Question

• The Hack assembler is a two-pass assembler. Explain why there is a need for going twice over the code.
• There is a method named backpatching that solves this issue with one pass only. Describe this method.
• What are the pros and cons for one pass assembly vs. two pass assembly?

Answer 1

Note: I will only provide the answer to the first question as I need to follow the Chegg guidelines.

Answer to the first question:
We know that the assembly program may contain jump instruction in the code. For eg;

JMP LABEL
.
.
.
.  
LABEL: ......

It can be clearly seen that there is a gap between JMP instruction and the LABEL. The assembler when scans the above code (line by line), then it obviously doesn't know where LABEL is when it first sees the JMP instruction. When assembler reaches JMP, it only knows that there is a jump instruction to be converted to machine-code but it doesn't know how far is the LABEL.
An easy solution would be to just guess the position of LABEL but that would be the absolutely wrong approach.

That's why there is a 2-pass assembler i.e. there is a need for going twice over the code: In the first pass, the assembler will find the symbols and create a symbol table. In symbol table, all the symbols and their corresponding addresses are stored. In the second pass, the assembler will generate the machine code, and use the symbol table to determine how far away labels are.