muCPU: an 8-bit MCU

In this revision of the code editor, I added line numbers and syntax highlighting (sort of). The line numbers consist of text on a canvas that is refreshed every time a key is pressed, a left click occurs, or a click and drag occurs. Syntax highlighting utilizes the Text.search() function supported by Tkinter. I also decided to use two files for this project to keep the main file (slightly) less cluttered. Code (with very few comments) below:

Main file (assemblerv2.py):

from Tkinter import *
from tkFileDialog import *
from tkFont import *
import os.path
import sys
import re

from assemblerlib import *

basefilename = "Untitled"
filename = ""
fileexists = False

canvaswidthdefault = 20
canvaswidth = canvaswidthdefault
currentlength = 1

mnemonicstring = "sll|add|sub|nand|nor|bez|bnez|bgez|blez|bgz|blz|lw|sw"
registerstring = "r0|r1|r2|r3|r4|r5|r6|r7"

def openFile():
    global filename
    global basefilename
    openfilename = askopenfilename()
    if openfilename is not None:
        filename = openfilename
        basefilename = os.path.basename(filename)
        asmfile = open(filename, "r")
        asmfile.seek(0)
        asmdata = asmfile.read()
        textArea.delete("1.0", "end - 1c")
        textArea.insert("1.0", asmdata)
        asmfile.close()
        filemenu.entryconfig(filemenu.index("Save"), state = NORMAL)
        frame.title("muCPU Assembler [" + basefilename + "]")
        frame.focus()
        initonOpen()
    
def saveFile():
    global filename
    asmdata = textArea.get("1.0", "end - 1c")
    asmfile = open(filename, "w")
    asmfile.seek(0)
    asmfile.truncate()
    asmfile.write(asmdata)
    asmfile.close()

def saveFileAs():
    global filename
    global fileexists
    global basefilename
    saveasfilename = asksaveasfilename()
    if saveasfilename is not None:
        filename = saveasfilename
        basefilename = os.path.basename(filename)
        fileexists = True
        asmdata = textArea.get("1.0", "end - 1c")
        asmfile = open(filename, "w")
        asmfile.seek(0)
        asmfile.truncate()
        asmfile.write(asmdata)
        asmfile.close()
        filemenu.entryconfig(filemenu.index("Save"), state = NORMAL)
        frame.title("muCPU Assembler [" + basefilename + "]")
        frame.focus()
    
        
def exitApp():
    frame.destroy()
    sys.exit()
    
def compileASM():
    global filename
    cpu_out = ""
    asm_in = textArea.get("1.0", END)
    asmlines = re.split("\n", asm_in)
    for i in range (len(asmlines)):
        if (asmlines[i] != ""):
            cpu_out += str(i) + " => x\"" + decode(asmlines[i]) + "\",\n"
    name, ext = os.path.splitext(filename)
    hexfilename = name + ".hex"
    hexfile = open(hexfilename, "w")
    hexfile.seek(0)
    hexfile.truncate()
    hexfile.write(cpu_out)
    hexfile.close()

def updateLinesEvent(event):
    drawLinenums()

def updateHighlightEvent(event):
    highlightSyntax((re.split("\.", textArea.index(INSERT))[0] + ".0"), INSERT)

def initonOpen():
    highlightSyntax("1.0", END)
    drawLinenums()


def drawLinenums():
    global canvaswidth
    global canvaswidthdefault
    linenumbers.delete("all")
    i = textArea.index("@0,0")
    while True:
        dline = textArea.dlineinfo(i)
        if dline is None: break
        y = dline[1]
        linenum = str(i).split(".")[0]
        linenumbers.create_text(canvaswidth, y, anchor=NE,text=linenum)
        i = textArea.index("%s+1line" % i)
        linenumbers.config(width = canvaswidth)
        

def highlightSyntax(start, end):
    global mnemonicstring
    global registerstring
    mnemoniclen = StringVar()
    registerlen = StringVar()
    numberlen = StringVar()
    pos = start
    while True:
        pos = textArea.search(mnemonicstring, pos, end, regexp = True, count = mnemoniclen)
        #print pos
        if not pos: break
        textArea.tag_add("mnemonic", pos, pos + " + " + str(mnemoniclen.get()) + "c")
        posarry = re.split("\.", pos)
        posarry[1] = str(int(posarry[1]) + 1)
        pos = posarry[0] + "." + posarry[1] + ("0" * (len(posarry[1]) - 2))
    pos = start
    while True:
        pos = textArea.search("-?\\d", pos, end, regexp = True, count = numberlen)
        #print pos
        if not pos: break
        textArea.tag_add("number", pos, pos + " + " + str(numberlen.get()) + "c")
        posarry = re.split("\.", pos)
        posarry[1] = str(int(posarry[1]) + 1)
        pos = posarry[0] + "." + posarry[1] + ("0" * (len(posarry[1]) - 2))
    pos = start
    while True:
        pos = textArea.search(registerstring, pos, end, regexp = True, count = registerlen)
        #print pos
        if not pos: break
        textArea.tag_add("register", pos, pos + " + " + str(registerlen.get()) + "c")
        posarry = re.split("\.", pos)
        posarry[1] = str(int(posarry[1]) + 1)
        pos = posarry[0] + "." + posarry[1] + ("0" * (len(posarry[1]) - 2))
    
        
Tk().withdraw()
frame = Toplevel(bg="#D8D8D8")

frame.bind("<Key>", updateLinesEvent)
frame.bind("<Button-1>", updateLinesEvent)
frame.bind("<B1-Motion>", updateLinesEvent)
frame.bind("<ButtonRelease-1>", updateLinesEvent)
frame.bind("<Key>", updateHighlightEvent)

scrollbar = Scrollbar(frame)
scrollbar.pack(side = RIGHT, fill = Y)
frame.title("muCPU Assembler [" + basefilename + "]")
textArea = Text(frame, height = 30, width = 100, padx = 3, pady = 3, yscrollcommand = scrollbar.set)
textArea.pack(side=RIGHT)
scrollbar.config(command=textArea.yview)

mnemonicfont = Font(frame, family = "Courier", size = 10, weight = "bold")
textArea.tag_config("mnemonic", foreground = "blue", font = mnemonicfont)
numberfont = Font(frame, family = "Courier", size = 10)
textArea.tag_config("number", foreground = "#585858", font = numberfont)
registerfont = Font(frame, family = "Courier", size = 10)#, slant = "italic")
textArea.tag_config("register", foreground = "red", font = registerfont)

linenumbers = Canvas(frame, width = canvaswidthdefault, height = 487, bg = "#D8D8D8", highlightbackground = "#D8D8D8")
linenumbers.pack()

menubar = Menu(frame)
filemenu = Menu(menubar, tearoff=0)
filemenu.add_command(label="Open", command=openFile)
filemenu.add_command(label="Save", command=saveFile, state = DISABLED)
filemenu.add_command(label="Save as...", command=saveFileAs)
filemenu.add_command(label="Exit", command=exitApp)
menubar.add_cascade(label="File", menu=filemenu)
runmenu = Menu(menubar, tearoff=0)
runmenu.add_command(label="Compile", command=compileASM)
menubar.add_cascade(label="Run", menu=runmenu)
frame.config(menu=menubar)

initonOpen()

frame.resizable(0,0)
frame.mainloop()

#Current code
"""
lw r4, 176(r0)
lw r3, 177(r0)
sub r2, r4, r1
bez r2, 8
sw r1, 252(r0)
bez r0, -8
add r1, r1, r3
sll r0, r0, r0
bez r0, -2
"""

Library file (assemblerlib.py):

import re

def asmtoint(asm):
    asm_split = re.split(" |, |\(|\)", asm)
    args = []
    for i in range (len(asm_split)):
        if (asm_split[i] != ""):
            args.append(asm_split[i])
    #print args
    opcode = 0
    func = 0
    rd = 0
    rs = 0
    rt = 0
    imm = 0
    if (args[0] == "sll"):
        if (len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 0
        func = 0
        rd = int(args[1][1:])
        rs = int(args[2][1:])
        rt = int(args[3][1:])
    elif (args[0] == "add"):
        if (len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 0
        func = 1
        rd = int(args[1][1:])
        rs = int(args[2][1:])
        rt = int(args[3][1:])
    elif (args[0] == "sub"):
        if (len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 0
        func = 2
        rd = int(args[1][1:])
        rs = int(args[2][1:])
        rt = int(args[3][1:])
    elif (args[0] == "nand"):
        if (len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 0
        func = 3
        rd = int(args[1][1:])
        rs = int(args[2][1:])
        rt = int(args[3][1:])
    elif (args[0] == "nor"):
        if (len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 0
        func = 4
        rd = int(args[1][1:])
        rs = int(args[2][1:])
        rt = int(args[3][1:])
    elif (args[0] == "bez"):
        if (len(args) != 3):
            return 0,0,0,0,0,0
        opcode = 1
        rt = 0
        rs = int(args[1][1:])
        imm = int(args[2])
    elif (args[0] == "bnez"):
        if (len(args) != 3):
            return 0,0,0,0,0,0
        opcode = 1
        rt = 1
        rs = int(args[1][1:])
        imm = int(args[2])
    elif (args[0] == "bgez"):
        if (len(args) != 3):
            return 0,0,0,0,0,0
        opcode = 1
        rt = 2
        rs = int(args[1][1:])
        imm = int(args[2])
    elif (args[0] == "blez"):
        if (len(args) != 3):
            return 0,0,0,0,0,0
        opcode = 1
        rt = 3
        rs = int(args[1][1:])
        imm = int(args[2])
    elif (args[0] == "bgz"):
        if (len(args) != 3):
            return 0,0,0,0,0,0
        opcode = 1
        rt = 4
        rs = int(args[1][1:])
        imm = int(args[2])
    elif (args[0] == "blz"):
        if (len(args) != 3):
            return 0
        opcode = 1
        rt = 5
        rs = int(args[1][1:])
        imm = int(args[2])
    elif (args[0] == "lw"):
        if (args[-1] == ''):
            args = args[0:-1]
        if (len(args) != 3 and len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 2
        rt = int(args[1][1:])
        if (len(args) == 3):
            imm = 0
            rs = int(args[2][1:])
        else:
            imm = int(args[2])
            rs = int(args[3][1:])
    elif (args[0] == "sw"):
        if (args[-1] == ''):
            args = args[0:-1]
        if (len(args) != 3 and len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 3
        rt = int(args[1][1:])
        if (len(args) == 3):
            imm = 0
            rs = int(args[2][1:])
        else:
            imm = int(args[2])
            rs = int(args[3][1:])
    else:
        return 0,0,0,0,0,0
    return opcode, rs, rt, rd, func, imm

def inttohex(opcode, rs, rt, rd, func, imm):
    if (opcode == 0):
        opstr = format(opcode, '02b')
        rsstr = format(rs, '03b')
        rtstr = format(rt, '03b')
        rdstr = format(rd, '03b')
        fnstr = format(func, '05b')
        #print opstr, rsstr, rtstr, rdstr, fnstr
        instruction = opstr + rsstr + rtstr + rdstr + fnstr
    else :
        opstr = format(opcode, '02b')
        rtstr = format(rt, '03b')
        rsstr = format(rs, '03b')
        if (imm < 0):
            imm2s = ((-imm) ^ 255) + 1
            immstr = format(imm2s, '08b')
        else :
            immstr = format(imm, '08b')
        #print opstr, rtstr, rsstr, immstr
        instruction = opstr + rsstr + rtstr + immstr
    return format(int(instruction, 2), '04x')

def decode(asm):
    opcode, rs, rt, rd, func, imm = asmtoint(asm)
    instruction = inttohex(opcode, rs, rt, rd, func, imm)
    return instruction

Using the assembler I wrote in Python, I generated some hex code to store in the synthesized ROM on the FPGA. I also added a clock divider to slow the system clock from 50MHz down to 25Hz. This results in about a half second period between increments of the output vector.

Assembler Code:

lw r4, 176(r0) //value to count to should be stored at this address
lw r3, 177(r0) //increment value should be stored at this address
sub r2, r4, r1
bez r2, 8
sw r1, 252(r0)
bez r0, -8
add r1, r1, r3
sll r0, r0, r0
bez r0, -2

Hex Code:

0 => x"84b0",
1 => x"83b1",
2 => x"2142",
3 => x"5008",
4 => x"c1fc",
5 => x"40f8",
6 => x"0b21",
7 => x"0000",
8 => x"40fe"

As I was writing some simple test code to run on my processor testbench (in iSim), I found that it was tedious and slow to write hex instructions from assembly mnemonics. What better solution to a problem that is tedious than a program that does the hex encoding over and over. I wrote a simple application in Python (because of its simple, powerful string manipulation functions) to "assemble" very rudimentary assembly language instructions. Screenshots and code below.

Input:

Output:

0 => x"84b0",
1 => x"83b0",
2 => x"2142",
3 => x"5008",
4 => x"c1fc",
5 => x"40f8",
6 => x"0b21",
7 => x"0000",
8 => x"40fe",

from Tkinter import *
import re

from tkFileDialog import *
import os.path

import sys

filename = "Untitled"
fileexists = False

def asmtoint(asm):
    asm_split = re.split(" |, |\(|\)", asm)
    args = []
    for i in range (len(asm_split)):
        if (asm_split[i] != ""):
            args.append(asm_split[i])
    #print args
    opcode = 0
    func = 0
    rd = 0
    rs = 0
    rt = 0
    imm = 0
    if (args[0] == "sll"):
        if (len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 0
        func = 0
        rd = int(args[1][1:])
        rs = int(args[2][1:])
        rt = int(args[3][1:])
    elif (args[0] == "add"):
        if (len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 0
        func = 1
        rd = int(args[1][1:])
        rs = int(args[2][1:])
        rt = int(args[3][1:])
    elif (args[0] == "sub"):
        if (len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 0
        func = 2
        rd = int(args[1][1:])
        rs = int(args[2][1:])
        rt = int(args[3][1:])
    elif (args[0] == "nand"):
        if (len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 0
        func = 3
        rd = int(args[1][1:])
        rs = int(args[2][1:])
        rt = int(args[3][1:])
    elif (args[0] == "nor"):
        if (len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 0
        func = 4
        rd = int(args[1][1:])
        rs = int(args[2][1:])
        rt = int(args[3][1:])
    elif (args[0] == "bez"):
        if (len(args) != 3):
            return 0,0,0,0,0,0
        opcode = 1
        rt = 0
        rs = int(args[1][1:])
        imm = int(args[2])
    elif (args[0] == "bnez"):
        if (len(args) != 3):
            return 0,0,0,0,0,0
        opcode = 1
        rt = 1
        rs = int(args[1][1:])
        imm = int(args[2])
    elif (args[0] == "bgez"):
        if (len(args) != 3):
            return 0,0,0,0,0,0
        opcode = 1
        rt = 2
        rs = int(args[1][1:])
        imm = int(args[2])
    elif (args[0] == "blez"):
        if (len(args) != 3):
            return 0,0,0,0,0,0
        opcode = 1
        rt = 3
        rs = int(args[1][1:])
        imm = int(args[2])
    elif (args[0] == "bgz"):
        if (len(args) != 3):
            return 0,0,0,0,0,0
        opcode = 1
        rt = 4
        rs = int(args[1][1:])
        imm = int(args[2])
    elif (args[0] == "blz"):
        if (len(args) != 3):
            return 0
        opcode = 1
        rt = 5
        rs = int(args[1][1:])
        imm = int(args[2])
    elif (args[0] == "lw"):
        if (args[-1] == ''):
            args = args[0:-1]
        if (len(args) != 3 and len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 2
        rt = int(args[1][1:])
        if (len(args) == 3):
            imm = 0
            rs = int(args[2][1:])
        else:
            imm = int(args[2])
            rs = int(args[3][1:])
    elif (args[0] == "sw"):
        if (args[-1] == ''):
            args = args[0:-1]
        if (len(args) != 3 and len(args) != 4):
            return 0,0,0,0,0,0
        opcode = 3
        rt = int(args[1][1:])
        if (len(args) == 3):
            imm = 0
            rs = int(args[2][1:])
        else:
            imm = int(args[2])
            rs = int(args[3][1:])
    else:
        return 0,0,0,0,0,0
    return opcode, rs, rt, rd, func, imm

def inttohex(opcode, rs, rt, rd, func, imm):
    if (opcode == 0):
        opstr = format(opcode, '02b')
        rsstr = format(rs, '03b')
        rtstr = format(rt, '03b')
        rdstr = format(rd, '03b')
        fnstr = format(func, '05b')
        #print opstr, rsstr, rtstr, rdstr, fnstr
        instruction = opstr + rsstr + rtstr + rdstr + fnstr
    else :
        opstr = format(opcode, '02b')
        rtstr = format(rt, '03b')
        rsstr = format(rs, '03b')
        if (imm < 0):
            imm2s = ((-imm) ^ 255) + 1
            immstr = format(imm2s, '08b')
        else :
            immstr = format(imm, '08b')
        #print opstr, rtstr, rsstr, immstr
        instruction = opstr + rsstr + rtstr + immstr
    return format(int(instruction, 2), '04x')

def decode(asm):
    opcode, rs, rt, rd, func, imm = asmtoint(asm)
    instruction = inttohex(opcode, rs, rt, rd, func, imm)
    return instruction

def openFile():
    global filename
    openfilename = askopenfilename()
    if openfilename is not None:
        filename = openfilename
        asmfile = open(filename, "r")
        asmfile.seek(0)
        asmdata = asmfile.read()
        textArea.delete("1.0", "end - 1c")
        textArea.insert("1.0", asmdata)
        asmfile.close()
        filemenu.entryconfig(filemenu.index("Save"), state = NORMAL)
        frame.title("muCPU Assembler [" + filename + "]")
        frame.focus()
    
def saveFile():
    global filename
    asmdata = textArea.get("1.0", "end - 1c")
    asmfile = open(filename, "w")
    asmfile.seek(0)
    asmfile.truncate()
    asmfile.write(asmdata)
    asmfile.close()

def saveFileAs():
    global filename
    global fileexists
    saveasfilename = asksaveasfilename()
    if saveasfilename is not None:
        filename = saveasfilename
        fileexists = True
        asmdata = textArea.get("1.0", "end - 1c")
        asmfile = open(filename, "w")
        asmfile.seek(0)
        asmfile.truncate()
        asmfile.write(asmdata)
        asmfile.close()
        filemenu.entryconfig(filemenu.index("Save"), state = NORMAL)
        frame.title("muCPU Assembler [" + filename + "]")
        frame.focus()
    
        
def exitApp():
    frame.destroy()
    sys.exit()
    
def compileASM():
    global filename
    cpu_out = ""
    asm_in = textArea.get("1.0", END)
    asmlines = re.split("\n", asm_in)
    for i in range (len(asmlines)):
        if (asmlines[i] != ""):
            #print asmlines[i]
            cpu_out += str(i) + " => x\"" + decode(asmlines[i]) + "\",\n"
    #print cpu_out
    name, ext = os.path.splitext(filename)
    hexfilename = name + ".hex"
    hexfile = open(hexfilename, "w")
    hexfile.seek(0)
    hexfile.truncate()
    hexfile.write(cpu_out)
    hexfile.close()
    
Tk().withdraw()
frame = Toplevel()

scrollbar = Scrollbar(frame)
scrollbar.pack(side = RIGHT, fill = Y)
frame.title("muCPU Assembler [" + filename + "]")
textArea = Text(frame, height = 30, width = 100, padx = 3, pady = 3, yscrollcommand = scrollbar.set)
textArea.pack(side=RIGHT)
scrollbar.config(command=textArea.yview)

menubar = Menu(frame)
filemenu = Menu(menubar, tearoff=0)
filemenu.add_command(label="Open", command=openFile)
filemenu.add_command(label="Save", command=saveFile, state = DISABLED)
filemenu.add_command(label="Save as...", command=saveFileAs)
filemenu.add_command(label="Exit", command=exitApp)
menubar.add_cascade(label="File", menu=filemenu)
runmenu = Menu(menubar, tearoff=0)
runmenu.add_command(label="Compile", command=compileASM)
menubar.add_cascade(label="Run", menu=runmenu)
frame.config(menu=menubar)

frame.minsize(750, 450)
frame.maxsize(750, 450)
frame.mainloop()

#Sample counting loop code
"""
lw r4, 176(r0)
lw r3, 177(r0)
sub r2, r4, r1
bez r2, 8
sw r1, 252(r0)
bez r0, -8
add r1, r1, r3
sll r0, r0, r0
bez r0, -2
"""

Going into this project, my vhdl experience consisted of code that made an led blink on my minispartan6+ dev board. With several youtube videos and a very useful book (Free Range VHDL), my skills quickly improved. The thing that frustrated me the most was the occasionally cryptic error messages that the Xilinx ISE Design Suite spat out. Oftentimes a mistake in one source file would lead to an error in another source file that used the original file, or an error in a file that used the file that used the original file. Anyway, I started the build with an ALU and instruction decoder; pretty much everything else stays similar even when the instruction set changes (I changed mine several times to allow for more functionality and an increased amount of registers). For the coding style of the project I used a mixture of behavioral (mostly for memory/registers), dataflow (for subcomponents), and structural (for the larger components). My intention of this project was more to learn about VHDL, and less to build a fast processor (duh...why would you build a slow, wheezing machine on a board that costs 5x as much as an MCU that can run circles around a soft processor), so the combination of styles is not likely very efficient. It was really cool to synthesize my code and view the RTL schematic that the synthesizer generated.