GNU C Compiler Internals

What is GCC? edit

GNU Compiler Collection (GCC) is a free software project that includes compilers for Ada, C, C++, Fortran, Java, and Objective-C, as well as libraries for these languages. It is capable of generating executables for a variety of platforms including x86, ARM, MIPS, PowerPC, etc.

History of GCC edit

The homepage of GCC is http://gcc.gnu.org. The modern history of GCC begins with the GCC 2.95 release. This version was released on July 31, 1999. GCC 3.0 which is considered modern history for the C++ compiler was released on June 18, 2001. Additional branches were created later on. As of now, the active development branches are GCC 3.4 with a latest release on November 30, 2005, and GCC 4.0 released last time on September 28, 2005. GCC 4.1 was released on Feb 28 2006. GCC 4.2 is the development branch of GCC. The source code repository is available online. GCC 4.2 (the trunk) is the only place where real development happens and new features can go.

GCC has become popular in industry and academia. The availability of its source code allows one to add new features to the compiler. GCC is used in several source-code based security projects, that is, in the tools that instrument the source code of the program to make it more secure. However, very few documents describing the GCC internals have been published so far.

When new functionality is implemented, the source code of GCC is modified directly. However, these compiler extensions are difficult to distribute because GCC is a really big program. A framework for creating modular extensions would greatly simplify the development of compiler extensions.

GCC 4.0 and above includes SSA optimizers and should be used for further high level optimizations and transformations. The RTL level should only be used for target specific optimizations and optimizations which are low level such as scheduling. Any GCC before 4.0 is in some minds an old dated compiler which was showing its age.

Purpose of this book edit

The purpose of this book is to address the demands of GCC hackers. We start with a description of GCC 3.4.1 architecture focusing on the source code parser. We chose this version of GCC because we used this version mostly. Then we address the problem of extension development. We present the GCC Extensibility Modules (GEM) project in the next chapter. GEM provides a number of hooks throughout GCC source code. It is implemented as a patch to GCC. A GEM-based compiler extension is developed as a stand-alone program. When the extension is completed, only its source code is distributed compared with distributing the source code of the GCC if GEM is not used. We give examples that demonstrate GEM programming at the end of the book.

Take home: What this book is: a guide to GCC internals for extension programmers. GEM framework for developing modular extensions is introduced. What this book is not: a programming language reference, GCC installation manual.


This book describes two releases of GCC: 3.4 and 4.1.