Index
> OS Construction > I am finishing the last portion of my OS |
| Author |
|
|
nkeck72 14 Jun 2015, 21:59
I want to post this OS code up here to show the forum what my work has created. Feel free to try this code at home and give suggestions if any issues come up. I personally have gotten this to work under bochs (DOSemu just quits spontaneusly when I set up the IVT for int 21h... sigh). If you cannot get it to work, let me know!
Source (made with emacs): Code: use16 mov ax, 9ch mov ss, ax mov sp, 4096d mov ax, 7c0h mov ds, ax mov ah, 02h mov dx, 0000h int 10h ;---------------------------------------- loadup: mov ah, 00h mov dl, 00h int 13h mov ah, 01h mov dl, 00h int 13h cmp al, 00h jne stop ;; Load the int 21h code (For later implementation) mov ah, 02h mov al, 01h mov ch, 00h mov dh, 00h mov cl, 01h mov dl, 00h mov bx, 1400h mov es, bx mov bx, 0000h int 13h mov ah, 01h mov dl, 00h int 13h cmp al, 00h jne stop ;; Set up the IVT to recognize my int 21h mov ax, 0000h mov es, ax mov al, 21h mov bl, 04h mul bl add ax, 02h mov bx, ax mov dx, 1400h mov [es:bx], dx mov al, 21h mov bl, 04h mul bl mov bx, ax mov dx, 0000h mov [es:bx], dx ;; Clear the screen mov ah, 02h mov dh, 00h mov dl, 00h int 10h mov ah, 07h mov al, 00h mov ch, 00h mov cl, 00h mov dh, 80d mov dl, 25d int 10h mov cx, 01h ;; Display 'NOS 1.0' message mov ah, 09h mov al, 4Eh int 10h mov ah, 02h inc dl int 10h mov ah, 09h mov al, 4Fh int 10h mov ah, 02h inc dl int 10h mov ah, 09h mov ah, 53h int 10h mov ah, 02h inc dl mov ah, 09h mov al, 20h int 10h mov ah, 02h inc dl int 10h mov ah, 09h mov al, 31h int 10h mov ah, 02h inc dl int 10h mov ah, 09h mov al, 2Eh int 10h mov ah, 02h inc dl mov ah, 09h mov al, 30h int 10h mov ah, 02h mov dl, 00h add dh, 01h int 10h ;; Begin setting up the environment in which the user will type typer: mov ah, 00h int 16h cmp al, 0Dh je print_enter cmp al, 08h je print_back mov ah, 09h int 10h push ax cmp dl, 80d je mcdn pop ax mov ah, 09h int 10h mov ah, 02h inc dl int 10h jmp typer print_enter: cmp dh, 25d je scr_dn mov ah, 02h add dh, 01h mov dl, 00h int 10h jmp typer print_back: mov ah, 02h dec dl int 10h mov ah, 09h mov al, 20h int 10h jmp typer scr_dn: mov ah, 07h mov al, 01h int 10h ret mcdn: cmp dh, 25d je scr_dn mov ah, 02h add dh, 01h mov dl, 00h int 10h ret stop: hlt ;---------------------------------------- times 510-($-$$) db 0 dw 0xAA55 Also have attatched the .asm file.
_________________ It may look hard, but it won't take long if you take it one byte at a time. NOS: www.github.com/nkeck720/nos |
|||||||||||
14 Jun 2015, 21:59 |
|
|
Trinitek 15 Jun 2015, 01:06
You should always disable interrupts before reprogramming the IVT. That might be what's upsetting DOSemu.
|
|||
|
15 Jun 2015, 01:06 |
|
|
Trinitek 15 Jun 2015, 01:50
I've been trying to get this thing to work in DOSbox, and you've got a lot more problems than you think you have.
Stop using DOSemu, first of all. Segmentation faults should never happen in real mode, unless you mean that DOSemu itself is segfaulting. In either case, it's a pile of smelly garbage. Second, look over your code very carefully. For example, you are supposed to check AH for the returned floppy drive status code, not AL, for your AH=0x01; INT 0x13 calls. |
|||
|
15 Jun 2015, 01:50 |
|
|
nkeck72 15 Jun 2015, 14:33
I always learned that AL is where INT 13h - 01h stores its result, I will check that out.
Also, am now switching over to bochs, DOSemu's segfault is actually a bug with the HLT instruction. Also, am adding appropriate CLI/STI instructons as suggested before and after modifying the IVT. Thanks for the feedback! |
|||
|
15 Jun 2015, 14:33 |
|
|
nkeck72 15 Jun 2015, 14:39
Code is now modified as follows:
Code: use16 org 7C00h mov ax, 9ch mov ss, ax mov sp, 4096d mov ax, 7c0h mov ds, ax mov ah, 02h mov dx, 0000h int 10h ;---------------------------------------- loadup: mov ah, 00h mov dl, 00h int 13h mov ah, 01h mov dl, 00h int 13h cmp al, 00h jne stop ;; Load the int 21h code (For later implementation) mov ah, 02h mov al, 01h mov ch, 00h mov dh, 00h mov cl, 01h mov dl, 00h mov bx, 1400h mov es, bx mov bx, 0000h int 13h mov ah, 01h mov dl, 00h int 13h cmp al, 00h jne stop ;; Set up the IVT to recognize my int 21h cli mov ax, 0000h mov es, ax mov al, 21h mov bl, 04h mul bl add ax, 02h mov bx, ax mov dx, 1400h mov [es:bx], dx mov al, 21h mov bl, 04h mul bl mov bx, ax mov dx, 0000h mov [es:bx], dx sti ;; Clear the screen mov ah, 02h mov dh, 00h mov dl, 00h int 10h mov ah, 07h mov al, 00h mov ch, 00h mov cl, 00h mov dh, 80d mov dl, 25d int 10h mov cx, 01h ;; Display 'NOS 1.0' message mov ah, 09h mov al, 4Eh int 10h mov ah, 02h inc dl int 10h mov ah, 09h mov al, 4Fh int 10h mov ah, 02h inc dl int 10h mov ah, 09h mov ah, 53h int 10h mov ah, 02h inc dl mov ah, 09h mov al, 20h int 10h mov ah, 02h inc dl int 10h mov ah, 09h mov al, 31h int 10h mov ah, 02h inc dl int 10h mov ah, 09h mov al, 2Eh int 10h mov ah, 02h inc dl mov ah, 09h mov al, 30h int 10h mov ah, 02h mov dl, 00h add dh, 01h int 10h ;; Begin setting up the environment in which the user will type typer: mov ah, 00h int 16h cmp al, 0Dh je print_enter cmp al, 08h je print_back mov ah, 09h int 10h push ax cmp dl, 80d je mcdn pop ax mov ah, 09h int 10h mov ah, 02h inc dl int 10h jmp typer print_enter: cmp dh, 25d je scr_dn mov ah, 02h add dh, 01h mov dl, 00h int 10h jmp typer print_back: mov ah, 02h dec dl int 10h mov ah, 09h mov al, 20h int 10h jmp typer scr_dn: mov ah, 07h mov al, 01h int 10h ret mcdn: cmp dh, 25d je scr_dn mov ah, 02h add dh, 01h mov dl, 00h int 10h ret stop: hlt ;---------------------------------------- times 510-($-$$) db 0 dw 0xAA55 And when I do so, bochs is complaining that there is an HLT instruction when IF=0. I see no issues here unless there is an issue loading the 512 bytes of INT 21h code (which does not yet exist and is actually filled with 00h on the disk)? |
|||
|
15 Jun 2015, 14:39 |
|
|
nkeck72 15 Jun 2015, 15:54
The new INT 21h has one function for now - to print a string. The code loaded for INT 21h is as follows:
Code:
use16
;; This is the int 21h code used in NOS.
;; Begin by checking the function and its parameters...
cmp ah, 00h
je print_string
;; Set the invalid function register (BL) to FF and return
mov bl, 0FFh
ret
print_string:
;; grab the char from RAM and process it
mov al, [es:bx]
cmp al, 24h
je line_done ; We want the terminating character to be $
cmp al, 00h ; If null then print a $
je print_ds
;; Go ahead and print the character since it is not a handled char
mov ah, 09h
int 10h
;; Move the cursor to the right
cmp dl, 80d
je line_dn ; If the end of a line is reached, we want to scroll down one line
inc dl ; If not, then move the cursor to the right
mov ah, 02h
int 10h
;; Increment the RAM address pointer and continue
inc bx
jmp print_string
print_ds:
;; Since the special char to terminate the line is a $, we want to print an actual $
;; if the stored char is null (00h)
mov ah, 09h
mov al, 24h
int 10h
cmp dl, 80d
je line_dn
mov ah, 02h
inc dl
jmp print_string
line_dn:
;; Scroll down a line and move the cursor to the beginning of said line
mov ah, 07h
mov al, 01h
int 10h
mov ah, 02h
mov dl, 00h
inc dh
int 10h
jmp print_string
line_done:
ret
;--------------------------------
times 512-($-$$) db 0
Am putting up the IMG file for the disk if you want it, other wise the .asm file for INT 21h is attatched.
_________________ It may look hard, but it won't take long if you take it one byte at a time. NOS: www.github.com/nkeck720/nos |
|||||||||||
|
15 Jun 2015, 15:54 |
|
|
nkeck72 15 Jun 2015, 16:04
The board will not let me put up the image, and so you will have to do the following (I use Linux, and so this is what i do to compile):
$ fasm myos.asm myos.bin $ fasm int21.asm int21.bin $ cat myos.bin int21.bin > NOS.img $rm *.bin Then, to write to a floppy disk: $ dd if=NOS.img of=/dev/fd0 _________________ It may look hard, but it won't take long if you take it one byte at a time. NOS: www.github.com/nkeck720/nos |
|||
|
15 Jun 2015, 16:04 |
|
|
Trinitek 16 Jun 2015, 02:58
I read through myos.asm very carefully this time, and I added a lot of comments. I haven't read through int21.asm yet.
https://gist.github.com/Trinitek/7d305d66b15ba90d427e |
|||
|
16 Jun 2015, 02:58 |
|
|
nkeck72 16 Jun 2015, 03:01
Thank you. Help is appreciated!
|
|||
|
16 Jun 2015, 03:01 |
|
|
nkeck72 16 Jun 2015, 03:09
I am going to modify myos.asm to include my INT 21h functions soon. I had never seen the 0x0E INT 10h function before, always learned that 0x09 puts a character and you had to move the cursor with 0x02. I wish I had better people teach me this stuff...
|
|||
|
16 Jun 2015, 03:09 |
|
|
Trinitek 16 Jun 2015, 03:10
Feast your eyes on all of the interrupts you would ever want and more: http://ctyme.com/rbrown.htm
|
|||
|
16 Jun 2015, 03:10 |
|
|
nkeck72 17 Jun 2015, 02:13
I have never seen Ralf Brown's list on the interweb, thanks! This will definitely come in handy.
|
|||
|
17 Jun 2015, 02:13 |
|
|
nkeck72 08 Jul 2015, 00:33
Have started a Git repository for this OS on SourceForge. Latest code will remain here, updated whenever possible. Help is still accepted. Thank you Trinitek for pointing out the bug in DosBox, I have made an image.asm file to circumvent this.
Git Repo is at: https://sourceforge.net/p/nosanopensourcefasmos/code/ci/master/tree/ _________________ It may look hard, but it won't take long if you take it one byte at a time. NOS: www.github.com/nkeck720/nos |
|||
|
08 Jul 2015, 00:33 |
|
|
PeExecutable 17 Jul 2015, 19:34
The problem with operating systems, especially OS's like Microsoft Windows is that it takes one man about five thousand years just to maintain it. What Microsoft's developers achieve in one year is what one man can achieve in five thousand years. If you begin working on an operating system with the extent and diversity similar to that of windows, you wouldn't even be finished with it if you began writing it in the time of the great pharaoh's of Egypt.
Windows, all together from it's start and until today have over 25 thousand years worth of work for a single man. What Microsoft is doing all together in just one year, is worth 60,000 years worth of work for one man. From January 2014 to January 2015, 60 thousand years worth of work was done by Microsoft. Even a simple game for windows, the amount of developers on that game is somewhere between 150 and 250 people. So even just a windows game takes 250 years to complete by a single man, and that is only if you're using very efficient tools. If you're using assembly, you can probably triple the amount of time. It takes about a thousand years for an assembly programmer to develope a hyper modern 3d shooter game with full content. If you're lucky, you'll finish a very simple program during the course of your life, but chances are you're dead before you finish that simple program. Last edited by PeExecutable on 17 Jul 2015, 19:52; edited 5 times in total |
|||
|
17 Jul 2015, 19:34 |
|
|
nkeck72 17 Jul 2015, 19:45
This topic is now closed. Thanks for the help! If I have any other questions, I will post them under their own threads.
_________________ It may look hard, but it won't take long if you take it one byte at a time. NOS: www.github.com/nkeck720/nos |
|||
|
17 Jul 2015, 19:45 |
|
< Last Thread | Next Thread > |
Forum Rules:
|