firefox/xpcom/threads/nsIThreadManager.idl

/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* vim:set ts=2 sw=2 sts=2 et cindent: */
/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */

#include "nsISupports.idl"

[ptr] native PRThread(PRThread);
      native ThreadCreationOptions(nsIThreadManager::ThreadCreationOptions);

interface nsIEventTarget;
interface nsIRunnable;
interface nsIThread;

%{ C++
#include "mozilla/Maybe.h"
%}

[scriptable, function, uuid(039a227d-0cb7-44a5-a8f9-dbb7071979f2)]
interface nsINestedEventLoopCondition : nsISupports
{
  /**
   * Returns true if the current nested event loop should stop spinning.
   */
  boolean isDone();
};

/**
 * An interface for creating and locating nsIThread instances.
 */
[scriptable, builtinclass, uuid(1be89eca-e2f7-453b-8d38-c11ba247f6f3)]
interface nsIThreadManager : nsISupports
{
  /**
   * Default number of bytes reserved for a thread's stack, if no stack size
   * is specified in newThread().
   *
   * Defaults can be a little overzealous for many platforms.
   *
   * On Linux and OS X, for instance, the default thread stack size is whatever
   * getrlimit(RLIMIT_STACK) returns, which is often set at 8MB. Or, on Linux,
   * if the stack size is unlimited, we fall back to 2MB. This causes particular
   * problems on Linux, which allocates 2MB huge VM pages, and will often
   * immediately allocate them for any stacks which are 2MB or larger.
   *
   * The default on Windows is 1MB, which is a little more reasonable. But the
   * vast majority of our threads don't need anywhere near that much space.
   *
   * ASan, TSan and non-opt builds, however, often need a bit more, so give
   * them the platform default.
   */
%{C++
#if defined(MOZ_ASAN) || defined(MOZ_TSAN) || !defined(__OPTIMIZE__)
  static constexpr uint32_t DEFAULT_STACK_SIZE = 0;
#else
  static constexpr uint32_t DEFAULT_STACK_SIZE = 256 * 1024;
#endif

  /**
   * Threads which may need large stacks (such as those running JS or JS helper
   * tasks), can instead specify nsIThreadManager::LargeStackSize(). This is
   * just short of 2MB to avoid the Linux kernel allocating an entire 2MB huge
   * page for the stack on first access.
   *
   * ASan and TSan builds are given a larger stack size due to extra data and
   * red-zones which consume stack space.
   */
  static uint32_t LargeStackSize();

  static const uint32_t kThreadPoolStackSize = DEFAULT_STACK_SIZE;

  struct ThreadCreationOptions {
    // The size in bytes to reserve for the thread's stack. A value of `0` means
    // to use the platform default.
    uint32_t stackSize = nsIThreadManager::DEFAULT_STACK_SIZE;

    // If set to `true`, any attempts to dispatch runnables to this thread
    // without `DISPATCH_IGNORE_BLOCK_DISPATCH` will fail.
    //
    // This is intended to be used for threads which are expected to generally
    // only service a single runnable (other than thread lifecycle runnables),
    // and perform their own event dispatching internaly, such as thread pool
    // threads or the timer thread.
    bool blockDispatch = false;

    // (Windows-only) Whether the thread should have a MessageLoop capable of
    // processing native UI events. Defaults to false.
    bool isUiThread = false;

    // If set, long task markers will be collected for tasks
    // longer than longTaskLength ms when profiling is enabled.
    // See https://www.w3.org/TR/longtasks
    mozilla::Maybe<uint32_t> longTaskLength;
  };
%}

  /**
   * Create a new thread (a global, user PRThread) with the specified name.
   *
   * @param name
   *   The name of the thread. If it is empty the thread will not be named.
   * @param options
   *   Configuration options for the newly created thread.
   *
   * @returns
   *   The newly created nsIThread object.
   */
  [noscript] nsIThread newNamedThread(in ACString name, in ThreadCreationOptions options);

  /**
   * Get the main thread.
   */
  readonly attribute nsIThread mainThread;

  /**
   * Get the current thread.  If the calling thread does not already have a
   * nsIThread associated with it, then a new nsIThread will be created and
   * associated with the current PRThread.
   */
  readonly attribute nsIThread currentThread;

  /**
   * This queues a runnable to the main thread. It's a shortcut for JS callers
   * to be used instead of
   *   .mainThread.dispatch(runnable, Ci.nsIEventTarget.DISPATCH_NORMAL);
   * or
   *   .currentThread.dispatch(runnable, Ci.nsIEventTarget.DISPATCH_NORMAL);
   * C++ callers should instead use NS_DispatchToMainThread.
   */
  [optional_argc]
  void dispatchToMainThread(in nsIRunnable event, [optional] in uint32_t priority);

  /**
   * Similar to dispatchToMainThread, but wraps the event with extra
   * runnable that allocates nsAutoMicroTask.
   */
  [optional_argc]
  void dispatchToMainThreadWithMicroTask(in nsIRunnable event, [optional] in uint32_t priority);

  /**
   * This queues a runnable to the main thread's idle queue.
   *
   * @param event
   *   The event to dispatch.
   * @param timeout
   *   The time in milliseconds until this event should be moved from the idle
   *   queue to the regular queue if it hasn't been executed by then.  If not
   *   passed or a zero value is specified, the event will never be moved to
   *   the regular queue.
   */
  void idleDispatchToMainThread(in nsIRunnable event,
                                [optional] in uint32_t timeout);

  /*
   * A helper method to dispatch a task through nsIDirectTaskDispatcher to the
   * current thread.
   */
  void dispatchDirectTaskToCurrentThread(in nsIRunnable event);

  /**
   * Enter a nested event loop on the current thread, waiting on, and
   * processing events until condition.isDone() returns true.
   *
   * If condition.isDone() throws, this function will throw as well.
   *
   * C++ code should not use this function, instead preferring
   * mozilla::SpinEventLoopUntil.
   */
  void spinEventLoopUntil(in ACString aVeryGoodReasonToDoThis, in nsINestedEventLoopCondition condition);

  /**
   * Similar to the previous method, but the spinning of the event loop
   * terminates when the quit application shutting down starts.
   *
   * C++ code should not use this function, instead preferring
   * mozilla::SpinEventLoopUntil.
   */
  void spinEventLoopUntilOrQuit(in ACString aVeryGoodReasonToDoThis, in nsINestedEventLoopCondition condition);

  /**
   * Spin the current thread's event loop until there are no more pending
   * events.  This could be done with spinEventLoopUntil, but that would
   * require access to the current thread from JavaScript, which we are
   * moving away from.
   */
  void spinEventLoopUntilEmpty();

  /**
   * Return the EventTarget for the main thread.
   */
  readonly attribute nsIEventTarget mainThreadEventTarget;
};