Book Image

Linux System Programming Techniques

By : Jack-Benny Persson
Book Image

Linux System Programming Techniques

By: Jack-Benny Persson

Overview of this book

Linux is the world's most popular open source operating system (OS). Linux System Programming Techniques will enable you to extend the Linux OS with your own system programs and communicate with other programs on the system. The book begins by exploring the Linux filesystem, its basic commands, built-in manual pages, the GNU compiler collection (GCC), and Linux system calls. You'll then discover how to handle errors in your programs and will learn to catch errors and print relevant information about them. The book takes you through multiple recipes on how to read and write files on the system, using both streams and file descriptors. As you advance, you'll delve into forking, creating zombie processes, and daemons, along with recipes on how to handle daemons using systemd. After this, you'll find out how to create shared libraries and start exploring different types of interprocess communication (IPC). In the later chapters, recipes on how to write programs using POSIX threads and how to debug your programs using the GNU debugger (GDB) and Valgrind will also be covered. By the end of this Linux book, you will be able to develop your own system programs for Linux, including daemons, tools, clients, and filters.
Table of Contents (14 chapters)

Exploring how processes are created

Before we go into the details of creating processes and daemons, we need a general understanding of processes. The best way to get this understanding is by looking at the processes already running on your system, which is what we are going to do in this recipe.

Every process on the system has started its life by being spawned—forked—from another process. The very first process to be used on Unix and Linux systems has historically been init. The init process has been replaced in modern Linux distributions with systemd. They both serve the same purpose; to start the rest of the system.

A typical process tree may look like this, where a user has logged on via a terminal (that is, if we skip the complexity of X Window logons):

|- systemd (1)
  \- login (6384)
    \- bash (6669)
      \- more testfile.txt (7184)

The process IDs are the numbers in parenthesis. systemd...