TL;DR User programs issue system calls to use the services provided by the operating system. Such services include interacting with the hardware devices.
The operating system is the interface between a user program and the hardware devices. The OS provides services to the user programs. The services are provided through system calls. To use hardware resources, users programs invoke system calls.
These are examples where systems calls are used:
- Wait for keyboard inputs
- Print messages on the monitor
- Send out a network packet.
The kernel provides code libraries or APIs for user programs to invoke system calls.
What happens exactly when a system call is invoked?
- Those libraries eventually trigger a software interrupt.
- There's a context switch from the user space to the kernel space. The kernel will then take control.
- There's a system call handler for the kernel to handle different kinds of system calls. The kernel takes action on behalf of the user programs.
- Once the operation is complete, the kernel passes the control back to the user program and the user program resumes execution.
How do we typically launch a program?
- We start by running bash commands or clicking some icons.
- The
shell program handles the user commands and starts a child process via a fork system call.
- The child process inherits quite a lot of attributes, including memory allocation, file descriptors, environment variables, etc.
- The parent and child process both proceed from the same code location. But what follows is a conditional statement that separates the two processes. The parent process moves on but the child process issues a
exec system call.
- With the
exec system call, the operating system finds the executable, loads the program into memory, and initiates the program states (memory, stack, heap).
- Now the child process jumps to the entry point of the program and starts running!