assembly newbie question

Valkyrie_of_valhalla · 11-05-2008, 05:10 PM

While self-learning assembly language (standard 80x86 syntax, using TASM as compiler) I can't seem to get what I'm doing wrong in an exercise taken from a book.

The exercise is like this: reverse the letters of the string like in the following examples:

qwer -> wqre
qwert -> wqret

Aka invert each 2 letters.

Here is my code:

Code:

   DOSSEG
   .MODEL SMALL
   .STACK 100h
   .DATA
MAXIMUM_STRING_LENGTH  EQU  1000
StringToReverse  DB  MAXIMUM_STRING_LENGTH DUP(?)
ReverseString    DB  MAXIMUM_STRING_LENGTH DUP(?)
i dw ?

   .CODE
   mov  ax,@data
   mov  ds,ax                       

; read string:
   mov  ah,3fh                      
   mov  bx,0                        
   mov  cx,MAXIMUM_STRING_LENGTH    
   mov  dx,OFFSET StringToReverse   
   int  21h   

;if nothing is read, end program                   
   and  ax,ax                       
   jz   Done                        
   mov  cx,ax                       
   push cx                          

   mov  bx,OFFSET StringToReverse
   mov  si,OFFSET ReverseString
 
;if only 1 character is read:
   cmp cx,1
   jne NoSingleChar
        mov al,[bx]
        mov [si],al
    jmp Print
    
    
   
NoSingleChar:
   
   mov i,1
   inc si
   
MainLoop:
    mov al,[bx]
    mov [si],al
    inc bx
    dec si
    mov al,[bx]
    mov [si],al
    inc bx
    add si, 3
    add i,1
    cmp cx,i
    je Print
    inc i
    cmp cx,i
    jne MainLoop
    dec si
    mov al,[bx]
    mov [si],al
  
Print:   
   
   pop  cx                          
   mov  ah,40h                      
   mov  bx,1                        
   mov  dx,OFFSET ReverseString     
   int  21h                         
Done:
   mov  ah,4ch                      
   int  21h                         
   END

It works for an odd number of letters, but not for an even number.

E.g. for qwert it shows tqre and I can't figure out why it does that.

What I also don't get... the AX register, which is supposed to contain the number of characters read after the 3fh dos function, according to the debugger, contains the number of characters +2... Aka for qwer it will hold 6, why is that?

Anyway, the main question: what am I doing wrong? Any suggestion or idea is welcomed.

bgeddy · 11-06-2008, 10:08 AM

While I find it odd posting Dos questions on the Linuxquestions forum I'll try to be helpful. A quick look at your code seems rather confused - this is after all a simple problem involving an increasing source and decreasing destination index - one initialised to the string beginning the other to the target start + string length-1. Try to map this process out into pseudo code and it should be easy to translate this to assembler.

I'm not familar with TASM but a powerfull tool is always a single stepping debugger - like Dos's DEGUG (or the more powerful SYMDEB - CODEVIEW came later from Micrsoft). TASM may have an equivalent allowing symbolic debugging. You can then step through your code examining registers and memory locations to easily spot mistakes.

As to the interrupt returning unexpected values - perhaps the string length includes a CR + LF on the end of the string - just a guess but debugging should make this obvious.

Good Luck with this.

Sergei Steshenko · 11-06-2008, 10:50 AM

Quote:

Originally Posted by bgeddy

While I find it odd posting Dos questions on the Linuxquestions forum I'll try to be helpful. A quick look at your code seems rather confused - this is after all a simple problem involving an increasing source and decreasing destination index - one initialised to the string beginning the other to the target start + string length-1. Try to map this process out into pseudo code and it should be easy to translate this to assembler.

I'm not familar with TASM but a powerfull tool is always a single stepping debugger - like Dos's DEGUG (or the more powerful SYMDEB - CODEVIEW came later from Micrsoft). TASM may have an equivalent allowing symbolic debugging. You can then step through your code examining registers and memory locations to easily spot mistakes.

As to the interrupt returning unexpected values - perhaps the string length includes a CR + LF on the end of the string - just a guess but debugging should make this obvious.

Good Luck with this.

This is what written at the page top (emphasis is mine):

Quote:

This forum is for all programming questions.
The question does not have to be directly related to Linux and any language is fair game.

bgeddy · 11-06-2008, 11:40 AM

Quote:

This is what written at the page top (emphasis is mine):

Heh - do you know I've never noticed that ! Anyway - I'm not wishing to be judgmental here just trying to help..

Hopefully this is how I came across.

To the OP - please ignore my ill informed comments about Dos questions in a Linux forum.

Valkyrie_of_valhalla · 11-06-2008, 12:48 PM

Thanks for your replies.

First of all,I got it fixed... it was related to the number of characters, the weird +2 value that cx had.

Yes, I am aware it is a Linux forum, and that it's the non-*nix section :P

You guys always helped me out, wether with programming or linux questions, so it didn't seem strange to me.

Plus, I'm learning the DOS-version of assembly only because you can find more resources about it, my final goal shall be, of course, some linux kernel device drivers and other fun stuff :P

How was that quote... "don't you miss the days when a man was a man and wrote his own device drivers?" or something like that

H_TeXMeX_H · 11-06-2008, 01:05 PM

Quote:

Originally Posted by Valkyrie_of_valhalla

How was that quote... "don't you miss the days when a man was a man and wrote his own device drivers?" or something like that

yeah, back when devices were connected via serial ports and used standard interfaces

jlinkels · 11-07-2008, 09:24 AM

Quote:

Originally Posted by H_TeXMeX_H

yeah, back when devices were connected via serial ports and used standard interfaces

This is what we still do in Linux... since 1970-ies we have /dev/ttyS0 and that has never changed since.

It was a (bad) habit introduced in DOS(*) that hardware was addressed immediately from any program, and yes, in some way you could call

Code:

mov ax,1F
out 378,ax

a standard way to talk to a serial device.

(*) Well maybe it was not really introduced in DOS, we had the AppleII and the Commodore etc before that, but DOS was the first platform were [commercial] programs became available in such vast quantities and still using those awkward techniques. Although the intention of the DOS architecture was to do it well-behaved thru a BIOS function.