Understanding the LINUX Kernel: From I/O Ports to Process Management

Understanding the Linux Kernel is intended to be read by those who are happy to check points off against the source code. The first thing you learn is how Linux, released from commercial constraints, is able to take advantage of the best ideas from other systems, implemented in wonderfully flexible ways. A good example is the Virtual File System (VFS), which has made it easy to add support for file systems from almost every other OS. It's fascinating to find out how such features are implemented. Then, there are loadable modules, I/O, scheduling, multitasking, multiprocessing, interrupts, spin locks, semaphores, and all of the other goodies that are involved in making a kernel work.

The authors are concerned primarily with the Linux 2.2 kernel. They discuss how Linus Torvald's decisions on kernel issues translate into architecture; for example, how the Linux memory management uses a slab allocator on top of a buddy system for greater efficiency. Similarly, at the cost of a little complexity, the decision to use three-level memory paging, when two work fine on 32-bit systems, makes it possible to port to 64-bit processors without changes. The tradeoffs between complexity and efficiency are discussed for most kernel features, and each chapter finishes with related new features in kernel 2.4.

Despite the lucid and knowledgeable writing, you'll come up against some brain-stretching complexity. Nevertheless, this book is an important addition to the Linux canon. --Steve Patient, Amazon.co.uk