🚀 go-pugleaf

Environment : MS-DOS 4.01 , 286 NEAT-Motherboard at 20 MHz with 1 WS ,

              BIOS from CHips & Technologies

              2 MB RAM (80 ns , 2 Banks of 1 MBit Chips )

              a Seagate ST296N SCSI Drive and

              a Seagate ST02 SCSI adaptor-card

              an EVA/1024 8-bit Vga-card with 512K , Tseng Chipset
                 ( max 1024 * 768 * 16)
              a noname Multisync-Monitor

My Problem sounds trivial,unbelievable but it's serious somehow.

I tried to simulate a virus in memory to force a warning when

scanning memory with McAffee's Scan V71c option /m for scanning

memory for all known viruses.

to do so , i wrote a very short assembler program , which does nothing than

1 ) get the old msdos interrupt ( Int 21h )

2 ) save this interrupt in it's code

3)  set the msdos interrupt to its own code

4)  if function 4b is called with 'FFh' in AL-register
    the answer should be '55aa' in the bx-register


sounds quite simple .

BUT when i try to get the int 21h pointer , I always receive junk from the

interrupt call ( AX = 2521h or 35 one is get the other set , then int 21h).

its the same rusult whether i watch with a debugger or not ( i tried Quaid's

Analyzer, Turbo Debugger V 1.0, MS Codeview 5.1 ) .

i even tried to read   the interupt vector table manually ,  but that also

failed.

When i tried something similar on my older system ( ST225 mfm instead of SCSi

and 640 KB , Running at 10 Mhz 0 WS  ) it worked well. NAturally all tools

which manipulate the interrupt-vectors work fine, just my self-written program

doesn't.

is there anybody out there with a solution , it's driving me crazy .

                              martin  zejma from vienna, austria



please e-mail to me , 'cause i'm no regular reader of this newsgroup