Moin Mike,

08.06.95 16:38, you wrote a message to John Poltorak:

[...]

 MB> The opinions I formed in due course are these:

 MB> 1.  Mastrianni's book is totally useless.  Much of it is an 
 MB> abbreviated regurgitation of the official IBM docs, but the 

Agree!

 MB> 2.  The official IBM docs are remarkably good.  They are in 
 MB> reference form, but they have very few inaccuracies and are 
 MB> quite reliable.

Agree

 MB> 3.  There are several books which device driver writers would 
 MB> find useful, and these are not obvious.  (I should probably 
 MB> sit down and compile a list.)  

Please, please do that!


[...4...]

 MB> 5.  If someone ever does write a device driver book, they need to 
 MB> concentrate on an architectural overview of OS/2 internals, 

Do you know such a book? Should be very technical, is 'Inside OS/2' a book
to think of?

 MB> not on the DevHelp API.  How does memory work?  When and how 
 MB> do you lock and unlock memory?  What are you really looking 
 MB> at with the debugging kernel?  What arbitrary conventions 
 MB> are used inside OS/2?  What happens from the driver point of 
 MB> view whan an application calls DosOpen(), DosRead(), 
 MB> DosDevIOCtl(), and so on?  

 MB> This is the stuff that is either lacking or buried in the 
 MB> IBM docs.

The IBM docs assume you have written some (real) MS-DOG drivers which would
lay a steady ground for all but protected mode issues. The IBM docs are on
the other hand the best references for IOCtl under MS-DOG (format, etc.).

The best information I ever had were 'Physical/Virtual Device Driver
Reference for OS/2 2.0' in _paper_. These books together with 4 disks from
IBM (beta release of DDK) explained all (my ADD [5 controller] is still in
version 1.22...).

Bye,
	Vitus

--- Sqed/32 0.999/r00228