muduo库网络部分

网络相关的文件简述

TcpServer.* : tcp server,用于处理客户端连接,管理客户端连接
Acceptor.* : 接受客户端连接
EventThreadLoop.* : 线程池,每个线程中有一个eventloop做事件循环
InetAddress.* : 包装struct sockaddr_in , 方便对struct sockaddr_in的使用
socketsOps.* : 对网络相关系统api封装,简化使用
socket.* : 封装socketfd,方便对socket的使用
TcpConnection.* : 客户端连接对象

TcpServer::TcpServer(EventLoop *loop, const InetAddress &listenAddr, const string &nameArg, Option option) : loop_(CHECK_NOTNULL(loop)), ipPort_(listenAddr.toIpPort()), name_(nameArg), acceptor_(new Acceptor(loop, listenAddr, option == kReusePort)), threadPool_(new EventLoopThreadPool(loop, name_)), connectionCallback_(defaultConnectionCallback), //将connect的fd添加到poller后调用,默认回调defaultConnectionCallback打印信息 messageCallback_(defaultMessageCallback), //将tcpconnect收到消息调用,默认defaultMessageCallback丢弃所有收到的消息 nextConnId_(1){ // newConnection ,acceptor 接收到连接调用,创建tcpconnect对象 acceptor_->setNewConnectionCallback( std::bind(&TcpServer::newConnection, this, _1, _2));}void TcpServer::start(){ if (started_.getAndSet(1) == 0) { threadPool_->start(threadInitCallback_); //启动事件循环 assert(!acceptor_->listening()); loop_->runInLoop( std::bind(&Acceptor::listen, get_pointer(acceptor_))); //将acceptor中fd添加到poller中,等待客户端连接 }}// 作为acceptor接收到连接后回调void TcpServer::newConnection(int sockfd, const InetAddress &peerAddr){ loop_->assertInLoopThread(); EventLoop *ioLoop = threadPool_->getNextLoop(); char buf[64]; snprintf(buf, sizeof buf, "-%s#%d", ipPort_.c_str(), nextConnId_); ++nextConnId_; string connName = name_ + buf; LOG_INFO << "TcpServer::newConnection [" << name_ << "] - new connection [" << connName << "] from " << peerAddr.toIpPort(); InetAddress localAddr(sockets::getLocalAddr(sockfd)); //获得socket服务端地址 // FIXME poll with zero timeout to double confirm the new connection // FIXME use make_shared if necessary TcpConnectionPtr conn(new TcpConnection(ioLoop, connName, sockfd, localAddr, peerAddr)); connections_[connName] = conn; conn->setConnectionCallback(connectionCallback_); // connectEstablished调用 conn->setMessageCallback(messageCallback_); //读取消息直接丢弃（默认read cb） conn->setWriteCompleteCallback(writeCompleteCallback_); conn->setCloseCallback( std::bind(&TcpServer::removeConnection, this, _1)); // FIXME: unsafe ioLoop->runInLoop(std::bind(&TcpConnection::connectEstablished, conn)); //添加任务到event loop 中}

Acceptor::Acceptor(EventLoop *loop, const InetAddress &listenAddr, bool reuseport) : loop_(loop), acceptSocket_(sockets::createNonblockingOrDie(listenAddr.family())), acceptChannel_(loop, acceptSocket_.fd()), listening_(false), idleFd_(::open("/dev/null", O_RDonLY | O_CLOEXEC)){ assert(idleFd_ >= 0); acceptSocket_.setReuseAddr(true); acceptSocket_.setReusePort(reuseport); acceptSocket_.bindAddress(listenAddr); //至此完成了socket的创建，以及初始化 acceptChannel_.setReadCallback( std::bind(&Acceptor::handleRead, this)); //设置channel中的回调，此刻完成了channel的初始化，但还没添加到poller中}Acceptor::~Acceptor(){ acceptChannel_.disableAll(); acceptChannel_.remove(); ::close(idleFd_);}// accepor开始工作void Acceptor::listen(){ loop_->assertInLoopThread(); listening_ = true; acceptSocket_.listen(); acceptChannel_.enableReading(); //添加socket_fd到poller中}//接受到新连接// socketfd上有数据，说明新链接到达void Acceptor::handleRead(){ loop_->assertInLoopThread(); InetAddress peerAddr; // FIXME loop until no more int connfd = acceptSocket_.accept(&peerAddr); // accep从该socket上获取新的fd用于和该连接的数据交互 if (connfd >= 0) { // string hostport = peerAddr.toIpPort(); // LOG_TRACE << "Accepts of " << hostport; if (newConnectionCallback_) { newConnectionCallback_(connfd, peerAddr); } else { sockets::close(connfd); } } else { // 当连接过多,导致失败时,关闭下一个连接 LOG_SYSERR << "in Acceptor::handleRead"; // Read the section named "The special problem of // accept()ing when you can't" in libev's doc. // By Marc Lehmann, author of libev. if (errno == EMFILE) //文件描述符过多,导致无法创建connfd { ::close(idleFd_); //关闭idlefd 留出fd 给accept使用 idleFd_ = ::accept(acceptSocket_.fd(), NULL, NULL); ::close(idleFd_); //然后关闭新的连接 idleFd_ = ::open("/dev/null", O_RDonLY | O_CLOEXEC); //继续打开fd } }}

这里讲解下idleFd_ 作用,我们知道对socket_fd accept可以获得一个新的fd用于和远端通信,如果远端连接过多,可能会消耗完所有的文件描述符fd。accept处理连接时，若出现 EMFILE 错误不进行处理，则内核间隔性尝试连接，导致整个网络设计程序崩溃。这时候关闭idleFd_,留出fd给accept使用,然后关闭fd,断开和远端的连接,最后再打开idleFd_ ,以此关闭超过最大数量的远端连接,保证网络不崩溃。