From 629f5b12d6d8278114d3b022b325f7d3b1e36ddf Mon Sep 17 00:00:00 2001
From: Anna Henningsen <anna@addaleax.net>
Date: Sat, 11 Jul 2020 01:59:21 +0200
Subject: [PATCH] worker: add option to track unmanaged file descriptors

Add a public option for Workers which adds tracking for raw
file descriptors, as currently, those resources are not cleaned
up, unlike e.g. `FileHandle`s.

PR-URL: https://github.com/nodejs/node/pull/34303
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com>
---
 doc/api/worker_threads.md                     | 12 ++++
 lib/internal/worker.js                        |  3 +-
 src/node_worker.cc                            |  7 +-
 src/node_worker.h                             |  1 +
 .../test-worker-track-unmanaged-fds.js        | 69 +++++++++++++++++++
 5 files changed, 89 insertions(+), 3 deletions(-)
 create mode 100644 test/parallel/test-worker-track-unmanaged-fds.js

diff --git a/doc/api/worker_threads.md b/doc/api/worker_threads.md
index 38167bfd0bf234..7dd74675585019 100644
--- a/doc/api/worker_threads.md
+++ b/doc/api/worker_threads.md
@@ -620,6 +620,10 @@ if (isMainThread) {
 <!-- YAML
 added: v10.5.0
 changes:
+  - version:
+    - REPLACEME
+    pr-url: ???
+    description: The `trackUnmanagedFds` option was introduced.
   - version:
      - v13.13.0
      - v12.17.0
@@ -679,6 +683,12 @@ changes:
     occur as described in the [HTML structured clone algorithm][], and an error
     will be thrown if the object cannot be cloned (e.g. because it contains
     `function`s).
+  * `trackUnmanagedFds` {boolean} If this is set to `true`, then the Worker will
+    track raw file descriptors managed through [`fs.open()`][] and
+    [`fs.close()`][], and close them when the Worker exits, similar to other
+    resources like network sockets or file descriptors managed through
+    the [`FileHandle`][] API. This option is automatically inherited by all
+    nested `Worker`s. **Default**: `false`.
   * `transferList` {Object[]} If one or more `MessagePort`-like objects
     are passed in `workerData`, a `transferList` is required for those
     items or [`ERR_MISSING_MESSAGE_PORT_IN_TRANSFER_LIST`][] will be thrown.
@@ -900,6 +910,8 @@ active handle in the event system. If the worker is already `unref()`ed calling
 [`WebAssembly.Module`]: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/WebAssembly/Module
 [`Worker`]: #worker_threads_class_worker
 [`cluster` module]: cluster.html
+[`fs.open()`]: fs.html#fs_fs_open_path_flags_mode_callback
+[`fs.close()`]: fs.html#fs_fs_close_fd_callback
 [`markAsUntransferable()`]: #worker_threads_worker_markasuntransferable_object
 [`port.on('message')`]: #worker_threads_event_message
 [`port.onmessage()`]: https://developer.mozilla.org/en-US/docs/Web/API/MessagePort/onmessage
diff --git a/lib/internal/worker.js b/lib/internal/worker.js
index fdd096552129e8..a0608277158735 100644
--- a/lib/internal/worker.js
+++ b/lib/internal/worker.js
@@ -151,7 +151,8 @@ class Worker extends EventEmitter {
     this[kHandle] = new WorkerImpl(url,
                                    env === process.env ? null : env,
                                    options.execArgv,
-                                   parseResourceLimits(options.resourceLimits));
+                                   parseResourceLimits(options.resourceLimits),
+                                   !!options.trackUnmanagedFds);
     if (this[kHandle].invalidExecArgv) {
       throw new ERR_WORKER_INVALID_EXEC_ARGV(this[kHandle].invalidExecArgv);
     }
diff --git a/src/node_worker.cc b/src/node_worker.cc
index fa68fe88dc89f4..803fea03de0b85 100644
--- a/src/node_worker.cc
+++ b/src/node_worker.cc
@@ -305,7 +305,7 @@ void Worker::Run() {
             context,
             std::move(argv_),
             std::move(exec_argv_),
-            EnvironmentFlags::kNoFlags,
+            static_cast<EnvironmentFlags::Flags>(environment_flags_),
             thread_id_,
             std::move(inspector_parent_handle_)));
         if (is_stopped()) return;
@@ -452,7 +452,6 @@ void Worker::New(const FunctionCallbackInfo<Value>& args) {
 
   std::vector<std::string> exec_argv_out;
 
-  CHECK_EQ(args.Length(), 4);
   // Argument might be a string or URL
   if (!args[0]->IsNullOrUndefined()) {
     Utf8Value value(
@@ -578,6 +577,10 @@ void Worker::New(const FunctionCallbackInfo<Value>& args) {
   CHECK_EQ(limit_info->Length(), kTotalResourceLimitCount);
   limit_info->CopyContents(worker->resource_limits_,
                            sizeof(worker->resource_limits_));
+
+  CHECK(args[4]->IsBoolean());
+  if (args[4]->IsTrue() || env->tracks_unmanaged_fds())
+    worker->environment_flags_ |= EnvironmentFlags::kTrackUnmanagedFds;
 }
 
 void Worker::StartThread(const FunctionCallbackInfo<Value>& args) {
diff --git a/src/node_worker.h b/src/node_worker.h
index 1b9ba322bed04f..658018bb0661f8 100644
--- a/src/node_worker.h
+++ b/src/node_worker.h
@@ -114,6 +114,7 @@ class Worker : public AsyncWrap {
   bool stopped_ = true;
 
   bool has_ref_ = true;
+  uint64_t environment_flags_ = EnvironmentFlags::kNoFlags;
 
   // The real Environment of the worker object. It has a lesser
   // lifespan than the worker object itself - comes to life
diff --git a/test/parallel/test-worker-track-unmanaged-fds.js b/test/parallel/test-worker-track-unmanaged-fds.js
new file mode 100644
index 00000000000000..88ac8ae751a656
--- /dev/null
+++ b/test/parallel/test-worker-track-unmanaged-fds.js
@@ -0,0 +1,69 @@
+'use strict';
+const common = require('../common');
+const assert = require('assert');
+const { Worker } = require('worker_threads');
+const { once } = require('events');
+const fs = require('fs');
+
+// All the tests here are run sequentially, to avoid accidentally opening an fd
+// which another part of the test expects to be closed.
+
+const preamble = `
+const fs = require("fs");
+const { parentPort } = require('worker_threads');
+const __filename = ${JSON.stringify(__filename)};
+process.on('warning', (warning) => parentPort.postMessage({ warning }));
+`;
+
+(async () => {
+  // Consistency check: Without trackUnmanagedFds, FDs are *not* closed.
+  {
+    const w = new Worker(`${preamble}
+    parentPort.postMessage(fs.openSync(__filename));
+    `, { eval: true, trackUnmanagedFds: false });
+    const [ [ fd ] ] = await Promise.all([once(w, 'message'), once(w, 'exit')]);
+    assert(fd > 2);
+    fs.fstatSync(fd); // Does not throw.
+    fs.closeSync(fd);
+  }
+
+  // With trackUnmanagedFds, FDs are closed automatically.
+  {
+    const w = new Worker(`${preamble}
+    parentPort.postMessage(fs.openSync(__filename));
+    `, { eval: true, trackUnmanagedFds: true });
+    const [ [ fd ] ] = await Promise.all([once(w, 'message'), once(w, 'exit')]);
+    assert(fd > 2);
+    assert.throws(() => fs.fstatSync(fd), { code: 'EBADF' });
+  }
+
+  // There is a warning when an fd is unexpectedly opened twice.
+  {
+    const w = new Worker(`${preamble}
+    parentPort.postMessage(fs.openSync(__filename));
+    parentPort.once('message', () => {
+      const reopened = fs.openSync(__filename);
+      fs.closeSync(reopened);
+    });
+    `, { eval: true, trackUnmanagedFds: true });
+    const [ fd ] = await once(w, 'message');
+    fs.closeSync(fd);
+    w.postMessage('');
+    const [ { warning } ] = await once(w, 'message');
+    assert.match(warning.message,
+                 /File descriptor \d+ opened in unmanaged mode twice/);
+  }
+
+  // There is a warning when an fd is unexpectedly closed.
+  {
+    const w = new Worker(`${preamble}
+    parentPort.once('message', (fd) => {
+      fs.closeSync(fd);
+    });
+    `, { eval: true, trackUnmanagedFds: true });
+    w.postMessage(fs.openSync(__filename));
+    const [ { warning } ] = await once(w, 'message');
+    assert.match(warning.message,
+                 /File descriptor \d+ closed but not opened in unmanaged mode/);
+  }
+})().then(common.mustCall());