Customer Reviews

The Basic Kernel: Source Code Secrets

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

 

 

This is a very nice book filling a long-empty niche: the low-level description of the logic and algorithms of an operating system. Probably the Lions' book was the last widely known good example of this kind. Well, 386BSD is dead but FreeBSD, NetBSD and OpenBSD are its direct descendants. And although they have diverged they still have a lot in common with 386BSD. The older versions of them are very close to 386BSD and still can be downloaded for free. It's not a general book on the principles of the OSes. It's not a high-level description of abstract Unix. It's a down-to-the-Earth discussion of the routines in the BSD kernel and the explanation of the design decisions in its base. It hits the information vacuum between the high-level descriptions, such as Vahalia's and McKusick et al. books and the raw source code. If you ever read the BSD source code and wondered what does such-and-such function do and when is it used and why does it do it in this way but not in some another way and how is it connected to the high-level picture then this is a just right book for you. But it's not an introductory reading, before reading it better get the general picture of what's going on in there from the McKusick's and Vahalia's books.

Few places where you can go and get a straight answer about how the code works
and why was it written that way. Hated the other BSD books because they promised
me to connect theory to practice, but never even got as far as clearly explaining
theory and need, let alone the details of making it work. A "Way things Work" for kernel's.

This book is destined to become a classic. Jolitz and operating
systems is like Knuth and algorithms. Its a shame they provoked
so much envy by sharing the secrets of kernel design that most
still don't know.

If you are looking for a book on OS design, this is most certainly not the one. However, this book may come in handy for UNIX-like system implementors on the Intel architecture. Despite what the introduction of the book reads, this is mostly just source code, and few comments, most of which I would expect to see /* */ around -- no incredible insight here. The only real reasons one might get this book is if one is a *BSD fan or if one is implementing an OS on the Intel architecture and is looking for ways other people have resolved related issues.

This was an interesting book. Its hard to know exactly what this book is trying to be. Is it trying to be like the Lions Commentary? nope.. since its not complete source code...

Basically they give you most of the core kernel functions etc, locore.s, and give you the design decisions, whys and hows.

I found it to be done pretty well, and it made a good reference, but one needs to remember the time in which 386BSD was written in.

Overall, a good book for any unix coder os design person.

The title of this book is somewhat misleading, this book is a discussion on some of the source code from the 386BSD kernal,unfortunately,that Operating System is of no real interest, and has for the most part been overshadowed by other,more up-to-date OS's such as FreeBSD, with development being abandoned by its author Bill Jolitz who is also this books author. No OS theory is present, and there are frequent references to articles published in old issues of Dr Dobbs journal, as well as another book called "386BSD From the Inside Out" which are neccessary to understanding key discussions. The most serious drawback is that 386BSD is dead, therefore reading about its kernal's implementation is for the most part a waste of time. Further, the kernal as presented in the book is not a complete kernal, and without the system's full source code, (which will cost you an additional 74$) the utility of this book is greatly diminished. I credit the author in wishing to p! rovide the reader with a sketch of a working kernal and documentation, but the discussion is rather unclear as it reads like a loosly organized set of notes. Most of the routines are presented by questions like What is such-and-such function, and how is it implemented? The actual kernal design is not mentioned nor is an overview given. There is a set of other books which are supposed to complement this book by describing virtual memory, file systems,etc; however, these have not yet been published, which makes for a giant time differential between the basic kernal discussion and other parts of the OS. In summary, avoid this book unless you want to learn about a dead OS and a couple of people struggling to document it.