Customer Reviews

Windows 2000 Kernel Debugging

5 star:		(0)
4 star:		(1)
3 star:		(1)
2 star:		(1)
1 star:		(5)

 

 

The book is essentially useless. Claiming to address itself to administrators and developers alike, it manages to satisfy neither.

The book explains on 160 (one hundred and sixty!) pages how to configure NT to produce a crash dump file; how to read a BSOD; how to run dumpexam; how to fire up a debugger; and how to get Windbag to run a debug session. Oh, I forget -- there are a few pages on the Driver verifier, too. The other 140 pages are a summary of Windbag commands (outdated) and a list of bugcheck codes and NTSTATUS values, both badly formatted, outdated versions of the corresponding header files.

This reviewer had expected all of the above to take, oh, 50 pages at the outside, with the rest of the book devoted to common debugging scenarios -- why does my driver go bang with a 0x1E bugcheck? how do I find and eliminate a deadlock? what did I do wrong in my IRP canceling code?

None of that is in there; and what _is_ in the book can be found in the DDK and Windbag docs, better written and more asily digested.

Felix Kasza.

This book does not teach you how to debug. It's essentially what the debugger documentation should have been 2 years ago.

If you have never done any kernel debugging, this is a good starting point that will give you an overall undertanding of the process and the tools. However, now that Microsoft has rewritten all the debugger documentation, most of this information comes with the online documentation.

The most unfortunate thing in my mind is that the most important chapter - remote debugging - has a major mistake in it: Figure 8-2 is wrong and will totally confuse the reader. Figure 8-2 should have the HOST machine located between the REMOTE and the TARGET machine.

This book is all about how to set up the debugger and get some basic information on the error.

WinDbg documentation is much much better than this. If you want some good introductory/advanced information about Kernel Debugging try getting hold of DebugFest materials from Microsoft, sells for some $200 as a kit. Thats a wonderful material on Kernel debugging.

This book deservs 0 stars. Only this i cant rate as zero. Complete watse of money.

If you are looking for some meat about debugging then this is NOT your book. IT's not beyond a debugger documentation.

This book isn't really about "how to debug" - it's more a reference on the Microsoft toolset. Prentice-Hall seems pretty bad about hype on the cover. But in that context, it isn't bad. The appendixes are fairly detailed on stop codes, error codes, and as a command reference. The chapters are pretty well written. It's obvious in places that the author moved from describing NT4 to describing Windows 2000 a little carelessly. He skims over some of OEMTOOLS stuff (like driver verifier), and is a little terse on how to write debugger extensions. The sections on setting up and using the tools is good, the sections on remote debugging seem useful. Lot of good stuff, even if not as deep as could be. If you're looking for tutorial on how to isolate deadlocks by tracking through thread structures for locks, then ain't it. Take the osr course ... . Is it worth my $40 as a driver developer? I think it is.

Don't even bother borrowing this book from the library. It is a waste of shelf space. My first clue should have been the book's website no longer exists.

The debugging tools mentioned throughout the book have been updated (by several years) so the detailed information about their use is useless.

The new versions of debugging tools come with help files that outclass this book at every turn. In fact, the book appears to be a poor interpretation of help files from previous versions of the tools.

Upon close examination, the book is also full of seemingly minor errors that would cause great confusion to a reader trying to learn about debugging for the first time.

I made it to Chapter 7 (skimming Chapters 5 & 6) before I decided to recycle (as in compost) my copy.

The book is essentially useless. Claiming to address itself to administrators and developers alike, it manages to satisfy neither.

The book explains on 160 (one hundred and sixty!) pages how to configure NT to produce a crash dump file; how to read a BSOD; how to run dumpexam; how to fire up a debugger; and how to get Windbag to run a debug session. Oh, I forget -- there are a few pages on the Driver verifier, too. The other 140 pages are a summary of Windbag commands (outdated) and a list of bugcheck codes and NTSTATUS values, both badly formatted, outdated versions of the corresponding header files.

This reviewer had expected all of the above to take, oh, 50 pages at the outside, with the rest of the book devoted to common debugging scenarios -- why does my driver go bang with a 0x1E bugcheck? how do I find and eliminate a deadlock? what did I do wrong in my IRP canceling code?

None of that is in there; and what _is_ in the book can be found in the DDK and Windbag docs, better written and more easily digested.

Felix Kasza.

This book was a real disappointment. The book is 300 pages long, over 130 pages is a reference on WinDbg and Stop Codes. The remainder of the book(about 170 pages) is divided into 11 chapters which gives only a basic overview of kernel debugging. If you have ANY experience using WinDbg you will find this book almost worthless.