Notes by Lex Toumbourou

When a process spawns a separate process to handle some work, it invokes the fork(2) system call. fork(2) duplicates the current process in memory, and both processes continue executing from the current location.

We refer to the new process as the child process. We refer to the calling process as the parent. See man two fork for information.

In the parent process, the fork(2) call returns the process id, aka the PID, of the child process. The same call from the child returns 0.

Python provides a wrapper around the call called fork() from the os module.

In this example, I write a simple script that calls fork and then prints whether it's in the child process or the parent.

from os import fork
from time import sleep

# Fork the process
pid = fork()

if pid == 0: # We're in the child process
    print("Child: Waddup?")
else: # We're in the parent process
    print("Parent: I just created child", pid)

    # Sleep for a second to avoid being dropped back to the shell when the parent finishes
    sleep(1)

Now, when I run the script in the terminal, we should see the child process' PID followed immediately by the code executed in the child process.

> python fork.py
Parent: I just created child 19478
Child: Waddup?

We can get the parent process' PID from the child process by calling the getppid(2) system call. Which Python provides a wrapper around called os.getppid().

from os import fork, getppid
from time import sleep

pid = fork()

if pid == 0: # We're in the child process
    print("Child: Waddup?")
    print("Child: My parent is", getppid())
else:
    print("Parent: I just created child", pid)
    # Sleep for a second to avoid being dropped back to the shell when the parent finishes
    sleep(1)

> python fork_ppid.py
Parent: I just created child 19741
Child: Waddup?
Child: My parent is 19740

Of course, we can have more than one child process; we can have as many as the operating system allows.

mpages (2021)

Backlinks

• Orphan Processes