build: speed up startup with V8 code cache

This patch speeds up the startup time and reduce the startup memory
footprint by using V8 code cache when comiling builtin modules.

The current approach is demonstrated in the `with-code-cache`
Makefile target (no corresponding Windows target at the moment).

1. Build the binary normally (`src/node_code_cache_stub.cc` is used),
  by now `internalBinding('code_cache')` is an empty object
2. Run `tools/generate_code_cache.js` with the binary, which generates
  the code caches by reading source code of builtin modules off source
  code exposed by `require('internal/bootstrap/cache').builtinSource`
  and then generate a C++ file containing static char arrays of the
  code cache, using a format similar to `node_javascript.cc`
3. Run `configure` with the `--code-cache-path` option so that
  the newly generated C++ file will be used when compiling the
  new binary. The generated C++ file will put the cache into
  the `internalBinding('code_cache')` object with the module
  ids as keys
4. The new binary tries to read the code cache from
  `internalBinding('code_cache')` and use it to compile
  builtin modules. If the cache is used, it will put the id
  into `require('internal/bootstrap/cache').compiledWithCache`
  for bookkeeping, otherwise the id will be pushed into
  `require('internal/bootstrap/cache').compiledWithoutCache`

This patch also added tests that verify the code cache is
generated and used when compiling builtin modules.

The binary with code cache:

- Is ~1MB bigger than the binary without code cahe
- Consumes ~1MB less memory during start up
- Starts up about 60% faster

PR-URL: https://github.com/nodejs/node/pull/21405
Reviewed-By: John-David Dalton <john.david.dalton@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Gus Caplan <me@gus.host>

2 files changed, 58 insertions, 1 deletions

diff --git a/lib/internal/bootstrap/cache.js b/lib/internal/bootstrap/cache.js
new file mode 100644
index 0000000000..a3d22ba020
--- /dev/null
+++ b/lib/internal/bootstrap/cache.js
@@ -0,0 +1,30 @@
+'use strict';
+
+// This is only exposed for internal build steps and testing purposes.
+// We create new copies of the source and the code cache
+// so the resources eventually used to compile builtin modules
+// cannot be tampered with even with --expose-internals
+
+const {
+  NativeModule, internalBinding
+} = require('internal/bootstrap/loaders');
+
+module.exports = {
+  builtinSource: Object.assign({}, NativeModule._source),
+  codeCache: internalBinding('code_cache'),
+  compiledWithoutCache: NativeModule.compiledWithoutCache,
+  compiledWithCache: NativeModule.compiledWithCache,
+  nativeModuleWrap(script) {
+    return NativeModule.wrap(script);
+  },
+  // Modules with source code compiled in js2c that
+  // cannot be compiled with the code cache
+  cannotUseCache: [
+    'config',
+    // TODO(joyeecheung): update the C++ side so that
+    // the code cache is also used when compiling these
+    // two files.
+    'internal/bootstrap/loaders',
+    'internal/bootstrap/node'
+  ]
+};
diff --git a/lib/internal/bootstrap/loaders.js b/lib/internal/bootstrap/loaders.js
index de911eb841..c141c9adcf 100644
--- a/lib/internal/bootstrap/loaders.js
+++ b/lib/internal/bootstrap/loaders.js
@@ -125,10 +125,15 @@
 
   const config = getBinding('config');
 
+  const codeCache = getInternalBinding('code_cache');
+  const compiledWithoutCache = NativeModule.compiledWithoutCache = [];
+  const compiledWithCache = NativeModule.compiledWithCache = [];
+
   // Think of this as module.exports in this file even though it is not
   // written in CommonJS style.
   const loaderExports = { internalBinding, NativeModule };
   const loaderId = 'internal/bootstrap/loaders';
+
   NativeModule.require = function(id) {
     if (id === loaderId) {
       return loaderExports;
@@ -229,7 +234,29 @@
     this.loading = true;
 
     try {
-      const script = new ContextifyScript(source, this.filename);
+      // (code, filename, lineOffset, columnOffset
+      // cachedData, produceCachedData, parsingContext)
+      const script = new ContextifyScript(
+        source, this.filename, 0, 0,
+        codeCache[this.id], false, undefined
+      );
+
+      // One of these conditions may be false when any of the inputs
+      // of the `node_js2c` target in node.gyp is modified.
+      // FIXME(joyeecheung):
+      // 1. Figure out how to resolve the dependency issue. When the
+      //    code cache was introduced we were at a point where refactoring
+      //    node.gyp may not be worth the effort.
+      // 2. Calculate checksums in both js2c and generate_code_cache.js
+      //    and compare them before compiling the native modules since
+      //    V8 only checks the length of the source to decide whether to
+      //    reject the cache.
+      if (!codeCache[this.id] || script.cachedDataRejected) {
+        compiledWithoutCache.push(this.id);
+      } else {
+        compiledWithCache.push(this.id);
+      }
+
       // Arguments: timeout, displayErrors, breakOnSigint
       const fn = script.runInThisContext(-1, true, false);
       const requireFn = this.id.startsWith('internal/deps/') ?