CS 270 PA10: Disassembler

due Sunday, May 10th at 10:00pm, no late submissions.

Goals

For this assignment, you will write a program dis.c, which will implement a crude LC-3 disassembler.

Sample Run

% cat max.asm
;; File:	max.asm
;; Description: Compute the maximum value of two numbers
;; Author:	Jack Applin
;; Date:	April 26, 2015

	.orig	x3000
max	add	r6,r6,#-3	; Make space for return value, r7, and r5
	str	r7,r6,#1	; Save return address
	str	r5,r6,#0	; Save frame pointer address
	add	r5,r6,#0	; Setup frame pointer
	ldr	r1,r5,#3	; Get one argument
	ldr	r2,r5,#4	; Get the other argument

	;; Which is greater, R1 or R2?
	not	r0,r1		; R0=~R1; first part of negation
	add	r0,r0,#1	; R0=-R1
	add	r0,r2,r0	; R0=R2-R1
	brn	over		; R0<0 => R2-R1<0 => R2<R1 => R1 is bigger
	add	r1,r2,#0	; R1=R2
over	str	r1,r5,#2	; Save our result

	ldr	r5,r6,#0	; Restore frame pointer
	ldr	r7,r6,#1	; Restore return address
	add	r6,r6,#2	; Don’t need this stack space any more
	ret

;; Some invalid instructions
	.fill	xdead
	.fill	x100f

	.end
% lc3as max.asm
STARTING PASS 1
0 errors found in first pass.
STARTING PASS 2
0 errors found in second pass.
% c99 -Wall dis.c -o dis
% ./dis max.obj
	.orig	x3000
	add	r6,r6,#-3
	str	r7,r6,#1
	str	r5,r6,#0
	add	r5,r6,#0
	ldr	r1,r5,#3
	ldr	r2,r5,#4
	not	r0,r1
	add	r0,r0,#1
	add	r0,r2,r0
	brn	pc+1
	add	r1,r2,#0
	str	r1,r5,#2
	ldr	r5,r6,#0
	ldr	r7,r6,#1
	add	r6,r6,#2
	ret
	.fill	xdead
	.fill	x100f
	.end

Object file format

The *.obj file is a sequence of sixteen-bit words. The first word is the location of the code, from .orig, and all subsequent words are instructions or data. Here’s how to read a sixteen-bit word: 
#include <stdio.h>
#include <arpa/inet.h>

// Read a single 16-bit word from the given file descriptor.
// Return the word, or EOF if no more data is available.

int read_word(FILE *f) {
    unsigned short s;
    if (fread(&s, sizeof(s), 1, f) != 1)
	return EOF;
    return ntohs(s);	// Convert file byte order to desired byte order
}

// Read all the words from stdin, and display them.
int main() {
    int w;
    while ((w = read_word(stdin)) != EOF)
	printf("%04x\n", w);
    return 0;
}

Requirements

  1. Your output must match the sample output exactly for that input.
  2. All output is lower case.
  3. Invalid instructions are translated to .fill
  4. .orig and .fill emit numbers in hexadecimal.
  5. All other numbers are #decimal
  6. Label operands for br*, jsr, ld*, st*, etc., become pc+digits or pc-digits
  7. Each output line begins with a tab, '\t'.
  8. If an opcode has operands, a tab separates the opcode from the operands.
  9. trap instructions are translated as trap #number, not as out, halt, etc.
  10. Treat our local instructions push & pop as invalid, and hence .fill.
  11. The n/p/z in the br* family of branch instructions are in alphabetical order: Emit brnpz, not brpnz. Emit brnp, not brpn.
  12. BR with nzp all 0's (i.e., 0000..01FF) is undefined behavior.
  13. BR with nzp all 1's (i.e., 0E00..0FFF) must emit brnpz, not br.
  14. Instruction C1C0 is ret.
  15. Emit .end at the end.

How to submit your program

Submit dis.c to the Checkin tab.

How to receive negative points

Turn in someone else’s work.