Ensures that audio tasks cannot run after AudioManager is deleted.

media/audio/audio_thread.h

Index: media/audio/audio_thread.h
diff --git a/media/audio/audio_thread.h b/media/audio/audio_thread.h
new file mode 100644
index 0000000000000000000000000000000000000000..9fc46f9825bbca8ea84a74c35c4eaace20e8015a
--- /dev/null
+++ b/media/audio/audio_thread.h
@@ -0,0 +1,36 @@
+// Copyright 2017 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#ifndef MEDIA_AUDIO_AUDIO_THREAD_H_
+#define MEDIA_AUDIO_AUDIO_THREAD_H_
+
+#include "media/base/media_export.h"
+
+namespace base {
+class SingleThreadTaskRunner;
+}  // namespace base
+
+namespace media {
+
+// This class encapulates the logic for the thread and task runners that the
+// AudioManager and related classes run on.
+class MEDIA_EXPORT AudioThread {
+ public:
+  virtual ~AudioThread() {}
+
+  // Synchronously stops all underlying threads.
+  virtual void Stop() = 0;
+
+  // Returns the task runner used for audio IO.

[o1ka, 2017/05/10 15:57:57]

State that both methods should always return valid task runners (not null)?

[alokp, 2017/05/10 18:04:03]

Done.

+  virtual base::SingleThreadTaskRunner* GetTaskRunner() = 0;
+
+  // Heavyweight tasks should use GetWorkerTaskRunner() instead of
+  // GetTaskRunner(). On most platforms they are the same, but some share the
+  // UI loop with the audio IO loop.

[o1ka, 2017/05/10 15:57:57]

Move audio_manager.h l.49 comment here instead? I think that one explains the
meaning of task runners better.

[alokp, 2017/05/10 18:04:03]

I do not see a difference in the comments. What exactly is missing here?

[o1ka, 2017/05/11 13:44:12]

The other comment makes more sense, since there is no concept of UI and IO loops
in media.

+  virtual base::SingleThreadTaskRunner* GetWorkerTaskRunner() = 0;
+};
+
+}  // namespace media
+
+#endif  // MEDIA_AUDIO_AUDIO_THREAD_H_