Improve memory usage during GCHeap snapshot

1) Write GCHeap snapshot directly to file
2) Override GCHeap snapshot directory with EVN variable

Author: emutavchi

Source/JavaScriptCore/heap/HeapSnapshotBuilder.cpp

@@ -39,6 +39,52 @@
 
 namespace JSC {
 
+namespace {
+
+class FileOutputStingBuilder {
+    WTF_MAKE_NONCOPYABLE(FileOutputStingBuilder);
+
+    FileSystem::PlatformFileHandle fileHandle;
+    StringBuilder builder;
+
+public:
+    FileOutputStingBuilder(FileSystem::PlatformFileHandle fileHandle)
+        : fileHandle(WTFMove(fileHandle))
+    {
+    }
+
+    ~FileOutputStingBuilder()
+    {
+        flush(true);
+    }
+
+    template<typename... StringTypes> void append(StringTypes... fragment)
+    {
+        builder.append(fragment...);
+        flush();
+    }
+
+    void appendQuotedJSONString(const String& string)
+    {
+        builder.appendQuotedJSONString(string);
+        flush();
+    }
+
+    void flush(bool force = false)
+    {
+        constexpr size_t kBufferLengthLimit = 4 * WTF::KB;
+        if ((force && !builder.isEmpty()) ||  builder.length() >= kBufferLengthLimit)
+        {
+            CString utf8String = builder.toStringPreserveCapacity().utf8();
+            FileSystem::writeToFile(fileHandle, utf8String.data(), utf8String.length());
+            builder.clear();
+            builder.reserveCapacity(kBufferLengthLimit);
+        }
+    }
+};
+
+} // namespace
+
 NodeIdentifier HeapSnapshotBuilder::nextAvailableObjectIdentifier = 1;
 NodeIdentifier HeapSnapshotBuilder::getNextObjectIdentifier() { return nextAvailableObjectIdentifier++; }
 void HeapSnapshotBuilder::resetNextAvailableObjectIdentifier() { HeapSnapshotBuilder::nextAvailableObjectIdentifier = 1; }
@@ -329,6 +375,20 @@ String HeapSnapshotBuilder::descriptionForCell(JSCell *cell) const
 }
 
 String HeapSnapshotBuilder::json(Function<bool (const HeapSnapshotNode&)> allowNodeCallback)
+{
+    StringBuilder json;
+    writeJson(WTFMove(allowNodeCallback), json);
+    return json.toString();
+}
+
+void HeapSnapshotBuilder::writeJsonToFile(FileSystem::PlatformFileHandle fileHandle)
+{
+    FileOutputStingBuilder json(fileHandle);
+    writeJson([] (const HeapSnapshotNode&) { return true; }, json);
+}
+
+template<typename OutputStingBuilder>
+void HeapSnapshotBuilder::writeJson(Function<bool (const HeapSnapshotNode&)>&& allowNodeCallback, OutputStingBuilder &json)
 {
     VM& vm = m_profiler.vm();
     DeferGCForAWhile deferGC(vm);
@@ -350,7 +410,6 @@ String HeapSnapshotBuilder::json(Function<bool (const HeapSnapshotNode&)> allowN
     HashMap<UniquedStringImpl*, unsigned> edgeNameIndexes;
     unsigned nextEdgeNameIndex = 0;
 
-    StringBuilder json;
 
     auto appendNodeJSON = [&] (const HeapSnapshotNode& node) {
         // Let the client decide if they want to allow or disallow certain nodes.
@@ -615,7 +674,6 @@ String HeapSnapshotBuilder::json(Function<bool (const HeapSnapshotNode&)> allowN
     }
 
     json.append('}');
-    return json.toString();
 }
 
 } // namespace JSC

Source/JavaScriptCore/heap/HeapSnapshotBuilder.h

@@ -31,6 +31,7 @@
 #include <wtf/HashSet.h>
 #include <wtf/Lock.h>
 #include <wtf/Vector.h>
+#include <wtf/FileSystem.h>
 
 namespace JSC {
 
@@ -130,6 +131,7 @@ class JS_EXPORT_PRIVATE HeapSnapshotBuilder final : public HeapAnalyzer {
 
     String json();
     String json(Function<bool (const HeapSnapshotNode&)> allowNodeCallback);
+    void writeJsonToFile(FileSystem::PlatformFileHandle fileHandle);
 
 private:
     static NodeIdentifier nextAvailableObjectIdentifier;
@@ -140,6 +142,9 @@ class JS_EXPORT_PRIVATE HeapSnapshotBuilder final : public HeapAnalyzer {
     bool previousSnapshotHasNodeForCell(JSCell*, NodeIdentifier&);
 
     String descriptionForCell(JSCell*) const;
+
+    template<typename OutputStingBuilder>
+    void writeJson(Function<bool (const HeapSnapshotNode&)>&& allowNodeCallback, OutputStingBuilder &json);
 
     struct RootData {
         const char* reachabilityFromOpaqueRootReasons { nullptr };

Source/WebCore/bindings/js/GCController.cpp

@@ -140,10 +140,26 @@ void GCController::deleteAllLinkedCode(DeleteAllCodeEffort effort)
 
 void GCController::dumpHeap()
 {
+    static const char* sHeapDumpDirOverride = []() {
+        return getenv("WEBKIT_HEAP_SNAPSHOT_DIR");
+    }();
+
     FileSystem::PlatformFileHandle fileHandle;
-    String tempFilePath = FileSystem::openTemporaryFile("GCHeap"_s, fileHandle);
+    String tempFilePath;
+    if (sHeapDumpDirOverride) {
+        char buf[32];
+        time_t now = time(nullptr);
+        struct tm* tstruct = localtime(&now);
+        std::strftime(buf, sizeof(buf), "%Y%m%d_%H%M%S", tstruct);
+
+        tempFilePath = WTF::String::fromUTF8(sHeapDumpDirOverride) + "/GCHeap_" + WTF::String::fromUTF8(buf);
+        fileHandle = FileSystem::openFile(tempFilePath, FileSystem::FileOpenMode::ReadWrite);
+    } else {
+        tempFilePath = FileSystem::openTemporaryFile("GCHeap"_s, fileHandle);
+    }
+
     if (!FileSystem::isHandleValid(fileHandle)) {
-        WTFLogAlways("Dumping GC heap failed to open temporary file");
+        WTFLogAlways("Dumping GC heap failed to open temporary file: %s", tempFilePath.utf8().data());
         return;
     }
 
@@ -152,19 +168,14 @@ void GCController::dumpHeap()
 
     sanitizeStackForVM(vm);
 
-    String jsonData;
     {
         DeferGCForAWhile deferGC(vm); // Prevent concurrent GC from interfering with the full GC that the snapshot does.
 
         HeapSnapshotBuilder snapshotBuilder(vm.ensureHeapProfiler(), HeapSnapshotBuilder::SnapshotType::GCDebuggingSnapshot);
         snapshotBuilder.buildSnapshot();
-
-        jsonData = snapshotBuilder.json();
+        snapshotBuilder.writeJsonToFile(fileHandle);
     }
 
-    CString utf8String = jsonData.utf8();
-
-    FileSystem::writeToFile(fileHandle, utf8String.data(), utf8String.length());
     FileSystem::closeFile(fileHandle);
 
     WTFLogAlways("Dumped GC heap to %s", tempFilePath.utf8().data());