Customer Reviews

Linux Internals

5 star:		(2)
4 star:		(3)
3 star:		(1)
2 star:		(3)
1 star:		(4)

 

 

I was curious how one book could be oriented to both "Linux internals" and "application development", and once I started reading this book, I realized my initial hunch was correct--it can't. This book is purely about the internals of the Linux kernel, period. That's not a bad thing, but people looking for a reference for app. development under Linux will be very disappointed.

The author has an extremely terse style, which I found very annoying. Reading this book is like spending hours listening to one of your arrogant, techie friends who loves to explain things just well enough to impress you with his/her knowledge, but not well enough to help you understand the topic fully.

The book's cover blurb says that it covers kernel 2.4, but it doesn't mention devfs, easily one of the more radical and significant changes in the release.

Finally, I found numerous minor glitches throughout the book, like the table that lists the "bite size" of various architectures, when clearly "word size" was the intention.

Here are some examples of the errors in this book; I gave up on it around page 50. It tries to distinguish between RISC and CISC but gets them backwards. Most of the underscores in the code listings have been replaced by spaces. It states that linked lists and arrays are the same thing. A table refers to the "Bite size" of processor architectures. Structurally, it jumps quickly into a deep description of x86 instruction scheduling without explaining why. There's some useful information in there, if you can find it through the errors and poor organization. There are better books on the topic.

This book was not edited. I believe they folded up the second draft and mailed it to bookstores. The author is also a very poor writer. Portions of the text remind me of the "All your base are belong to us" funnies or similar jokes about bad translations. I recommend that no one buy it; that the author be vigorously slapped with a large wet fish; and that the publishers and editors be exiled to some land without books or paper.

This is the worst technical book I have EVER purchased.

This book is a mess. It does not appear to have been seen by a copy editor, as it is full of spelling mistakes and figure captions like "Please provide figure caption."

It is also full of mistakes; even sample code is terribly defective, and wouldn't even compile properly with recent kernels. As one trivial example, the return type of the function "cleanup_module()" is given as an "int" instead of a "void".

A large fraction of the book is given over to pasting in large sections of the Linux kernel source; I wouldn't be surprised if these pages were one third of the total. Given the fact that the kernel source changes all the time, this code was already out of date when the book was published. (This wouldn't be a problem if the author had restricted himself to a stable production kernel of the 2.2 series, but the 2.4 kernel is prominently advertised on the cover.) This wholesale lifting and publishing of unaltered kernel code may not violate the GPL, but it is idiotic as any one following Linux kernel development knows one always has to look at the source itself to be accurate.

I believe the wholesale inclusion of out of date source was done to pad out this rather flimsy book.

Overall the author's discussion is incomprehensible, without any clear plan. I feel sorry for Ingo Molnar who must have been duped into writing a forward; surely someone of his talent could not have looked at the whole book and said something positive.

Don't buy this book. Complain to the publisher that Linux is not well served by the publication of pasted together quickie books.

This book is worse than worthless; it is misleading and inaccurate.

This is a surprising readable book considering that highly technical topics such as this are often difficult to write about. It is certainly a much better read than the other books about the Linux kernel currently in print. As the author points out in the preface, this book is not intended as a full source code commentary of the kernel, rather it focuses upon describing those parts of the kernel that are relevant to performance issues of the OS and user applications. Certainly some esoteric topics of particular interest to kernel hackers are likely missing but on the other hand more practical topics such as filesystems (including JFS and LVM) and signal handling are well covered.

The author's writing style can be a bit obtrusive at times and it would have been nice if a good technical editor had also corrected some of the more glaring errors such as "next'ed" instead of "nested" or the overuse of the adverb "excellently". The extracts of code from the kernel are generally well selected and short enough to comprehend along the with associated discussion.

Overall, I would recommend this book to someone already knowledgeable with Linux and familiar with C programming who would like to obtain a better grasp of how the kernel is designed and what the resulting implications are for the performance of the system.

Previous to this book I did not know what makes a Linux kernel. Sure, I had read various articles here and there. But this book really explained the inner workings of the new Linux 2.4.0 version to me in easy language. I had to read it several times because some of the concepts of kernels really are challenging, at least to me. Now, I know where to go look in the source code whenever I want to get deeper knowledge.

The author did a good job explaining even elementary kernel funcitonality and basic OS theory. The part on journalling file systme is the first comprehensible explanation of the subject I have seen.

This book is not a good reference for kernel internals in Linux. It has a very superficial treatment of the majority of topics. If you're expecting to use this book like the famous UNIX "Magic Garden", forget it. This book does not have any depth.

This book is for people who know the basics of an operating system. If you are one of these, then you will enjoy the no-frills to-the-point style of the author. He succeeds in explaining some of the darkest spots of Linux. It helps to be fluent in C. All the fundamental building blocks of Linux are mentioned. It's not easy reading, but then a kernel is not a trivial subject either.

This book has some good points, but mostly sucks. A lot of what you'll get out of it is dependent on what you know. If you're a newbie, it's going to screw you up. The book has obviously been through a very poor editing process, and there are plenty of mistakes that will confuse you and lead you down the wrong path. Some of this is extremely evident (in the beginning the author makes a personal note to remember to explain the origins of Linux w/ some background on Minix thrown in, then injects this info into the middle of his topic). If you're just past the novice user stage and entering development, there are a few gems of info, but really nothing that would warrant the books price, ... list, or even Amazon's discounted price ... However, if you can pick up a copy for less than 15 bucks, I'd recommend it. If your past the new developer stage, well, prepare to be disappointed. The author wrote a very short book then bulked it up by including code from the header files of a linux distribution. The rest isn't really worth reading, and overall not reliable enough to use as a reference guide.

Luckily, I bought this book from Borders. Started reading it and realized I'd been scammed. Finished reading it (takes about 15 hours to realy give it a fair shake), and after I got done laughing at all the mistakes make, ...

As the introduction to the book says (by a RedHat kernel developer), this book is a "reference work to be used to look up the Linux Kernel concepts." The author mostly breezes through the basics and jumps right into the source code and hi-powered concepts such as SMP.

For a primer to the Linux source code and concepts, I found "Linux Kernel Internals" by M. Beck et al. far more useful (although it only covers Linux 2.0). That book goes through such procedures as compiling and debugging the kernel, which I really needed to know. This is not to discredit "Linux Internals", it just seems like the two books target different audiences.