v8: support gc profile

Author: theanarkh

doc/api/v8.md

@@ -390,6 +390,82 @@ The API is a no-op if `--heapsnapshot-near-heap-limit` is already set from the
 command line or the API is called more than once. `limit` must be a positive
 integer. See [`--heapsnapshot-near-heap-limit`][] for more information.
 
+## `v8.collectGCProfile(options)`
+
+<!-- YAML
+added: REPLACEME
+-->
+
+* `options` {Object}
+  * `duration` {number} how long you want to collect the gc data.
+
+* Returns: {Promise}
+
+This API collects gc data over a period of time and return a object when
+the `Promise` resolve. The content is as follows.
+
+```json
+{
+  "version": 1,
+  "startTime": 1674059033862,
+  "statistics": [
+    {
+      "gcType": "Scavenge",
+      "beforeGC": {
+        "heapStatistics": {
+          "totalHeapSize": 5005312,
+          "totalHeapSizeExecutable": 524288,
+          "totalPhysicalSize": 5226496,
+          "totalAvailableSize": 4341325216,
+          "totalGlobalHandlesSize": 8192,
+          "usedGlobalHandlesSize": 2112,
+          "usedHeapSize": 4883840,
+          "heapSizeLimit": 4345298944,
+          "mallocedMemory": 254128,
+          "externalMemory": 225138,
+          "peakMallocedMemory": 181760
+        },
+        "heapSpaceStatistics": [
+          {
+            "spaceName": "read_only_space",
+            "spaceSize": 0,
+            "spaceUsedSize": 0,
+            "spaceAvailableSize": 0,
+            "physicalSpaceSize": 0
+          }
+        ]
+      },
+      "cost": 1574.14,
+      "afterGC": {
+        "heapStatistics": {
+          "totalHeapSize": 6053888,
+          "totalHeapSizeExecutable": 524288,
+          "totalPhysicalSize": 5500928,
+          "totalAvailableSize": 4341101384,
+          "totalGlobalHandlesSize": 8192,
+          "usedGlobalHandlesSize": 2112,
+          "usedHeapSize": 4059096,
+          "heapSizeLimit": 4345298944,
+          "mallocedMemory": 254128,
+          "externalMemory": 225138,
+          "peakMallocedMemory": 181760
+        },
+        "heapSpaceStatistics": [
+          {
+            "spaceName": "read_only_space",
+            "spaceSize": 0,
+            "spaceUsedSize": 0,
+            "spaceAvailableSize": 0,
+            "physicalSpaceSize": 0
+          }
+        ]
+      }
+    }
+  ],
+  "endtTime": 1674059036865
+}
+```
+
 ## Serialization API
 
 The serialization API provides means of serializing JavaScript values in a way

lib/v8.js

@@ -33,7 +33,7 @@ const {
 } = primordials;
 
 const { Buffer } = require('buffer');
-const { validateString, validateUint32 } = require('internal/validators');
+const { validateString, validateUint32, validateObject, validateNumber } = require('internal/validators');
 const {
   Serializer,
   Deserializer
@@ -63,7 +63,8 @@ const {
 } = require('internal/heap_utils');
 const promiseHooks = require('internal/promise_hooks');
 const { getOptionValue } = require('internal/options');
-
+const { setUnrefTimeout } = require('internal/timers');
+const { Promise, JSONParse } = primordials;
 /**
  * Generates a snapshot of the current V8 heap
  * and writes it to a JSON file.
@@ -397,6 +398,24 @@ function deserialize(buffer) {
   return der.readValue();
 }
 
+/**
+ * @param {{
+ *   duration: number,
+ * }} [options]
+ */
+function collectGCProfile(options) {
+  validateObject(options, 'options');
+  validateNumber(options.duration, 'options.duration', 1);
+  return new Promise((resolve) => {
+    const profiler = new binding.GCProfiler();
+    profiler.start();
+    setUnrefTimeout(() => {
+      const data = profiler.stop();
+      resolve(JSONParse(data));
+    }, options.duration);
+  });
+}
+
 module.exports = {
   cachedDataVersionTag,
   getHeapSnapshot,
@@ -416,4 +435,5 @@ module.exports = {
   promiseHooks,
   startupSnapshot,
   setHeapSnapshotNearHeapLimit,
+  collectGCProfile,
 };

src/node_v8.cc

@@ -33,6 +33,7 @@ namespace v8_utils {
 using v8::Array;
 using v8::Context;
 using v8::FunctionCallbackInfo;
+using v8::FunctionTemplate;
 using v8::HandleScope;
 using v8::HeapCodeStatistics;
 using v8::HeapSpaceStatistics;
@@ -210,6 +211,181 @@ void SetFlagsFromString(const FunctionCallbackInfo<Value>& args) {
   V8::SetFlagsFromString(*flags, static_cast<size_t>(flags.length()));
 }
 
+static const char* GetGCTypeName(v8::GCType gc_type) {
+  switch (gc_type) {
+    case v8::GCType::kGCTypeScavenge:
+      return "Scavenge";
+    case v8::GCType::kGCTypeMarkSweepCompact:
+      return "MarkSweepCompact";
+    case v8::GCType::kGCTypeIncrementalMarking:
+      return "IncrementalMarking";
+    case v8::GCType::kGCTypeProcessWeakCallbacks:
+      return "ProcessWeakCallbacks";
+    default:
+      return "UnKnown";
+  }
+}
+
+static void SetHeapStatistics(JSONWriter* writer, Isolate* isolate) {
+  HeapStatistics heap_statistics;
+  isolate->GetHeapStatistics(&heap_statistics);
+  writer->json_objectstart("heapStatistics");
+  writer->json_keyvalue("totalHeapSize", heap_statistics.total_heap_size());
+  writer->json_keyvalue("totalHeapSizeExecutable",
+                        heap_statistics.total_heap_size_executable());
+  writer->json_keyvalue("totalPhysicalSize",
+                        heap_statistics.total_physical_size());
+  writer->json_keyvalue("totalAvailableSize",
+                        heap_statistics.total_available_size());
+  writer->json_keyvalue("totalGlobalHandlesSize",
+                        heap_statistics.total_global_handles_size());
+  writer->json_keyvalue("usedGlobalHandlesSize",
+                        heap_statistics.used_global_handles_size());
+  writer->json_keyvalue("usedHeapSize", heap_statistics.used_heap_size());
+  writer->json_keyvalue("heapSizeLimit", heap_statistics.heap_size_limit());
+  writer->json_keyvalue("mallocedMemory", heap_statistics.malloced_memory());
+  writer->json_keyvalue("externalMemory", heap_statistics.external_memory());
+  writer->json_keyvalue("peakMallocedMemory",
+                        heap_statistics.peak_malloced_memory());
+  writer->json_objectend();
+
+  int space_count = isolate->NumberOfHeapSpaces();
+  writer->json_arraystart("heapSpaceStatistics");
+  for (int i = 0; i < space_count; i++) {
+    HeapSpaceStatistics heap_space_statistics;
+    isolate->GetHeapSpaceStatistics(&heap_space_statistics, i);
+    writer->json_start();
+    writer->json_keyvalue("spaceName", heap_space_statistics.space_name());
+    writer->json_keyvalue("spaceSize", heap_space_statistics.space_size());
+    writer->json_keyvalue("spaceUsedSize",
+                          heap_space_statistics.space_used_size());
+    writer->json_keyvalue("spaceAvailableSize",
+                          heap_space_statistics.space_available_size());
+    writer->json_keyvalue("physicalSpaceSize",
+                          heap_space_statistics.physical_space_size());
+    writer->json_end();
+  }
+  writer->json_arrayend();
+}
+
+static void BeforeGCCallback(Isolate* isolate,
+                             v8::GCType gc_type,
+                             v8::GCCallbackFlags flags,
+                             void* data) {
+  GCProfiler* profiler = static_cast<GCProfiler*>(data);
+  if (profiler->current_gc_type != 0) {
+    return;
+  }
+  JSONWriter* writer = profiler->writer();
+  writer->json_start();
+  writer->json_keyvalue("gcType", GetGCTypeName(gc_type));
+  writer->json_objectstart("beforeGC");
+  SetHeapStatistics(writer, isolate);
+  writer->json_objectend();
+  profiler->current_gc_type = gc_type;
+  profiler->start_time = uv_hrtime();
+}
+
+static void AfterGCCallback(Isolate* isolate,
+                            v8::GCType gc_type,
+                            v8::GCCallbackFlags flags,
+                            void* data) {
+  GCProfiler* profiler = static_cast<GCProfiler*>(data);
+  if (profiler->current_gc_type != gc_type) {
+    return;
+  }
+  JSONWriter* writer = profiler->writer();
+  profiler->current_gc_type = 0;
+  writer->json_keyvalue("cost", (uv_hrtime() - profiler->start_time) / 1e3);
+  profiler->start_time = 0;
+  writer->json_objectstart("afterGC");
+  SetHeapStatistics(writer, isolate);
+  writer->json_objectend();
+  writer->json_end();
+}
+
+GCProfiler::GCProfiler(Environment* env, Local<Object> object)
+    : BaseObject(env, object),
+      start_time(0),
+      current_gc_type(0),
+      state(GCProfilerState::kInitialized),
+      writer_(out_stream_, false) {
+  MakeWeak();
+}
+
+// This function will be called when
+// 1. StartGCProfile and StopGCProfile are called and
+//    JS land do not keep the object any more.
+// 2. StartGCProfile is called then the env exits before
+//    StopGCProfile is called.
+GCProfiler::~GCProfiler() {
+  if (state != GCProfiler::GCProfilerState::kInitialized) {
+    env()->isolate()->RemoveGCPrologueCallback(BeforeGCCallback, this);
+    env()->isolate()->RemoveGCEpilogueCallback(AfterGCCallback, this);
+  }
+}
+
+JSONWriter* GCProfiler::writer() {
+  return &writer_;
+}
+
+std::ostringstream* GCProfiler::out_stream() {
+  return &out_stream_;
+}
+
+void GCProfiler::New(const FunctionCallbackInfo<Value>& args) {
+  CHECK(args.IsConstructCall());
+  Environment* env = Environment::GetCurrent(args);
+  new GCProfiler(env, args.This());
+}
+
+void GCProfiler::Start(const FunctionCallbackInfo<Value>& args) {
+  Environment* env = Environment::GetCurrent(args);
+  GCProfiler* profiler;
+  ASSIGN_OR_RETURN_UNWRAP(&profiler, args.Holder());
+  if (profiler->state != GCProfiler::GCProfilerState::kInitialized) {
+    return;
+  }
+  profiler->writer()->json_start();
+  profiler->writer()->json_keyvalue("version", 1);
+
+  uv_timeval64_t ts;
+  if (uv_gettimeofday(&ts) == 0) {
+    profiler->writer()->json_keyvalue("startTime",
+                                      ts.tv_sec * 1000 + ts.tv_usec / 1000);
+  } else {
+    profiler->writer()->json_keyvalue("startTime", 0);
+  }
+  profiler->writer()->json_arraystart("statistics");
+  env->isolate()->AddGCPrologueCallback(BeforeGCCallback,
+                                        static_cast<void*>(profiler));
+  env->isolate()->AddGCEpilogueCallback(AfterGCCallback,
+                                        static_cast<void*>(profiler));
+  profiler->state = GCProfiler::GCProfilerState::kStarted;
+}
+
+void GCProfiler::Stop(const FunctionCallbackInfo<v8::Value>& args) {
+  Environment* env = Environment::GetCurrent(args);
+  GCProfiler* profiler;
+  ASSIGN_OR_RETURN_UNWRAP(&profiler, args.Holder());
+  if (profiler->state != GCProfiler::GCProfilerState::kStarted) {
+    return;
+  }
+  profiler->writer()->json_arrayend();
+  uv_timeval64_t ts;
+  if (uv_gettimeofday(&ts) == 0) {
+    profiler->writer()->json_keyvalue("endtTime",
+                                      ts.tv_sec * 1000 + ts.tv_usec / 1000);
+  } else {
+    profiler->writer()->json_keyvalue("endtTime", 0);
+  }
+  profiler->writer()->json_end();
+  profiler->state = GCProfiler::GCProfilerState::kStopped;
+  args.GetReturnValue().Set(
+      String::NewFromUtf8(env->isolate(), profiler->out_stream()->str().c_str())
+          .ToLocalChecked());
+}
+
 void Initialize(Local<Object> target,
                 Local<Value> unused,
                 Local<Context> context,
@@ -272,6 +448,14 @@ void Initialize(Local<Object> target,
 
   // Export symbols used by v8.setFlagsFromString()
   SetMethod(context, target, "setFlagsFromString", SetFlagsFromString);
+
+  // GCProfiler
+  Local<FunctionTemplate> t =
+      NewFunctionTemplate(env->isolate(), GCProfiler::New);
+  t->InstanceTemplate()->SetInternalFieldCount(BaseObject::kInternalFieldCount);
+  SetProtoMethod(env->isolate(), t, "start", GCProfiler::Start);
+  SetProtoMethod(env->isolate(), t, "stop", GCProfiler::Stop);
+  SetConstructorFunction(context, target, "GCProfiler", t);
 }
 
 void RegisterExternalReferences(ExternalReferenceRegistry* registry) {
@@ -281,6 +465,8 @@ void RegisterExternalReferences(ExternalReferenceRegistry* registry) {
   registry->Register(UpdateHeapSpaceStatisticsBuffer);
   registry->Register(SetFlagsFromString);
   registry->Register(SetHeapSnapshotNearHeapLimit);
+  registry->Register(GCProfiler::Start);
+  registry->Register(GCProfiler::Stop);
 }
 
 }  // namespace v8_utils

src/node_v8.h

@@ -3,8 +3,10 @@
 
 #if defined(NODE_WANT_INTERNALS) && NODE_WANT_INTERNALS
 
+#include <sstream>
 #include "aliased_buffer.h"
 #include "base_object.h"
+#include "json_utils.h"
 #include "node_snapshotable.h"
 #include "util.h"
 #include "v8.h"
@@ -34,6 +36,32 @@ class BindingData : public SnapshotableObject {
   SET_MEMORY_INFO_NAME(BindingData)
 };
 
+class GCProfiler : public BaseObject {
+ public:
+  enum class GCProfilerState { kInitialized, kStarted, kStopped };
+  GCProfiler(Environment* env, v8::Local<v8::Object> object);
+  inline ~GCProfiler() override;
+  static void New(const v8::FunctionCallbackInfo<v8::Value>& args);
+  static void Start(const v8::FunctionCallbackInfo<v8::Value>& args);
+  static void Stop(const v8::FunctionCallbackInfo<v8::Value>& args);
+
+  JSONWriter* writer();
+
+  std::ostringstream* out_stream();
+
+  SET_NO_MEMORY_INFO()
+  SET_MEMORY_INFO_NAME(GCProfiler)
+  SET_SELF_SIZE(GCProfiler)
+
+  u_int64_t start_time;
+  u_int8_t current_gc_type;
+  GCProfilerState state;
+
+ private:
+  std::ostringstream out_stream_;
+  JSONWriter writer_;
+};
+
 }  // namespace v8_utils
 
 }  // namespace node