tools: implement node_mksnapshot

Implements a node_mksnapshot target that generates a snapshot blob
from a Node.js main instance's isolate, and serializes the data blob
with other additional data into a C++ file that can be embedded into
the Node.js binary.

PR-URL: https://github.com/nodejs/node/pull/27321
Refs: https://github.com/nodejs/node/issues/17058
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>

Makefile

@@ -1226,6 +1226,8 @@ LINT_CPP_FILES = $(filter-out $(LINT_CPP_EXCLUDE), $(wildcard \
 	tools/icu/*.h \
 	tools/code_cache/*.cc \
 	tools/code_cache/*.h \
+	tools/snapshot/*.cc \
+	tools/snapshot/*.h \
 	))
 
 # Code blocks don't have newline at the end,

node.gyp

@@ -1156,7 +1156,47 @@
         }],
       ],
     }, # mkcodecache
-  ],  # end targets
+    {
+      'target_name': 'node_mksnapshot',
+      'type': 'executable',
+
+      'dependencies': [
+        '<(node_lib_target_name)',
+        'deps/histogram/histogram.gyp:histogram',
+      ],
+
+      'includes': [
+        'node.gypi'
+      ],
+
+      'include_dirs': [
+        'src',
+        'tools/msvs/genfiles',
+        'deps/v8/include',
+        'deps/cares/include',
+        'deps/uv/include',
+      ],
+
+      'defines': [ 'NODE_WANT_INTERNALS=1' ],
+
+      'sources': [
+        'src/node_code_cache_stub.cc',
+        'tools/snapshot/node_mksnapshot.cc',
+        'tools/snapshot/snapshot_builder.cc',
+        'tools/snapshot/snapshot_builder.h',
+      ],
+
+      'conditions': [
+        [ 'node_report=="true"', {
+          'conditions': [
+            ['OS=="win"', {
+              'libraries': [ 'Ws2_32' ],
+            }],
+          ],
+        }],
+      ],
+    }, # node_mksnapshot
+  ], # end targets
 
   'conditions': [
     ['OS=="aix" and node_shared=="true"', {

tools/snapshot/node_mksnapshot.cc

@@ -0,0 +1,51 @@
+#include <cstdio>
+#include <fstream>
+#include <iostream>
+#include <sstream>
+#include <string>
+#include <vector>
+
+#include "libplatform/libplatform.h"
+#include "node_internals.h"
+#include "snapshot_builder.h"
+#include "v8.h"
+
+#ifdef _WIN32
+#include <windows.h>
+
+int wmain(int argc, wchar_t* argv[]) {
+#else   // UNIX
+int main(int argc, char* argv[]) {
+#endif  // _WIN32
+
+  if (argc < 2) {
+    std::cerr << "Usage: " << argv[0] << " <path/to/output.cc>\n";
+    return 1;
+  }
+
+  std::ofstream out;
+  out.open(argv[1], std::ios::out | std::ios::binary);
+  if (!out.is_open()) {
+    std::cerr << "Cannot open " << argv[1] << "\n";
+    return 1;
+  }
+
+  int node_argc = 1;
+  char argv0[] = "node";
+  char* node_argv[] = {argv0, nullptr};
+
+  node::InitializationResult result =
+      node::InitializeOncePerProcess(node_argc, node_argv);
+  CHECK(!result.early_return);
+  CHECK_EQ(result.exit_code, 0);
+
+  {
+    std::string snapshot =
+        node::SnapshotBuilder::Generate(result.args, result.exec_args);
+    out << snapshot;
+    out.close();
+  }
+
+  node::TearDownOncePerProcess();
+  return 0;
+}

tools/snapshot/snapshot_builder.cc

@@ -0,0 +1,122 @@
+#include "snapshot_builder.h"
+#include <iostream>
+#include <sstream>
+#include "env-inl.h"
+#include "node_internals.h"
+#include "node_main_instance.h"
+#include "node_v8_platform-inl.h"
+
+namespace node {
+
+using v8::Context;
+using v8::HandleScope;
+using v8::Isolate;
+using v8::Local;
+using v8::Locker;
+using v8::SnapshotCreator;
+using v8::StartupData;
+
+std::string FormatBlob(v8::StartupData* blob,
+                       const std::vector<size_t>& isolate_data_indexes) {
+  std::stringstream ss;
+  size_t isolate_data_indexes_size = isolate_data_indexes.size();
+
+  ss << R"(#include <cstddef>
+#include "node_main_instance.h"
+#include "v8.h"
+
+// This file is generated by tools/snapshot. Do not edit.
+
+namespace node {
+
+static const uint8_t blob_data[] = {
+)";
+
+  for (int i = 0; i < blob->raw_size; i++) {
+    uint8_t ch = blob->data[i];
+    ss << std::to_string(ch) << ((i == blob->raw_size - 1) ? '\n' : ',');
+  }
+
+  ss << R"(};
+
+static const int blob_size = )"
+     << blob->raw_size << R"(;
+static v8::StartupData blob = {
+  reinterpret_cast<const char*>(blob_data),
+  blob_size
+};
+)";
+
+  ss << R"(v8::StartupData*
+NodeMainInstance::GetEmbeddedSnapshotBlob() {
+  return &blob;
+}
+
+static const size_t isolate_data_indexes_raw[] = {
+)";
+  for (size_t i = 0; i < isolate_data_indexes_size; i++) {
+    ss << std::to_string(isolate_data_indexes[i])
+       << ((i == isolate_data_indexes_size - 1) ? '\n' : ',');
+  }
+  ss << "};\n\n";
+
+  ss << "static const size_t isolate_data_indexes_size = "
+     << isolate_data_indexes_size << R"(;
+
+NodeMainInstance::IndexArray isolate_data_indexes {
+    isolate_data_indexes_raw,
+    isolate_data_indexes_size
+};
+
+const NodeMainInstance::IndexArray*
+NodeMainInstance::GetIsolateDataIndexes() {
+  return &isolate_data_indexes;
+}
+}  // namespace node
+)";
+
+  return ss.str();
+}
+
+std::string SnapshotBuilder::Generate(
+    const std::vector<std::string> args,
+    const std::vector<std::string> exec_args) {
+  // TODO(joyeecheung): collect external references and set it in
+  // params.external_references.
+  std::vector<intptr_t> external_references = {
+      reinterpret_cast<intptr_t>(nullptr)};
+  Isolate* isolate = Isolate::Allocate();
+  per_process::v8_platform.Platform()->RegisterIsolate(isolate,
+                                                       uv_default_loop());
+  NodeMainInstance* main_instance = nullptr;
+  std::string result;
+
+  {
+    std::vector<size_t> isolate_data_indexes;
+    SnapshotCreator creator(isolate, external_references.data());
+    {
+      main_instance =
+          NodeMainInstance::Create(isolate,
+                                   uv_default_loop(),
+                                   per_process::v8_platform.Platform(),
+                                   args,
+                                   exec_args);
+      HandleScope scope(isolate);
+      creator.SetDefaultContext(Context::New(isolate));
+      isolate_data_indexes = main_instance->isolate_data()->Serialize(&creator);
+    }
+
+    // Must be out of HandleScope
+    StartupData blob =
+        creator.CreateBlob(SnapshotCreator::FunctionCodeHandling::kClear);
+    // Must be done while the snapshot creator isolate is entered i.e. the
+    // creator is still alive.
+    main_instance->Dispose();
+    result = FormatBlob(&blob, isolate_data_indexes);
+    delete blob.data;
+  }
+
+  per_process::v8_platform.Platform()->UnregisterIsolate(isolate);
+  return result;
+}
+}  // namespace node

tools/snapshot/snapshot_builder.h

@@ -0,0 +1,15 @@
+#ifndef TOOLS_SNAPSHOT_SNAPSHOT_BUILDER_H_
+#define TOOLS_SNAPSHOT_SNAPSHOT_BUILDER_H_
+
+#include <string>
+#include <vector>
+
+namespace node {
+class SnapshotBuilder {
+ public:
+  static std::string Generate(const std::vector<std::string> args,
+                              const std::vector<std::string> exec_args);
+};
+}  // namespace node
+
+#endif  // TOOLS_SNAPSHOT_SNAPSHOT_BUILDER_H_