Go to main content

Oracle® Solaris 11.3 Programming Interfaces Guide

Exit Print View

Updated: April 2019

Pipes Between Processes

A pipe between two processes is a pair of files that is created in a parent process. The pipe connects the resulting processes when the parent process forks. A pipe does not exist in any file name space, so it is referred as anonymous. A pipe connects only two processes. A single pipe also connects multiple child processes to each other and their related parent.

A pipe is created in the process that becomes the parent by a call to pipe(). The call returns two file descriptors in the array passed to it. After forking, both processes read from p[0] and write to p[1]. The processes read from and write to a circular buffer that is managed for them. For more information, see the pipe(2) man page.

Calling fork() duplicates the per-process open file table. Each process has two readers and two writers. Closing the extra readers and writers enables the proper functioning of the pipe. For example, if the end of a reader is left open by the same process for writing, no end-of-file indication is returned. For more information, see the fork(2) man pages.

The following code shows pipe creation, a fork, and clearing the duplicate pipe ends.

#include <stdio.h>
#include <unistd.h>
        int p[2];
        if (pipe(p) == -1) exit(1);
        switch( fork() )
                case 0:                   /* in child */
                        close( p[0] );
                        dup2( p[1], 1);
                        close P[1] );
                        exec( ... );
                default:                  /* in parent */
                        close( p[1] );
                        dup2( P[0], 0 );
                        close( p[0] );

The following table shows the results of reads from a pipe and writes to a pipe, under certain conditions.

Table 6  Read/Write Results in a Pipe
Empty pipe, writer attached
Read blocked
Full pipe, reader attached
Write blocked
Empty pipe, no writer attached
EOF returned
No reader

Blocking can be prevented by calling fcntl() on the descriptor to set FNDELAY. This causes an error return (-1) from the I/O call with errno set to EWOULDBLOCK. For more information, see the fcntl(2) man page.