如何自己動手設計組譯器

訊息

參考文獻

簡體版

English

簡介

組譯器 (Assembler) 是用來將組合語言轉換成機器碼的程式,這是系統程式的一個重要主題,在本文中,我們將設計出 Beck 系統程式經典教科書中的 SIC 機器之組譯器,以下是一個簡單的 SIC 組合語言範例。

.-------------- SUM.sic -----------------------
SUMP START 0
LOOP LDA I . i++
 ADD ONE
 STA I
 ADD SUM . sum += i
 STA SUM
 LDA I
 COMP N . if (i < N) goto LOOP
 JLT LOOP 
 LDA SUM
 RSUB
I WORD 0
SUM WORD 0
N WORD 10
ONE WORD 1
END SUMP

實作

我們使用 C# 實作出 SIC 組合語言的組譯器,其程式原始碼如下所示。

// 共有兩個檔案
// 程式檔 : SicAssembler.cs
// 測試檔 : SUM.sic
// ------------------------------ SicAssembler.cs ------------------------------------
using System;
using System.Collections.Generic;
using System.Text;
using System.IO;
using System.Text.RegularExpressions;
using Microsoft.VisualBasic;
namespace SICAssembler
{
 public class SIC
 {
 static void Main(string[] args)
 {
 SicAssembler asm = new SicAssembler();
 asm.parse(args[0]);
 asm.pass2();
 SicMemory memory = new SicMemory();
 SicLoader.load(asm.codes, memory);
 SicVirtualMachine vm = new SicVirtualMachine(memory, 0);
 vm.run();
 }
 }
 public class SicVirtualMachine
 {
 public static OpTable opTable = new OpTable();
 SicMemory mem;
 int A = 0, X = 0, L = 0, SW = 0, PC = 0;
 public SicVirtualMachine(SicMemory pMemory, int pCounter) {
 mem = pMemory;
 PC = pCounter;
 }
 public void dump()
 {
 Debug.log("A=" + A + " X=" + X + " L=" + L + " SW=" + SW + " PC=" + PC);
 }
 public void run() {
 while (runCode())
 {
 }
 dump();
 }
 public int compare(int a, int b)
 {
 if (a > b) return 1;
 else if (a < b) return -1;
 else return 0;
 }
 public bool runCode()
 {
 byte op = mem.memory[PC];
 byte a1 = mem.memory[PC + 1];
 byte a2 = mem.memory[PC + 2];
 int m = ((a1 & 0x7F) << 8) + a2;
 int word = mem.getWord(m);
 String opStr = String.Format("{0:X2}", op);
 Debug.log(opTable.codeToOp[opStr]+"\tmem("+String.Format("{0:X2}", m)+")\t= "+word);
 switch (op) {
 case 0x18: // ADD m
 A += word;
 break;
 case 0x40: // AND m
 A &= word;
 break;
 case 0x28: // COMP m
 SW = compare(A, word);
 break;
 case 0x24: // DIV m
 A /= word;
 break;
 case 0x3C: // J m
 PC = m;
 break;
 case 0x30: // JEQ m
 if (SW == 0) PC = m;
 break;
 case 0x34: // JGT m
 if (SW > 0) PC = m;
 break;
 case 0x38: // JLT m
 if (SW < 0) PC = m;
 break;
 case 0x48: // JSUB m
 L = PC;
 PC = m;
 break;
 case 0x00: // LDA m
 A = word;
 break;
 case 0x50: // LDCH m
 A = (A & 0xFF0) | mem.memory[m];
 break;
 case 0x04: // LDX m
 X = word;
 break;
 case 0x20: // MUL m
 A *= word;
 break;
 case 0x44: // OR m
 A |= word;
 break;
// case 0xD8: // RD m
 case 0x4C: // RSUB
 if (L == 0) return false;
 PC = L;
 break;
 case 0x0C: // STA m
 mem.setWord(A, m);
 break;
 case 0x54: // STCH m
 mem.memory[m] = (byte)A;
 break;
 case 0x10: // STX m
 mem.setWord(X, m);
 break;
 case 0x1C: // SUB m
 A -= word;
 break;
// case 0xB0: // SVC x
// case 0xE0: // TD m
// break;
 case 0x2C: // TIX m
 X++;
 SW = compare(X, word);
 break;
// case 0xDC: // WD m
// break;
 default:
 Debug.log(String.Format("Error : op={0:X2} not found !", op));
 break;
 }
 PC += 3;
 return true;
 }
 }
 public class SicLoader
 {
 public static void load(CodeList codes, SicMemory m)
 {
 foreach (Code code in codes) {
 for (int i = 0; i < code.objCode.Length; i+=2)
 {
 String hex = code.objCode;
 int offset = code.offset + i/2;
 byte b = Byte.Parse(hex.Substring(i, 2), System.Globalization.NumberStyles.AllowHexSpecifier);
 m.memory[offset] = b;
 Debug.log(String.Format("{0:X2}:{1:X2}", offset, b));
 }
 }
 }
 }
 public class SicMemory
 {
 public byte[] memory = new byte[32768];
 public SicMemory() {
 for (int i = 0; i < memory.Length; i++)
 memory[i] = 0;
 }
 public void setWord(int word, int offset)
 {
 memory[offset] = (byte)(word >> 16);
 memory[offset + 1] = (byte)(word >> 8);
 memory[offset + 2] = (byte)(word);
 }
 public int getWord(int offset) 
 { 
 return memory[offset] << 16 | memory[offset + 1] << 8 | memory[offset + 2]; 
 }
 }
 public class SicAssembler
 {
 public static OpTable opTable = new OpTable();
 public CodeList codes = new CodeList();
 public SymbolTable symbolTable = new SymbolTable();
 public void parse(String pFileName)
 {
 Debug.log("======================= PARSE ================================");
 String text = IO.fileToText(pFileName);
 text = text.Replace('\r', ' ');
 String[] lines = text.Split('\n');
 int offset = 0;
 for (int i = 0; i < lines.Length; i++)
 {
 String line = lines[i];
 if (line.Trim().Length == 0) continue;
 if (line.StartsWith(".")) continue;
 Code code = new Code(line, opTable);
 code.lineNumber = i;
 if (code.offset >= 0)
 offset = code.offset;
 else
 code.offset = offset;
 Debug.log(code.ToString());
 if (code.label.Length > 0)
 symbolTable.Add(code.label, code);
 codes.Add(code);
 offset = offset + code.codeSize;
 }
 }
 public void pass2()
 {
 Debug.log("======================= PASS2 ================================");
 for (int i = 0; i < codes.Count; i++)
 {
 Code code = codes[i];
 if (!code.isPseudo)
 {
 if (code.arg.Length == 0)
 code.objCode += "0000";
 else
 {
 if (symbolTable.ContainsKey(code.arg))
 {
 int argCode = symbolTable[code.arg].offset;
 if (code.x == 'X')
 argCode += 0x8000;
 code.objCode += String.Format("{0:X4}", argCode);
 }
 else Debug.error("ARG " + code.arg + " not found !");
 }
 }
 Debug.log(code.ToString());
 }
 }
 }
 public class OpTable {
 public Dictionary<String, String> opToCode = new Dictionary<String, String>();
 public Dictionary<String, String> codeToOp = new Dictionary<String, String>();
 public static String opCodes = "ADD=18,ADDF=58,ADDR=90,AND=40,CLEAR=B4,COMP=28,COMPF=88,DIV=24,DIVF=64,DIVR=9C,FIX=C4," +
 "FLOAT=C0,HIO=F4,J=3C,JEQ=30,JGT=34,JLT=38,JSUB=48,LDA=00,LDB=68,LDCH=50,LDF=70,LDL=08," +
 "LDS=6C,LDT=74,LDX=04,LPS=D0,MUL=20,MULF=60,MULR=98,NORM=C8,OR=44,RD=D8,RMO=AC,RSUB=4C," +
 "SHIFTL=A4,SHIFTR=A8,SIO=F0,SSK=EC,STA=0C,STB=78,STCH=54,STF=80,STI=D4,STL=14,STS=7C,STSW=E8," +
 "STT=84,STX=10,SUB=1C,SUBF=5C,SUBR=94,SVC=B0,TD=E0,TIO=F8,TIX=2C,TIXR=B8,WD=DC," +
 "RESB=,RESW=,BYTE=,WORD=,START=,END=";
 public OpTable() {
 Debug.log("======================= OP TABLE ================================");
 String[] records = opCodes.Split(',');
 for (int i = 0; i < records.Length; i++)
 {
 String[] tokens = records[i].Split('=');
 opToCode.Add(tokens[0], tokens[1]);
 if (tokens.Length > 1 && tokens[1].Length>0)
 codeToOp.Add(tokens[1], tokens[0]);
 Debug.log(tokens[0] + "\t" + tokens[1]);
 }
 }
 }
 public class SymbolTable : Dictionary<String, Code> { }
 public class CodeList : List<Code> { }
 public class Code
 {
 public static String PSEUDO_OP = ",RESB,RESW,BYTE,WORD,START,END,";
 public String label = "", op = "", arg = "";
 public char x = ' ';
 public String objCode = null;
 public int offset = -1;
 public int codeSize = 0;
 public int lineNumber = 0;
 public bool isPseudo;
 public Code(String line, OpTable opTable)
 {
 String[] tokens = line.Split('\t');
 int argIdx;
 if (opTable.opToCode.ContainsKey(tokens[0]))
 {
 op = tokens[0].Trim();
 argIdx = 1;
 }
 else
 {
 label = tokens[0].Trim();
 op = tokens[1].Trim();
 argIdx = 2;
 }
 if (argIdx < tokens.Length)
 {
 String argStr = tokens[argIdx];
 String[] args = argStr.Split(',');
 arg = args[0].Trim();
 if (args.Length > 1)
 if (args[1].Trim().Equals("X"))
 x = 'X';
 }
 objCode = opTable.opToCode[op];
 if (objCode == null) Debug.error("op:" + op + " not found!");
 if (op.Equals("BYTE")) 
 {
 if (arg.StartsWith("C'")) // EOF BYTE C'EOF'
 {
 String str = arg.Substring(2, arg.Length - 3);
 codeSize = str.Length;
 objCode = "";
 for (int si = 0; si < str.Length; si++)
 {
 char ch = str[si];
 int chInt = ch;
 objCode += String.Format("{0:x2}", (uint) System.Convert.ToUInt32(chInt.ToString()));
 }
 objCode = objCode.ToUpper();
 }
 else if (arg.StartsWith("X'")) // OUTPUT BYTE X'05'
 {
 String str = arg.Substring(2, arg.Length - 3);
 objCode = str;
 codeSize = str.Length / 2;
 }
 else // THREE BYTE 3
 codeSize = 1; 
 }
 else if (op.Equals("WORD")) // FIVE WORD 5
 {
 codeSize = 3;
 objCode = String.Format("{0:X6}", Int32.Parse(arg));
 }
 else if (op.Equals("RESB")) // BUFFER RESB 1024
 codeSize = Int32.Parse(arg) * 1;
 else if (op.Equals("RESW")) // LENGTH RESW 1
 codeSize = Int32.Parse(arg) * 3;
 else if (op.Equals("START")) // COPY START 1000
 offset = Int32.Parse(arg, System.Globalization.NumberStyles.AllowHexSpecifier);
 else if (op.Equals("END"))
 codeSize = 0;
 else
 codeSize = 3; // SIC 中一個指令佔 3 byte
 isPseudo = (PSEUDO_OP.IndexOf(","+op+",") >= 0);
 }
 public override String ToString()
 {
 return lineNumber + "\t" + String.Format("{0:X}", offset) + "\t" + label + "\t" + op + "\t" + arg + "\t" + x+"\t"+objCode;
 }
 }
 public class IO
 {
 // 讀取文字檔,以字串形式傳回。
 public static String fileToText(String filePath)
 {
 StreamReader file = new StreamReader(filePath);
 String text = file.ReadToEnd();
 file.Close();
 return text;
 }
 }
 class Debug
 {
 public static void error(String msg)
 {
 Debug.log(msg);
 throw new Exception(msg);
 }
 public static void log(String msg)
 {
 Console.WriteLine(msg);
 }
 }
}

Facebook

[フレーム]

Facebook

[フレーム]

Wikidot

Post preview:

本網頁的作者、授權與引用方式

作者: 陳鍾誠,於金門大學資訊工程系,電子郵件:wt.ude.uqn|ccc#wt.ude.uqn|ccc,網站:http://ccckmit.wikidot.com。
授權: 本文採用創作共用 (Creative Common) 3.0 版的姓名標示─非商業性─相同方式分享授權條款,歡迎轉載或修改使用,但若做為商業使用時必須取得授權,引用本文時請參考下列格式。
中文版 (APA格式): 陳鍾誠 (28 Sep 2009 01:51),(網頁標題) 如何自己動手設計組譯器,(網站標題) 陳鍾誠的網站,取自 http://ccckmit.wikidot.com/code:assembler ,網頁修改第 12 版。
英文版 (APA格式): Chung-Chen Chen (28 Sep 2009 01:51), Retrieved from http://ccckmit.wikidot.com/code:assembler , Page Revision 12.

c# code

page revision: 12, last edited: 19 Oct 2010 09:09

Discuss (0)

金門大學:免費電子書、教材、程式、動畫

訊息

簡介

實作

Facebook

Facebook

Wikidot

Post preview:

Other interesting sites