muduo源码剖析之EventLoop事件循环类

简介: EventLoop.cc就相当于一个reactor,多线程之间的函数调用(用eventfd唤醒),epoll处理,超时队列处理,对channel的处理。运行loop的进程被称为IO线程,EventLoop提供了一些API确保相应函数在IO线程中调用,确保没有用互斥量保护的变量只能在IO线程中使用,也封装了超时队列的基本操作。

简介

EventLoop.cc就相当于一个reactor,多线程之间的函数调用(用eventfd唤醒),epoll处理,超时队列处理,对channel的处理。运行loop的进程被称为IO线程,EventLoop提供了一些API确保相应函数在IO线程中调用,确保没有用互斥量保护的变量只能在IO线程中使用,也封装了超时队列的基本操作。

成员及属性解析

一个事件循环,注意,一个创建了EventLoop对象的线程是workloop线程

主要接口

loop

死循环,阻塞在Poller的poll函数,等待唤醒唤醒后执行ChannelList中每个Channel的回调最后执行任务队列中的Functor

runInLoop

在IO线程中执行用户回调Functor,若调用者非IO线程,则会调用queueInLoop

queueInLoop

当调用者并非当前EventLoop所在线程时,将Functor存入EventLoop的任务队列从而保证Functor由IO线程执行,这是线程安全的保证之一

updateChannel与removeChannel

核心中的核心,通过这个公有接口建立起Channel和Poller沟通的桥梁Channel通过这个接口向Poller注册或者移除自己的fd实现了Poller和Channel两端的解耦

核心实现:handleEvent

遍历所有的activeChannelList_,并依次执行这些Channel中注册的回调函数这个环节非常非常关键,是一切事件派发机制中回调执行的地方

主要成员

wakeupchannel_

通过eventfd唤醒的channel

EventLoop可以通过这个Channel唤醒自己执行定时任务

activeChannelList_

通过一次poll获得的所有发生事件的Channel指针列表

pendingFunctors_

所有非IO线程调用的用户回调都会存放在这个队列中,通过mutex互斥量保护

poller_

一个多路复用实例

源码剖析

EventLoop.h

#ifndef MUDUO_NET_EVENTLOOP_H
#define MUDUO_NET_EVENTLOOP_H

#include <atomic>
#include <functional>
#include <vector>

#include <boost/any.hpp>

#include "muduo/base/Mutex.h"
#include "muduo/base/CurrentThread.h"
#include "muduo/base/Timestamp.h"
#include "muduo/net/Callbacks.h"
#include "muduo/net/TimerId.h"

namespace muduo
{
   
namespace net
{
   

class Channel;
class Poller;
class TimerQueue;

///
/// Reactor, at most one per thread.
///
/// This is an interface class, so don't expose too much details.
class EventLoop : noncopyable
{
   
 public:
  typedef std::function<void()> Functor;

  EventLoop();
  ~EventLoop();  // force out-line dtor, for std::unique_ptr members.

  //开启事件循环
  void loop();

  //退出事件循环
  void quit();

  //轮询返回的时间,通常意味着数据到达。
  Timestamp pollReturnTime() const {
    return pollReturnTime_; }

  int64_t iteration() const {
    return iteration_; }

  /// Runs callback immediately in the loop thread.
  /// It wakes up the loop, and run the cb.
  /// If in the same loop thread, cb is run within the function.
  /// Safe to call from other threads.
  ///在当前loop中执行cb
  void runInLoop(Functor cb);
  /// Queues callback in the loop thread.
  /// Runs after finish pooling.
  /// Safe to call from other threads.
  ///将cb放入队列中,唤醒loop所在的线程执行
  void queueInLoop(Functor cb);

  size_t queueSize() const;

  // timers

  ///
  /// Runs callback at 'time'.
  /// Safe to call from other threads.
  ///
  TimerId runAt(Timestamp time, TimerCallback cb);
  ///
  /// Runs callback after @c delay seconds.
  /// Safe to call from other threads.
  ///
  TimerId runAfter(double delay, TimerCallback cb);
  ///
  /// Runs callback every @c interval seconds.
  /// Safe to call from other threads.
  ///
  TimerId runEvery(double interval, TimerCallback cb);
  ///
  /// Cancels the timer.
  /// Safe to call from other threads.
  ///
  void cancel(TimerId timerId);

  // internal usage
  //唤醒loop所在的线程
  void wakeup();

  //调用poller的方法
  void updateChannel(Channel* channel);
  void removeChannel(Channel* channel);
  bool hasChannel(Channel* channel);

  // pid_t threadId() const { return threadId_; }
  void assertInLoopThread()
  {
   
    if (!isInLoopThread())
    {
   
      abortNotInLoopThread();
    }
  }
    //判断eventloop对象是否在自己的线程
  bool isInLoopThread() const {
    return threadId_ == CurrentThread::tid(); }
  // bool callingPendingFunctors() const { return callingPendingFunctors_; }
  bool eventHandling() const {
    return eventHandling_; }

  void setContext(const boost::any& context)
  {
    context_ = context; }

  const boost::any& getContext() const
  {
    return context_; }

  boost::any* getMutableContext()
  {
    return &context_; }

  static EventLoop* getEventLoopOfCurrentThread();

 private:
  void abortNotInLoopThread();
  void handleRead();  // waked up
  void doPendingFunctors();//在loop一次后执行pendingFunctors_中的所有方法(会清空队列)

  void printActiveChannels() const; // DEBUG

  typedef std::vector<Channel*> ChannelList;

  bool looping_; /* atomic */
  std::atomic<bool> quit_;//标识loop的退出
  bool eventHandling_; /* atomic */
  //标识当前loop是否需要有执行的回调操作
  bool callingPendingFunctors_; /* atomic */
  int64_t iteration_;
  const pid_t threadId_;//记录thread所在的线程pid
  Timestamp pollReturnTime_;
  std::unique_ptr<Poller> poller_;
  std::unique_ptr<TimerQueue> timerQueue_;

  //主要作用,当mainLoop获取到一个accept新用户的channel,通过轮询算法选择一个subloop,通过该成员唤醒subloop处理,使用eventfd
  int wakeupFd_;
  // unlike in TimerQueue, which is an internal class,
  // we don't expose Channel to client.
  std::unique_ptr<Channel> wakeupChannel_;
  boost::any context_;

  // scratch variables
  ChannelList activeChannels_;
  Channel* currentActiveChannel_;

  mutable MutexLock mutex_;//保证pendingFunctors_的线程安全操作
  std::vector<Functor> pendingFunctors_ GUARDED_BY(mutex_);//存储loop需要执行的所有操作
};

}  // namespace net
}  // namespace muduo

#endif  // MUDUO_NET_EVENTLOOP_H

eventloop.cc

// Copyright 2010, Shuo Chen.  All rights reserved.
// http://code.google.com/p/muduo/
//
// Use of this source code is governed by a BSD-style license
// that can be found in the License file.

// Author: Shuo Chen (chenshuo at chenshuo dot com)

#include "muduo/net/EventLoop.h"

#include "muduo/base/Logging.h"
#include "muduo/base/Mutex.h"
#include "muduo/net/Channel.h"
#include "muduo/net/Poller.h"
#include "muduo/net/SocketsOps.h"
#include "muduo/net/TimerQueue.h"

#include <algorithm>

#include <signal.h>
#include <sys/eventfd.h>
#include <unistd.h>

using namespace muduo;
using namespace muduo::net;

namespace
{
   
//保证一个线程只有一个loop
__thread EventLoop* t_loopInThisThread = 0;
//poll超时时间
const int kPollTimeMs = 10000;

int createEventfd()
{
   
  int evtfd = ::eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC);
  if (evtfd < 0)
  {
   
    LOG_SYSERR << "Failed in eventfd";
    abort();
  }
  return evtfd;
}

#pragma GCC diagnostic ignored "-Wold-style-cast"
class IgnoreSigPipe
{
   
 public:
  IgnoreSigPipe()
  {
   
    ::signal(SIGPIPE, SIG_IGN);
    // LOG_TRACE << "Ignore SIGPIPE";
  }
};
#pragma GCC diagnostic error "-Wold-style-cast"

IgnoreSigPipe initObj;
}  // namespace

EventLoop* EventLoop::getEventLoopOfCurrentThread()
{
   
  return t_loopInThisThread;
}

//创建了EventLoop对象的线程称为IO线程
EventLoop::EventLoop()
  : looping_(false),                                     //判断是否在loop
    quit_(false),                                        //判断是否退出的标志
    eventHandling_(false),                               //处理handevent的标志
    callingPendingFunctors_(false),                      //判断当前是不是在执行方法队列
    iteration_(0),
    threadId_(CurrentThread::tid()),                     //当前线程ID
    poller_(Poller::newDefaultPoller(this)),             //创建一个 poll 或 epoll 对象
    timerQueue_(new TimerQueue(this)),                   //创建一个计时器
    wakeupFd_(createEventfd()),                          //发送唤醒loop消息的描述符,随便写点消息即可唤醒
    wakeupChannel_(new Channel(this, wakeupFd_)),        //wakeupChannel_用来自己给自己通知的一个通道,该通道会纳入到poller来管理
    currentActiveChannel_(NULL)                          //当前活跃的channel链表指针
{
   
  LOG_DEBUG << "EventLoop created " << this << " in thread " << threadId_;
  if (t_loopInThisThread)                                //判断是否是本线程的loop,是一个loop类型的指针
  {
   
    LOG_FATAL << "Another EventLoop " << t_loopInThisThread
              << " exists in this thread " << threadId_;        //用LOG_FATAL终止abort它
  }
  else
  {
   
    t_loopInThisThread = this; //this赋给线程局部数据指针
  }

  //设定wakeupChannel的回调函数,即EventLoop自己的的handleRead函数
  wakeupChannel_->setReadCallback(
      std::bind(&EventLoop::handleRead, this));          //channel->handleEventWithGuard会调用到handleRead
  // we are always reading the wakeupfd
  wakeupChannel_->enableReading();    //注册wakeupFd_到poller
}

EventLoop::~EventLoop()
{
   
  LOG_DEBUG << "EventLoop " << this << " of thread " << threadId_
            << " destructs in thread " << CurrentThread::tid();
  wakeupChannel_->disableAll(); //从监听队列fd里移除
  wakeupChannel_->remove();  //移除epoll里面的channel
  ::close(wakeupFd_);
  t_loopInThisThread = NULL;
}

void EventLoop::loop()
{
   
  assert(!looping_);
  assertInLoopThread(); //事件循环必须在IO线程中,即创建该evenloop的线程
  looping_ = true; //是否正在循环
  quit_ = false;  // FIXME: what if someone calls quit() before loop() ?
  LOG_TRACE << "EventLoop " << this << " start looping";

  while (!quit_)
  {
   
    activeChannels_.clear();                            //activeChannels_是一个vector
    //等待io复用函数返回
    pollReturnTime_ = poller_->poll(kPollTimeMs, &activeChannels_); //调用poll返回活动的事件,有可能是唤醒返回的
    ++iteration_;

    //根据设置的日志等级打印跟踪信息
    if (Logger::logLevel() <= Logger::TRACE)
    {
   
      printActiveChannels();
    }
    // TODO sort channel by priority  按优先级排序
    //处理IO事件
    eventHandling_ = true;
    for (Channel* channel : activeChannels_)            //遍历通道来进行处理
    {
   
      currentActiveChannel_ = channel;
      currentActiveChannel_->handleEvent(pollReturnTime_);  //pollReturnTime_是poll返回的时刻
    }
    currentActiveChannel_ = NULL;                       //处理完了赋空
    eventHandling_ = false;

    //执行方法队列中的方法[方法队列functors,我们可以跨线程的往里面添加新的方法,这些方法会在处理完io事件后执行]
    doPendingFunctors();                                //这个设计也能够进行计算任务
  }

  LOG_TRACE << "EventLoop " << this << " stop looping";
  looping_ = false;
}

void EventLoop::quit()
{
   
  quit_ = true;  //设置退出标志
  // There is a chance that loop() just executes while(!quit_) and exits,
  // then EventLoop destructs, then we are accessing an invalid object.
  // Can be fixed using mutex_ in both places.
  if (!isInLoopThread())
  {
   
    wakeup(); //唤醒
  }
}

//在I/O线程中调用某个函数
//实际上就是如果是I/O线程主动调用该函数想要执行,那就同步执行该函数。如果是其他线程施加给I/O线程的任务,那么其他线程就需要把回调函数加入I/O线程的队列,等待异步执行
void EventLoop::runInLoop(Functor cb) 
{
   
  if (isInLoopThread())  //判断是否是本线程的loop
  {
   
    cb();
  }
  else
  {
   
    queueInLoop(std::move(cb)); 
  }
}

void EventLoop::queueInLoop(Functor cb)//把方法添加到队列中,该方法会出现在多个线程中,操作要加锁
{
   
  {
   
  MutexLockGuard lock(mutex_);
  pendingFunctors_.push_back(std::move(cb));//std::function支持移动初始化,所以这里用move提升性能。(减少一次拷贝)
  }

  if (!isInLoopThread() || callingPendingFunctors_)//如果调用的queneInLoop的线程不是IO线程,那么唤醒
  {
   //如果在IO线程调用queueInLoop(),而此时正在调用pending functor,由于doPendingFunctors()调用的Functor可能再次调用queueInLoop(cb),这是queueInLoop()就必须wakeup(),否则新增的cb可能就不能及时调用了
    wakeup();
  }
}

size_t EventLoop::queueSize() const
{
   
  MutexLockGuard lock(mutex_);
  return pendingFunctors_.size();
}

TimerId EventLoop::runAt(Timestamp time, TimerCallback cb)//在指定的时间调用callback
{
   
  return timerQueue_->addTimer(std::move(cb), time, 0.0);
}

TimerId EventLoop::runAfter(double delay, TimerCallback cb)//等一段时间调用callback
{
   
  Timestamp time(addTime(Timestamp::now(), delay));//微妙
  return runAt(time, std::move(cb));
}

TimerId EventLoop::runEvery(double interval, TimerCallback cb)//以固定的间隔反复的调用callback
{
   
  Timestamp time(addTime(Timestamp::now(), interval));
  return timerQueue_->addTimer(std::move(cb), time, interval);
}

void EventLoop::cancel(TimerId timerId) //取消timer
{
   
  return timerQueue_->cancel(timerId);
}

void EventLoop::updateChannel(Channel* channel)   //更新通道,用epoll_ctl更新fd
{
   
  assert(channel->ownerLoop() == this);  //判断channel的loop是不是当前loop
  assertInLoopThread();  
  poller_->updateChannel(channel);
}

void EventLoop::removeChannel(Channel* channel) //移除通道,将channel从ChannelMap移除并EPOLL_CTL_DEL掉fd
{
   
  assert(channel->ownerLoop() == this);  //表示当前的loop
  assertInLoopThread();
  if (eventHandling_)  //正在处理channel
  {
   
    assert(currentActiveChannel_ == channel ||   //当前的channel或不是活跃的channel
        std::find(activeChannels_.begin(), activeChannels_.end(), channel) == activeChannels_.end());
  }
  poller_->removeChannel(channel);
}

bool EventLoop::hasChannel(Channel* channel)//查找事件分发器是否在channels_中
{
   
  assert(channel->ownerLoop() == this);
  assertInLoopThread();
  return poller_->hasChannel(channel);
}

void EventLoop::abortNotInLoopThread()
{
   
  LOG_FATAL << "EventLoop::abortNotInLoopThread - EventLoop " << this
            << " was created in threadId_ = " << threadId_
            << ", current thread id = " <<  CurrentThread::tid();
}

void EventLoop::wakeup()
{
   
  uint64_t one = 1;
  ssize_t n = sockets::write(wakeupFd_, &one, sizeof one);   //随便写点数据进去就唤醒了
  if (n != sizeof one)
  {
   
    LOG_ERROR << "EventLoop::wakeup() writes " << n << " bytes instead of 8";
  }
}

void EventLoop::handleRead()      //读取唤醒的数据
{
   
  uint64_t one = 1;
  ssize_t n = sockets::read(wakeupFd_, &one, sizeof one);
  if (n != sizeof one)
  {
   
    LOG_ERROR << "EventLoop::handleRead() reads " << n << " bytes instead of 8";
  }
}

// 1. 不是简单的在临界区内依次调用functor,而是把回调列表swap到functors中,这一方面减小了
//临界区的长度,意味着不会阻塞其他线程的queueInLoop(),另一方面也避免了死锁(因为Functor可能再次调用quueInLoop)
// 2. 由于doPendingFunctors()调用的Functor可能再次调用queueInLoop(cb),这是queueInLoop()就必须wakeup(),否则新增的cb可能就不能及时调用了
// 3. muduo没有反复执行doPendingFunctors()直到pendingFunctors为空,这是有意的,否则I/O线程可能陷入死循环,无法处理I/O事件
void EventLoop::doPendingFunctors()
{
   
  std::vector<Functor> functors;
  callingPendingFunctors_ = true;

  //注意这里的临界区,这里使用了一个栈上变量functors和pendingFunctors交换
  {
   
  MutexLockGuard lock(mutex_);
  functors.swap(pendingFunctors_);  //pendingFunctors_是存放Functor的vector
  }
  //此处其它线程就可以往pendingFunctors添加任务

  for (const Functor& functor : functors)
  {
   
    functor();
  }
  callingPendingFunctors_ = false;
}

void EventLoop::printActiveChannels() const
{
   
  for (const Channel* channel : activeChannels_)
  {
   
    LOG_TRACE << "{" << channel->reventsToString() << "} ";
  }
}
目录
相关文章
|
7月前
muduo源码剖析之Socket类
封装了一个sockfd相关的设置。比较简单,已经编写注释。
62 0
|
7月前
|
API
muduo源码剖析之SocketOps类
对socket设置API的封装。比较简单,已经编写注释。
45 0
|
7月前
muduo源码剖析之Acceptor监听类
Acceptor类用于创建套接字,设置套接字选项,调用socket()->bind()->listen()->accept()函数,接受连接,然后调用TcpServer设置的connect事件的回调。listen()//在TcpServer::start中调用封装了一个listen fd相关的操作,用于mainLoop接受器封装,实质上就是对Channel的多一层封装监听连接 当新连接进入时,调用Socket::accept创建套接字,触发TcpServer的回调TcpServer通过该接口设置回调,
57 0
|
7月前
|
安全 Java
【Netty 网络通信】EventLoop 事件循环对象 解析
【1月更文挑战第9天】【Netty 网络通信】EventLoop 事件循环对象 解析
|
6月前
|
API 调度
协程源码剖析(二)
协程源码剖析(二)
34 1
|
7月前
|
前端开发
muduo源码剖析之AsyncLogging异步日志类
AsyncLogging是muduo的日志,程序如果直接让文件写日志可能会发生阻塞,muduo前端设计了2个BufferPtr,分别是currentBuffer_和nextBuffer_,还有一个存放BufferPtr的vector(buffers_)。多个前端线程往currentBuffer_写数据,currentBuffer_写满了将其放入buffers_,通知后端线程读。前端线程将currentBuffer_和nextBuffer_替换继续写currentBuffer_。
81 0
|
6月前
|
Java
线程池ThreadPoolExcutor源码剖析---工作原理
线程池ThreadPoolExcutor源码剖析---工作原理
|
6月前
|
调度
协程源码剖析(一)
协程源码剖析(一)
37 0
|
7月前
Muduo类详解之EventLoop
Muduo类详解之EventLoop
|
7月前
|
安全 C++ 容器
muduo源码剖析之TimerQueue类
通过timerfd实现的定时器功能,为EventLoop扩展了一系列runAt,runEvery,runEvery等函数TimerQueue中通过std::set维护所有的Timer,也可以使用优先队列实现muduo的TimerQueue是基于timerfd_create实现,这样超时很容易和epoll结合起来。等待超时事件保存在set集合中,注意set集合的有序性,从小到大排列,整个对TimerQueue的处理也就是对set集合的操作。
55 0
muduo源码剖析之TimerQueue类