examples/Kaleidoscope/Chapter7/toy.cpp

   1 #include "llvm/DerivedTypes.h"
   2 #include "llvm/ExecutionEngine/ExecutionEngine.h"
   3 #include "llvm/ExecutionEngine/JIT.h"
   4 #include "llvm/LLVMContext.h"
   5 #include "llvm/Module.h"
   6 #include "llvm/PassManager.h"
   7 #include "llvm/Analysis/Verifier.h"
   8 #include "llvm/Target/TargetData.h"
   9 #include "llvm/Target/TargetSelect.h"
  10 #include "llvm/Transforms/Scalar.h"
  11 #include "llvm/Support/IRBuilder.h"
  12 #include <cstdio>
  13 #include <string>
  14 #include <map>
  15 #include <vector>
  16 using namespace llvm;
  17
  18 //===----------------------------------------------------------------------===//
  19 // Lexer
  20 //===----------------------------------------------------------------------===//
  21
  22 // The lexer returns tokens [0-255] if it is an unknown character, otherwise one
  23 // of these for known things.
  24 enum Token {
  25   tok_eof = -1,
  26
  27   // commands
  28   tok_def = -2, tok_extern = -3,
  29
  30   // primary
  31   tok_identifier = -4, tok_number = -5,
  32
  33   // control
  34   tok_if = -6, tok_then = -7, tok_else = -8,
  35   tok_for = -9, tok_in = -10,
  36
  37   // operators
  38   tok_binary = -11, tok_unary = -12,
  39
  40   // var definition
  41   tok_var = -13
  42 };
  43
  44 static std::string IdentifierStr;  // Filled in if tok_identifier
  45 static double NumVal;              // Filled in if tok_number
  46
  47 /// gettok - Return the next token from standard input.
  48 static int gettok() {
  49   static int LastChar = ' ';
  50
  51   // Skip any whitespace.
  52   while (isspace(LastChar))
  53     LastChar = getchar();
  54
  55   if (isalpha(LastChar)) { // identifier: [a-zA-Z][a-zA-Z0-9]*
  56     IdentifierStr = LastChar;
  57     while (isalnum((LastChar = getchar())))
  58       IdentifierStr += LastChar;
  59
  60     if (IdentifierStr == "def") return tok_def;
  61     if (IdentifierStr == "extern") return tok_extern;
  62     if (IdentifierStr == "if") return tok_if;
  63     if (IdentifierStr == "then") return tok_then;
  64     if (IdentifierStr == "else") return tok_else;
  65     if (IdentifierStr == "for") return tok_for;
  66     if (IdentifierStr == "in") return tok_in;
  67     if (IdentifierStr == "binary") return tok_binary;
  68     if (IdentifierStr == "unary") return tok_unary;
  69     if (IdentifierStr == "var") return tok_var;
  70     return tok_identifier;
  71   }
  72
  73   if (isdigit(LastChar) || LastChar == '.') {   // Number: [0-9.]+
  74     std::string NumStr;
  75     do {
  76       NumStr += LastChar;
  77       LastChar = getchar();
  78     } while (isdigit(LastChar) || LastChar == '.');
  79
  80     NumVal = strtod(NumStr.c_str(), 0);
  81     return tok_number;
  82   }
  83
  84   if (LastChar == '#') {
  85     // Comment until end of line.
  86     do LastChar = getchar();
  87     while (LastChar != EOF && LastChar != '\n' && LastChar != '\r');
  88
  89     if (LastChar != EOF)
  90       return gettok();
  91   }
  92
  93   // Check for end of file.  Don't eat the EOF.
  94   if (LastChar == EOF)
  95     return tok_eof;
  96
  97   // Otherwise, just return the character as its ascii value.
  98   int ThisChar = LastChar;
  99   LastChar = getchar();
 100   return ThisChar;
 101 }
 102
 103 //===----------------------------------------------------------------------===//
 104 // Abstract Syntax Tree (aka Parse Tree)
 105 //===----------------------------------------------------------------------===//
 106
 107 /// ExprAST - Base class for all expression nodes.
 108 class ExprAST {
 109 public:
 110   virtual ~ExprAST() {}
 111   virtual Value *Codegen() = 0;
 112 };
 113
 114 /// NumberExprAST - Expression class for numeric literals like "1.0".
 115 class NumberExprAST : public ExprAST {
 116   double Val;
 117 public:
 118   NumberExprAST(double val) : Val(val) {}
 119   virtual Value *Codegen();
 120 };
 121
 122 /// VariableExprAST - Expression class for referencing a variable, like "a".
 123 class VariableExprAST : public ExprAST {
 124   std::string Name;
 125 public:
 126   VariableExprAST(const std::string &name) : Name(name) {}
 127   const std::string &getName() const { return Name; }
 128   virtual Value *Codegen();
 129 };
 130
 131 /// UnaryExprAST - Expression class for a unary operator.
 132 class UnaryExprAST : public ExprAST {
 133   char Opcode;
 134   ExprAST *Operand;
 135 public:
 136   UnaryExprAST(char opcode, ExprAST *operand)
 137     : Opcode(opcode), Operand(operand) {}
 138   virtual Value *Codegen();
 139 };
 140
 141 /// BinaryExprAST - Expression class for a binary operator.
 142 class BinaryExprAST : public ExprAST {
 143   char Op;
 144   ExprAST *LHS, *RHS;
 145 public:
 146   BinaryExprAST(char op, ExprAST *lhs, ExprAST *rhs)
 147     : Op(op), LHS(lhs), RHS(rhs) {}
 148   virtual Value *Codegen();
 149 };
 150
 151 /// CallExprAST - Expression class for function calls.
 152 class CallExprAST : public ExprAST {
 153   std::string Callee;
 154   std::vector<ExprAST*> Args;
 155 public:
 156   CallExprAST(const std::string &callee, std::vector<ExprAST*> &args)
 157     : Callee(callee), Args(args) {}
 158   virtual Value *Codegen();
 159 };
 160
 161 /// IfExprAST - Expression class for if/then/else.
 162 class IfExprAST : public ExprAST {
 163   ExprAST *Cond, *Then, *Else;
 164 public:
 165   IfExprAST(ExprAST *cond, ExprAST *then, ExprAST *_else)
 166   : Cond(cond), Then(then), Else(_else) {}
 167   virtual Value *Codegen();
 168 };
 169
 170 /// ForExprAST - Expression class for for/in.
 171 class ForExprAST : public ExprAST {
 172   std::string VarName;
 173   ExprAST *Start, *End, *Step, *Body;
 174 public:
 175   ForExprAST(const std::string &varname, ExprAST *start, ExprAST *end,
 176              ExprAST *step, ExprAST *body)
 177     : VarName(varname), Start(start), End(end), Step(step), Body(body) {}
 178   virtual Value *Codegen();
 179 };
 180
 181 /// VarExprAST - Expression class for var/in
 182 class VarExprAST : public ExprAST {
 183   std::vector<std::pair<std::string, ExprAST*> > VarNames;
 184   ExprAST *Body;
 185 public:
 186   VarExprAST(const std::vector<std::pair<std::string, ExprAST*> > &varnames,
 187              ExprAST *body)
 188   : VarNames(varnames), Body(body) {}
 189
 190   virtual Value *Codegen();
 191 };
 192
 193 /// PrototypeAST - This class represents the "prototype" for a function,
 194 /// which captures its argument names as well as if it is an operator.
 195 class PrototypeAST {
 196   std::string Name;
 197   std::vector<std::string> Args;
 198   bool isOperator;
 199   unsigned Precedence;  // Precedence if a binary op.
 200 public:
 201   PrototypeAST(const std::string &name, const std::vector<std::string> &args,
 202                bool isoperator = false, unsigned prec = 0)
 203   : Name(name), Args(args), isOperator(isoperator), Precedence(prec) {}
 204
 205   bool isUnaryOp() const { return isOperator && Args.size() == 1; }
 206   bool isBinaryOp() const { return isOperator && Args.size() == 2; }
 207
 208   char getOperatorName() const {
 209     assert(isUnaryOp() || isBinaryOp());
 210     return Name[Name.size()-1];
 211   }
 212
 213   unsigned getBinaryPrecedence() const { return Precedence; }
 214
 215   Function *Codegen();
 216
 217   void CreateArgumentAllocas(Function *F);
 218 };
 219
 220 /// FunctionAST - This class represents a function definition itself.
 221 class FunctionAST {
 222   PrototypeAST *Proto;
 223   ExprAST *Body;
 224 public:
 225   FunctionAST(PrototypeAST *proto, ExprAST *body)
 226     : Proto(proto), Body(body) {}
 227
 228   Function *Codegen();
 229 };
 230
 231 //===----------------------------------------------------------------------===//
 232 // Parser
 233 //===----------------------------------------------------------------------===//
 234
 235 /// CurTok/getNextToken - Provide a simple token buffer.  CurTok is the current
 236 /// token the parser is looking at.  getNextToken reads another token from the
 237 /// lexer and updates CurTok with its results.
 238 static int CurTok;
 239 static int getNextToken() {
 240   return CurTok = gettok();
 241 }
 242
 243 /// BinopPrecedence - This holds the precedence for each binary operator that is
 244 /// defined.
 245 static std::map<char, int> BinopPrecedence;
 246
 247 /// GetTokPrecedence - Get the precedence of the pending binary operator token.
 248 static int GetTokPrecedence() {
 249   if (!isascii(CurTok))
 250     return -1;
 251
 252   // Make sure it's a declared binop.
 253   int TokPrec = BinopPrecedence[CurTok];
 254   if (TokPrec <= 0) return -1;
 255   return TokPrec;
 256 }
 257
 258 /// Error* - These are little helper functions for error handling.
 259 ExprAST *Error(const char *Str) { fprintf(stderr, "Error: %s\n", Str);return 0;}
 260 PrototypeAST *ErrorP(const char *Str) { Error(Str); return 0; }
 261 FunctionAST *ErrorF(const char *Str) { Error(Str); return 0; }
 262
 263 static ExprAST *ParseExpression();
 264
 265 /// identifierexpr
 266 ///   ::= identifier
 267 ///   ::= identifier '(' expression* ')'
 268 static ExprAST *ParseIdentifierExpr() {
 269   std::string IdName = IdentifierStr;
 270
 271   getNextToken();  // eat identifier.
 272
 273   if (CurTok != '(') // Simple variable ref.
 274     return new VariableExprAST(IdName);
 275
 276   // Call.
 277   getNextToken();  // eat (
 278   std::vector<ExprAST*> Args;
 279   if (CurTok != ')') {
 280     while (1) {
 281       ExprAST *Arg = ParseExpression();
 282       if (!Arg) return 0;
 283       Args.push_back(Arg);
 284
 285       if (CurTok == ')') break;
 286
 287       if (CurTok != ',')
 288         return Error("Expected ')' or ',' in argument list");
 289       getNextToken();
 290     }
 291   }
 292
 293   // Eat the ')'.
 294   getNextToken();
 295
 296   return new CallExprAST(IdName, Args);
 297 }
 298
 299 /// numberexpr ::= number
 300 static ExprAST *ParseNumberExpr() {
 301   ExprAST *Result = new NumberExprAST(NumVal);
 302   getNextToken(); // consume the number
 303   return Result;
 304 }
 305
 306 /// parenexpr ::= '(' expression ')'
 307 static ExprAST *ParseParenExpr() {
 308   getNextToken();  // eat (.
 309   ExprAST *V = ParseExpression();
 310   if (!V) return 0;
 311
 312   if (CurTok != ')')
 313     return Error("expected ')'");
 314   getNextToken();  // eat ).
 315   return V;
 316 }
 317
 318 /// ifexpr ::= 'if' expression 'then' expression 'else' expression
 319 static ExprAST *ParseIfExpr() {
 320   getNextToken();  // eat the if.
 321
 322   // condition.
 323   ExprAST *Cond = ParseExpression();
 324   if (!Cond) return 0;
 325
 326   if (CurTok != tok_then)
 327     return Error("expected then");
 328   getNextToken();  // eat the then
 329
 330   ExprAST *Then = ParseExpression();
 331   if (Then == 0) return 0;
 332
 333   if (CurTok != tok_else)
 334     return Error("expected else");
 335
 336   getNextToken();
 337
 338   ExprAST *Else = ParseExpression();
 339   if (!Else) return 0;
 340
 341   return new IfExprAST(Cond, Then, Else);
 342 }
 343
 344 /// forexpr ::= 'for' identifier '=' expr ',' expr (',' expr)? 'in' expression
 345 static ExprAST *ParseForExpr() {
 346   getNextToken();  // eat the for.
 347
 348   if (CurTok != tok_identifier)
 349     return Error("expected identifier after for");
 350
 351   std::string IdName = IdentifierStr;
 352   getNextToken();  // eat identifier.
 353
 354   if (CurTok != '=')
 355     return Error("expected '=' after for");
 356   getNextToken();  // eat '='.
 357
 358
 359   ExprAST *Start = ParseExpression();
 360   if (Start == 0) return 0;
 361   if (CurTok != ',')
 362     return Error("expected ',' after for start value");
 363   getNextToken();
 364
 365   ExprAST *End = ParseExpression();
 366   if (End == 0) return 0;
 367
 368   // The step value is optional.
 369   ExprAST *Step = 0;
 370   if (CurTok == ',') {
 371     getNextToken();
 372     Step = ParseExpression();
 373     if (Step == 0) return 0;
 374   }
 375
 376   if (CurTok != tok_in)
 377     return Error("expected 'in' after for");
 378   getNextToken();  // eat 'in'.
 379
 380   ExprAST *Body = ParseExpression();
 381   if (Body == 0) return 0;
 382
 383   return new ForExprAST(IdName, Start, End, Step, Body);
 384 }
 385
 386 /// varexpr ::= 'var' identifier ('=' expression)?
 387 //                    (',' identifier ('=' expression)?)* 'in' expression
 388 static ExprAST *ParseVarExpr() {
 389   getNextToken();  // eat the var.
 390
 391   std::vector<std::pair<std::string, ExprAST*> > VarNames;
 392
 393   // At least one variable name is required.
 394   if (CurTok != tok_identifier)
 395     return Error("expected identifier after var");
 396
 397   while (1) {
 398     std::string Name = IdentifierStr;
 399     getNextToken();  // eat identifier.
 400
 401     // Read the optional initializer.
 402     ExprAST *Init = 0;
 403     if (CurTok == '=') {
 404       getNextToken(); // eat the '='.
 405
 406       Init = ParseExpression();
 407       if (Init == 0) return 0;
 408     }
 409
 410     VarNames.push_back(std::make_pair(Name, Init));
 411
 412     // End of var list, exit loop.
 413     if (CurTok != ',') break;
 414     getNextToken(); // eat the ','.
 415
 416     if (CurTok != tok_identifier)
 417       return Error("expected identifier list after var");
 418   }
 419
 420   // At this point, we have to have 'in'.
 421   if (CurTok != tok_in)
 422     return Error("expected 'in' keyword after 'var'");
 423   getNextToken();  // eat 'in'.
 424
 425   ExprAST *Body = ParseExpression();
 426   if (Body == 0) return 0;
 427
 428   return new VarExprAST(VarNames, Body);
 429 }
 430
 431 /// primary
 432 ///   ::= identifierexpr
 433 ///   ::= numberexpr
 434 ///   ::= parenexpr
 435 ///   ::= ifexpr
 436 ///   ::= forexpr
 437 ///   ::= varexpr
 438 static ExprAST *ParsePrimary() {
 439   switch (CurTok) {
 440   default: return Error("unknown token when expecting an expression");
 441   case tok_identifier: return ParseIdentifierExpr();
 442   case tok_number:     return ParseNumberExpr();
 443   case '(':            return ParseParenExpr();
 444   case tok_if:         return ParseIfExpr();
 445   case tok_for:        return ParseForExpr();
 446   case tok_var:        return ParseVarExpr();
 447   }
 448 }
 449
 450 /// unary
 451 ///   ::= primary
 452 ///   ::= '!' unary
 453 static ExprAST *ParseUnary() {
 454   // If the current token is not an operator, it must be a primary expr.
 455   if (!isascii(CurTok) || CurTok == '(' || CurTok == ',')
 456     return ParsePrimary();
 457
 458   // If this is a unary operator, read it.
 459   int Opc = CurTok;
 460   getNextToken();
 461   if (ExprAST *Operand = ParseUnary())
 462     return new UnaryExprAST(Opc, Operand);
 463   return 0;
 464 }
 465
 466 /// binoprhs
 467 ///   ::= ('+' unary)*
 468 static ExprAST *ParseBinOpRHS(int ExprPrec, ExprAST *LHS) {
 469   // If this is a binop, find its precedence.
 470   while (1) {
 471     int TokPrec = GetTokPrecedence();
 472
 473     // If this is a binop that binds at least as tightly as the current binop,
 474     // consume it, otherwise we are done.
 475     if (TokPrec < ExprPrec)
 476       return LHS;
 477
 478     // Okay, we know this is a binop.
 479     int BinOp = CurTok;
 480     getNextToken();  // eat binop
 481
 482     // Parse the unary expression after the binary operator.
 483     ExprAST *RHS = ParseUnary();
 484     if (!RHS) return 0;
 485
 486     // If BinOp binds less tightly with RHS than the operator after RHS, let
 487     // the pending operator take RHS as its LHS.
 488     int NextPrec = GetTokPrecedence();
 489     if (TokPrec < NextPrec) {
 490       RHS = ParseBinOpRHS(TokPrec+1, RHS);
 491       if (RHS == 0) return 0;
 492     }
 493
 494     // Merge LHS/RHS.
 495     LHS = new BinaryExprAST(BinOp, LHS, RHS);
 496   }
 497 }
 498
 499 /// expression
 500 ///   ::= unary binoprhs
 501 ///
 502 static ExprAST *ParseExpression() {
 503   ExprAST *LHS = ParseUnary();
 504   if (!LHS) return 0;
 505
 506   return ParseBinOpRHS(0, LHS);
 507 }
 508
 509 /// prototype
 510 ///   ::= id '(' id* ')'
 511 ///   ::= binary LETTER number? (id, id)
 512 ///   ::= unary LETTER (id)
 513 static PrototypeAST *ParsePrototype() {
 514   std::string FnName;
 515
 516   unsigned Kind = 0; // 0 = identifier, 1 = unary, 2 = binary.
 517   unsigned BinaryPrecedence = 30;
 518
 519   switch (CurTok) {
 520   default:
 521     return ErrorP("Expected function name in prototype");
 522   case tok_identifier:
 523     FnName = IdentifierStr;
 524     Kind = 0;
 525     getNextToken();
 526     break;
 527   case tok_unary:
 528     getNextToken();
 529     if (!isascii(CurTok))
 530       return ErrorP("Expected unary operator");
 531     FnName = "unary";
 532     FnName += (char)CurTok;
 533     Kind = 1;
 534     getNextToken();
 535     break;
 536   case tok_binary:
 537     getNextToken();
 538     if (!isascii(CurTok))
 539       return ErrorP("Expected binary operator");
 540     FnName = "binary";
 541     FnName += (char)CurTok;
 542     Kind = 2;
 543     getNextToken();
 544
 545     // Read the precedence if present.
 546     if (CurTok == tok_number) {
 547       if (NumVal < 1 || NumVal > 100)
 548         return ErrorP("Invalid precedecnce: must be 1..100");
 549       BinaryPrecedence = (unsigned)NumVal;
 550       getNextToken();
 551     }
 552     break;
 553   }
 554
 555   if (CurTok != '(')
 556     return ErrorP("Expected '(' in prototype");
 557
 558   std::vector<std::string> ArgNames;
 559   while (getNextToken() == tok_identifier)
 560     ArgNames.push_back(IdentifierStr);
 561   if (CurTok != ')')
 562     return ErrorP("Expected ')' in prototype");
 563
 564   // success.
 565   getNextToken();  // eat ')'.
 566
 567   // Verify right number of names for operator.
 568   if (Kind && ArgNames.size() != Kind)
 569     return ErrorP("Invalid number of operands for operator");
 570
 571   return new PrototypeAST(FnName, ArgNames, Kind != 0, BinaryPrecedence);
 572 }
 573
 574 /// definition ::= 'def' prototype expression
 575 static FunctionAST *ParseDefinition() {
 576   getNextToken();  // eat def.
 577   PrototypeAST *Proto = ParsePrototype();
 578   if (Proto == 0) return 0;
 579
 580   if (ExprAST *E = ParseExpression())
 581     return new FunctionAST(Proto, E);
 582   return 0;
 583 }
 584
 585 /// toplevelexpr ::= expression
 586 static FunctionAST *ParseTopLevelExpr() {
 587   if (ExprAST *E = ParseExpression()) {
 588     // Make an anonymous proto.
 589     PrototypeAST *Proto = new PrototypeAST("", std::vector<std::string>());
 590     return new FunctionAST(Proto, E);
 591   }
 592   return 0;
 593 }
 594
 595 /// external ::= 'extern' prototype
 596 static PrototypeAST *ParseExtern() {
 597   getNextToken();  // eat extern.
 598   return ParsePrototype();
 599 }
 600
 601 //===----------------------------------------------------------------------===//
 602 // Code Generation
 603 //===----------------------------------------------------------------------===//
 604
 605 static Module *TheModule;
 606 static IRBuilder<> Builder(getGlobalContext());
 607 static std::map<std::string, AllocaInst*> NamedValues;
 608 static FunctionPassManager *TheFPM;
 609
 610 Value *ErrorV(const char *Str) { Error(Str); return 0; }
 611
 612 /// CreateEntryBlockAlloca - Create an alloca instruction in the entry block of
 613 /// the function.  This is used for mutable variables etc.
 614 static AllocaInst *CreateEntryBlockAlloca(Function *TheFunction,
 615                                           const std::string &VarName) {
 616   IRBuilder<> TmpB(&TheFunction->getEntryBlock(),
 617                  TheFunction->getEntryBlock().begin());
 618   return TmpB.CreateAlloca(Type::getDoubleTy(getGlobalContext()), 0,
 619                            VarName.c_str());
 620 }
 621
 622 Value *NumberExprAST::Codegen() {
 623   return ConstantFP::get(getGlobalContext(), APFloat(Val));
 624 }
 625
 626 Value *VariableExprAST::Codegen() {
 627   // Look this variable up in the function.
 628   Value *V = NamedValues[Name];
 629   if (V == 0) return ErrorV("Unknown variable name");
 630
 631   // Load the value.
 632   return Builder.CreateLoad(V, Name.c_str());
 633 }
 634
 635 Value *UnaryExprAST::Codegen() {
 636   Value *OperandV = Operand->Codegen();
 637   if (OperandV == 0) return 0;
 638
 639   Function *F = TheModule->getFunction(std::string("unary")+Opcode);
 640   if (F == 0)
 641     return ErrorV("Unknown unary operator");
 642
 643   return Builder.CreateCall(F, OperandV, "unop");
 644 }
 645
 646 Value *BinaryExprAST::Codegen() {
 647   // Special case '=' because we don't want to emit the LHS as an expression.
 648   if (Op == '=') {
 649     // Assignment requires the LHS to be an identifier.
 650     VariableExprAST *LHSE = dynamic_cast<VariableExprAST*>(LHS);
 651     if (!LHSE)
 652       return ErrorV("destination of '=' must be a variable");
 653     // Codegen the RHS.
 654     Value *Val = RHS->Codegen();
 655     if (Val == 0) return 0;
 656
 657     // Look up the name.
 658     Value *Variable = NamedValues[LHSE->getName()];
 659     if (Variable == 0) return ErrorV("Unknown variable name");
 660
 661     Builder.CreateStore(Val, Variable);
 662     return Val;
 663   }
 664
 665   Value *L = LHS->Codegen();
 666   Value *R = RHS->Codegen();
 667   if (L == 0 || R == 0) return 0;
 668
 669   switch (Op) {
 670   case '+': return Builder.CreateFAdd(L, R, "addtmp");
 671   case '-': return Builder.CreateFSub(L, R, "subtmp");
 672   case '*': return Builder.CreateFMul(L, R, "multmp");
 673   case '<':
 674     L = Builder.CreateFCmpULT(L, R, "cmptmp");
 675     // Convert bool 0/1 to double 0.0 or 1.0
 676     return Builder.CreateUIToFP(L, Type::getDoubleTy(getGlobalContext()),
 677                                 "booltmp");
 678   default: break;
 679   }
 680
 681   // If it wasn't a builtin binary operator, it must be a user defined one. Emit
 682   // a call to it.
 683   Function *F = TheModule->getFunction(std::string("binary")+Op);
 684   assert(F && "binary operator not found!");
 685
 686   Value *Ops[] = { L, R };
 687   return Builder.CreateCall(F, Ops, Ops+2, "binop");
 688 }
 689
 690 Value *CallExprAST::Codegen() {
 691   // Look up the name in the global module table.
 692   Function *CalleeF = TheModule->getFunction(Callee);
 693   if (CalleeF == 0)
 694     return ErrorV("Unknown function referenced");
 695
 696   // If argument mismatch error.
 697   if (CalleeF->arg_size() != Args.size())
 698     return ErrorV("Incorrect # arguments passed");
 699
 700   std::vector<Value*> ArgsV;
 701   for (unsigned i = 0, e = Args.size(); i != e; ++i) {
 702     ArgsV.push_back(Args[i]->Codegen());
 703     if (ArgsV.back() == 0) return 0;
 704   }
 705
 706   return Builder.CreateCall(CalleeF, ArgsV.begin(), ArgsV.end(), "calltmp");
 707 }
 708
 709 Value *IfExprAST::Codegen() {
 710   Value *CondV = Cond->Codegen();
 711   if (CondV == 0) return 0;
 712
 713   // Convert condition to a bool by comparing equal to 0.0.
 714   CondV = Builder.CreateFCmpONE(CondV,
 715                               ConstantFP::get(getGlobalContext(), APFloat(0.0)),
 716                                 "ifcond");
 717
 718   Function *TheFunction = Builder.GetInsertBlock()->getParent();
 719
 720   // Create blocks for the then and else cases.  Insert the 'then' block at the
 721   // end of the function.
 722   BasicBlock *ThenBB = BasicBlock::Create(getGlobalContext(), "then", TheFunction);
 723   BasicBlock *ElseBB = BasicBlock::Create(getGlobalContext(), "else");
 724   BasicBlock *MergeBB = BasicBlock::Create(getGlobalContext(), "ifcont");
 725
 726   Builder.CreateCondBr(CondV, ThenBB, ElseBB);
 727
 728   // Emit then value.
 729   Builder.SetInsertPoint(ThenBB);
 730
 731   Value *ThenV = Then->Codegen();
 732   if (ThenV == 0) return 0;
 733
 734   Builder.CreateBr(MergeBB);
 735   // Codegen of 'Then' can change the current block, update ThenBB for the PHI.
 736   ThenBB = Builder.GetInsertBlock();
 737
 738   // Emit else block.
 739   TheFunction->getBasicBlockList().push_back(ElseBB);
 740   Builder.SetInsertPoint(ElseBB);
 741
 742   Value *ElseV = Else->Codegen();
 743   if (ElseV == 0) return 0;
 744
 745   Builder.CreateBr(MergeBB);
 746   // Codegen of 'Else' can change the current block, update ElseBB for the PHI.
 747   ElseBB = Builder.GetInsertBlock();
 748
 749   // Emit merge block.
 750   TheFunction->getBasicBlockList().push_back(MergeBB);
 751   Builder.SetInsertPoint(MergeBB);
 752   PHINode *PN = Builder.CreatePHI(Type::getDoubleTy(getGlobalContext()),
 753                                   "iftmp");
 754
 755   PN->addIncoming(ThenV, ThenBB);
 756   PN->addIncoming(ElseV, ElseBB);
 757   return PN;
 758 }
 759
 760 Value *ForExprAST::Codegen() {
 761   // Output this as:
 762   //   var = alloca double
 763   //   ...
 764   //   start = startexpr
 765   //   store start -> var
 766   //   goto loop
 767   // loop:
 768   //   ...
 769   //   bodyexpr
 770   //   ...
 771   // loopend:
 772   //   step = stepexpr
 773   //   endcond = endexpr
 774   //
 775   //   curvar = load var
 776   //   nextvar = curvar + step
 777   //   store nextvar -> var
 778   //   br endcond, loop, endloop
 779   // outloop:
 780
 781   Function *TheFunction = Builder.GetInsertBlock()->getParent();
 782
 783   // Create an alloca for the variable in the entry block.
 784   AllocaInst *Alloca = CreateEntryBlockAlloca(TheFunction, VarName);
 785
 786   // Emit the start code first, without 'variable' in scope.
 787   Value *StartVal = Start->Codegen();
 788   if (StartVal == 0) return 0;
 789
 790   // Store the value into the alloca.
 791   Builder.CreateStore(StartVal, Alloca);
 792
 793   // Make the new basic block for the loop header, inserting after current
 794   // block.
 795   BasicBlock *LoopBB = BasicBlock::Create(getGlobalContext(), "loop", TheFunction);
 796
 797   // Insert an explicit fall through from the current block to the LoopBB.
 798   Builder.CreateBr(LoopBB);
 799
 800   // Start insertion in LoopBB.
 801   Builder.SetInsertPoint(LoopBB);
 802
 803   // Within the loop, the variable is defined equal to the PHI node.  If it
 804   // shadows an existing variable, we have to restore it, so save it now.
 805   AllocaInst *OldVal = NamedValues[VarName];
 806   NamedValues[VarName] = Alloca;
 807
 808   // Emit the body of the loop.  This, like any other expr, can change the
 809   // current BB.  Note that we ignore the value computed by the body, but don't
 810   // allow an error.
 811   if (Body->Codegen() == 0)
 812     return 0;
 813
 814   // Emit the step value.
 815   Value *StepVal;
 816   if (Step) {
 817     StepVal = Step->Codegen();
 818     if (StepVal == 0) return 0;
 819   } else {
 820     // If not specified, use 1.0.
 821     StepVal = ConstantFP::get(getGlobalContext(), APFloat(1.0));
 822   }
 823
 824   // Compute the end condition.
 825   Value *EndCond = End->Codegen();
 826   if (EndCond == 0) return EndCond;
 827
 828   // Reload, increment, and restore the alloca.  This handles the case where
 829   // the body of the loop mutates the variable.
 830   Value *CurVar = Builder.CreateLoad(Alloca, VarName.c_str());
 831   Value *NextVar = Builder.CreateFAdd(CurVar, StepVal, "nextvar");
 832   Builder.CreateStore(NextVar, Alloca);
 833
 834   // Convert condition to a bool by comparing equal to 0.0.
 835   EndCond = Builder.CreateFCmpONE(EndCond,
 836                               ConstantFP::get(getGlobalContext(), APFloat(0.0)),
 837                                   "loopcond");
 838
 839   // Create the "after loop" block and insert it.
 840   BasicBlock *AfterBB = BasicBlock::Create(getGlobalContext(), "afterloop", TheFunction);
 841
 842   // Insert the conditional branch into the end of LoopEndBB.
 843   Builder.CreateCondBr(EndCond, LoopBB, AfterBB);
 844
 845   // Any new code will be inserted in AfterBB.
 846   Builder.SetInsertPoint(AfterBB);
 847
 848   // Restore the unshadowed variable.
 849   if (OldVal)
 850     NamedValues[VarName] = OldVal;
 851   else
 852     NamedValues.erase(VarName);
 853
 854
 855   // for expr always returns 0.0.
 856   return Constant::getNullValue(Type::getDoubleTy(getGlobalContext()));
 857 }
 858
 859 Value *VarExprAST::Codegen() {
 860   std::vector<AllocaInst *> OldBindings;
 861
 862   Function *TheFunction = Builder.GetInsertBlock()->getParent();
 863
 864   // Register all variables and emit their initializer.
 865   for (unsigned i = 0, e = VarNames.size(); i != e; ++i) {
 866     const std::string &VarName = VarNames[i].first;
 867     ExprAST *Init = VarNames[i].second;
 868
 869     // Emit the initializer before adding the variable to scope, this prevents
 870     // the initializer from referencing the variable itself, and permits stuff
 871     // like this:
 872     //  var a = 1 in
 873     //    var a = a in ...   # refers to outer 'a'.
 874     Value *InitVal;
 875     if (Init) {
 876       InitVal = Init->Codegen();
 877       if (InitVal == 0) return 0;
 878     } else { // If not specified, use 0.0.
 879       InitVal = ConstantFP::get(getGlobalContext(), APFloat(0.0));
 880     }
 881
 882     AllocaInst *Alloca = CreateEntryBlockAlloca(TheFunction, VarName);
 883     Builder.CreateStore(InitVal, Alloca);
 884
 885     // Remember the old variable binding so that we can restore the binding when
 886     // we unrecurse.
 887     OldBindings.push_back(NamedValues[VarName]);
 888
 889     // Remember this binding.
 890     NamedValues[VarName] = Alloca;
 891   }
 892
 893   // Codegen the body, now that all vars are in scope.
 894   Value *BodyVal = Body->Codegen();
 895   if (BodyVal == 0) return 0;
 896
 897   // Pop all our variables from scope.
 898   for (unsigned i = 0, e = VarNames.size(); i != e; ++i)
 899     NamedValues[VarNames[i].first] = OldBindings[i];
 900
 901   // Return the body computation.
 902   return BodyVal;
 903 }
 904
 905 Function *PrototypeAST::Codegen() {
 906   // Make the function type:  double(double,double) etc.
 907   std::vector<const Type*> Doubles(Args.size(),
 908                                    Type::getDoubleTy(getGlobalContext()));
 909   FunctionType *FT = FunctionType::get(Type::getDoubleTy(getGlobalContext()),
 910                                        Doubles, false);
 911
 912   Function *F = Function::Create(FT, Function::ExternalLinkage, Name, TheModule);
 913
 914   // If F conflicted, there was already something named 'Name'.  If it has a
 915   // body, don't allow redefinition or reextern.
 916   if (F->getName() != Name) {
 917     // Delete the one we just made and get the existing one.
 918     F->eraseFromParent();
 919     F = TheModule->getFunction(Name);
 920
 921     // If F already has a body, reject this.
 922     if (!F->empty()) {
 923       ErrorF("redefinition of function");
 924       return 0;
 925     }
 926
 927     // If F took a different number of args, reject.
 928     if (F->arg_size() != Args.size()) {
 929       ErrorF("redefinition of function with different # args");
 930       return 0;
 931     }
 932   }
 933
 934   // Set names for all arguments.
 935   unsigned Idx = 0;
 936   for (Function::arg_iterator AI = F->arg_begin(); Idx != Args.size();
 937        ++AI, ++Idx)
 938     AI->setName(Args[Idx]);
 939
 940   return F;
 941 }
 942
 943 /// CreateArgumentAllocas - Create an alloca for each argument and register the
 944 /// argument in the symbol table so that references to it will succeed.
 945 void PrototypeAST::CreateArgumentAllocas(Function *F) {
 946   Function::arg_iterator AI = F->arg_begin();
 947   for (unsigned Idx = 0, e = Args.size(); Idx != e; ++Idx, ++AI) {
 948     // Create an alloca for this variable.
 949     AllocaInst *Alloca = CreateEntryBlockAlloca(F, Args[Idx]);
 950
 951     // Store the initial value into the alloca.
 952     Builder.CreateStore(AI, Alloca);
 953
 954     // Add arguments to variable symbol table.
 955     NamedValues[Args[Idx]] = Alloca;
 956   }
 957 }
 958
 959 Function *FunctionAST::Codegen() {
 960   NamedValues.clear();
 961
 962   Function *TheFunction = Proto->Codegen();
 963   if (TheFunction == 0)
 964     return 0;
 965
 966   // If this is an operator, install it.
 967   if (Proto->isBinaryOp())
 968     BinopPrecedence[Proto->getOperatorName()] = Proto->getBinaryPrecedence();
 969
 970   // Create a new basic block to start insertion into.
 971   BasicBlock *BB = BasicBlock::Create(getGlobalContext(), "entry", TheFunction);
 972   Builder.SetInsertPoint(BB);
 973
 974   // Add all arguments to the symbol table and create their allocas.
 975   Proto->CreateArgumentAllocas(TheFunction);
 976
 977   if (Value *RetVal = Body->Codegen()) {
 978     // Finish off the function.
 979     Builder.CreateRet(RetVal);
 980
 981     // Validate the generated code, checking for consistency.
 982     verifyFunction(*TheFunction);
 983
 984     // Optimize the function.
 985     TheFPM->run(*TheFunction);
 986
 987     return TheFunction;
 988   }
 989
 990   // Error reading body, remove function.
 991   TheFunction->eraseFromParent();
 992
 993   if (Proto->isBinaryOp())
 994     BinopPrecedence.erase(Proto->getOperatorName());
 995   return 0;
 996 }
 997
 998 //===----------------------------------------------------------------------===//
 999 // Top-Level parsing and JIT Driver
1000 //===----------------------------------------------------------------------===//
1001
1002 static ExecutionEngine *TheExecutionEngine;
1003
1004 static void HandleDefinition() {
1005   if (FunctionAST *F = ParseDefinition()) {
1006     if (Function *LF = F->Codegen()) {
1007       fprintf(stderr, "Read function definition:");
1008       LF->dump();
1009     }
1010   } else {
1011     // Skip token for error recovery.
1012     getNextToken();
1013   }
1014 }
1015
1016 static void HandleExtern() {
1017   if (PrototypeAST *P = ParseExtern()) {
1018     if (Function *F = P->Codegen()) {
1019       fprintf(stderr, "Read extern: ");
1020       F->dump();
1021     }
1022   } else {
1023     // Skip token for error recovery.
1024     getNextToken();
1025   }
1026 }
1027
1028 static void HandleTopLevelExpression() {
1029   // Evaluate a top-level expression into an anonymous function.
1030   if (FunctionAST *F = ParseTopLevelExpr()) {
1031     if (Function *LF = F->Codegen()) {
1032       // JIT the function, returning a function pointer.
1033       void *FPtr = TheExecutionEngine->getPointerToFunction(LF);
1034
1035       // Cast it to the right type (takes no arguments, returns a double) so we
1036       // can call it as a native function.
1037       double (*FP)() = (double (*)())(intptr_t)FPtr;
1038       fprintf(stderr, "Evaluated to %f\n", FP());
1039     }
1040   } else {
1041     // Skip token for error recovery.
1042     getNextToken();
1043   }
1044 }
1045
1046 /// top ::= definition | external | expression | ';'
1047 static void MainLoop() {
1048   while (1) {
1049     fprintf(stderr, "ready> ");
1050     switch (CurTok) {
1051     case tok_eof:    return;
1052     case ';':        getNextToken(); break;  // ignore top-level semicolons.
1053     case tok_def:    HandleDefinition(); break;
1054     case tok_extern: HandleExtern(); break;
1055     default:         HandleTopLevelExpression(); break;
1056     }
1057   }
1058 }
1059
1060 //===----------------------------------------------------------------------===//
1061 // "Library" functions that can be "extern'd" from user code.
1062 //===----------------------------------------------------------------------===//
1063
1064 /// putchard - putchar that takes a double and returns 0.
1065 extern "C"
1066 double putchard(double X) {
1067   putchar((char)X);
1068   return 0;
1069 }
1070
1071 /// printd - printf that takes a double prints it as "%f\n", returning 0.
1072 extern "C"
1073 double printd(double X) {
1074   printf("%f\n", X);
1075   return 0;
1076 }
1077
1078 //===----------------------------------------------------------------------===//
1079 // Main driver code.
1080 //===----------------------------------------------------------------------===//
1081
1082 int main() {
1083   InitializeNativeTarget();
1084   LLVMContext &Context = getGlobalContext();
1085
1086   // Install standard binary operators.
1087   // 1 is lowest precedence.
1088   BinopPrecedence['='] = 2;
1089   BinopPrecedence['<'] = 10;
1090   BinopPrecedence['+'] = 20;
1091   BinopPrecedence['-'] = 20;
1092   BinopPrecedence['*'] = 40;  // highest.
1093
1094   // Prime the first token.
1095   fprintf(stderr, "ready> ");
1096   getNextToken();
1097
1098   // Make the module, which holds all the code.
1099   TheModule = new Module("my cool jit", Context);
1100
1101   // Create the JIT.  This takes ownership of the module.
1102   std::string ErrStr;
1103   TheExecutionEngine = EngineBuilder(TheModule).setErrorStr(&ErrStr).create();
1104   if (!TheExecutionEngine) {
1105     fprintf(stderr, "Could not create ExecutionEngine: %s\n", ErrStr.c_str());
1106     exit(1);
1107   }
1108
1109   FunctionPassManager OurFPM(TheModule);
1110
1111   // Set up the optimizer pipeline.  Start with registering info about how the
1112   // target lays out data structures.
1113   OurFPM.add(new TargetData(*TheExecutionEngine->getTargetData()));
1114   // Promote allocas to registers.
1115   OurFPM.add(createPromoteMemoryToRegisterPass());
1116   // Do simple "peephole" optimizations and bit-twiddling optzns.
1117   OurFPM.add(createInstructionCombiningPass());
1118   // Reassociate expressions.
1119   OurFPM.add(createReassociatePass());
1120   // Eliminate Common SubExpressions.
1121   OurFPM.add(createGVNPass());
1122   // Simplify the control flow graph (deleting unreachable blocks, etc).
1123   OurFPM.add(createCFGSimplificationPass());
1124
1125   OurFPM.doInitialization();
1126
1127   // Set the global so the code gen can use this.
1128   TheFPM = &OurFPM;
1129
1130   // Run the main "interpreter loop" now.
1131   MainLoop();
1132
1133   TheFPM = 0;
1134
1135   // Print out all of the generated code.
1136   TheModule->dump();
1137
1138   return 0;
1139 }