[ros-kernel] PipeSpy

KJK::Hyperion noog at libero.it
Wed Feb 11 00:11:21 CET 2004


At 09.14 09/02/2004, you wrote:
>I have just committed a named pipe spy program called PipeSpy, written by 
>Skywing.  It is intended to be used to help understand the Windows remote 
>kernel debugging protocol, so that ReactOS can implement it some day.

kd protocol? ask and you shall receive. I have good news and bad news, 
however. The good news is another article from our Cuban hero Albert Yulien 
Almeida - you may remember him for the best article about APCs ever 
written. Well, he did it again! this time he has documented most of the KD 
protocol and KD stub internals:

<http://www.vsj.co.uk/articles/display.asp?id=265>

The rest of the protocol is sort-of-documented in windbgkd.h, from the 
Windows 2000 DDK (not available in later versions). And now the bad news: 
I've had a look at said header, and found that the KD protocol is *highly* 
specific to the Windows NT kernel, to the point that many fields in the 
protocol structures contain the addresses of internal and obscure 
kernel-mode objects. KD may or may not depend on all or many of them to 
work correctly (or at all), but my guts say it surely does depend from them 



More information about the Ros-kernel mailing list