man.bsd.lv manual page server

Manual Page Search Parameters

FORK(2) System Calls Manual FORK(2)

forkcreate a new process

library “libc”

#include <sys/types.h>
#include <unistd.h>

pid_t
fork(void);

() causes creation of a new process. The new process (child process) is an exact copy of the calling process (parent process) except for the following:

Upon successful completion, fork() returns a value of 0 to the child process and returns the process ID of the child process to the parent process. Otherwise, a value of -1 is returned to the parent process, no child process is created, and the global variable errno is set to indicate the error.

fork() can create severe issues for multi-threaded programs due to the fact that the memory state of the child process will record the asynchronous state of the threads that are running in the parent. Fork() will only be synchronous for the specific thread making the call. In particular, locks used internally by pthread(3) and rtld(1) can be caught in a bad state. To deal with these issues, the pthreads library goes to great lengths to synchronize internal locks when a fork() call is issued. The threaded program itself as well as third party libraries used by the program might or might not properly handle these issues when it comes to their own internal state.

If at all possible, programs should use vfork(2) instead of () when forking for the purposes of issuing an exec of some sort. Attempting to fork a threaded program without issuing an exec is not recommended. Attempting to bypass pthreads and implement threading manually is also not recommended as it is doubtful that homegrown implementations could properly deal with rtld races.

Fork() will fail and no child process will be created if:

[]
The system-imposed limit on the total number of processes under execution would be exceeded. The limit is given by the sysctl(3) MIB variable KERN_MAXPROC. (The limit is actually ten less than this except for the super user).
[]
The user is not the super user, and the system-imposed limit on the total number of processes under execution by a single user would be exceeded. The limit is given by the sysctl(3) MIB variable KERN_MAXPROCPERUID.
[]
The user is not the super user, and the soft resource limit corresponding to the resource parameter RLIMIT_NPROC would be exceeded (see getrlimit(2)).
[]
There is insufficient swap space for the new process.

execve(2), rfork(2), setitimer(2), setrlimit(2), vfork(2), wait(2)

A fork() function call appeared in Version 6 AT&T UNIX.

June 4, 1993 DragonFly-5.6.1