• Please review our updated Terms and Rules here

Booting DOS from an 8086 simulation with Proteus

N

NicolasF

Experienced Member
Joined
Jun 28, 2006
Messages
255
Location
Argentina
Hi,

I've been trying to make an 8086 system with Proteus that would boot DOS. The system has a timer 8253, interrupt controller 8259, PPI 8255, and 2 USARTs 8251 (one is used as a video and keyboard). Software is loaded into memory using 2 ports (I used one port as input and the other one as output to avoid logic contentions) from the 8251 from an SD card.
So far I got these results when trying to load DOS:

CPM86-1.0 - Shows corrupted init message then hangs
CPM86-1.10 - Show init message, then Proteus hangs with "Internal Exception access violation"
MS-DOS 1.0 - Hangs, never stops loading
MS-DOS 3.10 - displays "Bad or missing Command Interpreter", then hangs
MS-DOS 3.30 - Hangs, never stops loading
DR-DOS 3.0 - Hangs, never stops loading
FREEDOS - Hangs, never stops loading

I'm not sure what DOS does when it boots and trying to debug it while loading is not an option because you can only debug your code.
I'm thinking there is something wrong with my BIOS implementation, but if only I knew what DOS is doing when it boots it would really help...

I'll leave the files in my onedrive: https://1drv.ms/f/s!ACnXr6aLzVwmg28
 
The IBM ROM BIOS wont work in the simulation because the IO Ports I'm using are not the same as the ones used in the XT. For example: in the XT the PIC is at address 0x20 and I'm using 0x40, PIT in the XT is at 0x40 and I'm using 0x80, PPI in the XT is 0x60 and I'm using 0xC0. I had to change them because I needed all even addresses for the registers so that I can use the lower 8 bits of the data bus.
 
Hello,

This may not be detailed enough for your purposes, but the book:

Advanced MSDOS by Ray Duncan (Microsoft Press) pages 14 - 18 gives a fairly detailed step-by-step account of how MSDOS is loaded, incl a number of memory maps as the stages progress. Various components move about in memory as the process progresses.

Have you seen this. May be online somewhere.

Geoff
 
Is your design similar to any of the other 8086 systems? Cribbing a working BIOS and OS with some minor patching for the port addresses might provide a guide as to what needs to be corrected.
 
In cases where I can't use an actual debugger, the next tool I tend to reach for is "printf debugging". Tweak your BIOS implementation to display the interrupt number and the values of registers for each BIOS interrupt before going on to the meat of the routine. That will tell you what DOS is doing so that you can check your assumptions and see what was the last thing it did before it crashed.
 
What about this idea: load DOS without using DMA. In that way you can keep track what your computer is loading and where it is saving the data. If it loads the wrong data or saves it at the same place, then the BIOS can be blamed. Fix that and if things work, try DMA.
 
I think there is some issue with Proteus simulation of the 8086, it seems to have some kind of bug so I'm going to cancel this project. Thank you all for the feedback!
 
If you search the Internet, you will find the source code for MS-DOS 6.0.
It can be used to single-step through the code and check if BIOS calls and CPU instructions work as designed.
Tracing the code with the original assembly code next to it is tedious, but possible.
FreeDOS is written in C and less useful for this purpose.
 
You must log in or register to reply here.
Back
Top