Introduction

The HC12 CPU is used to teach the CEG3136 class at the University of Ottawa. This document is a living personal reference to the CPU12 CISC instruction set. WIP!

Let’s get started with a practical example. Calculate:

z = a + b - c

A function written in C:

void addstuff() {
  int z;
  int a=5, b=6, c=8;
  z = a + b - c;
}

Now in HC12 ASM, using sections for organization:

consts: SECTION
var_a DC.W 5
var_b DC.W 6
var_c DC.W 8

data:   SECTION
z DS.W 1

code:   SECTION
addstuff:
  LDAA var_a
  ADDA var_b
  SUBA var_c
  STAA z
  RTS

Reading the Docs

When looking at documentation, you’ll need to know how to read the operands for any given function. Learning this will enable you to feed any given function the data it needs to run, and prevent you and the machine from diverging in your thinking!

Most manuals will use the following notation for operands:

oprx0_xysp functions can be given:

• oprx3, xys with xys pre or post incremented or decremented, like +X or Y-.
• oprx5, xysp which is indexed offset.
• abd, xysp, the X, Y, or Z registers offset by A, B, or D.

Great, now you can read the docs like a master! NXP has quite a few publicly available documents with the instruction set thoroughly detailed, like this one .

Using the Stack

The stack is an area of memory that we can place temporary variables in. It is good for functions or subroutines that need to manipulate more data than can be stored in the registers. First, we call the OFFSET 0 command to define a set of offset labels. Essentially, these labels allow data structures and variables to be reserved and accessed in the stack. See the Program Organization section for more information on OFFSET.

  OFFSET 0
space_to_store_x   DS.W 1
space_for_one_word DS.W 1
a_ten_byte_space   DS.B 10

store_something:
  PSHX
  LDD #$1234
  STAB space_for_one_word,SP

As you can see, variable_name, SP is used to reference areas of the stack memory. Here is a more complex example, where a subroutine preserves the registers before completing a set of operations:

; create_array.asm
; =========================================
;  Creates an array of one-byte decimals like so in the memory:
;   [ 3, 4, 5, 6, 7, 8 ]
;  at the address given in register Y.

  OFFSET 0
DS.W 1      ; Space to RTS address
DS.B 1      ; Space to store A
DS.B 1      ; Space to store B

create_array:
  PSHA  ; Save A to the stack
  PSHB  ; Save B to the stack

  ; Load 3 into the A register.
  LDAA #$3
  LDAB #$5

; Description below.
loop:
  STAA 0,Y
  INY
  INCA
  DECB
  BNE loop

  PULB  ; Restore B to previous status
  PULA  ; Restore A to previous status
  RTS

Let’s look carefully at the instructions in the loop.

1. STAA 0,X stores the contents of the register A into the memory address referenced in the register X. For the first run of the loop, the value 3 is loaded into the memory address [X].
2. Now that we are done with the X register for now, we can point it to the next address in memory with the INX command.
3. To ensure we’ll store 4 in the next address, we INCA. The A register is incremented and now contains the number 4.
4. DECB is called and B becomes 4. It will continue to be decremented until it becomes zero, and the BNE command refrains from jumping back to the loop label.
5. BNE is branch if not equal to zero, checking the CCR (Condition Code Registers) to see if the last operation set the Z (zero) bit. If the zero bit is not set, BNE will move the PC back to the argument label, in this case, loop.

Notably, we use PSHA and PSHB at the beginning of the subroutine to save the current contents of the A and B registers to the stack. At the end of the program, we use PULB and PULA to move the original contents of the registers back into place.

We do this in order to reduce the impact of the subroutine on the running machine code. Subroutines are responsible for cleaning up after themselves. If a subroutine messed up the registers in the middle of a long operation which made a subroutine call, that long operation would be very unhappy!

Addressing Modes

In short, the modes available on the HC12 CPU are:

1. Immediate
2. Direct
3. Indexed
4. Relative

Each of these can be expanded to include their operating modes (N is substituted here for instructions that can be all of the bit-modes indicated in the table above, and op for operation):

1. Immediate: op #oprNi
2. Direct: op opr8a

• Extended: op opr16a

1. Indexed:

• With N-Bit offset: op oprxN,xysp
• With 3-Bit offset, pre-decrement: op oprxN,-xysp
• With 3-Bit offset, pre-increment: op oprxN,+xysp
• With 3-Bit offset, post-decrement: op oprxN,xysp-
• With 3-Bit offset, post-increment: op oprxN,xysp+
• With accumulator offset: op abd,xysp
• Indexed-Indirect, 16-bit offset: [oprx16, xysp]
• Indexed-Indirect, D offset: [D, xysp]

1. Relative: op PCr, oprxN

These can be broken into three loose operating modes, with modifiers:

1. Directly providing the value, using the # character, directly gives the CPU the value in the instruction. There is no RAM access needed to proceed with the operation.
2. Using an offset, supplying two values without an addition or subtraction symbol (like,this), adds the two numbers together and allows the CPU to access things like the stack.
3. Incrementing or decrementing, supplying a + or - before or after the xysp allows the xysp register to be increased or decreased before or after the contents of the address are passed back to wherever the operation dictates.

Some examples might be nice:

; Inherent instructions have NO addressing mode.
inherent:
  NOP
  LSL

immediate:
  ; Note the #!
  LDAA #$64
  LDX #$1234
  BEQ inherent

direct:
  LDS

indexed_offset:



direct:
  LDS #AFE ; Init stack pointer.

Program Organization

A number of labels and directives can be used to sort the pieces of a program, ensuring they are placed in (or forced into,) the correct place in memory. Here are the commands to know for the midterm.

ORG is the origin command. It places assembly code at the specified location in memory. Generates an internal, absolute code section in memory.

; Next instruction will be placed at address 3000
  ORG $3000

OFFSET declares an offset section which is useful for stack frames or simulating data structures. Essentially, this allows you to pre-allocate blocks of memory around the stack pointer for use for a particular purpose. The command we use the most, OFFSET 0, tells the compiler to, starting at the stack pointer, preserve this much space for this set of variables. Order does matter: If you are going to push or pull anything to the stack, you need to reserve space for those at the ‘bottom’, and define those first.

SECTION declares a relocatable section. These can include data, constant data, or code. Define separate sections so these are placed in the correct memory location. Variables defined with DS (define storage,) constants defined with DC, and code all belong in different places.

consts: SECTION
const1: DC.B $8B
const2: DC.B $A4

data:   SECTION
var1:   DS.W 1
var2:   DS.W 1

code:   SECTION
loop:
  LDAA const1
  ADDA const2
  STAA var1
  BRA loop

SWITCH blocks assemble code according to case statements. After assembly, the switch block no longer exists; only one of the cases is preserved in the final .s19 file.

var EQU 5

; Written directive:
SWITCH var
CASE 0
  LDD #23
CASE 2
  LDD #89
CASE 5
  LDD #A8
DEFAULT
  LDD #19
ENDSW

; Generates:
  LDD #A8

Formatting Tests

The below are to ensure correct syntax highlighting.

A bit of assembly:

  OFFSET 0
  DS.W 1
  MYINT DS.W 1

loopTo3k:
  PSHX
  LDX MYINT,SP

Here’s a short C snippet I wrote while playing with strings.

#include <stdio.h>

void main(void) {

  short print(char *str) {

    while (*str != '\0') {
      printf("%x -> %c\n", str, *str);
      str++;
    }

  }

  // xyz is a pointer to the first character.
  char xyz[6] = {'h', 'e', 'l', 'l', 'o', '\0'};

  print(xyz);
}

…will return the following (address will vary):

83f76a4a -> h
83f76a4b -> e
83f76a4c -> l
83f76a4d -> l
83f76a4e -> o