即使连接不再可用，发送也会成功？

Question

我错过了什么？

我有非常简单的客户端和服务器。服务器使用recv 超时（使用select）3 秒。然后它shutdowns 和closes 连接套接字并退出。

客户端连接到服务器，休眠 30 秒并发送非常短的消息。服务器关闭套接字并退出后大约 27 秒发送。
而send 不会失败..？为什么？为什么它不返回-1 错误？

---

请注意：我取消了所有对返回代码的检查并删除了所有日志，以尽可能缩短此内容。此外，我删除了所有包含，以使其更短。但它是一个真实的代码。

客户端代码：

int main( int argc, char* argv[] )
{
    addrinfo hints;
    memset(&hints, 0, sizeof hints); // make sure the struct is empty

    hints.ai_family = AF_INET;
    hints.ai_socktype = SOCK_STREAM; // TCP stream sockets

    addrinfo *res;
    getaddrinfo( "127.0.0.1", "1313", &hints, &res );
    int nSocketFD = socket( res->ai_family, res->ai_socktype, res->ai_protocol );
    assert( -1 != connect( nSocketFD, res->ai_addr, res->ai_addrlen) );
    freeaddrinfo( res ); // free the linked-list, we don't need it anymore

    sleep( 30 );
    if( send( nSocketFD, "bla", 4, 0 ) > 0 )
    {
        printf( "Message successfully sent!\n" );
    }   

    close( nSocketFD );

    return 0;
}

和服务器：

int main()
{
    addrinfo hints;
    memset(&hints, 0, sizeof hints);
    hints.ai_family = AF_INET;
    hints.ai_socktype = SOCK_STREAM;
    hints.ai_flags = AI_PASSIVE;

    addrinfo *res;  // will point to the results
    getaddrinfo( NULL, "1313", &hints, &res );
    int nSocketFD = socket( res->ai_family, res->ai_socktype, res->ai_protocol );
    bind( nSocketFD, res->ai_addr, res->ai_addrlen );
    freeaddrinfo( res ); // free the linked-list
    listen( nSocketFD, 1 );

    sockaddr_storage their_addr;
    socklen_t addr_size = sizeof( sockaddr_storage );
    int nAcceptedSocket = accept( nSocketFD, (sockaddr*)&their_addr, &addr_size );
    assert( -1 != nAcceptedSocket );

    fd_set fds;
    FD_ZERO( &fds );
    FD_SET( nAcceptedSocket, &fds );
    timeval tv; 
    tv.tv_sec = 3;
    tv.tv_usec = 0;

    if( 0 == select( nAcceptedSocket + 1, &fds, NULL, NULL, &tv) )
    {   
        printf( "recv timeout! Exiting..\n" );
        shutdown( nSocketFD, SHUT_RDWR );
        close( nSocketFD );
        shutdown( nAcceptedSocket, SHUT_RDWR );
        close( nAcceptedSocket );
        return 1;
    }   
    assert( false );
    return 0;
}

当我执行它时，我会看到 recv 超时和成功发送消息的消息。

很抱歉这个冗长且可能很愚蠢的问题。

Answer 1

一般来说，你需要从一个套接字中读取，让它注意到远程端已经关闭了连接。

来自send 的手册页（只是write，但带有标志）：

No indication of failure to deliver is implicit in a send().  Locally
detected errors are indicated by a return value of -1.

Answer 2

您没有关闭客户端中的套接字 - 所以这是一个有效的套接字。来自 send() 的手册页

No indication of failure to deliver is implicit in a send().  Locally detected errors are indicated by a return value of -1.

我们的某些软件存在这个问题 - 如果网络硬件在某处出现故障并再次恢复正常 - 那么任一端仍然认为套接字是有效的 - 并且没有 keepalives 探测 - 它将保持这种状态。

添加： 看看套接字选项（man setsockopt 和 man tcp）SO_KEEPALIVE

来自 TCP 手册页 tcp_keepalive_intvl（整数；默认值：75；从 Linux 2.4 开始） TCP keep-alive 探测之间的秒数。

   tcp_keepalive_probes (integer; default: 9; since Linux 2.2)
          The  maximum  number  of  TCP  keep-alive  probes  to  send before giving up and killing the connection if no response is obtained from the other end.

   tcp_keepalive_time (integer; default: 7200; since Linux 2.2)
          The number of seconds a connection needs to be idle before TCP begins sending out  keep-alive  probes.   Keep-alives  are only  sent when the SO_KEEPALIVE socket option is enabled.  The default value is 7200 seconds (2 hours).  An idle connection is terminated after approximately an additional 11 minutes (9 probes an interval of 75  seconds  apart)  when  keep-alive is enabled.

          Note that underlying connection tracking mechanisms and application timeouts may be much shorter.