---

Intel 80386 addressing modes

---

• Real and Protected modes
  • The Intel 80386 can operate in 2 different modes:

    Real mode: No memory address translation          Protected mode: Uses memory address translation or virtual memory technique (explained in cs355)

  • Nowadays, programs are only written using the "Protected mode"

    The real mode in 80386 is provided only for backward compatibility

    ---

    The Protected Mode does not use segment registers

    I will omit this weird form of addressing mode.

  ---

  ---

• Instruction format
  • The instruction format is similar but opposite to M68000:

    Opcode DESTINATION, SOURCE

    Notes:

    The first operand is the destination The destination operand is also the first source operand in a binary operation The second operand is the second source operand

  ---

  ---

• Size of operands in an Intel assembler instruction
  • Specifying the size of an operand in Intel

    The size of the operand (byte, word, double word) is conveyed by the operand itself Example: EAX means: a 32 bit operand AX means: a 16 bit operand AL means: a 8 bit operand

  • Very important:

    The size of the source operand and the destination operand must be equal (Otherwise, you will get an error message)

  ---

  ---

  ---

  ---

• Register Operands
  • The name of a register is used to identify:

    The register used as an operand in an operation The size of the operand !!!

    ---

  • How the name is used to identify a register as operand:

    <----------- EAX ------------> +-------+-------+-------+-------+ | | AH | AL | +-------+-------+-------+-------+ <------ AX -----> EAX - 32 bits in register EAX AX - lower 16 bits in register EAX (Other parts of the EAX register is unchanged) AH - second last 8 bits in register EAX (Other parts of the EAX register is unchanged) AL - last 8 bits in register EAX (Other parts of the EAX register is unchanged) EBX - 32 bits in register EBX BX - lower 16 bits in register EBX BH - second last 8 bits in register EBX BL - last 8 bits in register EBX and so on...

    ---

  • Example instructions:

    mov EBX, EAX - copy 32 bits in register EAX to EBX mov BX, AX - copy 16 bits in register EAX to EBX mov BL, AL - copy 8 bits in register EAX to EBX

    ---

    ---

  • Examples with errors:

    mov EBX, AX - src operand = 16 bit register AX destinaton operand = 32 bit register EBX mov BX, AL - src operand = 8 bit register AL destinaton operand = 16 bit register BX

  ---

  ---

  ---

• Immediate Operands
  • Immediate (contant) operands:

    A numeric constant or a symbolic (numeric) constant is identified as an immediate operand

  • Examples:

    mov eax, 10 ;; Effect: eax = 00000000 00000000 00000000 00001010 MAX: equ 7 mov eax, MAX ;; Effect: eax = 00000000 00000000 00000000 00000111

  ---

  ---

  ---

• Memory Operand
  • Syntax of access memory variable:

    VarName: DD 0 ;; Integer variable named "VarName" mov EAX, VarName ;; Copy value store in "VarName" to register EAX mov EAX, offset VarName ;; Store the memory address of "VarName" to register EAX

    ---

    Example: read the value from variable "sum"

    int sum; sum = 1234; .data sum dd ? ;; Allocate memory space for variable "sum" .code mov sum, 1234 ;; sum = 1234

    ---

  • Example: get the memory address of variable "sum"

    int sum; eax = address of the variable sum .data sum dd ? ;; Defines sum .code mov eax, offset sum ;; eax = address(sum)

  ---

  ---

• The Hello World program in Intel Assembler
  • Hello World:

    ; ------------------------------------------- ; Code segment ; ------------------------------------------- .code start: ; invoke functionStdOut with parameter: addr HelloWorld mov eax, offset HelloWorld ;; eax = address "HelloWorld" push eax ;; Pass parameter to "StdOut" call StdOut ;; Call "StdOut" add esp, 4 ;; Pop parameter ; Exit.... invoke ExitProcess, 0 ;; Return to DOS ; ------------------------------------------- ; Data segment ; ------------------------------------------- .data HelloWorld db "Hello World !", 0

  ---

  ---

• Using the "Printf" library function in MASM
  • Printf:

    The printf function in MASM is structured like the printf() in C/C++ The first parameter of printf is a format string The format string can contain conversion symbols Commonly used conversion symbols: %d = print an integer value %f = print an floating point value The parameters 2, 3, ... are values that are used with the respective conversion symbols found in the format string (first parameter)

    ---

  • Example: invoke printf("Value = %d", 1234)

    ; ------------------------------------------- ; Code segment ; ------------------------------------------- .code start: mov eax, 1234 ;; eax = 1234 (binary) push eax ;; Pass 1234 mov eax, offset Str1 ;; eax = address of Str1 push eax ;; Pass parameter call crt_printf ;; Call "printf" add esp, 8 ;; Pop parameter ; Exit.... invoke ExitProcess, 0 ; ------------------------------------------- ; Data segment ; ------------------------------------------- .data Str1 db "Value = %d", 0

    ---

    ---

  • Example Program: (Demo above code)                                                
    • Prog file: click here

  ---

  ---

• Indirect and Indexed Addressing Mode
  • 80386 can use a register to access operands in memory:

    mov eax, offset VarName mov ebx, [eax] ;; Copy value stored at ;; memory location given by EAX ;; into register EBX

    ---

  • You can add various offset expressions to the base:

    mov ..., [eax] ;; Memory address given by EAX mov ..., [eax + 4] ;; Memory address (EAX + 4) mov ..., [eax + ebx] ;; Memory address (EAX + EBX) mov ..., [eax + 4*ebx] ;; Memory address (EAX + 4*EBX) mov ..., [eax + 4*ebx + 4] ;; Memory address (EAX + 4*EBX + 4)

    ---

    ---

    Program Example:

    ; ------------------------------------------- ; Data segment ; ------------------------------------------- .data ArrayA dd 1111, 2222, 3333, 4444 i dd 1 Str1 db "Result = %d", 0 ; ------------------------------------------- ; Program ; ------------------------------------------- .code start: ;; Access ArrayA[0] mov ebx, offset ArrayA ;; ebx = address(ArrayA) mov ebx, [ebx] ;; ebx = ArrayA[0] = 1111 push ebx push offset Str1 call crt_printf add esp, 8 ;; Pop parameter ;; Access ArrayA[1] mov ebx, offset ArrayA mov ebx, [ebx+4] ;; ebx = ArrayA[1] = 2222 push ebx push offset Str1 call crt_printf add esp, 8 ;; Pop parameter ;; Access ArrayA[i] mov ebx, offset ArrayA mov eax, i mov ebx, [ebx + 4*eax] ;; ebx = ArrayA[ i ] = 2222 push ebx push offset Str1 call crt_printf add esp, 8 ;; Pop parameter ;; Access ArrayA[i+1] mov ebx, offset ArrayA mov eax, i mov ebx, [ebx + 4*eax + 4] ;; ebx = ArrayA[ i+1 ] = 3333 push ebx push offset Str1 call crt_printf add esp, 8 ;; Pop parameter ; Exit.... invoke ExitProcess, 0 end start

    ---

  • Example Program: (Demo above code)                                                
    • Prog file: click here

  ---

  ---

  ---

  ---

• Converting byte -> short -> int -> long -> long long
  • Note:

    The conversion instruction will only work with the register EAX (accumulator)

    ---

  • Conversion instructions in 80386:

    Byte(8 bits) --> Word(16 bits) cbw - convert byte to word - Effect: sign-extends AL to AX Word(16 bits) --> Double Word(32 bits) cwd - convert word to double word - Effect: sign extends AX to DX:AX - Note: Obsolete instruction ! - This instruction is provide to - conserve portability with 8086 cwde - convert word to double word extended - Effect: sign-extends AX to EAX Double word(32 bits) --> Quad Word(64 bits) cdq - convert double word to quad - Effect: sign-extends EAX to EDX:EAX

  ---

  ---

• More Information
  • MASM Programmer's Guide - Chapter 3 Using Addresses and Pointers click here

    A note of warning: uses outdated addressing mode with segment registers (I have avoided this)