CS-220, Week 3A
Spring, 2006
R. Eckert

PREPARING AN 80X86 ASSEMBLY LANGUAGE PROGRAM

1. Enter the source assembly language program using an
editor like EDIT --> .ASM file

2. Translate the assembly language to machine language using an
assembler program like MASM --> .OBJ file (machine readable)

  You may want to request a human-readable version of the
  translation called a list file (.LST file)

  You may also want to request a cross reference listing to aid
  in debugging (.CRF file)

3. Link the object file to Operating System (and other) modules
using a linker program like LINK --> the executable, .EXE file

  You may also want to get a memory map showing where the various
  modules will be placed when the executable is loaded into
  memory (the .MAP file)

  You may also want to supply other .OBJ files from a library
  (.LIB file)

The MASM version 6.11 package allows you to combine steps 2 and 3
with the ml program which does both the assembling and linking.

4. Execute the program (.EXE file) by typing in its name at the
DOS prompt.

For details of how these steps are carried out, see the Lab 1
Handout from the "Lab 1 Assignment" link on the class Web pages.


OVERVIEW OF THE 80X86 ARCHITECTURE (REGISTERS, ACCESSING MEMORY)
Also see Chapter 2 of your Irvine textbook.

The following is a block diagram of internal organization of the 
Intel 8088 microprocessor. The 8088 is somewhat simpler than the 
8086. The main differences are the fact that the latter has a 16 
bit internal data and a six-byte-deep instruction prefetch queue.

The Internal Organization of the Intel 8088 Microprocessor


The following diagram shows the Intel 8086 16-bit registers that are
accessible to the assembly language programmer. Also see Figure
2-9 in your text book to see the Intel 80X86 32-bit registers.

ACCESSING MEMORY WITH AN INTEL 8086/8088-BASED COMPUTER -- REAL ADDRESS MODE
Memory is accessed by providing a 20-bit "physical address." Physical addresses
are generated by the bus interface unit part of the processor. Since the
processor's pointer registers (IP, SI, DI, BX, SP, BP) are only 16 bits wide
a method of storing a 20-bit physical address must be found. Intel came up with
the idea of memory "segments." A segment is a 64K byte block of memory, any
byte of which can be accessed with a 16-bit pointer register (2^16 = 64K). In
an 8086 or 8088-based computer, there can be up to four active memory segments
at any time, and a segment can be located anywhere in the 1 Mbyte memory 
space (2^20 = 1M). A physical address of any memory location is then given by
the following:
Physical address = address of base of segment + offset of location within segment
We say that a segment is "pointed to" by a segment register. In other words a    
segment register will always hold the information needed to get the base address
of the segment it points to. What is actually stored in a segment register is
the base address of the segment divided by 16. In other words, segments always
begin on "paragraph boundaries" -- addresses divisible by 16 (or that end in a
hexadecimal 0). Since this is the case, the least significant hex digit of the
base address does not need to be stored. So the segment register will always
contain the base address divided by 16 -- a 16-bit number. Any time the
processor accesses memory, it computes the physical address by taking the
contents of the appropriate segment register, left shifting it by 4 bits (i.e.,
multiplying it by 16) to get the segment's base address, and then adds the
appropriate offset:
Physical address = (contents of segment register)*16 + offset
Depending on the kind of memory access, different segment registers are used
to hold the segment base and other registers are used to hold the offset. The
following table shows the most common ways of accessing memory in an 80x80-based
system:
Type of memory access   Segment Register  Alternate Seg. Reg.  Offset
-----------------------------------------------------------------------
Instruction fetch          CS                none              IP
Stack access               SS                none              SP or BP
Data access                DS                CS, SS, ES        DI, SI, or BX
Symbolic data access       DS                CS, SS, ES        generated by assembler
See Section 2.3 of your textbook for a brief description of Real and Protected
mode memory addressing for the Intel 80X86 family of microprocessors.