errno set in child process after fork - OSX

Question

here's a weird thing I found today on Mac OSX.

After a fork, which has succeeded, errno is set at 0 in the father's process (as expected), but set at 22 in the child process. Here's the source-code :

#include <stdio.h> 
#include <stdlib.h> 
#include <unistd.h>
#include <errno.h>

int main(int nbArgs, char** args){
   int pid;
   errno = 0;
   printf("Errno value before the call to fork : %d.\n", errno);
   if ((pid = fork()) == -1){
      perror("Fork failed.");
      exit(1);
   }
   if (pid == 0){
      printf("Child : errno value : %d.\n", errno);
   }else{
      printf("Father : pid value : %d ; errno value : %d.\n", pid, errno);
      wait(NULL);
   }
   exit(0);
}

And the execution track :

Remis-Mac:TP3 venant$ ./errno_try
Errno value before the call to fork : 0.
Father : pid value : 9526 ; errno value : 0.
Child : errno value : 22.

As far as I know, and according to the Opengroup specifications, "The new process (child process) shall be an exact copy of the calling process (parent process) except as detailed below [...]", including the value of the global variable errno -_-

Does anyone have a clue to explain that undesired behavior ?

Answer 1

Curious...I can reproduce the problem on Mac OS X 10.9 Mavericks with GCC 4.8.2 and with Clang.

POSIX says that some functions that fail will set errno (and fork() is one of those functions), but does not say that functions that succeed will not set errno. For example, on Solaris, many standard I/O functions set errno if the output stream is not a terminal. However, resetting errno = 0; after the printf() doesn't alter the behaviour on Mac OS X.

POSIX 2008 (System Interfaces — General Information: 3. Error numbers):

Some functions provide the error number in a variable accessed through the symbol errno, defined by including the <errno.h> header. The value of errno should only be examined when it is indicated to be valid by a function's return value. No function in this volume of POSIX.1-2008 shall set errno to zero. For each thread of a process, the value of errno shall not be affected by function calls or assignments to errno by other threads.

If fork() failed, then errno would be set to indicate the failure. When it succeeds, it is not technically valid to inspect errno. And this is a demonstration of why.

Answer 2

POSIX says:

Upon successful completion, fork() shall return 0 to the child process and shall return the process ID of the child process to the parent process. Both processes shall continue to execute from the fork() function. Otherwise, -1 shall be returned to the parent process, no child process shall be created, and errno shall be set to indicate the error.

which means errno is specified to be set on success only. Otherwise nothing is specified, so you can expect an old value of errno or any unspecified value.