VLA's Assembly Tutorial #1
Hello there, this is Draeden typing this wonderful document. This is
an explanation of the basic assembler frame. This document assumes that you
know what hexdecimal is and somewhat how it works, that you have a copy of
TASM and TLINK, that you know what AX is, and how it relates to AL and AH,
and you know the commands:
MOV xx,xx
JMP xxxx
and INT xx
I'm also making the rash assumption that you want to learn ASSEMBLER. :)
To assemble ASM0.ASM into an executable do the following:
TASM ASM0
TLINK ASM0
Now you can exececute this wonderful program. Go ahead. Try it. In
case you are having problems figuring out how to execute this, just type:
ASM0 (followed by the enter key)
No, you did nothing wrong. This code (ASM0.ASM) does nothing. All it
does is return control to DOS. It is the basic frame for an assembler
program. All of the programs that I write use this frame. If you want to
know what each part does, read on. If you already know, just go read
ASM1.TXT.
The number followed by the colon means that this is from ASM0.ASM and
tells which line it is from.
1: DOSSEG
DOSSEG Sorts the segment using DOS standard, which is:
1) 'code' segments (in alphabetical order)
2) 'data' segments (in alphabetical order)
3) 'stack' segments (again, in alphabetical order)
Although it may not seem clear what this does, don't worry about it. Just
have it as the first line in your assembler programs, until you understand it.
2: .MODEL SMALL
MODEL ONLY needs to be used if you use the simplified segments, which I
strongly recommend.
In a nutshell, .MODEL Selects the MODEL to use. This is used so that this
code can be linked with C, PASCAL, ADA, BASIC, other ASSEMBLER program, and
other languages with ease. It also tells the compiler how to treat your
code and data segments.
NEAR means that the data/code can be reached using a 16bit pointer (offset)
FAR means that a SEGMENT:OFFSET pair must be used to access all the data/code
Possible MODELS are:
TINY: Code and Data must fit in same 64k segment.
Both Code and Data are NEAR.
SMALL: Code & Data have seperate segment, but must be each less than 64k
Both Code and Data are NEAR.
For most applications, this will suffice.
MEDIUM: Code may be larger than 64k, but Data has to be less than 64k
Code is FAR, Data is NEAR.
COMPACT: Code is less than 64k, but Data may be greater than 64k
Code is NEAR, Data is FAR.
LARGE: Both Code & Data can be greather than 64k. Both are FAR, but a
single array cannot be greater than 64k. Note that max array size
means nothing if you are just writing in assembler. This only
matters when you link to C or another high level language.
HUGE: Same as LARGE, but arrays can be greater than 64k.
What that means is that the array index is a far pointer, instead
of a NEAR one.
LARGE and HUGE are identicle to the assembler programmer.
3: .STACK 200h
Tells the compiler to set up a 200h byte stack upon execution of the
program. NOTE: the size you choose for the stack does not change the size
of the file on disk. You can see what I mean by changing the 200h to, say,
400h and then recompiling. The file sizes are identicle.
This could be replaced with:
: MyStack SEGMENT PARA PUBLIC STACK 'STACK'
: db 200h dup (0)
: MyStack ENDS
BUT, doing it this way makes your executable 512 bytes bigger. If you
were to double to 400h, the executable would be another 512 bytes bigger.
I think it's pretty obvious why the simplified version is preferred.
4: .DATA
Simplified, unnamed 'data' segment. This is where those simplified
segments become very handy. If you were to write out the segment declaration
the regular way, you'd have to write something like this:
: MyData SEGMENT PARA PUBLIC 'DATA'
:
: ... ;your data goes here...
:
: MyData ENDS
Where 'MyData' is the name of the segment, public means that its, well,
public, and PARA is the alignment of the start of the segment. 'DATA'
specifies the type of the segment. Instead of PARA, WORD or BYTE could
have been used. (PARA = segment will start on an adress that is a multiple
of 16, WORD = even addresses, BYTE = where ever it lands.)
5: .CODE
Pretty much the same story as above, but this is for the code segment.
Could be replaced with:
- IN MASM MODE -
: MyCode SEGMENT PARA PUBLIC 'CODE'
: ...
: MyCode ENDS
- IN IDEAL MODE -
: SEGMENT MyCode PARA PUBLIC 'CODE'
: ...
: ENDS MyCode ;the 'MyCode' is optional in IDEAL mode
6: START:
This is just a label. Labels just provide a way of refencing memory
easily. Like I could say "JMP START" which would jump to the label START and
execute the code immediatly after it. Or I could say MOV AX,[Start], which
would grab the WORD that was immediatly after the label START.
7: mov ax,4c00h
8: int 21h
This bit of code calls DOS function # 4ch, which returns control to DOS
and sends back the error level code that is in AL (which is zero).
Note that for all int 21h DOS functions, AH contains the function number.
THIS MUST BE AT THE END OF THE CODE! If it isn't, the code will continue to
run... right out of the end of your program and will execute whatever code
is there! The program will crash with out it!
9: END START
This tells the compiler that we are all done with our program and that it can
stop compiling, now. And it tells the compiler to put the entry point at
the label START. This means that DOS is effectivly starting your program by
executing this:
: JMP START
As you would probably guess, if you just put `END' instead of `END START'
and you compiled and linked the program, when you went to execute the code,
the computer will probably freeze because it does not know where to start
execution.
Ok, now that you know what the frame is/does, lets actually make the
program do something. Lets be wild and crazy, and PRINT A MESSAGE!
CONTINUED IN ASM1.TXT
ASM0.ASM
DOSSEG
.MODEL SMALL
.STACK 200h
.DATA
.CODE
START:
;
; Your code goes here...
;
mov ax,4c00h
int 21h
END START
; THIS CODE DOES ABSOLUTLY NOTHING EXCEPT RETURN CONTROL TO DOS!