Hello Alex,
One thing to consider is the amount of stack space you have while
servicing the interrupt requests. You should assume that you have
at most 100 bytes of stack available (most of the time you will have
more than this, but under certain circumstances you won’t).
As you suggest, you may also have problems with overrunning
the segment e.g. if the request is for more than 64K or if the
the buffer supplied starts at an offset near the end of a segment.
Shaun
Tuesday, March 1, 2005, 7:05:17 AM, you wrote:
AD> As for a full disk encryption the sytem is divided into 2 parts the (
AD> Real mode Preboot system , and the later drivers ) . Im right now
AD> stuck with the real mode pre boot driver the problem is that to test
AD> it im first checking it out on an AES ( 256 bit key and 128 bit block
AD> ) encrypted disk . But for some strage reason my code does not work
AD> fully . On version < DOS 7 which do not use int 13 extensions it works
AD> fine ( also when i implemented a XOR ( not AES ) based encryption it
AD> worked just fine i even loaded win 98 though it ran in compatibility
AD> mode ) , Also in a simple boot up mode ( after pressing F8 ) when safe
AD> mode is selected in command prompt it works fine ( currently testing
AD> E-MBR for dos only ) . I belive i have problem with how im using the
AD> segments in real mode .
AD> I have hooked int 13 extensions but they do not seem to work properly
AD> when AES comes in. Im using the windows startup floppy to create the
AD> test envoirnment ( so to get the preboot working first ) . If anyone
AD> has any idea what could be wrong or any specific procedure which is to
AD> be followed while loading this pre boot system ? .
AD> Thanking you