Question

I’m using FIFO for IPC. However they perform some strange behaviour.
As for demostration I post some code here which can be compiled and runned.
Additional info, I’m on Linux Ubuntu, with the basic g++ compiler comes with it, nothing special.

#include <iostream>
#include <sys/stat.h>
#include <stdio.h>
#include <fcntl.h>
#include <iostream>
using namespace std;

int main() {

    cout << "START" << endl;

    int fifo;
    int code;
    mkfifo("/tmp/FIFO", 666);
    if ((fifo = open("/tmp/FIFO", O_RDONLY | O_NONBLOCK)) < 0) {
        perror("open failed");
    }
    else {
        cout << "open successed" << endl;
        code = close(fifo);
        cout << "close: " << code << endl;
        if ((fifo = open("/tmp/FIFO", O_WRONLY | O_NONBLOCK)) < 0) {
            perror("reopen failed");
        }
        else {
            cout << "reopen successed" << endl;
            code = close(fifo);
            cout << "close: " << code << endl;
        }
    }

    cout << "END" << endl;
    return 0;
}

My expectation about the output is something like this, since I close it successfully:

START
open successed
close: 0
reopen successed
close: 0
END

However, I got this, the second open fails. Why? And why with this silly error message?

START
open successed
close: 0
reopen failed: No such device or address
END

I really would like to reopen the FIFO for write. And I’d like know the reason why the code above doesn’t work.

Answer 1

Remove the O_NONBLOCK flag from reopen or if you want to keep the flag, connect to a fifo that is already open for reading.
It is the correct behavior.

http://linux.die.net/man/7/fifo

A process can open a FIFO in nonblocking mode. In this case, opening for read only will >succeed even if no-one has opened on the write side yet, opening for write only will fail >with ENXIO (no such device or address) unless the other end has already been opened.