shutdown();
     ThreadState::shutdown();
-    ASSERT(Heap::allocatedSpace() == 0);
+    // FIXME: should be zero
+    //ASSERT(ThreadState::terminating()->allocatedSpace() == 0);
 }
 
 #if ENABLE(ASSERT)
 BasePage* Heap::findPageFromAddress(Address address)
 {
     MutexLocker lock(ThreadState::threadAttachMutex());
     for (ThreadState* state : ThreadState::attachedThreads()) {
         if (BasePage* page = state->findPageFromAddress(address))
             return page;
     }
     return nullptr;
 }
 #endif
 
 Address Heap::checkAndMarkPointer(Visitor* visitor, Address address)
 {
     ASSERT(ThreadState::current()->isInGC());
 
 #if !ENABLE(ASSERT)
-    if (s_heapDoesNotContainCache->lookup(address))
+    if (ThreadState::current()->heapDoesNotContainCache()->lookup(address))
         return nullptr;
 #endif
 
     if (BasePage* page = lookup(address)) {
         ASSERT(page->contains(address));
         ASSERT(!page->orphaned());
-        ASSERT(!s_heapDoesNotContainCache->lookup(address));
+        ASSERT(!ThreadState::current()->heapDoesNotContainCache()->lookup(address));
         page->checkAndMarkPointer(visitor, address);
         return address;
     }
 
 #if !ENABLE(ASSERT)
-    s_heapDoesNotContainCache->addEntry(address);
+    ThreadState::current()->heapDoesNotContainCache()->addEntry(address);
 #else
-    if (!s_heapDoesNotContainCache->lookup(address))
-        s_heapDoesNotContainCache->addEntry(address);
+    if (!ThreadState::current()->heapDoesNotContainCache()->lookup(address))
+        ThreadState::current()->heapDoesNotContainCache()->addEntry(address);
 #endif
     return nullptr;
 }
 
 void Heap::pushTraceCallback(void* object, TraceCallback callback)
 {
     ASSERT(ThreadState::current()->isInGC());
 
     // Trace should never reach an orphaned page.
-    ASSERT(!Heap::orphanedPagePool()->contains(object));
-    CallbackStack::Item* slot = s_markingStack->allocateEntry();
+    ASSERT(!ThreadState::current()->orphanedPagePool()->contains(object));
+    CallbackStack::Item* slot = ThreadState::current()->markingStack()->allocateEntry();
     *slot = CallbackStack::Item(object, callback);
 }
 
 bool Heap::popAndInvokeTraceCallback(Visitor* visitor)
 {
-    CallbackStack::Item* item = s_markingStack->pop();
+    CallbackStack::Item* item = ThreadState::current()->markingStack()->pop();

[haraken, 2015/11/30 02:54:41]

At the moment this is okay, but ThreadState::current() should be avoided as much
as possible (because it needs to look up thread-local storage). We'll need to
pass ThreadState* around or associate it with the Visitor.

[keishi, 2016/01/06 05:35:33]

Done.

     if (!item)
         return false;
     item->call(visitor);
     return true;
 }
 
 void Heap::pushPostMarkingCallback(void* object, TraceCallback callback)
 {
     ASSERT(ThreadState::current()->isInGC());
 
     // Trace should never reach an orphaned page.
-    ASSERT(!Heap::orphanedPagePool()->contains(object));
-    CallbackStack::Item* slot = s_postMarkingCallbackStack->allocateEntry();
+    ASSERT(!ThreadState::current()->orphanedPagePool()->contains(object));
+    CallbackStack::Item* slot = ThreadState::current()->postMarkingCallbackStack()->allocateEntry();
     *slot = CallbackStack::Item(object, callback);
 }
 
 bool Heap::popAndInvokePostMarkingCallback(Visitor* visitor)
 {
-    if (CallbackStack::Item* item = s_postMarkingCallbackStack->pop()) {
+    if (CallbackStack::Item* item = ThreadState::current()->postMarkingCallbackStack()->pop()) {
         item->call(visitor);
         return true;
     }
     return false;
 }
 
 void Heap::pushGlobalWeakCallback(void** cell, WeakCallback callback)
 {
     ASSERT(ThreadState::current()->isInGC());
 
     // Trace should never reach an orphaned page.
-    ASSERT(!Heap::orphanedPagePool()->contains(cell));
-    CallbackStack::Item* slot = s_globalWeakCallbackStack->allocateEntry();
+    ASSERT(!ThreadState::current()->orphanedPagePool()->contains(cell));
+    CallbackStack::Item* slot = ThreadState::current()->globalWeakCallbackStack()->allocateEntry();
     *slot = CallbackStack::Item(cell, callback);
 }
 
 void Heap::pushThreadLocalWeakCallback(void* closure, void* object, WeakCallback callback)
 {
     ASSERT(ThreadState::current()->isInGC());
 
     // Trace should never reach an orphaned page.
-    ASSERT(!Heap::orphanedPagePool()->contains(object));
+    ASSERT(!ThreadState::current()->orphanedPagePool()->contains(object));
     ThreadState* state = pageFromObject(object)->heap()->threadState();
     state->pushThreadLocalWeakCallback(closure, callback);
 }
 
 bool Heap::popAndInvokeGlobalWeakCallback(Visitor* visitor)
 {
-    if (CallbackStack::Item* item = s_globalWeakCallbackStack->pop()) {
+    if (CallbackStack::Item* item = ThreadState::current()->globalWeakCallbackStack()->pop()) {
         item->call(visitor);
         return true;
     }
     return false;
 }
 
 void Heap::registerWeakTable(void* table, EphemeronCallback iterationCallback, EphemeronCallback iterationDoneCallback)
 {
     ASSERT(ThreadState::current()->isInGC());
 
     // Trace should never reach an orphaned page.
-    ASSERT(!Heap::orphanedPagePool()->contains(table));
-    CallbackStack::Item* slot = s_ephemeronStack->allocateEntry();
+    ASSERT(!ThreadState::current()->orphanedPagePool()->contains(table));
+    CallbackStack::Item* slot = ThreadState::current()->ephemeronStack()->allocateEntry();
     *slot = CallbackStack::Item(table, iterationCallback);
 
     // Register a post-marking callback to tell the tables that
     // ephemeron iteration is complete.
     pushPostMarkingCallback(table, iterationDoneCallback);
 }
 
 #if ENABLE(ASSERT)
 bool Heap::weakTableRegistered(const void* table)
 {
-    ASSERT(s_ephemeronStack);
-    return s_ephemeronStack->hasCallbackForObject(table);
+    ASSERT(ThreadState::current()->ephemeronStack());
+    return ThreadState::current()->ephemeronStack()->hasCallbackForObject(table);
 }
 #endif
 
 void Heap::preGC()
 {
     ASSERT(!ThreadState::current()->isInGC());
     for (ThreadState* state : ThreadState::attachedThreads())
         state->preGC();
 }
 
@@ skipping 57 matching lines @@
         BlinkGCMemoryDumpProvider::instance()->clearProcessDumpForCurrentGC();
 
     // Disallow allocation during garbage collection (but not during the
     // finalization that happens when the gcScope is torn down).
     ThreadState::NoAllocationScope noAllocationScope(state);
 
     preGC();
 
     StackFrameDepthScope stackDepthScope;
 
-    size_t totalObjectSize = Heap::allocatedObjectSize() + Heap::markedObjectSize();
+    size_t totalObjectSize = ThreadState::current()->allocatedObjectSize() + ThreadState::current()->markedObjectSize();
     if (gcType != BlinkGC::TakeSnapshot)
-        Heap::resetHeapCounters();
+        state->resetHeapCounters();
 
     // 1. Trace persistent roots.
     ThreadState::visitPersistentRoots(gcScope.visitor());
 
     // 2. Trace objects reachable from the stack.  We do this independent of the
     // given stackState since other threads might have a different stack state.
     ThreadState::visitStackRoots(gcScope.visitor());
 
     // 3. Transitive closure to trace objects including ephemerons.
     processMarkingStack(gcScope.visitor());
 
     postMarkingProcessing(gcScope.visitor());
     globalWeakProcessing(gcScope.visitor());
 
     // Now we can delete all orphaned pages because there are no dangling
     // pointers to the orphaned pages.  (If we have such dangling pointers,
     // we should have crashed during marking before getting here.)
-    orphanedPagePool()->decommitOrphanedPages();
+    state->orphanedPagePool()->decommitOrphanedPages();
 
     double markingTimeInMilliseconds = WTF::currentTimeMS() - startTime;
-    s_estimatedMarkingTimePerByte = totalObjectSize ? (markingTimeInMilliseconds / 1000 / totalObjectSize) : 0;
+    ThreadState::current()->setEstimatedMarkingTimePerByte(totalObjectSize ? (markingTimeInMilliseconds / 1000 / totalObjectSize) : 0);
 
 #if PRINT_HEAP_STATS
     dataLogF("Heap::collectGarbage (gcReason=%s, lazySweeping=%d, time=%.1lfms)\n", gcReasonString(reason), gcType == BlinkGC::GCWithoutSweep, markingTimeInMilliseconds);
 #endif
 
     Platform::current()->histogramCustomCounts("BlinkGC.CollectGarbage", markingTimeInMilliseconds, 0, 10 * 1000, 50);
-    Platform::current()->histogramCustomCounts("BlinkGC.TotalObjectSpace", Heap::allocatedObjectSize() / 1024, 0, 4 * 1024 * 1024, 50);
-    Platform::current()->histogramCustomCounts("BlinkGC.TotalAllocatedSpace", Heap::allocatedSpace() / 1024, 0, 4 * 1024 * 1024, 50);
+    Platform::current()->histogramCustomCounts("BlinkGC.TotalObjectSpace", ThreadState::current()->allocatedObjectSize() / 1024, 0, 4 * 1024 * 1024, 50);
+    Platform::current()->histogramCustomCounts("BlinkGC.TotalAllocatedSpace", ThreadState::current()->allocatedSpace() / 1024, 0, 4 * 1024 * 1024, 50);
     Platform::current()->histogramEnumeration("BlinkGC.GCReason", reason, BlinkGC::NumberOfGCReason);
-    Heap::reportMemoryUsageHistogram();
+    ThreadState::current()->reportMemoryUsageHistogram();
     WTF::Partitions::reportMemoryUsageHistogram();
 
     postGC(gcType);
 
 #if ENABLE(ASSERT)
     // 0 is used to figure non-assigned area, so avoid to use 0 in s_gcGeneration.
-    if (++s_gcGeneration == 0) {
-        s_gcGeneration = 1;
-    }
+    ThreadState::current()->incrementGcGeneration();
 #endif
 
     if (state->isMainThread())
         ScriptForbiddenScope::exit();
 }
 
 void Heap::collectGarbageForTerminatingThread(ThreadState* state)
 {
     {
         // A thread-specific termination GC must not allow other global GCs to go
@@ skipping 23 matching lines @@
         processMarkingStack(gcScope.visitor());
 
         postMarkingProcessing(gcScope.visitor());
         globalWeakProcessing(gcScope.visitor());
 
         state->postGC(BlinkGC::GCWithSweep);
     }
     state->preSweep();
 }
 
+void Heap::collectGarbageForIsolatedThread(ThreadState* state)
+{
+    {
+        // A thread-specific termination GC must not allow other global GCs to go
+        // ahead while it is running, hence the termination GC does not enter a
+        // safepoint. GCScope will not enter also a safepoint scope for
+        // ThreadTerminationGC.
+        GCScope gcScope(state, BlinkGC::NoHeapPointersOnStack, BlinkGC::ThreadTerminationGC);

[haraken, 2015/11/30 02:54:42]

As commented in ThreadState.cpp, it is wrong to set
BlinkGC::NoHeapPointersOnStac. Heap::collectGarbageForIsolatedThread can be
called for conservative GCs.

[keishi, 2016/01/06 05:35:33]

Done.

+
+        ThreadState::NoAllocationScope noAllocationScope(state);
+
+        state->preGC();
+
+        // 1. Trace the thread local persistent roots. For thread local GCs we
+        // don't trace the stack (ie. no conservative scanning) since this is
+        // only called during thread shutdown where there should be no objects
+        // on the stack.
+        // We also assume that orphaned pages have no objects reachable from
+        // persistent handles on other threads or CrossThreadPersistents.  The
+        // only cases where this could happen is if a subsequent conservative
+        // global GC finds a "pointer" on the stack or due to a programming
+        // error where an object has a dangling cross-thread pointer to an
+        // object on this heap.
+        state->visitPersistents(gcScope.visitor());
+
+        // 2. Trace objects reachable from the thread's persistent roots
+        // including ephemerons.
+        processMarkingStack(gcScope.visitor());
+
+        postMarkingProcessing(gcScope.visitor());
+        globalWeakProcessing(gcScope.visitor());
+
+        state->postGC(BlinkGC::GCWithSweep);
+    }
+    state->preSweep();
+}
+
 void Heap::processMarkingStack(Visitor* visitor)
 {
     // Ephemeron fixed point loop.
     do {
         {
             // Iteratively mark all objects that are reachable from the objects
             // currently pushed onto the marking stack.
             TRACE_EVENT0("blink_gc", "Heap::processMarkingStackSingleThreaded");
             while (popAndInvokeTraceCallback(visitor)) { }
         }
 
         {
             // Mark any strong pointers that have now become reachable in
             // ephemeron maps.
             TRACE_EVENT0("blink_gc", "Heap::processEphemeronStack");
-            s_ephemeronStack->invokeEphemeronCallbacks(visitor);
+            ThreadState::current()->ephemeronStack()->invokeEphemeronCallbacks(visitor);
         }
 
         // Rerun loop if ephemeron processing queued more objects for tracing.
-    } while (!s_markingStack->isEmpty());
+    } while (!ThreadState::current()->markingStack()->isEmpty());
 }
 
 void Heap::postMarkingProcessing(Visitor* visitor)
 {
     TRACE_EVENT0("blink_gc", "Heap::postMarkingProcessing");
     // Call post-marking callbacks including:
     // 1. the ephemeronIterationDone callbacks on weak tables to do cleanup
     //    (specifically to clear the queued bits for weak hash tables), and
     // 2. the markNoTracing callbacks on collection backings to mark them
     //    if they are only reachable from their front objects.
     while (popAndInvokePostMarkingCallback(visitor)) { }
 
-    s_ephemeronStack->clear();
+    ThreadState::current()->ephemeronStack()->clear();
 
     // Post-marking callbacks should not trace any objects and
     // therefore the marking stack should be empty after the
     // post-marking callbacks.
-    ASSERT(s_markingStack->isEmpty());
+    ASSERT(ThreadState::current()->markingStack()->isEmpty());
 }
 
 void Heap::globalWeakProcessing(Visitor* visitor)
 {
     TRACE_EVENT0("blink_gc", "Heap::globalWeakProcessing");
     double startTime = WTF::currentTimeMS();
 
     // Call weak callbacks on objects that may now be pointing to dead objects.
     while (popAndInvokeGlobalWeakCallback(visitor)) { }
 
     // It is not permitted to trace pointers of live objects in the weak
     // callback phase, so the marking stack should still be empty here.
-    ASSERT(s_markingStack->isEmpty());
+    ASSERT(ThreadState::current()->markingStack()->isEmpty());
 
     double timeForGlobalWeakProcessing = WTF::currentTimeMS() - startTime;
     Platform::current()->histogramCustomCounts("BlinkGC.TimeForGlobalWeakPrcessing", timeForGlobalWeakProcessing, 1, 10 * 1000, 50);
 }
 
 void Heap::collectAllGarbage()
 {
     // We need to run multiple GCs to collect a chain of persistent handles.
     size_t previousLiveObjects = 0;
     for (int i = 0; i < 5; ++i) {
         collectGarbage(BlinkGC::NoHeapPointersOnStack, BlinkGC::GCWithSweep, BlinkGC::ForcedGC);
-        size_t liveObjects = Heap::markedObjectSize();
+        size_t liveObjects = ThreadState::current()->markedObjectSize();
         if (liveObjects == previousLiveObjects)
             break;
         previousLiveObjects = liveObjects;
     }
 }
 
-double Heap::estimatedMarkingTime()
-{
-    ASSERT(ThreadState::current()->isMainThread());
-
-    // Use 8 ms as initial estimated marking time.
-    // 8 ms is long enough for low-end mobile devices to mark common
-    // real-world object graphs.
-    if (s_estimatedMarkingTimePerByte == 0)
-        return 0.008;
-
-    // Assuming that the collection rate of this GC will be mostly equal to
-    // the collection rate of the last GC, estimate the marking time of this GC.
-    return s_estimatedMarkingTimePerByte * (Heap::allocatedObjectSize() + Heap::markedObjectSize());
-}
-
-void Heap::reportMemoryUsageHistogram()
-{
-    static size_t supportedMaxSizeInMB = 4 * 1024;
-    static size_t observedMaxSizeInMB = 0;
-
-    // We only report the memory in the main thread.
-    if (!isMainThread())
-        return;
-    // +1 is for rounding up the sizeInMB.
-    size_t sizeInMB = Heap::allocatedSpace() / 1024 / 1024 + 1;
-    if (sizeInMB >= supportedMaxSizeInMB)
-        sizeInMB = supportedMaxSizeInMB - 1;
-    if (sizeInMB > observedMaxSizeInMB) {
-        // Send a UseCounter only when we see the highest memory usage
-        // we've ever seen.
-        Platform::current()->histogramEnumeration("BlinkGC.CommittedSize", sizeInMB, supportedMaxSizeInMB);
-        observedMaxSizeInMB = sizeInMB;
-    }
-}
-
-void Heap::reportMemoryUsageForTracing()
-{
-#if PRINT_HEAP_STATS
-    // dataLogF("allocatedSpace=%ldMB, allocatedObjectSize=%ldMB, markedObjectSize=%ldMB, partitionAllocSize=%ldMB, wrapperCount=%ld, collectedWrapperCount=%ld\n", Heap::allocatedSpace() / 1024 / 1024, Heap::allocatedObjectSize() / 1024 / 1024, Heap::markedObjectSize() / 1024 / 1024, WTF::Partitions::totalSizeOfCommittedPages() / 1024 / 1024, Heap::wrapperCount(), Heap::collectedWrapperCount());
-#endif
-
-    bool gcTracingEnabled;
-    TRACE_EVENT_CATEGORY_GROUP_ENABLED("blink_gc", &gcTracingEnabled);
-    if (!gcTracingEnabled)
-        return;
-
-    // These values are divided by 1024 to avoid overflow in practical cases (TRACE_COUNTER values are 32-bit ints).
-    // They are capped to INT_MAX just in case.
-    TRACE_COUNTER1("blink_gc", "Heap::allocatedObjectSizeKB", std::min(Heap::allocatedObjectSize() / 1024, static_cast<size_t>(INT_MAX)));
-    TRACE_COUNTER1("blink_gc", "Heap::markedObjectSizeKB", std::min(Heap::markedObjectSize() / 1024, static_cast<size_t>(INT_MAX)));
-    TRACE_COUNTER1(TRACE_DISABLED_BY_DEFAULT("blink_gc"), "Heap::markedObjectSizeAtLastCompleteSweepKB", std::min(Heap::markedObjectSizeAtLastCompleteSweep() / 1024, static_cast<size_t>(INT_MAX)));
-    TRACE_COUNTER1("blink_gc", "Heap::allocatedSpaceKB", std::min(Heap::allocatedSpace() / 1024, static_cast<size_t>(INT_MAX)));
-    TRACE_COUNTER1(TRACE_DISABLED_BY_DEFAULT("blink_gc"), "Heap::objectSizeAtLastGCKB", std::min(Heap::objectSizeAtLastGC() / 1024, static_cast<size_t>(INT_MAX)));
-    TRACE_COUNTER1(TRACE_DISABLED_BY_DEFAULT("blink_gc"), "Heap::wrapperCount", std::min(Heap::wrapperCount(), static_cast<size_t>(INT_MAX)));
-    TRACE_COUNTER1(TRACE_DISABLED_BY_DEFAULT("blink_gc"), "Heap::wrapperCountAtLastGC", std::min(Heap::wrapperCountAtLastGC(), static_cast<size_t>(INT_MAX)));
-    TRACE_COUNTER1(TRACE_DISABLED_BY_DEFAULT("blink_gc"), "Heap::collectedWrapperCount", std::min(Heap::collectedWrapperCount(), static_cast<size_t>(INT_MAX)));
-    TRACE_COUNTER1(TRACE_DISABLED_BY_DEFAULT("blink_gc"), "Heap::partitionAllocSizeAtLastGCKB", std::min(Heap::partitionAllocSizeAtLastGC() / 1024, static_cast<size_t>(INT_MAX)));
-    TRACE_COUNTER1("blink_gc", "Partitions::totalSizeOfCommittedPagesKB", std::min(WTF::Partitions::totalSizeOfCommittedPages() / 1024, static_cast<size_t>(INT_MAX)));
-}
-
 size_t Heap::objectPayloadSizeForTesting()
 {
     size_t objectPayloadSize = 0;
     for (ThreadState* state : ThreadState::attachedThreads()) {
         state->setGCState(ThreadState::GCRunning);
         state->makeConsistentForGC();
         objectPayloadSize += state->objectPayloadSizeForTesting();
         state->setGCState(ThreadState::EagerSweepScheduled);
         state->setGCState(ThreadState::Sweeping);
         state->setGCState(ThreadState::NoGCScheduled);
     }
     return objectPayloadSize;
 }
 
 BasePage* Heap::lookup(Address address)
 {
     ASSERT(ThreadState::current()->isInGC());
-    if (!s_regionTree)
+    if (!ThreadState::current()->regionTree())
         return nullptr;
-    if (PageMemoryRegion* region = s_regionTree->lookup(address)) {
+    if (PageMemoryRegion* region = ThreadState::current()->regionTree()->lookup(address)) {
         BasePage* page = region->pageFromAddress(address);
         return page && !page->orphaned() ? page : nullptr;
     }
     return nullptr;
 }
 
-static Mutex& regionTreeMutex()
-{
-    AtomicallyInitializedStaticReference(Mutex, mutex, new Mutex);
-    return mutex;
-}
-
 void Heap::removePageMemoryRegion(PageMemoryRegion* region)
 {
     // Deletion of large objects (and thus their regions) can happen
     // concurrently on sweeper threads.  Removal can also happen during thread
     // shutdown, but that case is safe.  Regardless, we make all removals
     // mutually exclusive.
-    MutexLocker locker(regionTreeMutex());
-    RegionTree::remove(region, &s_regionTree);
+    //MutexLocker locker(regionTreeMutex());
+    ThreadState* state = ThreadState::current();
+    if (!state)
+        state = ThreadState::terminating();

[haraken, 2015/11/30 02:54:42]

I'm wondering why this is needed. We should make sure that
ThreadState::current() returns a valid pointer while the thread is operating on
Oilpan's heap (e.g., calling Heap::removePageMemoryRegion).

+    ThreadState::RegionTree* regionTree = state->regionTree();
+    ThreadState::RegionTree::remove(region, &regionTree);
+    state->setRegionTree(regionTree);

[haraken, 2015/11/30 02:54:42]

Help me understand: What are you doing by setting the RegionTree on the
ThreadState? Isn't it already set on the ThreadState?

 }
 
 void Heap::addPageMemoryRegion(PageMemoryRegion* region)
 {
-    MutexLocker locker(regionTreeMutex());
-    RegionTree::add(new RegionTree(region), &s_regionTree);
+    ASSERT(ThreadState::current() && !ThreadState::terminating());
+    //MutexLocker locker(regionTreeMutex());
+    ThreadState::RegionTree* regionTree = ThreadState::current()->regionTree();
+    ThreadState::RegionTree::add(new ThreadState::RegionTree(region), &regionTree);
+    ThreadState::current()->setRegionTree(regionTree);
 }
 
-PageMemoryRegion* Heap::RegionTree::lookup(Address address)
-{
-    RegionTree* current = s_regionTree;
-    while (current) {
-        Address base = current->m_region->base();
-        if (address < base) {
-            current = current->m_left;
-            continue;
-        }
-        if (address >= base + current->m_region->size()) {
-            current = current->m_right;
-            continue;
-        }
-        ASSERT(current->m_region->contains(address));
-        return current->m_region;
-    }
-    return nullptr;
-}
-
-void Heap::RegionTree::add(RegionTree* newTree, RegionTree** context)
-{
-    ASSERT(newTree);
-    Address base = newTree->m_region->base();
-    for (RegionTree* current = *context; current; current = *context) {
-        ASSERT(!current->m_region->contains(base));
-        context = (base < current->m_region->base()) ? &current->m_left : &current->m_right;
-    }
-    *context = newTree;
-}
-
-void Heap::RegionTree::remove(PageMemoryRegion* region, RegionTree** context)
-{
-    ASSERT(region);
-    ASSERT(context);
-    Address base = region->base();
-    RegionTree* current = *context;
-    for (; current; current = *context) {
-        if (region == current->m_region)
-            break;
-        context = (base < current->m_region->base()) ? &current->m_left : &current->m_right;
-    }
-
-    // Shutdown via detachMainThread might not have populated the region tree.
-    if (!current)
-        return;
-
-    *context = nullptr;
-    if (current->m_left) {
-        add(current->m_left, context);
-        current->m_left = nullptr;
-    }
-    if (current->m_right) {
-        add(current->m_right, context);
-        current->m_right = nullptr;
-    }
-    delete current;
-}
-
-void Heap::resetHeapCounters()
-{
-    ASSERT(ThreadState::current()->isInGC());
-
-    Heap::reportMemoryUsageForTracing();
-
-    s_objectSizeAtLastGC = s_allocatedObjectSize + s_markedObjectSize;
-    s_partitionAllocSizeAtLastGC = WTF::Partitions::totalSizeOfCommittedPages();
-    s_allocatedObjectSize = 0;
-    s_markedObjectSize = 0;
-    s_wrapperCountAtLastGC = s_wrapperCount;
-    s_collectedWrapperCount = 0;
-}
-
-CallbackStack* Heap::s_markingStack;
-CallbackStack* Heap::s_postMarkingCallbackStack;
-CallbackStack* Heap::s_globalWeakCallbackStack;
-CallbackStack* Heap::s_ephemeronStack;
-HeapDoesNotContainCache* Heap::s_heapDoesNotContainCache;
-bool Heap::s_shutdownCalled = false;
-FreePagePool* Heap::s_freePagePool;
-OrphanedPagePool* Heap::s_orphanedPagePool;
-Heap::RegionTree* Heap::s_regionTree = nullptr;
-size_t Heap::s_allocatedSpace = 0;
-size_t Heap::s_allocatedObjectSize = 0;
-size_t Heap::s_objectSizeAtLastGC = 0;
-size_t Heap::s_markedObjectSize = 0;
-size_t Heap::s_markedObjectSizeAtLastCompleteSweep = 0;
-size_t Heap::s_wrapperCount = 0;
-size_t Heap::s_wrapperCountAtLastGC = 0;
-size_t Heap::s_collectedWrapperCount = 0;
-size_t Heap::s_partitionAllocSizeAtLastGC = 0;
-double Heap::s_estimatedMarkingTimePerByte = 0.0;
-#if ENABLE(ASSERT)
-uint16_t Heap::s_gcGeneration = 0;
-#endif
-
 } // namespace blink