Hands-On System Programming with Linux

The Unix fork-exec semantic

So, that does not work. What we need, really, is a way for our simple shell simpsh to remain alive and well after the exec operation, but how can we achieve that?

The fork is the answer! Here's what we'll do: after the user supplies input (a command), we have our shell fork. We now have two identical shells alive: the original parent (let's say it has PID x) and the brand new child shell (PID y). The child shell is used as the sacrificial lamb: We have it exec the user command. So, yes, the child is the predecessor process that is impossible to return to; but that's OK as we have the parent shell process alive and well!

This well-known technique is called the fork-exec semantic. It combines what several other OSes ...