socket通信中EPIPE 错误

The send() function shall fail if:
[EPIPE] The socket is shut down for writing, or the socket is connection-mode and is no longer connected. In the latter case, and if the socket is of type SOCK_STREAM or SOCK_SEQPACKET and the MSG_NOSIGNAL flag is not set, the SIGPIPE signal is generated to the calling thread.

You can avoid this by using the MSG_NOSIGNAL flag on the send() call, or by ignoring the SIGPIPE signal with signal(SIGPIPE, SIG_IGN) at the beginning of your program. Then the send() function will return -1 and set errno to EPIPE in this situation.

 

A simple way to check would be to do a 0 byte write(2) to the pipe and check the return. If you're catching SIGPIPE or checking for EPIPE, you get the error. But that's just the same as if you go ahead and do your real write, checking for the error return. So, just do the write and handle an error either in a signal handler (SIGPIPE) or, if the signal is ignored, by checking the error return from write.   转载自：http://hi.baidu.com/tangzhenjiang/blog/item/9700f8ed7475434879f05570.html page1： 假设Server A上面有Process X，它有一个socket M，和另外的Server B上面的Process Y的 Socket N以TCP协议连接上了，那么，据我所知，有2种情况会出现RST包： （1）X没有close socket就退出了，然后Y继续向M send数据，A的内核就会发送RST 到 socket N； （2）X设置了SO_LINGER，其中l_onoff 非0， l_linger 为0，这样当A close socket M的时候，也会发送RST到socket N。 当socket N收到了RST，select的结果为socket可读，则： （a）如果这个时候调用recv，返回-1，errno为ECONNRESET，如果再次调用recv，返回-1，errno为EPIPE，同样产生EPIPE信号； （b）如果这个时候调用send，返回-1，errno为EPIPE，同时会产生SIGPIPE信号。   转载自：http://blogold.chinaunix.net/u/31357/showart_242605.html   好久没做过C开发了，最近重操旧业。 听说另外一个项目组socket开发遇到问题，发送端和接受端数据大小不一致。建议他们采用writen的重发机制，以避免信号中断错误。采用后还是有问题。PM让我帮忙研究下。 UNP n年以前看过，很久没做过底层开发，手边也没有UNP vol1这本书，所以做了个测试程序，研究下实际可能发生的情况了。   测试环境：AS3和redhat 9(缺省没有nc)   先下载unp源码: wget http://www.unpbook.com/unpv13e.tar.gz tar xzvf *.tar.gz; configure;make lib. 然后参考str_cli.c和tcpcli01.c，写了测试代码client.c     # include     "unp.h"  # define MAXBUF 40960 void processSignal( int signo)  {      printf ( "Signal is %d/n" , signo) ;      signal ( signo, processSignal) ;  }  void  str_cli( FILE * fp, int sockfd)  {      char     sendline[ MAXBUF] , recvline[ MAXBUF] ;      while ( 1) {          memset ( sendline, 'a' , sizeof ( sendline) ) ;          printf ( "Begin send %d data/n" , MAXBUF) ;          Writen( sockfd, sendline, sizeof ( sendline) ) ;          sleep (5 ) ;      }  }  int  main( int argc, char * * argv)  {      int                     sockfd;      struct sockaddr_in     servaddr;      signal ( SIGPIPE, SIG_IGN ) ;      //signal(SIGPIPE, processSignal);     if ( argc ! = 2)          err_quit( "usage: tcpcli [port]" ) ;      sockfd = Socket ( AF_INET , SOCK_STREAM , 0) ;      bzero( & servaddr, sizeof ( servaddr) ) ;      servaddr. sin_family = AF_INET ;      servaddr. sin_port = htons ( atoi ( argv[ 1] ) ) ;      Inet_pton( AF_INET , "127.0.0.1" , & servaddr. sin_addr) ;      Connect ( sockfd, ( SA * ) & servaddr, sizeof ( servaddr) ) ;      str_cli( stdin , sockfd) ;         /* do it all */      exit ( 0) ;  }    为了方便观察错误输出，lib/writen.c也做了修改，加了些日志：     /* include writen */  # include     "unp.h"  ssize_t                        /* Write "n" bytes to a descriptor. */  writen( int fd, const void * vptr, size_t n)  {      size_t         nleft;      ssize_t        nwritten;      const char     * ptr;      ptr = vptr;      nleft = n;      while ( nleft > 0) {          printf ( "Begin Writen %d/n" , nleft) ;          if ( ( nwritten = write ( fd, ptr, nleft) ) < = 0) {              if ( nwritten < 0 & & errno = = EINTR) {                  printf ( "intterupt/n" ) ;                  nwritten = 0;         /* and call write() again */              }              else                  return ( - 1) ;             /* error */          }          nleft - = nwritten;          ptr + = nwritten;          printf ( "Already write %d, left %d, errno=%d/n" , nwritten, nleft, errno ) ;      }      return ( n) ;  }  /* end writen */  void  Writen( int fd, void * ptr, size_t nbytes)  {      if ( writen( fd, ptr, nbytes) ! = nbytes)          err_sys( "writen error" ) ;  }    client.c放在tcpclieserv目录下，修改了Makefile，增加了client.c的编译目标   client: client. c                 ${ CC} ${ CFLAGS} - o $@ $< ${ LIBS}    接着就可以开始测试了。 测试1 忽略SIGPIPE信号，writen之前，对方关闭接受进程 本机服务端： nc -l -p 30000   本机客户端: ./client 30000 Begin send 40960 data Begin Writen 40960 Already write 40960, left 0, errno=0 Begin send 40960 data Begin Writen 40960 Already write 40960, left 0, errno=0 执行到上步停止服务端,client会继续显示: Begin send 40960 data Begin Writen 40960 writen error: Broken pipe(32) 结论:可见write之前，对方socket中断，发送端write会返回-1,errno号为EPIPE(32) 测试2 catch SIGPIPE信号，writen之前，对方关闭接受进程 修改客户端代码，catch sigpipe信号          //signal(SIGPIPE, SIG_IGN);         signal ( SIGPIPE, processSignal) ;  本机服务端： nc -l -p 30000   本机客户端: make client ./client 30000 Begin send 40960 data Begin Writen 40960 Already write 40960, left 0, errno=0 Begin send 40960 data Begin Writen 40960 Already write 40960, left 0, errno=0 执行到上步停止服务端,client会继续显示: Begin send 40960 data Begin Writen 40960 Signal is 13 writen error: Broken pipe(32) 结论:可见write之前，对方socket中断，发送端write时，会先调用SIGPIPE响应函数，然后write返回-1,errno号为EPIPE(32)   测试3 writen过程中，对方关闭接受进程 为了方便操作，加大1次write的数据量，修改MAXBUF为4096000 本机服务端： nc -l -p 30000   本机客户端: make client ./client 30000 Begin send 4096000 data Begin Writen 4096000 执行到上步停止服务端,client会继续显示: Already write 589821, left 3506179, errno=0 Begin Writen 3506179 writen error: Connection reset by peer(104) 结论:可见socket write中，对方socket中断，发送端write会先返回已经发送的字节数,再次write时返回-1,errno号为ECONNRESET(104) 为什么以上测试，都是对方已经中断socket后，发送端再次write，结果会有所不同呢。从后来找到的UNP5.12,5.13能找到答案 The client's call to readline may happen before the server's RST is received by the client, or it may happen after. If the readline happens before the RST is received, as we've shown in our example, the result is an unexpected EOF in the client. But if the RST arrives first, the result is an ECONNRESET ("Connection reset by peer") error return from readline. 以上解释了测试3的现象，write时，收到RST. What happens if the client ignores the error return from readline and writes more data to the server? This can happen, for example, if the client needs to perform two writes to the server before reading anything back, with the first write eliciting the RST. The rule that applies is: When a process writes to a socket that has received an RST, the SIGPIPE signal is sent to the process. The default action of this signal is to terminate the process , so the process must catch the signal to avoid being involuntarily terminated. If the process either catches the signal and returns from the signal handler, or ignores the signal, the write operation returns EPIPE. 以上解释了测试1，2的现象,write一个已经接受到RST的socket，系统内核会发送SIGPIPE给发送进程，如果进程catch/ignore这个信号，write都返回EPIPE错误. 因此,UNP建议应用根据需要处理SIGPIPE信号，至少不要用系统缺省的处理方式处理这个信号，系统缺省的处理方式是退出进程，这样你的应用就很难查处处理进程为什么退出。