Assignment

Write an assembler for the following instructions:

add, addi, lw, sw, beq, bgt, j.

Assume that: all registers are addressed by $register-name (like $s1, $t2, $a0, $sp; see your textbook for range of valid register names); the end of program is recognized by the .end instruction, the program is loaded into memory at location 0; each label has 1 to 4 characters; there is 1 or more space between the symbols in the program.

The input to your assembler is a text file consists of 1 to 50 assembly instructions. The last instruction is .end. You can write your assembler in any language as long as we are able to test it in the department.

The output of your program is a text file that contains the object code. The object code is represented in binary and also hex format.

Your assembler should consist of two passes.

First Pass,

During the first pass, the assembler generates a table that correlates all user-defined address symbols with their decimal equivalent value. The binary translation is done during the second pass. The content of PC (Program Counter) stores the value of the memory location assigned to the instruction or operand presently being processed. The assembler sets this counter to 0 initially. A line of symbolic code is analyzed to determine if it has a label (by the presence of a colon). If the line of code contains a label, it is stored in the address symbol table together with its decimal equivalent number specified by the content of PC. PC is then incremented by 4 and a new line of code is processed.

Second Pass,

Instructions are translated during the second pass by means of table-lookup or other procedures. There are three tables:

1. Pseudoinstruction table.
2. Instruction table.
3. Address symbol table.

The entries of the pseudoinstruction table are for pseudoinstructions such as bgt. Each entry refers the assembler to a subroutine that processes the pseudoinstruction when encountered in the program. The instruction table contains the symbols for the rest of instructions and their related information.

PC is initially set to 0. Lines of code are then analyzed one at a time. Labels are neglected during the second pass, so the assembler goes immediately to the instruction field and proceeds to check the first symbol encountered. It first checks the pseudoinstruction table. A match with an entry sends the assembler to the corresponding subroutine. If the symbol encountered is not a pseudoinstruction, the assembler refers to the instruction table. If a match occurs, the instruction is converted to its equivalent machine code by the use of address symbol table (if needed). The implementations of tables 1 and 2 are optional. You must use a hash function for the implementation of Table 3.

Error Diagnostics,

The assembler should check for possible errors in the symbolic program and if there is an error it should print out the proper error message.

Sample Program

main:
add $s1, $s2, $s3
loop: addi $s1,$zero, 4
sw $s3, 0($s2)
next: addi $a0, $zero, 0
lw $s4,0($s1)
j loop
addi $s4, $s1, 4
beq $s4, $s3, loop
bgt $s6, $s7, loop

.end