Project #1: Implementing an RISC-V (RV32I) Assembler

This project is to implement an RISC-V (RV32I) ISA assembler, and the given code is a basic skeleton for input and output.

Introductions

Basically, the code written in assembler.c operates only creation of files for input and output. Each student is required to implement parsing, converting, or any functions based on the skeleton code. The detailed instructions are described in handout/Project_1.pdf.

There are two directories for this project. The code will read a .s file from sample_input and it will produce .o file in the same directory.

The files in sample_output are used when you test your output with them. The following is an example for the skeleton code:

# Usage
$ ./assembler sample_input/example1.s   # This will produce sample_input/example1.o

# Test
$ make test # This will test all examples in sample_output

Description for Skeleton Code

The skeleton code we provided is part of a two-pass assembler. Basically, the two-pass assembler reads the assembly code twice and converts it to binary code (machine code).

RISC-V Assembler processing example

As shown in the figure above, the skeleton code takes an assembly code as an input. Then, in the 1st pass, you need to create a symbol table that specifies a key-value pair to map labels to memory addresses. Here, you are supposed to write code for make_symbol_table(). While you are building the symbol table, you may want to extract the data and code segments separately to simplify the 2nd pass.

In the 2nd pass, the skeleton code just reads the text segment and translates each line of code into binary (machine code). Here, you are supposed to fill out make_binary_code() code. While translating, you may encounter the labels. Now you can translate the labels to the memory addresses by referencing the symbol table you built in the 1st pass.

Hints

All students are supposed to implement the functions below.

1. void make_symbol_table(FILE *input)
   This function makes a symbol table for converting assembly code to binary. This function should be able to parse the line and fill the symbol table. Both data and text lines are added to their own segment.

2. void make_binary_file(FILE *output)
   Make a binary file using the previously created symbol table. You need to print binary lines to the .o file(=FILE *output) mentioned above. You can use record_text_section() and record_data_section. These functions will print binary code converted from assembly code.

3. void record_text_section(FILE *output)
   In this function, you need to translate a text segment line by line to binary code. The binary codes should be printed into an output file.

4. void record_data_section(FILE *output)
   Likewise text section, this function translates the data segment to binary code. It converts their initialized values to binary code and prints them into an output file.

The code that needs to be implemented is written as /* blank */ with short comments. If you do not want to use the skeleton code, it is allowed to write code from scratch. Functions you need to implement are all located in assembler.c (You should only modify assembler.c file). Also, we provide some utility functions, so feel free to use them (See util.c to check these functions).

Debug

We have a debug option. It is located in assembler.h header file. If you want to enable the debug option, define macro DEBUG as 1. If you want to disable, define macro DEBUG as 0. We hope this option helps you understand the code flow.

Reference functions

Description of functions used in this project.

1. FILE* tmpfile(void)
   Create a temporary file and return the file stream pointer. This function opens the temporary file in wb+ mode.

2. void rewind(FILE *fp)
   Point back to the start of the file stream.

Q & A

If you find any errors in this code or have a question, please leave it on LMS.